Research Programme

Engineering
Review of low adhesion research
 R&D Programme:
report commentary

Review of low adhesion research

Research aims the attributes of traction systems with

This work is part of a larger research respect to train performance on low
project titled ‘Understanding low adhesion’. adhesion are proposed. Recommendations
It was developed through a consultation on what could be done to existing traction
process with key industry stakeholders and control equipment and the costs for such
was completed in June 2004. changes are also proposed.

The objective of this work was to provide a Aerodynamic performance of rolling stock.
comprehensive overview of current and Compare the aerodynamic performance of
previous research on low adhesion with existing rolling stock designs and
respect to key areas chosen by determine what reasonable, practicable
stakeholders during the consultation methods can be employed on existing
process. trains to minimise leaf deposition on rails.
Rolling stock design in combination with
Research findings rail mounted leaf deflectors should also be
The research included a review of previous considered.
and past research and state of the art
knowledge on low adhesion management. ‘Whole train’ control systems.
Experts on adhesion and low adhesion Quantification of the potential benefits from
management were interviewed to gain an a closed-loop traction system and a
understanding of the current significant closed-loop braking system to optimise
contributory factors. The results of the vehicle performance during low adhesion
work are contained in an extensive report conditions. Identification of reasonably
including both a detailed commentary on practicable methods of achieving this on
the findings together with abstracts of existing vehicles where the benefits prove
previous research. to be cost-effective. Investigate the
conditioning effect of leading wheelsets on
The research has identified and the railhead during low adhesion.
recommends further work in the following
five areas: Creating and measuring low adhesion for
research and test purposes. The
Characteristics of leaf contamination determination of appropriate techniques for
properties. In particular; what binds leaf creating consistent and measurable low
contamination to the rail and determination adhesion conditions for required test
of what products can break this bond, how applications. Determine appropriate
hard is it, what moisture level is needed to techniques for measuring low adhesion
cause adhesion to be reduced and to what depending on the requirement for
level? These are areas where further measurement, including a review of the
research would be beneficial to suitability of existing measurement
understanding the low adhesion systems.
phenomenon.
Next steps
Influence of traction systems on low This work package has helped define the
adhesion performance. Quantification of future adhesion research that Rail Safety
potential benefits from improved traction and Standards Board (RSSB) will
system performance, and establishment of undertake.

Page 1 of 2
 R&D Programme:
report commentary

The scope of future projects resulting from

this work has been endorsed by the
Adhesion Working Group (AWG).

RSSB will work with AWG and Network

Rail to perform the research recommended
by this work.

Contact
Jim Lupton
Head of Engineering Research
Research and Development Programme
Rail Safety and Standards Board
jim.lupton@rssb.co.uk

Page 2 of 2
 Review of Low Adhesion Research

© Copyright 2004 Rail Safety and Standards Board

This publication may be reproduced free of charge for research, private study
or for internal circulation within an organisation. This is subject to it being
reproduced and referenced accurately and not being used in a misleading
context. The material must be acknowledged as the copyright of Rail Safety
and Standards Board and the title of the publication specified accordingly. For
any other use of the material please apply to RSSB's Head of Research and
Development for permission. Any additional queries can be directed to
research@rssb.co.uk. This publication can be accessed via the RSSB website
www.rssb.co.uk

Report: CRF04002
Issue: 1
Date: 14 May 2004
Author: C R Fulford, MSc(Eng) AMIMechE AMIOSH CMILT
 Report CRF04002
Issue 1
14 May 2004

Executive Summary
Rail Safety & Standards Board (RSSB) has consulted the UK rail industry on suitable areas
of engineering research as part of the continued RSSB Research Programme. Under the
Vehicle Track Interface (VTI) theme, a number of areas were discussed at an industry
workshop in January 2004, including further research into low wheel/rail adhesion.

High on the workshop’s priorities was ‘Understanding Low Adhesion’ and a number of key
questions have emerged through industry consultation which RSSB proposes to answer
through further research. The first part of this research work, the subject of this report, forms
the initial phase of the overall project, and is aimed at reviewing the industry's past research
and current knowledge on the specific areas required to answer the key questions posed.
The search has included an international dimension also.

This report details the results of a knowledge search conducted by the author to determine
what any past research has shown, and hence what the current state of knowledge is, with
respect to key questions posed by the industry workshop. The purpose of the knowledge
search is to ensure that existing, and often forgotten, work on this subject is not overlooked
when conducting further research hence avoid duplicating work already undertaken.

This workstream has included a review of reports already published and held in various
libraries, the results of interviews with key personnel and the results of questionnaires sent
to a range of administrations. It relates only to relevant knowledge associated with the key
questions posed in RSSB Research Project Specification T354.

Conclusions have been drawn from the knowledge search with regard to the current state of
knowledge on each particular subject area, and recommendations are made as to whether,
and broadly in what form, future research should take.

In addition to making conclusions and recommendations on the specific areas identified on

the RSSB Research Proposal Specification, further conclusions and recommendations have
been drawn on the subjects of simulating and measuring low adhesion, as these are key to
progressing research in the area of low adhesion.

It should be noted that some of the research recommended may already be being
undertaken, in part or whole, by other industry parties, including the Adhesion Working
Group (AWG) and Network Rail.

CR Fulford Associates 1
 Contents
1 Introduction ...................................................................................................................... 3

2 Purpose ........................................................................................................................... 3

3 Scope............................................................................................................................... 3
4 Methodology .................................................................................................................... 4

5 Organisations Approached .............................................................................................. 4

6 Low Adhesion Formation ................................................................................................. 6

7 Leaf Contamination Properties And Characteristics ...................................................... 17

8 Friction Modifiers ........................................................................................................... 22

9 Vehicle Design Influences ............................................................................................. 25

10 Whole Train Braking/Acceleration Concept ............................................................... 31

11 Adhesion Prediction & Communication...................................................................... 34

12 Conclusions................................................................................................................ 41

13 Recommendations ..................................................................................................... 52
14 Abbreviations ............................................................................................................. 57

15 References................................................................................................................. 58

Appendix 1 - Research Questionnaire .................................................................................. 64

Appendix 2 – Library Search Results AWG .......................................................................... 67

Appendix 3 - Library Search Results Network Rail ............................................................. 103

Appendix 4 - Library Search Results RSSB........................................................................ 120

Appendix 5 - Library Search Results AEA Technology Rail ............................................... 147

1 Introduction

Rail Safety & Standards Board (RSSB) has consulted the UK rail industry on suitable areas
of engineering research as part of the continued RSSB Research Programme. Under the
Vehicle Track Interface (VTI) theme, a number of areas were discussed at an industry
workshop in January 2004 including further research into low wheel/rail adhesion.

High on the workshop’s priorities was ‘Understanding Low Adhesion’ and a number of key
questions have emerged through industry consultation, which RSSB intends to answer
through further research. This would enable the industry to:

• gain a better understanding of the risks from low adhesion and the associated costs
of poor performance and safety incidents;

• gain a better understanding of the mechanics, duration and disposition of rail

contamination;

• determine whether new techniques may mitigate the effects of low adhesion such as
friction modifiers;

• understand what attributes of vehicle design impact on low adhesion performance

and whether different train braking strategies may assist;

• develop more sophisticated (automated) low adhesion prediction tools.

The first part of the research work, the subject of this report, forms the initial phase of the
overall project (Understanding Low Adhesion) and is aimed at reviewing the industry's past
and current knowledge on the specific areas required to answer the key questions posed.

2 Purpose

This report details the results of a knowledge search conducted by the author to determine
what the current state of knowledge is with respect to key questions posed by the industry
workshop. The purpose of the knowledge search is to ensure that existing, and often
forgotten, work on this subject is not overlooked when conducting further research, hence it
is aiming to avoid duplicating work already undertaken to a satisfactory level of
understanding.

3 Scope

This workstream has included a review of reports already published and available to the
author, as held in various libraries, interviews with key personnel and the issue of
questionnaires to other administrations. It relates only to relevant knowledge associated with
the key questions posed in RSSB Research Project Specification T354 [Ref 1].
4 Methodology

The methodology used in compiling this report has been restricted to a desktop exercise.
The author has approached relevant companies to seek their agreement to participate in the
knowledge search. RSSB’s ‘Key Questions’ were analysed to determine the key attributes
that were required to be understood, and a brief questionnaire (Appendix 1) was prepared
and sent to those companies who agreed to participate in order to direct the search in the
most effective manner, and to other appropriate companies who could not be visited.

Relevant reports made available were reviewed and a brief summary was written for each
extracting the knowledge relevant to the research subject areas. These summaries are
appended to this report (Appendices 2 to 5). From this, an overall summary has been
produced detailing the current knowledge with respect to the issues under consideration.

In addition, certain key personnel were interviewed to elicit more detailed considerations of
work conducted and future areas of work being undertaken by their companies. This also
afforded an opportunity to discuss future research in the areas of interest.

Conclusions have been drawn from the knowledge search with regard to the current state of
knowledge on each particular subject area, and recommendations are made as to whether,
and broadly in what form, future research should take.

5 Organisations Approached

The following organisations were approached and/or visited to review their research report
libraries and/or interview key personnel:

Organisation Key Personnel Method Response

AEA Technology Rail John Tunley/Bill Poole Interview Yes
Library Yes – 27
reports
reviewed
Ove ARUP & Partners Tim Armitage Interview Yes
ATOC & TOCs Rebeka Sellick Questionnaire # Part
AWG AWG members Questionnaire ## Part
Review of AWG Report Yes - 101
library reports
Deutsche Bahn Gunnar Baumann Questionnaire No
Interfleet Technology Kerry Schofield Interview Yes
Irish Rail Fred Johnston Questionnaire No
Japanese Railways Mr Kumagai Questionnaire Yes
Research Institute
 Organisation Key Personnel Method Response
London Underground Malcolm Dobell Questionnaire No
Nederlandse Geert Jan Bazuin Questionnaire No
Spoorwegen
Network Rail Peter Kent Interview Yes
library reviewed Yes – 25
reports
reviewed
Northern Ireland Seamus Scallon Questionnaire No
Railways
ProRail Theo te Brake Questionnaire No
RSSB Cliff Cork, Jim Lupton RSSB Insight intranet of Yes – 251
Archived Reports; reports
RSSB web site of reviewed
Research reports
RSSB safety Intelligence
Centre search
Scientifics Dr Ian McEwen Interview Yes
(formerly of Scientifics)
SEPTA Questionnaire No
TTCI Richard Reiff Questionnaire No
The University of Roger Lewis Questionnaire Yes
Sheffield
Laserthor Malcolm Higgins Questionnaire Yes

Notes:

# ATOC distributed the questionnaire to all TOCs.

## in addition to specific companies approached, AWG members include: Metronet BCV,

Transys Projects, Chiltern Railways, London Lines, Porterbrook Leasing and AMEC Spie
Rail.

The following sections distil the information obtained from the various reports reviewed,
interviews conducted and questionnaires returned. Fuller details of the reports are appended
to this report and should be referred to should more detail be required by the reader, or
ultimately the source material should be consulted further.
6 Low Adhesion Formation

British Rail Research (BRR) conducted a significant amount of research during the 1970s
and early 1980s into the problem of low adhesion as a result of an increase in the scale of
problems being encountered [Refs 2,3,4,5,6,7,8,9,10]. The research commenced with
developing a fundamental understanding of the nature of wheel/rail adhesion and the factors
which influence the lowering of the coefficient of friction. This consisted of a mixture of
theoretical studies, laboratory experiments and on-track data gathering and testing.
Whilst the main thrust of the question posed in the RSSB specification is towards the
properties and characteristics of leaf contamination, it is worth recording also the significant
amount of work undertaken by BRR to determine the properties and characteristics of other
contaminants to be found on the railhead, occurring as a result of natural processes,
activities on or about the lineside, and by activities of the railway's neighbours. It is also
considered that the subject of measuring adhesion values should also be drawn out in this
review, as the techniques applied can influence the readings obtained, possibly leading to
false conclusions.
[4] reports on a study of adhesion on BR in the early 1970s, determining the way in which
adhesion varied on BR. The quantity and variety of railhead contamination was studied,
including specific types of contaminant peculiar to certain individual sites. Generally, oil was
found to be surface active and present in such small quantities that boundary lubrication
conditions applied. The work built up a picture of how adhesion varies naturally on BR, but
did not look specifically at leaf contamination. Adhesion on each rail was found to vary
between about 0.04 and 0.55 and in reasonable dry conditions averages 0.3 (the
measurements were made by taking 10 measurements per day for 50 days). The amount of
solid debris varied widely and was mostly composed of rail wear debris and brake block
debris. Water usually had an immediate effect in reducing adhesion, by an average of 0.08
reduction. Average adhesion on a newly wetted rail was nearer 0.2 than 0.3. Steady rain
gradually washes excess oil from the rails leaving a few molecular layers. Adhesion
measured after rain is usually high. Changes in the humidity of air affects adhesion by a
small amount but the greatest change is due to water deposited as light rain or dew. A
smooth rail condensates when its surface temperature falls below the dew point but on a
rough or debris covered rail, water condensated in the cracks. It was also found that trains
control the spread of contamination and, together with track lubricators, were one of the
more prolific sources of oily fluids. However, the passage of just one long train could also
clean the railhead of any oil contamination. No correlation was found between rail roughness
and adhesion. Low adhesion arises due to the interaction between several factors. No one
cause was identified as the most important but the most widespread effect noted was that of
dew or newly falling rain on rails covered in tiny particles of oil or debris.
The effects of water were further studied [5]. The work identified that humidity and water can
control the friction of clean steel surfaces over a range from 0.6 to 0.15 even when the
surfaces are as free from oils as can be managed in a laboratory. The minimum friction was
seen to be associated with surfaces that are slightly wet, and is caused by changes in
strength and flow characteristics of a surface layer of wear debris mixed with water. On
debris free surfaces in the laboratory, the friction falls to 0.3 when the surfaces are wetted
with water. If small quantities of oil are present, as is invariably the case on rain affected
rails, the friction falls to lower values, which depend on the amount of oil present.
In the presence of a surface layer of debris, high humidity also reduced friction, the
parameter being the humidity expressed relative to the rail temperature. For ambient
humidity, cold rails were seen to absorb more water vapour than the rails warmed by the
sun, and the adhesion was consequently less. The changes in friction were almost
immediate. Large quantities of oil are removed by steady washing, such as by continuous
rain, but enough to form a few molecular layers remains tenaciously bound to the steel
surfaces. There was very strong evidence that the action of water in reducing friction on rails
is not by the mechanism of transporting debris from the rust covered shoulders of the rail to
the running band as had often been suggested.

[16] describes further analysis of data from the adhesion survey of a 280 km route [Ref 10].
It suggests that instead of low adhesion being due to oily contaminants on the railhead and
the effect of water being to enhance the lubrication in some way, water is most often the
prime cause of low adhesion, mixing with solid debris to form surface films with non-
Newtonian flow properties. With a critical water content the films are not readily squeezed
from underneath the wheels yet they provide a low resistance to shear. The major source of
debris is identified as rust. Surface layers reduce the initial slope of the friction
force/creepage curve. Experience with locomotives suggests that friction is increased by
sufficiently severe sliding/rolling action before being reduced by thermal softening under
more extreme slip. Evidence of the influence of debris on maximum slip shows how friction
varies from 0.05 to 0.3 depending only on visible changes in the consistency of surface films
of debris mixed with water.

These critical conditions arise when the quantity of water is minimal. With a more ample
supply of water, as in steady rain, the effective viscosity of the water/solid mixture is so low it
is readily squeezed aside and adhesion stays high as the wheel makes direct contact with
the rail. On a visibly rusty rail the film can initially be thick enough to support the wheel
completely, when extremely low adhesion can result. The concept explains why trains slip
more frequently the rustier the track and why the first train after Sunday's lack of use
experiences low adhesion. On well used rails there is little debris on the running band and
even though wear debris is spread from the rail shoulders the reduction in adhesion is not so
great. The small amounts of water which cause the lowest adhesion are most often supplied
in drizzle or dewy conditions, and persist longest in shade or along tree-lined sites where the
rails remain cool, the ambient humidity remains high, and any water film takes a
considerable time to evaporate.

6.1 Leaves
[7] identifies that the severe problems created by low adhesion during the autumn occur
largely where the route is tree-lined, where a large number of leaves can accumulate on the
track ballast. The turbulence of each passing train whips up the dead leaves which are
carried along in its slipstream swirling around the vehicle underside and getting crushed
under the wheels. It was confirmed by test that it is by this action that the leaves get trapped
under the wheels. The repeated crushing of a large number of leaves creates a hard, glazed,
black leaf film which can completely cover the running band on the railhead. The film of leaf
debris collects a high proportion of iron oxide rail debris forming a featureless black mass.
This material, with its relatively low shear strength, can totally prevent metal-to-metal contact
between wheel and rail, thus reducing adhesion (and forming an electrical insulating barrier).
Although the black film reduces adhesion to around 0.1 level, the small amount of
precipitation (dew, light rain or snow shower) will reduce the level of adhesion to 0.05 or
less. Spot values as low as 0.02 have been measured using the Tribometer Train. These
levels of adhesion should be compared with braking and traction demands of up to 0.09 and
0.2 respectively, from which it is apparent that there can be a substantial adhesion shortfall.
The influence of moisture on adhesion on leaf affected track is significant. As a result, early
morning and early evening trains frequently encounter low adhesion, and a light shower of
rain can have a sudden and dramatic effect on the train service. The leaf material does not
adhere to the rail surface in wet conditions and low adhesion was considered then to be due
rather to some component of the leaf material left on the railhead. This component had not
been identified and it was suggested it is in-situ on the railhead, due to oxidisation and/or
bacterial action.

[7] also analysed the leaf debris. A living leaf contains about 80% by weight of water. The
rest of the leaf is largely cellulose plus a complex mixture of other chemicals including some
highly polymerised fatty acids, a Class of compound well known for their effective lubricating
action. Other compounds include proteins, amino acids, saccharides, neutral fats and resins.
The dead leaf looses most of its water but retains almost all the basic elements of the other
constituents. The flakes of black deposit found on the railhead show some visible leafy
matter. Apart from this, the source of the black substance is not obvious when examine
separately. The main constituents were water and iron and its oxides. The most notable
feature is the large proportion of iron and iron oxides present, analysis indicating that they
provide almost all the non-organic solid matter in the black material. Leaf deposits viewed
under an optical microscope show that fibrous material which is translucent and still
identifiable as being part of a leaf, bears many small, embedded particles. It is presumed
that the carbon mostly originates from the organic molecules of the leaf structure in which it
was chemically combined. The other constituents of these molecules are largely hydrogen
and oxygen, which are lost in the analysis.
[22] reports on railhead contaminant sampling and analysis conducted as a feasibility study
for Merseyrail Electrics in October 1997. Sample swabs were analysed for oil, iron and
bacteria. Material scraped from the railhead was analysed for plant pigments, oil, iron and
cellulose. Sampling included following a number of safety incidents (both SPADs and station
overruns). Simple analytical techniques to determine railhead contamination were achieved
with respect to oil, iron, leaf mulch and bacteria, but further characterisation and
development was required. Bacterial counts proved useful in providing additional information
about the nature of railhead contaminants.
Banverket (Swedish infrastructure controller) conducted extensive investigations and trials
involving low adhesion in 1995 [24]. As part of this work, the Ecological Institute in Lund
analysed many railhead samples for chemical and biological content. In general plant
material was found. A connection between the Nitrogen content and plant material was
made. They reported that after a number of trains had run the plant material had become
charred. There was a possibility that in farming areas the slip problem could be caused by
airborne nitrogen compounds (eg. Ammonia), in which case they considered the track film
could be bacteria or algae. The collected material contained less concentrations than
growing material, possibly explaining a dilute mixture of plant material in a substance which
was otherwise dominated by rust and graphite grease.
Samples of the hard black film formed during laboratory tests of the effects of wheel load
were analysed by BR Research to determine their content [62]. Analytical services carried
out a Fourier Transform Infrared Spectroscopy to identify the organic content of the debris
and X-ray Fluorescence and X-ray Diffraction to identify the organic content. It was found
that all samples were very similar in that they contained cellulosic material, significant
amount of calcium oxalate and ferrous debris. The cellulosic material and calcium oxalate
are indicative of vegetable matter and ferrous material was from wheel and rail wear. Only
very small samples could be collected so only a very rough estimate could be made on the
relative proportions of the debris constituents. In all the samples there was a 60% metal and
a 40% vegetable content. The fact that the debris contact of vegetable and metallic material
may indicate why the hard black layer can form without further additional leaves in that
rolling is required to compact the two components together. The constituents found in the
laboratory debris were very similar to that determined from debris collected from the track.

Ove ARUP also conducted contaminant analysis at the sites forming the ‘deep dive’ and
‘shallow dive’ sites for Network Rail during the Autumn of 2003 [50]. The analysis of many
samples identified the presence of leaves to strongly correlate with the presence of
magnesium, and to be weakly correlated to the presence of calcium, phosphorous and
aluminium. The presence of copper, chromium or nickel was independent of the presence of
leaves.
A reply to the questionnaire by the University of Sheffield suggests other material that may
be relevant [65 and 66]. The University report that they have undertaken twin-disc tests to
study wear and adhesion with a number of different contaminants, and are to embark on
testing leaves as well.

6.2 Other Contaminants

Surface Solid Debris
[6] reports on a study of the physical characteristics of railhead debris particles and their
effect on friction when dispersed thinly and when forming a thick surface film. An important
parameter was the large surface area of the debris. Depending on this area, comparatively
large quantities of oil were necessary to cause low friction. Debris also helps maintain
adhesion against the effects of oil. Thick oil films reduce adhesion. Debris as found on rail
surfaces was Classified into two categories: the thinly disposed particles found on most well-
used lines, and thick flows which completely obscure the running band at a few particular
locations. Dispersed debris improves adhesion on dry rail. The particles provide a large
surface area on which spilled oil can be spread, thus reducing the amount of oil on the rail.
In humid conditions, the strength of the debris particles is reduced and adhesion is worst
affected when the surface density of particles is high. Adhesion on comparatively clean rails
is therefore more influenced by spilled oil, less by high humidity. Water, compared with oil,
covers the surfaces in overwhelmingly large proportions, and probably reduces adhesion
equally on clean or dirty rails. On thick films, adhesion is governed solely by the shear
properties of the debris layer. Humidity has an appreciable effect and water substantially
reduces adhesion. Minimum adhesion occurs when water is just on the point drying out.

[9] develops the concept of low adhesion caused by damp debris further. The physical
properties of railhead contamination were examined with the special emphasis on iron oxide
and water. These properties were used to calculate the limiting coefficients of friction on
heavily contaminated track. Track observations showed that significant debris coverage can
occur on mainline track under certain conditions. It was shown how very low adhesion can
be caused by slightly wet track debris. This acts as a viscous lubricant fully or partially
unloading the wheel from the rail. Under these conditions the Classical boundary lubrication
equation no longer holds. (Author's note: the report notes that coefficients of friction
measured with a lightly loaded portable Tribometer under debris contaminated conditions
cannot be transposed to describe wheel/rail adhesion - this has become a recurring theme in
the author’s search and has been drawn out as an issue in its own right). It identifies that
solid contaminant films reduce adhesion if their shear strength is low: leaves, oil and coal
contamination can form films of low shear strength. The rheological properties (flow and
deformation) of wet rail debris can change rapidly with water concentration, misty conditions
or light rain being necessary before severe loss of adhesion can occur.

[23] describes the 'paste effect' formed by rust (or other solid eg. coal, iron ore, clay) and
water. On thickly rusted rails adhesion coefficients as low as 0.03 have been measured in
drizzle. The proportion of water to solid matter is critical. Very low adhesion occurs as the
paste moves from viscous liquid behaviour to that of a solid (or vice versa). This is more
common on little used track or after a period of little traffic (the Monday morning effect). Also
more common at certain times of the day when humidity effects may give rise to
condensation on the rail. Humidity must be expressed relative to rail temperature and not air
temperature. As well as morning dew, condensation can occur when the rail temperature
falls, e.g. in a cutting, or when the local humidity is high, e.g. a wooded area.

Oils
[3] determined that friction depends both on the nature of the oil and on the quantity applied
to the surface. When varied in quantity, all fluids produce a similar curve in which friction falls
as more fluid is applied, until a constant low value, characteristic of the particular fluid, is
reached. This variation in friction can account for much of the variations in adhesion
observed on rails on dry days. For most lubricants the friction coefficient is about 0.13.
However, the nature of the oil can be substantially changed by oxidisation or shear such
that, in laboratory conditions, friction coefficients as low as 0.03 to 0.08 have been
measured. Water on oily surfaces reduces adhesion to a value almost as low as if excess oil
were present. This is probably the most frequent cause of low adhesion in practice. Debris
provides an extra surface in which oil may spread, so that more oil is needed on a debris
covered surface before the characteristic low friction is attained.

[5] points to large quantities of oil being removed by steady washing, such as by continuous
rain, but enough to form a few molecular layers remains tenaciously bound to the steel
surfaces. Further it identifies very strong evidence that the action of water in reducing friction
on rails is not by the mechanism of transporting debris from the rust covered shoulders of
the rail as is often been suggested. It concluded that humidity and water can control the
friction of steel surfaces over a range from 0.6 to 0.15 even when the surfaces are as free
from oils as can be managed in a laboratory. The minimum friction is associated with
surfaces that are slightly wet, and is caused by changes in strength and flow characteristics
of a surface layer of wear debris mixed with water. An unanswered question is how much
debris is necessary on rails before the lowest friction is encountered?

Measurement Techniques
The early research undertaken by BR Research in the 1970s used a mixture of portable
tribometer measuring and measurements on friction rigs, e.g. Amsler rig, in the laboratory.
Whether this was truly representative of the adhesion seen by a full size rail wheel on a steel
rail was questioned. A Tribometer Train was therefore developed [46] to measure adhesion
levels as seen by a full size wheel, with a typical full scale wheel rail contact patch, seeing
typical wheel loads. Many of the later tests were then conducted using the Tribometer Train.
The last extensive use of the Tribometer Train to measure naturally occurring adhesion was
in the early 1990s when a significant amount measurements were taken over many routes to
survey typical adhesion profiles on a number of BR routes (this data was subsequently used
in the WSPER). The final use of the Tribometer Train was in 1996 to compare the
performance of portable tribometers with a full-scale tribometer to calibrate the former. The
Tribometer Train has not been used since then and is shortly to be disposed of by AEA
Technology Rail (it is life expired).

Much reliance is currently being made on the performance of portable hand-operated small
scale wheeled tribometers, most notably the Salient Systems Inc equipment. Railtrack PLC
purchased a number of these equipments for use proactively for remedial treatment and
reactively in post-incident measurement. However, concern was expressed about their
performance and BR Research undertook comparative testing as explained above [47]. The
portable tribometer measured values of coefficient of friction which were considered to be
generally too high. Poor correlation existed between the mean values obtained with the
portable tribometer and the mean adhesion measured with a full sized wheel/rail contact.
There was also no consistent relationship found between the compared measurements on
different contaminants suggesting that the results of friction measurements taken on one
source of low adhesion could not be reliably compared with those on other contaminants.
The fact that the portable tribometer has a 1/10th scale wheel means that the contact area is
very much smaller than full size. This is more likely to be influenced by wear and surface
roughness effects, large debris and behave differently under lubricating conditions (e.g.
water/rust films, oil/metallic debris etc.). The results suggest that the Tribometer should not
be used for absolute measurement of the coefficient of friction due to poor correlation with
the full-scale wheel results. The results also suggest that using these devices to compare
the performance between different solutions may not be reliable either.

Ove ARUP have used the Salient Systems portable tribometer extensively when undertaking
testing for Network Rail. [48] reports on tests conducted on the Woking line using several
Salient Systems tribometers and different operators in a random sequence of use. The tests
established that the readings were independent of the operator or the tribometer used.
Readings were in excess of 0.5µ for clean rail and 0.25µ on oiled rail, which were higher
than would be expected. ARUP considered that the portable tribometer readings should not
be relied on for values lower than 0.15µ and higher than 0.65µ. Also, the tribometer was
seen to cut through the contamination if repeatedly used in same lateral position across the
railhead. They also noted that the train wheels have a wider contact band than the
tribometer wheel, concluding that the railhead friction experienced by the Tribometer is not
necessarily that experienced by the train.

Banverket used a Salient Systems Inc portable tribometer in their project TASS in 1995 [24].
They report findings that the tribometer was giving inaccurate, high readings when compared
to a vehicle-mounted friction measuring system.
A further report was produced by AEA Technology Rail for Railtrack in 1998 [64] which
compared various adhesion measuring techniques, and will give further insight into this
issue.

How quickly in terms of time and distance does adhesion change?

[4] studied the way in which adhesion varied on BR track. Variations along the track were
identified by measuring adhesion levels along 60ft rail sections using a portable tribometer.
Little variation was seen when adhesion was high (dry rail) or low (damp leaf) over five 60ft
rail lengths. On 2 occasions low adhesion was measured over a longer length through a leaf
affected cutting. Humid conditions in the shade of overbridges also reduced adhesion and
adhesion also varied laterally across the rail. Trains were seen to control the spread of
contamination and, together with track lubricators, were one of the more prolific sources of
oily fluids.
The BR Tribometer train was used to measure wheel/rail adhesion at regular intervals during
one year over a 280 km route commencing in November 1972 [10]. This particular survey
did not include leaf affected track. The overall mean adhesion level recorded during the 12-
month survey was 0.27, ranging from 0.086 to 0.40 (averages on a route not a spot
location). The report details the effects of traffic utilisation and type, the effects of track
curvature and gradient, the diurnal variation of adhesion, annual variation of adhesion, effect
of speed, effect of rain, incidence of low adhesion and effect of tunnels. A clear picture of the
statistics of how adhesion varies on BR track emerged. The survey confirmed that, by and
large, all sections of track are subject to the same variation, the magnitude of which tends to
swamp the large unexplained differences between different track sections. Over the 12-
month period considerable variations in adhesion were observed on all parts of the test
circuit. The work also found that a certain pattern of adhesion persisted for a number of
days. A number of factors were considered to contribute to the variations, and their
interactions made it difficult to ascribe precise values to particular effects. However, some
specific "low spots" were identified, and associated with features such as flange lubricators,
points and crossings, level crossings and cuttings where the local humidity may have been
high. The report highlights the most important conclusion being that low adhesion
encountered over substantial lengths of track is nearly always associated with water on the
rails.

Further extensive surveys on a number of routes were undertaken by BR Research in the

early 1990s which included leaf affected track. That work was used to populate the WSPER
test equipment and provides a representative picture of the autumn adhesion profile [40].

In terms of how quickly adhesion changes with time, [3] measured adhesion levels, surface
contamination and surface roughness over a 48 hour period, taking measurements every 30
minutes using a portable tribometer (humidity and rail temperature were also recorded). This
showed adhesion levels varying abruptly and gradually over the 48-hour period. Short-term
variations in adhesion were believed to be due to the passage of trains. Most trains alter
adhesion by a small amount. Humidity as distinct from rain may also affect adhesion
expressed as temperature difference between the rails and the dew point. The greater this
difference the drier the rails. This led to a general correlation between adhesion and
humidity.

[4] identifies the variation of adhesion with the passage of trains which were thought to
cause the small but abrupt changes in adhesion. Oil contaminated rails were quickly cleaned
by the passage of trains. A long freight train alone, and a passenger express, virtually
cleaned the rail of oil but a single locomotive was not enough. However, the trains carried oil
contamination further down the track. [17] identifies that the passage of an 8 car Class 508
EMU at 50 mph caused as many as 60% of the leaves present in four foot to be deposited
on the rails. [18] identifies that a 4-vehicle 16-axle train was observed to deposit on average
one leaf per yard of rail. The report also identifies that following rail cleaning, the continuous
black rolled film was observed to re-establish itself with the passage of a few trains,
amounting to no more than 200 axles, when suitable drying conditions returned. [49]
determined that as few as 50 axles passing affected a clean rail following high pressure
water jetting. [50] identifies that the action of just one train (16 axles) converted leaves on
the railhead to a black, durable contaminant. Thus, contamination can appear very quickly
depending on local conditions and train frequency.
[50] identified from intensive monitoring of a number of sites that the removal of leaf
contaminant was accomplished without special equipment. The same process of
accumulating contaminant is also wearing away contaminant if no new material is being
drawn in, but observed wear rate is slower than build-up. The wear rate depends on passing
trains, typically 300 axle passes were required for the leaf contaminant to be worn away.

Banverket undertook as similar exercise to BRR in plotting friction levels over time and
distance, creating a similar 3-D map to show how adhesion varies over time and distance
[24]. They identified rapidly alternating friction values as well as progressive reduction and
increase. They found that friction can vary considerably locally, but it does seem to change
in a similar way over a large geographical area. In an area where drizzle is falling friction can
fall to a very low level and it was demonstrated that high humidity could lower friction also.
[36] reports on extended trials of the Low Adhesion Warning System (LAWS) in Autumn
1998 giving a further insight into the incidence of low adhesion. The trial focussed on braking
and analysis was based on over 28,000 braking events. The majority of slip events were of
less than 50m in length but there were a significant number greater than 300m. The average
braking rate achieved was 4.3%g, and there was WSP activity for 0.65% of the total time
spent braking. Of the time with WSP activity, < 3%g was available for 14% of the total time
spent slipping.
[40] identifies from BR Research Tribometer train results (Autumns 1992 and 1993) that the
majority of low adhesion sections measured were less than 200m long and the frequency of
low adhesion sections identified below 0.06 was around one third of those below 0.09. It also
noted that there were no low adhesion lengths longer than 400m where the adhesion was
continuously 0.04 or below.

Metronet Rail BCV Limited’s response to the questionnaire reveals that adhesion profiles
were recorded on the London Underground Metropolitan Line in 1994 and revealed
significant variations in adhesion over very short distances. They also state from experience
of Central Line ATO that the adhesion levels can change over short time periods.

How long does the adhesion last for?

The simple answer to this question, in terms of time, is that low adhesion conditions can last
for a few minutes or a few hours depending on the prevailing conditions and traffic intensity.
[23] provides a useful overview. It states that adhesion is unpredictable and varies widely. It
is strongly influenced by the weather – rain temperature etc. and can fluctuate rapidly from
site to site. Low adhesion depends on there being critical amounts of water present which
may be short lived, the passage of only a few axles being sufficient to disturb the equilibrium.
Rain can affect significant lengths of track, as can dew or other condensation effects. There
are situations where the effect can be very localised, e.g. in short cuttings producing micro-
climates, over-spray from the sea or industrial/agricultural processes (cooling towers, crop
spraying etc.). Ice and snow has the same effect as it is melted by the action of passing
wheels. The report also notes that the effect of leaf extract can be seen many miles away
from the source of the leaves, being carried down the track by the wheels. The leaf related
low adhesion process was also seen to be similar to the 'paste effect' process described for
solid debris.
How long it lasts for in terms of distance is discussed above.

How localised is it?

Low adhesion is very site specific but the length of a site can vary from a very short section,
perhaps in a cutting or shade of an overbridge, or could be over extensive distances. Low
adhesion has been measured many miles away from its source. Specific features are
generally required to create low adhesion (eg. local topography, track geometry,
microclimates etc.), which by definition will be localised. However, the general onset of
drizzle can create low adhesion over very wide areas perhaps after a long dry spell.

[18] states that track geography can play a significant role in determining whether a given
location becomes a leaf site. The position of the trees in relation to the prevailing wind can
determine where the leaves land. Also railway cuttings are more liable to leaf contamination
than level ground or embankments. This is partly because they act as a leaf trap, and partly
because the moisture content of the air tends to be higher, encouraging the formation of dew
and delaying evaporation. The type of tree is also important as both the size of the leaf and
the volume of leaves are critical. Thus several Beech trees are required to produce the same
level of contamination as one Sycamore.
Report [24] identifies a general tendency for adhesion to reduce as the curve radius
decreases but not significant until curves of less than 670m are encountered.
7 Leaf Contamination Properties And Characteristics

What is it about the surface chemistry and physics of leaf contaminant “Teflon-like”
coating, which means that it adheres to the rail yet causes wheels to slide over it?
Despite the significant amount of work undertaken as described in section 6, no specific
evidence has been identified which answers the question of how the leaf film adheres to the
rail. [18] identifies that fallen leaves do not adhere to the rail during wet weather, partly
because the railhead is wet, the partly because the wet leaves adhere to each other on the
ballast and are not lifted into the trains' slipstream. This is backed by the fact that water acts
to remove the bond, softening the hard layer.

Regarding the surface chemistry, [23] reports that the leaf-related low adhesion process
could also be seen to be similar to the 'paste effect' process described for solid debris (see
section 6.2 above). It states that neither the mechanism of lowering adhesion by particulate
solid/water pastes nor that by leaf films can be described in Classical terms. In all probability,
it is reported that the former shares much with elastohydynamic lubrication and the latter
with the full film hydrodynamic lubrication. However, both of these mechanisms can only be
described mathematically assuming defined viscosity properties of the single-phase
lubricant. In the cases of pastes or leaf films on the rail, the lubricant is a solid/liquid mixture
of varying proportions with very complex non-Newtonian flow characteristics.

The usefulness in answering the basic question of how the leaf film adheres to the rail lies in
the possibility of identifying products that would break the bond between the leaf film and the
rail. Water is known to break the bond by softening the leaf film layer but, if used in a
practical way to aid removal of the hard coating of leaf film, such as prior to water jetting,
then it would need to break the bond very quickly. Tests conducted by AEA Technology Rail
have shown that there needs to be a time of between about 15 and 30 seconds to soften the
film sufficiently. This is too long for application from a water jetting train without requiring the
application speed to be significantly reduced.

Also, nothing has been found that describes the hardness of the black rolled-in leaf film and
hence the force required to scrape the film off the rail. The usefulness of this data would be
for assuring that artificial leaf film contamination used for testing, e.g. the AEA Technology
Rail 'paper tape' method or the Qunietiq 'Lignin and grass' method, is a true representation
for real leaf film contamination when assessing the efficacy of various rail cleaning systems.

What are the friction properties related to the leaf contaminant thickness?
[5] posed an unanswered question: how much debris is necessary on rails before the lowest
friction is encountered? This had not been seen to be answered in the series of research or
other reports reviewed. [6] identified that crushed leaf layers are usually about 100 microns
thick although some are twice that, and [12] identifies that the thickness of the leaf films
found during water spray testing varied along the track from between 5 to 10 microns and up
to 90 microns. Higher than average thickness values were obtained where fresh leaves were
on top of a black leaf film. The average thickness of the black leaf film was found to be in the
range 20 to 30 microns.

In discussions with AEA Technology Rail, anecdotal evidence materialises from brake
testing conducted on simulated leaf film using damp, rolled-in paper tape. Similar overall
deceleration rates were achieved over repeated runs on the paper tape, each run gradually
removing some of the simulated leaf film. Very low adhesion was still encountered when
there was little visible paper left, suggesting no linear relationship between contaminant
thickness and adhesion level.

The Ove ARUP monitoring for Network Rail conducted during Autumn 2003 [51 and 70]
found that the contaminant thickness was between 80 and 110 microns following initial
crushing of leaf onto rail (still recognisable as a leaf). As further trains rolled this in and
spread the individual leaf patches, the contamination thickness reduced to between 28 and
35 microns, spread over a considerable length. Fully contaminated rails were not found to
have a thickness greater than 35 microns suggesting a limiting thickness. Damp rails
exhibited a thinner contamination layer than dry rails and the passage of trains was seen to
wear away the contaminant, the rate at which this occurred varied with train length. Their
work also indicates that high pressure water jetting leaves behind a contaminant layer of 10
to 15 microns.

[50] also concluded that railhead friction, contaminant thickness and wheel wear appear to
be completely independent variables, but there is a correlation between contamination
thickness and electrical resistance.

[18] also noted that rails partially covered with leaf, and rails downstream of a leaf site, are
covered with a sticky leaf "juice" which can also exhibit very low adhesion values.

As an aside, care is needed when measuring thin layers of contaminant as the surface
roughness of the rail itself may start to break through the layer and affect the readings. This
may occur at around the 5 to 7 microns level.

Does water soften it such that there is an opportunity to agitate it off the rail?

It is clear from many sources that water acts to soften crushed leaf film allowing the natural
passage of trains to remove the contamination. [8] states that it has been noted in the past
that the combined action of traffic with rain is sufficient to remove all traces of leaf film from
the rail. Continued periods of rain also prevent film build-up since the saturated leaves in the
cess are less likely to be stirred up by traffic.
AEA Technology Rail tests undertaken for Balfour Beatty have shown that simulated leaf film
using rolled-in paper tape is softened by a water spray such that removal can be improved
by high pressure water jetting [57].

The rain affect was explored in a series of tests using artificial 'rain' to remove
contamination. [12] describes tests undertaken in 1977 using a trackside mounted water
spray system (the system is fully described in [13]). A section of trackside water spray
equipment was installed near Bekesbourne station. The purpose of this equipment was to
assess the effectiveness of simulating rainfall, which, when accompanied by the passage of
trains, had been observed to clean the rails most effectively. The spray equipment sprayed
water over 200m of track from jets situated in the middle of the 4 foot, every 18 minutes. The
spray rate was sufficient to ensure that the rails remained almost continuously wet in the
prevailing weather conditions, being just on the point of drying out at the end of each 18
minute period. The rails were inspected for areas of leaf contamination, being compared with
that of the preceding 200m of track, and the thickness of films found on the surface of the
rails was measured. Both the Tribometer Train and a portable tribometer were used to take
adhesion measurements.

It was found that when the trackside water spray was switched on the rails were almost
always completely clear of leaf debris, in marked contrast with the control section. When the
trackside water spray was switched off for a time, contamination was seen to develop, which
was quickly removed by the passage of trains once the system had been reactivated. The
mean adhesion coefficient on the water sprayed section was about 0.15 when the rails were
wet, increasing to about 0.21 as the rails dried out during the 18 minutes cycle (on heavily
contaminated track, in damp conditions, adhesion levels of < 0.05 are common).
Immediately beyond the sprayed section the adhesion was lower still. The rails wetted by the
trackside water spray system appeared at all times to be clean to the eye. Switching off
confirms the principle of trackside water sprays as effective means of maintaining rails free
of visible leaf debris.

Further trackside water spray tests were conducted at Bearsted Bank in 1978 [14], spraying
a 400m section of track. The spray operated for one minute every 15 minutes to keep the
rails continuously wet. The overall degree of leaf contamination was found to be in excess of
that experienced in the previous year that Bekesbourne. The rails were heavily contaminated
as far as the trackside water spray section where a marked change in the appearance of the
rails was seen. The rails in the sprayed section were noticeably cleaner than those in the
preceding control section, being free from gross leaf contamination and presenting a clear
wear band with only a few fresh leaves loosely attached to the surface. The rails immediately
beyond the sprayed section remained of clear appearance. On progressing further up the
Bank the rails became heavily contaminated with leaf debris. The rails in the sprayed section
became no cleaner in heavy rainfall. The efficiency of the system was badly affected by very
strong winds on November 15. This disturbed the spray pattern so that the rails were not
properly wetted. The wind conditions were exceptional.

It was concluded that the trackside water spray maintained considerably cleaner rails than
those which were not treated. The cleaning action was consistent and was judged to be far
more effective than the water cannon trains at their best. However the rails were not
completely clean, leaf contamination still being evident dotted about the wear band in the
small, pitted areas which are not contacted by the wheels. The same contamination pattern
was evident in heavy rainfall, the spray was therefore seen to be successful in simulating
rain. It is not known what effect these patches have on adhesion but it was assumed that at
worst the adhesion was no lower than that of rails cleaned by rain (note no adhesion
measurements were taken).

[18] details a cost benefit analysis undertaken on the leaf fall remedies then current or under
consideration. Trackside water sprays were rejected through their high cost.

[20] acknowledges that vehicle mounted water sprays may have a part to play under certain
emergency conditions such as those following the 1987 hurricane. There the volume of
green leaf on the track allowed the tree sap to be carried over long distances by the train
wheels and cause widespread adhesion problems, which were effectively dealt with by
washing the rails. However vehicle mounted equipment cannot deal with the more 'normal'
black leaf film. Heavy rainfall can disperse the leaf film by softening it to the extent that the
passage of trains and rain subsequently washes it away. Simulating rainfall from trackside
water sprays has been shown to have the same effect. However this is very costly as a
general treatment. Frost can be a problem since the probability of ice on the third rail
increases in electrified areas. This treatment is thus only suitable for particular locations
where there is a regular problem which cannot readily be cured by one of the other available
treatments. This could be because the trees are not on railway land, or because of the
logistics of including the location in a regular treatment train diagram.

[23] concludes that the mechanism by which gross water on the rail influences adhesion is
still not fully understood but the result is known. Water appears to act as a ‘weak lubricant’ in
that it reduces adhesion when it is initially high but not to the same level as an oil-based
lubricant. Water is readily pushed aside by passing wheels and does not form a boundary
layer like oil due to the molecule size. I McEwen’s view is water promotes surface oxidation
of the steel rail such that exceedingly thin oxide films protect the surface from metal-to-metal
contact, thus seeing oxide/oxide friction levels and not steel/steel friction levels.
Proprietary softening agents, such as Natrasolve and Orange Cleanse, are in current use on
Network Rail as a leaf remover and act much more quickly than water alone. These are
detailed further in [28] and [56].

Results from the questionnaire circulated indicate use by some TOCs of a chemical solution
(product not stated) for cleaning leaf debris off of wheel treads but the results appear to be
mixed.
8 Friction Modifiers

Can friction modifiers (other than Sandite) be used to mitigate the effects of low
adhesion? How effective would they be and how would they be used?
Applicators were originally introduced in 1958 to spread adhesion enhancing fluids along the
railhead by the action of the passage of trains. However, the usefulness of these devices for
friction control has long been doubted. During the late 1970s, BRR undertook a number of
reviews and tests using the Tribometer Train to investigate the effect on wheel/rail adhesion
of fluids applied to the railhead [11]. The fluids tested had previously been Classed as
adhesion improvers, either from work with the original trackside applicators, or from
theoretical considerations and laboratory tests. The Tribometer Train tests were designed to
see whether these fluids influence adhesion and to determine the optimum amounts (or
concentrations) needed for raising adhesion. The fluids tested were:

• Syton P - a proprietary "frictionising" agent 30% by weight suspension of colloidal silica

in an acquiescent media, the particle size being in the range 35-40 millimicrons,

• Ludox - a similar product to Syton having a higher silica concentration "40% by weight"
and is claimed to have similar properties,

• Ethyl Caprylate,

• Portec solution - a fluid consisting of 1% ethyl caprylate in a 50:50 water:Iso-propyl

alcohol mixture,

• Tertiary Butylamine Solution - a deflocculant which can markedly change viscosity

properties to a paste,

• Sodium Metasillcate Solution.

The fluids tested in this work either reduced adhesion considerably when freshly sprayed or
at best had little influence on the initial values. Only the Syton and Ludox materials produced
a noticeable improvement, and then only after drying out on the rail. Syton also lowered
adhesion when freshly applied but not to the same extent as with ethyl caprylate, in fact no
lower than with water alone. On drying out, Syton gave high values of adhesion (in some
cases greater than 0.5). When measurements were carried out subsequently in the best
conditions the adhesion dropped to a similar value to that on the control section. Although
this suggests that Syton would not be of use in wet weather, it should be noted that the
adhesion was still an acceptable value.
A further report was compiled by BRR in 1980 to provide a synthesis of historical tests [15].
The report details tests undertaken at Derby in 1969, between Barton and Walton in 1972, at
Wood Green in 1972, Woodhouse in 1972, on the Sheffield circuit in 1974, Beattock in 1973,
Derby Deadman's Lane in 1975, Bradford in 1975, Hemmerdon Bank in 1975, Buildwas
power station in June 1976, Taunton in March 1977 and in France between 1970 and 1977.
The evidence concerning the efficacy of the applicators was drawn from three areas: service
experience in combating specific problems, track trials in which success has been judged by
tribometer results, and the trials with fluids applied from the Tribometer Train. The evidence
obtained from the Tribometer Train suggested that Portec solution and ehtyl caprylate
solutions both cause low adhesion when freshly deposited and provided grounds for a
positive recommendation that neither of these fluids should be employed. There remained
the suspicion that the dispensing of Colloidal Silica or Sodium Metasillcate may be
beneficial, but the evidence as that time was so sparse it was not possible to justify any use.
In summary there was still no clear-cut evidence that fluid applicators were effective.

[24] identifies the possible use of lime on the track to improve adhesion, first suggested by
Drivers who experienced improved stopping where lime had come onto the track. The report
identifies successful tests in Denmark using lime.
[27] details a commercial product, Centrac HPF, which was aimed at reducing high adhesion
to reduce wheel wear and railhead wear, and to reduce the associated noise. HPF was said
to exhibit a positive friction characteristic, being designed to provide operation in a controlled
friction range of 0.17 to 0.35 from an uncontrolled range of 0.1 to 0.6. The report also details
brake tests conducted by Skytrain Vancouver using Centrac Sticklube HPF supplied by
Century Oils Limited. The tests concluded that there was no difference in the emergency
brake stopping distance between a dry rail, a wet rail and an HPF friction modified rail.

Trials with Centrac HPF friction modifier were undertaken by BRR in 1993 as part of a wider
review on the future of combating leaf fall problems [63]. [45] summarises the findings of the
tests and concluded that the claimed results were not found under typical autumn conditions.
The results showed little change in the average shape of the adhesion/slip curve and under
most test conditions a reduction in adhesion was seen throughout the slip range. Also, the
rate of HPF block wear was such that it was expected to need replacement after just 1700
miles, suggesting high labour costs for replacement outwith normal maintenance cycle,
despite Century Oils claiming 5650 to 8080 miles in service use elsewhere.
[28] reviewed the state of current knowledge about a number of potential friction improvers in
2003. This was in effect an update on previous knowledge gained by IJ McEwen based on
his earlier work in BRR. The review looked for advancements in the most promising fluids
previously tested: Colloidal Silica Fluids, Friction Modifiers, Sandite and derivatives, Kelsan
HPF Sticks, Keltrack Fluids and TracGlide Sticks. It was found that many of the products
were still available but no particular advances had taken place with the technology. No new
materials were identified and proprietary products were already on offer for railway use are a
friction/adhesion maintainer and are normally targeted at controlling high friction. The report
finds little published information regarding the effect of Kelsan HPF Sticks on contaminated
rails. In long term use it was still considered that they may override the influence of
contamination. Keltrack Fluid, a fluid development of HPF friction modifier, had been
formulated for vehicle and track application onto the railhead. For best effect it needed a
water carrier to dry out. As for Kelsan HPF, this product had been targeted as a high friction
controller rather than low adhesion improver. The TracGlide System is a vehicle mounted
liquid lubricant which does not affect braking. It is laid as a very thin film which degrades
leaving little residue. Once again was aimed at reducing lateral forces and saving energy for
situations of high friction rather than increasing low adhesion.

It is also thought that London Underground Limited had looked at using friction modifiers on
the railhead but abandoned their use due to concerns over reliable train detection and track
circuit operation.

Network Rail also studied friction modifiers (Keltrack friction modifier and Kelsan friction
enhancer) as part of their full-scale adhesion tests conducted at Bicester in March/April 2003
[69]. Keltrack is intended to create a coefficient of friction in the range 0.3 to 0.35 on the
running surface, and is currently used to reduce noise and vibration, to control vehicle
curving and to reduce lateral loads. The product can be applied using trackside applicators,
vehicle mounted and mobile systems. ARUP tested the Keltrack as an alternative to Sandite,
applied to the railhead by trackside applicators in small quantities. The transfer mechanism
was from the track to the wheels, and then from the wheels to the untreated section of track
one wheel revolution later. The results with an application on a clean dry rail showed
Keltrack maintained reasonably good braking and train detection performance. Excess of
Keltrack acted as a lubricant lowering friction levels. Very small quantities must be used to
ensure a film that modifies and maintains the coefficient of friction. They also concluded that
the results of the test runs may have been influenced by the test procedures adopted.

Kelsan VHFP (Very High Friction Product) was applied from a roller and brushed onto
contaminated track, again as if it were to be used as a Sandite replacement. On a dry
contaminated rail, Kelsan VHFP showed only a very small improvement to braking
performance and slightly improved axle detection compared to the contaminated track. It
was seen that Kelsan VHFP does not condition the rail such that contaminated track returns
to the same braking and detection performance as a clean rail. The results were not
successful with application of the film maintaining a very low braking rate.

A reply to the questionnaire from the University of Sheffield identifies a few references to the
use of friction modifiers [67 and 68] although it is not known whether these relate to reducing
high friction.
9 Vehicle Design Influences

What are the critical factors of vehicle characteristics that improve or worsen
adhesion performance (e.g. suspension, wheels, traction and braking system
design)?

Vehicle suspension and bogies/wheelsets

As part of the extensive research carried out in the 1970s and 1980s by BR Research, some
work explored the effect of rail and wheel profiles and wheel loads with respect to low
adhesion, but there is little evidence that much work has been conducted in the UK to
specifically characterise the attributes of vehicle suspension and how it interacts or
influences low adhesion performance.
It has been known for many years that a good riding vehicle with good wheelset steering
characteristics, will lead to little hunting of the bogie, and hence reduce the scrubbing
(cleaning) action on the railhead. [30] reports on work undertaken during 1991 and 1992 to
address the significant increase in WSTCF involving Class 158 DMU. The poor WSTCF
performance of this Class was attributed to the good vehicle ride qualities coupled with disc
brakes. Although the report does not directly detail the interaction between wheel and rail
with respect to failure to activate track circuits, it details the results of experimental fitment of
Auxiliary Tread Brakes (ATB) to a number of 3 car Class 158 DMUs. The results
demonstrated an approximate 50% reduction in incidence of Class 158 WSTCF for trailer
bogie fitted ATB only. A motor bogie modification was expected to add a further 5 to 10%
reduction. Costs were seen as high for the limited effectiveness and hence very little use has
ever been made by such devices. It is also known that ATB offers little improvement in low
adhesion braking performance thus reducing perceived benefits.
[62] reports on the effects of wheel load on the persistence of leaf film. The laboratory study
concluded that: with a 5 ton load it required up to 300 wheel passes to remove the leaf film
and give good electrical contact throughout the length of the contaminated rail, and the leaf
film was persistent under low creepage conditions; with a 10 ton load, a complicated result
occurred with an insulated layer being formed after leaves had ceased to be applied to the
railhead. This was considered to be partly due to the way in which leaves combine with
metallic wear debris to form the hard black insulated film under high load conditions. Hence
the passage of a freight train between low wheel-load passenger trains could exacerbate the
leaf film problem. Under low creepage conditions and a 10 ton load, about 100 passes of the
wheel was sufficient to break the film up. A doubling of load therefore remove the film in
about a third of the wheel passes.
An ORE report [25] documents a synthesis of current knowledge concerning adhesion and
includes work undertaken around the world, including on the influence of various vehicle
design parameters on braking. Adhesion is known to reduce with speed. The report states
that the evidence suggests that the reduction in adhesion with speed is largely due to the
vehicle suspension and to dynamic interactions between wheel and rail. It also states that it
has consistently been found that adhesion is reduced at higher axle loads, doubling the axle
load reduces adhesion by about 0.05. No reports were found on the effects of the body mass
or bogie mass, or on the effects of the rate of deceleration of the vehicle itself. As an
increase in axle load leads to a reduction in adhesion, it was concluded that a change in
wheel profile, or railhead profile, leads to a reduction in the contact ellipse dimensions and
hence increase in load. SNCF tests showed a reduction in adhesion from 0.066 to 0.054 just
by using a different wheel profile. It was also considered likely that the dimensions of the
contact ellipse change as a function of the hunting of the wheelset. The report found little
effect of wheel diameter on adhesion.
Metronet Rail BCV Limited’s response to the questionnaire states that the Central Line stock
introduced in 1992 had a significant wheel lock-up problem believed to be associated with
weight balance and weight transfer. This was corrected by fitting per-axle optimised WSP.

Traction and braking control

Also the work described in section 6 included understanding the effects of friction, wheel
creep and slip, and the effects of curving on traction, but this has been from the perspective
of good rail adhesion conditions and not to study the effects of low adhesion, i.e.
contaminants on the traction forces. Traction control systems can obviously have a
significant impact on performance under low adhesion conditions. [21] reports on trials with a
Class 47 locomotive modified by Brush Electrical Machines to give separately excited
(SEPEX) traction motor fields undertaken in November 1985. This work showed to be
beneficial in reducing the likelihood of gross wheelslip in particular when at least one axle is
not in wheelslip. [29] reviews the wheelslip control systems or various first generation sliding
door EMUs to determine the feasibility of improving the traction control element of these
systems. Some systems are very basic and interrupt traction when slip is detected.
Measures were proposed to improve the degree of control without interrupting power,
providing a more appropriate control under low adhesion conditions. Modern 3 phase ac
traction drives are known to exhibit improved performance under low adhesion conditions
due to the characteristics of ac motors.
Other aspects of the work described in section 6 looked extensively at the effects of
contamination on braking performance, with a great deal of work done with respect to
braking on low adhesion. Much of this related to the performance of WSP systems and rail
conditioning effect by operating with wheels in slip [31]. This included the development of the
WSP Evaluation Rig (WSPER), an industry best practice device, which utilises actual
recorded adhesion profiles to test and optimise WSP systems based on 'real' track
conditions. It is considered that the current knowledge on low adhesion with respect to WSP
equipment design is state-of-the-art.

Aerodynamics
Vehicle aerodynamics has a significant effect on the formation of leaf film. As detailed in
section 6, it is the turbulence of each passing train that whips up fallen leaves, carrying them
along in the train’s slipstream swirling around the vehicle underside and getting crushed
under the wheels. Attempts have been made in the past to influence this effect both by
trainborne and trackside solutions. [17] reports on tests carried out to investigate how
differences in rolling stock affected the rate of deposition of leaves on the railhead. The tests
were prompted by an increase in low adhesion problems found with Class 507 and 508 EMU
when introduced. The tests confirmed that Class 508 units deposit more leaves than Class
405s they replaced. Under the same conditions a Class 508 unit trapped more than twice as
many leaves on the rails as the older Class 405. The work identified that the length of the
train formation affects the number of leaves deposited on the railhead, but the number is not
proportional to train length, i.e. an 8-car unit does not deposit twice as many leaves as a 4-
car unit. Train speed is also a major factor affecting the number of leaves trapped by passing
trains.
As a result of the large number of track circuit operation failures in the autumn of 1991,
particularly by the Class 158, full-scale tests were carried out on the Old Dalby test track
[58]. Classes 156, 158 and 165 2 car DMUs, with different front-end configurations, were
tested, passing over sheets of paper laid on the track bed to represent leaves. Three colours
of paper were placed in four foot, six-foot and the cess so that after the train had passed the
numbers of leaves left, moved to and from each region and squashed on the railhead could
be counted. The tests showed that the number of leaves left squashed on the railhead after
a 2 car sprinter type DMU had passed over was highly dependent on both the presence and
type of air dam fitted to the front of the train. A snowplough fitted Class 156 squashed 3.5
times as many leaves to the railhead compared to the same unit without any such device.
The Class 158 fitted with standard fibreglass air dam squashes slightly less numbers of
leaves on the railhead than the snowplough fitted Class 156, but still a factor of 3 times
greater than the Class 156 without snowploughs. Removing the fibreglass air dam from the
Class 158 but leaving the substantial metal structure behind, which visually provides nearly
as much blockage, increased the number of leaves squashed on the railhead to the same as
the snowplough fitted Class 156. The newly designed integral snowplough type air dam fitted
to the Class 158 resulted in an even greater number of leaves deposited on the railhead.
This was worse by factor of 1.7 on the number deposited by the standard air dam fitted
Class 158 and hence a factor of 4 times worse than the Class 156 without snowploughs. The
Class 165 fitted with its standard air dam, which is significantly different to that of the Class
158 unit, only squashes a similar number of leaves to the railhead as the Class 156 without
snowploughs.
[59] reports on analysis of videos taken of the simulated leaf movement tests carried out on
the Old Dalby test track during 1992. It reveals that the leaf movement varies considerably
between different Classes and configurations. The variation was much larger than expected.
In particular the leaf movement that causes the Class 158 with obstacle deflector, all the
integral snowploughs, and the Class 156 with snowplough fitted, to squashed many more
leaves on the railhead compared to other units or configurations was identified. Leaves in
the cess are initially pushed out as the front of the train and first bogie passes and are drawn
back to the railhead as the open region of the latter half of the first vehicle passes. Two
strong small vortices in nip of the railhead caused the leaves in 4 foot and in the cess to be
captured and held in this region resulting in a number of them being run over by the vehicle
wheels. The Class 156 in the open configuration and Class 165 with its standard air dam
fitted, both produce completely different leaf movement from each other and from the Class
158 or 156 with snowplough, such that for the Class 165 and Class 156 in the open
configuration, few leaves are transferred across the railhead. A very limited understanding of
the leaf movement caused by the front air dam was developed: the more bluff the air
dam/obstacle deflector the stronger are the vortices created by it and hence the trapping of
more leaves; the actual size, position, strength and direction of rotation of the vortices
created underneath and to the size of the train are a consequence of the geometry of the air
dam.

BR Research went on to develop and test am experimental bogie-mounted leaf deflector

[60]. The report details tests conducted on a Class 165 fitted with the bogie-mounted leaf
defectors over one day. The conclusions of this report were that leaf defectors had been
demonstrated to be effective in reducing the amount of leaf coverage on the railhead. The
improvement obtained for real, large sycamore leaves was about 70 percent. The majority of
leaves rolled on the railhead originate from the four foot and leaf counts showed unequal
numbers of leaves crossed both rails from the four foot. Leaves crossing the railhead from
the cess side were not propelled into the wheel rail nip. The deflectors were left in situ for a
period of 6 weeks in service but became damaged due to design issues.

Laserthor approached the industry in 2002 with a proposal to develop further the concept of
vehicle mounted leaf deflectors. This followed anecdotal evidence from their trials with a
laser equipped vehicle. Laserthor tried out a deflector by way of fairings mounted on the
bogie which offered aerodynamic flow to reduce leaves in the wheel/rail nip rather than the
‘mechanical leaf blocker’ tested on the Class 165 by BR Research. Their tests began in a
wind tunnel and then progressed to live trials using Mark II coaches. Their results claim an
80% reduction in the number of simulated leaves being trapped in the wheel/rail nip, and
Laserthor suggested this could rise to 90% with further development.
A rail mounted application of a leaf deflector was developed by Vortok International, the
Vortok Leafguard, and tried out by Railtrack and Dutch railways. There was anecdotal
evidence from a number of sources that the Vortok Leafguard reduced the incidence of train
delays and WSTCFs. Vortok International reported on this following the Autumn of 2000 [43].
The report highlights some of the installation problems but claims (from Railtrack sources)
that Leafguard at Maxwell Park saw a 100% reduction in train delays; a total elimination in
train delays and WSTCF at Ingastone; no train delays or WSTCF at Ashford. Site visits to
Wylde Green on Midlands Zone showed a build up of leaf residue on the non-Leafguard
fitted up line, whilst the down line, fitted with Leafguard, remained clean. There were no
reports of train delays associated with the down line (the report does not state whether any
delays were associated with the up line either). Railhead samples were collected at
Manchester using tape which gave a visual representation of the state of the rail. The
Leafguard fitted rail appeared to be visibly cleaner.
A test programme was funded by AWG and was designed to investigate the use of Vortok
Leafguards to prevent leaves from being rolled onto the railhead by passing trains [44].
Tests were undertaken between Romsey and Eastleigh and showed that the Leafguard
approximately halved the quantity of leaf that is rolled onto the railhead. Although there were
insufficient test runs to provide data for statistical analysis, the observed trends were
consistent and supported the evidence from service trials during the previous autumn. The
majority of leaves from the initial position – an area of 4 sleeper bays – were lifted by the
train and became airborne over a distance of between 50m and 100m with only a few carried
for greater distances at speeds of 50mph or 60mph. Since the majority of large leaves that
were deposited on the railhead originated from the 4 foot, it was concluded that there may
be design modifications based on aerodynamic factors which would improve the Leafguard
performance.
However, there were many installation issues and it was a requirement that Leafguard be
removed from service after the autumn and refitted prior to the next autumn. This was very
labour intensive and costly. Leafguard was finally withdrawn from service on Railtrack due to
a length of the Leafguard becoming dislodged and destroyed under the wheels of a passing
train. Nevertherless, the principle, like leaf deflectors on trains, was shown to reduce leaf
deposition on the railhead. The main issue rested with the engineering of the product rather
than the concept. Experience had not provided detailed information regarding the braking
and traction performance benefits but the indications from the reduction in railhead
contamination indicates that benefits should accrue from the use of the leaf deflectors.
10 Whole Train Braking/Acceleration Concept

Are there benefits in varying the braking and traction demand along a train to
maximise the use of available local adhesion and thus take advantage of the
scrubbing action from the front wheels?

Whole train closed-loop traction control

No evidence has been found of any research that has looked at a ‘whole train traction
control’ concept. Most modern traction control systems regulate traction to the available
adhesion to one degree of sophistication or another. This is normally on a ‘per bogie’ or ‘per
vehicle’ basis. A ‘whole train traction’ concept would utilise the traction control system to
attempt to deliver the Driver’s traction demand selected by compensating for bogies or
vehicles subject to traction being ramped back due to low adhesion, by automatically
increasing traction on other bogies able to sustain a higher rate.

Whole train closed-loop brake control

Work has been undertaken by BR Research to look at a ‘whole train braking’ concept. [53]
describes a Train Brake Control (TBC) system which utilises a controller to monitor the
achieved train deceleration and attempts to meet the Driver's deceleration demand under all
adhesion conditions by modulating the train brake demand. The system would rely on the
existing WSPs on each vehicle to control the brake pressure of individual axles to avoid
excessive wheel slide or lockup. Tests were conducted on a modified WSPER to compare
the stopping performance of trains fitted with various braking systems: no WSP, state-of-the-
art WSP, one-shot sander, train braking control system, and idealised WSP. Conclusions
drawn from the test results were: the brake system with the longest stopping distances was
the no-WSP system; the one-shot sander achieved the shortest stopping distances. The
state-of-the-art WSPs performed very well, only allowing the wheelsets to lock up on
continuous lengths of very low adhesion. However, the train with idealised WSPs achieved a
mean stopping distance 12% shorter than that of the train with state-of-the-art WSPs. The
TBC system did not perform as well as expected, mainly, it was thought, because the
controller increases the brake demand above 9%g, and the state-of-the-art WSPs used are
only optimised for a maximum demand of 9%g. Even with WSPs optimised for a higher
brake demand, a TBC system was only considered to perform significantly better than a train
without a controller under a limited range of adhesion conditions. It was not considered worth
developing the TBC concept further solely to obtain improved stopping performance.
AEA Technology Rail later conducted tests on a Class 390 Pendolino unit to determine the
conditioning effect along the 9 car train length as part of an assessment of the need for
trainborne sanders. The work (report only available with permission from Alstom) involved
reducing adhesion levels to very extreme levels (as low as 0.01 was seen). The performance
of the WSP regulated brake system was monitored at the front, middle and rear of the train
thus measuring, with high confidence, the actual adhesion levels on the rail as seen along
the train length. The effect seen was an increase in adhesion levels along the length of the
train, as expected. The performance improvement, however, was not sufficient to raise
levels to that required for normal braking, but the tests were on extremely low adhesion
conditions. These tests confirmed that the WSP system is capable of regulating the brake
performance along the train to match the available adhesion. However, the available brake
demand was constrained by the maximum brake demand levels possible on individual
vehicle. The TBC concept was to raise individual vehicle brake demands above the ‘train’
maximum on certain vehicles to compensate for other vehicles unable to achieve the
nominal vehicle maximum (or that demanded).

Thus from the above Pendolino tests it can be concluded that a modern WSP can regulate
individual vehicle brakes to match available adhesion levels, and it can be inferred from this
that short trains on very low adhesion (trains of 4 vehicles or less most often involved in low
adhesion incidents) will never achieve normal levels of adhesion to support normal braking
through the concept of TBS. However, the extremely low adhesion conditions simulated in
these tests are not representative of the expected norm in the autumn. [36] analysed
deceleration rates from over 28,000 brake events. The average braking rate achieved was
4.3%g, and there was WSP activity for 0.65% of the total time spent braking. Of the time with
WSP activity, <3%g was available for only 14% of the total time spent slipping. Therefore,
the greatest opportunity would appear perhaps to be in reducing train delays where
conditions are poor but not extreme.

Interfleet Technology have advised of the Westinghouse EP2000 brake control system
which has a closed-loop brake control on a per vehicle basis but not on a per train basis.
However, Westinghouse doubted that there would be any more than a marginal gain in
performance to be achieved. However, they had not considered the potential benefit of
increasing brake demand at the rear of the train.

Further communication with AEA Technology Rail on this subject [55] elicits a potential way
forward to establish if a TBC utilising deceleration feedback would be beneficial to operating
performance in low adhesion conditions. The concept includes adding the trainborne sander
into the system. The operation of the TBC would be as follows: when the Driver's demanded
deceleration is not achieved the train brake demand would be increased to the current
design maximum in the hope that the rear of the train could support that higher brake
demand and increase the overall deceleration. The WSP would modulate the brake pressure
down the train to that sustainable at each wheel/rail interface, thus negating any further need
for intelligent control of brake demand down the train. If an insufficient increase in overall
train deceleration was achieved through the feedback loop, an optimum amount of sand
would be deployed to compensate for the adhesion deficit.

Metronet Rail BCV Limited’s response to the questionnaire identifies that some metro cross
brake blending systems apply an adhesion limit that can be varied depending on the location
of the vehicle within the formation. This then limits the extent to which any motored vehicle
can be over-braked to back off the friction brake on an unpowered or trailer vehicle.

Metronet Rail BCV Limited also identify that the Central Line ATO system achieves the
whole train braking concept by increasing the brake rate if the specified rate is not achieved
due to low adhesion. In this case the WSP along the train controls the braking to the
adhesion level that can be sustained. Therefore, the braking rate may well increase along
the length of the train.
11 Adhesion Prediction & Communication

Can current manual adhesion prediction models (e.g. Central Line ACCAT) be
automated to reliably predict the occurrence of low adhesion in real-time to minimise
disruption and if so how?

What would such a model look like and how should it be developed?

How could this be communicated to the train driver?

These questions are answered by way of charting the development of various systems of
low adhesion prediction.

The Low Adhesion Warning System (LAWS)

The Adhesion Working Group (AWG) has been interested in developing an adhesion
prediction system for some time now. In 1999, AWG developed a specification for a package
of measures which together constituted a system for predicting, detecting and advising the
appropriate people of the existence, or likely existence, of low rail adhesion conditions –
termed System for Advanced Warning of Low Adhesion. Using this information, warnings
would ultimately be provided to train Drivers to enable them to modify their driving technique
and so reduce the risk of an adhesion-related incident. However, this had to be appropriately
localised so as not to impact unnecessarily on train performance. Advice would also be
provided to the infrastructure controller and maintainer to enable corrective actions to be
taken and to optimise remedial programmes. Early work was undertaken by BR Research
under AWG funding to develop further the BRR Low Adhesion Warning System (LAWS) as a
start to this package. LAWS is described in [35].
Also adding to this was the systematic approach to categorising the lineside with respect to
vegetation type and density, pioneered by BR Research in the early 1990s [32]. This
process permits problem areas to be identified and categorised 1 to 5 in increasing levels of
low adhesion risk.

Other incidental (unrelated) work undertaken around the same time which points to the
ability to automatically gather salient data includes an ‘unattended data capture system’ as
described in [33]. Here, a monitoring and logging system was used to capture data on low
adhesion events on a Class 159 unit operating on South West Trains services. The data
captured included the braking performance under low adhesion conditions, the location of
low adhesion blackspots and all brake applications over 10 seconds duration where WSP
activity occurred. The data captured was manually retrieved from the train.

[35] details how the LAWS system performed under trial during Autumn/Winter 1997/98
(fitted to a single unit) and suggested a number of ways of improving the system which were
later adopted. The report details a methodology for interpreting and presenting data.
[36] reports on the findings of extended trial of LAWS in Autumn 1998. Trial involved 4 Class
165/166 DMU operating in October to December. Data was gathered from 2 Weatherchex
installations also. The trial focussed on braking and safety related issues and analysis was
based on over 28,000 braking events. The potential of LAWS to assist in improved targeting
of track treatment was demonstrated and the system showed potential to provide a similar
level of insight into performance related issues. The main findings were that braking sites
were clearly identified throughout the Thames Trains’ operating area and were almost all in
the vicinity of stations and junctions. Also low adhesion sites were identified and compared
with the Sandite/water jetting schedule for the North Downs route. It was concluded that a
reasonable level of coverage was provided by the 4 LAWS equipped units and should be
sufficient for making weekly revisions to track treatment schedules. Data from LAWS had
demonstrated to provide a clear indication of where track treatment is required.

[37] reports on an extended trial of the LAWS during Autumn 1998, extended into April 1999,
which was undertaken to evaluate the potential of trackside railhead moisture point and
temperature sensors and on-train humidity and temperature sensors. The railhead
temperature and moisture sensors were connected to a permanent weather station and on-
train sensors fitted to a Class 165 DMU. Data was gathered during April to correlate
readings between the on-train and trackside sensors. The most significant findings were that
railhead moisture as measured by the sensor was closely related to ambient humidity levels,
but there were a significant number of periods when ambient levels were high but measured
railhead moisture levels were low. It was possible to make accurate temperature and
humidity measurements from a moving vehicle (with a correction factor).
[38] describes the LAWS trial conducted in Autumn 1999. The work concentrated on
establishing the potential of on-train temperature and humidity measurement to determine
slip risk through improved understanding of the humidity/temperature conditions which
create low wheel rail adhesion. The main elements of analysis were to relate slip and spin
events on the train to measurements at the weather stations; using weather station
measurements, establish correction factors for on-train sensors; and relate occurrence of
slip and spin events recorded by LAWS over the Class 165 operating area to measurements
from the on-train sensors. Data was gathered for one month (November 1999). The main
findings were that on-train and trackside sensors indicate slip and spin events only when
recorded ambient humidity exceeded 80%. However, the ambient humidity levels remained
above this level for 89% of the readings taken by the trackside weather stations and 85% on
the on-train readings. The predicted formation of dew according to the dew point calculation
method did not provide a reliable indication of slip risk according to either on-train or
trackside sensors. There was very good correlation between occurrence of significant
spin/slip events and readings from a railhead moisture sensor. In contrast to the measured
ambient humidity, the rail moisture sensor was high for only a limited time during the autumn
period (38% of readings taken). It was concluded that monitoring humidity levels alone is of
limited value in predicting when trains are likely to slip and that the calculated dew point
method was not sufficiently reliable for this purpose. However, the railhead moisture sensor
was considered to offer a reliable and useful indication of the prevailing slip risk. Developing
an in-cab slip risk indicator, monitoring ambient humidity alone, could provide a reliable
indication of possible high slip risk conditions but it would report high for most of the autumn
even where slip risk was in fact relatively low. The railhead moisture sensor was not suited
to on-train measurements but was considered a reliable indication of actual slip risk if a
series of such sensors were mounted trackside along problematic routes and transmit
warnings to drivers by means of trackside warning signals or cab radio.

[38] also suggests that it should be possible to provide a live indication of the prevailing slip
risk to a Driver by monitoring ambient humidity levels using on-train humidity sensors.
Further work was recommended to determine the best position for the sensors and whether
prediction of dew point from the measurement of both temperature and ambient humidity
could provide a more accurate indication of prevailing slip risk than the measurement of
ambient humidity alone.

Therefore, AEA Technology Rail installed a number of moisture sensors around the London
area to monitor railhead moisture levels during the 2002 leaf fall season [39]. It was
expected that this data could be used to establish a link between the level of moisture on the
railhead and observed low adhesion conditions. Low adhesion events recorded by LAWS
equipped vehicles occurring within a set distance of the moisture sensors were compared
with the moisture level recorded by the sensing station. The amount of data available for
comparison was limited for a number of reasons therefore the experiment was unsuccessful
in establishing a verifiable link between moisture sensor data and low adhesion conditions.

As part of the LAWS development, [34] explored the performance of the digital Global
System for Mobile communications (GSM) system which was assessed in terms of reliability,
speed and data integrity. The conclusions were that the GSM communications was of
sufficient speed and reliability and integrity. The trial also demonstrated that SMS should be
fast and reliable enough to provide an in-cab live warning system.

Continuing with this theme, AWG commissioned a study [35] to investigate how adhesion-
related information could be communicated to train Drivers. The general findings of this
study were that it was technically feasible to construct a trackside warning system, based
either on in-cab indication or trackside signage. There is a lot of existing technology that
could be employed toward this end in order to keep development costs down, whilst
ensuring reliable operation. There remained a number of human factors and failure mode
issues that required further investigation and/or consideration.
The Dutch rail operator (NS) uses AEA Technology Rail in Utrecht to manage LAWS for
them. There are 17 LAWS fitted vehicles operating. AEA Technology Rail man the LAWS
terminal and if LAWS identifies a low adhesion problem then the AEA Technology Rail
person contacts the regional control office and provides details of where and when the low
adhesion event occurred. The regional controller has access to a look-up system and can
then determine which trains, and their associated GSM mobile number, will traverse the
route over the following 2 hour period. An SMS alert is then sent to these mobiles detailing
the route and kilometre position of the low adhesion so as the Driver may take defensive
action. Similarly, if a Driver radios in and reports low adhesion, then the regional controller
operates the same communications system.

The Adhesion Controllers Condition Assessment Tool (ACCAT)

Metronet BCV Rail operate the ACCAT to determine the acceptability from a risk perspective
of operating the Central Line ATO system outside of the tunnel sections during the autumn.
The process is semi-automated but manual assessments and manual inputs are relied on at
present. A variety of information is sourced from various inputs, including an emergency
brake test each day at defined sites to provide an idea of the average braking distance. The
ACCAT algorithm indicates the level of mitigation required and a manual feedback loop
determines whether it had been applied or not. The tool processes the information and
outputs this in the form of a decision matrix. The matrix contains a measure of the risk of
contamination, principally sourced from the ADAS information modified to reflect the effect of
moisture, and the 'railway condition' dominated by the railhead condition. From the matrix
comes a range of decisions from 'Favourable' requiring no further actions, 'Normal' requiring
say one run of the mitigation train, the 'Adverse' may require more than one run of the
mitigation train, and 'Severe (red)' would require virtual continuous running of the mitigation
train, and 'Severe (black) indicates no prospect of running in ATO. The system has been
developed year-on-year and forms a sound basis for moving forward, although aimed at a
slightly different situation than for a mainline application.

An Adhesion Management System on an ERTMS Railway

It is vital to know that there is sufficient adhesion available to sustain the required traction
and braking rates for operating a safe railway. Adhesion is therefore an important element
that requires incorporation into European Rail Traffic Management System (ERTMS). [41]
investigates the issues that need to be addressed in order for that to the feasible. The
feasibility of providing adhesion information to ERTMS has been established by reviewing
the critical factors that are required for a decision process. Critical adhesion factors were
categorised under Vegetation, Weather, Track, Rolling stock and Railhead Contaminants. It
is concluded in the report that an Adhesion Management System (AMS) should preferably
be able to automatically assess the current adhesion conditions in real-time and also predict
changes. In the current proposals for ERTMS, a reduced brake rate is specified as being
70% of that utilised under normal adhesion conditions. The AMS would provide a
mechanism by which the brake demand level can be switched between normal and low
adhesion mode and vice versa when adhesion conditions change. The system should also
be able to provide early warning of low adhesion to enable the management of mitigation
measures.

The essential process outlined is to monitor the factors that determine the wheel/rail
adhesion level and process this information to formulate a decision on level of brake force
demand that can be sustained. This would then be supplied to ERTMS. However, a decision
process could also be used to direct mitigation measures that will form part of a feedback
loop since they will influence the inputs to the decision processor. Furthermore, once
ERTMS has adjusted the brake rate it will be useful for this to be confirmed both as a safety
check and for use in the decision process. Feedback of the brake rate status will enable the
AMS to operate the switch in both directions, that is from high to low adhesion and vice
versa.
The process of assessing conditions, through adhesion inputs to a decision matrix, would
provide a semantic scale of adhesion levels, e.g. normal, adverse, severe, very severe. It is
unlikely that specific values of adhesion coefficients will be available to ERTMS on a
continuous basis. A number of building blocks are considered to currently exist to provide
the basis for a prototype European Train Adhesion Management System (ETAMS). LAWS
coupled with weather information, in particular moisture and rainfall levels, railhead
contamination levels and vegetation assessments will form significant inputs to the proposed
system. Experience with the ACCAT developed by Metronet BCV, provides valuable
information for building an AMS.

As a result of the original work done for AWG, and the encouraging results seen from
ACCAT, and the work undertaken by AEA Technology Rail for RSSB described above, AWG
commissioned AEA Technology Rail to review how such a model could be developed for a
conventionally signalled railway.

An Adhesion Management System on a Conventionally Signalled Railway

[42] provides a high-level review of the feasibility of applying the concept of European
Adhesion Management System (ETAMS) to a conventionally signalled railway. In particular,
the report considers the aspects associated with developing the communication and control
methodology for a real-time AMS. The required inputs and outputs, and the potential
methodology of achieving these are as described in [41]. The findings of [42] indicate that
application of a real-time AMS system to the conventionally signalled railway is possible,
utilising existing available technology. The AMS would provide primary benefits in terms of
more effective utilisation of traffic along a route, a reduction in workload for signalling staff,
and reinforcements of good driving practice.

It is considered that an AMS system which combines both predicted and reactive elements
within its control will provide the most effective strategy. In particular, the predictive elements
of the AMS will enable warnings to be dispatched prior to the passage of the first train of the
day over a particular section of low adhesion. In addition, reactive feedback, provided by
trains operating over low adhesion sites, will define very accurately the level of adhesion
present, the exact sites of low adhesion and any changes in condition. Additionally, the
reactive approach provides a method of defining low adhesion sites requiring mitigation
actions and the effectiveness of these actions over time.

A working system would rely primarily on satellite based positioning, supported by additional
positional information front inertial devices and the signalling system. A satellite based
positioning approach adopted should recognise future developments in this area from ESA
and the UIC in relation to application of satellite technology to railways. In particular, this
should consider EGNOS beacons and other Galileo satellite GNSS. Selection of an
appropriate approach will require further evaluation. In particular, the accuracy of the
GPS/GNSS selected and the survey of coverage along AMS operating routes must be
considered.

Communication between vehicles and a host computer will require a predictable, and
reliable network with good service levels with guarantees. The obvious candidate for this
application is GSM-R. If GSM-R is unavailable in the short or longer term, then consideration
should be given to a private mobile network, as this will offer levels of functionality and
performance required by AMS. The most likely candidate in this category is the Transcomm
Mobitex network. With more detailed testing and proving its may also be possible to achieve
a satisfactory level of service from a public network such as GPRS/SMS, as this approach
would potentially offer the lowest operational cost. It will be necessary to review the service
level agreements, and actual coverage applicable to the rail network in more detail for all
preferred options.

[61] identifies and prioritises the factors that are considered important for an AMS: weather
factors including rainfall, cloud cover, air temperature, wind speed & strength, atmospheric
pressure; rail conditions including railhead moisture, contamination type and persistency of
remedial treatment; lineside factors including leaf fall mobility & density on the ballast,
wetness of leaves, vegetation survey and leaf fall progress; train factors including braking
performance, traffic frequency and LAWS slip data.

How would it link into the VTI Model?

A specific question in the work specification [1] is how would an automated low adhesion
prediction model link to the VTI model? [52] defines the VTI model as: a Model that links
inputs such as track and vehicle characteristics to outputs such as rail life, wheel life and
maintenance regimes… the model will enable maintenance to be optimised through “what-if”
studies and the like.
Until the VTI model is fully understood it is difficult to identify how a low adhesion prediction
model would link into it. However, it is likely that certain outputs could be important to the VTI
model such as locations of wheelspin (indicating the possibility of rail burns). To define this
requires a knowledge of what parameters the VTI model would be looking for to determine
whether a low adhesion prediction model could provide these or not.
12 Conclusions

This report has distilled accessible knowledge on a number of specific areas related to low
adhesion. This has been based on a review of many reports and papers available to the
author; interviews of key personnel actively involved in low adhesion research and
development; and from responses to a questionnaire from other interested parties and
overseas railway administrations. The following conclusions with respect to the Project
Specification have been drawn from this knowledge search:

12.1 Low Adhesion Formation

The formation of low adhesion is a complex interaction of many factors, including the nature
of the contaminant, local topographical conditions (cuttings etc.), weather conditions (wind,
rain etc.) and vehicle characteristics, to state but a few. The specific influence of all of the
factors is not fully understood despite a significant amount of research, laboratory testing
and full-scale track measuring and testing. Neither is the proportion of mix of these factors
fully understood. However, the factors that will lead to low adhesion are known.

Creating low adhesion

The industry has long sought to simulate leaf film for test and development purposes. This
has been done by using a rolled-in paper tape for many years by BR Research and then
AEA Technology. Ove ARUP introduced the ‘grass and lignin’ method after research by
QuinetiQ, to simulate leaf film. Any such simulation needs to be appropriate to the need. For
example, skidpan training for Drivers requires the adhesion level to be low but it does not
matter how it is lowered. However, the testing of remedial treatments (e.g. Laserthor and
water jetting) needs the artificial contaminant to be representative of the actual leaf film
contaminant if the results are to be relied on. Ove ARUP are known to have undertaken
some steps forward in this area, but it is not known how representative these test processes
actually are of real leaf film.

How quickly in terms of time and distance does adhesion change and how long does
it last for?

The simple answer to this is that, in terms of time, low adhesion conditions have been shown
to materialise instantaneously or gradually over time. The onset of drizzle, or reaching dew
point on the railhead, on a heavily leaf contaminated railhead can immediately reduce what
were ‘adequate’ adhesion conditions (0.1) to extremely low conditions (0.03 or less). This
was demonstrated by a high profile buffer stop collision at Slough in 1994 as a train passing
over the same track only minutes before did not suffer any braking problems. Conversely,
the moisture content of the leaf film can dry out in the sun such as to very quickly restore
adhesion to a level sufficient to restore braking and traction conditions to a suitable level.
However, no evidence could be found about the precise amount of moisture, or rate of
drying, necessary to reduce adhesion to insufficient levels for the various contamination
types.

The formation and removal of contamination also depends on the traffic intensity. Oil
contaminated rails can be quickly cleaned by the passage of trains. Various observations on
this have been made in the past: the passage of an 8 car Class 508 EMU at 50 mph caused
as many as 60% of the leaves present in four foot to be deposited on the rails; a 4-vehicle
16-axle train has been observed to deposit on average one leaf per yard of rail. The
continuous black rolled film was been observed to re-establish itself with the passage of a
just a few trains, amounting to no more than 200 axles, when suitable drying conditions
returned. Rolled-in black leaf film has been observed to be removed by the passage of a few
trains. Also it has been identified from intensive monitoring of a number of sites that the
removal of leaf contaminant was accomplished by about 300 axles passes.
In terms of distance, again the simple answer is that an extremely low adhesion ‘blackspot’
can be directly encountered, particularly if the traffic flow is in one direction only.
Contamination is known to be picked up by passing wheels and deposited further along the
track by passing trains. Thus on bi-directional tracks, it would be expected to see a gradual
reduction in adhesion along the length of the rail until the specific problem site is
encountered. Similarly, as a train passes through a site of low adhesion (on uni- or bi-
directional operation) it would be expected that the adhesion levels would gradually improve
rather than have a sudden return to good conditions. However, sudden and abrupt changes
in adhesion have been measured in the restoring of conditions as well.

A significant amount of low adhesion track measuring has been conducted using a full-scale
Tribometer Train. This information has been plotted into 3-D plots showing the naturally
occurring adhesion profile of many miles of track with respect to time and distance. The
majority of low adhesion sections measured were less than 200m long and the frequency of
low adhesion sections identified below 0.06 was around one third of those below 0.09. There
were no low adhesion lengths longer than 400m where the adhesion was continuously 0.04
or below. Later data from LAWS trials identified that the majority of slip events were of less
than 50m in length but there were a significant number greater than 300m.

And how localised is it?

Low adhesion is very site specific but the length of a site can vary from a very short section,
e.g. a few metres in a cutting; through a medium section, e.g. a few miles over a tree-lined
route; or could be over extensive distances, e.g. tens of miles over a little used route after a
long dry spell. Specific features are generally required to create low adhesion (eg. local
topography, track geometry, microclimates etc.) which by definition will be localised.
However, the general onset of drizzle can create low adhesion over very wide areas perhaps
after a long dry spell. Low adhesion has also been measured many miles away from its
source due to transference down the track by the action of the wheels of passing trains.

Track geography can play a significant role in determining whether a given location becomes
a leaf site. The position of the trees in relation to the prevailing wind can determine where
the leaves land; railway cuttings are more liable to leaf contamination than level ground or
embankments. This is partly because they act as a leaf trap, and partly because the
moisture content of the air tends to be higher, encouraging the formation of dew and
delaying evaporation. The type of tree is also important as both the size of the leaf and the
volume of leaves are critical, e.g. several Beech trees are required to produce the same
level of contamination as one Sycamore.

In summary, the factors that create low adhesion can be very localised but can also extend
over significant distances. The existence and duration of low adhesion depends on local
features and climatic conditions, which themselves are changeable and unpredictable.

12.2 Leaf Contamination Properties And Characteristics

What is it about the surface chemistry and physics of leaf contaminant “Teflon-like”
coating, which means that it adheres to the rail, yet causes wheels to slide over it?
The formation of leaf contamination is a complex interaction of many factors including leaf
type, weather conditions and train type and frequency. The relationship of these factors in
creating the hard leaf film layer are not fully understood despite a significant amount of
research, laboratory testing and full-scale track measuring and testing. There are, however,
theories regarding how the surface properties of contamination affects friction - the known
'paste effect' formed by rust or other solid eg. coal, iron ore, clay, and water. On thickly
rusted rails adhesion coefficients as low as 0.03 have been measured in drizzle. The
proportion of water to solid matter is critical. Very low adhesion occurs as the paste moves
from viscous liquid behaviour to that of a solid (or vice versa). It is postulated that the same
effect is at work with leaf contaminated rails.

The usefulness in further understanding just how the wheel slides over the leaf
contamination is questioned, as it is fundamentally an issue of compacted leaf contaminant
and moisture. The exact moisture requirement is not understood neither is the amount of
leaf, and these factors will form an input into low adhesion prediction systems, as is
discussed under the low adhesion prediction and communication section.
No evidence could be found pointing to what it is about the hard leaf film layer that binds it to
the rail. It is known that water will soften it for ‘easy’ removal, and the layer is less likely to
form in a wet rail (although a damp rail is reported to be more likely to cause leaves to stick
to it). It is also known that LaserThor ablates the leaf contamination by acting at the
contaminant/rail interface rather than by ‘blasting’ the contaminant off the rail from the top.
The usefulness in answering the basic question of how the leaf film adheres to the rail lies in
the possibility of identifying products and processes (other than water softening) that would
break the bond between the leaf film and the rail. This could also provide a more scientific
basis as to how much water is required to soften what thickness of leaf film.

Also, nothing has been found that details the hardness of the black rolled-in leaf film and
hence the force required to scrape the film off the rail. The usefulness of this data would be
for assuring that artificial leaf film contamination used for testing, e.g. the AEA Technology
Rail 'paper tape' method or the Qunietiq 'Lignin and grass' method, is a true representation
for real leaf film contamination when assessing the efficacy of various rail cleaning systems.
Dr IJ McEwen described a relatively simple ‘pencil hardness’ test in his review of leaf lifting
products for AWG. The Pencil Hardness Test (ASTM D 3363) is a simple test which has
proved useful in assessing the hardness of paint and other organic thin films. The only
equipment needed is a set of pencils which come in a range from 9B (very soft) through to
9H (very hard) with HB and F grades in the middle, between B and H. The harder pencils are
the more useful. The pencil is held at 45o (a jig is sometimes used to set the pencil angle)
and pushed against the film to be tested. By moving from hard to softer pencils, the grade
which does not scratch through the film defines the hardness. Whilst the test can be
improved by angle jigs and constant pressure devices of various designs it in general shows
good repeatability in its simple hand-held pencil method. Examples of use showed that damp
leaf films on a test site were all scratched by a 3B pencil but not by a 5B pencil. On an open
site where the rails were dry, the leaf films were repeatedly scratched by a 4H pencil but not
by a 3H, the results showing good repeatability.

AEA Technology Rail have also described a simple ‘torque’ test whereby a stud is glued to
the contaminant the torque required to break the leaf contamination/rail bond is measured.

What are the friction properties related to the leaf contaminant thickness?

No definitive evidence has been found that answers the question: how much debris is
necessary on rails before the lowest friction is encountered? Neither has a relationship been
found between the leaf contaminant thickness and friction levels. Hence it is not yet
understood how much contamination can be allowed to build up before adhesion becomes
less than that required. Conversely, the same is true in determining how much contaminant
needs to be removed by railhead treatments. The average thickness of the black leaf film
has been found to be in the range 20 to 35 microns.
In discussions with AEA Technology Rail, anecdotal evidence materialises from brake
testing conducted on simulated leaf film using damp, rolled-in paper tape. Similar overall
deceleration rates were achieved over repeated runs on the paper tape, each run gradually
removing some of the simulated leaf film. Very low adhesion was still encountered when
there was little visible paper left, suggesting no linear relationship between contaminant
thickness and adhesion level.
The Ove ARUP monitoring for Network Rail conducted during Autumn 2003 concluded that
railhead friction and contaminant thickness appear to be completely independent variables.
The work also identified that high pressure water jetting still leaves a contaminant residue of
around 10 to 15 microns thick.

Does water soften it such that there is an opportunity to agitate it off the rail?
The simple answer to this is yes. It has been known for many years that water softens
crushed leaf film allowing its removal by passing trains, and also prevents leaf film build up.
These effects have been successfully demonstrated by using trackside water sprays which
were very effective, if not costly and practically difficult to manage.
Other products are in use to soften the leaf film layer, currently applied using by hand held
applicators, being based on diluted citrus cleaner (d-limonene). These products, sometimes
used with a hand-propelled rail scrubber, exhibit a faster softening process than water alone,
but add to the cost.

AEA Technology Rail have assessed the use of water as pre-wetting spray on a water jetting
vehicle to improve the effectiveness of the high pressure water jet, but with little success due
to the speed of train operation.

12.3 Friction Modifiers

Can friction modifiers be used to mitigate the effects of low adhesion? If so, how
effective would they be and how would they be used?
This review has concentrated on products other than Sandite that could be used to modify
adhesion levels. Quite a lot of research has been conducted over a number of years to
identify and test friction modifiers. The earliest tests conducted by BR Research used the full
scale Tribometer Train, with later tests relying on brake testing or portable tribometers to
determine their effectiveness. In 2003, AWG commissioned Dr IJ McEwen, formerly of
BRR/Scientifics, to review the current position with respect to research he himself had
conducted over many years. Network Rail’s recent Bicester trials have also tested two
friction modifying products. The results of almost all of the research found was that friction
modifiers have mainly been designed to control high friction situations and not for improving
low friction situations. Very little success has been seen in raising adhesion with adhesion
modifiers despite extensive trials. Test results have consistently shown that the application
of such products needs to be done with extreme care as they almost invariably can lead to a
lowering of the adhesion level. There is no evidence supporting the use of these products to
combat low adhesion despite extensive testing of a wide variety of potential products.
However, Banverket reported on the use of lime as a friction modifier as having good results
and referred to tests conducted by Danish State Railways. The results of this have not been
seen by the author.

12.4 Vehicle Design Influences

What are the critical factors of vehicle characteristics that improve or worsen
adhesion performance (e.g. suspension, wheels, traction and braking system
design)?

Vehicle body, suspension, bogies, wheelsets etc.

It is clear that the design of the vehicle, its suspension system including bogies and
wheelsets, has an effect of the wheel/rail interface with respect to low adhesion. Whilst little
research in this area has been undertaken in the UK, work reported by ORE identifies the
main influences and shows some of the relationships. Influential factors identified include:
instantaneous vertical load and effect of load transfer; axle load; weight of body; weight of
bogies; geometry of body; geometry of bogies; deceleration of the vehicle; surface of the
contact zone; diameter of wheels; wheel profile; and wheel hardness. Some of these factors
are quantified in the report, but referenced reports (not available to the author) contain the
detail. However, it can be inferred that the effect of vehicle suspension etc. is not significant
on overall low adhesion performance.
To be of value, research in this area would need to provide design data that could be
provided to vehicle designers/manufacturers such that they could use the data in their
designs. The vehicle dynamics simulation model used in the UK, VAMPIRE, does not have
the characteristics of low adhesion modelled within other than as a basic 'wet' or 'dry'
characteristic. In discussion with AEA Technology Rail, the amount of modelling and track
testing that would be required to provide this is considered to be extensive. Therefore, the
potential result of understanding this relationship in further detail, whilst acknowledging there
is a contribution to be made, is not considered to lead to a significant improvement in low
adhesion performance of vehicles and the value of further work in this area is questioned.
Further, it has already been seen that the goal of suspension designers in seeking a high
quality ride with reduced wear on components, is in itself opposed to the requirement for a
poor wheel/rail interaction to help clean the rail of contamination.

Traction systems
Some research has been conducted on the effect of traction control systems on low
adhesion performance, leading to recommendations for improving older systems. Modern 3
phase drive traction packages are known to exhibit inherently improved slip control
characteristics when compared to older traction systems. However, no evidence could be
found that any specific considerations have been given to traction design to maximise
performance under low adhesion conditions. Furthermore, anecdotal evidence from Network
Rail suggests that basically similar traction systems, but manufactured by different suppliers,
appear to have markedly different performance in the autumn season. The performance of
diesel electric and direct drive traction systems does not appear to have been studied either.
Any further work on this subject area would assist in reducing autumn delays due to
acceleration problems, although this has not been quantified.
(see section 12.5 for a whole train traction concept).

Braking systems
A significant amount of work has been conducted in the past to understand the influence of
braking systems (e.g. tread or disc brakes) on low adhesion and vice versa. AEA
Technology Rail possess what is state-of-the-art WSP optimisation equipment (WSPER)
which optimises WSP equipment to the vehicle characteristics. AWG has previously
constructed a cost benefit model to enable operators and vehicle owners to determine if
there is a business case to upgrade older WSPs with modern, optimised systems. A modern,
optimised, adaptive WSP system is understood to maximise available adhesion by operating
in a slip region which aids braking further by its wheel/rail interface conditioning effect. It is
concluded that conventional braking systems have matured over many years and further
research into their influence on low adhesion performance will yield little if any benefit.

(see section 12.5 for a whole train braking concept).

Aerodynamics
The shape of the train front end, the skirting or otherwise of the vehicle undersides and the
shape of the vehicle underside, all contribute to the train’s aerodynamics. The mechanism
for leaf deposition and crushing on the railhead is well documented as being due to the
vehicle turbulence that causes leaves lying on the track and trackside (mainly within the 4
foot) to be whipped up, trapped under the passing wheels and crushed onto the railhead. It
follows the aerodynamic performance of individual train types will affect the rate of leaf
contamination, either by drawing leaves towards the wheel/rail nip, or positively preventing
them being drawn into this area. It has been shown by test that specific types of vehicle can
deposit and crush more than 4 times the number of leaves as others. Limited experiments by
BR Research with bogie mounted leaf deflectors showed a 70% reduction in leaf build up.
Further testing by Laserthor claimed an 80% reduction in the number of leaves being
trapped in the wheel/rail nip, with potential to reach a 90% reduction. Vortok Leafguards
mounted around the rail itself have shown an approximate 50% reduction in leaf film build
up.
From this it can be concluded that the aerodynamic properties around the wheel/rail
interface can significantly affect the rate of deposition of leaves on the railhead. The rate of
deposition of leaves is obviously fundamental to the build up of leaf contamination. This can
be influenced by vehicle airflow considerations and by introducing equipment such as leaf
deflectors to reduce the leaves entering the wheel/rail nip.

12.5 Whole Train Braking/Acceleration Concept

Are there benefits in varying the braking and traction demand along a train to
maximise the use of available local adhesion and thus take advantage of the
scrubbing action from the front wheels?
Perhaps the greatest cost to the industry from the autumn effects is the delay and general
disruption caused by train slipping and sliding on contaminated rails. This leads to significant
adverse publicity and dents the railway’s reputation, as well as having a direct financial cost
in performance delays and lost revenue. Bearing this in mind, it is considered that there must
be benefit to be gained in improving the capabilities of trains to accelerate and brake in the
autumn to reduce these delays. This review has not set out to quantify the benefit that a
Train Braking Control (TBC) system or a Train Acceleration Control (TAC) system could
realise, but this would have to be one of the first outputs of any future research work.
No research work has been identified that has studied the effects and benefits of a closed-
loop whole train traction control system.
The initial concept of a TBC system was reviewed against other methods of reducing
stopping distances under low adhesion conditions, such as optimised WSP, trainborne
sanders etc. The review did not favour TBC as better performance could be achieved by the
use of these other controls. Since then, the performance of a modern, optimised WSP
systems has emerged as a device capable of controlling individual wheelsets at the limit of
available adhesion, hence there is no need for an intelligent brake control system controlling
each vehicle separately. However, the conventional brake systems are still limited to the
brake demand selected by the Driver and do not seek to automatically exceed this demand
to compensate for times when this demand is not achieved. The conditioning effect along a
train is known about but not fully understood. Further work to determine the benefit of a TBC
to improved autumn performance is considered to be worthwhile, albeit raising extremely low
levels of adhesion to more normal levels is not expected to be achievable.
AEA Technology Rail have suggested an alternative approach to TBC which includes the
trainborne sander into the control loop, yet does not rely on introducing any further
intelligence in to the braking system, making use of the WSP system performance and
existing levels of brake demand available. AEA Technology Rail have also suggested a
potential way to find out if a TBC could offer a real benefit would be to get more data on the
conditioning effect down the train. One way to find out if such a TBC could offer a real
benefit would be to get more data on the conditioning effect down the train. AEA Technology
Rail suggest that a longer train set should be instrumented for the autumn to measure the
mean WSP regulated brake pressure on (say) one axle of each vehicle (similar to the
Pendolino tests conducted on behalf of Alstom). With a good WSP, in low adhesion
conditions the brake pressure is a direct indication of the adhesion 'seen' by that axle. It
would then be possible to plot the effective adhesion down the train, taking into account the
direction of travel and the brake demand. Also, by disregarding the data for the vehicles at
the back of the train, the likely effect on shorter train sets could be established. AEA
Technology Rail recommend an initial trial should be conducted on one or more artificial
contaminants. The unit would then be released to service traffic with measuring equipment
on-board and operational. This would allow the conditioning effect to be established for a
range of natural autumn adhesion conditions and the TBC benefit could then be assessed.
Even if the answer is not positive for TBC, the conditioning data would be very useful
common domain information as this topic it is currently not well documented. A long train set
with a modern WSP would be essential for such a trial. There must be no sander fitted as
this will modify the result. Pendolino would be ideal for such an exercise or a 5-car Voyager
with ‘one shot’ sander would be an acceptable second choice as long as the (infrequent)
firing of the sander was recorded.

Also, it is necessary to establish how often we could benefit from such a system and if it is
only applicable to long trains or if shorter formations could also benefit.
12.6 Adhesion Prediction & Communication
Can current manual adhesion prediction models (e.g. Central Line ACCAT) be
automated to reliably predict the occurrence of low adhesion in real-time to minimise
disruption and if so how?

What would such a model look like and how should it be developed?

How could this be communicated to the train driver?

A recent study conducted by AEA Technology Rail for the Adhesion Working Group has
reported on the feasibility of an automated Adhesion Management System (AMS) for
conventionally signals lines. It is based on earlier work undertaken for RSSB, specifically
aimed at adhesion management for ERTMS. In particular, the report considers the aspects
associated with developing the communication and control methodology for a real-time AMS.
The required inputs and outputs, and the potential methodology of automating these are as
described in the RSSB work on ERTMS. These factors would require research and
development of appropriate techniques also. The findings of the feasibility study indicate that
application of a real-time AMS system to the conventionally signalled railway is possible,
utilising existing available technology. The AMS would provide primary benefits in terms of
more effective utilisation of traffic along a route, a reduction in workload for signalling staff,
and reinforcements of good driving practice.

How would it link into the VTI Model?

Until the VTI model is fully understood it is difficult to identify how a low adhesion prediction
model would link into it. However, it is likely that certain outputs could be important to the VTI
model such as locations of wheelspin (indicating the possibility of rail burns). To define this
requires a knowledge of what parameters the VTI model would be looking for to determine
whether a low adhesion prediction model could provide these or not.

12.7 Adhesion Measurement

The requirement to measure railhead friction levels includes for research and development
purposes, for equipment (e.g. brake system) testing purposes and for post-incident analysis.
The method of undertaking these measurements may vary as does the equipment used:

• Small scale wheel rigs,

• Small scale wheel Portable tribometers or vehicle mounted,

• Full-sized wheel on rail Tribometer Trains,

• Systems fitted to service vehicles, e.g. WSP, LAWS, OTMR etc.

AEA Technology Rail produced a report for Railtrack outlining the different adhesion
measuring techniques available and identifying their application and limitations. They
concluded that there were a number of options for measuring adhesion, including portable
tribometers for 'everyday' use, full-scale tribometers for research work and service trains
fitted with train data recorders (with optimised WSP) for post incident analysis or LAWS to
act as a service tribometer train. Regarding portable tribometers, they were considered more
applicable for Railtrack's adhesion measurement requirements, but they were considered to
be restricted in value because of the small contact area at the measuring interface, meaning
that adhesion values could be affected by surface roughness, wear and penetration of
contaminant films, all of which can result in false adhesion values. AEA Technology Rail
considered that if the portable devices were modified to make more accurate measurements
they would still only be valid for relatively high values (>0.1) of adhesion. However, they did
consider there was potential for the development of a larger scale portable device which may
overcome some of the limitations of the existing designs.

From the knowledge search conducted, it is quite clear that current research and
development is relying to a degree on the output from portable tribometers. The former BR
Research Tribometer Train is due to be scrapped shortly (life expired) and would not in any
case lend itself to some of the reactive measurements required. Neither does full scale train
testing. Therefore, reliance on portable tribometers will exist for the foreseeable future. It is
known that the results from small-scale tribometers tend to read high and are not reliable for
absolute measurements. Recent monitoring by Ove ARUP for Network Rail suggests that
consistent results can be provided by this equipment between individual operators and
individual devices. However, calibration of the Salient Systems portable tribometer
undertaken by BR Research using the full-scale Tribometer train, indicated that they may not
be a reliable comparator between different contaminants.
13 Recommendations

As with most research and development, especially in today’s UK rail industry, there has to
be a potential use for the knowledge gained to realise a payback for the investment put in.
Resources are finite and stakeholders expect their resources to be invested wisely, with an
ultimately realisable benefit. That is not always apparent when embarking on research and
development, but the author has attempted to recommend future research areas that will
likely lead to a benefit worth pursuing. In addressing the questions posed by the RSSB
Project Specification [1], it is not always considered that there is a benefit to be gained in
filling in the missing gaps in current knowledge, as the likely costs, timescales and difficulty
in arriving at the answer may not on first sight appear to be justified. However, no
quantification has been put to this, it is the opinion of the author, and the reasoning is
documented in the recommendations below for further consideration by RSSB.

13.1 Low Adhesion Formation

13.1.1 Despite the specific influence and proportions of the many factors that create low
adhesion not being fully understood, no further research is recommended to answer
the questions relating to the formation of low adhesion. The influential factors are
known. It is a fact that low adhesion can be very localised but can also extend over
significant distances; it can arise and subside very abruptly or over time and
distance. It is not considered that further knowledge in this specific area will provide a
significant benefit other than to a prediction system (see 13.6).

13.2 Leaf Contamination Properties And Characteristics

13.2.1 It is recommended that further research is undertaken to determine what it is that
binds dry leaf film onto a steel rail in order that products and processes may be
sought that will very quickly break this bond such that the performance of rail
cleaning treatments could be enhanced. The work should also attempt to quantify the
amount of water required to sufficiently soften the contaminant and the speed at
which this process works. Research of this nature would require the creation of ‘real
leaf film’ in the laboratory (a process BR Research have previously successfully
undertaken) and then analysis of the bonding structure via various analytical
techniques. This would be followed by a search of likely products and laboratory
testing to identify suitable products and processes. However, prior to this it would be
necessary to positively identify the potential use for any such processes as water is
already known to break the bond as are d-limonene solutions, both of which are in
current use.
13.2.2 It is recommended that further research be conducted to determine the hardness of
the leaf film with respect to the force required to remove it. The need for this stems
from gaining a basic understanding of how representative artificial leaf contaminants
are when using them to assess the efficacy of rail cleaning systems. Also, knowing
the force required to remove the contamination will assist in determining railhead
cleaning system requirements. A ‘best practice’ hardness measuring technique(s)
should be established. This should be a relatively simple piece of laboratory work,
followed by validation on real leaf contamination.
13.2.3 It is recommended that further work is undertaken to establish the relationship of
contamination thickness to friction levels, in particular to determine what the
minimum thickness of leaf contamination needs to be to reduce adhesion levels to an
unacceptable level. This will aid both the Network Rail railhead cleaning processes
and will form criteria for input to the low adhesion prediction system. Again, this could
be achieved in the laboratory followed by validation on real leaf contamination.
13.2.4 It is recommended that an ‘industry standard’ low adhesion simulation technique, or
the properties required of such techniques, be developed to ensure that development
and testing of equipment, systems and processes aimed at removing or modifying
leaf film contamination, are representative of the properties and characteristics of real
leaf film. This should include determination of moisture levels with respect to friction
levels, which would be useful in two respects: firstly as a potential input to low
adhesion predictive systems and secondly as a control measure for track testing on
artificial leaf contaminants. This would require a mixture of laboratory study and on-
track simulation and testing.

13.3 Friction Modifiers

13.3.1 No further research is recommended in the area of friction modifiers (improvers or
enhancers) due to the extent of current knowledge suggesting their use as a friction
enhancer has significant risks of actually lowering the friction levels.

13.3.2 However, the results of tests using lime conducted on Danish State Railways should
be sought and reviewed.
13.4 Vehicle Design Influences
Vehicle body, suspension, bogies, wheelsets etc.
13.4.1 No further research in the area of how vehicle suspension, bogies etc. influence low
adhesion performance is recommended as it is considered that a significant amount
of work would be required for little overall benefit. However, should RSSB wish to
pursue this subject area further then it is recommended that some of the reports
referenced in [25] be obtained to quantify the critical factors and the magnitude of
their influence.

Traction systems
13.4.2 Further research is recommended to establish what the attributes of various traction
drives should be with respect to controlling acceleration on low adhesion, as the
potential benefits from reduced train delay due to poor acceleration are likely to be
attractive (see section 13.5 for other aspects of traction control). The first step should
be to quantify the likely benefit to be realised. The main exercise is likely to require a
review of different traction systems and how they control wheelslip and compare this
to their performance in the autumn.

Braking systems
13.4.3 No further research is recommended in the area of how braking systems influence
low adhesion performance as it is considered that conventional braking systems
have matured and offer a state-of-the-art system with respect to managing low
adhesion. (see section 13.5 for other aspects of brake control).

Vehicle aerodynamics

13.4.4 It is recommended that the effects of vehicle aerodynamics on the deposition rate of
leaves on the railhead is further researched with respect to improved design of
airflows under the vehicle and across the railhead. The objective of the research
should be to identify whether the airflow can be successfully directed, and by what
means, such as to positively prevent, or minimise, leaves being presented into the
wheel/rail interface. This work should also review the use of rail mounted leaf
deflectors, either in isolation or combination with vehicle-mounted systems, that
effect a similar end result. This will in all probability would involve Computational
Fluid Dynamics modelling, require access to a wind tunnel, and ultimately track
testing.
13.5 Whole Train Braking/Acceleration Concept
13.5.1 It is recommended that further research be conducted to establish the potential
benefits to improved autumn performance that might be realised from a closed-loop
Train Brake Control system incorporating a trainborne sander system. The prospect
of retrospective application to existing vehicles with modern WSP systems should
also be investigated. Section 12.5 suggests a methodology for establishing this.

13.5.2 It is also recommended that further research is conducted to establish the potential
benefits of a ‘whole train’ closed-loop traction control system whereby individual
bogie traction drives could compensate for others not able to sustain the required
level. The prospect of retrospectively modifying existing vehicles should also be
investigated. The first step would be to determine the likely benefit to be realised
before proceeding with theoretical and practical studies.

13.6 Adhesion Prediction & Communication

13.5.2 Based on the findings of the AEA Technology Rail feasibility study, the development
of a trial AMS for conventionally signalled lines is recommended by AEA Technology
Rail and supported by the author, noting that it has yet to be discussed at the AWG
(note such a trial would also benefit the future ERTMS AMS system proposed-
ETAMS). The following recommendation is taken from the AEA Technology Rail
report: A trial will provide an opportunity to evaluate most of the concepts described
within the report. It would involve development of a minimum of two Autumn seasons,
and would be limited to a single route over a manageable route mileage. A suitable
communications bearer would be identified but it is not proposed to use GSM-R for
such a trial in view of the recent Network Rail announcement regarding further delays
in implementation. During the trial the positioning approach adopted will rely upon a
GPS/GNSS solution possibly supported by other devices. Due to the amount of
development required, the signalling interface, which is an option for vehicle
positioning and for the messaging strategy, is only likely to be a feature of a wider
area implementation. Following selection an assessment of the input factors used for
the decision-making process the total system will require optimisation over a 2 to 3
year period, although the adoption of a self learning software approach may minimise
this. It is recommended that the first year would involve shadow running of the
system, and it will include an opportunity to obtain comments on the operation of the
system from the parties concerned.
This work would also include developing the inputs and outputs identified in the work
conducted for RSSB for ERTMS.
13.7 Adhesion Measurement
13.7.1 It is recommended that the subject of ‘what should be used to measure low adhesion’
be explored further with the aim of creating an industry standard or best practice
guide. The objective is to prevent inappropriate measuring methods being used, or
inappropriate reliance on the results without understanding the calibration. This work
is likely to require a search of available systems, the categorisation of use of these
systems and the provision of calibration/scaling information, most likely gained
through practical testing. Whether a ‘new’ device is required will depend on the
outcome of the initial stages of work. AEA Technology Rail have previously identified
that the current small-scale portable tribometers could be improved or a more
accurate semi-portable device be developed.
14 Abbreviations

ACCAT Adhesion Controllers Condition Assessment Tool

AMS Adhesion Management System
ATB Auxiliary Tread Brake
ATO Automatic Train Operation
ATOC Association of Train Operating Companies
AWG Adhesion Working Group
BR British Rail
BRR British Rail Research
DMU Diesel Multiple Unit
EGNOS European Geostationary Navigation Overlay Service
EMU Electric Multiple Unit
ERTMS European Rail Traffic Management System
ESA European Space Agency
ETAMS European Train Adhesion Management System
GNSS Global Navigation Satellite System
GPRS General Packet Radio Service
GPS Global Positioning by Satellite
GSM Global System for Mobile communications
GSM-R Global System for Mobile communications - Railways
LAWS Low Adhesion Warning System
NS Nederlandse Spoorwagen
RSSB Rail Safety & Standards Board
SMS Short Messaging Service
TAC Train Acceleration Control
TBC Train Brake Control
TCA Track Circuit Actuator
TOC Train Operating Company
UIC Union Internationale des Chemin de Fer
VTI Vehicle Track Interaction
WSP Wheel Slide Protection
WSPER Wheel Slide Protection Evaluation Rig
WSTCF Wrong Side Track Circuit Failure
15 References

The page number shown at the end of each reference below refers to the relevant page
within the appendices in this report where the summary may be found.
[1] T354 RSSB Research Project Specification, Understanding
Low Adhesion, Part 1 Review of Current Knowledge.
[2] TM TRIB 1 A Basic Study of Wheel/Rail Adhesion 4 – Laboratory
Simulation Rigs, TM Beagley, R Jackson, C Pritchard &
JD Sewell, BR Research, December 1971
[3] TM TRIB 2 A Basic Study of Wheel/Rail Adhesion 5 – Laboratory
Studies of the Effects of Oily Fluids on Friction, TM
Beagley, IJ McEwen & C Pritchard BR Research,
November 1972 (page 121)
[4] TM TRIB 3 A Basic Study of Wheel/Rail Adhesion 6 - Adhesion
and Contamination on BR Track, M Broster & DA
Smith, BR Research, May 1973 (page 127)
[5] TM TRIB 4 A Basic Study of Wheel/Rail Adhesion 7 - Laboratory
Studies of the Effects of Water, TM Beagley & C
Pritchard, BR Research, 1973 (page 129)
[6] TM TRIB 5 A Basic Study of Wheel/Rail Adhesion 8 - Laboratory
Studies of the Effect of Solid Debris, TM Beagley, M
Broster, IJ McEwen & C Pritchard, BR Research, 1973
(page 130)
[7] TM TRIB 6 A Basic Study of Wheel/Rail Adhesion 9: Further
Observations On Leaves, C Pritchard & MA Tanvir, BR
Research, 1973 (page 122)
[8] TM TRIB 7 Adhesion and Leaves: Wet Abrasive Blasting Tests on
Southern Region, IJ McEwen & RK Taylor, BR
Research, June 1975 (page 123)
[9] TM TRIB 8 A Basic Study of Wheel/Rail Adhesion 10 - The
Rheology of Rail Contaminants Causes Low Friction,
TM Beagley, BR Research, May 1974 (page 131)
[10] TM TRIB 9 A Twelve Month Study on a 280 km Route, SG
Watkins, BR Research, June 1976 (page 124)
[11] TM TRIB 27 Track Testing of Adhesion Modifying Fluids, RK Taylor,
BR Research, May 1978 (page 132)
[12] TM TRIB 29 Assessment of The Cleaning of Leaf Affected Track on
Southern Region – 1977, RK Taylor & ID Pollicott, BR
Research, July 1978 (page 134)
[13] TM TRIB 30 Description of Trackside Water Spray Equipment, RK
Taylor, BR Research, August 1978 (page 126)
[14] TM TRIB 35 Assessment of The Cleaning of Leaf Affected Track on
Southern Region – 1978, RK Taylor & ID Pollicott, BR
Research, July 1979 (page 135)
[15] TM TRIB 36 Wheel/Rail Adhesion – Review of the Effectiveness of
Trackside Applicators, RK Taylor, BR Research,
November 1980 (page 136)
[16] TM TRIB 39 Brakes and Wheel/Rail Adhesion, C Pritchard, BR
Research, 1980 (page 137)
[17] TM WRI 5 A Comparative Study of The Trapping of Leaves by
Different Multiple Units, RK Taylor, BR Research,
March 1982 (page 143)
[18] TM TAG 87 Assessment of Proposals for Controlling the Autumn
Leaf Problem, E Rushton, BR Research, June 1983
(page 145)
[19] TM VTI 017 Adhesion and Leaves – A Review of the Problem and
Potential Solutions, TG Pearce & DJ Watkins, BR
Research, October 1987 (page 141)
[20] VDYN 162 Autumn Leaf Remedies – The Future Of, TG Pearce,
BR Research, March 1990 (page 140)
[21] TM ES 073 Computer Simulation of the Traction Systems Used in
the Class 47 SEPEX Adhesion Trials, S Quick, BR
Research, May 1986 (page 146)
[22] RT/DLA/0482/1 Influence of Leaf Mulch Contamination on Railhead
Adhesion, Hyder Environmental, 1997 (page 68)
[23] TM-MSU-408 Rail Contamination and It’s Influence on Wheel/Rail
Adhesion, IJ McEwen, BR Scientifics, January 1999
(page 69)
[24] TASS 1996-04-10 Track Adhesion Southern Sweden, Banverket, 1996
(page 72)
[25] B 164 Adhesion During Braking and Anti-skid Devices, ORE,
Draft Report No 1, September 1985 (page 75)
[26] - The Tribometer Train – A Brief Description, BR
Research, July 1996 (page 80)
[27] - Centrac HPF, Century Oils Limited (page 81)
[28] - Review of the Use of Chemicals in the Management of
Low Adhesion, Ian J McEwen, April 2003 (page 82)
[29] BK2050/SB253/01 EMU Wheelslip Control Study, D Pollard, WS Atkins
Rail, Issue 1 January 2000 (page 85)
[30] L/158LEAF2 Track Circuit Actuation Steering Group, J Mummery, A
Nelson, Regional Railways, 11 February 1993 (page
86)
[31] PRJ/10491-M81/RP011-JRW SPADRAM Adhesion Monitoring Using 319172, JR
Waring, BR Research, Issue 1 February 1995 (page
89)
[32] LR/DS/92/030 Autumn Leaf Fall: A Guide to Identifying Problem
Locations, W Poole and J B Feeney, BR Research, 1
December 1994 (page 91)
[33] VA-1467A Class 159 Unattended Data Capture System, R J
Hobson, Network Train Engineering Services,
November 1995 (page 90)
[34] RR-TRS-98-079 LAWS 1997/98 Trial, S Kent, AEA Technology Rail,
Issue 1 May 1998 (page 93)
[35] AEATR-T&S-2001-060 Low Adhesion Driver Communications Study, S Kent,
AEA Technology Rail, Issue 1 June 2001 (page 92)
[36] RR-TRS-99-037 LAWS Extended Trial ‘98, S Kent, AEA Technology
Rail, Issue 1 May 1999 (page 94)
[37] RR-TRS-99-103 LAWS On Train & Railhead Sensor Trial – April ‘99, S
Kent, AEA Technology Rail, Issue 1 October 1999
(page 97)
[38] RR-TRS-99-245 LAWS Trial ‘99, S Kent, AEA Technology Rail, Issue 1
January 2000 (page 95)
[39] PRD-60468-DOC-011 Comparison of LAWS Data With Moisture Sensor
Results, S MacDonald, AEA Technology Rail, Issue 2
20 November 2003 (page 98)
[40] - ERTMS and Low Adhesion, J Tunley & K Schofield,
2003 WCRR Paper (page 99)
[41] AEATR-VTI-2003-015 The Feasibility of Adhesion Control and Monitoring for
ERTMS, W Poole, AEA Technology Rail, Issue 1
March 2003 (page 175)
[42] AEATR-VTI-2004-011 Application of ETAMS to a Conventionally Signalled
Railway, SD Brown, AEA Technology Rail, March 2004
(page 101)
[43] - Vortok Leafguard Performance Autumn 2000, Vortok
International (page 88)
[44] AEATR-T&S-2001-152 An Assessment of the Effectiveness of Vortok
Leafguards, W Poole, AEA Technology Rail, Issue 1
July 2001 (page 87)
[45] LR-MEP-021 Part 5 Adhesion Modification Tests Autumn 1993 Part 5:
Lessons Learned, JR Waring, BR Research, March
1994 (page 161)
[46] - The Tribometer Train – A Brief Description, BR
Research, July 1996 (page 80)
[47] RR/TRS/96/062 Calibration of The salient Systems Portable Tribometer,
W Poole, BR Research, Issue 1 15 October 1996 (page
104)
[48] - Seasonal Preparedness - Adhesion Monitoring Salient
Systems Tribometer Usage, R Brookes, Ove ARUP &
Partners Ltd, Draft 1 12 April 2004 (page 107)
[49] RR-TRS-98-004 An Assessment of Low Adhesion Railhead Treatment
Methods Using Data From Service Vehicles, W Poole &
JD Tunley, AEA Technology Rail, March 1998 (page
110)
[50] - Autumn 2003 Monitoring - Deep Dive Report, R
Brookes, Ove ARUP & Partners Ltd, Draft 1 14 March
2004 (page 115)
[51] 113912/1/9999 Autumn 2003 Monitoring - Trials and Monitoring, I
Bruce Ove ARUP & Partners Ltd, Draft 1 25 March
2004 (page 111)
[52] 01 T353 Background Information for T353 Strategic VTI Model,
RSSB, 2003
[53] RR-SAM-016 Four Car Model and Train Braking Control, PD Horne,
BR Research, March 1995 (page 170)
[54] - Class 390 Brake Tests, AEA Technology Rail, 2003
[55] - Private communication, J Tunley, AEA Technology Rail, 21
April 2004
[56] - Dual Purpose Sandite Products, Dr IJ McEwen, January
2004 (page 78)
[57] AEATR-T&S-2002-134 High Pressure Water Jetting Trials, AEA Technology Rail,
August 2002 (page 159)
[58] LR-AER-042 Full Scale Leaf Tests at Old Dalby – Nose End Experiments
with Classes 156, 158 and 165, ND Humphreys & DW Read,
BR Research, 28 April 1992 (page 165)
[59] MB01183-07 Further Analysis of the Full-Scale Leaf Tests at Old Dalby
with Classes 156, 158 and 165, ND Humphreys, BR
Research, June 1992 (page 167)
[60] LR/DS/93/015 Report on the Leaf-Deflector Tests Conducted with a Class
165 Turbo Unit Between Princes Risborough and Bicester,
W Poole, BR Research, March 1993 (page 168)
[61] AEATR-VTI-2003-030 An Overview of Low Adhesion Factors for ERTMS, W Poole,
AEA Technology Rail, Issue 1 July 2003 (page 177)
[62] LR/VDY/92/087 An Investigation of the Effect of Wheel Load on the
Persistence of Railhead Leaf Film, W Poole, BR Research,
May 1992 (page 162)
[63] LR-MEP-021 Part 4 Adhesion Modification Tests Autumn 1993 Part 4: Centrac
HPF Results, JR Waring BR Research, February 1994 (page
160)
[64] RR-TRS-98-064 An Overview of Techniques For The Measurement of Wheel
on Rail Adhesion, W Poole AEA Technology Rail, March
1998 (page 105)
[65] - Wheel/Rail Adhesion and Analysis by Using Full Scale Roller
Rig, W Zhang, J Cheng, X Wu, X Jin, Wear, Vol. 253, pp82-
88
[66] - A Study of Adhesion Between the Rails and Running Wheels
on Main Lines: Results of Investigations by Slipping
Adhesion Test Bogie, K Nagase, 1989, Proceedings of the
IMechE Part F, Journal of Rail and Rapid Transit, Vol. 203,
pp33-43
[67] - Guidelines to Best Practice for Heavy Haul Railway
Operations: Wheel and Rail Interface Issues, The Heavy
Haul Association, 2001
[68] - Friction Control Between Wheel and Rail by Means of On-
Board Lubrication, M Tomeoka, N Kabe, M Tanimoto, E
Miyauchi and M Nakata, Wear, ,Vol. 253 pp124-129
[69] REP/112880/001 Autumn Performance Improvement Trials – March 2003
Bicester Tests, Ove ARUP & Partners Limited, June 2003
(page 117)
[70] 113912/1/9999 Autumn 2003 Monitoring – Trials and Monitoring, I Bruce,
Over ARUP & Partners Limited, draft 1 25 March 2004
Appendix 1 Questionnaire Report CRF04002
Issue 1
14 May 2004

LOW ADHESION FORMATION - How quickly in terms of time and distance does adhesion change? How long does the
adhesion last for? And how localised is it?
• Have you conducted any Tribometer train
track measurements of adhesion with respect
to time and distance?
• Have you conducted on track WSP testing that
would show this effect?
• Have you conducted monitoring (e.g. using
data recorders) of Adhesion events/incidents
at a specific location(s) over a time period that
would demonstrate this effect?
• Have you conducted railhead friction
monitoring and sampling at specific sites that
would show this effect?
LEAF CONTAMINATION PROPERTIES AND CHARACTERISTICS - What is it about the surface chemistry and physics of the
leaf contaminant “Teflon-like” coating, which means that it adheres to the rail, yet causes wheels to slide over it? What are the
friction properties related to the leaf contaminant thickness? Does water soften it such that there is an opportunity to agitate
it off the rail?
• Have you conducted any analysis of the
composition of leaf contamination, its make-
up, properties and characteristics?
• Have you conducted any analysis and/or
testing of the interaction of wheelsets on leaf
or other contaminants?
• Have you conducted any analysis and/or
testing of the effects of leaf contaminant
thickness on the coefficient of friction?
Appendix 1 Questionnaire Report CRF04002
Issue 1
14 May 2004

• Have you conducted any analysis and/or

testing of the effect of water on railhead leaf
contaminant and ease of removal?
• Are you aware of the application of other
materials used to remove the leaf contaminant
by their softening action?
FRICTION MODIFIERS - Can friction modifiers (other than Sandite) be used to mitigate the effects of low adhesion? If so, how
effective would they be and how would they be used?
• Have you conducted any analysis and/or
testing of the use of materials designed to
increase friction on low adhesion for traction
and/or braking, e.g. Centrac HPF?
• Are you aware of the use of any such
materials to raise adhesion levels?
VEHICLE DESIGN INFLUENCES - What are the critical factors of vehicle characteristics that improve or worsen adhesion
performance? (e.g. suspension, wheels, traction and braking system design)
• Have you conducted any analysis and/or
testing of vehicle suspension characteristics
and their effect on traction and braking
performance under low adhesion conditions, to
determine the effect of train accelerating and
braking?
• Have you conducted any analysis and/or
testing of wheel profiles and/or wheel
materials and their effect on braking and/or
traction performance under low adhesion
conditions?
Appendix 1 Questionnaire Report CRF04002
Issue 1
14 May 2004

WHOLE TRAIN BRAKING/ACCELERATION CONCEPT - Are there benefits in varying the braking and traction demand along a
train to maximise the use of available local adhesion and thus take advantage of the scrubbing action from the front wheels?
• Are you aware of any rolling stock that applies
this concept to braking and traction control?
• Have you conducted any analysis and/or
testing of individual vehicle/bogie/wheelset
sensing of adhesion and maximising of brake
and/or traction forces to match available
adhesion, increasing it at those that can
compensate for others on lower adhesion
conditions?
ADHESION PREDICTION & COMMUNICATION - Can current manual adhesion prediction models (e.g. Central Line ACCAT) be
automated to reliably predict the occurrence of low adhesion in real-time to minimise disruption and if so how? What would
such a model look like and how should it be developed? How could this be communicated to the train driver? How would it
link into the VTI Model (Project 1)?
• Are you aware of any applications of low
adhesion predictive models, either manual or
automated?
• What input systems are you aware of that
could contribute to an automated low adhesion
prediction tool?
• What methods are you aware of for
communication of predicted low adhesion data
to train drivers?
 Report CRF04002
Issue 1
14 May 2004

Review of Current Adhesion Knowledge

Appendix 2

Library Search Results

Adhesion Working Group

CR Fulford Associates 65
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Influence of Leaf Mulch Contamination on Railhead Adhesion
Report Author: Hyder Environmental
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: RT/DLA/0482/1
Issue and Date: 1997
Summary:

Reports on railhead contaminant sampling and analysis feasibility study conducted for
Merseyrail Electrics in October 1997. Sample swabs were analysed for oil, iron and
bacteria. Material scraped from the railhead was analysed for plant pigments, oil, iron and
cellulose. Sampling included following a number of safety incidents (both SPADs and
station overruns).

Some problems were encountered in the process of chemical analysis.

Test for pigment showed some success identifying 2 major types of leaf.
The process of analysing for oil proved successful but showed little oil content.
Iron measured presumed to reflect rust.
Bacterial measurements showed a wide range of colony count but did not allow the clear
identification of the causes of low adhesion.

Conclusions:

• Simple analytical techniques to determine railhead contamination were achieved

with respect to oil, iron, leaf mulch and bacteria, but further characterisation and
development was required.
• In some cases it was possible to unequivocally demonstrate the presence of leaf
mulch on the railhead
• In most cases sufficient material could be scraped off the railhead for sampling.
• Bacterial counts proved useful in providing additional information about the nature
of railhead contaminants

Further development of the techniques is required to establish their robustness was

recommended but no further work was undertaken.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Rail Contamination and It’s Influence on Wheel/Rail Adhesion
Report Author: IJ McEwen, BR Scientifics
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Friction Modifiers
Report Number: TM-MSU-408
Issue and Date: Issue 1, January 1999
Summary:

Provides a historical review of research, development and service practices relating to rail
contamination and its effect on wheel/rail adhesion.

Adhesion is unpredictable and varies widely, is strongly influenced by the weather – rain
temperature etc, adhesion is affected by contamination and can fluctuate rapidly from site
to site.

Outlines sources of contamination:

• weather influences adhesion, i.e. the presence of water which is chiefly
responsible for this influence, either as different degrees of precipitation
(heavy/light rain, snow, drizzle etc.) or combined effects of temperature and
humidity to trigger condensation on the rail. Rain on dry rails may reduce adhesion
to 0.2 to 0.25. In certain low adhesion conditions rain can increase adhesion.
Water present in small amounts on the rail is a major influence. Light drizzle, on a
dewy, misty morning or as wet rails dry out. Adhesion can fall to below 0.1 on
otherwise clean rails and with over night rusting down to 0.05. The formation of
small amounts of moisture or damp rails with rust or some other solid debris
presents problems. Low adhesion depends on there being critical amounts of
water present which may be short lived, the passage of only a few axles being
sufficient to disturb the equilibrium. Rain can affect significant lengths of track as
can dew or other condensation effects. There are situations where the effect can
be very localised, e.g. in short cuttings producing micro-climates, over spray from
the sea or industrial/agricultural processes (cooling towers, crop spraying etc.). Ice
and snow has the same effect as its is melted by the action of passing wheels.
• Oily matter on average 5 molecules thick is found on the railhead.
• Grease from badly adjusted flange lubricators
• Fuel oil and kerosene
• De-icing fluids
• Rust combined with water
• Leaves
• Sawdust
• Solid particulate e.g. coal dust
• Salt

Refers to an experiment with an 8 car multiple unit which picked up as many as 60% of
the leaves in the 4 foot and deposited them on the rails. The leaf like character to create a
Appendix 2 Library Search Results - Adhesion Working Group

black thin continuous film on the wear band. The film is hard and difficult to scrape off the
rail. The onset of light rain or drizzle results in very low adhesion down to 0.02. Heavy rain
will soften the film and passing wheels can then remove the film as quickly as it formed.
Heavy rain also helps prevent leaf build up as soggy leaves are less mobile and leaves do
not stick well to wet rails.

Refers to work undertaken by BR (Wheel/Rail Adhesion – Boundary Lubrication by Oily

Fluids, M Beagley, I McEwen & C Pritchard), Wear 31 1975 page 77) which identifies that
the oily residue on the rails has several hundred chemical components.

Outlines that organic matter (e.g. leaf) undergoes physical changes under high wheel
pressures which can be visually seen. But petrochemical contamination will change
rapidly by oxidative processes when spread out as a thin film.

Considers that the mechanisms for lower adhesion are few, are complex and in some
cases the true mechanism is not known. For oily contaminants a Boundary Lubrication
Layer is formed creating a very thin film between the 2 surfaces of tenaciously held active
lubricant molecules. Does not lead to low adhesion problems as surface roughness of
wheel/rail will cut through layer. Oily matter squeezed from leaf material and carried down
the track will have a similar influence on adhesion as petrochemical based oils.

Gross water – the mechanism by which gross water on the rail influences adhesion is still
not fully understood but the result is known. Water appears to act as a ‘weak lubricant’ in
that it reduces adhesion when it is initially high but not to the same level as an oil-based
lubricant. Water is readily pushed aside by passing wheels and does not form a boundary
layer like oil due to the molecule size. I McEwen’s view is water promotes surface
oxidation of the steel rail such that exceedingly thin oxide films protect the surface from
metal to metal contact, thus seeing oxide/oxide friction levels and not steel/steel.

The Paste Effect – rust (or other solid eg. Coal, iron ore, clay) and water. On thickly rusted
rails adhesion coefficients as low as 0.03 have been measured in drizzle. The proportion
of water to solid matter is critical. Very low adhesion occurs as the paste moves from
viscous liquid behaviour to that of a solid (or vice versa). This is more common on little
used track or after a period of little traffic (the Monday morning effect). Also more common
at certain times of the day when humidity effects may give rise to condensation on the rail.
Humidity must be expressed relative to rail and not air temperature. As well as morning
dew, condensation can occur when the rail temperature falls, e.g. in a cutting, or the local
humidity is high, e.g. a wooded area.

The effect of leaf extract can be seen many miles away from the source of the leaves,
being carried down the track by the wheels. Leaf process can also be seen as a paste
process with water.

Neither the mechanism of lowering adhesion by particulate solid/water pastes nor that by
leaf films can be described in Classical terms. In all probability the former shares much
with elastohydynamic lubrication and the latter with the full film hydrodynamic lubrication.
However, both of these mechanisms can only be described mathematically assuming
defined viscosity properties of the single-phase lubricant. In the cases of pastes or leaf
films on the rail the lubricant is a solid/liquid mixture of varying proportions with very
complex non-Newtonian flow characteristics.
Appendix 2 Library Search Results - Adhesion Working Group

Also reviews Friction Improvers – Stick Products – a range of polymeric materials in stick
or rod form described as friction modifiers. The low coefficient of friction product (LCF) for
reducing flange wear. High Positive Friction (HPF) versions are available claimed to
increase wheel rail friction. Have been successful to counter rail corrugation formation but
not sold as adhesion improvers. Little documented evidence of quantified friction
(adhesion) improvement.

Track water sprays – refers to tests with automatic water sprays. A central feed line with
fine spray jets in the 4 foot spraying at regular intervals or when a detector dried out.
Based on agricultural irrigation equipment and can include frost protection as well as a
number of other controls. Keeps leaves on ballast from drying out and prevents leaves
sticking to rails. Prototypes over several 100 m of track proved effective but cost of initial
set up found no real usage.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Track Adhesion Southern Sweden
Report Author: Banverket
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Adhesion Prediction & Communication
Report Number: TASS 1996-04-10
Issue and Date: 1996
Summary:

Reports on the Track Adhesion Southern Sweden (TASS) Project run by Banverket (BV)
Swedish National Rail Administration during 1995, to develop a strategy and methods to
combat leaf fall slip.

During the project BV:

• Put into operation 2 new equipments for combating slip
• Developed and tested a measuring carriage for friction measurements
• Carried out extensive measurement of friction, atmospheric humidity, rail and air
temperature over a 6 week period
• Compared driver reports and Traffic Management reports on slip with weather
statistics to investigate the possibility of a connection
• Gathered and analysed samples of substances found on railheads where leaf fall
slip was experienced
• Carried out tests with manual track preparation to simulate leaf-fall slip

Railhead friction measurements were taken several times a day along the untreated and
treated rails.

Ecological Institute in Lund analysed many railhead samples for chemical and biological
content. In general plant material was found. Connection between Nitrogen content and
plant material. After a number of trains had run the plant material had become charred.
There was a possibility that in farming areas the slip problem could be caused by airborne
nitrogen compounds (eg. Ammonia), in which case the considered the track film could be
bacteria or algae. The collected material contained less concentrations than growing
material, possibly explaining a dilute mixture of plant material in a substance which was
otherwise dominated by rust and graphite grease.

The macro-study showed on certain stretches friction was always lower than on others
and there are also ‘slippy spots’.

They looked for a connection between friction and atmospheric humidity, rail and air
temperature. They found a clear influence of atmospheric humidity. An almost linear
relationship between humidity and reduced friction, but noted that friction could be high
when humidity was high (CF as found by AEAT). They found no connection between
friction and rail or air temperature in the macro study but in some case studies rail
Appendix 2 Library Search Results - Adhesion Working Group

temperature may have had an effect. .

Continued research on slip and the factors involved would create the opportunity for
forecasting track conditions – developing a ‘slip warning algorithm.

The use of lime on the track to improve adhesion was suggested by drivers who
experience improved stopping where lime had come onto the track. The report identifies
successful tests in Denmark using lime.

They plotted friction levels over time and distance to create a 3-D map of how adhesion
varies over time and distance. They identified rapidly alternating friction values as well as
progressive reduction and increase. The report suggests measurements should be made
at 4 hour intervals. There was no analysis of what caused the variations in friction plotted.

They found that friction can vary considerably locally, but it does seem to change in a
similar way over a large geographical area. In an area where drizzle is falling friction can
fall to a very low level. It was demonstrated that high humidity could lower friction also.

Black spots and the film on the railhead are characteristic of autumn and falling leaves.
One theory was that the leaves act as a catalyst and together with other components
produce a coating which reduces friction with similar qualities to Teflon.

Recommended future work to:

• Develop slip forecasts
• Continued development of friction-measuring equipment
• Continued trials with track preparation and friction measurements
• Tests on different anti-slip agents
• International co-operation in studies and development of methods
• Further development of rolling stock
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Investigations by Banverket in Sweden into Methods of Rail Cleaning
Report Author: JD Tunley, BR Research
Subject Area: General
Report Number: RR-SAM-137
Issue and Date: January 1996
Summary:

Reviews findings of Banverket’s project to test different methods of rail cleaning. No

significant results were available at a presentation in Sweden.

Banverket used a Salient Systems Inc portable tribometer. Banverket considered the
tribometer was giving inaccurate (high) readings.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library (CF material)

Report Title: Adhesion During Braking and Anti-skid Devices
Report Author: ORE (ERRI)
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: B 164
Issue and Date: Draft Report No 1, September 1985
Summary:

A synthesis of current knowledge concerning adhesion.

Identifies parameters which might affect adhesion:

Macrogeometric state of surfaces in contact:

• System of braking (disc, cast-iron blocks etc.)
• Regeneration of adhesion by sustained slip
Pollution of the rail:
• Effect of traffic density
• Influence of mode of traction (diesel or electric)
• Environment
• Braking on wet dead leaves
Atmospheric conditions:
• Rain, drizzle, fog, snow, hygrometric conditions or air, temperature of rails
Vehicle speed
Specific pressure at wheel/rail contact point
• Instantaneous vertical load and effect of load transfer: Axle load, weight of body,
weight of bogies, geometry of body, geometry of bogies, deceleration of vehicle
• Surface of contact zone: diameter of wheels, wheel profile, wheel hardness, gauge
of running surfaces, railhead profile, rail hardness, track curvature.

The amount of rail contamination is usually small (Refs 9,10,26,27,28,29,30,31).

Measures to combat the effect leaves have been examined (Refs 10,11,12,30,31). On
thickly rusted track adhesion as low as 0.03 has been recorded in light rain (Refs 6, 7).

Identifies there is little evidence that environment has any effect on adhesion except
where there is obvious contamination. Most serious contamination is autumn leaves
aggravated by humidity (Refs 7,9,10,11,12,16,26,30,32,35). Rails may become covered
by a layer of leaf pulp 15 to 50 microns thick (Ref 12). Slight rain can reduce adhesion to
0.03 (Ref 9). Continuous rain softens the leaf pulp so that it is readily cleaned from the rail
by the action of passing wheels (Ref 12). Of all factors adversely affecting adhesion it is
unanimously agreed that dead leaves accompanied by humidity results in the lowest
adhesion values.

Atmospheric conditions:
Appendix 2 Library Search Results - Adhesion Working Group

• rain and drizzle – low adhesion is nearly always associated with wet rails, the
lowest values when light rain or condensation humidifies the rail surface.
• Air humidity and air temperature – most observers associate high humidity with
low adhesion. Measurements in traction suggest there is a sharp reduction when
relative humidity exceeds 85% (ref 4). Measurements on rails using a small
tribometer indicate a more gradual correlation (ref 16). It is considered important
that humidity is expressed relative to rail rather than air temperature, i.e. relative
humidity at rail determined from water vapour pressure in the air relative to the
saturated vapour pressure at the rail temperature (Refs 16, 17). Low rail
temperature is consequently associated with low adhesion (refs 14,32).

Vehicle speed:
• The evidence suggests that the reduction in adhesion with speed is largely due to
the vehicle suspension and to dynamic interactions between wheel and rail (Ref 1)

Specific pressure at wheel/rail interface:

• It has consistently been found that adhesion is reduced at higher axle loads (Refs
1,3) doubling the axle load reduces adhesion by about 0.05. Increasing axle load
also reduces the amount of slip at which adhesion was a maximum (Ref 1).
• No reports were found on the effects of the body or bogie mass or of the rate of
deceleration of the vehicle.
• Ref 20 identifies the friction coefficient to be in the same scatter range for axles
between 9t and 15.5t

Dimensions of wheel/rail contact ellipse:

• An increase in axle load leads to a reduction in adhesion, therefore a change in
wheel profile, or railhead profile, leads to a reduction in contact ellipse dimensions
and hence increase in load. SNCF tests showed a reduction in adhesion from 0.66
to 0.054 with different wheel profile.
• Also considered likely that dimensions of the contact ellipse change as a function
of the hunting of the wheelset.

Track curvature:
• Report B 44/RP 10 shows a general tendency for adhesion to reduce as the curve
radius decreases but not significant until less than 670m

Other considerations:
• Identifies little effect of wheel diameter
• Little effect of track vertical irregularities.
• Track horizontal irregularities cause bogie oscillation and involve some lateral slip
and reduce available longitudinal adhesion, which increases with speed.

References:

1) ORE B 44/RP14 Synthesis Report: the current state of knowledge about

wheel/rail adhesion, 1978
3) ORE 44/RP 11, Adhesion Tests of Autumn 1972, 1975
Appendix 2 Library Search Results - Adhesion Working Group

4) ORE B 64/RP 13, Adhesion Tests of Spring 1975, 1977

6) C Pritchard, Traction Between Rolling Steel Surfaces. A Survey of Railway and
Laboratory Experiments’
7) C Pritchard, Brakes and Wheel/Rail Adhesion, IMechE, Sept 1979
8) SG Watkins, A 12 Month Survey of a 280 Km Route, BR Report TM TRIB 9, 1976
9) TM Beagley, Adhesion Survey on Southern Region (Witley Bank), BR Report TM
TRIB 6, 1974
10) DB Report, Investigations into Adhesion at High Speeds for Vehicles Running
Over Rails Artificially Contaminated with Leaves, 1977
11) IJ McEwen, Adhesion and Leaves: Use of Sandite, BR Report TM TRIB 22, 1977
12) RK Taylor & ID Pollicott, Assessment of the Cleaning of Leaf Affected Track on
Southern Region – 1978, BR Report TM TRIB 35, 1979
14) H Bugarcic, H Ahlbrecht & O Lange, Optimisation of Auxiliary Brake Gear on
Tramway and Rapid Transit Stock, Verkehr und Technik, 1982, 4, 20
16) M Broster, C Pritchard & DA Smith, Wheel/Rail Adhesion: Its Relation to Rail
Contamination on BR, Wear, 1974, 29, 309
17) D Benz, Study of the Tribochemical Influence of Air Pollution on the Rolling
Friction of Certain Materials, ZEV-Glasers Annalen, 1982, 106, (7/8), 273
20) A Metzkov, The Conditions of Adhesion Between Wheel and Rail During Braking,
Organ f.d. Fortschr d. Eisenbahnwesens, 1934, 89, 247
26) SNCF Test Report, VTU Coach on Dead Leaves – Comparison Composite Brake
Discs Only, 1983
27) M Boiteux, Braking of Railway Vehicles, 1982
28) SNCF, Investigations into Adhesion During Braking, Morceux Tests, 1981
29) Knorr Bremse test report on SJ WSP tests, 1981
30) DB Tests on Tracks Covered with Dead Leaves Using Block-Braked and Disc-
braked Vehicles (1975)
31) SNCF Tests – 1981, WSP Devices in Operation or Out of use. Emergency Braking
on Dead Leaves
32) DJ Watkins, Exploring Adhesion with British Rail’s Tribometer Train, Railway
Engineering Journal, 1975, 4, 6
35) OHNO Kaoru at al, Low Adhesion Between Wheel and rail Caused by Fallen
Leaves and its Remedy, RTRI Tokyo, Sept 1983
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Dual Purpose Sandite Products
Report Author: Dr IJ McEwen
Subject Area: Low Adhesion Characteristics
Report Number: -
Issue and Date: January 2004
Summary:

A summary of progress made with an investigation commissioned by the Adhesion

Working Group into dual purpose Sandite products. The original remit was simply to
discuss the feasibility and benefits of such products with Lawrence Industries Limited,
however the opportunity to carry out limited field trials was also taken whilst the leaf
season offered suitable test opportunities.

The very limited tests carried out showed that citrus cleaners offer significant benefits in
softening leaf films on the rail for later removal by some other means – brushing, water
flushing or even traffic action. They appear to soften the films more quickly and more
effectively than water alone. The prevailing weather conditions on the day of test and the
test site features, however, meant that the rails at the time of test were damp with
moisture and the leaf films were relatively soft. This made it difficult to quantify the rate of
softening and no difference could be determined between the three cleaning solutions
used.

The use of a dual purpose Sandite material containing d-limonene was proposed for a
number of reasons:
a) when used as a replacement for either of the proprietary citrus cleaners for softening
leaf films it might be expected to lessen the danger that a slippery rail condition would
result.
b) it might soften leaf films sufficiently such that traffic action alone would remove the
films, thus obviating the need to follow up citrus cleaner treatment with a protective
Sandite treatment.
c) the use of a proprietary dual purpose product would ensure that track personnel always
use the correct mix and thus give better control to leaf busting operations.

In the event, the limited track testing which was possible did not establish whether traffic
action would help to remove the contaminant films or not and there appeared little
advantage in using d-limonene mixes as against the currently supplied Tractiongel mix.
Brushing and rinsing under reasonably controlled conditions again showed little benefit in
the 25% d-limonene mixes. It could be argued that under different conditions with harder,
more tenacious leaf films the results might be different but this would need further testing
to establish fact.

It is true that a product formulated to contain d-limonene would cost more than current
Sandite products to manufacture and thus there needs to be demonstrable benefit in
using such products. The current tests were useful but in truth did not give the indications
needed.

One useful additional feature which came out of the site trial was the demonstration of
Appendix 2 Library Search Results - Adhesion Working Group

Lawrence Industries applicator trolley to feed a fine uniform layer of dry sand to the
railhead with control to stop and start the sand flow at walking pace. This would appear a
very useful treatment to follow up citrus cleaner application – the sand adhering to the wet
railhead giving protection of a sort against a slippery rail being left behind.

The following conclusions can be made from the investigations and limited track tests
carried out:

Citrus cleaners have been shown to be more effective in softening railhead leaf films than
water alone. There was no difference between the two common cleaners Natrusolve and
Orange Cleanse.
Stable Sandite mixes containing 25% d-limonene cleaners have been manufactured in
quantities suitable for hand application at specific leaf-affected sites.
Under the prevailing conditions of the site trials, i.e damp conditions and already soft leaf
films, there was no difference in the action of the mixes containing d-limonene and the
currently supplied Sandite products.
As a by-product of the field trials the Pencil Hardness Test showed itself to be a useful
and repeatable indicator of leaf film condition at any particular time or location.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: The Tribometer Train – A Brief Description
Report Author: BR Research
Subject Area: Low Adhesion Measurement
Report Number:
Issue and Date: July 1996
Summary:

This report provides an outline of the methodology employed in the BR Tribometer train

No specific input to this work but provides description of test technique.

Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library (CF material)

Report Title: Centrac HPF
Report Author: Century Oils Limited
Subject Area: Friction Modifier
Report Number: -
Issue and Date: -
Summary:

Sales material from Century Oils for Centrac HPF.

Indicates Centrac HPF is aimed at reducing high adhesion to reduce wheel wear and
railhead wear and reduce associated noise. HPF exhibits a positive friction characteristic
and has been designed to provide operation in a controlled friction range of 0.17 to 0.35
from an uncontrolled range of 0.1 to 0.6.

Details brake tests conducted by Skytrain Vancouver using Centrac Sticklube HPF by
Century Oils Limited. Tests concluded that there was no difference in the emergency
brake stopping distance between a dry rail, a wet rail and an HPF friction modified rail.

Does not address increasing friction at all.

Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Review of the Use of Chemicals in the Management of Low Adhesion
Report Author: Ian J McEwen
Subject Area: Friction Modifiers
Report Number: -
Issue and Date: April 2003
Summary:

Details the results of a review of chemical products and processes for adhesion
management and control.

Reports on chemical growth inhibitors, defoliants but no positive result.

Historically chemicals have not been seen as offering much potential as preventive
treatments for low adhesion. Some attention had been given to their use in cleaning the
railhead but usually combined with mechanical methods such as high pressure water
blasting. Instead, the railway put significant effort into finding suitable friction improvers.
This was a wide ranging investigation into chemical and organic processes or products
used in the rail and other industries to see if those in use could be improved and others
successfully applied on the railway to help manage low adhesion.
Dr McEwen conducted extensive literature surveys and interviews with key companies
specialising in the research and application of chemical processes. His main objectives
were to identify chemicals and processes which could be used as preventative measures,
or in the removal of contamination or as friction improvers.
Preventative measures
Two main processes were examined: defoliants and growth inhibitors. The usefulness of
defoliants is very much dependent on the tree species and condition as well as the
weather. It is really not possible to drop leaves ‘to order’ and there are significant safety
problems with spraying defoliants in an open environment. He found little experience with
large scale usage. Growth inhibitors also have spray application problems and safety
issues and the retardation of growth does not guarantee less leaves, and sometimes the
leaves become smaller but thicker.
Removal of contaminants
The review took in a number of cleaning methods including detergents and other cleaners,
Biological cleaning, Citrus based cleaners, ‘Leaf Busters’ and Sodium Bicarbonate
particles. Discussions were had with companies like Uniqema, Ellis & Everard, Bio-Wise,
Rozone, Demon International and Kelsan. Some of the discussions were still ongoing with
suppliers of detergents but those with suppliers and users of Biological cleaners
determined that these are nearly all used in a closed environment for cleaning
components and not suited to an open environment such as the lineside.

Citrus based cleaners are already in use on the railway for removing contaminants. This
includes products such as Natursolve and Orange Cleanse, both used for localised leaf
removal being based on D-limonene. Current disadvantages of these are the labour-
intensive application methods. Alternatives to citrus cleaners may exist in the form of
‘safe’ paint strippers, oven cleaners, other aggressive cleaners and possibly even
compost accelerators.
Appendix 2 Library Search Results - Adhesion Working Group

Friction Improvers
A number of potential friction improvers were identified in the review including Colloidal
Silica Fluids, Friction Modifiers, Sandite and derivatives, Kelsan HPF Sticks, Keltrack
Fluids and Tracglide Sticks.
Colloidal silica products were looked at by British Rail many years ago and are still
available but no particular advances have taken place with this technology. No new
materials were identified and proprietary products are already on offer for railway use.
Kelsan HPF Sticks are a friction/adhesion maintainer and are normally targeted at
controlling high friction. Dr McEwen found little published information regarding their effect
on contaminated rails. In long term use he felt they may override the influence of
contamination.
The effectiveness of Friction modifiers may hold clues for further research:
Keltrack Fluids is a fluid development of HPF friction modifier and has been formulated for
vehicle and track application onto the railhead. For best effect it needs its water carrier to
dry out. Again, this product has been targeted as a high friction controller rather than low
adhesion.
TracGlide System is a vehicle mounted liquid lubricant which does not affect braking. It is
laid as a very thin film which degrades leaving little residue. Once again it is aimed at
reducing lateral forces and saving energy for situations of high friction.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Review of Initiatives – Wrong Side Track Circuit Failures
Report Author: RM Troup
Subject Area: General
Report Number: AEAT-T&S-2000-101
Issue and Date: Issue 1, August 2000
Summary:

Summary of Track Circuit operating issues with respect to adhesion. Provides source
references for further material
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: EMU Wheelslip Control Study
Report Author: D Pollard, WS Atkins Rail
Subject Area: Vehicle Design Influences
Report Number: BK2050/SB253/01
Issue and Date: Issue 1 January 2000
Summary:

Reviews the wheelslip control systems or various EMU to determine the feasibility of
improving the traction control systems. Some systems are very basic and interrupt traction
when slip is detected. Proposed measures to improve the degree of control without
interrupting power providing a more appropriate control under low adhesion conditions.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Track Circuit Actuation Steering Group
Report Author: J Mummery, A Nelson Regional Railways
Subject Area: Vehicle Design Influences
Report Number: Director of Operations Reference L/158LEAF2
Issue and Date: 11 February 1993
Summary:

Reports on work undertaken during 1991 and 1992 to address the significant increase in
WSTCF involving Class 158 DMU.

Although it does not directly detail the interaction between wheel and rail with respect to
failure to activate track circuits, it details the results of experimental fitment of Auxiliary
Tread Brakes (ATB) to a number of 3 car Class 158 DMUs.

The results demonstrated an approximate 50% reduction in incidence of Class 158

WSTCF for trailer bogie fitted ATB only. A motor bogie modification was expected to add
a further 5 to 10% reduction.

Costs were seen as high for the limited effectiveness

Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: An Assessment of the Effectiveness of Vortok Leafguards
Report Author: W Poole, AEA Technology Rail
Subject Area: Vehicle Design Influences
Report Number: AEATR-T&S-2001-152
Issue and Date: Issue 1 July 2001
Summary:

The Vortok Leafguard test programme was designed to investigate the use of Vortok
Leafguards to prevent leaves from being rolled onto the railhead by passing trains.

Tests were undertaken between Romsey and Eastleigh and showed that the Leafguard
approximately halved the quantity of leaf that is rolled onto the railhead. Although there
was insufficient test runs to provide data for statistical analysis, the observed trends were
consistent and supported the evidence from service trials during the previous autumn.

The majority of leaves from the initial position – an area of 4 sleeper bays – are lifted by
the train and become airborne over a distance of between 50 and 100m with only a few
carried for greater distances at speeds of 50 or 60mph. Since the majority of large leaves
that were deposited on the railhead originated from the 4 foot, there may be design
modifications based on aerodynamic factors which would improve the Leafguard
performance.

Experience to date had not provided detailed information regarding the braking and
traction performance benefits but the indications from the reduction in railhead
contamination indicates that benefits should accrues from the use of the guards.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Vortok Leafguard Performance Autumn 2000
Report Author: Vortok International
Subject Area: Vehicle Design Influences
Report Number:
Issue and Date:
Summary:

This report is from direct feedback from users of Leafguard in Autumn 2000 which
includes Railtrack and Dutch railways. 34.5 miles of Leafguard had been supplied.

The report highlights some of the installation problems.

The reports claims from Railtrack suggest that Leafguard at Maxwell Park saw a 100%
reduction in train delays; a total elimination in train delays and WSTCF at Ingastone; no
train delays or WSTCF at Ashford.

Site visits to Wylde Green on Midlands Zone showed a build up of leaf residue on the non-
Leafguard up line, whilst the down line, fitted with Leafguard, remained clean. There were
no reports of train delays associated with the down line (the report does not state whether
any delays were associated with the up line either).

Railhead sample were collected at Manchester using tape which gave a visual
representation of the state of the rail. The Leafguard fitted rail appears to be cleaner.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: SPADRAM Adhesion Monitoring Using 319172
Report Author: JR Waring, BR Research
Subject Area: Whole Train Braking/Acceleration Concept
Report Number: PRJ/10491-M81/RP011-JRW
Issue and Date: Issue 1 February 1995
Summary:

The SPARAM project identified that one way to mitigate some SPADs is to reduce the
extended stopping distances under low wheel/rail adhesion. This report details tests using
a BR Mark 2 Advanced Flywheel WSP and a DEMON monitoring system to establish the
degree of improvement due to the rail conditioning effect of wheels running in macro-slip
along the length of a 4 vehicle train.

Very limited data captured at the time of this report.

Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Class 159 Unattended Data Capture System
Report Author: R J Hobson, Network Train Engineering Services
Subject Area: Adhesion Prediction & Communication
Report Number: VA-1467A
Issue and Date: November 1995
Summary:

Describes an ‘unattended data capture system’ used to capture data on low adhesion
events on a Class 159 unit operating on South West Trains services. The data captured
included the braking performance under low adhesion conditions, the location of low
adhesion blackspots, all brake applications over 10 seconds duration which WSP activity
occurred, and the time histories of 21 representative incidents.

Data was transferred using PCMCIA memory cards and not remotely.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Autumn Leaf Fall: A Guide to Identifying Problem Locations
Report Author: W Poole and J B Feeney, British Rail Research
Subject Area: Adhesion Prediction and Communication
Report Number: LR/DS/92/030
Issue and Date: 1 December 1994
Summary:

Developed a systematic approach to categorising the lineside with respect to vegetation

type and density. Permits problem areas to be identified and categorised 1 to 5 in
increasing levels of low adhesion risk.

This process has been widely adopted within the industry and would provide an input to
an automated system of low adhesion prediction.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Low Adhesion Driver Communications Study
Report Author S Kent, AEA Technology Rail
Subject Area: Adhesion Prediction and Communication
Report Number: AEATR-T&S-2001-060
Issue and Date: 1 June 2001
Summary:

The AWG commissioned this study to investigate how adhesion related information could
be communicated to train drivers. The work was prompted by previous work sponsored by
AWG which demonstrated the potential of railhead moisture monitoring to determine the
likelihood of encountering low adhesion conditions.

The general findings of this study were as follows:

• It is technically feasible to construct a trackside warning system, based either on

in-cab indication or trackside signage.
• There is a lot of existing technology that could be employed toward this end and
this should keep development costs down, whilst ensuring reliable operation.
• Trackside signage based systems already exist and could easily be modified to be
automatically activated by a railhead moisture sensor assembly. However,
hardware and installation costs are high (providing sufficient power is a significant
issue) and there are a number of potential maintenance considerations.
• An in-cab indicator is potentially more appealing, due to the relative ease of
making the necessary trackside installations and low power requirement. In
addition, modern communications and computing equipment is become physically
smaller and cheaper to buy, although this needs to be offset against the additional
cost and effort required to install equipment in vehicle cabs.
• For both trackside and in-cab systems, there remain a number of human factors
and failure mode issues that require further investigation and/or consideration.
• The need for such systems is likely to increase with the move towards more
advanced signalling systems which rely on assumed braking performance levels
for their correct operation.

It was recommended that a trial be conducted of an in-cab warning system on a route

operated by LAWS equipped rolling stock and two suitable proposals were described in
the report. AEA Technology were also willing to work with existing suppliers of manually
activated low adhesion displays to interface their equipment to a railhead moisture sensor
assembly. Further work was also recommended to resolve human factor issues raised in
the report.

No further work was commissioned by AWG.

Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: LAWS 1997/98 Trial
Report Author: S Kent, AEA Technology Rail
Subject Area: Adhesion Prediction and Communication
Report Number: RR-TRS-98-079
Issue and Date: Issue 1 May 1998
Summary:

The report details how the LAWS system performed under trial during Autumn/Winter
1997/98 (fitted to a single unit) and suggested a number of ways of improving the system
which were later adopted.

The report details a methodology for interpreting and presenting data. The performance of
the digital communications system (GSM) was assessed in terms of reliability, speed and
data integrity.

The conclusions were that the GSM communications was of sufficient speed and reliability
and integrity. The trial also demonstrated that SMS should be fast and reliable enough to
provide an in-cab live warning system.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: LAWS Extended Trial ‘98
Report Author: S Kent, AEA Technology Rail
Subject Area: Adhesion Prediction & Communication
Report Number: RR-TRS-99-037
Issue and Date: Issue 1 May 1999
Summary:

Reports finding of extended trial of LAWS in Autumn 1998. Trial involved 4 Class 165/166
DMU operating in October to December. Data was gathered from 2 Weatherchex
installations also.

The trial focussed on braking and safety related issues and analysis was based on 28,000
+ braking events. The potential of LAWS to assist in improved targeting of track treatment
was demonstrated and the system showed potential to provide a similar level of insight
into performance related issues.

The main findings were:

• The main body of adhesion related problems occurred between 16th October and
end of November.
• The majority of slip events were of less than 50m in length but there were a
significant number of greater than 300m
• The average braking rate achieved was 4.3%g
• There was WSP activity for 0.65% of the total time spent braking
• Of the time with WSP< less than 3%g was available for 14% of the total time spent
slipping
• The analysis of braking step selection showed that Step 1 was selected for 89% of
the time, Step 2 for 9% and Step 3 and Emergency accounting for 2%.
• An average of 0.2kg of sand was consumed per unit per day, peak usage was
1.16kg per unit on 28 October
• Braking sites were clearly identified throughout the Thames Trains’ operating area
and were almost all in the vicinity of stations and junctions
• Low adhesion sites were also identified and compared with the Sandite/water
jetting schedule for the North Downs route

Conclusions:
• A reasonable level of coverage was provided by the 4 LAWS equipped units and
should be sufficient for making weekly revisions to track treatment schedules
• The risk of slip during the Autumn despite defensive driving) is about 2.5 times that
outside of the Autumn
• A significant amount of very low adhesion was encountered
• Data from LAWS demonstrated to provide a clear indication of where track
treatment is required.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: LAWS Trial ‘99
Report Author: S Kent, AEA Technology Rail
Subject Area: Adhesion Prediction and Communication
Report Number: RR-TRS-99-245
Issue and Date: Issue 1 January 2000
Summary:

Describes LAWS trial Autumn 1999. The work concentrated on establishing the potential
of on-train temperature and humidity measurement to determine slip risk through
improved understanding of the conditions which create low wheel rail adhesion.

2 weather stations were used and 3 LAWS equipped Thames Trains Class 165 DMU.
Main elements of analysis were:

• Relate slip and spin events on the train to measurements at the weather stations
• Using weather station measurements, establish correction factors for on-train
sensors
• Relate occurrence of slip and spin events recorded by LAWS over the Class 165
operating area to measurements from the on-train sensors.

Data gathered for one month (November 1999), main findings:

• On-train and trackside sensors indicate slip and spin events only occurred when
recorded ambient humidity exceeded 80%
• However, the ambient humidity levels remained above this level for 89% of the
readings taken by the weather stations and 85% on the on-train readings.
• The predicted formation of dew according to the dew point calculation method did
not provide a reliable indication of slip risk according to either on-train or trackside
sensors
• Very good correlation between occurrence of significant spin/slip events and
readings from the railhead moisture sensor installed adjacent to the Crowthorne
weather station
• In contrast to the measured ambient humidity, the rail moisture sensor was high for
only a limited time during the autumn period (38% of readings taken).

It was concluded that monitoring humidity levels alone is of limited value in predicting
when trains are likely to slip and that the calculated dew point method was not sufficiently
reliable for this purpose.

The railhead moisture sensor was considered to offer a reliable and useful indication of
the prevailing slip risk for trackside installations.

Re developing an in-cab slip risk indicator, monitoring ambient humidity alone could
provide a reliable indication of possible high slip risk conditions but it would report high for
Appendix 2 Library Search Results - Adhesion Working Group

most of the autumn even where slip risk was in fact relatively low.

The railhead moisture sensor was not suited to on-train measurements but was
considered a reliable indication of actual slip risk if a series of such sensors were mounted
trackside along problematic routes and transmit warnings to drivers by means of trackside
warning signals or cab radio.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: LAWS On Train & Railhead Sensor Trial – April ‘99
Report Author: S Kent, AEA Technology Rail
Subject Area: Adhesion Prediction & Communication
Report Number: RR-TRS-99-103
Issue and Date: Issue 1 October 1999
Summary:

An extended trial of the LAWS during Autumn 1998, extended into April 1999 to evaluate
the potential of

• Trackside railhead moisture point and temperature sensors

• On-train humidity and temperature sensors

The railhead temperature and moisture sensors were connected to a permanent weather
station and on train sensors to a Class 165 DMU. Data was gathered during April to
correlate readings on-train to trackside.. The most significant findings were:

• Railhead moisture as measured by the sensor was closely related to ambient

humidity levels, but there were a significant number of periods when ambient
levels were high but measured railhead moisture levels were low.
• It is possible to make accurate temperature and humidity measurements from a
moving vehicle (with a correction factor).

The work suggests it should be possible to provide a live indication of the prevailing slip
risk to a driver by monitoring ambient humidity levels using on-train humidity sensors.

Further work was recommended to determine the best position for the sensors and
whether prediction of dew point from the measurement of both temperature and ambient
humidity could provide a more accurate indication of prevailing slip risk than the
measurement of ambient humidity alone.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Comparison of LAWS Data With Moisture Sensor Results
Report Author: S MacDonald, AEA Technology Rail
Subject Area: Adhesion Prediction & Communication
Report Number: PRD-60468-DOC-011
Issue and Date: Issue 2 20 November 2003
Summary:

AEAT installed a number of moisture sensors around the London area to monitor railhead
moisture levels during the 2002 leaf fall season. It was expected that this data could be
used to establish a link between the level of moisture on the railhead and observed low
adhesion conditions. Low adhesion events recorded by LAWS equipped vehicles
occurring within a set distance of the moisture sensors were compared with the moisture
level recorded by the sensing station.

The amount of data available for comparison was limited for a number of reasons
therefore the experiment was unsuccessful in establishing a verifiable link between
moisture sensor data and low adhesion conditions. The report provides details of the
results.

Future work was recommended to be focussed on the Central Line work with Metronet
BCV Tail Limited.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: ERTMS and Low Adhesion
Report Author: J Tunley & K Schofield, 2003 WCRR
Subject Area: Adhesion Prediction & Communication
Report Number:
Issue and Date: 2003
Summary:

Identifies from BR Research Tribometer train results (Autumns 1992 and 1993):
• The majority of low adhesion sections measured were less than 200m long
• The frequency of low adhesion sections identified below 0.06 was around one third
of those below 0.09
• There were no low adhesion lengths longer than 400m where the adhesion was
continuously 0.04 or below.

Identifies that with ERTMS GSM-R radio will provide packet 44 data transmission 2 ways
an ERTMS train could use this
• allowing adhesion information to be transmitted to appropriate trains (those
approaching or within the low adhesion section), as a single data bit indicating high
or low adhesion operating mode should be selected.
• Refinements to an adhesion management system. An advanced warning of
changing adhesion could be transmitted and displayed in the cab as a predicted
adhesion level (adverse, severe, very severe) for a specified location.

Also proposes that the adhesion management system could benefit a non-ERTMS
railway:

The essential process is to monitor the factors that determine the wheel-rail adhesion
level and process this information to formulate a decision on the level of brake force
demand that can be sustained, which is then supplied to the train control system.

This strategy is developed by giving consideration to:

• INPUTS (automated assessment of the factors affecting adhesion),

• A DECISION MATRIX
• OUTPUT(S) (to ERTMS or other train control systems.)

The integration of LAWS, and other key adhesion related elements into an adhesion
management strategy require the necessary inputs to be defined.

• Railhead Contaminants: There is a need to understand the relationship between rail

surface appearance, the contaminants present and the prevailing adhesion level
• Vegetation: There is need to identify the criticality of the line side tree density, the tree
type, distance from the running rails. Systems to monitor leaf fall and the state of
leaves on the track need to be developed to provide input for low adhesion risk
Appendix 2 Library Search Results - Adhesion Working Group

assessment.
• Weather: The relative importance of general weather forecasts, local micro-climatic
conditions and railhead moisture needs to be better understood. Heavy rain is
believed to be beneficial since it prevents very low adhesion and this effect needs to
be taken into account within any adhesion management system. Also a greater
understanding is required of the factors that influence drying conditions at low
adhesion locations.
• Track: the influence of gradient, cuttings, number of tracks, operating speed, railhead
cleaning measures etc needs to be understood
• Rolling Stock: the influence of brake and traction demand levels, train length,
operating speed, aerodynamic design, the effect of adhesion improving measures
such as WSP and sanders, needs to be understood. Data about the current adhesion
status can be supplied by train-borne systems such as the Low Adhesion Warning
System, initially developed by AEA Technology Rail under BR funding.
• Human Factors: any related human factors that may adversely affect safety must be
understood and factored into the total system solution.

Decision Matrix
The above parameters will be input into a intelligent decision process, similar to some
aspects of the system developed by Infraco BCV, the Adhesion Controller’s Condition
Assessment Tool (ACCAT), The inputs will be modified by predefined weighting factors
and an assessment will be made on the ability of the railway to support normal braking
rates. The decision matrix will then provide adhesion warnings to the train control system
and to the adhesion mitigation processes. It will also predict when normal operations can
be resumed.

Output
Potentially, the output from the decision matrix would be used by splitting the route into
Risk Sections that are used to instruct the train management system on the appropriate
brake rate.
Appendix 2 Library Search Results - Adhesion Working Group

Source: AWG library

Report Title: Application of ETAMS to a Conventionally Signalled Railway

Report Author: SD Brown, AEA Technology Rail

Subject Area: Adhesion Prediction & Communication

Report Number: AEATR-VTI-2004-011

Issue and Date: Issue 1.1 March 2004

Summary:

This report provides a high-level review of the feasibility of applying the concept of a
European Train Adhesion Management System (ETAMS) to a conventionally signalled
Railway as an Adhesion Management System (AMS). In particular, the report considers
the aspects associated with developing the communication and control methodology for a
real-time AMS. The findings of the report indicate that application of a real-time AMS
system to the conventionally signalled railway is possible, utilising existing available
technology. The AMS will provide primary benefits in terms of more effective utilisation of
traffic along a route, a reduction in workload for signalling staff, and reinforcements of
good driving practice.

An AMS system which combines both predicted an reactive elements within its control will
provide the most effective strategy. In particular, the predicted elements of the AMS will
enable warnings to be dispatched prior to the passage of the first train of the day over a
particular section of low adhesion. In addition, reactive feedback provided by trains
operating over low adhesion sites will define very accurately the level of adhesion present,
the exact sites of location of low adhesion and any changes in condition. Additionally, the
reactive approach provides a method of defining low adhesion sites requiring mitigation
actions and the effectiveness of these actions over time.

A working system would rely primarily on satellite based positioning, supported by

additional positional information front inertial devices and the signalling system. A satellite
based positioning approach adopted should recognise future developments in this area
from ESA and the UIC in relation to application of satellite technology to railways. In
particular, this should consider EGNOS beacons and other Galileo satellite GNSS.
Selection of an appropriate approach will require further evaluation. In particular, the
accuracy of the GPS/GNSS selected and the survey of coverage along AMS operating
routes must be considered.

Communication between vehicles and a host computer will require a predictable, and
reliable network with good service levels with guarantees. The obvious candidate for this
application is GSM-R. If GSM-R is unavailable in the short or longer term, then
consideration should be given to a private mobile network, as this will offer levels of
functionality and performance required by AMS. The most likely candidate in this category
is the Transcomm Mobitex network. With more detailed testing and proving its may also
be possible to achieve a satisfactory level of service from a public network such as
GPRS/SMS, as this approach would potentially offer the lowest operational cost. It will be
Appendix 2 Library Search Results - Adhesion Working Group

necessary to review the service level agreements, and actual coverage applicable to the
rail network in more detail for all preferred options.

The development of a trial AMS is recommended. A trial will provide an opportunity to

evaluate most of the concepts described within the report. It would involve development of
a minimum of to Autumn seasons, and would be limited to a single route over a
manageable route mileage. A suitable communications bearer would be identified based
upon the recommendations of section 6.3 and the availability of the service at the time of
trial. It is not propose to use GSM-R for such a trial in view of the recent Network Rail
announcement regarding further delays in implementation. During the trial the positioning
approach adopted will rely upon a GPS/GNSS solution possibly supported by other
devices. Due to the amount of development required, the signalling interface, which is an
option for vehicle positioning and for the messaging strategy, is only likely to be a feature
of a wider area implementation. Following selection an assessment of the input factors
used for the decision-making process the total system will require optimisation over a 2 to
3 year period, although the adoption of a self learning software approach may minimise
this. It is recommended that the first year would involve shadow running of the system,
and it will include an opportunity to obtain comments on the operation of the system from
the parties concerned.
 Report CRF04002
Issue 1
14 May 2004

Review of Current Adhesion Knowledge

Appendix 3

Library Search Results

Network Rail

CR Fulford Associates 101

Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Calibration of The salient Systems Portable Tribometer
Report Author: W Poole, BR Research Limited
Subject Area: General
Report Number: RR/TRS/96/062
Issue and Date: 1 15 October 1996
Summary:

The BRR Tribometer Train was used to calibrate the small scale wheeled Salient Systems
portable tribometer. The results suggest that the Tribometer should not be used for
absolute measurement of the coefficient of friction due to poor correlation with the full
scale wheel results.

The report offered the following guidance if the equipment were to be used:

• It is recommended that railhead and weather conditions be noted when the

tribometer is used at site
• Between 25 and 75 (average 50) measures of coefficient of friction should be
made with some readings being taken for both rails
• Use of adhesion charts should be made to assess the likely adhesion level that
trains will experience

Recommended further research to try and resolve the problem of high coefficients of
friction measured by the portable tribometer.
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: An Overview of Techniques for the Measurement of Wheel on Rail
Adhesion
Report Author: W Poole, AEA Technology Rail
Subject Area: General
Report Number: RR-TRS-98-064
Issue and Date: Issue 1 March 1998
Summary:

A report produced for Railtrack detailing options for measuring low adhesion.

The available methods of measuring adhesion each have their own limitations so that they
should only be used for the purpose for which they were designed.

The Tribometer Train and load measuring wheels/axles are sophisticated tools for
measuring wheel-rail forces and require considerable expertise to operate, maintain and
analyse the measurements. Therefore, these methods are not ideally suitable for
purposes other than adhesion research.

Portable tribometers were considered more applicable to Railtrack's requirements for

adhesion measurement but they are restricted in value because of the small contact area
at the measuring interface. This means that adhesion values can be affected by surface
roughness, wear and penetration of contaminant films all of which can result in false
adhesion values. The portable tribometers at that time in use with Railtrack gave
unrealistic values of measured adhesion. If they were modified to make more accurate
measurements they may still only be valid for relatively high values (>0.1) of adhesion.
However, there is potential for the development of a larger scale portable device which
may overcome some of the limitations of existing designs.

Contaminant sampling may be used to give an indication of the likely range of adhesion
values which may be experienced on a section of track, depending particularly upon the
prevailing humidity. Intelligently gathered sample may, in some operational situations, be
the only method of assessing the likely adhesion conditions.

Where trains which are fitted with a train data recorder and modern WSP equipment are
involved in incidents, this equipment has the potential to allow accurate analysis of the
data to indicate the prevailing adhesion conditions.

LAWS can be used on in-service trains to provide a vehicle based monitoring system. The
present system gives an indication of low adhesion locations based on the detection of
WSP activity. This is invaluable for the identification of low adhesion sites for railhead
cleaning. However, there is potential to develop this system to give a measure of the
adhesion level under low adhesion conditions, thus potentially providing a tribometer on
every train.

It was recommended that Railtrack should, if possible, obtain improvements to their

portable tribometers. These could then be used for post incident assessments of adhesion
Appendix 3 Library Search Results - Network Rail

and comparison of adhesion remedies. Railtrack should also ensure that users of portable
tribometers understand the limitations of this method of measuring adhesion. As an
alternative, a device should be developed which more closely matches the full-sized
contact area and that could be fitted to a road/rail vehicle for ease of transport.

A record of conditions should be made when measuring adhesion with portable

tribometers; the weather and railhead conditions and any other factors should be noted.
With or without the availability of a portable tribometer, where local adhesion data is
required samples should be taken from the railhead and analysed and a recording
technique should be developed and used throughout Railtrack Zones.

To assist with post-incident investigation, Railtrack should encourage the fitment of

modern WSP systems and train data recorders to new and re-conditioned stock which,
amongst numerous other advantages, would facilitate the availability of adhesion data. A
development of the LAWS system for measuring adhesion should be investigated to give
on-line adhesion data from operational train services. An investigation of the relationship
between adhesion values and the brake cylinder pressure would be an appropriate
starting point.

The widespread adoption of LAWS could be conducted as a collaborative venture

between train operators and Railtrack to provide real time adhesion data for driver
training, low adhesion warnings and targeted railhead cleaning.
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Seasonal Preparedness - Adhesion Monitoring Salient Systems
Tribometer Usage
Report Author: R Brookes, Ove ARUP & Partners Ltd
Subject Area: General
Report Number: -
Issue and Date: Draft 1 12 April 2004
Summary:

Reports on background knowledge that should accompany the usage of the Salient
Systems portable tribometer, identifying advantages and drawback of its use.

Reports on tests at Woking line using several tribometers and several operators in a
random sequence of use. Tests established that the readings were independent of the
operator or the tribometer used.

Readings were in excess of 0.5µ for clean rail and 0.25µ on oiled rail (CF note suggests
absolute levels are high).
Observations:
• Tribometer readings should not be relied on for low or high values of co-efficient of
friction. Values lower than 0.15µ and higher than 0.65µ must be regarded with
suspicion.
• Rail appears to be clean when the reading exceeds 0.40µ (and coating thickness
gauge shows <10µm. The gauge R&R (6 Sigma Repeatability & Reliability) shows
that a scrubbed, clean rail produces tribometer readings in excess of 0.50µ.
• Contaminated rails at Bicester site produced readings lower than 0.30µ and
between 0.30µ and 0.40µ for partially contaminated rails. Moisture, notably
morning dew, produced readings as low as 0.11µ.
• Slight WSP activity on test train was seen when tribometer readings dropped
below 0.35µ, and WSP was certain to operate when tribometer readings dropped
below 0.30µ with the train experiencing extended stopping distances.

Tribometer was seen to cut through the contamination if repeatedly used in same lateral
position across the railhead.
Concluded:

• Tribometer readings are only valid with the calibration of the instrument, extreme
low and high values should be used to indicate a trend not a comparative value.
• Narrow contact patch may cut through contaminant if run repeatedly.
• Train wheels have a wider contact band than the Tribometer. The railhead friction
experienced by the Tribometer is not necessarily that experienced by the train.

Note also Tribometer cannot be used on bullhead rail profile.

Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Autumn 2003 Monitoring - Depot Audits
Report Author: T Armitage, Ove ARUP & Partners
Subject Area: General
Report Number: REP/113912/002
Issue and Date: January 2004
Summary:

Reviews railhead treatment train fleet operational processes and reports on depot audits
conducted.
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Summary Report Project NADIR
Report Author: M Lee, B Grant, R Phillips, AEA Technology Rail
Subject Area: General
Report Number: -
Issue and Date: February 2000
Summary:

Summarises Project NADIR findings 1999

Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: An Assessment of Low Adhesion Railhead Treatment Methods Using
Data From Service Vehicles
Report Author: W Poole & JD Tunley, AEA Technology Rail
Subject Area: Low Adhesion Formation
Leaf Contamination Properties
Vehicle Design Influences
Report Number: RR-TRS-98-004
Issue and Date: March 1998
Summary:

Report details comparative tests of water jetting railhead cleaning and use of Sandite.
Produced for Railtrack.

Indicates possible mechanism for longer lasting effectiveness of Sandite alone or Sandite
after jetting may be in part due to the fact that very clean, oxide free, surface will be highly
reactive to the fatty acids from crushed leaves.

Indicates that each of the rail treatments is very effective for a very short initial period, up
to 50 axles since treatment.

Indicates that the wheel/rail dynamics, which may have a vertical unloading effect, may
influence the dynamic properties of the contaminant or treatment film.
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Autumn 2003 Monitoring - Trials and Monitoring
Report Author: I Bruce Ove ARUP & Partners Ltd
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Friction Modifiers
Adhesion Prediction & Communication
Report Number: 113912/1/9999
Issue and Date: Draft 1 25 March 2004
Summary:

Reviews studies undertaken development trials undertaken at Bicester in Spring 2003 and
monitoring across the Network Rail network in Autumn 2003.

Development trials included water jetting, Sandite laying, Kelsan friction improver,
LaserThor, high speed grinding and abrasive blasting.

Monitoring (4 'deep dive' and 6 'shallow dive' sites) undertaken to confirm track conditions,
confirm treatment effectiveness, obtain chemical samples of autumn track contamination
and compare predictions with actual conditions. Monitoring a deep dive sites included
track conditions, climatic conditions, evaluating LaserThor and observing performance of
railhead treatment fleet. Additionally under deep dive study, wheel and rail profiles were
collected to provide information on wheel/rail contact patch.

Deep dive sites allowed the behaviour and nature of the contaminant to be observed over
a long period of time. Report indicates that the accumulation of contaminant can be a fast
process that may take as little as 100 high-speed axles passes (half a day) to form,
remarking that there was a perception that it was a slow process in which the contaminant
layer gradually formed over several days.

Contaminant thickness 80 to 110 microns following initial crushing of leaf onto rail (still
recognisable as a leaf). Further trains roll this in and spread the individual leaf patches to
28 to 35 microns, spread over a considerable length. Fully contaminated rail not found to
have a thickness greater than 35 microns suggesting a limiting thickness.

Leaves on the trackside need to be 'fresh' i.e. not saturated that they congeal to a heavier
mulch, but those on ground for several days can still be a source of contamination.

Train speed needs to be high enough for their vortices to pick up leaves from cess or
ballast. Effect best seen between a stopping and non-stopping train.

Same process of accumulating contaminant is also wearing away contaminant if no new

material being drawn in, but observed wear rate is slower than build-up. Wear rate
depends on passing trains and heavy rainfall which softens the contaminant. Typically
about 300 axles passes for contaminant to be worn away. Heavy, prolonged rainfall has a
Appendix 3 Library Search Results - Network Rail

pronounced effect in quickly cleaning the railhead.

Salient Systems hand-operated tribometers used (note limitations). The use of portable
tribometer showed consistent results between different operators and different rail
conditions. Readings were in excess of 0.5µ for clean rail and 0.25µ on oiled rail (CF note
suggests absolute levels are high).

Contaminant adhesion levels found to be a range of levels rather than one consistent
value. No statistical relationship was found. A rail can be contaminated without causing
low adhesion conditions (dry rail), but a small amount of water (dew, damp etc.) on a
contaminated rail can lead to very low levels of adhesion.

Analysis of railhead swabs from deep dive sites found that magnesium is found in swab
sample. Strong correlation between leaf presence and magnesium and iron. Magnesium
still found when leaves are not present. Magnesium may be element in Sandite. Chemical
analysis proposed (CF note may be from brake blocks).
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Identification of Contaminant for Railhead Contamination
Management Trials
Report Author: Dr MA French, QinetiQ
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: QINETIQ/FST/CR032267
Issue and Date: June 2003
Summary:

Describes bench tests and field trials of artificial contaminants to be used as part of the
Bicester tests (Lignin (Norlig G) solution with grass).

Recommended:
• full chemical analysis of the real contaminants taken from several regions of the
country to identify the chemical constituents and physical properties and structure
of different contaminants and examine possible methods of disrupting the
contaminant with the aim of removing it.
• Carry out friction tests on a real contaminated rail to determine the similarities
between this and those recorded on the trial.
• Identify the influence of moisture and temperature on slip of real leaves and the
leaf substitute to allow the effect of different environmental conditions to be
evaluated. The influence of moisture on slippage should also be assessed under
controlled conditions.
• Lab trials to identify how leaves are transformed into the contaminant by the
action of the wheels passing over the leaves thereby inducing both pressure and
temperature change on the leaves.
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Autumn 2003 Monitoring - Shallow Dive Report
Report Author: Ove ARUP & Partners Ltd
Subject Area: Leaf Contamination Properties and Characteristics
Report Number: -
Issue and Date: Draft 1 February 2004
Summary:

Report describes the Shallow Dive monitoring conducted through Autumn 2003. The
shallow dive monitoring concentrated on looking at the effectiveness of the range of
existing railhead treatment trains operated by Network Rail (Sandite EMU/DMU, Class 37,
RHHT, MPV, water jetting unit, static Sandite applicators, GUV, TRAMM).

Chemical samples were taken from railhead before treatment train pass - little leaf
evidence mostly iron, Sandite, grease and coolant
Contamination position, resistance and thickness measured
Climate data obtained.
Tribometer readings before and after.

Contaminant electrical resistance using hand held Metrix MX20 DMM was not found to be
a good indication of likelihood of WSTCF due to variable pressure using probe, binary
readings close to zero or very high, patchy contaminant, zero on wet rail.
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Autumn 2003 Monitoring - Deep Dive Report
Report Author: R Brookes, Ove ARUP & Partners Ltd
Subject Area: Leaf Contamination Properties and Characteristics
Report Number: -
Issue and Date: Draft 1 14 March 2004
Summary:

Monitoring (4 'deep dive' and 6 'shallow dive' sites) undertaken to confirm track conditions,
confirm treatment effectiveness, obtain chemical samples of autumn track contamination
and compare predictions with actual conditions. Monitoring a deep dive sites included
track conditions, climatic conditions, evaluating LaserThor and observing performance of
railhead treatment fleet. Additionally under deep dive study, wheel and rail profiles were
collected to provide information on wheel/rail contact patch.

Sites monitored: Wanborough, Oxshott, Riddlesdown and Stonegate.

Chemical samples taken every 4 days increased to every day in mid-leaf fall season.
Measurements taken for every train of contaminant position, thickness, electrical
resistance, railhead friction, running band.
Weather data from weather stations (Campbell Scientifics), recording data every minute:
ambient temperature, relative humidity, wind speed and direction, rainfall, solar radiation,
plus thermocouples attached to rail.

Site survey - gradient, tree position and density.

Leaf fall survey undertaken using subjective interpretation of % leaves remaining.
Railhead profiles taken using Miniproof profile recorder.
Wheel profile measurements taken using Miniproof profile recorder.

Daily data gathered from sources listed above plus ADAS leaf fall reports, NR Adhesion
Index Forecasts and NR Adhesion Control Logs, daily statistics of weather and track
variables.

Concluded:

• There is a correlation between contamination thickness and electrical resistance

• There is no correlation between railhead friction and contamination thickness or
between railhead friction and contamination resistance.
• Max ambient pressure is negatively correlated to both daily rainfall and maximum
humidity.
• Daily rainfall is weakly correlated with maximum relative humidity.
• Max ambient temperature is weakly positively correlated with maximum wind
speed.
• Daily rainfall is independent of the ranges of ambient pressure change or relative
humidity change.
Appendix 3 Library Search Results - Network Rail

• Ambient pressure range is correlated with relative humidity change.

• The presence of leaves strongly correlates with the presence of magnesium.
• The presence of leaves is weakly correlated to the presence of calcium,
phosphorous and aluminium
• The presence of copper, chromium or nickel is independent of the presence of
leaves.
• Worn rails could give rise to an increased area of contact with the wheel.
• Railhead friction, contaminant thickness and wheel wear appear to be completely
independent variables.
• Leaves sucked from the cess onto the railhead did not stay on the railhead unless
wind speed was low or there was a small amount of precipitation.
• Gentle rain provides a wet railhead for leaves to stick to.
• Heavy rain acts as a track cleaning mechanism.
• The action of one train (16 axles) converted leaves on the railhead to a black,
durable contaminant.
• With significant leaf fall in the cess, a few through trains and some light rainfall, the
railhead contaminant was observed to accumulate into a continuous black band.
The process was faster than expected.
• Removal of the contaminant was accomplished without special equipment. After
20 typical trains (4 car units) the contaminant layer was worn away (a few hours on
a busy route).
• Weak correlation between ADAS Low Adhesion Forecasts and delay minutes due
to the site conditions.
• NR Adhesion Forecasts have no correlation to delay minutes.
• Poor correlation between daily change in leaf fall and ADAS Low Adhesion
Forecast.

NOTE THERE ARE MANY QUESTIONS POSED BY NR IN THE REPORT THAT ARE
YET TO HAVE AN ANSWER PROVIDED.
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Autumn Performance Improvement Trials - March 2003 Bicester Test
Report Author: R Brookes, Ove ARUP & Partners Ltd
Subject Area: Friction Modifiers
Adhesion Prediction & Communication
Report Number: REP/112880/001
Issue and Date: June 2003
Summary:

Reports on detailed full-scale adhesion engineering tests conducted at Bicester in

March/April 2003

Kelsan Friction Modifier

Keltrack is intended to create a coefficient of friction in the range 0.3 to 0.35 on the
running surface. Currently used to reduce noise and vibration, to control vehicle curving
and to reduce lateral loads. Can be applied using trackside applicators, vehicle mounted
and mobile systems.

Kelsan Keltrack product tested as an alternative to Sandite. Applied to railhead by

trackside applicators applying small quantity (<0.001kg/axle). Transfer mechanism from
track to wheels then to untreated section of track one wheel revolution later.

Results with application on a clean dry rail showed Keltrack maintained the braking
performance at 89%+/-2% and the detection performance at 90%+/-4%.

Excess of Keltrack acted as a lubricant lowering friction. Very small quantities must be
used to ensure a film that modifies and maintains the coefficient of friction.

The first tests resulted in high braking tests suggesting subsequent runs were affected by
contamination and product build-up on the MPV wheels. Multiple and bi-directional passes
over an applicator site are not indicative of how this product is normally applied. The
results of these demonstrations may have been influenced by the test procedures
adopted.

Kelsan Friction Enhancer

Kelsan VHFP (Very High Friction Product) applied from a roller and brushed onto
contaminated track (less well regulated than Keltrack tests). Intended as Sandite
replacement.

On dry contaminated rails VHFP showed only a very small improvement to braking
performance and slightly improved axle detection compared to the contaminated track.
VHFP does not condition the rail such that contaminated track returns to the same braking
and detection performance as clean rail.
Appendix 3 Library Search Results - Network Rail

Results were not successful with the film producing a very low braking rate.

Recommended that the results be fed back to the supplier, review the suppliers
development plans and undertaken controlled uni-directional off-network assessment.
Appendix 3 Library Search Results - Network Rail

Source: Network Rail

Report Title: Analysis of Autumn Performance:2003
Report Author: P Dunkerley, AEA Technology Rail
Subject Area: Adhesion Prediction & Communication
Report Number: LD03709_Report_2_Issue 1
Issue and Date: Issue 1 February 2004
Summary:

Update of Project NADIR work in 1999/2000.

Describes findings of a study of Autumn 2003 performance.

Conclusions:

• Vegetation levels adjacent to the track influence the degree of delay, although the
measure of the influence of sections with very high vegetation risk appears
anomalous (suspected due to high vegetation sections do not generally coincide
with high braking or acceleration requirement or that the high risk in such sections
is recognised and addressed well by local managers).
• Predicted weather conditions (ADAS forecast combining predicted rainfall, wind
and temperature into a single composite risk factor for each day) influence delay,
but this effect is mitigated on very bad days.
• Treatment of the railhead using a combination of Sandite and Water jetting
appears to be very effective in reducing expected delays to 48% of their value
without treatment.
• Rail grinding has not demonstrated to be an effective treatment for Autumn
adhesion effects.
• The effectiveness of on-train sanders has given a marked reduction in delay.
Combination of direct adhesion improvement and driver confidence to drive more
sharply.
 Report CRF04002
Issue 1
14 May 2004

Review of Current Adhesion Knowledge

Appendix 4

Library Search Results

Rail Safety & Standards Board

CR Fulford Associates 118

Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Track Contamination
Report Author: BR Research
Subject Area: Low Adhesion Formation
Leaf Contamination Properties And Characteristics
Report Number: TM TRIB 2
Issue and Date: -
Summary:

Friction has been shown to depend both on the nature of the oil and on the quantity
applied to the surface.

When varied in quantity, all fluids produce a similar curve in which friction falls as more
fluid is applied, until a constant low value, characteristic of the particular fluid, is reached.
This variation in friction can account for much of the variations in adhesion observed on
rails on dry days.

The characteristic friction coefficient is dependent on the nature of the particular fluid. For
most lubricants it is about 0.13. However, the nature of the oil can be substantially
changed by oxidisation or shear such that, in laboratory, friction coefficients as low as
0.03 to 0.08 have been measured.

Experiments on mixed use show that all components have an effect on friction and that
the effects of the components with the lowest characteristic friction is not necessarily
dominant.

Water on oily surfaces reduces adhesion to a value almost as low as if excess oil were
present. This is probably the most frequent cause of low adhesion in practice.

Debris provides an extra surface in which oil may spread, so that more oil is needed on a
debris covered surface before the characteristic low friction is attained.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: A Basic Study of Wheel/Rail Adhesion 9: Further Observations On
Leaves
Report Author: C Pritchard & MA Tanvir, BR Research
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: TM TRIB 6
Issue and Date: 1973
Summary:

The problem of low adhesion during the autumn occurs largely in tree lined cuttings where
a large number of leaves can accumulate on the track ballast. Each passing train whips
up these leaves which are carried along in its slipstream, and it has been confirmed that
its is by this action that the leaves get trapped under the wheels. The repeated crushing of
a large number of leaves can completely cover the wear band on the railhead with a film
of leaf pulp which collects a high proportion of iron oxide rail debris, forming a featureless
black mass. This material, with its relatively low shear strength, can totally prevent metal
to metal contact between wheel and rail, thus reducing adhesion. Rain or condensation
reduces adhesion to even lower values and occasionally to values as low as 0.03. The
leaf pulp does not adhere to the rail surface in wet conditions and the low adhesion is then
due rather to some component of the leaf material left on the railhead. This component
had not been identified and it was suggested it is in-situ on the railhead, due to oxidisation
and/or bacterial action.

Analysis of leaf debris – a living leaf contains about 80% by weight of water. The rest of
the leaf is largely cellulose plus a complex mixture of other chemicals including some
highly polymerised fatty acids, a Class of compound well known for their effective
lubricating action. Other compounds include proteins, amino acids, saccharides, neutral
fats, resins etc. The dead leaf looses most of its water but retains almost all the basic
elements of the other constituents. The flakes of black deposit found on the railhead show
some visible leafy matter. Apart from this the source of the black substance is not obvious
when examine separately. Samples were analysed. The main constituents were water:
and iron and its oxides. The most notable feature is the large proportion of iron and iron
oxides presence, analysis indicating that they provide almost all the non-organic solid
matter in the black material. Crushed leaves layers are usually about 100 microns thick
although some are twice that. Leaf deposits viewed under an optical microscope show
that fibrous material which is translucent and still identifiable as being part of a leaf bears
many small embedded particles. It is presumed that the carbon mostly originates from the
organic molecules of the leaf structure in which it was chemically combined. The other
constituents of these molecules are largely hydrogen and oxygen which are lost in the
analysis.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Adhesion and Leaves: Wet Abrasive Blasting Tests on Southern
Region
Report Author: IJ McEwen & RK Taylor, BR Research
Subject Area: Low Adhesion Formation
Leaf Contamination Properties and Characteristics
Report Number: TM TRIB 7
Issue and Date: June 1975
Summary:

States that it has been noted in the past that the combined action of traffic with rain is
sufficient to remove all traces of leaf film from the rail. Continued periods of rain also
prevent film build-up since the saturated leaves in the cess are less likely to be stirred up
by traffic.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: A Twelve Month Study on a 280 km Route
Report Author: SG Watkins
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: TM TRIB 9
Issue and Date: June 1976
Summary:

The BR Tribometer train was used to measure wheel/rail adhesion at regular intervals
during one year over a 280 km route commencing in November 1972. Claimed results are
presented and the influence of those factors expected to affect adhesion are discussed.
The most important cause of low adhesion was found to be rainfall. Most other
differences, including those between different lengths of track, are swamped by the
random spatial and temporal variations which occur on all track. However, some specific
"low spots" were identified, and associated with rail side features such as flange
lubricators and cuttings where the local humidity may be high.

The BR Tribometer train was used.

The report details the effects of traffic utilisation and type, the effects of track curvature
and gradient, at diurnal variation of adhesion, annual variation of adhesion, effect of
speed, effect of rain, incidence of low adhesion, effect of tunnels.

The most important conclusion is that low adhesion encountered over substantial lengths
of track is nearly always associated with water on the rails.

A clear picture of the statistics of how adhesion varies on BR track has emerged. Except
where identifiable extraneous influences are diurnal work the survey has confirmed that,
by and large, all sections of track are subject to the same variation, the magnitude of
which tends to swamp the large unexplained differences between different track sections.
Clearly a number of factors contribute to this variation, and their interactions make it
difficult to ascribe precise values to particular effects.

Apart from general low adhesion, usually due to rain, localised low spots have been
shown to be associated with a number of line side features, many of which are
unavoidable. Perhaps the most serious low spots are those which seem to be caused by
flange lubricators, raising doubts about the continued use of these devices in their present
form.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Water Spray Treatment for Conditioning Wheel/Rail Adhesion II The
Effect of Additives
Report Author: IJ McEwen, ID Pollicott & A Starbuck
Subject Area: Leaf Contamination Properties and Characteristics
Report Number: TM TRIB 16
Issue and Date: February 1977
Summary:

A report "Water Spray Treatment for Conditioning Wheel/Rail Adhesion" describes a water
spray treatment system for conditioning wheel/rail adhesion. The spray equipment,
mounted on the Tribometer train, is used to reduce adhesion and is particularly useful for
the testing of low adhesion remedies. The use of a water mist to give a very thin moisture
film on the rail has advantages in that the effect is instantaneous and, because of the
small volume of water involved, is short lived such that the chances of subsequent rail
traffic encountering low adhesion is minimal.

However, two slight disadvantages to the system had shown themselves over the first
year of usage. These are:

(A) not all sites react in the same way to the water film applied, some sites being more
susceptible to low adhesion than others. (However, adhesion is always reduced from the
drying "as found" condition).

(B.) the adhesion level falls to some characteristic level, dependent on the sites, and is
often not as low as the naturally occurring low value sometimes recorded and which are
required to be simulated.

It would be of great advantage to be able to produce adhesion conditions of 0.1 particular

when assessing remedies. To this in small percentages of other materials were added to
the water used in the mist generator in attempts to give a further consistent reduction in
the adhesion coefficient the system is then less realistic of normal railway conditions
(water being a natural contaminant).

The effectiveness of the water spray system in reducing track adhesion is increased by
the addition of small amounts (1-2%) of a soluble cutting oil to the water.

Some sections of track are more susceptible to low adhesion produced in this way than
others.

The method provides a useful tool for the testing of remedies to counteract low adhesion.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Description of Trackside Water Spray Equipment
Report Author: RK Taylor, BR Research
Subject Area: Leaf Contamination Properties and Characteristics
Report Number: TM TRIB 30
Issue and Date: August 1978
Summary:

Referenced in report RR-TRS-98-004 ‘An Assessment of Low Adhesion Railhead

Treatment Methods Using Data From Service Vehicles’ Issue 1 March 1998, produced for
Railtrack PLC.

REPORT NOT AVAILABLE

Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: A Basic Study of Wheel Rail Adhesion 6 – Adhesion and
Contamination on BR Track
Report Author: M Broster, C Pritchard and DA Smith, BR R&D
Subject Area: Leaf Contamination Properties and Characteristics
Report Number: Technical Note TM TRIB 3
Issue and Date: May 1973
Summary:

The way in which adhesion varies on BR track has been determined. The quantity and
variety of railhead contamination was studied, including specific types of contaminant
peculiar to certain individual sites. Generally, oil was found to be surface active and
present in such small quantities that boundary lubrication conditions apply. The amount of
solid debris varied widely. Water usually had an immediate effect in reducing adhesion.
Trains control the spread of contamination and, together with track lubricators, are one of
the more prolific sources of oily fluids. The knowledge gained has enabled a full
programme of tests simulating track conditions to be undertaken in the laboratory.

Variation on rails – adhesion levels quoted are ave. along a 60ft rail section using a
portable tribometer. Little variation seen when adhesion was high and low (damp leaf)
over 5 lengths. On 2 occasions adhesion measured over a longer length through a leaf
affected cutting. Humid conditions in shade of overbridges reduced adhesion. Adhesion
also varies laterally across the rail.

Variation with time – adhesion & humidity measurements taken over 48 hour period. Short
term variations in adhesion believed to be due to the passage of trains. Most trains alter
adhesion by a small amount. Humidity as distinct from rain may also affect adhesion
expressed as temperature difference between the rails and the dew point. The greater this
difference the drier the rails. This led to a general correlation between adhesion and
humidity.

Level of adhesion – average adhesion over 50 days measurement 10 measurements/day

was 0.3 with the lowest 0.06 in November.

Variation with train passage – trains thought to cause the small but abrupt changes in
adhesion. But other factors at play, weather or rail temperature. Oil contaminated rails
quickly cleaned by passage of trains. Long freight train alone and a passenger express
virtually cleaned rail by single loco not enough. Trains carried oil further down the track.

Effect of humidity and rail temperature – humid/moist conditions required for low
adhesion. Smooth rail condensates when surface temperature falls below the dew point
which is 100% humidity, but on rough or debris covered rails water may condensate in
cracks. It is the humidity relative to rail surface temperature that is important.

Effect of water – studied by spraying tap water on rails. Adhesion immediately reduced by
0.05 and 0.15, ave. reduction being 0.08.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Bad adhesion sites – 30 locations of reported bad adhesion were visited. 50% had no
visible contamination but rail burns present. Adhesion levels recorded were normal. Sites
characterised by gradient or curve. Other sites obvious cause of contamination (leaves,
coal, sawdust, grease, clay, oils, thick rust). Requires moisture to lower adhesion to a
minimum. 50% of leaf pulp was made up of rail debris (iron and iron oxides).

The work built up a picture of how adhesion varies naturally on BR. Adhesion on each rail
varies between about 0.04 and 0.55 and in reasonable dry conditions averages 0.3. The
passage of trains maintains this average presumed to be by helping to clean
contaminated rails but spreads contamination along the rails. Occasionally trains depress
the adhesion markedly suggesting oil spillage. The weather affects both rails and also
many miles of track. Changes in the humidity of air affects adhesion by a small amount
but the greatest change is due to water deposited as rain or dew. Ave. adhesion on newly
wetted rail is nearer 0.2 than 0.3. Steady rain gradually washes excess oil from the rails
leaving a few molecular layers. Adhesion measured after rain is usually high. Debris on
the rail varies but is mostly composed of rail wear debris and brake block debris. No
correlation found between rail roughness and adhesion.

No one cause was identified as the most important. Low adhesion arises due to the
interaction between several factors but the most widespread is the effect of dew or newly
falling rain on rails covered in tiny particles of oil.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: A Basic study of wheel/Rail Adhesion – Laboratory Studies of the
Effects of Water
Report Author: TM Beagley & C Pritchard, BR Research
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: TM TRIB 4
Issue and Date: 1973
Summary:

Perhaps the most important conclusion of this report is that humidity and water can control
the friction of steel surfaces over a range from 0.6 to 0.15 even when the surfaces are as
free from oils as can be managed in a laboratory. The minimum friction is associated with
surfaces that are slightly wet, and is caused by changes in strength and flow
characteristics of a surface layer of wear debris mixed with water. An unanswered
question is how much debris is necessary on rails before the lowest friction is
encountered.

In the presence of a surface layer of debris, high humidity also reduces friction, the
parameter being the humidity expressed relative to the rail temperature. For ambient
humidity, cold rails absorb more water vapour than the rails warmed by the sun, and the
adhesion is consequently less. The changes in friction are almost immediate.

On debris free surfaces in the laboratory, the friction falls to 0.3 when the surfaces are
wetted with water. If small quantities of oil are present, as is invariably the case on rain
affected rails, the friction falls to lower values which depend on the amount of oil present.

Large quantities of oil are remove by steady washing, such as by continuous rain, but
enough to form a few molecular layers remains tenaciously bound to the steel surfaces.

Very strong evidence that the action of water in reducing friction on rails is not by the
mechanism of transporting debris from the rust covered shoulders of the rail as is often
been suggested.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: A Basic Steady Off Wheel Rail Adhesion 8 Laboratory Studies of the
Effects of Solid Debris
Report Author: TM Beagley, M Broster, IJ McEwen & C Pritchard BR Research
Subject Area: Leaf Contamination Properties and Characteristics
Report Number: TM TRIB 5
Issue and Date: 1973
Summary:

The study has been made of the physical characteristics of railhead debris particles and
their effect on friction when dispersed thinly and when forming a thick surface film. An
important parameter is their large surface area. Depending on this area, comparatively
large quantities of oil are necessary to cause low friction. On rails, debris helps maintain
adhesion against the effects of oil. Thick films reduce adhesion. Rain supplies so much
water its effect then predominates.

Debris as found on rails surfaces can roughly be Classified into two categories, the thinly
disposed particles found on most well used lines, and thick flows which completely
obscure the wear band at a few particular locations.

Dispersed debris improves adhesion on dry rail. The particles provide a large surface area
on which spilled oil can be spread, thus reducing the amount of oil on the rail. In humid
conditions, the strength of debris particles is reduced and adhesion is worst affected when
the surface density of particles is high. Adhesion on comparatively clean rails is therefore
more influenced by spilled oil, less by high humidity. Water, compared with oil, covers the
surfaces in overwhelmingly large proportions, and probably reduces adhesion equally on
clean or dirty rails.

On thick films, adhesion is governed solely by the shear properties of the debris layer.
Thick layers on rails mostly found some fluid constituents and the shear strength, and
therefore the adhesion, as low. Adhesion becomes lower as the layers become thicker.
Humidity has an appreciable effect and water substantially reduces adhesion. Minimum
adhesion occurs when water is just on the point drying out.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: A Basic Steady Of Wheel/Rail Adhesion10: The Rheology of Rail
Contaminants Causes Low Friction
Report Author: TM Beagley, BR Research
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: TM TRIB 8
Issue and Date: May 1974
Summary:

Slightly wet, contaminated rails are reputed to cause wheel slip. The physical properties of
railhead contamination are examined with the special emphasis on mixed use of iron
oxide and water. These properties are then used to calculate the limiting coefficients of
friction on heavily contaminated track. It is shown how very low adhesion can be caused
by slightly wet track debris. This acts as a viscous lubricant fully or partially unloading the
wheel from the rail. Track observations have shown that significant debris coverage can
occur on mainline track under certain conditions.

The causes of low adhesion on the track due to solid contamination have been analysed.
Under these conditions the Classical boundary lubrication equation no longer holds.
Coefficients of friction measured with lightly loaded Tribometer under debris contaminated
conditions cannot be transposed to describe wheel/rail adhesion.

A permanent build-up of track debris is only possible when the contaminated material has
specific properties combining both yield strength and pliability.

Solid contaminant films reduce adhesion if their shear strength is low. Leaves, oil and coal
contamination can form films of low shear strength.

Very low adhesion can be caused by slightly wet track debris. This act as a viscous
lubricant fully or partially unloading the wheel from the rail according to the degree of
contamination present.

Track observations have shown that significant debris coverage can occur on mainline
track under certain conditions.

The rheological properties of wet rail debris can change rapidly with water concentration,
misty conditions or light rain being necessary before severe loss of adhesion can occur.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Track Testing of Adhesion Modifying Fluids
Report Author: RK Taylor, BR Research
Subject Area: Friction Modifiers
Report Number: TM TRIB 27
Issue and Date: May 1978
Summary:

In recent years tests have been carried out using the Tribometer train to investigate the
effect on wheel/rail adhesion of a series of fluids applied to the railhead. The fluids tested
had previously been Classed as adhesion improvers, either from work with the original
trackside applicators or from all recent theoretical considerations and laboratory tests. The
Tribometer train tests were designed to see whether these fluids influence adhesion and
to determine the optimum amounts (or concentrations) needed for raising adhesion. Some
tests were made excepting the "natural" as found adhesion levels on the test sites; others
were made using the water mist system fitted to the train to depress the adhesion levels.

The investigation was planned in order to gain information relevant to the use of trackside
fluid applicators, i.e devices which deliver fluids to the railhead and wheels as trains pass,
conditioning the track for a certain distance beyond the device. The applicator at present
commercially available is called a "Maxi-mu" and makes use of a trackside pump operated
by train wheels to scrape conditioning fluid.

Due to the mechanical unreliability in service of these fluid applicators, research and
development division developed a more robust system. It is gravity fed using a trackside
switch which simply opens a valve. Results to support the basic principles are still few and
indecisive. It still has to be established which fluid, if any, is effective in improving
adhesion or in maintaining adhesion above a minimum safe level.

The Tribometer train was used to gain information rapidly about the effect on adhesion of
fluids which had already shown promise. The fluids were applied to the rails between the
measuring axles of the Tribometer vehicle, thus enabling a measure of adhesion to be
made before and after treatment in one pass, though it was realised this would not show
the long-term conditioning claimed for some of the fluids.

The fluids tested were:

Syton P - a proprietary "frictionising" agent 30% by weight suspension of colloidal silica in

an acquiesce media, the particle size being in the range 35-40 millimicrons.

Ludox - a similar product to Syton has a higher silica concentration "40% by weight" and is
claimed to have similar properties.

Ethyl Caprylate - solutions of this ester containing ethyl caprylate. In spite of its long
usage there is still little evidence that the fluid raises adhesion it is claimed to work by a
conditioning effect which takes a good of time. Although normally using diluted solutions it
has been suggested that for it to be defective by the original mechanism proposed the
Appendix 4 Library Search Results - Rail Safety & Standards Board

neat ester should be used.

Portec solution - this is a fluid for use in the maxi-mu applicators. It consists of 1% ethyl
caprylate in 50:50 water: Iso-propyl alcohol mixture. No change in adhesion due to this
solution had been detected by the Tribometer train survey of trackside applicators, even
when the site had been conditioned for a considerable length of time.

Tertiary Butylamine Solution - it has been suggested that for track contaminated with solid
debris the most likely cause of low adhesion is the formation of a solid/liquid paste with
the rheological properties to support a locomotive wheel and prevent or reduce wheel/rail
contact. It was further suggested that a remedy to overcome this situation would be the
tertiary butylamine, a deflocculant which can markedly change viscosity properties of such
a paste.

Sodium Metasillcate Solution -- this material was also tested in early applicators and is
being used by SNCF, in maxi-mu applicators. SNCF claimed to use a 5% solution in
water, but, at Derby, difficulty has been found in making a stable solution any more
concentrated than 4%, except by the use of demineralised water.

The fluids tested in this work either reduced adhesion considerably when freshly sprayed
or at best had little influence on the initial values. Only the Syton/Ludox materials
produced a noticeable improvement, and then only after drying out on the rail. Syton also
lowered adhesion when freshly applied but not to the same extent as with ethyl caprylate,
in fact no lower than with water alone. On drying out, it gave high values of adhesion in
some cases greater than 0.5. When measurements were carried out subsequently in best
conditions the adhesion drops to a similar value to that on the control section. Although
this suggests that Syton would not be of use in wet weather, it should be noted that the
adhesion was still an acceptable value.

Ethyl caprylate and Portec solution can both produced an unacceptably low level of
adhesion < 0.1.

20% Syton and 20% Ludox reduced adhesion when initially applied but, when allowed to
dry, reduce high levels of adhesion. These high levels of adhesion were not maintained in
mist conditions.

Other fluids tested were tertiary butylamine and sodium Metasillcate. Neither fluid was
tested extensively and the results were inconclusive.

The only fluids that shown promise of improving adhesion were Portec, Ludox and Syton,
and they would seem to be the only ones worth using in applicators. Since Portec solution
and ethyl caprylate can produce very low adhesion, one must doubt the wisdom of using
such materials in trackside applicators.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Assessment of The Cleaning of Leaf Affected Track on Southern
Region – 1977
Report Author: RK Taylor & ID Pollicott, BR Research
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: TM TRIB 29
Issue and Date: July 1978
Summary:
A section of trackside water spray equipment was installed near Bekesbourne station. The
purpose of this equipment was to assess the effectiveness of simulating rainfall, which
when accompanied by the passage of trains had been observed to clean the rails most
effectively.

The spray equipment sprayed water on a total length of rail of 200m from jets situated in
the middle of the 4 foot, spraying both rails of the down line every 18 minutes. The spray
rate was sufficient to ensure that the rails remained almost continuously wet in the
prevailing weather conditions, being just on the point of drying out at the end of each 18
minutes. The rails were inspected for areas of leaf contamination. The thickness of films
found on the surface of the rails was measured. To assess the effectiveness of water
spray in maintaining leaf free rails, the appearance of the treated section was compared
with that of the preceding 200m of track. Both the Tribometer Train and a portable
tribometer were used to take measurements.

It was found that when the trackside water spray was switched on the rails were almost
always completely clear of leaf debris, in marked contrast with the control section. When
the trackside water spray was switched off for a time, there was seen to be black on the
rail which was quickly removed by the passage of trains once the system had been
reactivated. The mean adhesion coefficient on the water sprayed section was about 0.15
when the rails were wet, increasing to about 0.21 as the rails dried out during the 18
minutes cycle. Immediately beyond the sprayed section the adhesion was lower still.

The thickness of the leaf films found varied along the track from between 5 to 10 microns
and up to 90 microns. Higher than average thickness values were obtained where fresh
leaves were on top of a black leaf film. The general thickness of the black leaf film was
found to be in the range 20 to 30 microns.

The rails wetted by the trackside water spray system at all times appeared clean to the
eye. Occasionally, single leaves became crushed on the rail and were soon removed by
the following traffic. Far less contamination was observed on the preceding control
section. It was also possible to demonstrate the cleaning effect by switching off the spray
system to allow some contamination to form. Switching off confirms the principle of
trackside water sprays as effective means of maintaining rails free of visible leaf debris.
The lowest mean adhesion measured on the sprayed section was 0.15 the lowest mean
adhesion recorded for the control section was 0.13 on rails wetted by mist generator. On
heavily contaminated track, in damp conditions, adhesion levels of < 0.05 are common.
Bearsted Bank trackside water spray was able to keep the rails and these sites as clean
as it has been seen, the minimum mean adhesion of 0.15 would give Bearsted Bank
considerable improvement.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Assessment of The Cleaning of Leaf Affected Track on Southern
Region – 1978
Report Author: RK Taylor & ID Pollicott, BR Research
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: TM TRIB 35
Issue and Date: July 1979
Summary:

This report identifies the results of further water spray testing conducted in 1978 out
Bearsted Bank. The water spray sprayed a 400m section of track on the down line. The
spraying nozzles situated in four foot operated for one minute every 15 minutes to keep
the rails continuously wet. The degree of leaf contamination on the Bank was found to be
in excess of that experienced in the previous year that Bekesbourne. The rails on the
Down Line were heavily contaminated as far as the trackside water spray section where a
marked change in the appearance of the rails was seen in the dry weather conditions. The
rails in the sprayed section were noticeably cleaner than those in the preceding control
section, being free from gross leaf contamination and presenting a clear wear band with
only a few fresh leaves loosely attached to the surface. The rails immediately beyond the
sprayed section remained of clear appearance. On progressing further up the Bank the
rails became heavily contaminated with leaf debris. The rails in the sprayed section
became no cleaner in heavy rainfall.

The efficiency of the system was badly affected by very strong winds on November 15.
This disturbed the spray pattern so that the rails were not properly wetted. This was
exceptional and are not normally encountered.

The trackside water spray maintained considerably cleaner rails than those which were
not treated. The cleaning action was consistent and was judged to be far more effective
than the water cannon trains at their best. However the rails were not completely clean.
Leaf contamination was still evident dotted about the wear band in the small pitted areas
which are not contacted by the wheels. The same contamination pattern was evident in
heavy rainfall, the spray was therefore seen to be successful in simulating rain. It is not
known what effect these patches have on adhesion but it is reasonable to assume that at
worst the adhesion is no lower than that of rails cleaned by rain. No adhesion
measurements could be obtained at Bearsted Bank but the results from 1977 showed the
lowest individual adhesion measurements obtained on the water spray section was 0.12
compared with 0.07 on untreated rail.

Longer sections of track will contaminated with leaf debris and remained so for longer
than in previous years. The trackside water spray maintain markedly cleaner rails than
those that were untreated, and appeared considerably more effective than the water
cannon at its best.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Wheel/Rail Adhesion – Review of the Effectiveness of Trackside
Applicators
Report Author: RK Taylor, BR Research
Subject Area: Friction Modifiers
Report Number: TM TRIB 36
Issue and Date: November 1980
Summary:

Over the past ten years various tests have been carried out to establish the effectiveness
of trackside applicators which dispense fluids to improve wheel/rail adhesion. Applicators
were originally induced in 1958 to spread the fluids along the railhead by the action of the
passage of trains. However, the usefulness of these devices has long been doubted,
though it has proved difficult to devise tests which would convincingly settle the matter
one way or another. This report sets out to outline the more recent tests which have been
done, and to summarise the information gained.

This report details tests undertaken at Derby in 1969, Barton and Walton in 1972, Wood
Green in 1972, Woodhouse in 1972, Sheffield circuit in 1974, Beattock in 1973, Derby
Deadman's Lane in 1975, Bradford in 1975, Hemmerdon Bank in 1975, Buildwas power
station in June 1976, Taunton in March 1977, and in France between 1970 and 1977.

The evidence concerning the efficacy of applicators is drawn from three areas: service
experience in combating specific problems, track trials in which success has been judged
by Tribometer results, and the trials with fluids applied from the Tribometer Train.

The evidence obtained with fluids deposited from the Tribometer Train, that Portec
solution and ehtyl caprylate solutions both cause low adhesion when freshly deposited,
provides grounds for a positive recommendation that neither of these fluids should be
employed.

In summary therefore there is no clear-cut evidence the applicators are effective. There
remains the suspicion that the dispensing of Colloidal Silica or Sodium Metasillcate may
be beneficial, but the evidence at present is so sparse it is impossible to justify any new
installation.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Brakes and Wheel/Rail Adhesion
Report Author: C Pritchard, BR Research
Subject Area: Formation of Low Adhesion
Report Number: TM TRIB 39
Issue and Date: 1980
Summary:

Describes work reported elsewhere on the adhesion survey of a 280 km Route.

Causes of Low Adhesion - low adhesion has generally been considered to be due to oily
contaminants on the railhead, the effect of water being to enhance the lubrication in some
way. However, recent work suggests that, instead, water is most often the prime cause of
low adhesion, mixing with solid debris (rust) to form surface films with non-Newtonian flow
properties. With a critical water content the films are not readily squeezed from
underneath the wheels yet they provide a low resistance to shear. These critical
conditions arise when the quantity of water is minimal. With a more ample supply of water,
as in steady rain, the effective viscosity of the water/solid mixture is so low it is readily
squeezed aside and adhesion stays high as the wheel makes direct contact with the rail.
On a visibly rusty rail the film can initially be thick enough to support the wheel completely,
when extremely low adhesion can result.

The concept explains why trains slip more frequently the rustier the track and why the first
train after Sunday's lack of use experiences low adhesion. On well used rails there is little
debris on the wear band and even though wear debris is spread from the rail shoulders
the reduction in adhesion is not so great. The small amounts of water which cause the
lowest adhesion are most often supplied in drizzle or dewy conditions, and persist longest
in shade or along tree-lined sites where the rails remain cool, the ambient humidity
remains high, and any water film takes a considerable time to evaporate.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Traction Between Rolling Steel Surfaces. A Survey of Railway and
Laboratory Experiments
Report Author: C Pritchard, BR Research
Subject Area: Low Adhesion Formation
Report Number: TM TRIB 46
Issue and Date: September 1980
Summary:

This report is the text of an invited paper to be presented at the seventh Leeds – Lyon
symposium Tribology, “Friction and Traction", in September 1980.

A survey is presented, also many studies undertaken, to measure the friction between
railway wheels and rails, and of laboratory investigations aimed at understanding the
fractional processes involved. Friction on rails varies widely, caused according to
laboratory tests by changes in the rheological properties of solid railhead debris as
modified by water and small amounts of oil. The major source of debris is identified as
rust. Surface layers reduce the initial slope of the friction force/creepage curve, the
experience with locomotives suggests that friction is increased by sufficiently severe
sliding/rolling action before being reduced by thermal softening under more extreme slip.

It is clear that there still remains a great deal to be learnt about the fractional processes in
all three regions of behaviour defined by the notional creep curve; low creepage,
maximum friction, and high slide/role ratios.

Evidence of the influence of debris on maximum friction is more convincing. It shows how
friction varies from 0.05 to 0.3 depending only all visible changes in the consistency of
surface films of debris mixed with water.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Update of the Review of Autumn Leaf Remedies
Report Author: TG Pearce, BR Research
Subject Area:
Report Number: VDYN 148
Issue and Date: February 1990
Summary:

This report provides an update on earlier reviews undertaken into potential remedies for
solving the problem of low adhesion. The report identifies the potential for trackside water
sprays although these are not considered during and testing. Also contains a table
identifying many methods of addressing the problem and ranking their efficiency,
practicality. This includes train mounted water sprays, trackside water sprays end water
troughs around rail.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Autumn Leaf Remedies – The Future of
Report Author: TG Pearce, BR Research
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: VDYN 162
Issue and Date: March 1990
Summary:

This report follows the earlier report title “Update of the Review of Autumn Leaf remedies”
addressing the specific recommendations of the report.

It acknowledges that vehicle mounted water sprays have a part to play under certain
emergency conditions such as those following the 1987 hurricane. There the volume of
green leaf on the track allowed the sap to be carried over long distances on the train
wheels and cause widespread adhesion problems, which were effectively dealt with by
washing the rails. However vehicle mounted equipment cannot deal with black leaf film.
Heavy rainfall can disperse the leaf by softening it to the extent that the passage of trains
and rain subsequently washes away. Simulating rainfall from trackside water sprays has
been shown to have the same effect. However this is very costly as a general treatment.
Nevertheless it is very effective in keeping the rails continuously clean. Naturally the rails
stay damp, so the adhesion does not rise as high as that found on dry rail, but remain
sufficiently high to support normal braking and traction demands. Frost can be a problem
since the probability of ice of the third rail increases in electrified areas.

This treatment is thus only suitable for particular locations where there is a regular
problem and cannot readily be cured by one of the previous treatments. This could be
because the trees of Railway land, or because of the logistics of including the location in a
regular treatment train diagram. There are no locations of this nation for which a trackside
water spray is currently being proposed.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Adhesion and Leaves – A Review of the Problem and Potential
Solutions
Report Author: TG Pearce & DJ Watkins, BR Research
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: TM VTI 017
Issue and Date: October 1987
Summary:

This report reviews in detail the mechanism whereby the rails become contaminated,
discusses the many effects of this contamination and renewal measures that have been
suggested to overcome the problem. The more promising of the remedies have been
subjected to a cost benefit analysis which is updated and extended in this report. This
report then concludes the summary of the most cost-effective ways of combating the
problem.

The mechanism of deposition is described. During dry weather, leaves lying on the
lineside become airborne by the turbulent air around passing trains. Many these are going
to and fro across the rails, and are trapped by the vehicle and squeezed on to the rail. For
examples, a 4 vehicle 16 axle train has been observed to deposit an average leaf per yard
of rail. A succession of trains compress the deposited leaves and turning them into a
glazed black film covering almost the whole of the running surface of the rail.

Although the black rolled film reduces adhesion to around 0.1 level, the small amount of
precipitation (dew, light rain or snow shower) will reduce the level of adhesion to 0.05 or
less. Spot values as low as 0.02 have been measured using the Tribometer Train. These
levels of adhesion should be compared with braking and traction demands of up to 0.09
as 0.2 respectively, from which it is apparent that there can be a substantial adhesion
shortfall. The influence of moisture on adhesion on leaf affected track is significant. As a
result, early morning an early evening trains frequently encounter low adhesion, and a
light shower of rain can have a sudden dramatic effect on the train service.

A number of other important observations have been made:

(A) continuous rain together with the passage of trains quickly clears heavily contaminated
sections of track.
(B) fallen leaves do not adhere to the rail during wet weather. This is partly because the
railhead is wet, and partly because the wet leaves adhere to each other on the ballast and
are not lifted into the trains' slipstream.
(C) the transfer mechanism was noted when a pile of fallen leaves were placed on an
uncontaminated railhead; after the passage of a train contamination was noted not only at
the initial position of the leaves but at successes wheel revolutions along the rail.
(d) following rail cleaning, the continuous black rolled film has been observed to re-
establish itself with the passage of the few trains, amounting to no more than 200 axles,
when suitable drying conditions return.
(e) leaf fall areas where tread brakes are repeatedly applied do not exhibit the degree of
contamination expected. Here it appears that the action of tread brakes prevents the
Appendix 4 Library Search Results - Rail Safety & Standards Board

transfer mechanism from producing the levels of contamination.

(f) coniferous needles have never been identified as significant contamination on the
railhead.
(g) some newer train sets have been found to deposit twice the number of leaves on the
railhead. This was shown by simulating leaves with pieces of paper used alternatively by
the newer Class 508 units and the older Class 405 units.
(h) track geography can play a significant role in determining whether a given location
becomes a leaf site. The position of the trees in relation to the prevailing wind can
determine where the leaves land. Also Railway cuttings are more liable to leaf
contamination than level ground or embankments. This is partly because they act as a
leaf trap, and partly because the moisture content of the air tends to be higher,
encouraging the formation of dew and delaying evaporation.
(i) The type of tree is important. Both sizes of leaf and volume of leaves are critical. Thus
several beech trees are required to produce the same level of contamination as one
sycamore, and the authors are not aware of any site where Hawthorn bushes have given
any problem.
(j) whilst the above comments have concentrated on heavily contaminated track, rails
partially covered with leaf, and rails downstream of a leaf site, are covered with a sticky
leaf "juice" can also exhibit similarly low adhesion values to those quoted.

The report details many remedies for removing the source of contamination preventing its
build up, cleaning the rails or other methods of adhesion raising.

The report also provides a detailed cost benefit analysis of various remedies.

Source: RSSB Library

Appendix 4 Library Search Results - Rail Safety & Standards Board

Report Title: A Comparative Study of The Trapping of Leaves by Different Multiple

Units
Report Author: RK Taylor, BR Research
Subject Area: Low Adhesion Formation
Vehicle Design Influences
Report Number: TM WRI 5
Issue and Date: March 1982
Summary:

Tests were carried out to investigate how differences in stock affected the rate of
deposition of leaves on the railhead. Since the leaf fall season is relatively short and once
commenced the continued fallen leaves would make the execution of control test difficult,
the tests were carried out outside of the autumn season using pieces of paper to
represent leaves. Further tests were then done with fallen leaves to compare with the
paper tests. The tests were prompted by the increased problems with low adhesion
conditions found with the new type of the electrical multiple units such as the 507s and
508s.

The length of the train formation obviously affect the numbers of leaves deposited on the
railhead but the number is not proportional to train length. 8 car units do not deposit twice
as many leaves as 4 car units.

Train speed is also a major factor affecting the numbers of leaf trapped by passing trains.

The tests with real leaves confirmed that Class 508 deposit more leaves than Class 405s.
Under the same conditions a Class 508 unit trapped more than twice as many leaves on
the rails as a Class 405. The passage of an 8 car Class 508 at 50 mph caused as many
as 60% of the leaves present in four foot to be deposited on the rails.

Train speed and layer of both influence the number of leaves rolled onto the rails, but
number rolled on increases more slowly than the increasing train speed or length.

Use of paper is a viable test procedure for comparative tests.

Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: British Rail’s Autumn Leaf Problem – 1980
Report Author: RK Taylor, BR Research
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: TM DA 38
Issue and Date: May 1981
Summary:

This report tends to give an informed insight into the leaf problem and reviews the way in
which various areas are attempting to combat it. The report details recent history on
Southern region, between Marylebone and Aylesbury and on Merseyrail. It details
remedial action is being taken for those lines and the results of the 1980 season.

When assessing the effects of the leaf fall season it is difficult to compare one-year with
another as a number of factors combine to create the critical adhesion conditions. One of
these factors is steadily increasing growth of trees on the lineside. Another key factor
affecting the build-up of leaf films on the rails is the prevailing weather conditions. The
films quickly build up in dry weather, which when followed by early morning mist or light
rain will give rise to low adhesion conditions. In steady rain, on the other hand, the leaves
less readily become airborne and the passage of trains effects a cleaning of the rails.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Assessment of Proposals for Controlling the Autumn Leaf Problem
Report Author: E Rushton, BR Research
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: TM TAG 87
Issue and Date: June 1983
Summary:

This report provides an approximate comparison of the cost per year for the different
methods of controlling autumn leaves. Eight proposals are examined including for
trackside water sprays, leaf guards, Sandite, dry sand, tread brakes, tree clearance, tree
management, aerodynamic deflections.

Trackside water sprays, dry sand, aerodynamic deflections are rejected through cost.
Proposals for further consideration are vehicle leaf guards, Sandite, tree clearance and
tree management.
Appendix 4 Library Search Results - Rail Safety & Standards Board

Source: RSSB Library

Report Title: Computer Simulation of the Traction Systems Used in the Class 47
SEPEX Adhesion Trials
Report Author: S Quick, BR Research
Subject Area: Vehicle Design Influences
Report Number: TM ES 073
Issue and Date: May 1986
Summary:

Reports on the use of computer modelling to predict locomotive performance under poor
adhesion conditions. Trials with a Class 47 locomotive modified by Brush Electrical
Machines to give separately excited traction motor fields were undertaken in November
1985. Models of the SEPEX modified and standard locomotives were constructed for
these trials. The model results are presented and compared with examples from the field
trials. Limitations of the models are discussed. The performance of the separately excited
motor locomotive and of the various control schemes proposed was then evaluated with
regard to the utilisation of available adhesion.

Computer modelling of control and motor variants of a Class 47 locomotive has enabled
performance and stability predictions to be made. The use of separately excited motors
for diesel electric locomotives subjected to poor adhesion conditions has shown to be
beneficial in reducing the likelihood of gross wheelslip in particular when at least one axle
is not in wheelslip. in particular, SEPEX motors introduce the possibility of using a simple
locomotive control scheme using current feedback to maintain good use of adhesion.
 Report CRF04002
Issue 1
14 May 2004

Review of Current Adhesion Knowledge

Appendix 5

Library Search Results

AEA Technology Rail

The majority of AEA Technology Rail reports available for review were already reviewed
through either the AWG library or RSSB library searches.

CR Fulford Associates 145

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Low Adhesion Due to Leaf Fall: Response of Non UK Administrations
Report Author: JR Waring, BR Research
Subject Area: General
Report Number: LR-MEP-020
Issue and Date: 13 December 1993
Summary:

Nothing of use – relates to factors in use such as defensive driving, Sanditing and water
jetting
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: The Distribution of Adhesion on BR
Report Author: JD Tunley, BR Research
Subject Area: Low Adhesion Formation
Report Number: VS-278
Issue and Date: June 1990
Summary:

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Fallen Leaves During the Autumn and Their Effect on Railway
Operation – A Bibliography
Report Author: TG Pearce, BR Research
Subject Area: Low Adhesion Formation
Report Number: LR-SAM-032
Issue and Date: June 1995
Summary:

Referenced in AEAT Report AEAT-T&S-2000-101 Issue 1 August 2000 ‘Review of

Initiatives – Wrong Side Track Circuit Failures’ produced for the Adhesion Working Group.

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Adhesion Measurement on London Underground Track : Part 1 –
Adhesion Data
Report Author: JD Tunley, BR Research
Subject Area: Low Adhesion Formation
Report Number: RR-SAM-027
Issue and Date: March 1995
Summary:

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: A Basic Study of Wheel/Rail Adhesion 3 : Observations on the Effect
of Fallen Leaves
Report Author: C Pritchard & DA Smith, BR Research
Subject Area: Low Adhesion Formation
Report Number: CR 47
Issue and Date: July 1970
Summary:

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: 12 Month Adhesion Survey - Low Adhesion Spots
Report Author: SG Watkins & FGR Zobel, BR Research
Subject Area: Low Adhesion Formation
Leaf Contamination Properties & Characteristics
Report Number: IM TRIB 23
Issue and Date: May 1976
Summary:

The correlation of low adhesion (not leaf affected sites) the lineside features has been
attempted in this report largely without visiting the sites in question, and so the correlation
should be treated in some cases with caution. However certain common features occur on
a number of occasions. Probably the most significant is the flange lubricators, which is
known to have adverse effects of adhesion. Out of a total of 37 lubricators on the sections
of the survey route where measurements were made, 7 and possibly as many as 14 can
almost certainly the associated with low adhesion at some time, then clearly this depends
primarily on the state of adjustment of the lubricator. In three cases the effect was
particularly marked.

Points and crossings also appear to affect the measured adhesion. This may be due to
dynamic effects (unloading of wheels) as the train passes, or to excessive or angry is
being present.

In some cases level crossings are implicated. Road vehicles passing over crossings tend
to spread contamination on rails, and the effect is usually localised, and is present where
major roads crossed the track.

Other low spots can be correlated tentatively with cuttings and bridges over the track.
These locations are likely to remain damp and shaded, possibly water training directly
from certain bridges makes a contribution.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Description of the Trackside Water Spray Equipment Installed at
Bearsted Bank – 1978
Report Author: D Pollicott & RK Taylor, BR Research
Subject Area: Leaf Fall Properties & Characterisation
Report Number: TM TRIB 34
Issue and Date: July 1979
Summary:

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: The Effectiveness of Trackside Water Sprays on Leaf Affected Track
at Bearsted Bank 1979
Report Author: RK Taylor, BR Research
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: TN TRIB 43
Issue and Date: September 1980
Summary:

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Adhesion and Leaves: Use of Sandite
Report Author: I J McEwen
Subject Area: Leaf Contamination Properties and Characteristics
Report Number: Technical Memorandum Trib 22
Issue and Date: June 1977
Summary:

Referenced in report RR-TRS-98-004 ‘An Assessment of Low Adhesion Railhead

Treatment Methods Using Data From Service Vehicles’ Issue 1 March 1998, produced for
Railtrack PLC.

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Failure of Track Circuits due to Rail Contamination
Report Author: GG Graver, K Kendall, BRB Chemical Research Division
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: CR53
Issue and Date: May 1971
Summary:

Referenced in AWG Report AEAT-T&S-2000-101

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: 1989 Leaf Fall and Associated Track Circuit Problems Investigation
Report Author: M Brewer, BR SR Signalling Performance Section
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: 13-LEAF
Issue and Date: January 1990
Summary:

Referenced in AWG Report AEAT-T&S-2000-101

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: High Pressure Water Jetting Trials' for Balfour Beatty
Report Author: AEA Technology Rail
Subject Area: Leaf Contamination Properties & Characteristics
Report Number: AEATR-T&S-2002-134
Issue and Date: August 2002
Summary:

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Adhesion Modification Tests Autumn 1993 Part 4: Centrac HPF
Results
Report Author: JR Waring, BR Research
Subject Area: Friction Modifiers
Report Number: LR-MEP-021 Part 4
Issue and Date: February 1994
Summary:

This document describes tests carried out during November 1993 with Centrac HPF, a
solid substance running on the wheel tread, which is claimed to modify adhesion to a
characteristic that increases from a value of 0.17 at to 2.5% slip to a value of 0.38% slip.

The track trials reported had the objective of investigating whether the Centrac HPF could
improve the prevailing level of adhesion during braking in autumn conditions with various
levels of leaf contamination, and if so determined the magnitude of the improvement and
its variation with adhesion level. The results indicate that overall it does not have a
beneficial effect in overcoming low adhesion due to Leaf contamination of the railhead.

The wear rate of the Centrac HPF blocks would result in a replacement interval of about
1680 miles. This is unacceptably short, as a typical EMU examination interval this to 2800
miles.

Unfortunately, weather and leaf fall patterns during the available test period resulted in a
few measurements being made at very low adhesion levels (less than 0.06), which are the
most significant for poor braking performance. It is time is available during any future low
adhesion tests then it would be useful to obtain further results.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Adhesion Modification Tests Autumn 1993 Part 5: Lessons Learned
Report Author: JR Waring, BR Research
Subject Area: Friction Modifiers
Report Number: LR-MEP-021 Part 5
Issue and Date: March 1994
Summary:

Summary report for various work streams

Tested Centrac HPF friction modifier – Concluded that the claimed results were not found
under typical autumn conditions. Results showed little change in average shape of the
adhesion/slip curve and under most test conditions a reduction in adhesion throughout the
slip range.

Rate of HPF block wear expected replacement after 1700 miles suggesting high labour
costs for replacement outwith normal maintenance cycle.

Century Oils (supplier) state 5650 to 8080 miles in service use elsewhere.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: An Investigation of the Effect of Wheel Load on the Persistence of
Railhead Leaf Film
Report Author: W Poole, BR Research
Subject Area: Vehicle Design Influences
Report Number: LR/VDY/92/087
Issue and Date: May 1992
Summary:

The conclusions are based on a single test for each load case and should be interpreted
carefully.
• A compacted layer of leaf debris and metal/metal oxide has been formed on a railhead
in the laboratory which is similar to that fall on real track.
• Tests with 5 ton load showed that it required up to 300 wheel passes to remove the
leaf film and give good electrical contact throughout the length of the contaminated
rail.
• The leaf film in the 5 ton load tests was persistent under low creepage conditions
• tests with a 10 ton load produced a complicated result within insulated layer been
formed after leaves had ceased to be applied to the railhead. This may be partly due
to the way in which leaf combines with metallic wear debris to form the hard black
insulated film under high load conditions. Hence the passage of every freight train
between low wheel load passenger trains could exacerbate the leaf film problem
• under low creepage conditions and a 10 ton load, about 100 passes of the wheel was
sufficient to break the film up. A doubling of low therefore remove the film in about a
third of the wheel passes
• a heavy large freight train of about 25 bogie vehicles would be required to treat
railhead leaf film. Repeated passes of the train over the leaf affected side would
probably be required in order to maintain a clean surface
• laboratory testing cannot simulate the rate at which leaves contaminated railhead.
Contamination depends heavily on the weather conditions at the time of leaf fall, the
trackside vegetation and topography, aerodynamic characteristics and frequency of
passing trains. Therefore freight trains could contaminated rail at a faster rate than the
removal rate.

Samples of the hard black film form during the tests were analysed to determine their
content. Analytical services carried out a Fourier Transform infrared spectroscopy to
identify the organic content of the debris and x-ray fluorescence and x-ray Diffraction to
identify the organic content. It was found that all three samples were very similar to that
they contained cellulosic material, significant amount of calcium oxalate and ferrous
debris. The cellulosic material and calcium oxalate are indicative of vegetable matter and
ferrous material will be from wheel and rail wear. Only very small samples could be
collected surly a very rough estimate could be made on the relative proportions of the
debris constituents. In all the samples there was a 60 percent metal and a 40% of
vegetable content. The fact that the debris contact of vegetable and metallic material may
indicate why the hard black layer can form without further additional leaves in that rolling
is required to compact the two components together. The constituents found in the
laboratory debris is very similar to that determined from debris collected from the track.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Plan View Vehicle Dynamic Behaviour and Track Circuit Operation
Report Author: RG Gostling & BM Eickhoff, BR Research
Subject Area: Vehicle Design Influences
Report Number: IM-CPC-009
Issue and Date: November 1989
Summary:

Referenced in AEAT Report AEAT-T&S-2000-101 Issue 1 August 2000 ‘Review of

Initiatives – Wrong Side Track Circuit Failures’ produced for the Adhesion Working Group.

Reports on improved riding characteristics of modern vehicles showing an effect on the

wheel/rail interface leading to less breakdown of the surface contamination.

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: A Summary of the Fundamentals of Track Circuit Behaviour
Report Author: TG Pearce, BR Research
Subject Area: Vehicle Design Influences
Report Number: VDYN 254
Issue and Date: February 1991
Summary:

Referenced in AEAT Report AEAT-T&S-2000-101 Issue 1 August 2000 ‘Review of

Initiatives – Wrong Side Track Circuit Failures’ produced for the Adhesion Working Group.

Reports on improved riding characteristics of modern vehicles showing an effect on the

wheel/rail interface leading to less breakdown of the surface contamination, specifically
tread surface roughness on rust films.

States that careful measurement of wheel surface revealed that peak-to-trough tread
roughness was a factor, with rougher surfaces making better electrical contact. Disc
braked vehicles had treads typically 10 microns peak-to-trough roughness while tread
braked vehicles had typical roughness of 20 microns. Tyre turning specification is for the
roughness after turning to be less than 50 microns. This work relates to rust films and not
leaf films.

REPORT NOT AVAILABLE

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Full Scale Leaf Tests at Old Dalby – Nose End Experiments with
Classes 156, 158 and 165
Report Author: ND Humphreys & DW Read, BR Research
Subject Area: Vehicle Design Influences
Report Number: LR-AER-042
Issue and Date: 28 April 1992
Summary:

As a result of the large number of track circuit operation failures in the autumn of 1991,
particularly by the Class 158, full-scale tests have been carried out on the Old Dalby test
track. Classes 156, 158 and 165 2 car DMUs, with different front-end configurations, were
tested, passing over sheets of paper laid on the track bed representing leaves.

Three colours of paper were placed in four foot, six-foot and the cess so that after the train
had passed the numbers of leaves left, moved to and from each region and squashed on
the railhead could be counted.

It was found the presence and type of obstacle deflector/air dam fitted to the front of the
train had a significant effect on the numbers of leaves squashed on the railhead and in the
movement of leaves from one area to another.

The presence of a bluff obstacle deflector, has fitted to the Class 158, or the snowplough
as fitted to the Class 156, increases the number of leaves squashed on the railhead by a
factor of about 4 compared to the open configuration of the Class 156 without its
snowplough or any other obstacle deflector.

The Class 165 with its curved obstacle deflector was successful in not increasing the
number of leaves squashed on the railhead compared to the open configuration of the 156
without its snowplough or any other obstacle deflector.

The newly fitted integral snowplough obstacle deflector to the Class 158 results in even
larger numbers of leaves being squashed on the railhead compared to a standard
configuration.

The number of leaves left squashed on the railhead after a 2 car sprinter type DMU has
passed is highly dependent on both the presence and type of air dam fitted to the front of
the train. A snowplough fitted Class 156 squashed 3.5 times as many leaves to the
railhead compared to the same unit without any such device. The Class 158 fitted with
standard fibreglass air dam squashes slightly less numbers of leaves on the railhead
found the snowplough fitted Class 156, but still a factor of three times greater than the 156
without snowploughs. Removing the fibreglass air dam from the Class 158 but leaving the
substantial metal structure behind, which visually provides nearly as much blockage,
increases the number of leaves squashed on the railhead to the same as the snowplough
fitted Class 156. The newly designed integral snowplough type air dam fitted to the Class
158 results in an even greater number of leaves deposited on the railhead. This is worse
by factor of 1.7 on the number deposited by the standard air dam fitted Class 158 and
hence a factor of four times worse than the Class 156 without snowploughs. The Class
Appendix 5 Library Search Results – AEA Technology Rail

165 fitted with its standard air dam, which is significantly different to that of the Class 158
unit, only squashes a similar number of leaves to the railhead as the Class 156 without
snowploughs.

All the types of unit and configuration tested, the effect on the leaves seemed to be only
significant at least half a vehicle length downstream, the leaves remaining on the ground
until half vehicle has passed. The air dam must therefore condition the flow much further
downstream that may be expected.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Further Analysis of the Full-Scale Leaf Tests at Old Dalby with
Classes 156, 158 and 165
Report Author: ND Humphreys, BR Research
Subject Area: Vehicle Design Influences
Report Number: MB01183-07
Issue and Date: June 1992
Summary:

Analysis of the videos taken of the simulated Leaf movement tests carried out on the Old
Dalby test track during 1992 has revealed that the leaf movement varies considerably
between different Classes and configurations. This variation is much larger than expected.
However a systematic difference has been identified to explain some of the differences.

In particular the leaf movement that causes the Class 158 with obstacle deflector, all the
integral snowplough, and Class 156 with snowplough fitted, to squashed many more
leaves on the railhead compared to other units or configurations has been identified.
Leaves in the cess are initially pushed out as the front of the train and first bogie passes at
other drawn back to the railhead as the open region of the latter half of the first vehicle,
beneath the sole bar, passes. Two strong small vortices in the nip of the railhead caused
the leaves in four foot and in the cess to be captured and held in this region resulting in a
number of them being run over by the vehicle wheels.

The Class 156 in the open configuration and Class 165 with its standard air dam fitted,
both produce completely different leaf movement from each other and from the 158 or 156
with snowplough, such that for the Class 165 and Class 156 in the open configuration, few
leaves are transferred across the railhead.

The present very limited understanding of the leaf movement caused by the front air dam
are as follows:
1. The more bluff the air dam/obstacle deflector the stronger vortices created by it and
hence the trapping of more leaves. For Class 158 these are trapped in the wheel/rail nip
area and are therefore run over by the train. It to alleviate this problem the radius of the air
dam/obstacle deflector sites and base should be increased. Only a relatively small
increase in radius may be required to substantially reduce the strength of these vortices in
the wake of the obstacle deflector.
2. The actual size, position, strength and direction of rotation of the vortices created
underneath and to the size of the train are a consequence of the geometry of the air dam.
It is felt that sweeping the air dam back, as on the Class 165, may be the main
mechanism that causes the vortices to be completely different regarding position and
direction of rotation to that on the Class 158. This may be a general method that stops
leaves being aerodynamically moved into the wheel/rail nip region.

The report suggests standard aerodynamic model tests in which different designs of air
dam and other devices can be tried with the aim of producing vortex strengths and
positions that will not track leaves in the wheel/rail nip region.
The aerodynamics of general wheel/nip leaf guards are also discussed.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Report on the Leaf-Deflector Tests Conducted with a Class 165 Turbo
Unit Between Princes Risborough and Bicester
Report Author: W Poole, BR Research
Subject Area: Vehicle Design Influences
Report Number: LR/DS/93/015
Issue and Date: March 1993
Summary:

The report details tests conducted a Class 165 fitted with bogie mounted leaf defectors.

The conclusions of this report are:

1. Leaf defectors have been demonstrated to be effective in reducing the amount of leaf
coverage on the railhead. The improvement obtained for real, large sycamore leaves was
about 70 percent.
2. Results have only been obtained for a vehicle in at air condition.
3. The majority of leaves rolled of the railhead originate from the four foot. Leaf counts
showed unequal numbers of leaves crossed both rails from the four foot.
4. Leaves crossing the railhead from the set-aside will not propelled into the wheel rail nip.
5. Removal of leaf from the wheel tread may have taken place because of contact
between the deflector at wheel.
6. The defect to design requires some strengthening to prevent the tapering section
becoming damaged.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Report on the Further Development of Leaf-Deflectors
Report Author: BR Research
Subject Area: Vehicle Design Influences
Report Number: 009/02
Issue and Date: June 1993
Summary:

The report details the results of the extended trials on the prototype leaf defect was and
problems revealed with the existing design. Solutions to these problems have been
addressed at a set of modifications to the deflector and the support rackets have been
proposed.

Finally further investigations into alternative materials of revealed to more which may be
worth considering. The
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Four Car Model and Train Braking Control
Report Author: PD Horne, BR Research
Subject Area: Whole Train Braking/Acceleration Concept
Report Number: RR-SAM-016
Issue and Date: March 1995
Summary:

The established British Rail Research Wheel Slide Protection Evaluation Rig (WSPER)
has recently been enhanced and can now model trains of up to four vehicles, one-shot
sanders, and train braking control systems.

A train braking control system utilises a controller which monitors the achieved train
deceleration and attempts to meet the driver's deceleration demand, under all adhesion
conditions, by modulating the train brake demand. The system relies on the existing
WSPs on each vehicle to control the brake pressure of individual axles to avoid excessive
wheel slide or lockup.

This report describes tests conducted with the enhanced WSPER to compare the
stopping performance of trains fitted with various braking systems: no WSP, state-of-the-
art WSP, one-shot sander, train braking control system, and idealised WSP.

Conclusions drawn from the test results are:

• The brake system with the longest stopping distances was the no-WSP system. The
one-shot sander achieved the shortest stopping distances.

• The state-of-the-art WSPs performed very well, only allowing the wheelsets to lock
up on continuous lengths of very low adhesion. However, the train with idealised
WSPs achieved a mean stopping distance 12% shorter than that of the train with
state-of-the-art WSPs.

• The train braking control system did not perform as well as expected, mainly, it is
thought, because the controller increases the brake demand above 9%g, and the
state-of-the-art WSPs used are only optimised for a maximum demand of 9%g. Even
with WSPs optimised for a higher brake demand, a train braking control system will
only perform significantly better than a train without a controller under a limited range
of adhesion conditions.

The following recommendations are made for braking in low adhesion conditions:

• A sander should be considered for implementation for use in very low adhesion
conditions, subject to a cost-benefit analysis and a successful outcome of the
planned fleet trials.

• More responsive brake systems (hydraulic or electric) and lower inertia wheelsets,
Appendix 5 Library Search Results – AEA Technology Rail

together with an accurate speed reference, should be considered for new rolling
stock. This would allow the WSP to control the slip level more accurately and reduce
stopping distances.

It is not worth developing the Train Braking Control concept further solely to obtain
improved stopping performance
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Email communication from John Tunley
Report Author:
Subject Area: Whole Train Braking/Acceleration Concept
Report Number:
Issue and Date: 21 April 2004
Summary:
It is necessary to establish if an ABC utilising deceleration feedback would be beneficial to
operating performance in low adhesion conditions. The operation of an ABC could be as
follows: When the driver's demanded deceleration is not achieved the brake demand
would be increased in the hope that the rear of the train could support that higher brake
demand and increase the overall deceleration. The WSP would modulate the brake
pressure down the train to that sustainable at each wheel/rail interface, thus negating any
further need for intelligent control of brake demand down the train. If an insufficient
increase in deceleration was achieved, the optimum amount of sand would be deployed to
compensate for the adhesion deficit.

One way to find out if ABC offers a real benefit would be to get more data on the
conditioning effect down the train. Also, it is necessary to establish how often we could
benefit from such a system and if it is only applicable to long trains or if shorter formations
can also benefit.

We suggest that a longer train set should be instrumented for the autumn to measure the
mean WSP regulated brake pressure on (say) one axle of each vehicle (similar to the
Pendolino tests we conducted for Alstom). With a good WSP, in low adhesion conditions
the brake pressure is a direct indication of the adhesion 'seen' by that axle. It would then
be possible to plot the effective adhesion down the train, taking into account the direction
of travel and the brake demand. Also, by disregarding the data for the vehicles at the back
of the trainset, the likely effect on shorter train sets could be established.

An initial trial would be conducted on one or more artificial contaminants (paper tape and
detergent say). The unit would then be released to service traffic with measuring
equipment on-board and operational. This would allow the conditioning effect to be
established for a range of natural autumn adhesion conditions and the ABC benefit could
then be assessed.

Even if the answer is not good for ABC, the conditioning data would be very useful
common domain information as this topic it is currently not well documented.

A long train set with a modern WSP would be essential for such a trial. There must be no
sander fitted as this will modify the result. Pendolino would be ideal for such an exercise
and 5-car Voyager with one shot sander would be an acceptable second choice as long
as the (infrequent) firing of the sander was recorded.

A parallel work stream could be to look at the possibilities for acceleration feedback as
this must be sufficiently accurate in very low adhesion (1%g) and should also compensate
for gradients. The latter could introduce a 2%g error (on a 1:50 gradient - there are a few
locations even steeper).
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: A Review of the Costs and Benefits of a Whole Train Braking System
Report Author: TR Davidson, BR Research
Subject Area: Whole Train Braking/Acceleration Concept
Report Number: LR-MEP-014
Issue and Date: October 1993
Summary:

Only considers cost benefit issues

Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: Verbal report from John Tunley
Report Author:
Subject Area: Adhesion Prediction & Communication
Report Number:
Issue and Date: 20 April 2004
Summary:

John Tunley advises that the Dutch rail operator uses AEA Technology Rail in Utrecht to
manage LAWS for them. There are 17 LAWS fitted vehicles operating. AEA man the
LAWS terminal and if LAWS identifies a low adhesion problem then the AEA person
contacts the regional control office and provides details of where and when the low
adhesion event occurred. The regional controller has access to a look-up system and can
then determine which trains and an associated GSM mobile will traverse the route over
the following 2 hour period. An SMS alert is then sent to these mobiles detailing the route
and kilometre position of the low adhesion so as the Driver may take defensive action.
Similarly, if a Driver radios in and reports low adhesion, then the regional controller
operates the same communications system.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: The Feasibility of Adhesion Control and Monitoring for ERTMS
Report Author: W Poole, AEA Technology Rail
Subject Area: Adhesion Prediction & Communication
Report Number: AEATR-VTI-2003-015
Issue and Date: Issue 1 March 2003
Summary:

Report produced for RSSB.

It is vital to know that there is sufficient adhesion available to sustain the required traction
and braking rates for operating a safe railway. Adhesion is therefore an important element
that requires incorporation into ERTMS and this report investigates the issues that need to
be addressed in order for that to the feasible.

The feasibility of providing adhesion information to ERTMS has been established by

reviewing the critical factors that are required for a decision process. Critical adhesion
factors have been categorised under Vegetation, Weather, Track, Rolling stock and
Railhead Contaminants.

It is concluded that an adhesion management system should preferably be able to

automatically assess the current adhesion conditions in real-time and also predict
changes. In the current proposals for ERTMS a reduced brake rate the specified as being
70% of that utilised under normal adhesion conditions. The adhesion management system
would then provide a mechanism by which the brake demand level can be switched
between normal and low adhesion mode and vice versa when adhesion conditions
change. The system should also be able to provide early warning of low adhesion to
enable the management of mitigation measures.

The essential process is to monitor the factors that determine the wheel/rail adhesion level
and process this information to formulate a decision on level of brake force demand that
can be sustained. This is then supplied to ERTMS. However, a decision process can also
be used to direct mitigation measures that will form part of a feedback loop since they will
influence the inputs to the decision processor. Furthermore, once ERTMS has adjusted
the brake rate it will be useful for this to be confirmed both as a safety check and for use
in the decision process. Feedback of the brake rate status will enable the AMS to operate
the switch in both directions, that is from high to low adhesion and vice versa.

The process of assessing conditions, through adhesion inputs to a decision matrix, will
provide a semantic scale of adhesion levels, e.g. normal, adverse, severe, very severe. It
is unlikely that specific values of adhesion coefficients will be available to ERTMS on a
continuous basis.

The consideration of the above concept has identified a number of areas that require
development before such a system can be made available to railway administrations. A
number of building blocks currently exist that will provide the basis for a prototype
ETAMS. LAWS coupled with weather information, in particular moisture and rainfall levels,
railhead contamination levels and vegetation assessments will form significant inputs to
Appendix 5 Library Search Results – AEA Technology Rail

propose system. Experience with the ACCAT developed by Metronet BCV, provides
valuable information for building an adhesion management system.

The system should preferably shadow a trial route during the ERTMS early deployment
scheme. This will permit the relative weightings of the various inputs to be adjusted to
replicate as closely as possible the actual service conditions. The issue of the distance
over which to match the brake rate so adhesion conditions will need to be addressed
during this development phase of the system. Also any new inputs or knowledge of
managing from other investigations may be incorporated where appropriate at this stage.
Predicted algorithms may be developed from the space data.

It is recommended that automatic methods should be developed for monitoring leaf fall
and railhead contamination. Some techniques have been identified that may readily
provide the means for automating monitoring of these factors, but a detailed assessment
should now be carried out to determine if the capability of the systems is sufficient for
application in ETAMS. Once a prototype system has been demonstrated, a longer term
objective should be to establish the hardware platform on which to build a standard
system for application to any route selected for ERTMS.
Appendix 5 Library Search Results – AEA Technology Rail

Source: AEA Technology Rail

Report Title: An Overview of Low Adhesion Factors for ERTMS
Report Author: W Poole, AEA Technology Rail
Subject Area: Adhesion Prediction & Communication
Report Number: AEATR-VTI-2003-030
Issue and Date: Issue 1 July 2003
Summary:

Report produced for RSSB.

This report reviews the results of the completed parts of the RSSB ERTMS project
"ERTMS & Adhesion Management" to bring together the significant conclusions of each
work area. The measures required to address the adhesion demands have been identified
and these are now ranked in importance at ease or difficulty of achieving them. The
review has included the results from the following work streams:
• automatic train protection, automatic train operation and ERTMS experience
• low adhesion slip risk
• odometry issues
• adhesion management system
• sander and WSP systems

The report also includes new information that deals with:

• additional analysis of slip risk data supporting the basis for the proposals to reduce
the number of low adhesion SPADs to ALARP
• prioritisation of the developments required for measuring the key inputs to an
adhesion management system
• a review of the railway group standards, carried out in the light of the above
summary in order to identify any changes necessary all the requirement for a new
standard
• any issues pertinent to the conventional railway and of interest to the adhesion
working group

3 important recommendations for predicting and managing low adhesion conditions are
made from the results of this work:

1 based on the ERTMS proposals for a 70% of low adhesion brake rate, the following
should be adopted for ERTMS low adhesion operation to reduce the risk of a SPAD or
overrun to ALARP:
• a low adhesion brake rate of 6% g
• incorporate an additional 200 m margin in advance of the 6% g braking point and,
where required, include an allowance for gradient.
The use of the 200 m margin should be reviewed once operational experiences gained

2 a prototype adhesion management demonstration system should be developed in the

basis of existing data from system such as LAWS, weather stations etc. The development
Appendix 5 Library Search Results – AEA Technology Rail

should be based on approach is given in section 3 1 of this report.

3 it is suggested that most of the safety and performance benefits as detailed above could
be achieved on a non ERTMS railway line:
• the adoption of low adhesion braking point signs placed at the 6% g braking point
(plus a 200 m margin) at critical low adhesion signals. This would reduce the
reliance on individual judgement of distances and braking performance required by
every driver, which is bound to introduce an increased SPAD risk. Technical,
human factors and liability issues should be considered before advisory braking
signs are introduced
• the adoption of a European train adhesion management system ( ETAMS) to
provide clear and concise warnings to drivers of the prevailing adhesion conditions
and allowing them to drive accordingly. However, reliable communications
between host and train and vice versa are likely to be the greatest barrier to the
adoption of such a system.

Other recommendations made into the following:

• The use of trainborne Sanders to improve adhesion will enable the 200 m margin
to be reduced. However, the sander system may then began safety critical and
reliability of the systems would have to be considered in the context.
• For braking case, optimise WSP systems should be useful trains operating under
ERTMS to give the require odometry accuracy and to ensure reliable sander
operation. It is also recommended that the wheel diameter calibration factor should
be utilised to ensure the accuracy of the odometry system.
• The traction case, it is recommended that only non-powered axles should be used
for ERTMS odometry purposes, again taking into account possible wheel diameter
errors. Where all acts as a powered, it is suggested that other adhesion
independent odometry sources should be considered.
• It is recommended that the need for developing a professional driving policy for the
industry should be evaluated. ERTMS will introduce new driving techniques
because improved low adhesion information will be available to the drivers from
ETAMS. The human factors study will be required to determine how drivers should
use this information.
• The group standards shall continue to be based upon normal operating adhesion
levels, but additions and amendments should be incorporated to clarify the
standards with reference to the requirements for ERTMS operation.
 Rail Safety and Standards Board Evergreen House 160 Euston Road London NW1 2DX
Reception Telephone +44 (0)20 7904 7777 Facsimile +44 (0)20 7904 7791
www.rssb.co.uk

Rail Safety & Standards Board Registered Office: Evergreen House 160 Euston Road London NW1 2DX. Registered in England and Wales No. 04655675.

Rail Safety & Standards Board is a not-for-profit company limited by guarantee.