Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Codified Methods To Analyse The Failures of Water Pipelines: A Review

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/281603098

Codified methods to analyse the failures of water pipelines: A Review

Conference Paper · November 2014

CITATIONS READS

2 1,672

5 authors, including:

D.J. Robert Pat Rajeev


RMIT University Swinburne University of Technology
144 PUBLICATIONS 1,710 CITATIONS 189 PUBLICATIONS 3,477 CITATIONS

SEE PROFILE SEE PROFILE

Chun Qing Li Jayantha Kodikara


RMIT University Monash University (Australia)
245 PUBLICATIONS 5,123 CITATIONS 395 PUBLICATIONS 6,985 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by D.J. Robert on 09 September 2015.

The user has requested enhancement of the downloaded file.


Codified methods to analyse the failures of water pipelines: A
Review1 1 2 1 3
P. Forsyth , D.J.Robert , P. Rajeev , C.Q.Li and J. Kodikara
1
School of Civil, Environment and Chemical Engineering, RMIT University, VIC 3001, Australia.
2
Department of Civil and Construction Engineering, Swinburne University of Technology, VIC 3122,
Australia.
3
Monash University

Abstract: Pipelines used for the transportation of water and other services are very important lifelines
in modern society. The important role that they play in our present economy is reflected by thousands
of kilometres of service laid in urban centres worldwide. As many of these pipelines have been laid
few decades earlier, in most cases, their condition has been deteriorated mainly by electro-chemical
and (or) micro-biological corrosion. There are several codes of practice available at present to
analyse the condition of such deteriorated pipes. This paper identifies and reviews all such relevant
standards applicable to corroded water mains with the use of some case studies. Three dimensional
finite element analyses are also conducted to investigate the capabilities of simulating the failures of
corroded pipelines and thus to ascertain the validity of codified methods. By careful assessment of the
current codes of practice in use, it is possible to understand where these codes are lacking and which
codes can rightly predict the realistic water pipe failures observed in the past.

Key words: Finite element modelling, pipeline failure, cast iron pipeline, corrosion failure, ABAQUS,
fitness for service, Failure assessment diagrams

Introduction
Water authorities around the world are faced with the issue of ageing water distribution networks, with
pipeline failure a more prevalent issue today, leading to a severe loss of delivery efficiency. The
factors that contribute to a specific pipe failure can be categorized in three principal groups: (a) pipe
geometry, material type, pipe-soil interaction and quality of installation, (b) internal loads due to
operational and transient pressure and external loads due to soil overburden, traffic loads, frost loads
(in cold climate) and third party interference (catastrophic loads), and (c) material deterioration due
largely to the external and internal chemical factors; this includes bio-chemical microbiological and
electro-chemical activities that lead to corrosion [9]. In addition, the failures can manifest as a result of
the combination of one or more factors among such principal groups. For instance, heavy transient
event can burst the pipe at the corroded location where there is a deteriorated wall thickness. If these
failures can be predicted accurately in advance, not only will water utilities benefit on cost savings, but
also society will be served with improved efficiency.
Many authorities and bodies have collaborated and developed a set of codes [2,3,4], which can be
used to assess the ‘fitness’ of a pipeline network for operation. They have developed and set out
uniquely, in order to convey their own method of assessments to analyse corroded water mains. The
corrosion in these water mains can be categorised into three types; general corrosion (also referred to
as uniform corrosion), patch corrosion and pitting corrosion [8], with each type of corrosion having
differing effects on the structural integrity of the pipeline. The codes of practice choose to assess
these types of corrosion in many different ways, adding to the problem of what is the best assessment
method.
The current study investigates the applicability of the codes of practice to predict the failure of
corroded cast iron water pipes. Due to the nature of cast iron and the size of these pipes, their failure
mode is generally attributed to material deterioration through corrosion which could fall into one of
aforementioned categories. Codified methods identify ways of analysing such material deterioration
and hence pipeline’s fitness for service. This paper identifies and reviews all relevant standards
applicable to corroded cast iron water mains which fall into pre-identified corrosion patterns. The
codified methods are assessed against a past pipeline failure to explore the capabilities of the codes

529
to predict the actual pipe failures. Three dimensional finite element (FE) analyses are also conducted
to simulate the failures of corroded pipelines and thus to ascertain the validity of codified methods. By
careful assessment of the current codes of practice in use, it is possible to understand where these
codes are lacking and which codes can rightly predict the realistic water pipe failures observed in the
past.

Methodology
This section identifies the description of the relevant standards for cast iron water mains that are
corroded under three different corrosion patterns as identified in Rajeev et al [8]. Corrosion is one of
the main reasons for failure in piping systems. The type of corrosion will determine how the pipeline is
assessed and how it can be treated in order to maximise its service life.

General Corrosion
General corrosion or uniform corrosion refers to reasonably uniform reduction of thickness over the
surface of the pipeline wall. Figure 1 shows an example of a corroded cast iron water main for general
corrosion type.

Figure 1 Example of general corrosion in cast iron water mains [8]

API/ASME 2007
API code [2] suggests different levels of assessments for general corroded pipelines. Level 1
assessment, which is suggested as an initial assessment, is proposed to undertake either using point
thickness reading (PTR) measurements or critical thickness profile measurements. The type of
method is determined using the coefficient of variation of the thickness readings (i.e. former method if
coefficient of variance, COV<10% and later method if COV>10%). If the pipe is not satisfied under
Level 1 Assessment conditions, the code suggests to undertake Level 2 assessment (acceptance
criteria for Level 2 is given in Table 4.4 in API [2]). Level 3 assessments are proposed if the pipe does
not satisfy the level 1 & 2 assessment criteria. The detailed descriptions of each level of assessments
are given in section 4.4.2 4.4.3 of API [2].
BRITISH STANDARD
The British Standard [4] discusses about the assessment of general corrosion in a slightly different
way by providing an annex (annex G in [4]) to deal with all types of corrosion. This standard
categorises general corrosion as a corroded region in which metal loss is less than 10% of the original
wall thickness. The code gives suggestions to determine the safe working pressure. This is suggested
in a few different ways depending on the flaw type and condition. For a single flawed section, safe
working pressure is determined by first calculating the failure pressure of the unflawed pipe, by using
2𝐵0 𝜎
the equation 0 = , where 𝐵0 = Original measured pipe wall thickness, 𝜎 = 𝑡𝑒𝑛𝑠 𝑙𝑒 𝑠𝑡𝑟𝑒𝑛 𝑡
( 𝐵0 )
and D= diameter of the pipe. Then, the failure pressure of the corroded pipe is determined by
multiplying the unflawed failure pressure with the calculated reserve strength factor. Finally, a factor of
safety is applied to produce a safe working pressure, where the factor of safety is determined by
multiplying the modelling factor of the corrosion flaw by the original design factor of the corrosion flaw.
530
ASME B31.G
ASME B31.G [3] is an ageing code of practice but still is an effective tool when undertaking corrosion
assessments and subsequent remaining life assessments. However, it does not have separate
assessment procedures in order to assess the different types of corrosion flaws. The code can be
utilised for general type of corrosion by assessing the amount of corroded area. If the corroded area
under assessment results to be less than 10% of the nominal wall thickness, then the code suggests
no action and the pipeline can be returned to service and considered to be safe. On the other hand, if
it is found to be 80% or greater, immediate replacements or repair actions are proposed. For the range
in between, it suggests further assessments by determining the longitudinal extent of the corroded
area. Part 3 of the code suggests tables of corrosion limits which can be an effective tool to quickly
determine whether a pipeline is suitable for continued service under the required maximum allowable
operational pressure (MAOP).

Patch Corrosion
Patch corrosion (Fig. 2) is identified as a patch of corrosion due to graphitization or cluster of
geometrically interacting pits, which can be approximated as a patch of corrosion.

Figure 2 Example of patch corrosion in cast iron water mains [8]

API/ASME 2007
Part 5 in API code [2] deals with patch corrosion or as it is known in this standard as local metal loss.
The code suggests Level 1 assessment to accommodate for patch corrosion. The first step in this
assessment is to determine the critical thickness profile (CTP) for the affected area. In the case of a
local thinned area (LTA), a grid is suggested in the code to determine thickness readings and
establish a CTP in both the longitudinal (s) and circumferential (c) directions. Once the minimum wall
thickness is determined, the remaining thickness ratio (Rt) and longitudinal flaw length parameter ()
𝑡 𝐹𝐶𝐴 12
can be determined from the equations 𝑅𝑡 = 𝑛𝑑  = as suggested in the code.
𝑡 𝑡

The acceptable criteria for Level 1 assessment are 𝑅𝑡 0 20 and 𝑡 − 𝐶 25 and 𝑠𝑑


1 8 𝐷𝑡 , where Rt is the remaining thickness ratio, FCA is future corrosion allowance, Lmsd is the
distance to the nearest major structural discontinuity and tc is the corroded wall thickness. If Level 1 is
satisfied, the code suggests assessing MAWP as elaborated in Annex A of API [2]. Figure 5.6 in API
[2] determines whether the calculated MAWP is acceptable for the corroded pipe. Further details of
this method can be found in section 5.4.2 in API [2]. The code suggests level 2 & 3 if the pipe in
concern is not satisfied Level 1 criteria. The detailed descriptions of level 2 & 3 can be found in section
5.4.3 and 5.4.4 respectively in API code.

BRITISH STANDARD
With regards to the patch corrosion, similar strategy has been suggested in British standard as
applicable with general corrosion explained earlier (Annex G in [4]). Additional assessments are

531
suggested if the results from the safe working pressure assessment are found to be inadequate for the
operational specifications.

ASME B31.G
As this code does not differentiate the assessment procedures for different types of corrosion flaws,
the assessment procedure for patch corrosion is similar to that applicable for general corrosion.

Pitting Corrosion

Pitting corrosion can be defined as localized regions of metal loss such as shown in Fig. 3.

Figure 3 Example of pitting corrosion in cast iron water mains [8]

API/ASME 2007
Part 6 of API code [2] discusses the assessments in relation with pitting corrosion.
First step here is to determine the required initial data, that being diameter of the pipe, future corrosion
allowance, nominal thickness and previous recorded metal loss. A ratio is determined between the
𝑡 𝐹𝐶𝐴
remaining wall thickness to the future wall thickness; 𝑅 𝑡 = , where 𝑡 = 𝑡𝑛 − 𝐶 −
𝑡
and 𝑎𝑥 is the maximum pit depth in most serious pitting damage, where FCA is the Future
corrosion allowance and LOSS is the amount of wall thickness lost reported from the previous
inspection. If 𝑅 𝑡 < 0 2, then the level 1 assessment criteria is not satisfied. If the criterion is satisfied,
MAWP is calculated using the same methods explained earlier. From the pit charts and the tables
provided in section 6-21 through to 6-28 in [2] (dependant on the severity of the pitting damage), RSF
(remaining strength factor) can be calculated. If the pitting damage is more extensive than the
provided figures in the code, RSF is suggested to take similar with Rwt. If 𝑅 𝑅 𝑎 (allowable
remaining strength factor which is 0.90), then the section is proposed as acceptable to operate under
the MAWP, and if 𝑅 < 𝑅 𝑎 , then the section is acceptable to operate under 𝑀 𝑟 which is
𝑅𝑆𝐹
determined by the equation, 𝑀 ( ). Level 2 and 3 assessment procedures can be found in
𝑅𝑆𝐹
section 6.4.3 and 6.4.4 respectively in the API code [2].
The code also suggests undertaking stress intensity fracture oriented analyses which are based on
failure assessment diagrams (FAD) for cracked-like flaws. In FAD diagrams, the horizontal axis (Lr) is
ratio of the applied stress to the stress to cause plastic yielding of the structure containing a flaw, and
vertical axis (Kr) is the ratio of the applied linear elastic stress intensity factor to the material fracture
toughness. Failure is described by limiting line which is a nonlinear function of Lr. The method can be
used to asses fully brittle failure (Kr=1) as well as plastic collapse of material (Lr=Lrmax). Level 3
assessments are proposed if the pipe does not satisfy the level 1 & 2 assessment criteria. Hence, the
code provides different levels of assessments depending on the level of conservativeness required.
Higher assessment levels need more inputs and involve complex calculations in contrast to lower level
assessments. The detailed descriptions of each level of assessments are given in section 9 of API [2].

532
BRITISH STANDARD
The same assessment procedure detailed in patch corrosion (single patch) is suggested to undertake
for pitting corrosion in British standard [4]. The code also suggests three levels of fracture assessment
methods for pit corroded pipes, depending on the input data available and conservatism required.
Further details of these methods can be found in section 7 of British standard [4].

ASME B31.G
The same assessment procedure detailed in general corrosion is proposed for pitting corrosion in [3].

Case Study
A case study of actual pipeline failure has been studied to investigate the capabilities of predicting the
failures using codes. The data of the failed pipe, which was shared by a water utility [8], have been
presented in Table 1. As it is shown in the table, the case study was on the basis of a cast iron water
main, which was installed in 1955, with the diameter and thickness of 375mm and 13.7mm. The
selected pipe has been failed through a single through wall pit.

Table 1 Pipe geometry details and failure information

D (mm) t (mm) H (m) L (m) L’ X w Nature of failure Cause of failure


(mm)
375 13.7 0.8 6.0 140 X 85 Longitudinal Severe
fracture corrosion &
valve operation

L
’ w

Figure 4 Failed pipe (Rajeev et al, 2013)


The results of the assessments are summarised in Table 2. It can be seen that, following API
assessments [2], the pipe was predicted to fail under both Level 1 and 2 assessment procedures. Pipe
is failed in Level 1 & 2 assessment due to the calculated remaining wall thickness becomes null
(𝑅 𝑡 < 0 2). Detailed assessments were conducted to investigate the pipe failure state under stress
intensity factor based API Level 2 assessments as shown in Fig. 5 (assumed fracture toughness is
10MPa m ). The pipe was failed at a water pressure of 0.6MPa. Detailed FE assessments were
conducted in the next section to investigate the pipeline response as recommended in API [2]. Similar
assessments were conducted using British Standard [3] and found that pipe is not suitable to operate
under any circumstances. Furthermore, it is no longer fit for service under ASME B31G assessments
[3], as the pipe component under assessment has lost greater than 85% of its wall thickness.
Therefore, all the relevant standards predict the pipe failure due to the observed corrosion in field.

533
Table 2 Failure assessment results for the case study

Code Level 1 Level 2 Level 3

API Pipe failed Fracture failure as explained Detailed FE analysis are


in Fig. 5 conducted in next section

BS Pipe failed

ASME Pipe failed

Pipe failure
P=0.7MP
P=0.6MP
P=0.5MP

P=0.3MP

P=0.2MP

Figure 5 Failure state of the pipe using API Level 2 assessment

Finite Element Analysis


Model description

Three dimensional (3D) finite element analyses were carried out using ABAQUS 6.11 [1] to obtain the
stress state of the corroded pipes and to predict the failure observed in the field. The soil and pipe
were represented by 8-noded brick reduced integration elements. The behaviour of soil was assumed
as a linear elastic material. i.e., soil is assumed to be over-consolidated and behave elastically during
the light traffic load (or no traffic) anticipated in the selected case study. The behaviour of the pipe was
modelled using both linear elastic model as well as using a specific elasto-plastic constitutive model
available in ABAQUS 6.11 [1] for grey cast iron materials. The soil side boundaries of the FE model
were assumed to be smooth and are located far (i.e., 5m) from the pipe (& traffic loads) to eliminate
534
any boundary effects. Figure 1 shows the mesh discretization (pipe elements = 41216, soil
elements=41310) and model dimensions. The appropriate dimensions and the mesh density of the
model were selected after a number of trials to minimise mesh and boundary effects on the calculated
pipeline stresses. The FE model idealisation of the corroded pipe used to simulate the failures case
has been shown in Fig. 6.

0.8m
6m

2m
6m

Figure 5 Geometry and mesh discretisation of the FE model

L 85mm
140mm
w

(a) Field observation (b) FE model idealisation


Figure 6 Field observed corrosion of the failed pipe along with the FE model idealisation

Material properties

The behaviour of the pipe was modelled using both linear elastic model as well as using a specific
elasto-plastic constitutive model available in ABAQUS 6.11 [1] for grey cast iron materials. The results
from the linear elastic analysis can be served as a basis of comparison with the predictions from the
standards which are mostly based on linear elastic material. The analyses conducted using elasto-
plastic constitutive behaviour for cast iron provide much robust results which can be effectively utilised
to predict the pipe failures.

The behaviour of soil was assumed as a linear elastic material. i.e., soil is assumed to be over-
consolidated and behave elastically during the light traffic load (or no traffic) anticipated in the selected
case studies.

The properties of the pipe and soil used in the current study are shown in table 3. The elastic-plastic
properties of the cast iron material were obtained on the basis of Rajani [12]. The initial yield stress
has been assumed as 20MPa as observed from stress-strain characteristics of cast iron.

535
Table 3 Material properties of the pipe and soil

Pipe
Material E (MPa)   (kg/m3) Behaviour
type
Cast iron 761 0.32 78002 Elastic1/Elasto-plastic1
Soil
E (MPa)   (kg/m3) Behaviour
2
10 0.32 19002 Elastic
2
10 0.32 19002 Elastic
Note 1 Obtained from Rajani [XX]
2
Assumed

Results

The results obtained from the FE analyses for the pipe used in the case study have been presented
and discussed here. The results are compared between elastic and plastic analyses and the prediction
of failures is discussed on the basis of the tensile strength criteria.

Pipe stress states

The evolution of the maximum principal stresses with applied water pressure at the critical element of
the pipe is shown in Fig. 7a. It can be seen that the stress increase in pipe modelled using elastic
approach is linear with applied water pressure. However the elasto-plastic modelling of the pipe
behaviour has less stress increase compared to the results from elastic analyses. This is because the
strain hardening modelling adopted in plasticity analyses in contrast to linear elastic approach used in
the elastic analyses. The stress observed in the nominal pipe has not been affected by the plasticity
as the resulted stresses due to the applied water pressure are less than the initial yield stress (20MPa)
assumed in the current analyses.

The stress contours of the corroded pipe are shown in Fig. 7b, at the water pressure of 1.5MPa. It can
be seen that the maximum stresses are concentrated at the location of the patch in both the cases.
This reveals that there is a substantial stress concentration (factor of 7.5) can result due to operating
water pressures in corroded pipes. i.e. this stress concentration causes the plastic analyses to reach
the failure stresses of the pipe (shown in Fig. 7a).

Pipe failure prediction

The stress paths obtained from the plastic analyses have been shown in Fig. 8. Each of these plots
shows the stress path of the critical element in the pipe in principal stress space as well as deviatoric
stress space. Stress paths are derived from FE analyses data feeding into a Matlab coding that
incorporates the composite yielding of the cast iron plastic model in deviatoric and principal stress
spaces. The stresses of the critical element reveal that they remain elastic until reaching 20MPa and
afterwards subjecting to yield (strain harden) with the evolution of plastic strains. The failure is initiated
once the element stress reaches the tensile strength of cast iron material (104MPa in the current
case). It is to be noted here that significant evidences are available from literature for a lower bound of
the tensile strength for spun type cast iron pipes than what showed in the current study [6,7, 11]. This

536
would argue the fact that the pipes can be failed at lower water pressures than resulted in the current
FE analyses.

The results from plastic analyses depict that the pipe stresses reach the ultimate tensile strength of
the cast iron (~104MPa in current study) at the water pressures of 1.5MPa. Pipes modelled using
linear elastic model showed that a water pressure of 0.9MPa is sufficient for the pipe to reach the
tensile strength capacity. Unfortunately, there is no record made in the field about the water pressures
at the time of failure, but it has been reported that the failure has occurred during a valve operation
event. This raises possibilities of generating high transient pressures. Ruus et al [10] has shown that
the maximum pressure head rise due to sudden valve closures can be as large as
 1.0  static head (rise will be higher for pipes with large wall friction and smaller vale operational
times). Brunone et al [5] also reported that significant transient pressures can be generated due to
valve operations (as high as 1MPa).

250
Elastic Pipe
Elasto-Plastic Pipe
Maximum principal stress (MPa)

200
Nominal Pipe

150

100
Pipe maximum stress reached
failure tensile strength
50

Failure initiated pressure


0
0 0.5 1 1.5 2
Applied water pressure (MPa)

(a) Maximum principal stress evolution (b) Maximum principal stress


with applied water pressure contours of the pipe pressure

Figure 7 Stresses of the pipe using FE analyses


150 0
Stress path of the Stress path of the
330 30
critical element critical element
100
Major Principal Stress (MPa)

300 60
50
Failure tensile strength
Initial yielding at 20MPa (Ultimate yielding) 100 150
200 250
0 50
270 90

-50 Initial yielding at 20MPa

240 120
-100
Failure tensile strength
Subsequent yielding Subsequent yielding
(Ultimate yielding) 210 150
-150
-150 -100 -50 0 50 100 150 180
Minor Principal Stress (MPa)

(a) Principal stress space (b) Deviatoric stress space


Figure 8 Stress path of the critical element in elasto-plastic pipe of case 1 plotted in; (a) Principal
stress space, (b) Deviatoric stress space

537
Conclusion

With the ever growing need to preserve our water delivery infrastructure, it is becoming more critical
that we monitor and control these systems effectively and efficiently. The pipelines, which serve a
huge role within these systems, are required to assess for anticipated failures so that any unwanted
costs and social impacts can be excluded. Unfortunately, the deterioration to pipes by corrosion
cannot be prevented, but is possible to assess and thus to determine their fitness for service using the
codified methods. Several codes of practice are available at present (such as [2,3,4]) to determine the
suitability of pipes for operation. This paper identifies and reviews all relevant standards applicable to
corroded cast iron water mains with the use of some case studies. Three dimensional finite element
(FE) analyses are also conducted to investigate the capabilities of simulating the failures of corroded
pipelines and thus to ascertain the validity of codified methods.

Although each code under analysis successfully predicted the failure of the pipe in the selected case,
it is clear that some codes are more superior to others. For instance, the assessment methods of
ASME B31.G propose no alternative methods for flaw type or state. On the contrary, the British
standard [4] differentiates between single corrosion and interaction/composite corrosion assessments,
but it is lacking to provide differing methods for various corrosion types. On the other hand, API [2]
proposes more comprehensive assessment procedures for all three corrosion types with differing
levels in assessment, making it the superior fitness for service code among all relevant standards. The
three dimensional FE analyses conducted herein (as a part of API suggestions) ascertained that the
guidelines by API can reliably predict the corrosion induced pipe failure. It is to be noted that the
current study is only limited on the failure assessments when considering common corrosion types,
but not considering fracture resistance methods in detail. More rigours assessments methods are
required (such as based on fracture toughness) to assess the effect of sharp flaws as recommended
in [2] & [4].

References

1. ABAQUS, Inc. (2011), ABAQUS V.6.7 User’s Manual, Providence, Rhode Island.
2. American Petroleum Institute/American Society of Mechanical Engineers,Fitness-For-Service-
API579/ASME FFS-1, (2007) “Recommended Practice for Fitness for Service”, API, Washington.
3. American Society of Mechanical Engineers, ASME B31G,1991 “Manual For Determining the
Remaining Strength in Corroded Pipelines- a supplement to ASME B31 code”, New York.
4. British Standard, BS7910,2005, “Guide for methods assessing the acceptability of flaws in
metallic structures”, British Standards Institution, UK.
5. Brunone, B and Morelli, L (1999), “Automatic control valve-induced transients in operative pipe
system”, ASCE-Journal of Hydraulic Engineering, Vol. 125, No. 5,pp.534-542.
6. Caproco Corrosion Prevention Ltd. (1985), ‘‘Underground corrosion of water pipes in Canadian
cities. Case: The city of Calgary.’’ Rep. Prepared for CANMET, Ottawa.
7. Ma, Z., and Yamada, K. (1994), ‘‘Durability evaluation of cast iron water supply pipes by sampling
tests.’’ Proc., Structural Engineering, Japan Society of Civil Engineers, Tokyo, 40A.
8. Rajeev. P, Kodikara. J., Robert. D.J, Zeman. P. and Rajani. B. (2014), “Factors Contributing to
Large Diameter Water Pipe Failure as Evident from Failure Inspection”, IWA Publishing’s
newsletter, Water Asset Management International Journal” (accepted).
9. Rajani, B. and Kleiner, Y. (2001). Comprehensive review of structural deterioration of water main:
physically based modes, Urban Water, Vol. 3, 151-164.
10. Ruus, E and El-Fitiany F A (1980), “Water hammer in pipelines resulting from valve closure”,
Canadian Journal of Civil Engineering, Vol. 7, pp-243-255.
11. Rajani, B and Tesfamariam, S. 2007, Estimating time to failure of cast-iron water mains,
Proceedings of the Institution of Civil Engineers, Water Management 160, Issue WM2, pp- 83–88.

538
12. Rajani, B., 2012, Nonlinear stress–strain characterization of cast iron used to manufacture Pipes
for water supply, Journal of Engineering Materials and Technology, Vol, 134 (4), pp-1-8.

539

View publication stats

You might also like