Sabp A 004
Sabp A 004
Electricity
Energy Calc Prices:
Steam Feed
Data engine
Fuel Product
Energy
Process Fluid
Flows Data
Densities
Recon Balanced
Model Electricity
Steam
Heat & Fuel
Mass Refinery Total Energy
Balance Model Indices
Table of Contents
Page
1. Introduction 3
1.1 Purpose and Scope 3
1.2 Definitions 3
1.3 Conflicts with Other Standards 4
1.4 References and Related Documents 4
2 General 5
2.1 Classification 6
2.2 Data Availability and Quality 7
2.3 Calculation Strategy 9
2.4 Simulation models 10
2.5 Trend Chart Formatting 11
3 Equipment EPIs 14
3.1 Pumps 14
3.2 Compressors 15
3.3 Fired Heaters 20
3.4 Fired Boilers 25
3.5 Unfired Boilers 26
3.6 Steam Turbines 31
3.7 Gas Turbines 36
3.8 Cogeneration Systems 50
5 Product EPIs 65
5.1 Methodology 65
5.2 Examples – Oil Refineries 68
5.3 Example – Juaymah Gas Plant 69
6 On-line Implementation 74
@Seismicisolation
@Seismicisolation Page 2 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
1.0 INTRODUCTION
Energy is a significant operating cost parameter for Saudi Aramco, and one that is important to
management. The Company’s energy conservation program is overseen by the Energy
Management Steering Committee (EMSC), which determined that the potential existed to
double the company’s energy efficiency (i.e., to halve its energy consumption index) in the
major industrial manufacturing plants. One of the key elements in the EMSC’s strategy was to
develop and monitor effective Energy Performance Indices (EPIs), and to report the results
regularly.
After conducting an exhaustive survey of industry practice, it was determined that while the
indices in common industrial use were indeed of some value for competitive benchmarking, they
are lacking in other important capabilities, chiefly:
• Monitoring trends in energy efficiency by product, by process unit, and by major
equipment.
• Serving as a diagnostic tool for process troubleshooting and operational efficiency
improvement.
Ideally, the EPIs should also screen out the effect of variations in uncontrollable external factors
such as feed rate, feed composition, product mix, and ambient conditions.
This manual is an explanation of the methodology for calculating the new EPIs that were
developed internally within Saudi Aramco to meet the EMSC’s specifications.
The purpose of this DeskTop Standard is to describe proven Best Practices, consistent with
guidelines endorsed by the Company’s Energy Management Steering Committee (EMSC), in
developing and deploying an integrated system of EPIs. This Manual provides the theoretical
basis for developing EPIs, the methodology for calculating them, and examples of final output.
It is intended as an aid to engineers working in Saudi Aramco plants who are responsible for
maintaining and reporting the departmental EPIs.
It is important to also understand what this manual is not. It is not intended to provide a
comprehensive review of all current or past practices – only to describe the best ones. It
focuses on practices that have been proven to be useful in the field, not on the theoretical ideal
way of doing things. The Best Practices therefore incorporate necessary compromises that
invariably must be made to adjust to the reality of existing instrumentation, raw data quality,
available software, limited manpower resources, etc. Finally, despite the high level of detail in
some sections, the manual is not intended to be a cookbook giving step by step procedures for
every possible EPI for every facility. The objective is to outline the general approach and
procedure for the major classes of EPIs, and to illustrate how these EPIs were implemented in
their “best” manifestation.
@Seismicisolation
@Seismicisolation Page 3 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
1.2 Definitions
Best Practice: A process or method that, when correctly executed, leads to enhanced system
performance.
Energy Performance Index: An operating policy that distributes the load between parallel
networks of multiple machines/equipment in a way that minimizes their energy (fuel + power)
consumption, without compromising safety or reliability
There are no other published Standards, Procedures, or General Instructions that address the
subject of Energy Performance Indices, and therefore no conflicts are expected.
This manual is based on an unpublished course that was developed and delivered in June 2005
by the Energy Systems Unit of Consulting Services Department to energy engineers from all
proponent facilities. The software referenced in this Manual was originally developed by Energy
Systems Unit, and is available to all Saudi Aramco engineers upon request.
@Seismicisolation
@Seismicisolation Page 4 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
2.0 GENERAL
The concept of EPIs for monitoring process and equipment efficiency is not new. In fact, many
different types of EPIs are in use throughout the oil and gas industry. EPIs have two primary
applications:
• Benchmarking
o Historical
o Competitive
o Absolute
• Process Improvement (viz. energy efficiency)
Historical benchmarking is when the current energy efficiency is compared to previous energy
efficiency, usually at some “baseline” period in time.
Competitive benchmarking is when the plant performance is compared against the performance
of other similar plants, usually on an industry-wide basis, although it can also be applied to
different plants making the same products within a single company.
Absolute benchmarking is when the equipment or plant energy efficiency is compared against a
theoretical target or budget.
EPIs can be used to measure and monitor the efficiency of process equipment, process units, or
entire process plants. The proper formulation depends upon the objective.
In all cases it is imperative that the EPI should accurately reflect the efficiency of energy
consumption. For equipment, the EPI is generally formulated as the thermodynamic efficiency.
The higher the thermodynamic efficiency, the better. For process plants, there are several
formulations, but in general they are all expressed as an “energy intensity”, i.e., energy
consumption per unit of throughput. In general, a high numerical value for the process EPI
indicates high energy intensity, whereas a small EPI indicates low energy consumption per unit
of output. In short, the lower the EPI, the better.
For oil refining, the most widely accepted process EPI is the Solomon Energy Intensity Index
(EII), which was developed in the mid 1970s as a tool for comparing the energy efficiencies of
different oil refineries. This type of index works well when the processes being used are
comparable in all plants. Its primary use is in competitive benchmarking.
In the gas processing industry, the most widely accepted EPI is CGEY’s* Energy Efficiency
Index, which is expressed as energy consumption per unit of feed per GPU (Gas Processing
Unit), or $/MMscfd per GPU. The GPU is a parameter developed by CGEY to represent the
complexity of a plant. Each process unit (e.g., amine treating, dehydration, NGL recovery, etc.)
is characterized by a GPU value, which is determined based on its theoretical energy
consumption per unit of feed flow. The GPU values of all the units in a plant are summed up to
get the overall GPU. The capacity parameter is combined raw feed gas (MMscfd). CGEY’s EEI
was modeled after the Solomon EII, and is also designed primarily for use as a competitive
benchmarking tool.
*
Cap Gemini Ernst and Young, a management consulting firm based in Canada
@Seismicisolation
@Seismicisolation Page 5 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Plants operating under different feedstock conditions, product specs, and process technologies
are impossible to compare unless there is a mechanism for reducing them to a common basis.
The artifice of creating a GPU is an attempt to partially compensate for some of the differences
in feedstock compositions and pressures, product slates, and product specification. CGEY
freely concedes that the GPU technique is approximate at best and cannot provide truly
accurate benchmarks, but they claim it is the best available (which at the moment of writing was
true).
There is no comparable industry-wide energy efficiency index applicable to oil & gas exploration
and production.
An even more important application of EPIs than benchmarking is to serve as a diagnostic tool
for process troubleshooting and identifying opportunities for process efficiency improvement.
Neither the Solomon nor the CGEY indices have this capability. To fill the need for timely
actionable information, CSD’s Energy Systems Unit has developed a different approach to
Energy Performance Indices (EPIs) with the following desired features:
• Energy consumption and cost are expressed per unit of product rather than feed
• Different indices for different user groups – managers, engineers, and operators
• Capable of meeting all EMSC objectives – report card, accurate benchmarking,
diagnostic tool for process trouble-shooting and identifying efficiency improvement
opportunities, and an operator alarm generation system to flag significant deviations
from target performance.
In addition the output from these EPIs can be used for accurate product pricing as well as
calculating a global energy KPI for the entire company, which is of great value and interest to
both the Finance Department and to senior management.
The main structural difference between EMSC’s overall product indices and existing industry
indices is that they are formulated as output-based instead of input-based. The problem with
feed-based indices is that even a reduction in energy consumption associated caused by
declining process yield (something undesirable) will show up as an improved energy index,
which is misleading. Product-based indices, on the other hand, would clearly show that the
index has deteriorated, and reliably indicate both favorable and adverse trends.
2.1 Classification
The first thing to understand is that there is not just a single EPI but sets of EPIs, each
formulated to meet a specific application. In fact, one can have as many EPIs as one would
like, as long as each one provides some useful actionable information. The key here is
“actionable”, which means the engineer or manager should be able to take some logical action
to improve the index, based on its numerical value.
@Seismicisolation
@Seismicisolation Page 6 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Four categories of EPIs are described in this manual, with different objectives and functions as
follows:
Intended Users
Category Management Engineers Operators
Equipment EPIs x x
Process Area EPIs x x
Individual product EPIs x x
Corporate Energy KPI x
The new EPIs are expected to be developed and deployed by each Department in 3 phases:
Phase 1: Initial rollout, including on-line implementation, based on simplified models.
Phase 2: Improved data conditioning and model refinement.
Phase 3: Further improvements, e.g., comparison against dynamic targets, segregation of
indices by shift, etc.
This manual covers implementation of Phase 1 only. Additional manuals will be prepared for
Phases 2 and 3 after they are implemented (scheduled for 2007 and beyond).
A large number of flow, pressure, temperature and economic data are needed as input to the
EPI models. The primary data sources are expected to be as follows:
• Readings from the PI system
• Manual readings and logs
• Official raw material, product and energy prices as specified by Business Analysis
Department of Corporate Planning
One of the problems with using live PI data directly in the EPI models is that even if one of the
data values is “bad” (i.e., a non-numerical value), the model fails to complete the calculation and
give a numerical result. When we have hundreds of such PI inputs, the probability of at least
one input being “bad” is quite high. Bad data such as a textual error message or a numerical
error (e.g., division by zero) are easy to recognize. Therefore, it is necessary to screen live data
and substitute default values when bad readings are encountered.
The second and more difficult problem we must deal with is to recognize when the numerical
data that we do get are not correct, and to make necessary corrections. Two simple error
detection techniques are recommended to start with. One is to set upper and lower bounds for
each parameter. These limits could either be fixed values, or set as a function of some other
parameter(s). If the measured data value is outside these limits, the default value is used, and
an error message is transmitted to the operator and logged. The other technique is to do a
consistency check on a set of inter-related parameters using material and heat balances and/or
physical property correlations. For example, unless the pressure and temperature
measurements of saturated steam are consistent with the Vapor-Liquid Equilibrium (VLE)
equations, we would know that one or the other is incorrect. When such inter-related data
cannot be reconciled, and fail the consistency checks, an error message should be
automatically transmitted by the system to the operator and the engineer’s logs. All of this logic
@Seismicisolation
@Seismicisolation Page 7 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
(see Exhibit 2-1) must be embedded within the EPI models, reviewed periodically, and revised
as needed.
There are several options for selecting default values, in order of increasing accuracy and
computational effort:
a) Use a fixed number, based on manual input
b) Use the last good numerical value
c) Use an average over some reasonable period (could be hours, days, or weeks) prior to
encountering the bad value
d) Use a computed value based on statistical regressions
e) Use a computed value derived inferentially (using simulation models) from other measured
data that are known to be good.
The appropriate default value must be selected for each parameter on a case-by-case basis.
For input data that do not change much over time, e.g., HHV of the fuel, a fixed number is
probably the right choice. For parameters that typical vary from period to period, such as flow,
temperature, and pressure, the last good value may be the right choice for the following 2-4
periods, after which we might switch to the use of an average value. In the case of power or
fuel consumption readings, the correct choice would be a computed value based on statistical
correlations of historical data, since the energy consumption can be expected to be a function of
throughput as well as ambient temperature. And so on.
@Seismicisolation
@Seismicisolation Page 8 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The general form for expressing energy EPIs (or any KPI for that matter) is
Resource Consumption
EPI =
Throughput
Ideally, the throughput should be expressed in terms of the desired output from that process,
which is the product produced. However, when one has multiple products from a single feed, as
in an oil refinery or gas plant, it becomes difficult to decide how much of the resource consumed
should be allocated to each product. That is why standard industry practice has been to
express the indices on the basis of feedstock processed.
Unfortunately, the consequent indices are not particularly helpful, and can even be misleading,
as we shall demonstrate, using a simple illustrative example. Consider the simplified process
model illustrated in Exhibit 2-3.
Waste, W
Feed, F Product, P
PROCESS
Energy, E
E input1 E input
FB = = PB = =
F input2 P output
Note: FB = Feed-based, PB = Product-based
Suppose the yield of the process has been improved through continuous process innovation
since it was started up. Suppose that a 10% increase in yield requires a 5% increase in energy
consumption. In this case the feed-based index would go up by 5%, while the product based
index would go down by 4.5% (= 100-105/110). The feed-based EPI would indicate to
management that the plant has become less efficient, when in fact it has become more efficient.
A similar argument can be made to demonstrate that a feed-based index will not change if part
of the product is used as fuel (which in fact is common practice in Aramco’s gas plants as well
as refineries), and might even improve, if we account only for purchased fuels and power. A
product-based index will highlight the revenue loss, because using product as fuel is equivalent
@Seismicisolation
@Seismicisolation Page 9 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
to a loss in yield. A product-based EPI gives accurate indications of efficiency trends, while a
feed-based EPI does not. More generally, they could be thought of as input-based versus
output-based. As noted earlier, the CGEY indices are all input-based. The EMSC has
concluded that we need something better.
The type of model required depends upon the type of index being calculated. For the product
indices we need a fairly sophisticated approach – combining a relatively rigorous process
simulation model (including heat and material balances) with a relatively sophisticated economic
resource-allocation model. For the process unit (area) indices we can use a simpler approach
for calculating EPIs (no economics involved), but which uses the output from the same rigorous
process simulation model. For the equipment indices we do not use these models at all.
Instead, we develop equipment-specific models for directly calculating energy efficiency.
The general data collection and computational strategy is depicted in Exhibit 2-4.
Archive of “Good”
data set (input
data for Model)
Run calc engine to
generate EPIs for
current data set
Archived database
of historical EPI
values
Generate trend
charts and
Reports
All EPIs require some sort of calculations based on measured process parameters. Equipment
EPIs require only very simple models, most conveniently done using spreadsheets. Process
and product EPI calculations, though somewhat more complex in concept, are also fairly simple
to model using spreadsheets. However, to give meaningful results, process and product EPIs
require relatively good (i.e., consistent and accurate) mass and energy balance data. Direct
measured data from the PI system is seldom of sufficiently good quality for our purposes, and
must be validated using a “data reconciliation” software package that goes well beyond the rudi-
mentary data conditioning techniques described in section 2.2. Such data reconciliation soft-
@Seismicisolation
@Seismicisolation Page 10 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
ware is not currently installed at any Saudi Aramco facility. Therefore it is recommended that a
stop-gap approach be used on an interim basis.
The proposed solution is to employ commercial software packages such as AspenPlus, Pro-II,
or Hysys to develop simplified heat and mass balance simulation models for each process unit
and also for the plant as a whole. These simplified models would not do detailed simulations of
reactions and VLE separations (such as distillation) – just the heat and mass balances.
Reactors and separations would be modeled as mixers and splitters – with the split ratios for the
cut streams specified as manual input. The models would have to be “tuned” using accurate
feed and product flow data (which are usually metered using very high-accuracy “custody-
transfer” meters, and certain key temperature/pressure/composition measurements at “anchor”
points within the process. The simulation models essentially serve as a “virtual metering”
system for intermediate flows between process units, and thereby minimize the reliance on
measured data of dubious quality. While the virtual metering approach cannot guarantee
accuracy, it does guarantee consistency, which is perhaps more important.
Output from these simulation models is then used as input for the EPI models.
EPI trend charts must be formatted to show time scale on the X axis and the index or indices on
the Y axis.
The time scale should be selected to cover the period of interest. For product and process EPIs
being used within the department, a monthly or weekly EPI calculation is probably appropriate,
and the recommended period is 12 months preceding the current date. If, on the other hand,
the EPI trend chart is being used for a presentation to management on long-term trends, the
time scale should be on an annual basis. For equipment EPIs, the index should be calculated
on an hourly, shift, or daily basis, depending on the equipment and process variability for that
particular application. The recommended display period could be the previous 3-10 days (for
EPIs calculated frequently) or the previous 30 days (for EPIs calculated daily).
The Y-axis scale should be selected to provide a balance between perspective and discrimina-
tion. Too narrow a range magnifies variations at the cost of perspective. A wide range restores
perspective, but may lose discrimination. A good rule of thumb is that the Y-axis scale should
be approximately 2-4 times the difference between the highest and lowest values for the
parameter being displayed.
@Seismicisolation
@Seismicisolation Page 11 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
1.0 2500
0.8 2000
MMBtu/MB
0.6 1500
KWH/MB
0.4 1000
Power Index
0.0 0
Jan- Feb- Mar- Apr- May- Jun- Jul-03 Aug- Sep- Oct- Nov- Dec-
03 03 03 03 03 03 03 03 03 03 03
70
60
50
Energy Cost, $/MB
40
30
20
10
0
Jan- Feb- Mar- Apr- May- Jun- Jul-03 Aug- Sep- Oct- Nov- Dec-
03 03 03 03 03 03 03 03 03 03 03
@Seismicisolation
@Seismicisolation Page 12 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
30 0.20
25 0.18
MMBtu/bbl crude oil feed
20 0.16
10 0.12
0 0.08
1999 2000 2001 2002 2003 2004 2005
100%
95% K-003A
90% K-003B
85%
80%
75%
70%
65%
60%
3/29/05 4/3/05 4/8/05 4/13/05 4/18/05 4/23/05 4/28/05 5/3/05
@Seismicisolation
@Seismicisolation Page 13 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
EPIs are generally recommended only for major energy consumers, i.e., power > 2 MW, or fuel
> 15 MMBtu/h. The definitions of efficiency for these two categories are correspondingly
different. For consumers, energy efficiency = useful work accomplished on the process divided
by the energy that must be supplied. For converters, energy efficiency = useful energy output
divided by energy input.
Exhibit 3-1: Definitions of Equipment Energy Efficiency
Equipment Type Input (Denominator) Output (Numerator)
Pump Mechanical shaftwork Liquid kinetic energy
Compressor Mechanical shaftwork Vapor/gas pressure energy
Fired process heater Chemical (fuel) energy Process enthalpy
Distillation column Thermal energy Chemical separation of mixture
Electric motor Electrical power Mechanical shaftwork
Generator Mechanical shaftwork Electrical power
Steam Turbine Steam kinetic energy Mechanical shaftwork
Boiler Chemical (fuel) energy Thermal energy (as steam)
Nowhere does a capacity term appear in the calculation of energy efficiency. One common
mistake worth mentioning is that some people refer to condensing steam turbines as being
more “efficient” than back-pressure steam turbines because they generally produce more
shaftwork (kwh) per lb of steam; this is inconsistent with fundamental thermodynamic principles.
An index expressed as “kwh/lb of steam” could perhaps be considered an indicator of
generating capacity, but certainly not of energy efficiency.
3.1 Pumps
To calculate efficiency on-line it is necessary to have four measured data points – the flow rate,
in gpm, the suction and discharge pressures in psi, and the horsepower (or kilowatt) reading.
Normally, we have instrumentation for the first three but not the last. Without all four readings, it
is not possible to calculate the efficiency. Sometimes, a pump may have an ammeter instead of
a kw meter. The power can then be calculated as:
I .V . 3. cos(φ ).η M
HP =
746
@Seismicisolation
@Seismicisolation Page 14 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Normally the mechanical efficiency of a pump does not fluctuate in the short term. Rather, it
degrades slowly over a period of months or even years. Therefore on-line efficiency
measurement is not especially valuable. For both of these reasons, pump efficiency monitoring
is not a high priority in most cases. It could be included for selected applications in the future if
desired, after the power meters have been installed.
An alternative diagnostic tool on the pump’s mechanical condition and performance is to plot the
actual pump characteristic curve (flow rate versus ΔP) and compare it against the design curve
as in Exhibit 3-2. If the pump is delivering less TDH (total discharge head) than design at the
measured flow, it means that the performance has fallen off and needs corrective action. The
efficiency loss can be estimated very roughly as the ratio of actual head to design head at that
particular flow rate.
Centrifugal Pump
1000
800
600
Head, ft
400
200 Design
Actual
0
0 2000 4000 6000 8000 10000
Flow, gpm
3.2 Compressors
Large gas compressors are good candidates for efficiency monitoring. Overall Energy
Efficiency of an individual compressor is defined as:
absorbed energy into process gas isentropic HP
ηo = =
delivered energy to the driver brake HP
@Seismicisolation
@Seismicisolation Page 15 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
For compressors, there are two types of efficiencies: the adiabatic efficiency and polytropic
efficiency. As a practical matter, the adiabatic efficiency can be calculated from process data as
T2' − T1
ηa =
T2 − T1
where T1 = Suction temperature, °F
T2 = Actual discharge temperature before any cooling, °F
T2’ = isentropic (adiabatic) discharge temperature, °F, calculated as:
⎡ k −1
⎤
⎛P ⎞
T2 = (T1 + 460 ) ⎢⎜⎜ 2 − 1⎥ + T1
k
'
⎟⎟
⎢⎝ P1 ⎠ ⎥
⎢⎣ ⎥⎦
P1 & P2 = suction and discharge pressures, psia
k = specific heat ratio Cp/Cv
The polytropic constant can be calculated from on-line process measurements by the equation:
ln( P2 / P1 )
n =.
{ln( P2 / P1 ) − ln(T2 / T1 )}
Adiabatic efficiency varies with compressor inlet conditions, whereas the polytropic efficiency is
constant for a particular gas mixture, being a function primarily of mechanical design. The
polytropic efficiency is therefore a better indicator of compressor mechanical condition and
performance.
Spreadsheet templates for calculating compressor efficiency are shown in Exhibits 3-3 and 3-4.
It is recommended that compressor efficiencies should be monitored on-line (at 4 hour intervals)
to detect any adverse trends in a timely manner and to take appropriate corrective action.
@Seismicisolation
@Seismicisolation Page 16 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
To see a sample trend chart, refer back to Exhibit 2-7. The approximate polytropic efficiency for
large compressors (suction flow >5000 acfm) in good condition is Eff = 61 + 1.31 ln (acfm).
@Seismicisolation
@Seismicisolation Page 17 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Exhibit 3-4a: Efficiency Calculation Template for Two-Stage Compressors (LP section)
@Seismicisolation
@Seismicisolation Page 18 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Exhibit 3-4b: Efficiency Calculation Template for Two-Stage Compressors (HP section)
For gas mixtures, the compressibility and specific heat data are most conveniently determined
off-line at average inlet/outlet conditions (using Hysis, VMGsim or other physical properties
simulation software), and entered manually. For pure component gases (e.g., refrigeration), a
higher level of modeling flexibility and accuracy can be obtained without too much effort by
using correlations (developed off-line for the expected range of operating conditions) for the
compressibility and specific heat, calculating the values separately at inlet and outlet conditions,
and then taking their arithmetic average.
@Seismicisolation
@Seismicisolation Page 19 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Observe that several checks are built into the EPI models – to check consistency of measured
vs calculated mixture temperatures (for 2-stage compressors), and measured vs calculated
power. The model also checks to ensure that the measured flow rate does not exceed com-
pressor capacity and that the calculated horse power does not exceed motor size.
Three-stage compressors are not very common, but can be handled in the same manner as two
stage compressors: i.e., efficiency must be calculated for each section.
Under field conditions, mechanically identical machines may develop differences in polytropic
efficiencies of as much as 5%. As a general rule, if they are within 3% of one another, and the
trends tend to track in parallel, it is an indication that the machines are in comparable
mechanical condition, and that the instruments are reading correctly. Wide discrepancies in
efficiency or non-parallel trends could indicate that one machine has deteriorated significantly
compared to others, the instrument readings are not accurate, or that something abnormal is
going on process wise.
Fired heaters or furnaces are found in oil refineries, gas plants (usually in the Sulfur Recovery
Units), and GOSPs.
Losses
Fuel, F
For fired heaters that are supplying sensible heat to the process, the fuel efficiency is most
simply calculated by the heat balance method, also known as the “direct” method:
W .Cp.(T2 − T1 )
η=
F .HHV
where W = mass flow rate of process fluid, lb/h
Cp = average specific heat, Btu/lb°F (varies with composition and temperature)
T1, T2 = inlet and outlet process temperatures, °F
F = fuel gas flow rate, scfm
HHV = higher heating value, about 1050 Btu/scf on average (spec is 1080). If the
lower heating value (LHV) used instead of HHV, η is called thermal efficiency
instead of fuel efficiency.
If the process stream being heated is undergoing both sensible heating and evaporation, the
equation becomes:
W .( H 2 − h1 )
η=
F .HHV
where h1, H2 = inlet and outlet enthalpies of the process stream, Btu/lb.
@Seismicisolation
@Seismicisolation Page 20 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
While simple is theory, the heat balance method has a major deficiency – it gives accurate
results only when the data quality is extremely good (generally less than 1% error), which is
hardly ever the case. A more accurate estimate of furnace efficiency can be obtained by the
heat loss method (also known as the “indirect“ method) as follows:
Choose the formulation according which measured value (absorbed duty of fuel input) is likely to
be more accurate.
A spreadsheet template for calculating furnace/heater efficiency is shown in Exhibit 3-5, with
provision for calculating the efficiency for both gaseous and liquid fuels. It also has the option to
calculate either single-phase or two-phase process heating duty.
@Seismicisolation
@Seismicisolation Page 21 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Observe the messages saying “OK”. These represent the results of built-in consistency checks.
If the consistency check failed, there would be an warning error message instead. It may
appear that running a full set of calculations including consistency checks for each set of data is
too much work. However, it is necessary in order to ensure that the results are as accurate and
reliable as possible.
Furnace efficiency monitoring using the heat balance method, without any consistency checks,
is illustrated in Exhibits 3-6 and 3-7. It is an acceptable first step, but ultimately the efficiency
calculation method should be upgraded to the more rigorous approach.
@Seismicisolation
@Seismicisolation Page 22 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
PI DATA
80 OUT 400
Thermal Eff, %
Stack Temp, F
60 300
40 200
20 Efficiency 100
Stack Temp
0 0
Jan-03
Feb-03
Mar-03
May-03
Jun-03
Jul-03
Oct-03
Nov-03
Dec-03
Sep-03
Apr-03
Aug-03
@Seismicisolation
@Seismicisolation Page 23 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The spreadsheet templates presented in this and the following section on boilers are intended to
serve as guidelines only. There could be many variations from the illustrated case, depending
on the specific situation. The engineer must apply his knowledge and understanding of the
basic principles to adapt the “base-case” EPI calculation engine. The information provided in
Exhibits 3-8 to 3-10 may be useful in this effort.
20000
y = -0.4303x 2 + 68.919x + 17671
Heating Value, Btu/lb
19000
18000
y = -0.3774x 2 + 57.906x + 16786
LHV
17000 HHV
P o ly. (HHV)
P o ly. (LHV)
16000
0 10 20 30 40 50 60 70
deg API gravity
@Seismicisolation
@Seismicisolation Page 24 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
In general, the heat loss method is preferred over heat balance method because it is less
sensitive to errors in measurement. For example, let’s say that the true fuel firing rate is 100
MMBtu/h and the true efficiency is 85%. i.e., Absorbed duty = 85 MMBtu/h and Losses = 15
MMBtu/h. Now suppose the fuel meter reading is low by 10%, ie it measures 90 MMBtu/h
instead of 100 MMBtu/h. By the heat balance method, the calculated efficiency would be 85/90,
or 94.4% (error of 9.4%), whereas by the heat loss method, it would be (90-15)/90 = 83.3%, an
error of only 1.7%. Clearly, the latter is more robust.
A fired boiler converts chemical energy in the fuel to thermal energy in steam. Saudi Aramco
plants appear to have standardized on steam generation pressures of 600 psig, 375 psig, and
150 psig. However, there is no reason boilers could not be designed for other pressures. In
fact, there are significant advantages in terms of power generation potential (employing back-
pressure steam turbines) to higher steam pressures, e.g., 900-1200 psig. Usually, the steam is
superheated by at least 200°F in order to avoid condensation inside the turbine casing.
Flue gas
Water BOILER
Steam
In, W Out, S
As for fired heaters, boiler efficiency can be calculated by either the heat balance method or the
heat loss method. The heat balance method is straightforward:
S .( H 2 − h1 )
η=
F .HHV
where the parameters are identical to those described earlier. ASME recommends the heat
loss method, which though more complicated method, is far more accurate:
F .HHV − ∑(losses) S .ΔH
η= =
F .HHV S .ΔH + ∑(losses)
The input data required are usually measured on-line in any case:
@Seismicisolation
@Seismicisolation Page 25 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
In addition, it is necessary to provide the following details about boiler geometry and design as
one-time manual inputs: surface area exposed to the environment (viz. wind), total area of
openings through which heat can be radiated out (e.g., peepholes), wall thickness, insulation
thickness, and type of wall material – whether metal or brick.
A spreadsheet template for calculating boiler efficiency is shown in Exhibit 3-11, and a sample
trend chart is shown in Exhibit 3-12. An iterative procedure is required. Assuming the gas flow
is more accurate; the steam flow is calculated by heat balance, and compared against the
measured value. Although the spreadsheet template presented is for gas-fired boilers, which
are the most common type used in Saudi Aramco facilities, it can be easily adapted to other fuel
types, in a manner similar to that shown for oil-fired furnaces in Exhibit 3-5.
Waste heat boilers (WHBs) are similar in function to conventional boilers, except that the heat
source is a hot process stream (e.g., reactor outlet or incinerator flue gas), and no fuel-firing is
employed. Because they generally operate at much lower temperatures than fired boilers, the
construction is quite different – no refractory linings are required. In fact, they are very similar in
design to shell-and-tube kettle reboilers, with steam generation either on the shell side or the
tube side, depending upon process considerations.
Steam, H2
Water, h1
For waste heat boilers (WHBs) where the objective is recovery of steam for process use, the
thermal efficiency is given by:
S .( H 2 − h1 )
η=
W .Cp.(T2 − T1 )
where W = mass flow rate of tail gas, lb/h
T1, T2 = inlet and outlet process temperatures, °F
S = mass flow rate of steam generated, lb/h
h1, H2 = enthalpies of boiler feedwater makeup and steam respectively, Btu/lb.
For WHBs, we can only use the heat balance method, because the heat loss method does not
apply. As the flow data are usually extremely unreliable, one has to check for consistency using
material balances on the water side, i.e., to ensure steam plus blowdown flow rates add up to
the boiler feedwater supply.
@Seismicisolation
@Seismicisolation Page 26 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
@Seismicisolation
@Seismicisolation Page 27 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
@Seismicisolation
@Seismicisolation Page 28 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
94% Boiler-1
Boiler-2
93%
Boiler-3
92% Boiler-4
91%
90%
89%
88%
87%
86%
85%
3/29/05 4/3/05 4/8/05 4/13/05 4/18/05 4/23/05 4/28/05 5/3/05
@Seismicisolation
@Seismicisolation Page 29 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The calculation template and sample trend chart for WHBs are presented in Exhibits 3-13 and
3-14. The consistency checking method is illustrated in Exhibit 3-15. It should be obvious there
is a metering problem from June 2004 to January 2005, as the generated steam cannot possibly
exceed boiler feedwater flow.
Exhibit 3-14: Sample Trend Chart for Waste Heat Boiler Efficiency
100
80
Thermal Eff, %
60
40
20
0
O 4
De 4
De 4
04
Ju 4
A 4
Se 4
5
Ja 4
Fe 4
04
No 4
Fe 5
M 4
A 4
M 04
M 5
A 5
M 5
0
0
l-0
-0
-0
-0
0
0
-0
0
0
-0
0
-0
-0
p-
v-
c-
c-
n-
n-
n-
n-
b-
b-
ug
ct
ay
ay
ar
pr
ar
pr
Ju
Ja
Ja
140%
120%
Stm / BFW mass frac
100%
80%
60%
40%
20%
0%
O 4
De 4
De 4
04
A 04
Ju 4
Se 4
5
Ja 4
Fe 4
04
No 4
Fe 5
M 4
A 4
M 04
05
A 5
M 5
0
0
0
-0
-0
-0
0
0
-0
0
0
-0
-0
-0
p-
v-
c-
c-
l-
n-
n-
n-
n-
b-
b-
ug
ct
ay
ay
ar
pr
ar
pr
Ju
Ja
Ja
@Seismicisolation
@Seismicisolation Page 30 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
In case the steam is generated at saturated conditions (quite common in the case of WHBs) for
use directly in a process application, there is a risk of steam hammer problems in the pipework
due to condensation. To avoid this problem the steam should be blended back in to the main
header or superheated, even if by only 20°F, e.g., against flue gas from any nearby furnace.
Examples of these two potential solutions are shown in Exhibits 3-16 and 3-17 (next page). In
the worst case, some HP steam could be injected.
Plant engineers should always be asking themselves if the results they see make sense, e.g.,
by comparing them against known industry standards. Consider, for example, the trend chart of
Exhibit 3-12, which shows boiler efficiency values in excess of 90%. Examination of typical
efficiency values for well operated boilers, as listed in Exhibit 3-18, should lead us to conclude
that the input data are probably wrong, not that the boilers are working exceptionally well.
Waste heat boilers, on the other hand, should have thermal efficiencies in the range of 95-99%,
as the only losses are due to radiation and convection.
The most common steam turbine design used in industrial applications is the single-exhaust
back-pressure type, because it is the most efficient thermodynamically.
@Seismicisolation
@Seismicisolation Page 31 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
@Seismicisolation
@Seismicisolation Page 32 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
HP Steam
WORK
BPST
LP Steam
PROCESS
Two kinds of efficiency are important. One is the isentropic efficiency of the turbine, which is an
indicator of how well the machine was designed to begin with, and its present mechanical
condition. This “machine” efficiency is calculated as:
H1 − H 2
η=
H1 − H 2 '
where H1 = Enthalpy of HP inlet steam, Btu/lb
H2 = Actual enthalpy of exhaust LP steam, Btu/lb
H2’ = Enthalpy of exhaust LP steam assuming isentropic expansion, Btu/lb
The other kind of efficiency is the overall or “cycle” energy efficiency, which is calculated as:
3413 x kw + W.H 2
η=
W.H 1
Since enthalpy cannot be measured directly, it must be inferred from pressure and temperature
measurements using a steam properties database. This is called Method 1. The recommended
properties software is called SteamTab© version 3, which is an Excel add-in software package,
available for about $150 per copy from ChemicalLogic Corp (Woburn, Mass). It is also possible
to manually read these properties off a Mollier Diagram, but this is not practical to do when
calculating and trending results using a computer.
Some plants have “extraction” steam turbines with two exhausts – at medium and low pressure.
The thermodynamic efficiency is calculated for each stage, and the overall efficiency is derived
from the stage-efficiency results. A spreadsheet template for calculating isentropic steam
turbine efficiency according to Method 1 is presented in Exhibit 3-19a.
The template shown can be used for either single- or dual-exhaust turbines. Observe that there
are two parallel calculation columns, entitled “Approx” and “Exact”. The Exact calculation
requires the SteamTab© software add-in to be installed on the computer being used. If it is not,
then the Approx calculation results can be used.
@Seismicisolation
@Seismicisolation Page 33 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
@Seismicisolation
@Seismicisolation Page 34 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
It may appear that there is not much difference in results between the two sets of data for the
steam pressures selected, but if the pressures are significantly different, the Approx correlations
may no longer be in such close agreement. The overall system (“cycle”) energy efficiency for
this same turbine is shown in Exhibit 3-19b.
Notice that the overall cycle efficiency is extremely high – approaching 100%. This is typical of
back-pressure steam turbines. Condensing steam turbines, by contrast, have overall cycle
efficiencies in the range of 20-30%, which is why they should never be considered except in
certain extreme conditions, e.g., for extremely remote locations where electrical power is either
not available at all or is very expensive, or if the fuel is a waste material that requires disposal.
The former condition exists only at Shaybah in Saudi Aramco facilities. The latter is unlikely to
be encountered anywhere in the Kingdom within the foreseeable future.
The isentropic efficiency is of more practical use from an operational viewpoint, because it
provides a warning of developing mechanical problems. The cycle energy efficiency is more
useful for design and decision-making purposes, such for calculating the process unit energy
balance when conducting plant energy audits or choosing between project alternatives.
Exhibit 3-20: Sample Trend Chart for Back-pressure Steam Turbine Efficiency
80
Isentropic Efficiency, %
KT-361
78
KT-461
76
74
72
4/1/05 4/8/05 4/15/05 4/22/05 4/29/05
If the turbine is equipped with a torque meter it is possible to determine the actual power output
directly, from which the overall energy efficiency can be calculated by energy balance. This is
called Method 2. Most existing steam turbines in Saudi Aramco plants do not have such torque
meters, but a revision to the standards is under consideration that will in the future require
torque meters on all turbines larger than about 5 MW.
@Seismicisolation
@Seismicisolation Page 35 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Energy efficiency for condensing type steam turbines cannot be calculated by method 1; only by
method 2 (based on torque meter reading). Fortunately, the point is moot because as a matter
of policy, the Company should not be using condensing turbines due to their extremely low
cycle efficiency, as noted earlier.
Gas Turbines, also called Combustion Gas Turbines (CGTs) are widely used in Saudi Aramco
plants, both as direct process drivers and as electrical power generators. Two kinds of energy
efficiency are important – the equipment efficiency, and the cycle efficiency (not the same as
“process” efficiency, which is covered in the next section). The energy efficiency of individual
components can be estimated using the methods of sections 3.2 and 3.6. This section focuses
on cycle efficiency.
A gas turbine is an internal combustion engine that operates on a thermodynamic cycle known
as the Brayton cycle. The principal components consist of an air compressor, a combustion
chamber, and a turbo-expander. Atmospheric air is drawn into the compressor and compressed
to several times atmospheric pressure. The pressure of the compressed air is further increased
by burning it in a confined space (the combustion chamber). The hot, pressurized combustion
gases are then expanded through a series of stationary nozzles and rotating turbine wheel and
blade assemblies, which results in rotation of the output shaft. The mechanical energy of the
shaft rotation is used to drive the gas turbine compressor and gas turbine accessories, as well
as the “process load” such as a generator, pump, or compressor. After giving up energy in
expansion, the gases may be discharged directly to the atmosphere or to a heat recovery
system. The flow of air, fuel, and combustion products through the gas turbine cycle is shown
schematically in Exhibit 3-21.
@Seismicisolation
@Seismicisolation Page 36 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The ideal Brayton cycle consists of the four processes that are shown on the pressure – volume
and temperature – entropy diagrams in Exhibit 3-22. The points identified in Exhibit 3-22
correspond to the similarly labeled points on the turbine in Exhibit 3-21, and they indicate where
the processes are occurring.
@Seismicisolation
@Seismicisolation Page 37 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
There are several different variations of the Brayton cycle, and each has a different calculation
procedure for determining the efficiency. Therefore a brief overview of cycle types and the
governing thermodynamic equations is provided here as background. For additional details, the
reader is referred to Saudi Aramco’s PEDD course MEX-214.
The Brayton cycle for power generation is usually employed in one of the following four
configurations:
• Simple (or open)
• Regenerative
• Combined Cycle
• Steam Injection (Cheng Cycle)
Only the simple cycle and combined cycle configurations are currently used in Saudi Aramco
plants. Nevertheless it is important to be aware of the other configurations in case they are
used in the future.
For the ideal simple cycle, the optimum pressure ratio to produce maximum net output is that for
which the compressor discharge and turbine exhaust temperatures are the same. In real life, the
pressure ratio at which maximum net work is produced is considerably lower, due to the effect
of compressor and turbine inefficiencies and combustion section pressure drop. Consequently,
the turbine exhaust temperature is considerably higher than the compressor discharge
temperature. For example, the typical exhaust temperature in a heavy duty industrial turbine is
about 1000°F, while the compressor discharge temperature is only about 650°F.
Simple cycle energy efficiency can be increased by as much as 25% using a “regenerator”,
which recovers some of the waste heat in the exhaust gas to preheat the compressor discharge
air before it enters the combustor, as shown in Figure 3-23.
@Seismicisolation
@Seismicisolation Page 38 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Although the regenerative cycle exhibits improved efficiency over the simple cycle, it is not
widely used in industrial gas turbine applications because the recuperative heat exchanger
required is invariably too large and expensive. In addition to capital cost, the large heat
exchanger size offsets one of the main advantages of using a gas turbine, which is its compact
size and small footprint. Pressure drops occurring through the heat exchanger further reduce
the theoretical improvement in efficiency that could be obtained.
The Combined Cycle is a combination of the Brayton cycle and the Rankine cycle, as illustrated
schematically in Exhibit 3-24. Heat in the CGT exhaust is used to generate HP steam in a type
of boiler called a Heat Recovery Steam Generator (HRSG). The HP steam is used to generate
additional power in a condensing steam turbine according to the Rankine cycle.
Because CGT exhaust usually contains about 15% oxygen, it is capable of sustaining
combustion. This characteristic can be exploited to generate additional steam and power by
employing supplementary firing in duct burners (not shown).
Because much of the waste heat in the gas turbine exhaust is used, the efficiency of the
combined cycle is considerably greater than that of the simple cycle, as follows:
ηCC = ηGT + ηHRSGηST - ηGTηHRSGηST
@Seismicisolation
@Seismicisolation Page 39 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
where:
ηGT = Gas turbine simple cycle efficiency
ηHRSG = Heat recovery steam generator efficiency
ηST = Steam turbine cycle (Rankine cycle) efficiency
Thus, for example, the combined cycle efficiency of a plant that utilizes a 30% efficient gas
turbine with a heat recovery steam generator that is 80% efficient and a 32% Rankine cycle
efficiency will be:
ηCC = 0.30 + (0.80)(0.32) - (0.30)(0.80)(0.32) = 0.479 = 47.9%
This is typical, although advanced cycles have been reported with operating efficiencies in the
range of 50-55% even under field conditions.
One of the drawbacks of the combined cycle is that the efficiency drops off significantly at part
load. Another is that the steam/power ratio is fixed. This makes it unsuitable for processes that
have variable steam and power demands. For such applications, the Cheng cycle can be a
better solution, as it has the flexibility to produce varying amounts of steam and power to match
plant process loads (see Exhibit 3-25). This is accomplished by injecting a portion of the
generated HP steam back into the combustion chamber, which is why it is also known as a
Steam Injection Gas Turbine (or STIG). Most of the applications have been in the 2-12 MW size
range, with process steam demands of 0-100 Klb/h. The drawbacks are (a) the water content of
the injected steam is lost up the stack, which can becomes uneconomic when makeup water
costs are high, and (b) higher maintenance costs due to increased corrosion of turbine internals.
It should be emphasized that the thermodynamic cycles described here are for power
production only, whether mechanical or electrical. If the turbine exhaust gases are used for
process heating, that is called cogeneration, which is covered in section 3.8.
The efficiency of the gas turbine is equal to the ratio of turbine power output to fuel heat input,
as shown in the following equation:
w net c p (TC − TD − TB + TA ) TC − TD − TB + TA T − TA
η= = = = 1− D
q c p (TC − TB ) TC − TB TC − TB
@Seismicisolation
@Seismicisolation Page 40 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Since the compression (A to B) and the expansion (C to D) are both isentropic in the ideal cycle
of Exhibit 3-22, the pressure ratio for these two processes is the same, viz.
k −1
TB ⎛ PB ⎞ k TC
=⎜ ⎟ =
TA ⎜⎝ PA ⎟⎠ TD
where:
PB = Compressor discharge pressure, psia or kPa abs
PA = Compressor inlet pressure, psia or kPa abs
k = Cp/Cv = 1.4 for air (assumed to be constant)
The substitution of these variables in the efficiency equation results in the following equation:
T − TA 1
η = 1− D
k −1 k −1
= 1− k −1
⎛P ⎞ k ⎛P ⎞ k ⎛ PB ⎞ k
TD ⎜⎜ B ⎟⎟ − TA ⎜⎜ B ⎟⎟ ⎜⎜ ⎟⎟
⎝ PA ⎠ ⎝ PA ⎠ ⎝ PA ⎠
Thus, for the ideal simple cycle, the efficiency is a function only of the pressure ratio developed
by the gas turbine compressor, and the performance of the compressor is extremely important
in overall gas turbine efficiency. In the real simple cycle, turbine efficiency is also dependent on
the turbine inlet temperature (Point C), and decreases as turbine inlet temperature increases for
a given pressure ratio.
Using the formula presented above for the efficiency of an ideal Brayton cycle, the calculated
efficiency of an ideal cycle with a pressure ratio of 11.5 would be
1
η = 1− 1.4 −1
= 0.5 = 50%
11.5 1.4
A real cycle that is operating at this pressure ratio might only have an efficiency of 30 – 32%,
because of the following critical differences between the real simple cycle and the ideal cycle:
• In the real cycle, the compression in the compressor section and the expansion in the
turbine are not isentropic due to compressor and turbine inefficiencies. Consequently the
real cycle turbo-expander provides less power, and the compressor uses more.
Compressor efficiency is especially important. For each 1% decrease in compressor
efficiency, there is about a 1.5 to 2% reduction in net power output.
• Factors such as compressor fouling can rapidly reduce compressor efficiency. Because of
the importance of compressor efficiency in overall CGT performance, compressor
performance should be routinely monitored during operation so that corrective action,
such as a compressor section water wash, can be taken in a timely manner.
@Seismicisolation
@Seismicisolation Page 41 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
• In the real cycle, a pressure drop through the combustion system between the
compressor discharge and the turbine inlet occurs, which reduces the available pressure
ratio for expansion across the turbine.
• Due to frictional resistance in the exhaust system ductwork, the gases do not exhaust at
atmospheric pressure, but rather at a slightly higher pressure, typically about 4 to 10
inches WC (0.14 to 0.36 psi, or 1 to 2.5 kPa) above atmospheric. This exhaust pressure
loss also contributes to a smaller pressure ratio that is available across the turbine. Each
4-inch WC increase in exhaust pressure results in a decrease of about 0.4% in both
output power and efficiency.
• There are also inlet pressure losses due to the ducting and filters, and these losses are
typically also about 4 inches WC (0.14 psi, or 1 kPa). Each 4 inch WC of inlet pressure
drop results in a decrease of about 1.4% in output and 0.5% in efficiency.
• The properties of the air and combustion gases, CP and k, are not constant, but rather
vary with temperature in a way that makes the real performance worse than the ideal
predicted performance.
@Seismicisolation
@Seismicisolation Page 42 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The most common parameter for measuring thermal energy efficiency of a power generation
cycle is the “heat rate”, defined as Btu of fuel input (usually HHV) per MW of net power
generated. Net power is defined as the gross power output of the expander minus the parasitic
power consumption of the air compressor and other essential components. The lower the heat
rate, the more thermally efficient the machine.
It is customary to measure the power output of the turbine in kilowatts (kW) if the turbine is used
as a generator drive and horsepower (HP) if it is used as a direct mechanical drive for a pump
or compressor. The two measures are related as follows:
1 HP = 0.7457 kW
For power generation applications, the same manufacturer is generally responsible for
supplying both the gas turbine and the generator as a set, and the power output is specified as
the kW output at the generator terminals; therefore, the performance measured also includes
the efficiency of the generator. Because large generators are generally very efficient (~ 98%),
the measured performance is still mainly that of the gas turbine.
For mechanical drive units, the power output is measured as the power delivered at the turbine
shaft coupling to the driven piece of equipment, even if the same manufacturer is supplying the
driven equipment; therefore, only the performance of the turbine is being measured.
Measuring the power output of a gas turbine for a generator drive is straightforward, as the
electric metering instrumentation measures electric power directly in kW. For mechanical drive
turbines, SAES-K-502, “Combustion Gas Turbines,” requires that the load coupling be a torque-
metering coupling designed for continuous operation. The shaft torque at the coupling is
measured in ft-lb (Nm in SI units). The power output is then calculated as follows:
2π NT
HP =
33,000
where:
HP = shaft horsepower
T = measured torque, ft-lb
N = shaft rotational speed, rpm
Since the theoretical minimum heat rate (at 100% efficiency) of any thermodynamic cycle is
3412.14 Btu/kwh, the thermal efficiency can also be expressed as:
3412.14
η=
Heat Rate
These equations are general, and apply to all power generation cycles.
@Seismicisolation
@Seismicisolation Page 43 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Industrial gas turbines normally are normally manufactured in certain standard sizes and for
certain fixed design parameters, e.g.
Normally, the published design and performance parameters are for “ISO” conditions, defined to
be:
Parameter Assumption
Suction air pressure 14.7 psia
Suction air temperature 59°F (15°C)
Suction air Relative Humidity 60%
Inlet pressure losses none
Exhaust pressure losses none
Gearbox and transmission losses none
Mechanical deterioration none
These conditions are almost never found in practice. Therefore the published ISO performance
has to be re-rated at actual site conditions. The site ambient conditions to be used in
calculating site turbine performance can be found in SAES-A-112, “Meteorological and Seismic
Design Data.” The procedure for re-rating a turbine is described in the next section.
The actual heat rate over any operating period is easy to determine from measured power
output and fuel consumption. The design heat rate at site conditions requires adjustment of the
ISO heat rate for changing site conditions of temperature, pressure, humidity, and mechanical
conditions, as described below.
Since the power produced by a gas turbine is directly related to the mass flow through the
machine, anything that reduces the density of the inlet air will reduce the mass flow through the
machine and, as a result, the power output.
@Seismicisolation
@Seismicisolation Page 44 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Inlet air temperature has the greatest impact on gas turbine performance. Gas turbine power
output will decrease approximately 0.5 % for each 1°F increase in inlet air temperature (0.9% for
each 1°C). This is because as the air temperature rises, the density of the air decreases, which
results in reduced mass flow through the turbine. Also, the pressure ratio developed by the
compressor will decrease, which results in a lower expansion ratio available across the turbine.
The reduced pressure ratio results in lower turbine efficiency and therefore lower power output.
Barometric pressure also has a significant effect on power output. At higher altitudes, the lower
barometric pressure means lower air density, which results in reduced mass flow through the
turbine and consequent reduction in gas turbine power output. However, there will be no net
effect on overall turbine efficiency, as the benefit of lower discharge pressure (to atmosphere)
will offset the penalty due lower suction pressure.
Humidity variations generally have an insignificant minor effect on gas turbine performance and
are usually neglected in performance calculations. For a given total atmospheric pressure, the
density of the mixture decreases as the amount of water vapor in the air increases, resulting in
reduced mass flow through the turbine and consequently lower power output. An increase in
the amount of water vapor increases the heat capacity of the mixture also, which reduces the
efficiency of the machine, but only very slightly.
Turbine manufacturers provide curves, charts, or other data that show the effects of ambient
conditions on gas turbine performance (see Exhibit 3-27 as an example).
@Seismicisolation
@Seismicisolation Page 45 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Instead of charts, it may be more convenient for computerized calculations to use the following
correlations that represent an acceptably accurate approximation for most industrial gas
turbines:
Temperature Correction factor for Power Output, TCFPWR = 1.22 – 0.375 (T/100)
Temp Correction factor for Heat Rate, TCFHR = 0.9656 + 0.0304 (T/100) + 0.0473 (T/100)2
The correction factors for changes in atmospheric pressure, generally due to altitude are more
straightforward:
Pressure Correction factor for Power Output, PCFPWR = PATM / 14.7 = Altitude (ft) /7500
@Seismicisolation
@Seismicisolation Page 46 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
As indicated earlier, the correction factors for humidity variations are very small, but can be
estimated either from the chart or the equations given below.
Obstructions to air flow in the inlet air path (inlet filters, silencers, and duct work) cause the total
pressure at the compressor inlet to decrease, which reduces the gas turbine's power output in
two ways. First, the lower density reduces the mass flow rate of the working fluid (air). Second,
it causes a decrease in compressor discharge pressure, which in turn results in a lower turbine
pressure ratio, and therefore power output. The decrease of the compressor inlet pressure also
affects the heat rate. A lower turbine pressure ratio reduces thermal efficiency, which increases
the heat rate.
The magnitude of the inlet pressure correction as a percentage of machine rating will vary,
generally being a larger percentage for smaller machines. Typically, a 4 in. H2O decrease in
inlet air pressure will cause a 1.7% decrease in the turbine power output, a 0.7% increase in the
heat rate, and a 2°F increase in the exhaust temperature. Exhibit 3-30 shows the effect of inlet
pressure loss on turbine power and heat rate.
@Seismicisolation
@Seismicisolation Page 47 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
@Seismicisolation
@Seismicisolation Page 48 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
In equation form, the correction factors for inlet pressure drop are:
Flow restrictions in the exhaust system (heat recovery equipment, silencers, and duct work)
have a similar effect, for comparable reasons. Exhibit 3-31 shows the effect of outlet pressure
loss on turbine power and heat rate; the corresponding equations for correction factors are:
@Seismicisolation
@Seismicisolation Page 49 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
SAES-K-502 requires a power correction factor for mechanical transmission losses as well, if
applicable. A power correction factor will typically be a data item submitted by the turbine
vendor for base load. For example, if the reduction gear has a 97.5% mechanical efficiency, the
power correction factor would be 0.975.
Since the mechanical transmission losses are a direct reduction of output for a given fuel flow, a
heat rate correction factor can be calculated simply as the reciprocal of the power correction
factor: 1/0.975 = 1.026 in the example cited.
The vendor’s stated ratings are invariably for a brand new gas turbine. Some deterioration of
performance occurs naturally with turbine operation due to mechanical wear. To account for this
deterioration, SAES-K-502 requires that the following factors be applied:
• 0.90 for generator drive, both single shaft and multiple shaft units.
• 0.90 for mechanical drive - multiple shaft units.
• 0.85 for mechanical drive - single shaft units.
Gas turbine heat rate will also deteriorate (increase) as the turbine ages. The percentage
change in efficiency or heat rate will typically be less than the percentage change in power
output. No standard correction for heat rate deterioration is specified in SAES-K-502. If the
engineer wants to predict future heat rate for the turbine, a percentage change equal to about
one-half of the percentage change in output is a reasonable estimate.
Finally, if any of the shaft-driven auxiliaries, other than those auxiliaries included in the turbine
manufacturer’s rating, are driven by the turbine, the power they require must be subtracted from
the turbine power output.
Once all the correction factors have been determined, they are applied as multipliers to the ISO
rated conditions, with the exception of the auxiliary power correction. The auxiliary power
correction is subtracted from the turbine power output after the other corrections have been
applied. Thus, the site rated power is calculated as follows:
Site Rated Power = ISO Rated Power x Temperature Correction x Altitude Correction x
Humidity Correction x Inlet Losses Correction x Exhaust Losses
Correction x Transmission Loss Correction x Deterioration and
Contingency Correction - Auxiliary Power
Site Rated Heat Rate = ISO Rated Heat x Temperature Correction x Humidity Correction x
Inlet Losses Correction x Exhaust Losses Correction x Transmission
Loss Correction x Deterioration and Contingency Correction
@Seismicisolation
@Seismicisolation Page 50 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
An EPI trend chart for the cogeneration system at Abqaiq Plants is presented in Exhibit 3-32 as
a sample. Observe that the heat rate is considerably lower than the typical value of 10,000
Btu/kwh for simple cycle gas turbine installations.
8000
7000
6000
Heat Rate, Btu/kwh
5000
4000
3000
2000
1000
0
Jan. Feb. March April May June July Aug. Sep. Oct. Nov. Dec.
A detailed description of the methodology and instructions on how to calculate the net heat rate
will be added during the next revision.
@Seismicisolation
@Seismicisolation Page 51 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The Solomon Energy Intensity Index (EII) for oil refining builds up the overall plant energy index
from the energy indices for individual process units. The process unit EIIs are an acceptable
measure of energy efficiency, provided that the data being used are accurate. No further model
development work is required. All that each refinery has to do is calculate and monitor the EII
values for the major energy-consuming process units.
The CGEY Energy Efficiency Index (EEI) for gas processing, on the other hand, is formulated
as Fuel Consumption per GPU for the entire plant, where:
This formulation cannot be disaggregated into EEIs for individual process areas such as inlets,
stabilization, gas treating, DDPC, Sales Gas compression, and sulfur recovery. Therefore, the
process area EPIs have to be computed on the basis of output from the HMB simulation model.
Process area indices for E&P plants also have to be computed from HMB simulation model
output, because there is no established industry-standard energy index for such process
operations.
In general, each process unit or area will have only two energy indices: a fuel index and a power
index:
@Seismicisolation
@Seismicisolation Page 52 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
FuelConsumption PowerConsumption
FuelIndex = and PowerIndex =
CapacityParameter CapacityParameter
In cases where the process unit uses both steam and direct fuel, it may be desirable to have
three indices – a process fuel index, a steam index (= boiler fuel index), and a power index.
The energy consumed in the Utilities area of the plant and by common facilities (eg admin
buildings, perimeter lighting) must be properly allocated to the process units. The process area
indices are expressed in terms of a single “capacity parameter”, which should be chosen to
represent the operating rate of each process area – e.g., combined feed gas flow, combined
condensate flow, sweet wet gas, sweet dry gas, sales gas, or dry crude oil flow. The process
areas do not have a cost index, as the capacity parameter is not a distinct salable “product”.
All Saudi Aramco refineries subscribe to the benchmarking service from Solomon Associates
(Dallas, Texas). As a subscriber, Saudi Aramco has access to their proprietary equations and
calculation procedures, which have been encoded into an Excel spreadsheet available from the
Energy Systems Unit of Consulting Services Department in Dhahran. The basic data required
are shown in Exhibit 4-1a. The calculation for Complexity Factor is shown in Exhibit 4-1b. The
EII calculations for the process units are shown in Exhibit 4-1c.
@Seismicisolation
@Seismicisolation Page 53 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
@Seismicisolation
@Seismicisolation Page 54 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
To effectively use the EII for process efficiency improvement, we should monitor the EII of each
unit preferably every shift, but at least on a daily basis. The EII should be calculated for each
set of data (easily done using a macro). Once done, however, the results can be archived for
future reference as a “flat file”, so that the calculation does not have to be repeated for past
periods. A sample trend chart of monthly data is shown in Exhibit 4-2.
Whole Refinery
180
160
140
120
Solomon E I
100
80
60
40
A ctual Index
20 Wo rld's B est
0
2/13/02 5/24/02 9/1/02 12/10/02 3/20/03 6/28/03
Saudi Aramco has five gas-processing plants (some with NGL recovery and some without) and
two NGL fractionation plants. Fairly accurate spreadsheet-based HMB simulation models have
been developed by the GO Technical Support Department (Center of Excellence) in Dhahran.
These models require only basic feed rate, product flow, and process configuration data as
input, from which they can compute all internal streams with a very high degree of accuracy.
The model is effectively an “inferential meter”, in which the unknown parameter (usually a flow)
is inferred by calculation from known accurate values of other parameters using conservation
laws. It is recommended that these computed internal stream flow rates be used for calculating
the EPIs rather than the actual metered values, as the inferred values are far more reliable.
The “process area” indices are designed to monitor the energy efficiency of groups of process
units that together produce recognizable intermediate products – such as sweet gas, dry gas,
HP gas, etc., and whose capacities can be measured by the flow rate of those products. Each
of these intermediate products is characterized by the allocated amount of energy for boiler fuel
(for steam), direct process fuel (in furnaces), and electrical power.
@Seismicisolation
@Seismicisolation Page 55 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
CONFIDENTIAL Page 56 of 79
@Seismicisolation
@Seismicisolation
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
• Divide the plant into major process areas (as illustrated in Exhibit 4-3) which have clearly
defined feeds coming from other process units and clearly defined intermediate “products”
that go to other units. These process areas may include more than one process unit, or
may be only part of a process unit.
• Use feed and product data (flow rates, compositions, temperatures, pressures) and flow
rates + concentrations of key intermediate streams from the PI system to run the steady-
state HMB simulation models developed by GO. Extract the capacity parameters.
• Extract the steam and fuel consumption calculated by the simulation model. Convert
steam usage to proportional fuel usage in the boilers. Add the fuel for steam plus direct
fuel use in the process unit (area) to get the total fuel consumption. Divide this number by
the selected capacity parameter to get the Fuel Index.
• These models do not currently calculate the power consumption. Instead they allocate
total measured plant power import to each unit based on installed motor HP. This is not
acceptable for calculating the power indices. The recommended procedure is to calculate
the power consumption of the major compressors and pumps from PI data using the
equipment spreadsheet models described in sections 3.1 and 3.2. Let us say that the
total is X megawatts, and that the measured total power import is Y megawatts. Then the
balance of unallocated power usage for the process units is (Y-X) megawatts. This
amount must be allocated to each process unit based on installed HP of non-major
equipment. The allocated and calculated power consumption for each unit is then totaled,
and used to compute the Power Index.
Sample results for HGP following this approach are presented in Exhibit 4-4. These indices
clearly indicate which units are the major energy consumers, and can pinpoint specific areas of
the plant that need attention both in terms of operating problems and opportunities for major
cost reduction; in short they provide actionable information not available from the CGEY indices.
The plant HMB simulation models are quite complex and not practical to display in toto; but
sample input data screens are shown in Exhibits 4-5a and 4-5b. The output from this simulation
model is used as the input for the EPI model, which has to be run in batch mode (off-line) for
each set of data. It is not possible to run the EPI model using live data, as the measured input
parameters will be changing continuously, and they will not be mutually consistent.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 57 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 58 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Sample output data and trend charts from the EPI model are shown in Exhibits 4-6 and 4-7.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 59 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
22.0 1100
21.8 1060
MMBtu/MB Condensate Feed
21.4 980
Fuel Index
21.2 940
Power Index
21.0 900
5/1 5/8 5/15 5/22 5/29
10 60
8 56
Fuel Indices, MMBtu/tonne
6 52
Boiler Fuel
Process fuel
4 Power
48
2 44
0 40
5/1 5/8 5/15 5/22 5/29
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 60 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Although relatively simple in terms of process complexity when compared to oil refineries or gas
processing plants, GOSPs consume a huge amount of energy. The company’s experience at
Abqaiq and Safaniya Onshore Producing Plants has shown that there is significant potential for
economically feasible energy cost savings, and that the EPIs are well worth monitoring.
The recommended procedure is identical to that described for Gas Plants in section 4.2, except
that HMB models are generally not available. Therefore, we can either use reconciled data
directly from the PI (Plant Information) system, or HMB models must be developed from scratch.
The EPIs for Safaniya Onshore Plants are described here as a guide.
The block flow diagram for Safaniya is typical of large GOSPs, except that some may include oil
stabilization, and others may have a lower degree of gas processing.
We can choose the process areas in various degrees of detail. For example, Exhibit 4-9 shows
a slightly more detailed version of Exhibit 4-8. The Safaniya EPI team decided to divide the
plant into three process areas, shown below. However, they could have elected to choose six,
by breaking the Gas Plant into Inlets, DDPC, and Compression.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 61 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Exhibit 4-9: Simplified Process Flow Diagram for Safaniya Onshore Plants
CONFIDENTIAL Page 62 of 79
@Seismicisolation
@Seismicisolation
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Safaniya Plants has an extensive PI data system. The raw data flow and energy consumption
for each process area must be compiled on a daily basis. Ideally these data should be checked
for consistency by going through a formal Data Reconciliation procedure, but this is time-
consuming to do manually, and not practical without installation of sophisticated software.
Nonetheless, a quick data consistency check for material balance is easy to do, as illustrated in
Exhibit 4-10, and should be done.
Exhibit 4-10: Daily Average Feedstock and Product Flow Rates (from PI)
The raw flow data show that more dry AH product is being produced than wet feed (including
water) is coming in – a clearly impossible situation. This is a red flag to the process engineer
that the data being collected in PI is not to be trusted, and requires manual investigation to
correct the error. The material balance for AM crude, on the other hand is totally consistent.
They show a “yield” of around 95%, which is consistent with the observation that approximately
5% of the incoming wet crude (expressed on a water-free basis) flashes off as gas due to
pressure let down.
Once the data are reconciled, meaningful EPIs can be calculated, and be used as the basis for
troubleshooting or identifying areas of opportunity for energy efficiency improvement.
The EPI calculations and trend charts for the gas plant are shown in Exhibits 4-11 and 4-12, as
samples.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 63 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Exhibit 4-11: Weekly Process EPIs for Gas Plant (R-74), sample calculations
Exhibit 4-12: Weekly Process EPIs for Gas Plant (R-74), trend chart
0.60 6
0.50 5
0.40 4
0.30 3
0.20 2
Fuel Index
0.10 1
Pow er Index
0.00 0
1/1
2/1
3/1
4/1
5/1
6/1
7/1
8/1
9/1
10/1
11/1
12/1
It is not enough to merely report these EPI trends; they need to be analyzed and interpreted
correctly. For example, it may be observed that there is no correlation with the ambient
temperature, and some times the fuel and power indices more in parallel, and other times not.
What does this tell us? It is the answers to such questions that will lead to process insights and
point the way towards efficiency improvement.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 64 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
For oil refineries, the Best Practice committee (energy team) has decided that the Solomon
Energy Intensity Index (EII) for the overall refinery, as described in section 4.1, is satisfactory for
their needs and will continue to be followed. Therefore new product EPIs will not be developed.
For the gas processing plants, Saudi Aramco currently subscribes to the plant benchmarking
service offered by Cap-Gemini Ernst & Young (CGEY), which includes their overall Energy
Efficiency Index. This will continue to be used for comparing Saudi Aramco plants against
global competitors in the gas industry. However, they are in agreement that it would be useful
to be able to determine the energy efficiency of producing individual products in addition. The
objective of the product EPIs described here is to fulfill that need.
5.1 Methodology
The basic concept underlying EMSC’s model for Product EPIs is to properly allocate energy
consumption in each of the process units to the appropriate product. There are many possible
allocation strategies – by volume, by market value, by Btu content, by value-added, etc. We
have chosen the “value-added” method, because it yields the most meaningful and actionable
results, and is consistent with the approach used to calculate financial and other corporate KPIs.
Waste W
Product A
E1 E2
Energy
input E
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 65 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Step one, as always, is to produce simplified Process Flow Diagrams (PFDs) and a validated
Heat and Material Balance simulation model.
Once the material and heat balances have been determined, the “Value-Added” allocation
procedure is applied to calculate the overall product EPIs using the equations cited below.
Although these equations are presented for only a single stream in each of the categories –
feed, product, byproduct, waste, and energy input, they can be easily generalized to the case
where there are multiple streams in these categories.
F = A+B+W
E = E1 + E2 + ES = EA + EB
A B
XA = ; XB =
F−W F−W
⎛ VA ⎞ ⎛ VB ⎞
E A = ⎜⎜ ⎟⎟.E; E B = ⎜⎜ ⎟⎟.E
⎝ VA + VB ⎠ ⎝ VA + VB ⎠
PA = VA – EA; PB = VB – EB
A, B, F, and W are the mass flow rates for product, byproduct, feed, and “waste” streams.
CA, CB, and CF are the product sales prices/values and feedstock cost respectively.
XI = Mass fraction of F that becomes product I.
VI = gross value added by product I.
EI = amount of energy allocated to product I.
En = amount of energy consumed by process unit “n”
ES = combined energy consumed by other site facilities unrelated to a specific process unit.
PI = net value added (or “profit”) from sales of product I = gross value added minus the allocated
energy cost.
For our purposes, a product is defined as any desired output from the process which makes a
profit, ie PA > 0. Sales gas and condensate fall into this category. A byproduct is defined as an
unavoidable output from the process of making a product, which does have some market value,
but which may or may not make a profit, for example sulfur. In cases when PB < 0, the energy
allocations to the product and byproduct should be adjusted such that PB = 0, i.e., we set
EB’ = VB, and EA’ = E – VB
One additional clarification is in order. A “waste” stream is defined as that portion of the feed
stock that does not eventually become product or byproduct, and includes streams such as flare
gas and off-spec (unsalable) product that can only be used as fuel. It should be noted that the
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 66 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
allocation procedure does not make use of E1 and E2. This is not an oversight; it is a deliberate
simplification in which the energy consumption for the plant as a whole (including non-process
users such as buildings, site lighting, etc.) is distributed rationally among the profit-making
products. The proposed model gives reasonable and meaningful results for all combinations of
cost scenarios, assuming that 0< CF < CA, CB):
There is a more sophisticated version of this procedure, which considers each process unit to
be a self-standing plant, treating intermediate streams as products from upstream units and
feedstocks to downstream units. Energy consumption is accrued for each product as it passes
through the various units and ultimately shows up in the product tanks. Although rigorously
more correct, it is also much more complicated, and difficult to implement. Accordingly, a policy
decision was made to adopt the simplified approach. Perhaps, at some later date, after all the
business lines have implemented the simplified EPI models, the EMSC might consider
recommending an upgrade to the more rigorous models.
The mathematical model of the plant required to calculate the EPI does not necessarily follow
the “process unit” modular structure. Rather, the goal is to set up a sequence of modules that
are punctuated by splits and mixes in various intermediate streams in such a way that they can
be easily attributed to a specific product. In some cases, therefore, we may group multiple
process units together into a module, and in other cases, a process unit may be sub-divided into
a number of sub-systems, or even individual items of equipment. The structure of the overall
EPI model for JGP is shown in Exhibit 5-2 for illustrative purposes.
The heat and material balance is determined using simulation models. It does not matter which
software platform (Hysys, Excel, AspenPlus, PIMS, etc.) is used as long as it provides
consistent and accurate results.
Fortunately, validated spreadsheet-based models have been developed for all Gas Plants by
the Gas Operations Center of Excellence in Dhahran, and are freely available to all proponent
facilities. They have “switches” to turn process units or equipment on and off, closely following
actual plant operations.
For E&P facilities, rigorous HMB models using Hysys are currently being developed by Process
& Control Systems Department (P&CSD) for all major GOSPs. Validation is expected to be
completed by mid-2006. For pipelines, SWID, and production wells, HMB models do not
currently exist, and will have to be developed by each proponent facility.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 67 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The major input data for these models are the inlet flow rates, pressures, and compositions. In
addition we need measured steam flow rates for individual users, and fuel consumption for the
boilers, fired heaters, and thermal oxidizer. Regarding electrical power, current practice in
Company plants is to meter only the total power imported from or exported to the national grid;
consumption by individual equipment or process unit is not measured. The power consumption
in individual process units can be approximated by allocating the total imported power in
proportion to the horsepower of the running motors and equipment in that area. The calculated
product flow rates are compared against measured values by the custody flow meters to ensure
that there are no major errors. For intermediate streams, we recommend using calculated flows
from the model rather than metered values, as the former are considered to be more reliable.
The oil refineries are already calculating and reporting their Solomon EIIs on a monthly basis,
and have had the capability to do so for several years.
Reported results for the past years are presented in Exhibit 5-3.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 68 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
120
100
Solomon EII
80
60
40 RTR
RR
JR
20 RB R
YR
0
1999 2000 2001 2002 2003 2004
Exhibits 5-4 through 5-8 show the spreadsheet calculations for the JGP product EPI model, to
illustrate the general computational approach and procedure.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 69 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 70 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Note that the calculated total production matches total feed inputs to the plant, which it must.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 71 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
All these calculations have to be done for each set of data. The recommended time interval is 1
day (24 hours). A macro can be set up within Excel to collect data over 1-day intervals for any
desired period, say 6 months to a year, and provide an output table listing each product EPI on
a daily basis. Of course PI tag numbers have to be provided for referring to the correct data
values, per Exhibit 5-9.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 72 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Sample product EPIs on a monthly basis for 6 months in 2002 for the two grades of butane
product are illustrated in Exhibit 5-10. They show very clearly that almost as much energy is
required for refrigerating the product during storage as in producing it. Therefore one way to
reduce energy costs would be to minimize the residence time of product in the storage tanks
through better inventory management. Such actionable insights cannot be gained from the
feed-based indices offered by CGEY or Purvin & Gertz.
100 5000
80 4000
K Btu / bbl
kwh / bbl
60 3000
40 2000
Un-Refg Un-Refg
20 1000
Refg (export) Refg (export)
0 0
Nov Dec Jan Feb Mar April Nov Dec Jan Feb Mar April
1.40
1.20
1.00
$ / bbl
0.80
0.60
0.00
Nov Dec Jan Feb Mar April
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 73 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Abqaiq Plants and Hawiyah Gas Plant (HGP) have been the pioneers in implementing their
EPIs on-line and making them available through their web-site. Abqaiq Plants rolled out their
EPIs in 2004, and HGP in 2005.
This is the main menu that lists all available process KPIs (there is another similar menu for
Maintenance KPIs and Balanced scorecard KPIs). To get further details on the item of interest,
e.g., Energy Usage, just click on it. This will display the trend chart for the default period
(previous 2 days), as in Exhibit 6-2. If you want to see the trend charts for some other period,
click on the Index Tag Name, enter the desired start and finish dates, and you will see a chart
such as Exhibit 6-3.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 74 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Exhibit 6-3: Energy Usage Trend Chart for User Specified Period (Abqaiq)
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 75 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The EPI deployment at HGP took considerably less time, about 2 man-months for off-line model
development, and another 2 man-months for online implementation (data conditioning and web-
page design/programming):
(a) Writing logic (into Excel cells) for detection of gross errors in measured PI data
(b) Selection and specification of default values, when PI data were determined to be in error
(c) Web-page design and programming
(d) Testing and sorting out IT communication protocols
Sample screen shots are provided in Exhibits 6-4 to 6-8 for illustrative purposes.
Any authorized company employee can access HGP’s departmental web-site via the intranet by
typing http://hgpis to reach the page where key departmental performance parameters (feed
rates, production rates, energy consumption, etc.) are listed, as shown in Exhibit 6.4.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 76 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
The user can view the instantaneous EPIs for any product (e.g., Exhibit 6-5), any process (e.g.,
Exhibit 6-6) or any equipment (Exhibit 6-7), by clicking on the appropriate button. There is a
button ( ) next to each index name which will display the trend chart for that EPI over any
specified period (e.g., February 26th to 28th) and sampling interval (e.g., two hours) as shown in
Exhibit 6-8 for Sales Gas Cost EPI.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 77 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 78 of 79
Document Responsibility: CSD/ESD/Energy Systems Unit SABP-A-004
Issue date: 19 March 2006
Next Update: 19 March 2009 Energy Performance Indices
Revision Summary
19 March 2006 New Saudi Aramco Best Practice.
CONFIDENTIAL @Seismicisolation
@Seismicisolation Page 79 of 79