Compressor Optimization

EXECTUIVE BRIEF
Maintaining Productivity
Yields Through
Instrumentation Analysis
Instrumentation repair, maintenance, or
replacement is a key essential of any project
or plant operation analysis. Whether you are
measuring existing energy consumption norms in
gas pipelines, scoping a maintenance turnaround,
or defining criteria design and conditions, all
procedures and resulting analysis are measured
to safety and maximize your capital return.
Through an Instrument Failure Analysis, FAST
sponsored by:
engineers determine the root cause of your
failed gauge – whether it is a WIKA product or
not – and recommend a solution to help prevent
further breakdowns. Equipment inspections,
by company regulation or government
mandate, are built in measures to assure your
productivity.
3
The pursuit of
‘zero leaks’ 6 Compressor
Optimization 16 Compressor
Optimization 23 Analysis yields
turnaround
Part 1 Part 2 benchmarks
for allowance,
contingency
More than half your
engineers are retiring... SM
who’s handling your next “Oh

$#%!” moment?
SM
Instrument Audit Turnaround Instrument Planning Instrument Failure Analysis Instrument Safety Training
Did you know that an average of 8 failing gauges are within 20 feet of every
employee working in your plant? These ticking time bombs make your team unsafe,
less productive and can even lead to serious disasters.
WIKA can take the worry out of instrumentation with our Full Service Audit Team
(FAST). Using our proven process, WIKA’s experts can lower your costs, make you
safer and reduce downtime with our FAST Total Care Program.
Let us show you how today by downloading our free eBook at

www.WIKA-FAST.com/ebook or by calling us at 855-651-FAST (3278).
Originally published February 6, 2012
The pursuit of ‘zero leaks’
by Christopher E. Smith
U
S President Barack Obama signed the Pipeline Safety, Regulatory
Certainty, and Job Creation Act of 2011 into law Jan. 3, reauthorizing
the Department of Transportation’s existing pipeline safety programs
through 2015 while also placing new requirements on both pipeline
operators and regulators.
On the operators’ side of the ledger, the law increases maximum penalties for
individual violations to $200,000 from $100,000 and for a series of violations to
$2 million from $1 million. It also requires gas transmission pipeline operators
to report within 18 months any pipeline segments with insufficient maximum
allowable operating pressure (MAOP) records, to report incidents pushing
operating pressure beyond MAOP within 5 days of their occurrence, and to
consider seismic activity when evaluating pipeline threats.
DOT requirements
Among the requirements placed on DOT regulators, meanwhile, was maintaining
a map of high-consequence areas (HCAs) on the National Pipeline Mapping
System (NPMS) and to develop an NPMS awareness program within a year. It also
requires DOT within 18 months to both develop guidance for operators to share
system-specific information with emergency responders and establish time limits
on leak and accident notifications to both emergency responders and other state
and local officials.
The timeline to establish requirements for gas transmission pipeline operators

to confirm their pipelines’ physical and operational characteristics, including
MAOP, for pipelines in Class 3 and 4 zones and HCAs is even shorter at 6 months.
This relatively short timeframe recognizes increased public awareness of pipeline
safety in the wake of significant accidents on both oil and gas pipelines.
3
Oil & Gas Journal :: EXECUTIVE BRIEF :: sponsored by
The law also requires DOT’s Pipeline and Hazardous Materials Safety
Administration to issue new pipeline safety standards requiring operators to
install automatic or remote-controlled shut-off valves and excess flow valves
in new or replaced transmission pipelines. It also authorizes $110 million/
year in safety related grants for use by states in damage prevention programs,
emergency response training, technical outreach to local communities, and
one-call system improvements.
The promulgation of such programs on a state level has leapt in importance in

areas like the Eagle Ford and Marcellus shales, where pipeline development has
gained momentum as operators seek to bring their gas and liquids to market.
Cooperation and coordination among states also is important, particularly in
areas like the Marcellus where a single, relatively small project can often cross
multiple state boundaries.
Pennsylvania acts
Pennsylvania Gov. Tom Corbett signed his state’s “Gas and Hazardous Liquids
Pipelines” Act into law in December 2011, authorizing its public utility
commission (PUC) to conduct pipeline safety inspections in coordination with
PHMSA and to regulate pipelines without declaring them a public utility. This
latter point was particularly important given concerns from property owners
regarding imminent domain, which public utilities can exert. If the PUC had been
allowed to regulate only pipelines designated as utilities many would have gone
uncovered (bad for safety) or had to have been reclassified as utilities (bad for
property owners).
Now that the property owners’ rights have been preserved, it is incumbent on
them and their communities to live responsibly in the company of the new
pipelines. The safety-related money authorized for disbursement to states
through PHMSA under the new federal pipeline safety law can help this happen.
Both the federal law and its Pennsylvania counterpart are encouraging. Not
just because they help codify the importance of pipeline safety, but because
they recognize that it is best achieved as a partnership: between regulators and
operators, between the federal government and smaller jurisdictions, and finally,
4 between all of these and the citizens at large.

Colonial Pipeline Co. Chief Executive Officer Tim Felt once aptly described “zero
leaks” as the only reasonable goal for the US pipeline industry (OGJ Online, Mar.
25, 2009). Without each of these parties’ active participation, this goal cannot
even be seriously approached, much less attained.
Christopher E. Smith is Pipeline Editor for Oil & Gas Journal.
5
Originally published January 9, 2012
COMPRESSOR OPTIMIZATION—1
EnergyTRANSPORTATION
recovery guides natural
gas pipeline system efficiency
COMPRESSOR OP TIMIZ ATION —1
Energy recovery guides natural g

by Shaghayegh Khalaji
E pipeline system efficiency

conomic analyses demonstrate the operational benefits of using
particular energy recovery and scrubber technologies on natural gas
transmission systems. The first article of this series, presented here,
Shaghayegh
details theseKhalaji
analyses as applied to Iran’s gas system. The concluding
National Iranian Gas Co.
article will focus on the effect of scrubbers and other equipment in attempting to
Tehran
maximize capital return.
Economic analyses dem- across the c

Background NATURAL GAS VOLUMES Table 1
onstrate the operational compressor s
Consumed
Iran is one of the largest
benefits of using particular Shipped as fuel Consumed, ly 100 km a
energy recovery and scrub- Country –––––––––– MMscfd –––––––––– % information
consumers of natural gas in
ber technologies on natural Austria 8,118 223 2.7 consumption
US 652,373 17,638 2.7
the world. Several transmission
gas transmission systems. Italy 84,897 591 0.7 sor stations in
Turkey 36,599 190 0.5
The first article
pipelines move natural gas of this se- Czech Republic 6,216 538 8.6 not be direc
ries, presented here, details Poland 16,202 389 2.4 Iranian statio
across the country, withas applied to
these analyses ences in line l
Iran’s gas system.
compressor stations roughly The con- need to first
cluding article will focus on the effect of scrubbers and other ber of consumption indices.
100 km apart. Existing
equipment information
in attempting regarding
to maximize capitalgas consumptionTable
return. by compressor
1 shows the energy consumed trans
stations in Europe cannot be directly applied to Iranian stations somedue differences
countries.
Background
in line length and the need to first define a number of consumption indices. Comparison, however, requires developme
Iran is one of the largest consumers of natural gas in the indices. This article compares the power req
world. Several transmission pipelines move natural gas transfer via pipeline by first calculating the
Table 1 shows the
by each station energy consumed
per 1 million cu m gas/
km/day. SIMULATIONEXAMPLE
SCHEMATIC PIPELINES Table 2
transmitting gas Natural Power

Table 2 shows this power index for gas Compressor
Gas compressor
index,
Pressure transferred, station Maximum Mw/
in some countries.
a variety of countries. Gas inlet drop Length,
= 58.95 kPa 1,000Scrubber out power,
outlet
pressure, (km × 1,000Gas outlet To
From Country km cu m/day Mw bar cu m/day)
Table 3 shows the calculated power
previous Gas compressor Air cooler
Scrubber power50
requirement for gas transfer of 1station
mil- China 1,084 5 24.88 0.00459
Comparison,
lion cu m gas/km/day on three Iranian
The Netherlands
Fuel
Romania
720
100
8.7
51
32.1
25
69
63
0.00512
0.00491
Gas
US 1,200 14.3 92 77
Stack 0.00537
however,
pipeline networks. Iranianrequires
power re- Algeria 520 Combustion
335 Hot compressor
822 72 0.00472
6 Pressure
Global averageratio set gas 0.004773
quirement are development
about 60% moreofthan Exhaust
the other coun-
tries. Liquid Gas compressor
power
Oil &GAS
IRANIAN NATURAL Gas Journal
PIPELINES
Inlet air :: EXECUTIVE
High-pressure BRIEF
Table 3 :: TEMPERATURE,
sponsored by Power = 2.456 Mw
Table 4
Investigation air
Power REQUIRED POWER
Country km cu m/day Mw bar cu m/day)
able 3 shows the calculated power
irement for gas transfer of 1 mil- China 1,084 5 24.88 50 0.00459
The Netherlands 720 8.7 32.1 69 0.00512
cu m gas/km/day on three Iranian Romania 100 51 25 63 0.00491
US COMPRESSOR
1,200 OPTIMIZATION—1
14.3 92 77 0.00537
line networks. Iranian power re- Algeria 520 335 822 72 0.00472
Global average 0.004773
ement are about 60% more than
other coun-
.
IRANIAN NATURAL GAS PIPELINES Table 3 TEMPERATURE, Table 4
estigation Power REQUIRED POWER
ameters Trans- index, Compressor
ferred gas, Nominal Mw/ Input gas power
classic defi- Length, 1,000 power, (km × 1,000 temperature, requirement,
n of effi- Pipeline km cu m/day Mw cu m/day) °C. kw
cy of ther- Line 1 400 46 160 0.008696 10 39,651
per
d y n1 amillion
m i c cu mLinegas/
Line 2
3
656
656EXAMPLE 91.7
90PIPELINES 478
440
0.007943
0.007461
15 40,679
Table 2
20 41,693
Average 0.007905
ems centers Natural 25 43,692
30 43,678 Power
ws this power
ncreasing en- index for gas Compressor
35 44,653 index,
transferred, station
40 Maximum 45,617 Mw/
per unit of
ntries. Length, 1,000 power, pressure, (km × 1,000
ECONOMIC
fuel. suitable INVESTIGATION RESULTScompares
power indices.Country This article the
km Table 5
cu m/day Mw bar cu m/day)
umed
ws the calculated Compressor Equival- Annual
eases in ei- power ScrubberrequiredChina for gas transfer
power via pipeline
1,084
ent gas gas 5by EQUATION
24.88 50 0.00459
r gas transfer of 1pressure
mil- Thesaved,
Netherlands saving,720 saving,
8.7 32.1 69 0.00512
pressure
m/day on threefirst
temperature calculatingRomania
change
Iranian the kw power needed cu m/day 100 by each station perEf1=25
cu m51 wp \ Qf + E
million cu m 63 gas/km/day.
(1)
0.00491
10 psi decrease US 1,682 1,200
12,615 14.3
4,604,475 92
where: 77 0.00537
rks. Iranian
increase the power 5 psire-
decrease Algeria845 6,338520 335
2,313,188 Ef 822
= desired efficiency72 0.00472
5 psi increase Global
–853average –6,398 –2,335,088 wp = desired work, or work done for pres- 0.004773
about
er 60% more
produced 10than
Table 2increase
psi shows this power index
–1,714 –12,855for a variety –4,692,075 of countries. sure increase
compressor Qf = quantity of energy from the fuel
n- E = station’s electric power consumption
ons. Pres-
is a desired Table 3 shows the calculated power requirement for gas transfer of 1 million cu
IRANIAN TEMPERATURE,
ease in this
mNATURAL
context, GAS PIPELINES
temperature Table 3 Table 4
gas/km/day on three
HAIRPINIranian
COOLER-DESIGN pipeline networks.
EFFECTS Iranian
REQUIRED powerPOWER requirement Table 6are
ot. Power Optimum
quation 1 definesabout desired60% more
efficien- than the other countries.
Trans- index, Compressor
cooler
ferred gas, Nominal Mw/ Input gas Status quo designpower
fi- index for comparing
an different
Length, 1,000 power, (km × 1,000 temperature, requirement,
ons.
fi- Pipeline km cu m/day Input gas flow,MwMMscfd cu m/day) °C. 100 100kw
Investigation parametersOutlet gas pressure leaving origin station, psi 1,000 1,000
alculating Line
r- required
1 power
400 in- 46 Pressure drop
160in air coolers, %0.008696 9.6 1.5
Gas pressure in downstream station inlet, psi 10 780 789.739,651
es extracting 2 The
Lineenthalpy classic
656 definition
of output 91.7 Downstream478of efficiency of
station compressor thermodynamic
0.007943
power consumption, kw systems
15 centers 34,875on 33,04340,679
c Line 3 656 90 Downstream440 station compressor 0.007461 20 kw
power consumption decrease, 1,833 41,693
pressure and input
Average gas tempera-
increasing energy per unitannual
Equivalent of consumed savingfuel.
natural gas 0.007905 Increases
under optimum design 25in either
conditions, cu m pressure
5,017,199 or 43,692
rsfrom thermodynamic tables and 30 43,678
rmining gas flow rate. Thermal ef-can increase the power produced by compressor
n- temperature 35 stations. Pressure is
44,653
40 45,617
of and energy wasted
ncy a desired increase in this context, temperature
in different
ECONOMIC INVESTIGATION STRAIGHT COOLER-DESIGN EFFECTS
RESULTS Table 5 Table 7
ons
el. of the compressor
is not. are part of
Optimum
calculation. Compressor Equival- Annual
ei- Scrubber power ent gas gas EQUATION Status quo
cooler
design
re pressure saved, saving, saving,
rational conditions
changeEquation 1 kw
defines desired cu flm/day
Input gas efficiency,
ow, MMscfd ancuindex
m Ef = wp \ Qf + E 100 100 (1)
re
ulating a station
Outlet gas pressure leaving origin station, psi 1,000 1,000
10 psi using
for HYSYS soft-
comparing
decrease Pressure12,615
different
1,682 drop in air coolers, %4,604,475
stations. where: 14.5 1.4
he
e and existing
5 psidata allowed study845
decrease
Gas pressure
Downstream
in downstream station inlet, psi
6,338
station compressor 2,313,188
power consumption, kw
Ef = desired effi 808
ciency 823.7
5 psi increase –853 –6,398 –2,335,088 = desired work,29,672
wp kw 26,867
or work done for pres-
ed
hanging operating conditions
10 psi increase on
–1,714
Downstream
–12,855
Equivalent
station compressor power consumption decrease,
savings using optimum design, cu m sure increase
annual natural gas–4,692,075
2,805
7,678,365
pressor station performance
or Calculating, in-required power involves extracting Qf = quantity of energy from the fuel
E = station’s electric power consumption
ing
s- the pressureenthalpy
drop in scrubber
of output gas pressure and input gas
air coolers (Fig. 1).
ed temperature
he following conditions governedfrom thermodynamic tables
the simulation: • Output andgas pressure, 1,000 psi.
s Flow
context, temperature
HAIRPIN
flowCOOLER-DESIGN
rate passing through the station, 80 MMscfd.
determining gas EFFECTS and energy wasted in different Table 6
rate. Thermal efficiency Table 4 shows input gas temperature’s effect on compres-
Input gas pressure, 700 psi. sor station power under constant pressure conditionsOptimum
for
defines desiredsections
efficien- of the compressor are part of this calculation. cooler
Status quo design
or comparing different
Operational
Gas Journal | Jan. 9, 2012 conditions
Input gas flow, MMscfd 100 105100
required power in- a station
Simulating Pressureusing
drop in air coolers, %
HYSYS software and existing data allowed 9.6 of
study 1.5
7
Gas pressure in downstream station inlet, psi 780 789.7
ng enthalpy ofchanging
output operating
Downstream station compressor
conditions power consumption,
on compressor kw performance , including
station 34,875 33,043
the
nd input gas tempera- Downstream station compressor power consumption decrease, kw 1,833
Equivalent annual natural gas saving under optimum design conditions, cu m 5,017,199
modynamic tables and
s flow rate. Thermal ef- Oil & Gas Journal :: EXECUTIVE BRIEF :: sponsored by
Iran’s gas system. The con- need to first define a num-
cluding article will focus on the effect of scrubbers and other
ber of consumption indices.
equipment in attempting to maximize capital return. Table 1 shows the energy consumed transmitting gas in
some countries.
Background COMPRESSOR OPTIMIZATION—1 Comparison, however, requires development of suitable
Iran is one of the largest consumers of natural gas in the indices. This article compares the power required for gas
world. Several transmission pipelines move natural gas transfer via pipeline by first calculating the power needed
SIMULATION SCHEMATIC FIG. 1
Pressure Gas compressor

Gas inlet drop = 58.95 kPa outlet Gas outlet To next station
From Scrubber out
previous Gas compressor Air cooler
station Scrubber power
Fuel Gas
compressor Stack
Combustion Hot
Pressure ratio set gas
Exhaust
Liquid Gas compressor

power
High-pressure Power = 2.456 Mw
z120109OGJtkh01
Inlet air
air
To air Turbine
Air compressor Power set
compressor
From recycle Recycle valve To recycle
pressure drop in scrubber and air coolers (Fig. 1).

104 Oil & Gas Journal | Jan. 9, 2012
The following conditions governed the simulation:

:: Flow rate passing through the station, 80 MMscfd.
:: Input gas pressure, 700 psi.
:: Output gas pressure, 1,000 psi.
Table 4 shows input gas temperature’s effect on compressor station power under
constant pressure conditions for each period. Decreasing input gas temperature
allows pressurization with smaller amounts of power. Each 1° C. decrease in input
gas temperature causes roughly a 0.5% decrease in required power.
A scrubber causes pressure drop of 5-6.5 psi when clean and 18.5-24.5 psi when
dirty. Compressor station simulation calculations show a 5 psi pressure drop
reduces power requirements by 850 kw. Each 1 psi drop decreases compressor
power requirements by 0.39%.
Compensating for the pressure drop caused by air coolers requires increasing
compressor station pressure. Holding both input gas pressure to the compressor
and the output gas pressure from the station constant, each 1 psi pressure drop
caused by air coolers increases power requirement by about 0.27%.
8
The classic defi- Length, 1,000 power, (km × 1,000 tempe
nition of effi- Pipeline km cu m/day Mw cu m/day) °C.
ciency of ther- Line 1 400 46 160 0.008696 10
Line 2 656 91.7 478 0.007943
modynamicCOMPRESSOR OPTIMIZATION—1
Line 3 656 90 440 0.007461
15
20
Average 0.007905
systems centers 25
30
on increasing en- 35
40
ergy per unit of
consumed fuel. ECONOMIC INVESTIGATION RESULTS Table 5
Sensitivity analysis Compressor Equival- Annual
Increases in ei- Scrubber power ent gas gas EQUAT
Sensitivity analysis used a
ther pressure pressure saved, saving, saving,
change kw cu m/day cu m Ef =
gas thermal valueorof 36,000
temperature
wher
10 psi decrease 1,682 12,615 4,604,475
kilojoule/cu m andcan gasincrease
turbinethe 5 psi decrease
5 psi increase
845
–853
6,338
–6,398
2,313,188
–2,335,088
Ef =
wp =
power produced 10 psi increase –1,714 –12,855 –4,692,075
efficiency of 32% (based
by compressor Qf =
E =
on the average of installed
stations. Pres-
sure is a desired
compressors) to investigate
increase in this context, temperature
solutions for decreasing compressor power requirements. HAIRPIN COOLER-DESIGN
Cooling input gas EFFECTS
by 1°
is not.
C. yielded roughly 55,000 cu 1m/year
Equation defines gas savings.
desired efficien-
cy, an index for comparing different
stations. Input gas flow, MMscfd
Researches also investigated pressure drop in the scrubber stemming
Outlet fromorigin station, psi
gas pressure leaving
Calculating required power in- Pressure drop in air coolers, %
nonobservance of design principlesenthalpy
volves extracting and, inofsome output cases, failure to replace
Gas pressure thestation inlet, psi
in downstream
Downstream station compressor power consumption, kw
scrubber filter in agas
timely manner.
pressure Tablegas
and input 5 shows
tempera- scrubberDownstream
pressure drop’s
station effects
compressor power consumption decrease, k
Equivalent annual natural gas saving under optimum design con
ture from thermodynamic tables and
on turbine performance.
determining gas flow rate. Thermal ef-
ficiency and energy wasted in different
This article only investigates STRAIGHT COOLER-DESIGN EFFECTS
sections of thepressure drops
compressor up to
are part of 25 psi, but field observations
show drops as large thisascalculation.
30-50 psi in scrubbers.
Operational conditions Input gas flow, MMscfd

Study included investigating reduced
Simulating a station pressure
using drop in air
HYSYS soft- coolers,
Outlet holding
gas pressure input
leaving origin
Pressure drop in air coolers, %
station, psi
Gas pressure in downstream station inlet, psi
gas pressure to the compressor
ware and
and existing output
data pressure
allowed study from the station
Downstream constant.
station compressor power consumption, kw
of changing operating conditions on Downstream station compressor power consumption decrease,
Calculations show each 1 psi pressure drop in air coolersEquivalent savingannual315,000 cusavings
natural gas m/ using optimum design, cu
compressor station performance , in-
year of natural gas.
cluding the pressure drop in scrubber
and air coolers (Fig. 1).
Energy loss The following conditions governed the simulation: • Output gas pressure, 1
Determining energy •balance
Flow rate
ispassing
the best through thecompare
way to station, 80design
MMscfd.conditionsTable and 4 shows input gas t
• Input gas pressure, 700 psi. sor station power under co
optimum operational conditions and to calculate energy loss.
After investigatingOilglobal energy

& Gas Journal consumption
| Jan. 9, 2012 norms in gas transfer projects
and determining the essential indices and parameters affecting operational
conditions, researchers sought to determine energy loss points and determine
energy savings potentials.
Investigating efficiency of selected compressor stations along the three pipelines

showed it to be lower than designed. Some stations with 10-25 Mw gas turbines
installed were producing <4 Mw, prompting a 3-year study to determine energy
9 loss by each station component. Investigating turbine efficiency showed that only

TRANSPORTATION
COMPRESSOR THERMAL EFFICIENCY FIG. 2
45
Minimum
40 Average
Maximum
35 Design
30
Thermal efficiency, %
25
20
15
10
z120109OGJtkh02
5
0
Zorya Nuovo Pignone Nevsky Siemens Ideal
Compressor type
each period. Decreasing input gas temperature allows pres- Sensitivity analysis
surization with smaller amounts of power. Each 1° C. de- Sensitivity analysis used a gas thermal value of 36,000 kilo-
Siemens
crease in inputturbines closely
gas temperature approximated
causes roughly a 0.5% de- to design
joule/cu m and operations.Other
gas turbine efficiency of 32%turbines
(based on thehad
crease in required power. average of installed compressors) to investigate solutions for
lower performance in most cases (Fig. 2).decreasing compressor power requirements. Cooling input
A scrubber causes pressure drop of 5-6.5 psi when clean
and 18.5-24.5 psi when dirty. Compressor station simula- gas by 1° C. yielded roughly 55,000 cu m/year gas savings.
tion calculations show a 5 psi pres- Researches also investigated pres-
Turbines lose
sure drop reduces most
power of their power through the exhaust
requirements suresystem.
drop in theInvestigation
scrubber stemming
by 850 kw. Each 1 psi drop decreases STATION COMPARISON Table 8
from nonobservance of design prin-
showed Siemens
compressor power turbines
requirements by losing less than others Power loss through this path.
ciples and, in some cases, failure to
Thermal through
0.39%. Station efficiency, chimney, replace the scrubber filter in a timely
type % kw
Compensating for the pressure manner. Table 5 shows scrubber pres-
Compressors
drop caused by air coolers requires Nevskey
Zorya
20.37
18.73
44,450
40,975 sure drop’s effects on turbine perfor-
increasing compressor station pres- Nuovo Pignone 21.13 49,210 mance.
Siemens 29.30 27,220
Investigating compressors
sure. Holding both input gas pressure revealed
Optimum designthe40 following characteristics: This article only investigates pres-
to the compressor and the output gas sure drops up to 25 psi, but field ob-
:: Compressor
pressure 260-13-1.
from the station constant, Installed in old Russian stations, real
servations showpower
drops asconsumption
large as 30-
each 1 psi pressure drop caused by air coolers increases 50 psi in scrubbers.
was higher than designed power consumption
power requirement by about 0.27%.
by 400-1,000 kw, growing as
Study included investigating reduced pressure drop in
rpm’s increased. air coolers, holding input gas pressure to the compressor
and output pressure from the station constant. Calculations
:: C
ompressor PCL 802-3. Operates with Nuovo Pignone turbo compressors.
COMPRESSOR COMPARISON SCRUBBER PERFORMANCE Table 10
Power consumption Polytrophic
wasConsumed400-1,000
Table 9
kw higher than designed and Out efficiency
of
–––––– efficiency, % –––– –––––– power, Mw –––––– allowable
5-10%Operational
Compressor
lower.
avg.
Design
avg.
Operational
avg.
Design
avg. Station
Clean,
0-10 psi, %
Dirty,
10-25 psi, %
range,
25-100 psi, %
Siemens 70.15 76.65 10.53 9.54 No. 1 48 37.3 14/7

:: BCL605
Compressor
68 BCL
74.9 605. Also
17.3 uses
16.3Nuovo Pignone
No. 2
No. 3
turbo compressors.
31.1
24.5
46.4
41.3
Power
22.5
34.2
PCL802-3 71.3 78.6 7.9 7.2
260-13-1 –– –– 4.5 3.9 No. 4 15.2 57.1 27.7
consumption is 200-1,300 kw higher than designed and the difference between
designed polytrophic efficiency and operational efficiency 5-9%.
106 Oil & Gas Journal | Jan. 9, 2012
:: S
imner-design compressors. Used 0.5-1.5 Mw more power than designed.
Operational efficiency in Sinner-design E compressors was 5-7% lower than
10 designed polytrophic efficiency in most instances.

Scrubbers
Scrubbers are responsible for nearly 10% of compressor stations’ energy loss.
More than 40% of scrubbers were dirty (resulting in a 10-25 psi pressure drop)
and nearly 10% were completely out of commission (resulting in >25 psi pressure
by each station per 1 million cu m gas/ EXAMPLE PIPELINES Table 2
drop). Pressure drop under optimum conditions is 5-10
km/day. psi.
Natural Power
Table 2 shows this power index for gas Compressor index,
transferred, station Maximum Mw/
a variety of countries. Length, 1,000 power, pressure, (km × 1,000
Air coolers
Table 3 shows the calculated power Country km cu m/day Mw bar cu m/day)
requirement for gas transfer of 1 mil- China 1,084 5 24.88 50 0.00459

Aspen-HTFS+ software-based simulation
The Netherlands compared
720 air
8.7 cooler function
32.1 with
69 0.00512
lion cu m gas/km/day on three Iranian Romania 100 51 25 63 0.00491
design
pipeline parameters.
networks. Iranian power re-
US
Algeria
1,200
520
14.3
335
92
822
77
72
0.00537
0.00472
Global average 0.004773
quirement are about 60% more than
the other coun-
Researchers used designed gas flow conditions to compare air cooler efficiency
tries.
with international standards.
IRANIAN NATURAL GASThe actual pressure drop
PIPELINES Table 3 of 9.6TEMPERATURE,
psi exceeded Table 4
Investigation Power REQUIRED POWER
expectations of 8.5 psi.
parameters Trans- index, Compressor
ferred gas, Nominal Mw/ Input gas power
The classic defi- Length, 1,000 power, (km × 1,000 temperature, requirement,
nition of effi- Pipeline km cu m/day Mw cu m/day) °C. kw
Tested
ciency of air
ther-coolersLinehad
1 a hairpin
400 arrangement,
46 160 with total output
0.008696 10 gas of the 39,651
Line 2 656 91.7 478 0.007943 15 40,679
mturbocompressor
odynamic flowing
Line 3 into a common
656 90 header and
440 0.007461 then being 20 divided 41,693
Average 0.007905
systems centers 25 43,692
30 43,678
onbetween
increasing en-air coolers by an input header before entering transformer 35 tubes. 44,653
40 45,617
ergy per unit
Other air coolersof have aINVESTIGATION
straight arrangement, with Table total output gas of each
consumed fuel. ECONOMIC RESULTS 5
turbocompressor
Increases in ei- directly entering
Scrubber
Compressor an air
power
cooler bank.
Equival-
ent gas
Annual
gas EQUATION
ther pressure pressure saved, saving, saving,
change kw cu m/day cu m Ef = wp \ Qf + E (1)
or temperature
where:
Pressure
can increase thedrops more rapidly in
10 psi decrease
5 psi decrease a straight
1,682
845 configuration.
12,615
6,338
4,604,475
2,313,188 Researchers
Ef = desired found
efficiency a
5 psi increase –853 –6,398 –2,335,088 wp = desired work, or work done for pres-
power produced
14.3 psi pressure10drop in straight-designed
psi increase –1,714 –12,855air coolers.–4,692,075 Using hairpin sureair coolers
increase
by compressor Qf = quantity of energy from the fuel
E = station’s electric power consumption
reducesPres-
stations. pressure drop, decreasing gas consumption. Table 6 shows the effects of
sure is a desired
a reduced pressure loss
increase in this context, temperature
HAIRPIN COOLER-DESIGN EFFECTS
isin
not.hairpin air coolers.
Table 6
Optimum
Equation 1 defines
Optimizing air desired
coolerefficien- cooler
Status quo design
cy, an index for comparing different
design in a straight
stations. Input gas flow, MMscfd 100 100
Calculating required
arrangement reduces power in- Pressure drop in air coolers, % 9.6 1.5
volves extracting enthalpy of output Downstream station compressor power consumption, kw 34,875 33,043
pressure
gas pressure and lossinput
even gasfurther
tempera- Downstream station compressor power consumption decrease, kw
Equivalent annual natural gas saving under optimum design conditions, cu m 5,017,199
1,833
ture from 7).

(Table thermodynamic tables and
determining gas flow rate. Thermal ef-
ficiency and energy wasted in different
STRAIGHT COOLER-DESIGN EFFECTS Table 7
Savings
sections of thepotential
compressor are part of
Optimum
this calculation. cooler
Identifying energy loss Status quo design
Operational conditions
points allowed definition
Input gas flow, MMscfd
Outlet gas pressure leaving origin station, psi
100
1,000
100
1,000
Simulating a station using HYSYS soft- Pressure drop in air coolers, % 14.5 1.4
of potential
ware and existing savings and
data allowed study
11
Downstream station compressor power consumption, kw 29,672 26,867
ofrecovery.
changing operating Downstream station compressor power consumption decrease, kw 2,805
Energy savingson
conditions Equivalent annual natural gas savings using optimum design, cu m 7,678,365
compressor station performance , in-
cluding the pressure drop in scrubber
and air coolers (Fig. 1).
Oil &governed
Gas Journal :: EXECUTIVE BRIEF :: sponsored by
The following conditions the simulation: • Output gas pressure, 1,000 psi.
20 crease in input gas temperature causes roughly a 0.5% de- joule/cu m and gas turbin
Thermal
crease in required power. average of installed compr
15
A scrubber COMPRESSOR
causes pressureOPTIMIZATION—1
drop of 5-6.5 psi when clean decreasing compressor po
10 and 18.5-24.5 psi when dirty. Compressor station simula- gas by 1° C. yielded rough
tion calculations show a 5 psi pres- R
z120109OGJtkh02
5 sure drop reduces power requirements sure
by 850 kw. Each 1 psi drop decreases STATION COMPARISON Table 8
from
0possibilities included reducing:
Power loss
compressor Nuovo
Zorya power requirementsNevsky
Pignone by Siemens
Thermal
Ideal
through ciple
:: Energy loss from scrubber pressure drop. Station efficiency, chimney,
0.39%. Compressor type repla
type % kw
:: Energy lossCompensating
from air coolerfor pressurethe pressure
drop. man
Nevskey 20.37 44,450
each period. drop
Decreasing inputcaused by air allows
gas temperature coolers requiresSensitivity analysis Zorya 18.73 40,975 sure
:: Electrical energy consumption inpres-
air coolers. Nuovo Pignone 21.13 49,210
surization with smallerincreasing
amounts of compressor
power. Each 1° station
C. de- pres- Sensitivity analysis used a gas thermal
Siemens 29.30 value of27,220
36,000 kilo- man
crease in input gas temperature causes roughly a 0.5%
sure. Holding both input gas pressure de- joule/cu m and gas turbine
Optimum design efficiency 40 of 32% (based on the Th
crease in required power.
to the compressor andreducing
the output gas average of installed compressors) to investigate solutions for sure
Potential
A scrubber causesenergy
pressurerecovery lay
drop of 5-6.5 psiinwhen clean decreasing compressor power requirements. Cooling input
energy
and 18.5-24.5 whenpressure
psileaving dirty. from the
compressor
Compressor station
stations
station constant,
through
simula- gas by 1° C. yielded roughly 55,000 cu m/year gas savings. serva
tion calculations show each5 1psipsi pressure drop caused by air coolers increases 50 psi in scrubbers.
the chimney.a Energy pres-
captured by reducing pressure drop canResearches be applied also
toinvestigated pres-
sure drop reduces powerpower requirement by about 0.27%.
requirements sure drop in the Study included
scrubber stemming investi
transmission
by 850 kw. Each 1 psi dropinstead.
decreases STATION COMPARISON Table 8
from nonobservance of design prin- input
Power loss
air coolers, holding
compressor power requirements by Thermal through ciples and, in some cases, failure to
0.39%. Station efficiency, chimney, replace the scrubberand output
filter in apressure
timely from
Table 8 shows
Compensating for thethe two parameters used in calculating energy
pressure
type % kw savings
manner. Table and
5 shows scrubber pres-
Nevskey 20.37 44,450
thermal
drop caused by airrecovery values.
coolers requires Zorya 18.73 40,975 sure drop’s effects on turbine perfor-
increasing compressorCOMPRESSOR
station pres- COMPARISON
Nuovo Pignone
Siemens
21.13
29.30
49,210
27,220
mance.
Table 9 SCRUBBER PERFORMANCE
sure. Holding both input gas pressure Optimum design
Polytrophic 40 Consumed This article only investigates pres-
Energy lost
to the compressor and through
the output a gasstation’s chimney
–––––– efficiency, % ––––reduces gaspower,
–––––– turbine
Mw sure thermal
–––––– drops upefficiency.
to 25 psi, but field ob-
Operational Design Operational Design Clean,
pressureStudying
from the the
station constant,
polytrophic
Compressor efficiency and
avg. avg. consumption avg.
servations
poweravg. show
of compressors drops as large as
also
Station 30-
0-10 psi, %
each 1 psi pressure drop caused by air coolers increases 50 psi in scrubbers.
showed potential
power requirement Siemens
by about to save70.15
0.27%. large volumes 76.65 of fuel gas
10.53
Study (Table 9).
included 9.54
investigating reducedNo. 1
pressure drop in48
BCL605 68 74.9 17.3 16.3 No. 2 31.1
PCL802-3 71.3 78.6 air coolers,
7.9 holding input
7.2 gas pressure to the
No. 3 compressor
24.5
260-13-1 –– –– and output 4.5pressure from No. 4 Calculations
3.9the station constant. 15.2
COMPRESSOR COMPARISON Table 9 SCRUBBER PERFORMANCE Table 10

Polytrophic Consumed Out of
106 %
–––––– efficiency, –––– –––––– power, Mw –––––– allowable
Operational Design Operational Design Clean, Dirty, range,
Compressor avg. avg. avg. avg. Station 0-10 psi, % 10-25 psi, % 25-100 psi, %
Siemens 70.15 76.65 10.53 9.54 No. 1 48 37.3 14/7

BCL605 68 74.9 17.3 16.3 No. 2 31.1 46.4 22.5
PCL802-3 71.3 78.6 7.9 7.2 No. 3 24.5 41.3 34.2
260-13-1 –– –– 4.5 3.9 No. 4 15.2 57.1 27.7
106 Table 10 shows scrubber performance. Oil & Gas Journal | Jan. 9, 2012
Energy consumption
This article initially outlined existing energy consumption norms in gas pipelines and
describED essential parameters and criteria for comparing operational conditions
with design conditions and optimum conditions. It then discussed the potential
energy savings found in each compressor station component. It will now turn to
available means for realizing these savings and an economic analysis of the results.
12
Gas turbines
Methods for energy savings and recovery in compressor stations’ gas turbines
include:
:: Using a recuperator.
Air entering a combustion chamber is preheated by hot air leaving the turbine
and entering a recuperator. This system, however, is suitable only for gas
turbines with low power and low pressure ratios. High-pressure turbines suffer
from low efficiency when using a recuperator since the pressure loss is greater.
Recuperators could only operate at those Iranian compressor stations using
Nevsky systems.
:: Cooling the input air.
:: Recovering energy lost through the chimney.
:: Replacing the gas turbine.
Inlet air TRANSPORTATION
Practical methods for cooling

show each 1 psi pressure drop in air
a compressor station’s inlet POWER REQUIREMENTS OF SIEMENS COMPRESSOR, KW Table 11
coolers saving 315,000 cu m/year of Underground Underground
airnatural
include:
gas. Cooling type Chiller
FOG
system
Underground
channel
channel,
FOG
channel,
chiller
:: Energy
irect loss
D evaporation system. Ambient, 35° C.
29° C.
20,097 20,097 20,097
21,252
20,097 20,097
27° C. 21,559
Determining energy balance is the best
:: way
Absorption, mechanical
to compare design conditions and
21° C.
15° C. 23,100
23,407
23,100
chillers.operational conditions and

optimum
Outlet power increase
Consumed power
3,003
1,200
1,462
17
1,155
100
3,310
117 TRANSPORTATION
3,003
820
Net outlet power 1,803 1,445 1,055 3,193 183
to calculate energy loss.
:: Underground
After investigating channels.
global energy
show each 1 psi pressure
consumption in airin gas POWER
dropnorms transferREQUIREMENTS OF SIEMENS KALIANCOMPRESSOR,
CYCLE EFFICIENCY, KW NEVSKY UNITS Table 1112
Table
coolers saving 315,000
projectscu andm/year of the essential indices and param-
determining Turbine outlet power, kw Underground 874.1
Underground
eters affecting operational conditions, researchers sought to Net system outletUnderground
FOG power, kw channel, channel, 853.5
natural gas. Table 11 summarizes theCooling effectstype of each on Siemens
Chiller compressors.
System
system efficiency, %
channel FOG chiller14.52
determine energy loss points and determine energy savings Gas turbine efficiency under design conditions, % 28
Gas turbine efficiency equipped with Kalian system, % 30
potentials. Ambient, 35° C. 20,097 20,097 20,097 20,097 20,097
Energy loss 29° C. compressor stations
Investigating efficiency of selected 21,252
Cooling
Determining energy balancetheis inlet
the air has a27°negligible
best C. effect on system 21,559 efficiency, and is suggested
along the three pipelines showed it to be lower than de-
21° C. 23,407
way to compare only
design for
signed. stations
conditions
Some facing
and
stations operational
with 10-25 15°Mw limitations
C. gas turbines
Outlet power increase
23,100 in
installed
3,003
Scrubbers
hot1,462
months of
1,155
the year. 3,310
23,100
3,003
optimum operational were conditions
producing <4and Mw, prompting a 3-year
Consumed study to deter-
power 1,200 Scrubbers 17 are responsible100 for nearly 10% of compressor
117 820 sta-
mine energy loss by each Netcomponent.
station outlet power Investigating
1,803 tions’1,445
energy loss. 1,055
More than 40% 3,193
of scrubbers 183
were dirty
to calculate energy loss.
In-chimney recovery
turbine efficiency showed that only Siemens turbines closely (resulting in a 10-25 psi pressure drop) and nearly 10% were
After investigating global energy
approximated to design operations.Other turbines had lower completely out of commission (resulting in >25 psi pressure
consumption normsEitherina Kalian cycle or organic
gas intransfer KALIAN CYCLE EFFICIENCY, NEVSKY UNITSconditions isTable 12 psi.
performance most cases (Fig. 2). drop). Pressure drop under optimum 5-10
projects and determining
Rankine thelose
cycle
Turbines essential
most ofindices
(ORC) chimney and through
their power param-the exhaust Turbine outlet power, kw 874.1
Net system outlet power, kw 853.5
eters affecting operational
unit conditions,
system. Investigation
can recover researchers
showed
energy fromsought
Siemens to losing less
turbines
the Air
System effi coolers
ciency, % 14.52
13
than points
determine energy loss others through this path. energy savings
and determine Aspen-HTFS+
Gas turbine efficiency under software-based simulation
design conditions, % compared air 28 cool-
chimney. Gas turbine
er efficiency with
function equipped with parameters.
design Kalian system, % 30
potentials.
Compressors
Investigating efficiency of selected compressor stations
Researchers used designed gas flow conditions to com-
Investigating compressors revealed the following character- pare air cooler efficiency with international standards. The
along the three pipelines showed it to be lower than:: de- :: sponsored
istics: Oil & Gas Journal EXECUTIVE BRIEF actual pressure drop of by9.6 psi exceeded expectations of 8.5
signed. Some stations• with 10-25260-13-1.
Compressor Mw gasInstalled
turbinesininstalled Scrubbers
old Russian stations, psi.
Net outlet
signed. Somepowerstations 1,803
with 10-25 1,445 Mw gas turbines 1,055 installed 3,193 Scrubbers 183
al energy were producing <4 Mw, prompting a 3-year Compressors

study to deter- Scrubbers are responsible for n
mine energy loss by each station component. Investigating
Investigating compressors revealed loss. the following
More thanch
s transfer KALIAN
COMPRESSORCYCLE EFFICIENCY,
OPTIMIZATION—1 NEVSKY UNITS tions’ energy Table 12
turbine
e essential indices andefficiency
param- showed that only Siemensistics:
Turbine outlet power, kw
turbines closely (resulting874.1 in a 10-25 psi pressu
approximated
onditions, researchers sought to to design operations.Other
Net system outlet power,turbines
kw • Compressor
had lower 260-13-1. Installed
completely out in
853.5 of old Russian
commission
System efficiency, % 14.52
s and determineperformance
energy savings in most cases Gas (Fig.
turbine2).
efficiency underreal power
design consumption
conditions, % drop). was higher
Pressure28 thanunder
drop designed
op
:: Kalian
Turbinescyclelose
is useful Gas
most when turbine
of their effi
thepower ciency
exhaust-gas equipped
through with
consumption Kalian
temperature
the exhaust system,
byfrom %
400-1,000 30
kw, growing as rpm’s in
the chimney
system. Investigation
is low, limiting
of selected compressor stationspossible showed
energySiemens
recycling. turbines• Compressor
In Iran losing less PCL Air
such conditions 802-3.coolers
occur Operates
only with Nuovo
howed it to bethan when
lower others
thanthrough
using Nevsky
de- this path.
gas turbines. Since these turbo compressors.
turbines have very Power low consumption
Aspen-HTFS+
efficiency, was 400-1,
software-based
however,
10-25 Mw gas turbines they are not
installed good candidates forhigher
Scrubbers energythan designed
recycling and
er
equipment efficiency
function andwith 5-10%
design lower.
param
mpting a 3-yearCompressors
study instead
should Scrubbers
to deter-be replaced are responsible for
by higher-efficiency
• Compressor BCL 605.
nearly 10% of compressor
turbines.
Also
Researchers usessta- Nuovo Pignon
used designed
ation component. Investigating
Investigating compressors
tions’ revealed
energy loss. compressors.
the following
More thancharacter- Power
40% of scrubbers consumption
pare air werecooler is
dirty 200-1,300with
efficiency kw
istics: than designed and theactual difference between drop designed
at only SiemensTableturbines closely Kalian
12 summarizes (resulting in aeffect
cycle’s 10-25on psiNevsky
pressure drop)
units nearlypressure
and design
under 10% were of 9.6 psi
ations.Other turbines • Compressor
hadThe lower 260-13-1. Installed
completely out in old phic
Russian
of commission
efficiency
stations,
(resulting
and operational
inispsi.
>25 psi
efficiency 5-9%.
pressure
conditions. authors
real power consumption do not recommend their use. ORC
• Simner-design cycle the
compressors.best Used 0.5-1.5hadMwa mo
Fig. 2). drop).was higherdrop
Pressure thanunderdesignedoptimumpowerconditions Tested
is 5-10 air psi.
coolers ha
energy recovery
consumption solution. Fluid selection
by 400-1,000 kw, growing aserrpm’s is based on gas
thanincreased. turbine conditions,
designed. Operational
output gasefficiency in Sinner-d
of the turbocompre
eir power through the exhaust
local •conditions,
Compressor and the802-3.
technology used bycompressors
the manufacturer.was 5-7% But lower
average thanORC designed
beingpolytro
d Siemens turbines losing less PCLAir Operates
coolers with Nuovo Pignone header and then divide
cycle efficiency at
turbo compressors. Powercompressor station
consumption chimney ficiency in
temperature
was 400-1,000 most instances.
is about
kw compared 10%.
input header The before entering
h. Aspen-HTFS+ software-based simulation air cool-
lowhigher than designed
temperature erand
in Nevsky efficiencywith5-10%
gas turbine
function designlower.
parameters. coolers have a straight arrange
• Compressor
chimneys, however,BCL 605.
precludes Alsouseuses
Researchers of the Nuovo
used Pignonegas
designed turbo each turbocompressor
flow conditions to com- directly
compressors. Power pare
consumption TURBINE REPLACEMENT, Table 13 TURBINE
ORC
evealed the following cycle on efficiency
character- air coolerisefficiency
grounds. 200-1,300
40% with kwinternational
higher
THERMAL EFFICIENCY
Pressure drops
standards. The more rapidl
WITH SIE
than designed and theactual difference
pressurebetweendrop designed
of 9.6 psipolytro- Researchersoffound
exceeded expectations 8.5 a 14.3 psi
Gas savings,
phic efficiency
nstalled in oldCalculations
Russian stations, and operational
psi.
show recoverable
efficiency 5-9%.
power in
signed air coolers.
million cu Using hairpi
• Simner-design compressors. air Used 0.5-1.5 Station
Mw more pow- m/year
drop, decreasing Station
as higher than designed
compressor stations usingTested
power ORC cycles coolers had a hairpin arrangement, with total gas consump
kw, growing as er thanincreased.
designed. Operational gasefficiency in Sinner-design
No. 1 E ofinto
a reduced 95 pressure loss No.in1 h
of rpm’s
about 120 kw/hr. Since output
the recoveredof the turbocompressor
No. 2 flowing a common
76 No. 2
Operates with compressors
Nuovo was 5-7%
Pignone lowerand
header thanthen designed
being polytrophic
No. 3
divided ef-
between air coolers
air cooler93design
by an in a straightNo. 3 a
energy in the ORC cycle is electrical No. 4 48 No. 4
onsumption was ficiency
400-1,000 in mostkwinstances.
input header before entering loss even100
No. 5 transformer tubes. Other
further air (Table 7).
No. 5
and electric energy use in compression
ficiency 5-10% lower. coolers have a straight arrangement, with total output gas of
stations is low, the saved electric
lso uses Nuovo Pignone turbo each turbocompressor directly entering an air cooler bank. INITIAL C
TURBINE
energy REPLACEMENT,
is best used to start another Table 13gas TURBINE REPLACEMENT, Table 14
mption is 200-1,300 kw
40% THERMAL higher Pressure
EFFICIENCY drops more rapidly
WITH in
SIEMENS a straight configuration. DESIGN
compressor by electric motor. Oil & Gas Journal | Jan. 9, 2012
ence between designed polytro- Researchers found
Gas savings,
a 14.3 psi pressure drop in Gasstraight-de-
savings, Gas outl
nal efficiency 5-9%. signedmillion
air coolers.
cu Using hairpin air coolers reduces pressure
million cu Pressure
Turbine
Stationreplacement m/year Station m/year by air c
rs. Used 0.5-1.5 Mw more pow- drop, decreasing gas consumption. Table 6 shows the effects Cost, $
Gas No. 1
turbines normally 95 No. 1 46
al efficiency in Sinner-design
No. 2 E ofoperate
a reduced at
76 pressure loss No.in2 hairpin air coolers.Optimizing31
lower
r than designed efficiency
polytrophic
No. 3 than their
ef- design
air cooler design in a straight
93 No. 3 arrangement reduces pressure 47
No. 4 48 No. 4 3
condition.
No. 5 Reconstructing loss even further (Table 7).
compressor
100 No. 5 50
stations and replacing gas turbines can
increase thermal efficiency. This article
Table 13 TURBINE REPLACEMENT, Table 14 INITIAL COOLER Table 15
considers two turbine replacement
WITH SIEMENS
Oil & Gas Journal | Jan. 9, 2012 DESIGN PARAMETERS
scenarios. Table 13 shows energy Gas outlet temp., °C. 15
s savings, Gas savings,
million cu savings stemming from replacing million cu Pressure drop caused
m/year Station m/year by air cooler, psi 9.5
existing turbines with units having 40% Cost, $ 9,183,981
95 No. 1 46
76 better
No.thermal
2 efficiency. In the 31 second
14
93
48
No. 3
scenario,
No. 4 Siemens turbines replace
47
3 the
100 No. 5 50

TRANSPORTATION
existing units, resulting

ANNUAL SAVINGS, AMBIENT TEMP. BELOW GAS FEED Table 16
Iranian com
in the energy savings • Cooli
Compressor power requirement drop, kw 6,250
shown in Table 14. Station fuel consumption drop, million cu m 2.44 • Recov
• Repla
Feed cooling
SAVINGS, COST COMPARISON Table 17 Inlet air
Cooling a compressor Fuel Practical m
consumption
station’s gas feed can decrease include:
potential,
decrease its power million Overall Annual • Direc
Station cu m/year cost, $ cost, $
requirement. Either a • Abso
No. 1 1.1 5,013,213 251,000 • Unde
chiller or an air cooler No. 2 4.3 8,059,140 536,775
No. 3 5.9 8,693,586 816,680 Table 11
can cool the gas feed. No. 4 3.7 11,018,083 616,260
No. 5 3.3 8,254,959 397,550 pressors.
Preliminary calculations Cooling
show a chiller uses too ficiency, an
limitations
much energy to be economical. Table 15 shows general specifications of an air
Savings potential
cooler used in cooling a compressor station’s
Identifying energygaslossfeed.
points allowed definition of potential In-chimn
savings and recovery. Energy savings possibilities included Either a Ka
Table 16 shows the reduced power reducing:
needs and fuel requirement realized by using ney unit ca
• Energy loss from scrubber pressure drop. • Kalia
an air cooler for a year, given that the ambient temperature is cooler than the gas
• Energy loss from air cooler pressure drop. ture from t
feed. Table 17 shows the cost savings potential
• Electrical achieved
energy by using
consumption in airair coolers at
coolers. cling. In Ir
five other stations in differing climates.
Potential energy recovery lay in reducing energy leaving gas turbin
compressor stations through the chimney. Energy captured however, t
by reducing pressure drop can be applied to transmission equipment
instead. ciency turb
Shaghayegh Khalaji Table 8isshows
a senior expert
the two of energy
parameters used management
in calculating en- Table 1
at National Iranian Gas Co. She earned an MS in
ergy savings and thermal recovery values. chemical units unde
EnergyTehran
engineering (2002) from lost through a station’s chimney reduces gas tur-
University. mend thei
bine thermal efficiency. Studying the polytrophic efficiency lution. Flu
and consumption power of compressors also showed poten- local cond
tial to save large volumes of fuel gas (Table 9). turer. But a
Table 10 shows scrubber performance. chimney te
Nevsky ga
Energy consumption ORC cycle
This article initially outlined existing energy consumption Calcula
norms in gas pipelines and describED essential parameters tions using
and criteria for comparing operational conditions with de- ered energ
sign conditions and optimum conditions. It then discussed use in com
the potential energy savings found in each compressor sta-
tion component. It will now turn to available means for real-
15 izing these savings and an economic analysis of the results.
Gas turbines
Methods
Oil & Gas Journal for energyBRIEF
:: EXECUTIVE savings::and recoverybyin compressor sta-
sponsored
tions’ gas turbines include:
Originally published March 5, 2012
(Conclusion): Energy
recovery, scrubbers offer
keys to efficient operation
by Shaghayegh Khalaji
E
nergy recovery and efficient scrubber technologies can help
minimize energy loss on natural gas transmission systems.
Iran’s natural gas pipeline system requires roughly 60% more power/
unit shipped than the global average. Iranian gas does not have to
travel particularly long distances, but its compressor units are of relatively low
efficiency and the booster compressor stations somewhat underpowered. The
system’s design also occurred when domestic gas prices in Iran were very low,
reducing the emphasis placed on efficiency.
The first article of this series (OGJ, Jan. 9, 2012, p. 104) outlined existing energy
consumption norms in gas pipelines, describing essential parameters and criteria
for comparing operational conditions with design conditions and optimum
conditions. This concluding article will focus on the effect of scrubbers and other
equipment in attempting to maximize capital return.
Gas turbines normally work at lower efficiency than their design condition.
Reconstructing compressor stations and replacing gas turbines can increase
thermal efficiency. Lost heat can be recovered to generate electricity via a
Rankine power cycle. Turboexpanders can use reduction in gas pressure to
generate power.
Cooling a compressor station’s gas feed, meanwhile, can decrease its power
16 requirement. Either a chiller or an air cooler can cool the gas feed. Inlet air

cooling can prevent loss of output when ambient temperatures are high.
Preliminary calculations, however, showed a chiller uses too much energy to
be economical.
After investigation determined energy loss levels, potentials for mitigating
ION— 2 this
( Closs
o or
n crecovering
l u s i othe
n lost
) energy were defined. Energy saving possibilities
included reducing:
:: Energy loss from scrubber pressure drop.
scrubbers :: Energy loss from air cooler pressure drop.
:: Electrical energy consumption in air coolers.
ent operation Scrubbers

Scrubbers help separate excess liquids or undesirable particles from the gas
more power/unit shipped than the global average. Iranian
stream as it passes through a compressor station, keeping the stream properly
gas does not have to travel particularly long distances, but
dehydrated.
its compressor units are of relatively low efficiency and the
booster compressor
Scrubber performance stations
is best improved somewhat underpowered. The
through:
ologies can ::help system’s
Redesign, design also occurred when domestic gas prices in
replacement.
ission systems. Iran were very low, reducing the emphasis placed on effi-
:: Timely replacement of filters.
es roughly 60% ciency.
The first article of this series (OGJ,
As mentioned in the first article of Jan. 9, 2012,
this series, p. 104)
scrubber outlined
pressure drop canexist-
sometimes exceed 25 psi. This scaleing of pressure drop requires either
energy consumption normsscrubber
in gas
replacement or use of a new scrubber in parallel with the existing one.
pipelines, describing essential param-
t
(B t - Ft - V t)
/ - I 0 = 0 show gas (3)
Calculations
(1 + rt) t
savings when scrubber performance was optimized of 6.1
t= 0
million cu m/year.
where:
t = CapitalTable
return time five test stations’
SCRUBBER Table 1
1 shows
OPTIMIZATION SAVINGS
rt = Capitalinterest
gas savings rate duringvia
achieved t
Gas savings,
scrubber
B t = Profits optimization.
resulting from the Savings million
are based on the unit’s average Station cu m/year
project during t at fixed
thermal efficiency and 1,388 No. 1 3.3
prices
working hr/year (Mar. 21, 2009-
No. 2 3.4
No. 3 4.5
Ft = Fixed expenses during t at
Mar. 20, 2010). No. 4 2.4
17 fixed prices
No. 5 6.8
V t = Annual variable expenses

during t at fixed prices GAS SAVINGS, OPTIMIZED Table 2
re very low, reducing theproject
ed
during
emphasis t at fixed
placed on effi- No. 1 3.3
prices No. 2 3.4
No. 3 4.5
TheFfirst article of thisCOMPRESSOR
series (OGJ, OPTIMIZATION—2
t = Fixed expenses during t at No. 4 2.4
equire - Jan. 9, 2012,fixed
p. prices
104) outlined exist- No. 5 6.8
ing energy consumption norms in gas

V t = Annual variable expenses
pipelines, describing essential param-
(3) during t at fixed prices GAS SAVINGS, OPTIMIZED Table 2
(2) Air coolers AIR-COOLER PERFORMANCE
I 0 = Initial expenses and capi -
Both redesign
SCRUBBER taland replacement and a simple bypass of air coolers can reduce Potential
requirement Table 1 savings,
ate of cap - gas consumption.
OPTIMIZATION Table 2 shows the volumes of gas saved optimizing air million
SAVINGS cooler
ng t Station cu m/year
performance
/ (B(1t - +Funder
t - V t) average Gas working conditions.
t
savings,
he - I 0 = NPVmillion
t (4) No. 1 3.72
rt) t
ng from t= 0
Station cu m/year No. 2 0.89
ed where: No. 3
When ambient conditions are cool enough, the temperature of gas leaving the
1.68
ng t at No. 1 3.3 No. 4 1.63
No. NPV
2
compressor t = Net present value3.4
may be low enough of the No. 5
to enter the transfer 0.7
No. 3 4.5 No. 6 line without first being
2.02
at investment
No. 4 eliminating
cooled, during
air cooler2.4 t
pressure No.37 shows gas savings associated
drop. Table 2.26
ses during No. 5 6.8 No. 8 1.67
with bypassing rate during t
the air cooler.
es B t = Profits resulting from the
penses
s GAS SAVINGS,project
OPTIMIZED
during t at Table 2 GAS SAVINGS, Table 3
prices
i- AIR-COOLER PERFORMANCE
fixed prices AIR-COOLER BYPASS
d cap - Potential Gas savings,
savings, million
million Station cu m/year
Station
fixed prices cu m/year
V t = Annual variable expenses No. 1 1.07
(4) No. 1 3.72 No. 2 0.31
No. 2 during t at fixed 0.89
prices No. 3 0.32
No. 3 1.68 No. 4 0.42
No. 4I 0 = Initial expenses and1.63
capi - No. 5 0.22
the No. 5 0.7 No. 6 0.59
No. 6 tal requirement 2.02 No. 7 0.59
t No. 7 2.26 No. 8 0.58
No. 8 1.67
ng t
he
GAS SAVINGS, Table 3 Oil & Gas Journal | Mar. 5, 2012
Turboexpanders
AIR-COOLER BYPASScan also reduce the amount of gas required to do the same
quantity of work. Tables 4,Gas
5,savings,
and 6 show the savings realized by this approach,
at million
respectively
Station showing power cu use
m/year when outlet-gas temperature is the same as
ambient
No. 1 temperature, when 1.07outlet-gas temperature is cooled as much as possible,
es No. 2 0.31
andNo.
when
3
inlet gas is preheated.
0.32
es No. 4 0.42
i- No. 5 0.22
Economic
No. 6 analysis 0.59
No. 7 0.59
Energy
No. 8 optimization requires0.58
advantages
and costs to be reported and assessed in
financial terms. The indices such assessments use are:
:: Cash flow (Equation 1).
:: I nternal Oil
rate& of
Gasreturn
Journal | Mar. 5, 2012
(Equation 2).
:: Capital return time (Equation 3).
18 :: Net present value of investment (Equation 4).

TRANSPORTATION
TRANSPORTATION
STATION PERFORMANCE, OUTLET GAS TEMPERATURE SAME AS AMBIENT Table 4

Station capacity, cu m/hr <10,000 10,000-50,000 50,000-100,000 100,000-200,000 200,000-500,000 500,000-1 million
Average throughput, cu m/hr
Average power production, Mw
9,000
0.20
37,000
0.81
C O M P 72,380
R 1.59
ES S OR OP T3.07
IM
139,730 I Z A T I371,000
O 8.16
N— 2 ( C o866,000
n19.04
clu sion)
Heat demand, Mw 0.249 1.02 2.00 3.86 10.25 23.92
STATION PERFORMANCE, MAXIMUM OUTLET GAS COOLING Energy recovery, scrubbers Table 5
offer keys to efficient operation

Average throughput, cu m/hr 9,000 37,000 72,380 139,730 371,000 866,000
Average power production, Mw 0.18 0.76 1.48 2.86 7.6 17.74
Heat demand, Mw 0.14 0.57 1.11 2.14 5.67 13.24
Shaghayegh Khalaji more power/un

National Iranian Gas Co. gas does not ha
STATION PERFORMANCE, PREHEATED INLET GAS Table 6
Tehran its compressor u
Average throughput, cu m/hr 9,000 37,000 72,380 139,730 371,000 booster compre
866,000
Heat demand, Mw 0.43 1.75 Energy recovery
3.43 and efficient scrubber
6.63 technologies
17.60 can help system’s design
41.07
minimize energy loss on natural gas transmission systems. Iran were very
Iran’s natural gas pipeline system requires roughly 60% ciency.
SOLUTION CHARACTERISTICS BY STATION Table 7
Assumptions
Heat
recovery EQUATIONS
The external parameters used
system,
t = /
Organic
/ (B(1- +F r-)Scrubber
V)
t t
Rankine Renovated CFNatural (B t - Ft - V t) - I 0 (1) t t
-I =0
t
(3)
in calculating
Cycle project economics
–––––––– gas turbine –––––––
t =0
––––––– gas cooling –––––––– –––––––– Air cooling ––––––––– ––––––––
t
–––––––
0
Million cuwhere:
t= 0 t
Station Gw-hr/yr Million cu m/yr $, thousand m/yr $, thousand Million cu m/yr $, thousand Million cu m/yr $, thousand
where:
include:
No. 1 119 49 25,000 ––CF1 = Cash flow–– 3.72 3,710 2.7 700
No. 2 25 –– –– 1.07 –– t = Capital
0.6 return time
1B t = Profits during time =0.89
t––at 3,200 350
:: E
No.
xpenses
3
No. 4
and incomes
––
64
10
26
set at
25,000
25,000 1.65
1,834
3,079 1.68
––
2,300
1.4
1.2
700
rate
350 during t
No. 5 40 –– –– 1.63
fixed prices
3,146 1.63 2,877 0.8 350
2010-11
No. 6 levels.
–– 10 25,000 1.46V t = Annual1,834 –– - –– B t = Profits
1.7 resulting700
from the
No. 7 70 42 25,000 2.92
variable costs dur
3,560 –– –– 1.9 700
No. 8 35 –– ––- 1.54 3,299 0.68 2,050 project
0.9 during 350at fixed
t
ing time = t at fixed
:: R
eal interest rate for net value
No. 9 72 –– –– 2.44 9,184 2.02 2,966 1.1
prices 350
No. 10 –– 12 25,000 0.75 prices
1,780 –– –– 2.8 700
No. 11 99 41 25,000 1.69 3,329 2.26 2,410 3.2 1,050
calculation of 5%.
No. 12 26 –– –– 0.9I 0 = Basic 3,146
and capital require
1.67 - 2,266 0.9 350
fixed prices
ments
:: O il price in international V t = Annual variable expenses
eters and criteria
offor comparingyielding
operationalaconditions with - F Scrubbers during t at fixed prices
markets $60/bbl, / (B -V )
t
t t T
-I = 2 (2) 0
design conditions and optimum conditions. This concluding Scrubbers
(1 + IRR )
t =0
help separate
t
t
excess liquids
I = or undesirable
Initial expenses andpar-
capi - 0
natural
article gasonprice
will focus of of$0.25/cu
the effect scrubbers and other equip-
where: ticles from the gas stream as it passes through a compressor
tal requirement
ment
m inand
attempting
a power to maximize
supplycapital
pricereturn.
of station,
IRR = Internal
t
keeping
return thecap
rate of stream
- properly dehydrated.
Gas turbines normally work at lower efficiency than their ital during Scrubbert performance is best improved through:
/ (B(1- +F r-)V ) - I = NPV
t
t t t
$0.09/kw-hr. (4) 0 t
design condition. Reconstructing compressor stations and • Redesign, replacement. t
B t = Advantages resulting from t= 0 t
replacing gas turbines can increase thermal efficiency. Lost • Timely replacement of filters.where:
::heat
Iranian thermal
can be recovered to generate plant via a Rankine the project
power electricity during t at
As mentioned in the first article
NPV oft =this
Netseries,
presentscrubber
value of the
power cycle. Turboexpanders can use reduction in gas pres- fixed prices
pressure drop can sometimes exceed 25 psi. This scale of t
efficiency of 35%. F1 = Annual fixed expenses
investment
during either scrubber replacement or use
during
sure to generate power. pressure drop requires
rt = Capitalinterest rate during t
:: RCooling
ankine a compressor station’s gas feed, meanwhile, can t at
cycle technical of fixed
a newprices
scrubber in parallel with the existing one. Calcu-
B t = Profits resulting from the
decrease its power requirement. Either a chiller or anVtair = Annuallations show
variable gas savings when scrubber
expenses performance was
data
cooler canyielding a design-cycle
cool the gas feed. Inlet air cooling can prevent during optimized of 6.1 million cu m/year. project during t at
t at fixed prices
lossefficiency
of output when fixed prices
of ambient
17.4% temperatures are high. Prelimi- Table 1 and
I 0 = Initial expenses shows
capfive
- test stations’ gas savings achieved via
nary calculations, however, showed a chiller uses too much scrubber optimization. Savings are Ft based
= Fixedon expenses during
the unit’s aver-t at
ital requirement
energy to be economical. age thermal efficiency and 1,388 workingfixed hr/year (Mar.
prices 21,
After investigation determined energy loss levels, poten- 2009-Mar. 20, 2010). V t = Annual variable expenses
Cost
tials foranalysis
mitigating this loss or recovering the lost energy during t at fixed prices
were defined. Energy saving possibilities included reducing: Air coolers I 0 = Initial expenses and capi -
Five basic energy optimization
• Energy loss from scrubber pressure drop. Both redesign and replacement and a simple bypass of air
19 • Energy emerged
solutions loss from air through
cooler pressure
flowdrop.
• Electrical energy consumption in air coolers.
tal requirement
coolers can reduce gas consumption. Table 2 shows the vol-
umes of gas saved optimizing air cooler performance under
Oil & Gas Journal :: EXECUTIVE

118 BRIEF :: sponsored by
Oil & Gas Journal | Mar. 5, 2012 119
process study at the sample compressor stations:

:: Heat recovery from chimney gases and electric power production.
:: Gas turbine reconstruction.
:: Cooling natural gas feed.
:: Cooling natural gas leaving compressor. TRANSPORTATION
:: Scrubber optimization.
STATION PERFORMANCE, OUTLET GAS TEMPERATURE SAME AS AMBIENT Table 4
Economic investigation exposed the characteristics of each feasible solution
Heat demand, Mw 0.249 1.02 2.00 3.86 10.25 23.92
(Table 7).
STATION PERFORMANCE, MAXIMUM OUTLET GAS COOLING Table 5
Table 8 shows the results of economic analysis of each solution using the
Average power production, Mw
Heat demand, Mw
0.18
0.14
0.76
0.57
1.48
1.11
2.86
2.14
7.6
5.67
17.74
13.24
assumptions presented earlier. Energy recovery in chimneys and its use for
electric power production
STATION PERFORMANCE, andGAS
PREHEATED INLET scrubber optimization are economically feasible inTable 6
all Average
stations,
throughput,and
cu m/hrthe return
9,000 rate37,000
of the project72,380during the exploitation
139,730 term (less866,000
371,000
than 8 years) is more 0.43
Heat demand, Mw
than 31%. 1.75 3.43 6.63 17.60 41.07
SOLUTION CHARACTERISTICS BY STATION Table 7

Heat
recovery
system,
Organic
Rankine Renovated Natural
Cycle –––––––– gas turbine ––––––– ––––––– gas cooling –––––––– –––––––– Air cooling ––––––––– –––––––– Scrubber –––––––
Station Gw-hr/yr Million cu m/yr $, thousand Million cu m/yr $, thousand Million cu m/yr $, thousand Million cu m/yr $, thousand
No. 1 119 49 25,000 –– –– 3.72 3,710 2.7 700

No. 2 25 –– –– 1.07 –– 0.89 3,200 0.6 350
No. 3 –– 10 25,000 1 1,834 –– –– 1.4 700
No. 4 64 26 25,000 1.65 3,079 1.68 2,300 1.2 350
No. 5 40 –– –– 1.63 3,146 1.63 2,877 0.8 350
No. 6 –– 10 25,000 1.46 1,834 –– –– 1.7 700
No. 7 70 42 25,000 2.92 3,560 –– –– 1.9 700
No. 8 35 –– ––- 1.54 3,299 0.68 2,050 0.9 350
No. 9 72 –– –– 2.44 9,184 2.02 2,966 1.1 350
No. 10 –– 12 25,000 0.75 1,780 –– –– 2.8 700
No. 11 99 41 25,000 1.69 3,329 2.26 2,410 3.2 1,050
No. 12 26 –– –– 0.9 3,146 1.67 2,266 0.9 350
TRANSPORTATION
eters and criteria for comparing operational conditions with

SOLUTION RESULTS BY STATION
Scrubbers Table 8
design conditions and optimum conditions. This concluding Scrubbers help separate excess liquids or undesirable par-
–––––––––– Heat recovery system, ORC –––––––––– Gas turbine
article will focus on the effect ofInternalscrubbers and Net other equip-
replacement ticles from the gas stream as it passes through a compressor
rate of present –––– GT 10 B –––– – Gas cooling – –– Air cooling –– –– Scrubbers ––
ment in attempting
Capital, to maximizereturn,capital return.
value, IRR, station, keeping
NPV, IRR, the
NPV,stream properly
IRR, dehydrated.
NPV, IRR, NPV,
Station
Gas turbines $1,000normallyMwwork at % $
lower efficiency than%their $ Scrubber % performance
$ is% best improved
$ %
through: $
design
No. 1 condition.
25,040 Reconstructing
17.0 29compressor
33,871 stations 41 and 44,095 • Redesign, –– replacement.
–– 14 1,523 91 3,678
No. 2 5,261 3.6 29 7,116 –– –– –– –– –– –– 36 566
replacing
No. 3 gas turbines
–– can
–– increase –– thermal efficiency.
–– ––Lost –– • Timely –– replacement
–– of––filters. –– 43 1,472
No. 4 13,467 9.1 29 18,216 13 8,610 –– –– –– –– 36 1,132
heat
No.can
5 be 8,417
recovered to 5.7generate 29electricity
11,385via a Rankine
–– –– As mentioned
–– in
–– the first
–– article of –– this series,
51 scrubber
906
power
No. 6 cycle. Turboexpanders
–– –– can––use reduction –– in gas––pres- pressure
–– drop
6 can110sometimes–– exceed–– 25 psi. 55 This 1,981
scale of
No. 7 14,730 10.0 29 19,924 33 33,295 7 359 –– –– 62 2,320
sure
No.to8 generate
7,365power. 5.0 29 9,962 –– pressure
–– drop requires
–– –– either
–– scrubber –– replacement
59 or use
1,075
No. 9 15,150 10.3 29 20,493 –– –– –– –– –– –– 31 963
Cooling
No. 10 a compressor
–– ––station’s––gas feed, meanwhile,
–– –– can of––a new –– scrubber –– in parallel
–– with the –– existing95 one.3,847
Calcu-
20
No. 11 its20,832
decrease 14.1
power requirement. 29Either a28,178
chiller or an32 air 31,752 lations ––
show ––
gas savings 12
when 724
scrubber 109
performance 4,526was
No. 12 5,471 3.7 29 7,400 –– –– –– –– –– –– 59 1,075
cooler can cool the gas feed. Inlet air cooling can prevent optimized of 6.1 million cu m/year.
loss of output when ambient temperatures are high. Prelimi- Table 1 shows five test stations’ gas savings achieved via
nary calculations, however, showed a chiller uses too much scrubber optimization.
• Internal Savings
rate of return are based
(Equation 2). on the unit’s aver-
EXPANSION
energy to be TURBINE DATA
Oil & Gas Journal :: EXECUTIVE •BRIEF
economical. Table 9 age Capital ::
thermal efficiency and
sponsored
return 1,388
by 3).
time (Equation working hr/year (Mar. 21,
After
Gas flowinvestigation
<10,000 cu m/hr determined energy loss levels, poten- 2009-Mar.
• Net 20, 2010).
present value of investment (Equation 4).
Power capacity Mw 0.2
Station $1,000 Mw % $ % $ % $ % $ % $
No. 1 25,040 17.0 29 33,871 41 44,095 –– –– 14 1,523 91 3,678

No. 2 5,261 3.6 29 7,116 –– –– –– –– –– –– 36 566
No. 3 –– –– –– –– –– –– –– –– –– –– 43 1,472
No. 4 13,467 9.1 29 18,216 13 8,610 –– –– –– –– 36 1,132
No. 5 8,417 5.7 COMPRESSOR OPTIMIZATION—2
29 11,385 –– –– –– –– –– –– 51 906
No. 6 –– –– –– –– –– –– 6 110 –– –– 55 1,981
No. 7 14,730 10.0 29 19,924 33 33,295 7 359 –– –– 62 2,320
No. 8 7,365 5.0 29 9,962 –– –– –– –– –– –– 59 1,075
No. 9 15,150 10.3 29 20,493 –– –– –– –– –– –– 31 963
No. 10 –– –– –– –– –– –– –– –– –– –– 95 3,847
No. 11 20,832 14.1 29 28,178 32 31,752 –– –– 12 724 109 4,526
No. 12 5,471 3.7 29 7,400 –– –– –– –– –– –– 59 1,075
Scrubber optimization is the best approach, significantly lowering costs and
TRANSPORTATION
energy consumption in natural gas transfer lines. Cooling
• Internal ratenatural gas before
of return (Equation 2). and
EXPANSION TURBINE DATA Table 9
after compressors is justifiable in some stations,• and
Gas flow <10,000 cu m/hr
Capital
gasreturn time (Equation
turbine 3).
reconstruction
• Net present value of investment (Equation 4).
SOLUTION
and demand RESULTS BY
Power capacity
Heatreplacement STATION Mw in older,0.25
Mw
is justifiable
0.2
less efficient units.
Capital costs $/kw 950
–––––––––– Heat recovery system, ORC –––––––––– Assumptions
Gas turbine
Gas flow: 500,000 to 1 million cu m/hr Internal Net replacement
The external parameters used
– Gasin calculating project
–– Air econom-
Cost-benefit
Power capacity
Heat demand Capital, analysis Mw of
rate
Mw
20
present
25
––––
ics
GT 10 B –––– cooling – cooling –– ––
Capital costs
return,
$/kw
value, 650
IRR,include: NPV, IRR, NPV, IRR, NPV, IRR
Station $1,000 Mw % $ % $ % $ % $ %
Energy is gradually lost in gas pressure reduction• stations, Expenses and incomes set at
dissipating 2010-11 levels.
the
No. 1 25,040 17.0 29 33,871 41• Real interest
44,095 rate for–– net value calculation
–– of 5%. 1,523
14 9
mechanical
No. 2 work potential
5,261 3.6 produced
29 by compressors.
7,116 Thein
––• Oil price recovery
–– –– ofmarkets
international mechanical
–– of $60/bbl,
–– yielding–– 3
No. 3 –– –– –– –– –– –– –– –– –– –– 4
average
No. 4working
potential andconditions.
13,467its transformation
9.1 work andathe
29 into 18,216 natural
13 gas8,610
price of $0.25/cu
production of –– m and
electric –– a power––
energy supply price––
is, 3
When
No. 5 ambient conditions5.7
8,417 are cool enough,
29 the 11,385
tempera- of––$0.09/kw-hr.–– –– –– –– –– 5
therefore,
No.
ture a suitable
of 6gas leaving mechanism
––the compressor
–– for enough
may––be low energy
–– to optimization in
–– a natural
––• Iranian thermal power gas
6 plant system.
110
efficiency of––
35%. –– 5
No. 7 14,730 10.0 29 19,924 33 33,295 7 359 –– –– 6
enter
No.the
8 transfer line without5.0
7,365 first being 29
cooled, eliminating
9,962 ––• Rankine cycle
–– technical –– data yielding
–– a design-cycle
–– ef-–– 5
air No. 9 pressure
cooler 15,150drop. Table10.3 3 shows 29gas savings20,493
associ- ––
ficiency of 17.4%–– –– –– –– –– 3
Table
No. 109 contains
atedNo.
with11 bypassing
–– expansion
20,832 the air 14.1
––
cooler. turbine
––
29
data ––for both
28,178
–– small (less
32
––
31,752
than –– 10,000––cu m/hr
–– ––
––
12
––
724
9
10
No. 12
Turboexpanders 5,471 can also 3.7
reduce the 29
amount of 7,400
gas re- Cost
––
normal gas flow) and large units (500,000 to 1 million cu m/hr normal gas flow).analysis –– –– –– –– –– 5
quired to do the same quantity of work. Tables 4, 5, and Five basic energy optimization solutions emerged through
6 show the savings realized by this approach, respectively flow process study at the sample compressor stations:
showing power use when outlet-gas temperature is the same • Heat recovery
Table • from chimney economic
10 shows
Internal rategases and electric power
of return (Equation 2).
asEXPANSION TURBINE DATA
ambient temperature, when outlet-gas temperature is production.
Table 9
cooled as much as possible, and when inlet gas is preheated. •reconstruction.
analyses
• Gas turbine Capital return
based ontime
$60/(Equation 3).
Power capacity Mw • Coolingbbl
0.2 •
natural Net
crude present
gas feed. value of investment (Equatio
oil. The internal
Economic analysis
Heat demand Mw 0.25• Cooling natural gas leaving compressor.
Capital costs $/kw 950 return rate in a 4-year
Energy optimization requires advantages and costs to be re- • Scrubber Assumptions
optimization.
ported
Gas and
flow: assessed
500,000 toin1financial
million cuterms.
m/hr The indices such as- Economicexploitation
investigation exposed
periodthewill
characteristics of
bein calculating
Power capacity
sessments use are: Mw The external
20 feasible solution
each
parameters
(Table 7).
used pr
Heat demand Mw 25 more
•Capital
Cashcosts
flow (Equation 1). $/kw 650Table 8 shows thethan
ics include:results26%/year.
of economic The
analysis of each
• Expenses and incomes set at 2010-11 lev
• Real interest rate for net value calculation
ECONOMIC ANALYSIS, $60/BBL CRUDE • Oil price in international markets
Table 10 of $60
Years of operations –1 0 1 2 3 4 5 6 7 8
average working conditions. a natural gas price of $0.25/cu m and a powe
When ambient
Capacity Mw conditions –– are
0.2 cool enough,
0 the
0.2 tempera- 0.2 of
0.2 $0.09/kw-hr.
0.2 0.2 0.2 0.2
Cash flow $1,000 –190 72 72 72 72 72 72 72 72 72
ture
IRR of gas leaving %/year the –– compressor –– may –17 be low 7enough19to 26 • Iranian
30 thermal
33 power
34 plant36efficiency o
NPV $1,000 –– –115 –53 6 63 117 168 217 263 308
enter the transfer line without first being cooled, eliminating • Rankine cycle technical data yielding a d
airGas flow, 500,000 to 1 million cu m/hr
cooler
Capacity pressure
Mw drop. –– Table 20 3 shows 0gas savings 20 associ-20 ficiency
20 of
20 17.4% 20 20 20
Cash flow $1,000 –– 7,519 7,519 7,519 7,519 7,519 7,519 7,519 7,519 7,519
ated
IRRwith bypassing
TRANSPORTATION %/year the ––air cooler.
–42 10 34 45 50 53 55 56 57
NPV $1,000 –– –5,561
Turboexpanders can also reduce the amount of gas re-
933 7,119 13,010
Cost analysis
18,620 23,964 29,053 33,899 38,515
quired to do the same quantity of work. Tables 4, 5, and Five basic energy optimization solutions em
ECONOMIC ANALYSIS, VARIABLE CRUDE PRICES Table 11
6 show the savings Gas
realized by this approach, respectively flow process study at the sample compressor s
120 Oil & Gas Journal | Mar. 5, 2012
showing powermillion use cu when
savings, outlet-gas temperature is the same
Savings,
• Heat recovery from chimney gases and
as ambient temperature, m/year $ millionwhenYears outlet-gas temperature
–––––––––––––––––– is
IRR, % ––––––––––––––––production.
–––––––––––––––––––––– NPV, $ ––––––––––––––––––
cooled as$/bbl
Oil price, much as––possible,
ORC cycle, electric
–– and –– when 10inlet gas
25 is preheated.
50 65 80 • Gas 10 turbine 25 reconstruction.
50 65 80
gas compressor 44 27.089 10 –7 5 20 28 35 –12,513 579 22,400 35,492 48,584

Gas turbine, • Cooling natural gas feed.
Economic
40% efficiency
analysis
Siemens gas turbine
85
49
––
25
––
10
––
–23
––
5
––
32
––
46
––
59
––
•–18,904
Cooling ––
natural
–4
–– ––
31,495 gas50,395
leaving ––
compressor.
62,294
Air cooler in
Energy
stationoptimization
inlet –– requires –– advantages
–– –– and–– costs––to be ––re- –– • Scrubber –– –– optimization.
–– –– ––
21
Optimized scrubber 2.7 0.7 10 0 32 75 100 124 –140 1,005 2,914 4,059 5,205
ported and assessed
Optimized air cooler 3.72 in financial
3.7 10 terms.– The indices
12 8such 17as- 25 Economic
–3,737 investigation
–2,159 471 exposed
2,049 3,627 the cha
Bypass route 1.07 –– –– –– –– –– –– –– –– –– –– –– ––
sessments use are: each feasible solution (Table 7).
• Cash flow (Equation 1). Table 8 shows the results of economic an
solution using the assumptions
Oil & Gas presented
Journal earlier. Energy re-
:: EXECUTIVE Table 9::contains
BRIEF expansion
sponsored by turbine data for both small
covery in chimneys and its use for electric power production (less than 10,000 cu m/hr normal gas flow) and large units
present value of the investment in a small system for the 4-year exploitation
period will be $117,000 and in large systems $18.62 million. Mechanical recovery
is economically justifiable and provides a basic solution for energy recycling in
compressor stations.
Solution prioritization
As shown in Part 1, only Organic Rankine cycle for energy recovery and
optimization of scrubbers are economically feasible in all instances, the rest of
the solutions being viable only in particular situations.
Table 11 shows detailed economic analyses for sample stations using five different
crude prices.
Shaghayegh Khalaji is a senior expert of energy management at

National Iranian Gas Co. She earned an MS in chemical engineering
(2002) from Tehran University.
22
Originally published April 2, 2012
Analysis yields turnaround

benchmarks for allowance,
contingency
by Gordon Lawrence
T
urnarounds—also known as planned outages or shutdowns—
are major events for refineries and other petrochemical facilities.
They typically cost large sums of money to execute. Cost estimates
for turnarounds, however, have historically been rather inaccurate.
This can often be traced to an inability accurately to calculate allowances and
contingency for “unknowns” in the estimate.
This article examines the allowances and contingencies that are needed, the
different methods used for calculating them, and how much money is typically
allocated and required. From this, it provides some “rule of thumb” benchmarks
for turnaround estimators to use.
It then will discuss how the benchmarks might be refined, the advantages
of tracking the use of allowances and contingencies during execution of a
turnaround, and finally some recommendations for steps that estimators can
take to improve their estimating capabilities for allowances and contingencies.
Turnaround estimate accuracy

As an example of how turnaround cost estimates are notoriously inaccurate.
Fig. 1 shows the average budget overrun for a sample of recent turnarounds.
This sample is taken from the AP-Networks turnaround database, which
contains execution data on more than 750 turnarounds from almost 40 firms at
plants worldwide. It includes schedule data, execution information, and scope
characteristic data, as well as the cost data referred to here.1
23
Analysis yields turnaround benchmarks for allowance, contingency
ACTUAL TURNAROUND COST FIG. 1

165
160 P90 159%
155
150
Portion of total budget, %
145
140
135
130
125
120
115 Mean 116%
110 110% Median = 109%
z120402OGJPLA01
105
100 100%
95
90 90% P10 89%
85
Source: Expected outcome for 10% accuracy Actual outcome
From this sample we can see that turnarounds typically overrun their estimates
by an average value of 16%. Furthermore, the variability around that average is
on the order of –27% to +43%, which is far wider than the ±10% at 80% confidence
that most turnaround teams claim to be their estimate-accuracy level.
The reasons for this weakness in estimating can partly be explained by a lack of
knowledge among owner companies’ estimators as to how much money to allow
in the estimate for the unknowns in their turnaround scope.
What are ‘unknowns’?

Any turnaround will include a maintenance scope, consisting of:
:: I nspection of equipment to company regulations or government mandatory
rules.
:: I nspection of pipework for corrosion and erosion damage, both internal (for
example, process weak points) and external (for example, corrosion under
insulation, or CUI).
:: C
leaning, repair, and maintenance of equipment, pipework, and
24 instrumentation (for example, pulling and cleaning heat exchanger tube

bundles, repairing leaks in pipework, or checking of pressure-relief valves).

:: M
inor upgrades and modifications to the plants (items controlled under the
“management of change” procedures).
Most turnarounds will also include a project scope, consisting of tie-ins (for future
capital projects) that can only be installed when the plant is shutdown.
The scope of both maintenance and project that is known at the time of “scope
freeze” can generally be estimated by the turnaround estimator, based on
historical knowledge. Scope freeze is the date set by the team, by which time all
interested parties—inspection department, plant operators, project teams, so
forth—are to have offered all tasks they would like included in the turnaround.
Those tasks have been challenged and accepted as being unable to be done during
normal operations. This typically occurs around 9 months before the shutdown.
Problems arise, however, with allowing for unknowns. These consist of the
additional work that inevitably creeps into the scope after the scope freeze date
and the costs that grow beyond the original estimate, due to underestimation of
quantities, optimistic estimation of productivity, changes in material costs, and
so forth.
The unknowns are typically covered by turnaround estimators by their including

in the estimate some kind of allowance or contingency funds. Some estimators
simply include a single line item to cover all allowance and contingency
eventualities. The likely areas needing to draw from that fund, however, will fall
into one of four categories:
:: F
orgotten, overlooked work. This is work known before the scope freeze date but
which, for whatever reason, was not included in the basis of estimate.
:: E
merging work. This is work arising from equipment that breaks down or
pipework that leaks after the scope freeze but before the start of the shutdown
for the turnaround.
:: Discovery work. This is work that was unknown (or only suspected) at the time
of scope freeze, which could not be verified because access is impossible while
the plant is running and which is only discovered when equipment or pipework
25 is opened during the shutdown.

:: T
rue contingency. This refers to additional funds for items that simply end up
costing more than originally allowed for in the estimate or for completely
unexpected activities.
The categories of forgotten, overlooked, and emerging work are usually lumped
together by turnaround estimators. This combined category is then typically
known simply as either emerging work or additional work.
Calculating allowances, contingency

The problem for the estimators is how much money to include as allowance or
contingency. There are effectively four basic methods for calculation:
:: Using a predetermined percentage.
:: Using expert judgment.
:: Using a probabilistic risk analysis method such as Monte Carlo simulation.
:: Using a regression model.
We can consider how each of these might be employed by a turnaround

estimator.
Predetermined percentages can be a good “rough-and-ready” way of

calculating contingency. But the issue always arises that, in order to know
what the percentage should be, one needs some benchmark of what was
needed in the past.
Using the “gut-feel” judgment of an expert can also be a good “rough-and-ready”

way of calculating contingency. But, of course, one first needs the expert. Recent
history of turnaround cost outcomes suggests that these experts are few and far
between for turnarounds.
Probabilistic risk analysis uses a Monte Carlo simulation or similar method to

calculate the contingency requirement and estimate range.
At first glance, Monte Carlo appears to offer the most scientific and by extension
the most accurate method for calculating allowances and contingency. One study,
26 however, casts some doubt on that assumption.2 It study looked at Monte Carlo

in comparison to the other three methods discussed here. It focused on capital

project estimates and therefore may not be fully transferable to turnaround
estimates; nevertheless, its findings are instructive.
It was discovered that, contrary to general perceptions in the industry, for

early estimates predetermined percentages and especially expert judgment
provided results that were generally much better than the results from the
supposedly more rigorous Monte Carlo method. Monte Carlo did provide better
results for very detailed estimates, but the advantage it gave over predetermined
percentages and expert judgment was slight.
The referenced article does not discuss in detail why Monte Carlo does not
provide the expected superior results to the other methods, but it seems
reasonable to assume that part of the reason is that Monte Carlo analysis is not
always carried out with sufficient rigor and skill. The study looks at whether
Monte Carlo was used, not whether Monte Carlo was used correctly.
A proper Monte Carlo analysis requires a well facilitated session to identify,

analyze, and evaluate the risks and uncertainties in scope and estimate. That
session in turn, requires sufficient detail in the estimate to allow the risks to be
identified. It also requires the team to abandon the usual over-optimism that
prevails in most discussions of contingency requirements.
Monte Carlo therefore may well be useful for contingency setting but only if
used with sufficient rigor. Since most turnaround teams lack the skills and
training in-house, they would need to look externally for support to implement
this method successfully.
Developing a regression model can be a relatively accurate method for calculating

contingency requirements, but it requires a large data set on which to build the model.
Most owner-company estimators do not have such a data set available to them.
Benchmarks for allowances, contingency

What follows are some benchmark data that could be used by owners’ cost
estimators as rules-of-thumb in applying the predetermined-percentage method
27 or the expert-judgment method to their turnaround cost estimates.

ELEMENTS OF COST ESTIMATE FIG. 2
Turnaround planning and Includes: owners’ turnaround team and contractors’ planning and
preparation (office work) preparation costs
Pre-shutdown Includes: direct & indirect costs associated with pre-turnaround

construction & pre-turnaround work including scaffolding and
field execution pre-turnaround insulation removal
Includes: shutdown support; mechanical-piping (cleaning, bolt ten-

sioning, heat treatment); civil (structural steel, scaffolding, insulation,
Direct labor painting, refractory); electrical & instrumentation; rotating; non-
destructive testing; field machining; industrial cleaning
Labor
Maintenance Includes: turnaround management; planners, supervisors; inspectors;
base cost HSE; schedulers; engineering, technologists; operation reps; logistics
Indirect labor
labor (infrastructure, catering, transportation, accommodation, security,
IT, communication)
Includes: equipment; equipment parts; piping, fittings; manual valves;

Total Materials control valves; E&I; insulation; consumables & supplies;
catalyst; spare parts electrical & instrumentation
maintenance
budget
Total Equipment rental Includes: cranage, bundle puller, etc. (e.g., rolling equipment)
turnaround
budget
Includes: costs associated with provisioning of temporary facilities;
Logistics accommodation; infrastructure
Shutdown field
execution
Forgotten, emerging
Total
allowances
and Discovery
contingency
Contingency
z120402OGJpLa02
Direct labor
Project (turnaround Labor
execution Indirect labor
only) base cost
Materials
These benchmarks present the allowances and contingency as percentages of

elements of the turnaround’s cost estimate. Understanding them requires being
clear about what is meant when we talk about an element of “the cost estimate.”
We have used the definitions as laid out in Fig. 2. The benchmarks are generally
shown as percentages of one of the following:
:: T
he total maintenance and project base cost (i.e., excluding allowances and
contingency).
:: The total maintenance base cost.
:: The total maintenance budget.
The data sets used to calculate the benchmarks are drawn from the AP-Networks
28 turnaround database mentioned earlier.

Allowances, contingency as a block value

These two benchmarks (Fig. 3) show the total allowances and contingency (i.e.,
forgotten work + emerging work + discovery work + true contingency) as one
block, in comparison with different elements of the turnaround budget.
TOTAL ALLOWANCE, CONTINGENCY FIG. 3
As portion of total base cost* FIG. 3a

55
+1 std. dev. 52%
50
45
40
35
30
25
20 20%
%
Mean
15 15%
10 9% Median = 13%
5 Median = 8%
4%
0
–5
–10
–1 std. dev. –12%
–15
Included Used
As a portion of maintenance base cost FIG. 3b

50
45 +1 std. dev. 46%
40
35
30
25
20 Mean 20%
%
15 15% Median = 15%

10 10%
6% Median = 9%
z120402OGJpLa03
–5
–1 std. dev. –6%
29 –10
*Maintenance and projects. Included Used

:: F
ig. 3a shows the total allowances and contingency as a percentage of
maintenance base cost plus project base cost. Two benchmarks are given: how
much teams included in their estimates and how much was actually used by
those teams to complete the turnaround.
As can be seen, the teams included around 9% but actually used about 13-20%,
suggesting that teams were consistently underestimating how much they would need.
In addition, the variability (the vertical line shows the standard deviation) in
the value of how much was actually used, compared with the variability in the
value of how much was included suggests that teams are thinking in terms of
preconceived percentages that take insufficient notice of the amount of risk and
uncertainty in their estimates.
Note: For those attempting to take a “back-bearing” and relate Figs. 3a and
3b to Fig. 1, bear in mind first that Fig. 1 is calculated as a percentage of total
turnaround budget, whereas Figs. 3a and 3b are calculated as percentages of total
base cost and of maintenance base cost, respectively. Secondly, bear in mind that
Fig. 1 uses a larger data set than is used for Figs. 3a and 3b.
We were able to use a larger set for Fig. 1 because, while most turnaround teams
are able to state their total budget and actual costs, far fewer teams track costs in
sufficient detail to provide accurate information about the use of allowances and
contingency within those totals. This in turn raises another interesting discussion
point: The teams that can differentiate their allowances and contingency seem to
overrun their budgets by less than the teams that cannot.
:: F
ig. 3a gives a benchmark comparing the allowances and contingency to
the total base cost for maintenance and projects. The amount of project
involvement in a turnaround, however, can vary considerably. Therefore, in
Fig. 3b, we look at total allowances and contingency as a percentage of the
maintenance base cost only (i.e., without the project base cost).
Despite this change in comparison basis, the results shown in Fig. 3b remain
similar to those in Fig. 3a. This time the amount included is 10%, with the used
amount around 15-20%. Once again, the variability in the amount used is much
30 greater than the variability in the amount included.

Percentage breakdown
The previous two benchmarks looked at allowances and contingency as a single unit
amount. This section looks at each allowance or contingency category separately.
:: F
ig. 4 shows each allowance
TOTAL ALLOWANCE, CONTINGENCY FIG. 4
or contingency category as a
percentage of the total allowances
Forgotten,
and contingency. The pie chart True contingency 39% emerging 19%
shows the average percentages

and then the medians appear
as separate notes in each pie Median = 19%
segment. The average and the

Median = 43%
median are reasonably similar
in each case, suggesting that the
variability for each segment is only
Median = 36%
slightly skewed.
:: F
ig. 5 shows each allowance
z120402OGJpLa04
or contingency category as Discovery 42%
a percentage of the total
maintenance budget (maintenance
base estimate plus all allowances
and contingency). Again, the mean and median are not greatly different.
Greater granularity
The benchmarks given in the earlier section are offered to guide turnaround
estimators who have no other in-house data to begin using expert judgment or a
predetermined percentage in their turnaround estimates.
These benchmarks, however, simply look at the overall average for turnarounds.
From previously published research,3 4 we are already aware that certain
characteristics of a turnaround and of turnaround planning affect turnaround
cost predictability. These characteristics are listed below.
Since these affect cost predictability, they will, by extension, affect the amount
of allowance and contingency required. Hence a next step in providing greater
31 granularity in the benchmarks will be to provide benchmarks that take account

ALLOWANCES, CONTINGENCY INCLUDED* FIG. 5

10
9 +1 std. dev. 8.6%
8.3%
8
7
6
5.5% Mean 5.4%
5 Median = 5.3%
%
4.6% Median =
4 4.4%
Median = 2.9%
3 2.8% 2.7%
z120402OGJpLa05
2 –1 std. dev. 2.1%
1 0.9%
0
Forgotten, emerging work Discovery work Contingency
*Each category expressed as a percentage of total maintenance budget.
of these characteristics.
:: T
urnaround complexity. For example, as shown in Fig. 6, if we examine the
complexity of the turnaround (where complexity is measured as a function of
COMPLEXITY, COST OVERRUN* FIG. 6

40
30
20
10
%
0
z120402OGJpLa06
–10
–20
Low complexity Medium complexity High complexity
32 *Each category expressed as percentage of total turnaround budget.
Source: Reference 4

TURNAROUND READINESS INDEX, COST OVERRUN* FIG. 7

70
60
50
40
30
%
20
10
0
z120402OGJpLa07
–10
–20
Suboptimal Borderline Optimal
Improving turnaround readiness index (TRI)
*Each TRI category expressed as a percentage of total turnaround budget.
Source: Reference 4
the percentage of the turnaround that involves project work, the total field-
labor hours, and the interval time between turnarounds),5 we can see that
more complex turnarounds appear to overrun more (as a percentage of budget)
than less complex turnarounds.
:: T
urnaround readiness. Similarly, as shown in Fig. 7, the level of turnaround
readiness (i.e., the level of planning and preparation carried out for the
turnaround, before the shutdown), as measured with the AP-Networks
Turnaround Readiness Index (a quantitative measure, arising from the
turnaround readiness pyramid questionnaire6), affects cost predictability.
Better prepared turnarounds unsurprisingly have fewer cost overruns.
:: C
ontract strategy. One other interesting point, raised by many turnaround
teams, is whether field-labor contract payment type (e.g., unit rates, time,
and materials, and so forth) affects turnaround cost predictability. One study
suggests that, in comparison with complexity and readiness, contract type
appears to have little to no effect on cost predictability in turnarounds.7
33
Data gathering, differentiation

Clearly, the more historical data one can gather on the use of allowances and
contingencies in turnaround costs, the more benchmarks there are that can be
refined. At AP-Networks, we continue to gather cost and characteristic data on
every turnaround in which we are involved.
We routinely urge every turnaround team to differentiate the allowances and

contingency in the estimate and clearly to track its use in the actual costs.
The cost estimator will frequently find his or her calculations challenged by
management. If the estimator can produce a logical argument for how each
allowance and contingency was calculated, this strengthens his or her argument
for retaining the funds in the estimate.
Using benchmarks gives backing to any argument as to why a certain amount of

allowance and contingency has been included. A further refinement in strengthening
the argument, however, is if less of the estimate needs to be contingency, unallocated
to a specific task or action. Therefore, let us examine all four categories of allowance
and contingency to see whether we can better define them.
If we take each of the four categories in turn, we can see that in fact only one of
the four categories, “true” contingency, really needs to be subject to a calculation
technique such as expert judgment, predetermined percentage, or probabilistic
risk analysis.
:: F
orgotten, overlooked work. If a team is conducting a thorough scope-gather exercise,
involving all groups with a stake in adding scope to the turnaround, then this
category should ideally be zero in every estimate. Tracking its actual value will
give teams a clue as to how effective are their scope gathering techniques.
:: E
merging work. It should be possible to estimate emerging work, based on
knowledge of the reliability of the plant. For example, “We have X months
between now and the shutdown. Typically, we have Y valves/month that fail.
Therefore we should allow for X*Y failures in the estimate as emerging work.”
:: Discovery work. It should be possible to make an educated guess at discovery
work, based on the operator’s knowledge of how well the plant is operating,
34 before the shutdown. For example, “Column X is only operating at 50% of

capacity; therefore, we fully expect that when we open it there will be tray
damage to repair.”
:: ( True) contingency. True contingency is the only category that will still need to be
left to expert judgment, predetermined percentage, or probabilistic risk analysis.
Forgotten, overlooked work, emerging work, and discovery work all now become
“allowances” rather than “contingency” and can be assigned to specific line items
in the cost estimate. They no longer need to be amorphous buckets of money,
subject to suspicion and deletion. A clear case can be made for each cost item.
Following the same reasoning, just as with capital projects, the true contingency
now becomes money that, by its very nature cannot be assigned to any specific
line items, but which history or experience shows will be required somewhere
during execution of the project for some element of the scope that was
underestimated or overlooked.8
This then brings turnaround estimating more in line with capital cost estimating,
where, if you can assign it to a specific line item in the estimate, it’s an allowance;
and if you can’t but historical experience tells you that you’ll need the money
somewhere, it’s a contingency.
Looking further now makes clear that the only one of the four categories that
should be affected by the presence of capital projects in the shutdown scope
is true contingency, since the three allowance categories all clearly relate to
the maintenance scope only. Recognizing this should also help to make the
calculation and use of contingency in turnaround budgets less opaque.
AP-Networks will continue adding to its database and examining this topic,
delving deeper into the variables (complexity, readiness, contract strategy, etc.)
that may affect the need for allowances and contingency.
References
1. http://www.turnaroundbenchmarking.com/database.html.
2. Burroughs, S.E., and Juntima, G., Exploring Techniques for Contingency Setting. AACE
International, Washington, June 13-16, 2004.
35 3. Alkemade, J.M.G., “Key Leading Indicators of Turnaround Performance Outcomes,”

Plant Maintenance Middle East Conference, Dubai, Nov 26-28, 2007.

4. O’Kane, M., “In Search of Excellence: Benchmarking Turnaround Outcomes and
Practices,” Turnaround Industry Networking Conference, Houston, Sept. 24-26, 2011.
5. http://ap-networks.com/tools/turnaround-network.html.
6. http://ap-networks.com/services/turnaround-assist-program-t-ap.html.
7. Lawrence, G.R., “Contract Strategies for Turnarounds—A comparison of the use of
Unit Rates versus Time & Material,” Turnaround Industry Networking Conference,
Houston, Sept. 24-26, 2011.
8. The exact definition of contingency, given by the Association for the Advancement
of Cost Engineering—International in its Recommended Practice 10S-90 “Cost
Engineering Terminology,” is “An amount added to an estimate to allow for items,
conditions, or events for which the state, occurrence, or effect is uncertain and that
experience shows will likely result, in aggregate, in additional costs.”
Gordon Lawrence (glawrence@ap-networks.com) is a senior

consultant based in the Amsterdam office of Asset Performance
Networks, a turnaround and capital-project consultancy. He has
more than 25 years’ experience in process industries. In his career, he
has progressed from project engineer to project manager and lately
senior project manager at several owner and contractor organizations,
including the US contractor Jacobs Engineering and the Swiss pharmaceutical firm
Novartis. He also previously worked as a consultant in capital project benchmarking.
Lawrence is a chartered engineer in the UK, a registered professional engineer in the
European Union, and a fellow of the UK Institution of Chemical Engineers. He is also
a member of the American Institute of Chemical Engineers. He holds a bachelors in
chemical engineering (1984) from Heriot-Watt University, Edinburgh, a masters (1985)
in biochemical engineering from Birmingham University, England, and an MBA (1991)
from Strathclyde University Business School in Glasgow.
36
Company Description:
With almost 70 years of experience, WIKA Instrument Corporation is the leading
global manufacturer of pressure and temperature measurement instrumentation,
producing more than 43 million pressure gauges, diaphragm seals, pressure
transmitters, thermometers and other instruments annually. WIKA’s extensive
product line, including mechanical and electronic instruments, provides
measurement solutions for any application in a large variety of industries. A
global leader in lean manufacturing and instrumentation experience, WIKA
offers a broad selection of stock and custom instrumentation as well as dedicated
services to provide customers with the right solutions, at the right time, wherever
they need us.
These value-added services include the Full Audit Service Team (FAST) for the
downstream petroleum industry. Our FAST services include: Instrument Audit,
Turnaround Instrument Planning, Instrument Failure Analysis and Instrument
Safety Training. During an Instrument Audit, FAST engineers evaluate your
gauges to identify and correct issues before a serious incident happens. FAST
engineers also perform a Storeroom Audit to reduce your SKUs and ensure only
the right gauges are on the shelf – eliminating guesswork when a replacement
gauge is needed. With Turnaround Instrument Planning, FAST engineers help reduce
discovery during the project to avoid cost over-runs, scope creep and leaks
upon start-up. With our Instrument Safety Training, FAST engineers teach your
maintenance team to spot gauge issues before they develop into serious problems.
links:
Navigating a Mine Field and Staying Safe — How Overlooking

Instrumentation Could Cost You at Every Turn
37

Compressor Optimization

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Compressor Optimization

Transféré par

Droits d'auteur :

Formats disponibles

EXECTUIVE BRIEF

who’s handling your next “Oh

Let us show you how today by downloading our free eBook at

The pursuit of ‘zero leaks’

The timeline to establish requirements for gas transmission pipeline operators

The promulgation of such programs on a state level has leapt in importance in

Oil & Gas Journal :: EXECUTIVE BRIEF :: sponsored by

Christopher E. Smith is Pipeline Editor for Oil & Gas Journal.

Energy recovery guides natural g

E pipeline system efficiency

Economic analyses dem- across the c

transmitting gas Natural Power

SIMULATION SCHEMATIC FIG. 1

Pressure Gas compressor

Liquid Gas compressor

From recycle Recycle valve To recycle

pressure drop in scrubber and air coolers (Fig. 1).

The following conditions governed the simulation:

:: Input gas pressure, 700 psi.

:: Output gas pressure, 1,000 psi.

Operational conditions Input gas flow, MMscfd

After investigatingOilglobal energy

Investigating efficiency of selected compressor stations along the three pipelines

Oil & Gas Journal :: EXECUTIVE BRIEF :: sponsored by

COMPRESSOR THERMAL EFFICIENCY FIG. 2

Siemens 70.15 76.65 10.53 9.54 No. 1 48 37.3 14/7

Oil & Gas Journal :: EXECUTIVE BRIEF :: sponsored by

requirement for gas transfer of 1 mil- China 1,084 5 24.88 50 0.00459

ture from 7).

COMPRESSOR COMPARISON Table 9 SCRUBBER PERFORMANCE Table 10

Siemens 70.15 76.65 10.53 9.54 No. 1 48 37.3 14/7

:: Recovering energy lost through the chimney.

:: Replacing the gas turbine.

Inlet air TRANSPORTATION

Practical methods for cooling

chillers.operational conditions and

al energy were producing <4 Mw, prompting a 3-year Compressors

Oil & Gas Journal :: EXECUTIVE BRIEF :: sponsored by

existing units, resulting

15 izing these savings and an economic analysis of the results.

Oil & Gas Journal :: EXECUTIVE BRIEF :: sponsored by

After investigation determined energy loss levels, potentials for mitigating

scrubbers :: Energy loss from air cooler pressure drop.

:: Electrical energy consumption in air coolers.

ent operation Scrubbers

V t = Annual variable expenses

ing energy consumption norms in gas

Oil & Gas Journal :: EXECUTIVE BRIEF :: sponsored by

STATION PERFORMANCE, OUTLET GAS TEMPERATURE SAME AS AMBIENT Table 4

offer keys to efficient operation

Shaghayegh Khalaji more power/un

Oil & Gas Journal :: EXECUTIVE

process study at the sample compressor stations:

:: Gas turbine reconstruction.

:: Cooling natural gas feed.

:: Cooling natural gas leaving compressor. TRANSPORTATION

SOLUTION CHARACTERISTICS BY STATION Table 7

No. 1 119 49 25,000 –– –– 3.72 3,710 2.7 700

eters and criteria for comparing operational conditions with

No. 1 25,040 17.0 29 33,871 41 44,095 –– –– 14 1,523 91 3,678

gas compressor 44 27.089 10 –7 5 20 28 35 –12,513 579 22,400 35,492 48,584

Shaghayegh Khalaji is a senior expert of energy management at

Analysis yields turnaround

Turnaround estimate accuracy

:: Input gas pressure, 700 psi.

:: Output gas pressure, 1,000 psi.

:: Recovering energy lost through the chimney.

:: Replacing the gas turbine.

scrubbers :: Energy loss from air cooler pressure drop.

:: Electrical energy consumption in air coolers.

:: Gas turbine reconstruction.

:: Cooling natural gas feed.

:: Cooling natural gas leaving compressor. TRANSPORTATION

:: Using expert judgment.

:: Using a probabilistic risk analysis method such as Monte Carlo simulation.

:: Using a regression model.

:: The total maintenance budget.