SPE 94373

Optimize Separator Operating Pressures to Reduce Flash Losses

B.E. Boyer, SPE, COMM Engineering, and S.C. O'Connell, Devon Energy Corp.

Copyright 2005, Society of Petroleum Engineers Inc.

system or do not operate a low-pressure booster compressor,
This paper was prepared for presentation at the 2005 SPE/EPA/DOE Exploration and the low-pressure separators and/or heater treaters may vent to
Production Environmental Conference held in Galveston, Texas, U.S.A., 7 – 9 March 2005.
the atmosphere or burn the gas in a flare.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in a proposal submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to Companies can optimize the process by reducing operating
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at pressures of low-pressure separators and heater treaters such
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
that less flash gas is vented to the atmosphere. For example,
for commercial purposes without the written consent of the Society of Petroleum Engineers is less flash will be generated from a oil storage tank provided
prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300
words; illustrations may not be copied. The proposal must contain conspicuous the facility reduces the operating pressure of the low-pressure
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
separator or heater treater just upstream of the oil storage tank
and the flash gas from the low-pressure separator or heater
Abstract treater is routed back into the system (e.g., suction of
Flashing losses from crude oil and condensate storage are compressor, fuel gas system, etc.). This may require a lower
routinely vented to the atmosphere. By minimizing operating operating pressure for the booster compressor's first stage
pressure of low-pressure separators, the amount of flashing suction or for the fuel gas system. These optimizations can be
losses can be reduced, resulting in increased profits at a accomplished by adjusting operating pressures with minimal
minimal cost of implementation and an immediate payback. capital and operational costs.
Companies can be report these methane reductions to the
Environmental Protection Agency's Natural Gas STAR A survey of the operating parameters of a facility can reveal
Program. Devon Energy realized savings of approximately opportunities to reduce flash losses and increase profits by
$7000 per year after separator optimization at one of its oil increasing gas sent to sales. Such a survey may reveal
and gas production facilities. opportunities that warrant a vapor recovery unit to recover the
remaining flash gas that is vented to the atmosphere before or
Introduction after optimization.
Devon Energy Production Company, L.P. (Devon) surveyed
its G. A. Ray No. 93 oil and gas production facility and These flash gas savings result in a reduction in the emissions
increased profits approximately $7000 per year by optimizing of methane, a greenhouse gas. Companies can report these
the operating pressures of the three-phase, low-pressure methane reductions to the United States Environmental
production separators. The primary goal of the optimization Protection Agency's (EPA) voluntary program known as
was to increase profits for the facility by putting more gas into Natural Gas STAR Program. The Natural Gas STAR's
the sales pipeline and to reduce emissions of methane with Lessons Learned document entitled, "Installing Vapor
minimal costs to the facility. This paper describes methods for Recovery Units on Crude Oil Storage Tanks”1 gives
optimizing separator pressures at an oil and gas production information on the sizing and economics of vapor recovery
facility and Devons' results for optimizing its G. A. Ray No. units for oil storage tanks.
93.
Statement of Theory
In the oil and gas production, processing and transmission There are many areas in the production, gas processing and
operations, processes that involve crude oil and condensate transmission processes where flash gases are generated and
undergoing pressure drops result in the separation of natural vented to the atmosphere. These include:
gas from the oil fraction. This liberation of natural gas is
commonly referred to as "flashing" of natural gas from the oil. 1. Intermediate Flash: high/intermediate pressure
These flash gases typically have a higher BTU value than the separators that send crude oil and condensate to a
inlet gas and have more value. These gases from crude oil and low-pressure separator. This low-pressure separator
condensate (oil) storage tanks are generally vented to the operates at a pressure greater than atmospheric
atmosphere (although sometimes they are recovered for use as pressure and may vent directly to the atmosphere. A
fuel gas, gas lift gas, sent to the sales pipeline or burned in a heater treater may also be used in an intermediate
flare). For facilities that do not have a low-pressure gathering flash situation. (See Figure 1)
2 SPE 94373

2. Fixed Roof Storage Tanks: high/low-pressure crude pressure of 80 psig. If the operator reduced the operating
oil and condensate are flashed into tanks operating at pressure of the low-pressure separator to 50 psig, then the
atmospheric pressure. (See Fig. 1) suction pressure of the compressor must be less than 50 psig to
3. Pipelines: gas lines that are "pigged" or physically ensure the low-pressure gas is recovered by the compressor.
purged of condensate have the potential to vent
natural gas to the atmosphere.
4. Inlet Separators: gas plant inlet separators that dump
into storage tanks operating at atmospheric pressure.

Note that for storage tanks, emissions from flashing are in

addition to tank working losses and breathing losses and
loading losses from tank trucks and barges. Flashing
emissions can be several orders of magnitude greater than the
combined working, breathing and loading losses, depending
on the pressure drop and the volatility of the oil or condensate.

Fig. 1-Process flow and possible destination of flash gas

The optimization technique is to operate the low-pressure Procedure

Below are the five steps that Devon used to optimize operating
separators or heater treaters at the lowest possible pressures
pressures of the low-pressure separators flowing to oil storage
such that less flash gas is generated from the pressure drop the
tank battery.
oil takes between the separator and the downstream vessels
that vent to the atmosphere or burned in flare. For this
1. Choose a flash gas volume estimation method
technique to be successful, the operating pressure of the
2. Collect needed process data
upstream vessel (e.g., low-pressure separator to be optimized)
3. Determine existing gas-to-oil ratio (GOR) and gas
must be of greater pressure than the downstream vessel (e.g.
volume of flash gas lost
compressor) that receives the flash gas. For example, a low-
4. Determine the optimal operating pressures and
pressure separator operated at 85 psig sends its natural gas to
implement
the suction of an onsite booster compressor with a suction
SPE 94373 3

5. Determine the new GOR and volume of flash gas lost Pressurized oil sample and GOR. To estimate evolved gas
after lowering operating pressures and calculate the losses using this method, the technician collects a pressurized
monetary volume of recovered methane oil sample from the separator using a water-filled cylinder.
For a multiple staged flash, an operator can collect a sample
Step 1. Choose a Vent Gas Flowrate Estimation Method. from the high pressure separator and determine the flash for
Flash gas liberated due to crude oil and/or condensate each stage of separation of interest. The recommended
undergoing a pressure drop between separators, heater treaters procedure for collecting the samples is Gas Processors
and storage tanks can be estimated using the following Association (GPA) Standard 2174 -93. Care must be taken to
methods: ensure that no water from the separator enters the sample
cylinder. The laboratory determines the corresponding gas-to-
1. Direct measurement oil ratio (GOR) in standard cubic feet per stock tank barrel and
2. Collection and analysis of a pressurized oil the compositional analysis of the flashed gas at each stage of
sample to obtain a gas-to-oil ratio (GOR) in separation.
standard cubic feet per stock tank barrel
produced Process Simulators. Process simulators such as HYSYS
3. Process simulators such as HYSYS or PROSIM (Hyprotech at www.hyprotech.com), PROSIM® (Bryan
4. Vasquez-Beggs Equations2 Research & Engineering at www.bre.com) and WinProp
5. American Petroleum Institute's (API) E&P (Computer Modeling Group at http://www.cmgroup.com) are
TANK Version 2 software3 frequently used to estimate flash losses, especially for
6. Griswold and Ambler GOR Chart Method4 designing new facilities. Equations of state commonly used
for tank flash estimates include the Peng-Robinson and Soave-
For a given set of process conditions, each of the above Redlich-Kwong Equations of State. Process simulators can
methods can have a large variation in results. Direct use compositional analyses from pressurized oil and gas
measurement of the vent flowrate, pressurized oil sampling samples to simulate flash generation at each stage of
and chemical analysis and process simulations are often used separation and from storage tanks.
to obtain the most accurate data.
Vasquez-Beggs Equations. The Vasquez-Beggs Equations
Flashing losses from an intermediate separator should use are based on a paper entitled "Correlations for Fluid Physical
direct measurement, pressurized oil sampling/analysis or Property Prediction" by Vasquez and Beggs (1980) and are
process simulation. The Griswold and Ambler GOR Chart widely used by companies and regulatory agencies to estimate
Method, Vasquez-Beggs Equations, and the E&P TANK flash losses from oil storage tanks. The equations yield the
software were developed to determine a single flash between GOR of natural gas liberated in standard cubic feet per barrel
separators and/or heater treaters and storage tanks operating of oil produced. The user estimates the amount of methane
nearly at atmospheric pressure. based on analytical data for the flash gas. Input variables
include stock tank oil API gravity, separator pressure,
Direct Measurement. Direct measurement gives the most separator temperature, gas specific gravity at separator
accurate results for evaluating flash gas flow rates. The types conditions and volume of produced hydrocarbons.
of meters used for direct measurement include in-line meters,
insertion meters and strap-on meters. Each of these meters API E&P TANK Version 2. The American Petroleum
can use different technologies ranging from ultrasonic transit- Institute (API) developed the E&P TANK Version 2 program
time meters, turbine meters, thermal mass flow meters, to estimate oil storage tank flashing losses along with storage
positive displacement and vortex flow meters. The accuracy tank working and standing losses. The model uses the Peng-
and precision of each measurement device will vary Robinson equation of state to estimate tank-flashing losses.
depending on the system pressure and flow rate. E&P TANK also allows the user to input compositional
analysis from pressurized oil and gas samples to simulate flash
Prior to measurement of a storage tank, the facility must generation in storage tanks.
route all vent gas from the storage tanks to the vent line, repair
anything leaking from or bleeding into the storage tanks, Griswold and Ambler GOR Chart Method. This method
replace or repair all worn or faulty tank hatch seals and repair uses graphical curves of API gravity of stored crude
all holes. Then collect a sample of the vent gas and oil/condensate and separator pressure to determine the GOR.
chemically analyze for components so that the mass and BTU These curves were constructed using empirical flash data from
value of the vent gas can be determined. laboratory studies and field measurements (Griswold and
Ambler, 1978). Use the graph to approximate total potential
Taking measurements frequently (e.g., 1-minute averages) vapor emissions from a barrel of oil based on the pressure
over the course of a 24-hour period will yield data on the differentials given for the x-axis of the graph.
variations in flow rates. Such data are valuable for sizing a
vapor recovery unit to handle expected flow variations of the Step 2 - Process Data Collection. Collect facility process
vent/flare gas remaining after optimization. data to determine the potential for lowering the operating
pressure of the separator(s) just upstream of the oil storage
tanks. Determine the average throughput of crude oil and/or
4 SPE 94373

condensate for the facility. Determine the destination of the METHANE = Volume of methane vented in scf per day
flash gas from the low-pressure separator(s), heater treater and MWM = Mole percent of methane in vent gas
storage tank(s). For example, determine if the gas is routed to
the booster compressor suction, the low-pressure gathering From this data, the operator can calculate the total volume of
system, fuel gas system, flare or to the atmosphere. flash gas per day (scf/day), the dollar value of the flash gas
Optimization will be effective only if gas from the low- based on the dollar value of the flash gas per 1,000,000 BTUs
pressure vessel just upstream of the storage tank is not routed ($/MMBTU) and the methane content for reporting under the
to a flare or to the atmosphere. Gas STAR Program.

Collect data on the operating pressure of the low-pressure These formulas were used to develop Table 1 on page 7.
separator or heater treater just upstream of the storage tanks
and for the all stages of the onsite booster compressor or other CASE STUDY
low-pressure system (e.g., fuel gas). Determine if the Fig. 2 displays a simplified flow of gas and oil for the process
operating pressures of the separators/heater treaters and at Devon's G.A. Ray No. 93 Facility. Multiple oil and gas
booster compressor or low-pressure system can be reduced production wells flowed to the respective high-pressure
while maintaining the recovery of flash gas from the separators. System No. 1 wells flowed to High-Pressure
separator/heater treater. Adjust as necessary and determine Separator No. 1 operating at 425 psig at 86°F, and System No.
flash gas emissions after optimization. 2 wells flowed to High-Pressure Separator No. 2 operating at
295 psig at 85°F. The booster compressor collected flash gas
Step 3 - Determine the existing GOR and volume of flash from the high-pressure separators for compression to the sales
gas lost. Determine the GOR and volume of flash gas and pipeline pressure of 800 psig.
methane emitted per year for the existing operating pressure
conditions. Estimate the amount of flash gas emissions vented Oil from the high-pressure separators flowed to the
to the atmosphere or burned in a flare prior to optimization respective low-pressure separators (Nos. 1 and 2) for further
using the vent gas flowrate estimation methods in Step 1. separation of gas, oil and water. Gas from the low-pressure
separators flowed to suction of the booster compressor for
Step 4. Determine the optimal operating pressures and ultimate compression to the gas sales pipeline pressure of 800
implement. Determine if the operating pressures of the psig. The low-pressure separators operated at 40 psig. The
separators/heater treaters and booster compressor or low- booster compressor first stage suction was set at 35 psig.
pressure system can be reduced while maintaining the Oil from the low-pressure separators flowed to the oil
recovery of flash gas from the separator/heater treater. Adjust storage tanks.
as necessary and determine flash gas emissions after
optimization. Devon conducted sampling, chemical analyses and computer
simulations for several operating scenarios for separators and
Step 5. Determine the new GOR and volume of flash gas tanks operating at its G.A. Ray No. 93 Production Facility
lost after lowering operating pressures. Use one of the located in Bee County, TX. The sampling consisted of a
methods described in Step 1 to determine the GOR and pressurized oil sample taken from each of the two high-
volume of flash gas lost after optimization of separator pressure separators operating at the facility.
pressures.
FESCO, Ltd conducted a laboratory analysis of each
Step 6. Calculate the reduction in vent gas and the pressurized oil sample to determine the gas-to-oil ratio (GOR)
monetary value of the vent gas. The following formulas5,6 and chemical make-up of the flash gas. The analyses were
can be used to calculate vent gas volumes and the monetary conducted for the intermediate flash between each high-
values of the gas. pressure separator's operating conditions and the low-pressure
separator conditions of 20 psig and 80 psig for a total of four
VOL = ( PROD )(GOR ) intermediate flash analyses. The samples from each separator
were also analyzed to determine the GOR chemical makeup
for flash losses from 20 psig to storage tanks conditions and
VALUE = (GOR )(VOL)( BTU )( P ) from 80 psig to storage tank conditions for a total of four final
flash analyses.
METHANE = (GOR)(VOL)( MWM )
The GOR and chemical makeup of flash gas for several
VOL = Volume of vent gas in standard cubic feet (scf) per day other operating conditions were also simulated using WinProp
PROD = Volume of oil produced in barrels per day 2004. The simulations used the Soave-Redlich-Kwong
GOR = Gas-to-oil ratio in standard cubic feet flash gas per Equations of State tuned to actual laboratory analyses from the
barrel (scf/bbl) field samples. The results of the testing and analyses for
VALUE = Dollar value per day of vent gas various operating conditions are summarized in Table 1.
BTU = BTU value of vent gas in BTUs per scf
P = $5.00 per 1,000,000 BTU
SPE 94373 5

To optimize the operations, Devon adjusted the operating

pressures of the low-pressure separator and booster
compressor suction to 20 psig. This reduced the pressure drop
by 20 psig. The average oil production to each of the
separators was 100 barrels of oil per day (BOPD).
6 SPE 94373

Fig. 2-Process Flow before and after optimization

SPE 94373 7

Total estimated flash gas emitted to the atmosphere from the

storage tanks before optimization was 2,650,000 scf/year and
total estimated gas flashed emitted to atmosphere after
optimization was 1,789,000 scf/year. This resulted in an
increase of potential gas to sales of approximately 861,000
scf/year. The gas would be worth $6,896 per year based on
the BTU content of the flash gas and a gas value of $5.00 per
MMBTU (Table 1).

Total estimated methane emitted to the atmosphere from the

storage tanks before optimization was 653,000 scf/year and
total estimated methane emitted to atmosphere after
optimization was 317,000 scf/year. This resulted in an
increase of potential methane gas to sales of approximately
336,000 scf/year.

Table 1. Results for flashes from low-pressure

separator conditions to atmospheric storage tank
conditions

Flash Analysis for Pressure Total Flash Gas Total Flash Flash Gas Methane Methane
Gas-to-Oil Ratio drop (psig) GOR BTU Value VolumeA Value per GOR Ratio Flash
(GOR) (scf/bbl) (BTU/scf) (scf/year) yearB (scf/bbl) VolumeC
(scf/year)
Sample for LP (Low
Pressure Separator) 20 20 2303 730,000 $8,406 4.5 164,000
No. 1 to storage tanks
Sample for LP No. 2
to storage tanks 20 29 2548 1,059,000 $13,492 4.2 153,000

TOTALS 1,789,000 $21,898 317,000

Simulation for LP No.
1 to atmospheric tank 40 30.4 2079 1,110,000 $11,538 8.6 314,000

Simulation for LP No.

2 to storage tanks 40 42.2 2241 1,540,000 $17,256 9.3 339,000

TOTALS 2,650,000 $28,794 653,000

Simulation for LP No.
1 to storage tanks 60 47.7 2083 1,741,000 $18,133 13.5 493,000

Simulation for LP No.

2 to storage tanks 60 63.5 2225 2,318,000 $25,788 14.7 537,000

TOTALS 4,059,000 $43,921 1,030,000

Sample for LP No. 1
to storage tanks 80 75 2172 2,738,000 $29,734 19.9 726,000

Sample for LP No. 2

to storage tanks 80 90 2384 3,285,000 $39,157 19.2 701,000

TOTALS 6,023,000 $68,891 1,427,000

A,C
Estimate based on average production rate of 100 barrels of
oil per day.
B
Estimate based on average production rate of 100 barrels of
oil per day and flash gas valued at $5.00 per 1,000,000
BTU/scf.
8 SPE 94373

Conclusions
Devon realized a $6,896 per year increase in gas to sales by SI Metric Conversion Factors
reducing the operating pressure of the low-pressure separators.

This demonstrates that optimization of operating pressures at Barrels (bbl)X 0.1589 = m3

facilities can recover more product for sales and can reduce cubic feet X 0.02832 = m3
the amount of methane gas lost to the atmosphere. This (°F-32)/1.8 = °C
optimization technique can be implemented at negligible costs psi X 6.895 = kPa
to the facility. By obtaining buy-in from company
engineering and field operations, Devon is implementing
optimization opportunities company wide. The reductions in
methane will be reported to the EPA's Gas STAR program.

The flash gas recovered will also reduce emissions of

volatile organic compounds (VOCs) and hazardous air
pollutants (HAP) such as n-hexane, benzene, toluene, ethyl
benzene, and xylenes. If the facility operates above major
source emissions levels specified in Title V of the Clean Air
Act, it is possible that optimizing the operating pressures
would result in reducing facility emissions below Title V
levels, thereby qualifying the facility for permit exemptions or
minor source air permits. To ensure that Title V permits are
not required, federally enforceable permit limits are needed to
make a facility a "synthetic minor" source of emissions.

As a further benefit, the volume of flash gas vented or flared

after optimization might be large enough to warrant the
installation of a vapor recovery unit.

References
1. Lessons Learned; Installing Vapor Recovery Units on Crude Oil
Storage Tanks (see http://www.epa.gov/gasstar/install.htm)

2. Vasquez, M. and Beggs, H.D.: “Correlations for Fluid Physical

Property Predictions,” Journal of Petroleum Technology (June
1980- 968-970) SPE Paper 6719.

3. Radian International LLC,: Evaluation of a Petroleum

Production Tank Emission Model, API Publication No. 4662,
American Petroleum Institute, Washington, D.C., (1997).

4. Griswold, John A., Power Services, Inc. and Ambler, Ted C., A
& N Sales, Inc.: “A Practical Approach to Crude Oil Stock Tank
Vapor Recovery,” paper SPE 7175 presented at the 1978 SPE
Rocky Mountain Regional Meeting, Cody, WY, May 7-9.

5. Harris, G., Shires, T., and Loughran, C.: “Compendium of

Greenhouse Gas Emissions Estimation Methodologies for the Oil
and Gas Industry,” American Petroleum Institute (2001).

6. USEPA Emission Inventory Improvement Program (EIIP)

Volume II, Point Sources, Chapter 10 - Preferred and Alternative
Methods for Estimating Air Emissions from Oil and Gas Field
Production and Processing Operations. September 1999. (see
www.epa.gov/ttn/chief/eiip/techreport/volume02/ii10.pdf)