Logo
About us banner

Print Printable View

banner
High Temp In Heavy Oil - Logo

SPE/PS-CIM/CHOA 98009

PS2005-433

High Temperature Multiphase Flowmeters in Heavy-Oil Thermal Production

P. Mehdizadeh, SPE, Consultant, Production Technology Inc.

Abstract

The accurate measurement of Oil, Water and Gas/Steam in heavy oil thermal production (SAGO and other Steam Flood Processes) is a very diffcult task faced by the heavy oil industry. The accuracy of these measurements is critical for reservoir management and production diagnostics. Mulitphase flow meter technology has been used successfully around the world for over 10 years and in heavy oil "cold" production in Venezuela and other countries. But multiphase technology has never been used in Extra Heavy Oil Thermal Production. The Canadian heavy oil thermal producers regularly see production temperatures exceeding 200 C (392 F) and some wells are approaching 232 C (450 F). New technology is required to accurately measure wells producing at these elevated temperatures. The first field tests using a multi phase flow meter in a heavy oil thermal project was conducted by one of the major Canadian producers in the fall of 2004. Additional tests were completed during the summer of 2005 with another heavy oil producer. This paper will review the unique problems encountered with testing heavy oil in high temperature applications. The test results from multiple well tests and the accuracy of the multi phase flow meters when compared to the field reference will be presented.

Introduction

An estimated six trillion barrels of heavy oil and bitumen is available worldwide. A majority of these reserves are located in United States, Canada and Venezuela (Ir Thermal-based recovery methods have been used since 1950's to recover oil from these reservoirs. Significant changes related to reservoir management and production facilities have been made (2). In the past decade new thermal recovery techniques such as the steam assisted gravity drainage (SAGO), two cyclic steam stimulation (CSS), steam and gas push (SAP) and vapor extraction (Vapex) processes have been developed and used to enhance the recovery of very heavy oil (i ,3,4). Among these processes, the steam assisted gravity drainage has emerged as an effective technology for recovering oil from sand deposits that are too deep to be recoverable by surface mining(3,4). In the SAGO process, steam is injected continuously down one well while the mobilized bitumen and condensate steam are produced continuously up a second welL. The injector and producer are drilled approximately 5 m apart. In some developments, horizontal wells are used to enhance reservoir access and well productivity (4). In the injector well, steam injection creates a chamber that grows as the steam condenses on the chamber walls and releases heat. Heated bitumen and condensed steam drain by gravity into the lower producing well and are pumped out.

The oil-in-water emulsion produced by the SAGO process is very stable and requires chemical treatment and processing to separate the oil and water (5). Conventional gravity based test separators used in measuring well rates are not able to deal with this stable emulsion, in the presence of steam condensate and produced gas, as well as the high temperatures. Reference 5 has reported on a SAGO plant that processes a reverse emulsion of about 360 m3/day of oil and 1200 m3/day of water at 195-200 C and 1800 kPa (260 psig). At the operating temperature of the separator (about 200 C) the produced water is less dense than bitumen. Wet oil exits from the bottom of the high temperature separator and enters a flash treater operated at temperatures above 145 C to flash out the steam. These conditions make it diffcult for conventional gravitybased test separators, with limited retention time, to produce accurate measurements of the produced fluids.

Multiphase metering technology has been used successfully for well testing in the production of heavy oil (stable emulsions) in cold processes (6). Multiphase measurement technology has been an enabling technology for the owners of the Petrozuata heavy oil operation (6). The installation of 37 multiphase meters in conjunction with the multi phase pumps in place of separators, liquid pumps, and gas compressors paid significant dividends in CAPEX and continue to pay year on year in aPEX. Multiphase metering techniques have also been used successfully in steam flood operations (7, 9) to improve field allocation factors. The field tests described in this paper were undertaken to assess the capability and advantages of using multiphase metering technology to overcome the shortcomings of the conventional gravity based test separators used in a number of high temperature thermal recovery operations such as CSS and SAGO.

Fluid Processing and Facilities

The two sites (A and B) used in the field tests are operated by different companies. Both operators employed SAGD or a modification thereof to produce the oil bitumen. All wells at Site A are produced by long stroke rod pumps. The combination of lift method, operating pressure and temperature as well as the SAGO process used at Site A produced a stream that had fairly low gas (steam and hydrocarbon) volume fraction. Site B had 2 pairs of SAGO wells. Well I produced a muItiphase stream with low gas volume fraction and used ESP for lift. Well 2 was produced using steam lift. This well had. high gas (steam and hydrocarbon) volume fraction. Table 1 summarizes the flow rates and fluid conditions, pertinent to production measurements, encountered in these test sites.

Table 1 - Fluid Conditions at the Test Sites

High Temp In Heavy Oil - Fig. 1
Fig 1 - Schematic of the installation of the multiphase meter (l\PFl\) and associated equipment for obtaining tank test reference now rates
Location Site A Site B
Liquid / Emulsion Rate 53 - 350 m3/day 200 - 550 m3/d
% WC Range 22 - 85 70 - 90
Temperature - °C 87 - 166 130 - 180
Pressure - kpa 770 - 1800 400 - 1100
% Gas Volume Fraction 14 - 67 10 - 98

Field Testing of Multiphase Meter - Site A

Field qualification of multi phase meters (11) is a challenging task since many field test separators and facilities, which have to provide reference data, lack the accuracy and consistency to be used as the "reference". Previous experience with qualification tests of multiphase meter performance in heavy oil production (6, 8, and 9) had indicated difficulties with obtaining accurate reference data using conventional measurement procedures. Considerable amounts of effort were therefore given to the planning of the field tests - specifically related to the gathering of reference data. Since the conventional test separator used in the facility had diffculties with producing accurate well test data, it was decided to measure the performance of the multi phase meter against tank tests. A number of different procedures were tried to obtain reference flow rate and water cut data to check the performance of the multi phase meter. These methods are summarized in Table 2. The method that proved accurate enough as a reference measurement for the multiphase meter performance verification is shown schematically in Figure 1. In the process shown in Figure 1, the produced emulsion was cooled before entering the pressure tank which acts as a separator. The fluid is then dumped into the atmospheric tank where pressure transducers were utilized to provide head pressure. The head pressure was then converted to volume of liquid using density of oil and water as well as the water cut data from the manual and or automatic sampling. These measurements could then be converted to flow rates that were used to verifY the volumetric reading generated by the multi phase meter.

Table 2 - Various Methods Used to Get Reference Fluid Flow Rate Data

Method Measurement Difficulties/Inaccuracy
Test Separator Foaming, water salinity and fluid density changes affect WC and mass measurement meter. High viscosity requires high retention time for seoaration
Truck Unloadking Tickets Same as above with WC and volumetric measurements. Batch process, not unifonn due to human involvement.
Atmospheric Tank Gas foaming and flashing of water into steam which caused foaming. Error with float gauge measurements
Atmospheric Tank and Cooler Cooling corrected tlashing but foaming still occurs due to gas, defoamer not effective
Pressure Tank -Sight Glass Hard to read the sight glass with foaming. Small volume of the tank limits the length of the well test -short well tests
Atmospheric Tank plus a pair of differential pressure transmitters Pair of Delta P devices to measure liquid hydrostatic pressure and density, detennine the liquid volume. If foam level is below device I, measurement requires WC and density values for water and oil

A multi phase meter manufactured by Agar Corporation (MPFM 30 I -20) was installed at the location marked as "MPFM" in the schematic shown in Figure 1. The principle of operation of this multi phase meter is described in reference I O. The photograph in Figure 2 shows the installed multiphase metering skid. The skid shown in Figure 2 was designed to handle fluid rates up to 600 M"3/0 (3750 BPD), gas rates up to 765 AM"3/D (27,000 ACFO) and operating temperatures as high as 232°C (450 OF).

Field verification data obtained at Site A indicated that the multi phase meter can perform under the high temperature operating conditions shown in Table I for Site A and provided reliable flow rate and water cut data as shown in Figures 3 and 4. The accuracy for the multi phase meter, as specified by the vendor, used in these field tests is outlined by the dotted line in Figure 3. Since the absolute accuracy of tank tests, used as the reference in the plot can not be determined under field conditions, one can only state that the accuracy of the multiphase meter is as good as the "best" tank test procedures that can be attained in practical field conditions.

Watercut verification - especially at high water cuts, using manual sampling, is always diffcult due to inherent problems associated with obtaining a representative sample and analysis. The manual samples in Figure 4 were taken from horizontal line and a 2-inch tee. Getting a representative sample in high water cut flow streams is highly dependant on the flow condition and how tight the emulsion is. We believe that the composite automatic sampler provided a more representative and accurate water cut data under the conditions at this site, although the hardware required some maintenance.

It should be noted that to obtain the quality of flow rate and water cut data from conventional measurement methods in the above field tests, the operator had to go through considerable amount of effort and equipment modifications as well as revisions to measurement practices. This level of attention is not normally afforded to measurement and well testing activities in the field.

Using the multiphase meter, the same quality of information was made available to the field on-line without the need for all the equipment modifications and personnel intervention, to perform the tank testing and manual/automatic water cut sampling. The justification for using the multi phase meter in a project should consider these factors as well as the accuracy of the data and capital cost and maintenance of the equipment(6).

A 12 well field testing campaign was conducted after the field qualification tests to access the capability of the multiphase metering skid to handle different wells. The results are shown in Figure 5. If one assumes that the tank test are the best reference measurements that can be practically obtained in the field, the multi phase meter data is in good agreement with the tank tests. The results form historical well tests on these wells are also plotted in Figure 5. The historical data, based on field records, was obtained from conventional test separator and production measurements

For this site the allocation factor obtained by comparing the total production from the reference tank tests vs. multi phase meter is 0.97, where as the allocation factor for the historical field data vs. tank test is i .40. This improvement in allocation factor, through the use of multiphase metering was noted in other applications (6, 7, and 9) and is an important measurement parameter for monitoring production processes.

Field Testing of Multiphase Meter - Site B

As noted in Table i, the flow and fluid conditions at Site B required well tests measurements to be conducted at higher level of gas volume fraction (GVF). This site had 2 wells (pairs). Well i produced a stream with 98% gas volume fraction containing saturated steam. Well 2 produced a stream with a gas volume fraction of 30% containing undersaturated compressed water. Since high GVF is one of the major parameters affecting the accuracy of measurements in the multiphase meters (11, i 2), an Agar type 401-20-40 meter was used to handle the range of gas volume fractions encountered at Site B.

The skid installed at Site B is shown in Figure 6. This skid has the capability to handle 600 MA31D (3750 BPO) of liquid and 2 i 200 AMA3/0 (750,000 ACFO) operating at 232 °C (450 OF). The principle of operation for of this multiphase meter is covered in references 14 and 15.

The hardware in the Agar 401-20-40 multi phase meter was modified to withstand the high temperatures expected in this service. The software and measurement algorithm were modified by the vendor to account for the very large portion of the gas phase being steam rather than hydrocarbon.

A 30 day well testing campaign was conducted at Site B on 2 pairs of SAGO wells. Figure 7 shows the test layout at Site B. For this test campaign, there was no reference tank test data available as was at Site A. Well rate data was compared with the output of a Coriolis meter that measured the emulsion flow rate at the liquid leg of the test separator and the flow rates of bitumen and water from HTS#2 separator as shown in figure 7. The Bitumen/oil dump flow rates at HTS#2 were measured using a Coriolis meter and inline water cut monitoring device to obtain WC correction. The water dump flow rates at the HTS#2 were measured with a turbine meter. Admittedly the measurements from all these devices have inaccuracy levels that introduce uncertainty in the data that is used as "the reference". Since the absolute accuracy of oil and water measurement from the separator dump is not available, it is impossible to claim absolute accuracy of the well rate data from multiphase meter under these field conditions. Fortunately from tank tests conducted at Site A (Figures 3 and 4), we can assume that the multiphase meter can provide accurate flow rate and water cut determination.

High Temp In Heavy Oil - Fig. 2
Fig 2 - Multiphase Metering Skid Installed at Site A
High Temp In Heavy Oil - Fig. 3
Fig 3 - Fluid flow rates rrom multiphase meter and truck loading tickets are compared with tank tests.
High Temp In Heavy Oil - Fig. 4
Fig 4 - Water cut data from the multiphase meter compared with data from manual and automatic sampling
High Temp In Heavy Oil - Fig. 5
Fig 5 - Comparison of Emulsion now rate data obtained by Multiphase Meter and Tank tests from 12 wells at Site A. Historical Data from conventional measurements are also plotted
High Temp In Heavy Oil - Fig. 6
Fig 6 - Multiphase metering skid installed at Site B

Figure 8(page 7) shows the comparison of the well rate data from multi phase meter with the data from the emulsion meter. There is close agreement with the emulsion meter. Figure 9 (page8) shows the comparison with the HTS #2 separator output. As noted in Figure 9, the multiphase meter test data is consistent in trending the fluid outputs from the separator to better than 50 M"3/0.

Some amount of over reporting of oil and under reporting of water was noted - especially after the shut down on May 23 for workover operations. This could be partly due to the uncertainty introduced by the reference water cut device. This device had previously shown discrepancies with water cut from the field production data. Additionally, the water cut device was observed to flat line at 10% water cut regardless of the water cut in the oil/bitumen stream that was supposed to run at 3-5% water cut. A second source of error may have been the deposition of heavy oil components on the probes in the multi phase meter, due to equipment cooling down, during the shut down of May 27-28 for workover. We noted that the discrepancy in the water cut, which would be the cause of over reporting oil and under reporting water, was high immediately after the restart but diminished substantially as the test progressed. These issues will be the subject of further investigations.

As in the tests conducted at Site A, the ability to assign absolute accuracy level to the multiphase meter data, obtained in a field test, is limited by the quality of the reference data. In practical terms, the quality and the consistency of the measurements that was provided by this multi phase meter under very challenging field conditions is remarkable and the accuracy of the measurements are felt to be very adequate for well testing.

Conclusions

The field tests described in this paper were undertaken to assess the capability and advantages of using multiphase metering technology to overcome the shortcomings of the conventional gravity based test separators used in a number of high temperature thermal recovery operations such as CSS and SAGO. The SAGO thermal recovery process produces measurement conditions, shown in Table I, that challenge multiphase measurement technology to the limit. The hardware must be able to operate at very high temperatures. The measurement strategy and techniques must be able to handle the tight emulsion - made up of water and bitumen that have very close densities - as well as a gas phase that is made up of produced hydrocarbons and superheated steam. The superheated steam can condense to water with reduction in temperature. The superheated water can flash to steam with reduction in pressure. The performance of a multi phase meter made by Agar Corporation was evaluated under these challenging conditions and is reported in this paper. Field tests were conducted at two sites (A and B) operated by different companies. Both operators employed SAGO or a modification thereof to produce the oil bitumen. Consequently, Sites A and B had different fluid conditions as described in Table I.

At Site A, a number of different "tank gauging" procedures were tried to obtain reference flow rate to check the performance of the multiphase meter. These methods are summarized in Table 2. The method that proved accurate enough as a reference measurement for the multiphase meter performance verification is shown schematically in Figure 2. Subsequent performance tests of the multiphase meter against the tank data indicated that the accuracy of the multiphase meter is as good as the "best" tank test procedures and water cut sampling that can be attained under practical field conditions as shown in Figures 3 and 4.

A 12 well field test campaign was conducted after the initial qualification tests at Site A. The results are shown in Figure 5. The multiphase meter data is in good agreement with tank test, which are the best reference measurements that can be practically obtained in the field. The results form historical well tests on these wells are also plotted in Figure 5. The historical data is obtained from conventional test separator and production measurements. For this site, the allocation factor obtained by comparing the total production from the reference tank tests vs. multi phase meter is 0.97, where as the allocation factor for the historical field data vs. tank test is 1.40. This improvement in allocation factor, through the use of multiphase metering was noted in other applications (6, 7, and 9) and is an important measurement parameter for improving the monitoring of production processes.

A 30 day well testing campaign was conducted at Site B on 2 pairs of SAGO wells. Figure 7(page 7) shows the test layout at Site B. Figure 8 shows the comparison of the well rate data from multi phase meter with the data from the emulsion meter. There is close agreement with the emulsion meter. Figure 9 shows the comparison with the HTS #2 separator output as shown in Figure 7.

As in the tests conducted at Site A, as well as other field tests (6, 7, 8, 10, and 13), the ability to assign absolute accuracy level to the multi phase meter data, obtained in the field, is limited by the quality of the reference field data. Figure 9 shows that the multi phase meter test data is consistent and trends the fluid outputs from the separator to better than 50 M"3/D. Some amount of over reporting of oil and under reporting of water was noted - especially after the shut down on May 23 for workover operations. This could be partly due to the uncertainty introduced by the reference water cut device. This device had been known previously to have discrepancies with water cut from the field production data. A second source of error may have been the deposition of heavy oil components on the probes in the multiphase meter during the shut down of May 27-28 for workover. Having conducted the qualification tests at Site A (Figures 3 and 4), one can reasonably assume that multiphase meter is reporting well rate data that is close to the actual values at Site B.

Field verification data obtained at Sites A and B indicated that the multi phase meter used in these field tests can perform well under the challenging fluid and high temperature operating conditions shown in Table i and provided reliable flow rate and water cut data. The multi phase meter provided consistent measurements. The accuracy of the measurements matched the level of accuracy that can be attained with the very rigorous tank measurements, shown in Figure i, and was felt to be very suitable for well testing. In addition, the multiphase meter provided online data as shown in Figure 10 (page 8) that can be used for production and well diagnostics (6). The use of multi phase meters should therefore be considered as a viable and improved measurement alternative to conventional test separators for well testing under SAGO process conditions.

It should be also noted that to obtain the quality of flow rate and water cut data from conventional measurement methods in the above field tests, the operators had to go through considerable amounts of effort and equipment modifications as well as revisions to measurement practices as shown in Table 2 and Figure i. Normally this level of attention can not be afforded to measurement and well testing activities in the field. Using the multiphase meter, the same, or perhaps better, quality of information was made available to the field on-line without the need for all the equipment modifications and extra personnel interventions, to perform the tank testing and manual/automatic water cut sampling. The justification for using the multi phase meter in a project should consider these factors as well as the accuracy of the data and capital cost and maintenance (6) of the equipment.

Acknowledgement

The author would like to acknowledge the input and technical discussions provided by Michael Olugan, Long Lake Project, NEXEN, in preparing this paper. The author also wishes to acknowledge the technical support provided by personnel from Zirco and Agar Corporation during the field tests.

References

1. Kannaker K. and Maini Brij B. "Applicability of Vapor Extraction Process to Viscous Oil Reservoirs" SPE 84034 presented at SPE A TCE, Denver, Colorado, 5-8 October, 2003.

2. Hong K.C. "Recent Advances in Steamflood Technology" SPE 54078 presented at International Thennal Operations and Heavy Oil Symposium, Bakersfield, California, 17-19 March 1999.

3. Nasser T.N.; Golbeck H.; Korpany G. "SAGD Operating Strategies" SPE 50411 presented at International Conference on Horizontal Well Technology, Calgary, Canada, 1-4 November 1998.

4. Scott G.R. "Comparison of Cyclic Steam Stimulation and Steam Assisted Gravity Drainage" 1. Pet. Tech. (June 2003):68-69.

5. Wang S., Axcell E. : et al. '"Application of a Reverse - Emulsion Breaker at SAGD Pilot Plant in Northern Alberta" SPE 86932 presented at International Thennal Operations and Heavy Oil Symposium, Bakersfield, California, 16-18 March 2004.

6. Bertolin L., Mehdizadeh P., and Stobie G.," Petrozuata - An Application of Multiphase Metering Technology" SPE paper 89870 presented at the SPE- A TCE 2004, Houston, TX, Sept. 28-29, 2004.

7. Means S.R. and Mehdizadeh P. "New Technology Improves Well Testing Units" Oil&Gas Journal, Oct. 30, 2000.

8. Mehdizadeh P., Tuss B., "Field Test of High Viscosity Multiphase Meter" paper presented at the IPC 96, Calgary, June 10-12, 1996.

9. Buell R.S. and Turnipseed S.P., "Application of Lean Six Sigma in Oilfield Operations" SPE 84434 paper presented at SPE-A TCE 2003, Denver, Colorado, 5-8 October 2003.

10. Colmenares, J., et aI, "Venezuelan Experience in Multiphase Pumping and Metering Technologies: Heavy Oil Field Applications - Present and Future", Multiphase Measurement & Production Testing User Roundtable, May 3, 2000, Houston, TX

11. Mehdizadeh P. and Farchy D. - "Multiphase Flow Metering Using Dissimilar Flow Sensors -Theory and Field trials Results" SPE 29847 presented at the SPE Middle East Oil Show, Bahrain, 1995.

12. Wiliamson J., and Mehdizadeh P., "Alaska Regulatory Guidelines for Qualification of Multiphase Metering Systems for Well Testing" SPE paper 97240 presented at SPE Western Regional Meeting held in Irvine, CA, U.S.A., 30 March - I April 2005.

13. Hasebe B., Hall A, Smith B, Brady J., and Mehdizadeh P., " Field Qualification of Four Multiphase Flowmeters on North Slope, Alaska "SPE paper 90037 presented at the SPE- A TCE 2004, Houston, TX, Sept. 28-29, 2004.

14. Tuss, Bernie, et aI, "Field Tests of the High Gas Volume Fraction Multiphase Meter", SPE 36594, 1996 SPE Annual Technical Conference and Exhibition, October 6-9, 1996, Denver, CO.

15. Ngai C.c., Brown M.D., and Mehdizadeh P. - "Perfonnance Test of a High Gas Volume fraction Multiphase Meter in a Producing Field" SPE 38784 presented at the 1997 SPE Annual Technical Conf. & Exhibition in San Antonio, TX.

High Temp In Heavy Oil - Fig. 7
Fig 7 - Schematic of the process train at Site B. The location and devices used to measure emulsion, bitumen oil, and water that are used as the reference to compare with the multiphase meter data are shown in the schematic. Note that at the process conditions of high temperature separator, the produced water is less dense than the bitumen. Wet bitumen/oil exits from the bottom of the separator.
High Temp In Heavy Oil - Fig. 7
Fig 8 - Comparison of the emulsion flow rate from multphase meter with emulsion flow rate from the Coriolis meter located at the liquid leg of the Inlet Test Separator shown in Figure 8. There was very good agreement between the emulsion flow rate data from the multiphase meter and the Inlet Test Separator.
High Temp In Heavy Oil - Fig. 7
Fig 9 - Comparison of well mte data form multiphase meter with the output from the field separator (Tr 2 Bitumen and Water Dump in Figure 8). Well 2 was on multiphase meter test during the May 21-22 period and again during the June 7-8. Weill was on multiphase meter test during the May 25 to June 6. There were 2 shut downs during the field tests on May 24 and 27 for workovcr operations. The upper graph shows the water dump data and the water rates (points) as determined by the multiphase meter. The lower graph shows bitumen dump rates and oil rates (points) as determined by the multiphase metcr.
High Temp In Heavy Oil - Fig. 7
Fig 10 - Comparison of the oil (bitumen) now rates from the multiphase meter with the output of the production separator. The online capabilty of the multiphase meter to trend well production can be used for production and well diagnostics as was reported in reference 6.
Addresses