Samuel S. MOFUNLEWI, Joseph A. AJIENKA and D. APPAH
Department of Petroleum & Gas Engineering, University of Port Harcourt, Nigeria
Emails: Samlewi@engineer.com, jajienka@yahoo.com, dappah@yahoo.co.uk
The aim of field testing of Multiphase Flow Meter (MPFM) is to show whether its accuracy compares favourably with that of the Test Separator in accurately measuring the three production phases (oil, gas and water) as well as determining meter reliability in field environment. This study evaluates field test results of the MPFM as compared to reference conventional test separators. Generally, results show that MPFM compares favourably with Test Separator within the specified range of accuracy.
At the moment, there is no legislation for meter proving technique for MPFM. However, this study has developed calibration charts that can be used to correct and improve meter accuracy.
Multiphase Flow Meter (MPFM); Test Separator (TS); Correction Charts; Accuracy; Water Liquid Ration (WLR); Oil; Gas.
Until recently, large expensive test separators have been used to separate the oil, gas and water which are then measured using conventional technology [1, 2].
Multiphase meter is a device that can be used to measure individual fluid flow rates of oil and gas when more than one fluid is flowing through a pipeline. A multiphase meter provides accurate readings even when different flow regimes are present in the multiphase flow [3]. When using singlephase meters, the fluid mixture (oil and gas) coming from the wellbore must pass through a fluidseparation stage (separator) prior metering. Otherwise, the readings of the singlephase meters will be inaccurate. Separators are not necessary for multiphase metering, and the meters can support different proportions of gas and oil. Multiphase meters provide the advantage of continuous well monitoring, which is not possible using singlephase meters. Additionally, multiphase meters cost less, weigh less and require less space. Multiphase meters are more common in deepwater operations, where wellintervention operations are often prohibitively expensive.
The problem now arises as to whether the accuracy of multiphase meter (MPFM) compare well with that of test separator. How can the MPFM accuracy be improved? This paper proposes solutions to these probes.
The test reference loop consists of a threephase separator. Gas and liquid are separated in the test separator. In order to achieve the desired steady water cut concentration, the oil/water volume in the separator and different draw points are adjusted. On separation, the liquid is pumped through a liquid measurement line. In this line, volumetric measurement is performed with PD meters and water cut measurement is performed with the oil/water meter. A vortex meter and rotameters are used to measure gas after compression.
Following the separation is a recombination of gas and liquid phases. On recombination, the combined stream then passes through the multiphase meter and measurement taken accordingly. Figure 1 shows the diagram of the test reference loop [4].
Below is a list of the procedure for the main testing of MPFM [5]. Test separator is validated as a reference to the multiphase flow meter.
Figure 1. Schematic of test reference loop. Adapted from Tests at Agar Corporation (1999)
1. A purge is time is assigned to each well to be tested.
2. Data review from test separator to ensure that a steady production condition is attained before starting test.
3. Data collection from the MPFM and test separator at the same time when test starts.
4. Initial test result is reviewed.
5. Enquiry from vendor for any modification for improvement of meter performance was made
6. Validity of all data collected with a test separator. This involves comparing test results with historical data.
7. Test is invalidated if major discrepancies are observed.
8. MPFM inputs were reviewed.
9. A final list of valid comparison tests was prepared.
10. Crossplots of MPFM against test separator were produced.
Process and performance specifications are given in Tables 1 and 2, respectively.
Description 
Process Conditions 
Watercut 
10 – 90% 
GVF 
24 – 85% 
Liquid flow rate 
1,000 – 5,000 BPD 
Salinity of water 
1.5% by weight 
Oil viscosity 
360cp 
Temperature 
40^{o}C 
Description 
Specification 
Liquid (oil and water) 
± 5% 
Crude Oil 
± 5% 
Water 
± 5% 
Gas 
± 5% 
The formula below was then used to compute accuracies in each case from the total flow rate and total deviations.
_{} 
(1) 
The test result summary is presented in Table 3 for clarity.
No. 
Reference Measurements (BPD) 
Test Measurement (BPD) 
Deviations 


Oil 
Water 
Gas (ACFD) 
Liquid 
WLR 
Oil 
Water 
Gas (ACFD) 
Liquid 
WLR 
Oil 
Water 
Gas 
Liquid 
WLR 
1 
1054 
129 
40229 
1183 
0.109 
1128 
84 
39954 
1212 
0.069 
74 
45 
275 
29 
0.04 
2 
2701 
363 
20080 
3064 
0.118 
2819 
305 
19026 
3124 
0.098 
118 
58 
1054 
60 
0.02 
3 
3276 
433 
6835 
3709 
0.117 
3382 
363 
5427 
3745 
0.097 
106 
70 
1408 
36 
0.02 
4 
786 
701 
47241 
1487 
0.471 
730 
729 
47376 
1459 
0.5 
56 
28 
135 
28 
0.029 
5 
2373 
2575 
30098 
4948 
0.521 
2480 
2488 
28839 
4968 
0.501 
107 
87 
1259 
20 
0.019 
6 
180 
1319 
48773 
1498 
0.881 
218 
1204 
50196 
1423 
0.846 
38 
115 
1423 
75 
0.035 
7 
487 
4504 
29236 
4991 
0.902 
661 
4493 
27048 
5154 
0.872 
174 
11 
2188 
168 
0.03 
Total 
10857 
10024 
222492 
20880 
3.119 
11418 
9666 
217866 
21805 
2.983 
561 
358 
4626 
205 
0.135 
The following results in Table 4 were obtained using equation 1 above.
It can be inferred from the results in Table 4 that the MPFM compare well with test separator. The percent accuracy falls within the specifications in Table 2. This means that the overall performance of the meter was excellent.
Table 4. Percent accuracy of Oil, Water, Gas, Liquid and Watercut
Description 
Oil 
Water 
Gas 
Liquid 
Watercut 
Accuracy 
5.17% 
3.57% 
2.08% 
0.98% 
4.34% 
The correction charts below are developed from the test result summary (Table 4). They are to be used for improving meter accuracy. These charts are developed by selecting the best trend line for oil, gas water, liquid and water liquid ration (WLR) separately. They are presented in Figures 2 – 6. Descriptive statistics are also presented.
Correction Charts for Oil Rate
Figure 2. Cross plot for oil rate for test separator versus oil rate for MPFM
Table 5a. Descriptive Statistics
Regression Statistics 

R Square 
0.9969 
Standard Errors 
73.90 
Observations 
7 
Table 5b. Descriptive Statistics

Coefficients 
Standard Error 
t Stat 
PValue 
Lower 95% 
Upper 95% 
Intercept 
43.39 
48.58 
0.89 
0.41 
163.26 
81.48 
X Variable 
0.98 
0.02 
40.12 
1.81E07 
0.91 
1.04 
Correction Charts for Water Rate
Figure 3. Cross plot for water rate for test separator versus water rate for MPFM
Table 6a. Descriptive Statistics
Regression Statistics 

R Square 
0.9996 
Standard Errors 
51.8 
Observations 
7 
Table 6b. Descriptive Statistics

Coefficients 
Standard Error 
t Stat 
PValue 
Lower 95% 
Upper 95% 
Intercept 
56.90 
26.81 
2.12 
0.087 
12.02 
125.82 
X Variable 
1.00 
0.013 
75.06 
7.95E09 
0.96 
1.02 
Correction Charts for Liquid Rate
Figure 4. Cross plot for liquid rate for test separator versus liquid rate for MPFM
Table 7a. Descriptive Statistics
Regression Statistics 

R Square 
0.999 
Standard Errors 
57.6 
Observations 
7 
Table 7b. Descriptive Statistics

Coefficients 
Standard Error 
t Stat 
PValue 
Lower 95% 
Upper 95% 
Intercept 
64.17 
47.26 
1.36 
0.23 
57.31 
185.65 
X Variable 
0.97 
0.014 
69.6 
1.16E08 
0.93 
1.00 
Correction Charts for Gas Rate
Figure 5. Cross plot for gas rate for test separator versus gas rate for MPFM
Table 8a. Descriptive Statistics
Regression Statistics 

R Square 
0.997 
Standard Errors 
839.36 
Observations 
7 
Table 8b. Descriptive Statistics

Coefficients 
Standard Error 
t Stat 
PValue 
Lower 95% 
Upper 95% 
Intercept 
2438.41 
738.53 
3.30 
0.021 
539.96 
4336.86 
X Variable 
0.94 
0.021 
44.00 
1.14E07 
0.89 
1.00 
Correction Charts for WLR Rate
Figure 6. Cross plot of WLR for test separator versus WLR rate for MPFM
Table 9a. Descriptive Statistics
Regression Statistics 

R Square 
0.996 
Standard Errors 
0.025 
Observations 
7 
Table 9b. Descriptive Statistics

Coefficients 
Standard Error 
t Stat 
PValue 
Lower 95% 
Upper 95% 
Intercept 
0.019 
0.016 
1.24 
0.27 
0.021 
0.06 
X Variable 
1.00 
0.029 
34.18 
4.03E07 
0.92 
1.07 
The plots in Figures 2 – 6 can be used to predict what the rate (oil, gas, water, liquid or WLR) of a MPFM will be if that of test separator is known.
For example, if the oil rate for test separator is 2000BPD, then the predicted value of MPFM will 2200BPD. Also, if the gas rate for test separator is 1000ACFD, the predicted MPFM rate will be 800ACFD. The closer the value of R^{2} is to unity (1), the better. For rates that fall outside those presented in the charts above, the corresponding correlations can be used to determine the predicted values. That is if the value of test separator is know, make substitution into the appropriate equation to get the corresponding value of MPFM. For example, if the test separator rate for liquid is 10,000BPD, it will be better to substitue into the liquid rate equation to obtain the value for MPFM. Doing this, we will get 10254BPD.
The equations, R^{2} and P values are summarised below:
Table 10. Equations and R^{2} values for different rates
S/No. 
Description 
Equation 
R^{2} Value 
PValue 
1 
Oil rate 
y = 0.9775x – 43.388 
0.9969 
0.412 
2 
Water rate 
y = 0.9958x + 56.899 
0.9991 
0.087 
3 
Liquid rate 
y = 0.969x + 64.169 
0.999 
0.233 
4 
Gas rate 
y = 0.9429x + 2438.4 
0.9974 
0.0214 
5 
WLR rate 
y = x + 0.0194 
0.9957 
0.269 
This study has been able to show that the MPFM accuracy compare favourable with that of test separators. Hence, due to the economic benefits and the dependability of its accuracy, it is important to spread the expertise in MPFM through the oil industry. Both field and laboratory testing should be conducted to determine meter accuracy for added confidence.
Also, the correction charts developed in this study are useful tools for predicting the values of MPFM fluid flow rates when the flow rates of test separators are known. However, the charts are limited to the ranges shown on them. For fluid flow rates outside those obtainable on the charts, the equations developed are recommended for use.
1. Mofunlewi S. S., Ajienka J. A., Economic Evaluation of Multphase Meters, Leonardo J. Sci, 2007, 11, p. 2.
2. Mofunlewi S. S., Evaluating the Efficiency of Multiphase Meters, M. Eng. Thesis, Department of Petroleum & Gas Engineering, University of Port Harcourt, Port Harcourt, Nigeria, 2003, p. 2.
3. Schlumberger Oilfield Glossary, Multiphase Meter, http://www.glossary.oilfield.slb.com/Display.cfm?Term=multiphase%20meter, 2008, ID 11235.
4. Sonatrach, Anardarko, Lasmo, Quobba, Multiphase Meter and Watercut Meter Test, Tests at Agar Corporation, Houston, Texas, 1999, p. 13.
5. AlTaweel A. B., Barlow S. G., FieldTesting of Multiphase Meters, Saudi Aramco Journal of Technology, http://www.saudiaramco.com/irj/go/km/docs// SaudiAramcoPublic/Publications/EN/Journal%20of%20Technology/Spring2000/p50to59.pdf, Spring, 2000, p. 5059.