Application of a Mass Flowmeter for Allocation Measurement of Crude Oil Production

Abstract

Summary Mass flowmeters have been used to improve crude oil measurement atindividual wells throughout the Little Knife field. This type of meter uses theprinciple of the Coriolis effect to measure the mass flow rate of fluid. Massmeasurement eliminates problems encountered by conventional volumetricflowmeters because of the presence of free gas in the liquid phase. After 5 years of operation, mass flowmeters have proved to be a significantimprovement over conventional flowmeters. Excellent agreement has beenconsistently achieved between the sum of the measurement from the allocationmeters and the custody transfer meters, which receive the commingled oil. Inaddition to the performance improvement, significant capital and operating costsavings were realized compared with a more conventional oilfield approach. Introduction Accurate measurement of crude oil production from individual wells is neededfor calculating royalty and working-interest payment, guiding daily productionoperation, and managing the reservoir. In some instances, this information isalso used to evaluate EOR projects. Measurement philosophy and equipmentarrangement often differ widely, depending on the operator, productionconditions, contractual requirements, and government regulations. The Little Knife field, located in western North Dakota, consists of 89wells that produce a total of about 10,000 B/D [1590 m 3/d] of crudeoil and 30×106 scf/D [0.85×106 and m 3/d] ofnatural gas. The production wells are grouped into three central tankbatteries, each with a lease automatic custody transfer (LACT) unit to measurecrude oil sales. The gravity of the crude oil ranges from 38 to 42°API [0.83 to0.82 g/cm3], with an average GOR of 3,000 scf/bbl [540 stdm3/m3]. The produced gas contains about 10% H2S and 2.5% CO2. The 89 wells have different working-interest and mineral owners, thusrequiring accurate measurements at each well for working-interest and royaltycalculations. Each production well is equipped with a dedicated three-phaseseparator to separate produced gas and free water. Each separator is containedin a small enclosure maintained at 40 to 60°F [4.4 to 15.6°C] during extremelycold weather. Oil dumped from the separator, which typically contains <3%basic sediment and water (BS&W), is measured with a flowmeter. Aproportional sampler is used to obtain a BS&W correction factor. Aftermetering, oil streams from a group of separators are commingled into a commonheater-treater for final BS&W separation. Oil then moves through adegassing vessel to storage tanks where it is flashed to atmospheric pressurefor removal of the remaining solution gas before sales through a LACT meter. Allocation factors, defined as the ratio of net LACT unit oil sales volume tothe sum of BS&W well production measured at the individual well separators, are then computed for each battery. In the original installation, the oil stream from the separators wasmeasured with a paddle-wheel-type volumetric flowmeter. Problems with themetering system were discovered in 1980. Of primary concern was therepeatability of individual meter factors. A review of monthly meter factorsshowed deviations >2%. An extensive study concluded that three majorfactors affect measurement accuracy: presence of solution gas and/or entrainedgas; intermittent, cyclic flow; and inadequate meter-proving practices. Because the volatile crude oil was maintained at its equilibrium vaporpressure in the separator, pressure drops in the flow piping and in the meteritself allowed significant amounts of solution gas to break out. In addition, entrained gas, resulting from incomplete gas/liquid separation in theseparator, was suspected to be present in the oil stream. The presence of thegas phase severely affected the accuracy of the volumetric flowmeter, causingoverregistration of the liquid flow. The small separator dump volume resultedin a significant transient flow effect, which reduced meter accuracy. Theinherent design of the separator and level controller resulted in a dischargeof approximately 0.1 bbl [0.016 m3] of crude oil per dump cycle. Therefore, the acceleration and deceleration periods were a significant portionof the measurement period. Metering accuracy was hampered because the effect ofcyclic flow could not be taken into account properly. The master meter used inmeter proving was the same type of meter as used on each separator and wassubject to the same transient flow problems and even more free gas than themeter being proved. Two approaches were considered to improve measurement accuracy. The firstapproach entailed an extensive modification of the existing equipment to"condition" the live, volatile crude oil to a "dead," stable, pipeline-qualityoil before it was metered. This involved the addition of a heater-treater toremove more dissolved gas at an elevated temperature, a heat exchanger toreduce the temperature of the oil coming out of the treater by heat exchangewith the incoming oil stream, and a turbine meter to measure the stabilizedoil. With these treatment facilities, the crude oil could then be metered at apressure above its vapor pressure. A larger separator would also be used toincrease the dump volume. The second approach involved a direct replacement ofthe existing flowmeter with a mass flowmeter that would tolerate cyclic flowand the presence of gas. No other modifications or equipment would benecessary. The mass-flowmeter approach was estimated to cost about $1.1million, compared with $17 million for the more complicated approach. Becauseof its simplicity and cost effectiveness, the mass-flowmeter approach wasexplored in detail. Evaluation tests of several mass flowmeters weresimultaneously conducted in the laboratory and in the field. On the basis ofthe favorable test results obtained, this approach was subsequently accepted bythe State of North Dakota and other royalty owners.