Effects of flow measurement errors on oil and gas production forecasts

Abstract

Different flow meter technologies are used to monitor the output of oil and gas wells. Although flowmeter accuracy has generally improved over time, there remain substantial uncertainties, particularly in multiphase flow. These errors could potentially be greater where older meters are being used for calibration, and/or maintenance is difficult. Consequently, the associated errors with the recorded data could be out of specification in such cases. One use of the well flow data is to improve parameter estimates for important characteristics of reservoirs such as porosity and permeability. Therefore, any errors in flow measurement influence the results of a reservoir simulation and production forecasts. However, the impact of flow measurement errors on the forecast of oil and gas production has not been considered before. In this study, the effects of using out-of-specification errors on the predicted reservoir production have been investigated. As a test case, the simulated production results of a reservoir with known characteristics were considered to be the actual flow rate values. Then, two sets of data were generated by applying errors up to 5% and 10%, respectively, to the flow rates and the resulting values were used in a history matching exercise to modify the predictions of the simulations for the same reservoir with incorrect porosity and permeability parameters. The errors in the first and second sets of data were considered to be within and without the specification, respectively. The results show that when errors are within the specification, the corrected porosity and permeability values have less than a 2.2% and 2.5% error, which cause minor deviations of up to 2.3% in the production forecast. However, for the second set of data, when the errors are increased up to 5% more than the specification, the corrected porosity, permeability and production forecast deviate significantly up to 10.8 %, 10.1% and 12.4% from their respective reference values.