A New Approach to P/Z Analysis in Abnormally Pressured Reservoirs

Abstract

Abstract Future performance of gas reservoirs can be predicted if one identifies the main drive mechanisms and the original gas-in place. The extrapolation of the p/z versus cumulative gas production plot is a commonly accepted method that is used to estimate original gas in-place in volumetric gas reservoirs. However, abnormally pressured reservoirs usually exhibit initial pressure gradients in excess of 0.65 psi/pt at depths greater than 10,000 ft. As a result, in addition to fluid expansion, gas is produced by sand grain expansion, rock compaction and water expansion. As formation compressibility approaches gas compressibility, the standard p/z curve tends to bend downward, making "straight line extrapolation" questionable. To account for this compaction phenomenon, some authors suggested to modify the material balance equation by incorporating constant pore and water compressibilities. Unfortunately, by assuming a constant pore compressibility, the original gas in-place can be overestimated by as much as 60%. These errors are caused by pore compressibilities continually changing as fluids are withdrawn and as pressures decline to a normal range. To overcome this shortcoming, we propose to rigorously account for the variation of pore compressibility with pressure. Assuming that pore compressibility as a function of pressure is available, a straight-line method which allows estimation of original gas in-place is presented. If pore compressibility is not readily available, we show that by adopting simplified models for pore compressibility versus pressure, we can regress on measured gas production versus reservoir pressure data to obtain initial gas in-place and parameters for pore compressibility model. Regression is performed using a public-domain Levenberg-Marquardt algorithm, and the model parameters are constrained using simple previously-published algorithms. We demonstrate the validity of our approach by considering synthetic data from a finite difference simulator. We also successfully apply our analysis to data from two abnormally-pressured fields. Introduction The observed behavior of abnormally pressured reservoirs has been discussed in the literature. P. 249