Well test analysis and reservoir modeling of geopressured-geothermal systems (includes topical reports on analyses for: Pleasant Bayou Well No. 2, Gladys McCall Well No.1). Final report

摘要

Automated inversion of pressure transient data from the Pleasant Bayou and Gladys McCall geopressured test wells has allowed the resolution of several outstanding questions concerning the two reservoirs. The added accuracy of the interpretation of the various data sets from each well provides convincing evidence that the formation response of each of the reservoirs behaved in an essentially linear fashion throughout its multi-year depletion/recovery test history. The only exception was in the immediate vicinity of the sandface; the skin factor in each of the wells varies during the course of its test history (see Appendices for detailed discussions of the variations and the likely causes of the variations for each well). (a) The six bottomhole pressure transient tests of Pleasant Bayou Well No.2 over a twelve year period (1980-1992) yield estimates for the permeability-thickness product that are in excellent agreement (kh = 3.665 D-m). Inversion of the 45-day buildup portion of the 1980 RLT yields an estimate of X{sub a} = 869 m as the distance from the well to the nearest boundary. The inversion of the data from the other five tests, of much shorter duration, yield estimates in the range X{sub a} = 427-503 m. The RLT estimate would normally be given much more credence. The failure, repair and resetting of the gauge after the first ten days of the start of the buildup test, however, may have affected the RLT data. Calculations are employed to estimate bottomhole pressures from surface recordings subsequent to removal of the bottomhole gauge following the 1992 70-hr buildup test. These estimated values were combined with the 70-hr data to form the 111-day (from September 12, 1992 through January 1, 1993) LTT buildup test data set. The LTT data were inverted to yield an estimate of X{sub a} = 452 m. Since this is in agreement with the range of estimates from the five short-term bottomhole test, we believe the range X{sub a} = 427-503 m to represent the best estimate for the distance from the well to the nearest boundary. The simulation model presented in Appendix B used L1 = 990 m (Figure 3 in Appendix B) as the distance from the test well to the nearest boundary. The simulation model could probably be improved using a smaller value for L1. The actual location of the boundary to the south of the well is not known from the available geological information. (b) The five bottomhole pressure transient tests of Gladys McCall Well No.2 over an eight-year period (1983-1991) yield estimates for the permeability-thickness product that differ significantly from each other. The differences, however, can be explained by the partial plugging of the near-well formation during an aborted attempt to inject a scale-inhibitor pill in May 1985, followed by subsequent scouring out the pill precipitates by the flowing fluid during the May 1985-August 1987 production testing. The partial plugging of the near-well formation in conjunction with a known shale stringer that prevents vertical communication in sand 8 (Figure 2 in Appendix A) would reduce the apparent value of kh to about sixty percent of the original RLT value of 13.44 D-m. Inversion of the 79-hr (1985) and 92-hr (1986) buildup tests yield kh estimates of 7.69 and 8.86 D-m, respectively. Inversions of the valid MRT (1987) and FDT (1991) pressure transient data yield estimates very near the original RLT value. The reinterpretation of the MRT data represents a correction of our earlier conclusion (in Appendix A); we now believe that the data from the first and second gauges employed during the MRT are incompatible. Data from the first (''old'') gauge during the drawdown portion of the MRT could not be adequately corrected for the observed drift in the recorded data. Only the second (''new'') gauge data are considered reliable and inversion of the buildup data it recorded yields the estimate for kh that is in agreement with estimates obtained from inversion of the RLT and FDT pressure transient data.