Thin Bed Information Resolution of Well-logs And Its Applications

Yongxian Qian and Xingshui Zhong: Department of Geophysical Exploration, Jianghan Petroleum University

Abstract: We develop the theoretical potential of well logs for distinguishing thin beds, and introduce a new vertical resolution concept for logs, thin bed information resolution (TBIR). This is developed by combining Walsh function theory, linear systems theory, and logging instrument response functions. The relationship between minimum bed thickness and Walsh functions is established. Two definitions and three theorems are introduced to support the relationship. TBIR is defined and explained using two examples, and is then applied to enhance the vertical resolution of a sonic interval transit time log and an induction conductivity log. The results show that both logs have information resolution of 0.25m when the sampling interval in depth is 0.125m. TBIR is higher than the vertical resolution as usually defined. Moreover, TBIR can also be helpful in the development of new logging instruments that are desired to possess both deep radial investigation and high vertical resolution.

Michael L. Cheng and Marco A. Leal: Canadian Petroleum Ltd., Calgary, Canada David McNaughton: Mincom Inc., Houston, Texas U.S.A.

Abstract: The Upper Qishn Clastics of Cretaceous age are the primary producing reservoirs in the Masila Block Development area. The underpressured, low gas-oil ratio reserves require artificial lift from initial completion. Electric submersible pumps are used to produce these reservoirs; consequently, knowledge of initial reservoir productivity is essential to the well completion designs and pump sizing. Conventional core and log evaluation methods used to predict reservoir productivity have not been reliable because of variation in reservoir quality and facies changes between pools. A simple and cost effective prediction technique using commonly available open-hole log data has been developed. The method uses log-derived normalized resistivity ratios (Rn = log {(Rt/Rw)/(Rxo/Rmf)}), that characterize reservoir fluid mobility, and predict well productivity indices (PIs). The correlation was developed using well test data from 20 oil bearing zones in 9 wells, and is routinely applied to predict initial PIs in new development wells. The method has been proven effective over four years of field development and production in the Masila Block. The Rn technique is a natural extension of the conventional moved-oil plot method in log analysis that is used to infer zones of maximum permeability and movable hydrocarbon. The model is simple and appears to be grain size independent. Further, the technique does not require complex petrophysical and geological analysis; it utilizes data sets (i.e., dual laterolog with micro-spherically focused resistivity devices) that are consistent between wells.