Application of Computer Technology to Reservoir Studies in Canada

Abstract

Abstract Three-dimensional block models of petroleum reservoirs can be mathematicallyconstructed and solved by high-speed digital computers. Machine programs havebeen written which will solve a reservoir model consisting of more than 5,000 blocks. This permits the effect of areal variations of thickness, porosity and permeability to be studied in much more detail than was formerly possible. Themodels are particularly applicable in pressure maintenance and secondary recovery studies, but are also useful for optimizing well spacing during development operations. Introduction In the oil-producing areas of Canada, the use of mathematical model studies performed by high-speed computers is feasible for more major fields and wouldproduce more advantageous information than in any comparable oil-producing areaof the world today. This is primarily because the Canadian oil industry is ayoung one; most fields have been developed since the importance of applyingreservoir engineering principles to production was realized. Also, thedevelopment has taken place during a period when advanced technology permitted the acquisition of adequate data on reservoir and formation characteristics.That is, diamond coring and methods of regular and special core analysis were available, and modern logging methods, with logs which would give quantitative information, were offered. During the major development period, members of both industry and the Conservation Board recognized the need for obtaining sufficient reservoir data, not only from cores and logs but also from performance, so that well production tests, pressure surveys and build-up or drawdown data are usually available. Also, the importance of early pressure maintenance projects to obtain maximum recovery, and cooperation in the unitization of fields to obtain this end, has been well recognized. Maximum use of the available data for reservoir performance studies and predictions can be made today through numerical models designed for modern computers. We have used these models successfully in a number of large fields, and their value has been proved.