An Empirical Approach to Waterflood Predictions

Abstract

This prediction approach is based on an analysis of the actual performance of eight waterflood projects in Southern California. With the charts resulting from this analysis, such waterflood parameters as time to initial response and peak production rate can be related to the fluid and rock properties. The method permits the prediction of maximum, average, and properties. The method permits the prediction of maximum, average, and minimum recovery. Introduction The actual performance of a waterflood seldom matches the theoretically predicted behavior. The main reason for the deviation is that it is virtually impossible and impractical to incorporate all the variables affecting the performance of a waterflood into a theoretical prediction technique. Some of the factors contributing to the deviation are:the validity of the basic geologic, reservoir and production data, and allowance for the type of reservoir drive mechanism; andoperating problems such as well failures, poor injection profiles, thief sands, early water breakthroughs, degree of water-oil ratio (WOR) control and failure to keep wells pumped off. Empirical waterflood prediction techniques, on the other hand, inherently take all the parameters into account. However, these techniques may be geographically or geologically limited. A number of empirical prediction methods have appeared in the literature. Guthrie and Greenberger have published a method based on a correlation between oil recovery by waterflooding and fluid and rock properties using the data published by Craze and properties using the data published by Craze and Buckley on 73 sandstone reservoirs. Guerrero and Earlougher have developed a convenient method for predicting the performance of waterfloods, using data on Mid-Continent projects. Their method is strictly based on an average performance of the flood. performance of the flood. Bush and Helander have published an empirical method based on a statistical analysis of 86 successful Oklahoma floods. Very useful generalized rules of thumb for predicting waterflood performance were developed. Also, the authors have published qualitative comments on the effect of reservoir heterogeneity, fillup pore volume, and well spacing on the various flood parameters. However, they do not specifically relate the fluid and rock properties to the values of these parameters. properties to the values of these parameters. The approach used for the development of the Empirical Correlation Method (ECM) relies on dividing the flood performance into time periods similar to the ones used by Bush and Helander (Fig. 1). However, the previous work is extended to include a quantitative measure of the effect of fluid and rock properties on the performance of a flood. The following properties on the performance of a flood. The following reservoir and rock properties were found to have statistically significant influence on flood behavior:permeability variations;oil and gas saturation at the start of the flood;oil-water viscosity ratio;injection rate; andan average distance from first-line producers. Development of ECM The ECM is based on a statistical analysis of actual waterflood performance of eight Southern California floods. Environmental, reservoir, and production data for these floods are summarized in Tables 1 through 3. JPT P. 565