Decline and Forecast Studies Based on Performances of Selected California Oilfields

Abstract

Abstract A study was made of the production decline carves of the oil produced from 347 zones in California reservoirs. These fields have been operating at the maximum efficient rate for many years. The best decline curve using the least-squares technique for the linear, semilog, log-log and hyperbolic relationships was tested to determine the best relationship. Such a thorough testing of the data was possible because of the use of the electronic computer. Using percentiles, the study included a comparison of the slopes of decline curves with such variables as areal extent, feet of pay, depth, acre-feet, and API gravity and viscosity of the oil in each zone. Introduction Over the years many different types of decline curves have been used by petroleum engineers to estimate the performance and the future oil production that may be obtained from oil fields. These curves have been used for making long-term forecasts. The purpose of this study was to use production decline curves to estimate the potential production if an emergency should arise whereby the large foreign supply would be interrupted for a period from two to five years or longer, which would require higher domestic production from prorated fields. This paper presents the results of a computer analysis of the composite effect of important factors influencing the decline performance of 347 California reservoirs which have been produced virtually at the maximum efficient rate for many years. The performance of these reservoirs provided an excellent source to test the merits of different decline equations for short-term forecasting. They could also be used for longer-term forecasts if interest demanded. Yearly oil production, reserves and reservoir data were conveniently available for the period 1946 through 1961. Thus the maximum production period for the study was 15 years, and no reservoir was selected that had a production history of less than five years. The relationships between production and the related variables were programmed using the least-squares technique. This was done for the linear, semilog, log-log and one hyperbolic relationship. Arps' decline relationship was used, but the characteristics of the equation are such that even with double precision (16 significant figures) some zones present computer program debugging problems that have not been completely eliminated. First the linear relationship was tried, then the classical semilog, the log-log and the hyperbolic. Other relationships are being tested. The study pertained not only to the use of conventional decline curves which treat the production-time data as empirical points, but also to the application of as many fluid flow relationships as the quantity and quality of the additional field data permitted. The results of the use of some fluid flow relationships are presented in this report and others are in progress. In the event of an immediate need for more domestic production, oil from prorated fields would be produced at the maximum efficient rate. Several methods of forecasting the performance of these fields were tried. By grouping the performances of California fields into quantiles, a relationship was discovered that would have predicted the performances of California fields had they been prorated at any time. The inference is that the same principles would be applicable to prorated fields in other states. Because of the variable nature of the field data, all the results in this report are applicable only to the average performances of groups of reservoirs. In the study the calculations were made for each zone and the results were then grouped according to the parameter to be studied. Basic Equations The general form of the equations used when treating the production- time data as empirical points on a graph is: Linear,(1) Semilog,(2) or(2a) Log-Log,(3) (3a) Hyperbolic,(4) The polynomial,(5) was not used because the end portion dominates the extrapolation. JPT P. 959ˆ