Improved Gas-In-Place Determination for Coal Gas Reservoirs

Abstract

Abstract The Upper Cretaceous Fruitland Formation of the San Juan Basin of Colorado and New Mexico has been a very active natural gas play in recent years. Case studies of coal gas-in-place volume reassessments have revealed that gains of up to 74% are possible based upon the use of improved analysis methods recently developed by the Gas Research Institute (GRI). The greater gas-in-place estimates were consistent with production history and provide a new perspective upon the producible coal gas resources. Coal gas-in-place volume is proportional to the reservoir drainage area and three petrophysical parameters: gross reservoir rock thickness (consisting of both coal and other rock types), average reservoir rock density, and average in-situ sorbed gas content. The GRI procedure evaluates each of these parameters with a combination of core and density log data for specific reservoirs. Quantitative errors and causes of errors in these petrophysical parameters have been determined. Errors in coal gas-in-place volume estimates are caused by:geologic structural and stratigraphic variations that disrupt the lateral continuity of coalbedsusing a too low maximum density cut-off limit value when determining gross reservoir rock thickness with density log data;by basing average reservoir rock density estimates upon "rules of thumb" or bounding rock densities;by performing gas desorption measurements at ambient surface temperature; andby basing in-situ sorbed gas content estimates on gas desorption data collected from drill cuttings rather than whole core samples. Introduction Coal seam gas reservoirs hold approximately 13 percent (134 Tscf) of U.S. natural gas resources1 and in recent years have been one of the most active natural gas plays in the U.S. There are currently about 6,300 coal seam gas wells throughout the U.S. which accounted for nearly 5 percent (858 Bscf) of annual domestic gas production during 1994. The success of coal seam gas production in the U.S. has sparked intense interest worldwide in this gas resource, particularly among several coal-rich nations in Eastern Europe and Asia. One of the keys to reliably determining the economic value of coal seam gas reserves is to accurately estimate the volume of gas-in-place. However, this critical analysis presents some unique data acquisition and interpretation challenges. A key difference between coal and conventional gas reservoirs is that, in the former, the vast majority of the gas-in-place volume is stored by physical sorption whereas in the latter the gas is stored by compression. The coal gas-in-place volume is proportional to three petrophysical parameters: the reservoir rock thickness (consisting of both coal and other rock types), the average reservoir rock density, and the average in-situ sorbed gas content. These three parameters are generally determined using data obtained from open-hole geophysical logs and core samples. Little has been published concerning the accuracy, comparability, and limitations of the most commonly used methods for determining these three crucial parameters. Two observations indicate that the petrophysical data used for calculating coal gas-in-place volume may frequently be inaccurate. First, accurately estimating the volume of gas stored in-situ by sorption requires making measurements on coal samples. However, widely different in-situ sorbed gas content estimates are obtained from different types of coal samples or data analysis methods. Recent Gas Research Institute (GRI) research has revealed that some commonly used coal sample type and data analysis methods have inherent shortcomings which result in in-situ sorbed gas content estimates that are low by 50%. P. 439