Impact of reservoir heterogeneity on inventory verification in natural gas underground storage reservoirs subject to water drive

Abstract

This thesis demonstrates how the current method used for inventory verification in natural gas underground storage fields subject to water drive can be impacted by reservoir heterogeneity. The current method of performing inventory verification assumes average (homogeneous) reservoir properties and a relatively consistent operation of the field comprising of similar injection/withdrawal volumes from year to year (Tek 1996). Throughout the text, this method will be referred to as Tek’s method. For high quality reservoirs (high permeability/porosity), the assumptions made in Tek’s method may hold true, but at higher levels of reservoir heterogeneity, these assumptions may lead to large errors in the calculated inventory, resulting in an incorrect understanding of how the storage horizon and aquifer interact. This misunderstanding could lead to large differences between how the field/reservoir is expected to perform and how it actually performs. These differences could be seen in various forms such as a loss in deliverability, increase in water production, change in the field’s pressure-content (hysteresis) plot, reduction in field peak rate, etc. The degree of difference between the field’s anticipated performance and the actual performance is directly related to the number and location of the observation wells used to perform the inventory verification. Tek’s method of inventory verification incorporates data from two observation wells, one being completed in the storage horizon to provide gas pressure, and the other being completed in the water leg to provide pressure data on the aquifer. With only two observation wells being used in the inventory verification calculations, one or both wells may be impacted by the heterogeneity of the reservoir. Many storage fields have additional observation wells that could be incorporated into the inventory verification analysis to help alleviate the error caused from the underlying assumptions. However, the current trend in the underground gas storage (UGS) industry, due to regulatory emphasis on well bore integrity, is to plug and abandon these older observation wells. The results from this study will demonstrate the importance of using additional observation well data for inventory verification, providing storage operators evidence to support maintaining their observation wells instead of plugging them. The methodology of this study began with a conceptual 2D homogeneous gas/water model. From the homogeneous model, various degrees of reservoir heterogeneity were applied. The pressure results produced from the model were used in Tek’s method for calculating the gas inventory. The error between the calculated inventory and the simulation model inventory was determined for each case. A maximum allowable error was defined enabling the establishment of an applicable window of reservoir heterogeneity for inventory verification using Tek’s method. Finally, additional observation wells were added to the inventory verification methodology to determine if the error could be reduced. The major findings of this study can be grouped into two categories; first, the impact that reservoir heterogeneity has on the calculated inventory and in turn on the percent-error, and second, the impact of incorporating additional observation wells into the equations used to calculate inventory has on the percent-error. For the impact of reservoir heterogeneity, the results demonstrate two findings. The first finding is that, as the degree of heterogeneity increases throughout the reservoir, the percent-error in the calculated inventory increases as well. The second finding demonstrates how reservoir heterogeneity between the observation well and the adjacent injection/withdrawal wells increases the percent-error in the calculated inventory. With the incorporation of additional observation wells, the results show that in cases of high reservoir heterogeneity, using data from additional observation wells can help to reduce the percent-error. For most cases, two additional observation wells were able to decrease the percent-error to an acceptable level. For the cases with the highest degree of heterogeneity and percent-error, more observation wells were required to reduce the error to an acceptable level.