Effects of Observation Well Location in Lumped Parameter Models

Abstract

Lumped Parameter modeling is used as a widespread tool for predicting the performance of a geothermal reservoir. Being a simple and an efficient method lumped parameter models can give performance predictions in a short time. By history matching the past data, the behavior of the reservoir can be characterized, which is done especially by determining recharge indices and storage capacities of each tank that the geothermal system is represented by. Hence, if error is introduced to the average reservoir pressure and/or temperature values, these errors will propagate into the parameters that will be estimated. Having estimated the parameters with error will lead to an incorrect future performance prediction which would then lead to poor reservoir management. In this study, for the liquid dominated isothermal geothermal reservoirs containing single-phase liquid water, the errors made in estimation of the model parameters, caused by the errors in the average reservoir pressure inputs, are characterized. First of all, for the synthetic data generation, a 2D numerical model is generated. Then lumped parameter models are used to history match to the synthetic data in order to obtain the recharge index and storage capacity values for the representing lumped parameter model. By using the inverse model and numerical model, errors are investigated. Based on the results, the reservoir systems that have aquifer connection will lead to drastic errors if the proper observation well is not chosen.