Dynamical modeling of a geothermal system to predict hot spring behavior

Abstract

In this paper, a lumped-parameter modeling approach is proposed for predicting spring flow, temperature and chemistry variations as a result of geothermal reservoir exploitation. Based on the assumption that a geothermal reservoir can be coarsely discretized into a few homogeneous blocks, the model combines a non-isothermal tank model of liquid-dominated geothermal reservoirs with the mass and thermal balance equations of hot springs. A state augmented Ensemble Kalman filter is employed for parameter estimation. The proposed model is applied to simulate the impact of geothermal reservoir utilization on an important hot spring in North West Sabalan geothermal system in Iran. This low-fidelity model can provide preliminary insight into the underlying processes of a geothermal system in the early life of a reservoir when data is scarce. Furthermore, the minute simulation time of the model makes it an appealing low-fidelity option for uncertainty analysis and simulation–optimization procedures, where a high number of model simulations are often required.