A simplified method to evaluate geothermal storage in an aquifer with consideration of heat transfer between aquifer and caprock/baserock

¹ Department of Civil and Environmental Engineering, UC Berkeley, Berkeley, CA 94704 USA
² Department of Engineering, University of Cambridge, Cambridge CB2 1PZ UK
³ Energy Geosciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720 USA

^* Corresponding author: alpcinar@berkeley.edu

Abstract

Storing and extracting heat during different seasons of the year is possible through the utilization of a ground aquifer with an open loop Ground Source Heat Pump (GSHP) system. Being able to predict the hydrothermal performance of geothermal storage is required for an efficient operation of the system for cooling and heating of buildings. Complex 2D and 3D hydrothermal numerical models can simulate the thermal performance of geothermal storage accurately but often lack the desired computational speed for conducting large number of simulations for performance optimization. Instead, a 1D radial model can be used to conduct fast evaluation. However, it is important that the model computes the amount of heat loss from an aquifer into the overburden and underlying layers accurately to evaluate the amount of geothermal storage in the aquifer at different times. In this study, a source term is introduced into a 1D model to simulate the heat transfer between the aquifer and caprock/baserock in the vertical direction. The following two heat loss models are introduced in the heat advection-conduction equation: (i) Newton’s heating/cooling law, which leads to a closed form solution, and (ii) a conduction-based semi-analytical model, which requires a 1D finite element solution. When compared to a full 2D axisymmetric simulation result, it was found that the Newton’s heating/cooling law model with a constant heat transfer coefficient works well in cases of fast heat flow rate in thick aquifers of around 100 meters. But large errors in estimating heat dissipation are observed in cases with low heat flow rate in thin aquifers, especially for simulations exceeding two to five years. On the other hand, the model with the conduction-based semi-analytical solution gives a better match for these conditions.