Characterization, testing, and optimization of load aggregation methods for ground heat exchanger response-factor models

Abstract

Ground heat exchangers (GHEs) are often simulated using response factor-type models that rely on the GHE load history. As the simulation progresses, the number of GHE load history terms that need to be included in the temperature response computation grows. For extended, multiyear simulations the required computation time due to these history terms can be excessive, which limits the usefulness of response factor-type models. To cope with this, load aggregation procedures have been developed that preserve the net effect the individual loads but reduce the overall number of computations. The previously developed load aggregation methods can be characterized as two different methods, but with different parameters. The parameters each affect the accuracy and computational efficiency of the methods, yet there has been no systematic investigation of the parameter space. In this article, both methods are implemented and compared for a wide range of parameters. A final recommendation is made that results in a 73-fold reduction in simulation time when compared to nonaggregated simulations. For a worst-case situation with an imbalanced load profile on a densely packed borehole field, the MBE is less than 0.13 °C for 5- and 10-year simulations. The resulting predicted heat pump energy consumption error is less than 0.3%.