Adaptive model-based optimal control of hybrid deep borehole ground source heat pump systems with integrated latent heat thermal energy storage

Compared with conventional deep borehole ground source heat pump (DB-GSHP) systems, integrating latent heat thermal energy storage (LHTES) and borehole passive heating into the DB-GSHP system has greater potential in achieving energy savings and increasing demand flexibility. This study presented an adaptive model-based optimal control strategy for hybrid DB-GSHP systems with integrated LHTES and passive heating. The optimal control problem was solved using adaptive performance models, quantile regression, online identification, and a genetic algorithm (GA), to identify the optimal control settings of the hybrid system. To predict system energy performance, novel adaptive models for the deep borehole heat exchanger (DBHE), LHTES tanks, and heat pump were proposed, and the model parameters were continuously updated using an adaptive forgetting factor recursive least squares estimation algorithm. A quantile regression technique was integrated with a GA optimizer to dynamically narrow down the search space of the decision variables. The proposed control strategy was tested along with two benchmarking scenarios using a co-simulation approach. The results showed that the DBHE control-oriented adaptive model, combining discrete transfer functions and online identification technique, can effectively predict the outlet temperature of the borehole under dynamic working conditions. By integrating quantile regression models, the average computational costs of the GA optimizer were reduced by 32.9 %. The proposed control strategy achieved 11.9 % energy savings and 11.5 % electricity cost savings for the integrated system over a heating season with respect to a baseline control strategy. Compared to the system without LHTES, the system with integrated LHTES saved 6.4 % in energy use and 35.2 % in electricity costs, when the proposed control strategy was applied to both systems.