Imitation learning with artificial neural networks for demand response with a heuristic control approach for heat pumps

TL;DR

The paper addresses demand response for heat pumps amid volatile renewables by introducing PSC-ANN, a forecast-free imitation-learning controller trained on optimal actions from a MILP. The method integrates a neural regressor with a Price-Storage-Control heuristic to produce fast, real-time control decisions that maintain comfort. Key contributions include forecast-free imitation learning, cross-building generalization using data from other buildings, and demonstrated cost reductions and faster execution compared with conventional and some intelligent controls; optimal control remains an upper bound. The work supports scalable deployment of heat-pump flexibility across buildings to better utilize renewables and aid grid stability.

Abstract

The flexibility of electrical heating devices can help address the issues arising from the growing presence of unpredictable renewable energy sources in the energy system. In particular, heat pumps offer an effective solution by employing smart control methods that adjust the heat pump's power output in reaction to demand response signals. This paper combines imitation learning based on an artificial neural network with an intelligent control approach for heat pumps. We train the model using the output data of an optimization problem to determine the optimal operation schedule of a heat pump. The objective is to minimize the electricity cost with a time-variable electricity tariff while keeping the building temperature within acceptable boundaries. We evaluate our developed novel method, PSC-ANN, on various multi-family buildings with differing insulation levels that utilize an underfloor heating system as thermal storage. The results show that PSC-ANN outperforms a positively evaluated intelligent control approach from the literature and a conventional control approach. Further, our experiments reveal that a trained imitation learning model for a specific building is also applicable to other similar buildings without the need to train it again with new data. Our developed approach also reduces the execution time compared to optimally solving the corresponding optimization problem. PSC-ANN can be integrated into multiple buildings, enabling them to better utilize renewable energy sources by adjusting their electricity consumption in response to volatile external signals.

Figures (9)

• Figure 1: PSC-ML training
• Figure 2: PSC-ML applied
• Figure 3: ANN schema
• Figure 4: Training process with data from other buildings
• Figure 5: Improvement regarding the cost reduction compared to the Conventional Control approach averaged over 20 buildings per building type for 20 weeks and three runs