18 May 2017
Since an increasing number of unpredictable and undispatchable local renewable energy sources like pv-installations or wind turbines are coupled to the electrical distribution grids, balancing of these grids become more and more challenging. Demand response is seen as a way to deal with this problem.
To assess the demand response performance of a residential heat pump, a lab setup was built. This setup consists of an 11 kW heat pump, combined with two buffers: one for storage of domestic hot water and one for storage of space heating. In addition, the house also has a building integrated pv-installation or a small scale wind turbine and some uncontrollable load. By means of a market-based multi-agent system (MAS), the heat pump is controlled. The ambition was to enable the heat pump when local renewable electricity production was available and to switch the heat pump off at times when there already was a high electricity demand from the household (peak shaving). The tests were performed for both the coldest winter week and an average winter week in a Belgian climate, whereby first the heat pump was controlled by means of a standard heat driven control algorithm and afterwards by means of active control.
The tests show that active control is able to reduce the peak power extraction from the grid significantly. Also the self-consumption of local produced electricity of the house increased by 8 to 30%. However, the results also indicate that the active controlled heat pump consumes more electricity (8-12%).