Data Driven Electricity Management for Residential Air Conditioning Systems: An Experimental Approach

Effective control of air conditioning systems (ACs) has the potential of significant electricity savings and demand response management for an entire power system. In this context, this paper demonstrates some key experimental results on controlling the electricity consumption of ACs. In particular, the degree to which energy can be throttled for energy management purposes without affecting the end user's comfort level is described. The testbed was set up in a residential building, in which the set point temperature of ACs installed within each apartment unit is controllable from a remote server. The main objectives were to reduce the consumption of electricity by the compressors, and to investigate the feasibility of having residential ACs as interruptible loads to participate in the electricity market. The algorithm used for controlling is explained in detail, and the users' experiences during the experiments are briefly discussed. Extensive data collected throughout the experiment are provided to show the effectiveness of having ACs as flexible loads in reducing the power consumption by the compressors as well as the potential of offering ACs as interruptible load into the market without compromising user comfort.