Model Predictive Control for Thermostatically Controlled Appliances Providing Balancing Service

This paper studies the modeling and control methods for the provision of ancillary services through aggregating thermostatically controlled appliances (TCAs). A model predictive control (MPC) scheme is presented. Minimum ON/OFF time for protecting the TCA unit is explicitly described in the MPC scheme. A novel method of converting time-integrated interdependent logic conditions, which are for constraining the lockout effect, into inequalities is proposed in this paper. A case study is conducted with realistic data to validate the feasibility of the modeling and control approaches. Parametric studies in terms of temperature deadband width, ambient temperature, and lockout time are conducted to investigate the service quality in different circumstances. The test results reveal that a population of diversified TCAs can become a major source of providing ancillary service. The proposed approach can be specialized to deal with the pure ON/OFF control.