Artificial-Neural-Network-Based Autonomous Demand Response Controller for Thermostatically Controlled Loads

Abstract

Thermostatically controlled loads (TCLs) possess an inherent potential for demand-side management (DSM). Space heating (SH) demands offer added flexibility due to their heat-storing capability and fast response times that may be unnoticed, but it has significant value for DSM. In this article, a communication-free demand response (DR) controller has been presented that harnesses the heat-storing capacity of residential buildings by utilizing TCLs. The proposed controller addresses customer voltage violations, comfort constraints, and energy savings. The DR controller attempts to control the TCLs by relying on voltage and voltage-to-load sensitivity. By deploying an artificial-neural-network-based model, the proposed control scheme ensures that the SH loads connected at spatially distributed nodes in the power system participate fairly to mitigate the residential grid voltage violations. Simulation results verify the proposed controller's performance to revamp the system voltage profiles while maintaining the desired comfort of end users. The controller is successfully tested with distributed generation integration, system reconfiguration, severe ambient temperature, and noise in the measured signals.