Using smart meters for diagnostics and model-based control in thermal comfort systems

This paper describes how smart meters can be used to for automatic diagnostics and control in HVAC systems. The specific smart meter used in this paper is the non-intrusive load monitor (NILM). The NILM can disaggregate the operating schedule of individual loads from aggregate current measurements. Combining load operating schedules with measurements from several environmental sensors, the NILM builds a representative thermal model for the building. Two different model formulations are proposed. The first is designed to detect refrigerant leaks in existing thermostatically controlled HVAC systems; the other is intended for use with next-generation variable-capacity systems. Experimental and simulation results are presented.