Regression-Based Thermodynamic Model Identification of a Zone with a Closed-Access Air Conditioner

Abstract

Smart buildings have a great potential in the energy regulation market. One of the levers that are used for the flexibility provision of buildings are the thermal comfort systems. This paper deals with a thermodynamic model identification for a comfort-regulated zone of a smart home with an installed air conditioner. The intended end-use of the model is model predictive control of comfort with electricity demand response over a collection of objects with such similar configuration. Typical commercial setups of closed-access air conditioners found in residential objects are considered, where there is no possibility of any data communication from the air conditioner and where the sensory equipment is quite limited due to an intended large-scale deployment. This drives the specific input-output model form where the air conditioner electricity consumption is the selected model input. The performance of different mathematical models for temperature prediction in the smart home in the heating season is analyzed and the results with quantitative measures are provided. The complete analysis is based on measurement data collected on a smart home experimental setup.