Estimation of physical buildings parameters using interval thermostat data

Significant energy savings can be achieved by retrofitting the enclosures and HVAC systems of existing residential buildings. Identification of building retrofit opportunities currently requires on-site energy assessments that are inconvenient to homeowners, expensive, and are of variable accuracy, making it challenging to deliver cost-effective retrofit opportunities at scale. Massive deployment of communicating thermostats provides a possibility for remote energy assessment by analyzing the associated interval indoor temperature and heating system run-time data. In this paper, we present a methodology to estimate the overall building insulation level, HVAC system efficiency, and building airtightness from the communicating thermostat data. The methodology uses a grey-box model of a residential building and includes identification of basic model parameters, followed by estimation and non-parametric modeling of generally variable external and internal heat gains/losses. In this way, it is also possible to predict indoor temperature and energy consumption of the building under various retrofit scenarios and user behaviors. Preliminary results demonstrate the feasibility of the proposed method.