Algorithm of Tunning Heating Source Thermophysical Parameters in Smart Home

A development of the efficient smart home energy management systems requires understanding of the heating behavior. Part and parcel of this process is thermal parameters of heating sources which define the effectiveness of energy usage. In this paper we introduce a novel approach for the estimation of thermophysical parameters of heating elements. Based on the formulation of dynamic heat transfer problem we present a model that describes the heating process in a building with multiple heating sources. Within this model we propose an algorithm of determination the effective thermophysical parameters, where historical temperature and aggregated energy data are used in order to estimate unknown parameters. Also we present the algorithms for data preparation and renormalization. The latter means a transition from integral building parameter of heating sources to separate parameters corresponding to particular room heating element. Due to occurred errors during the computation, the presence of some incorrect values in provided time series and the ambiguity of calculation scenarios, the resulting parameters can be tuned to achieve greater accuracy. To do that, we propose an approach which involves the multivariate minimization of loss function using the genetic algorithm. The whole algorithmic pipeline is discussed in this report.