Detection of consumers with electric heating devices and temperature normalization of their electric energy consumption

In the present paper, the procedure to detect consumers with the electric heating devices is shown. In this procedure, the AdaBoost machine learning algorithm is used to detect heat pumps. For the input data, three main sources of data are used from which the telemetry of the 15-minute electric energy consumption shows the most informativeness. The accuracy of the results is estimated from the internal data of the consumer's market actions for heat pumps. Further, bilinear regression is used to detect consumers with any electric heating devices and the results of both analyses are compared. Finally, the algorithm of temperature normalization of energy consumption of these consumers is presented.