STATISTICAL ANALYSIS OF THE THERMAL PARAMETERS OF SMART HOMES

Abstract

A statistical analysis of the annual smart home sensor data obtained from REFIT smart homes project is considered. These data, in particular, consist of the high frequency external, internal, radiator surface temperatures and the gas usage values selected from 20 buildings in a period between March 7, 2014 and March 7, 2015. The data were obtained using sensors installed in twentyUKdwellings under REFIT programme dedicated to the research problem of energy savings for the smart homes. Firstly, a sqlite database for the raw data readings was created. Secondly, data preprocessing and resampling of the data were made in order to fill missing values, smooth and decrease the frequency. The correlation analysis of different data readings showed relations which allow one to use the data not only for forecasting problems, but also to formulate inverse problems of the determination of the effective thermophysical coefficients.

Key words: smart home, home automation, data acquisition, data mining