Research of Building Heat Inertia Cycle Based on Data Mining

Abstract

In this paper, a periodic regression prediction model of building thermal inertia was established, based on the time series historical data of the actual monitoring of intelligent heating system, which took outdoor temperature, indoor temperature and historical heat consumption as independent variables. MATLAB software was used to conduct regression prediction analysis on the heat consumption of radiator heating buildings with different thermal performance. Through analysis, it was also found that the heat consumption was affected by the thermal inertia period of buildings, and then, the concept of thermal inertia period of buildings was put forward. Through the analysis of the relative error of regression results, it was found that the cycle of building thermal inertia went through three periods, namely fluctuation period, stationary period and fluctuation divergence period. Buildings with different thermal properties usually have different optimal thermal inertia cycles, and with the improvement of thermal performance, the optimal thermal inertia cycle becomes longer. The study of building thermal inertia period provides valuable theoretical reference for the accurate prediction of heating parameters in heating system.