Modelling transitions on heating usage in buildings with multivariate statistical monitoring

In this paper Principal Component Analysis (PCA) is proposed for monitoring and controlling the heating system of a building. PCA allows modelling correlations between independent variables weather and energy consumptions of the distinct dwellings.This approach allows defining simple statistic indices T² and SPE to be used in monitoring charts. These indices can be used to detect abnormal behaviours but also as proposed in this paper they can be used for controlling the heating system. Also PCA is proposed as energy forecasting technique. Finally a case study based on real data from a real building with 96 dwellings is presented.