Analysis of Power consumption behavior based on the seasonal daily pattern of buildings in district heating domain

Abstract: The power consumption analysis in district heating system is quite interesting research area. The findings of such analysis could provide assistance to energy distributors for power consumption optimization and prediction. In this study, we provide an analysis of clustering process by two different methods including Intensity based and Shape-Intensity based clustering. Shape-Intensity clustering provides the information from two different views, one is power intensity consumed and Secondly, the way by which the power is consumed by the customer through out the day. The data set used for the experimentation of this work belongs to Helsinborg, Sweden. Moreover, the analysis is carried out on different categories of customers such as Tillverkande industri, Flerfamiljshus, Offentlig förvaltning, Skola, Sjukvård, Sociala tjänster etc. The work includes analysis for two seasons: winter and summer for each customer. The study basically provides the descriptive analysis by using visualization tools by which we try to draw the merits and demerits of both type of methods. We also try to understand the effect of outdoor temperature on power consumption, which was the only additional parameter provided for this data set. The proposed method of Shape-Intensity clustering provide some merits over Intensity based clustering which could provide assistance for the task of prediction in District Heating Systems.