作者: Marc Azar
DOI:
关键词: Principal component analysis 、 Simulation 、 Statistics 、 Energy consumption 、 Consumption (economics) 、 Engineering 、 Trend analysis 、 Apartment 、 Data collection 、 Cluster analysis 、 Visualization
摘要: This thesis investigates the influence of energy visualization on hot water consumption, as well asbuilds up framework for analysis electricity multi-apartment buildingsin Sweden, . 115 apartments in Nynashamn have been scheduled to be equipped with feedback visualization,through use television sets, allowing monitoring and waterconsumption a monthly basis. One year consumption data prior this introductionwas acquired statistical analysis. The results were then displayed analysed, allowingfor composition generalized conclusion whilst revealing need further investigationand future work.To achieve end, clustering was performed according followingcharacteristics: Number tenants per apartment, Area Location theapartment, age tenants. Principal Component Analysis used select dominantcharacteristics, through eliminating highly correlated components, after which trend wasperformed each separate clusters seasonal change model.Finally, Multivariate Variance utilized paired identify any significantchange, along post-analysis tests specify groups significant wasdetected, is presented applied work.The preliminary clearly show that characteristic can grouped into three distinctclusters distinct. Moreover thedata reveals correlation between apartment’s characteristics its consumption.However collection will required before strong beidentified, power conclude change.Nonetheless initial demonstrate promising signs ought investigated inthe future.2