摘要
分析了建筑逐时能耗数据,研究了检测异常建筑逐时能耗的方法.该方法包括两个步骤,首先根据数据属性将相同条件下产生的逐时能耗数据进行分类,然后对同一类能耗数据进行分析,确定异常能耗点.实验证明该方法是有效的,可以优化建筑管理方式,达到节能的目的.
A method for detecting abnormal building electricity consumption is described using collected hourly energy consumption data. It includes two steps： firstly, the data are classified depending on their attribute, namely the data in the same set should be generated in almost the same situation ; then the outliers are detected in the same set based on statistic method. It determines whether abnormal energy consumption is existent. Experiments show the data analysis method is effective and promising since it can optimize building operation systems and the purpose of energy saving can be achieved.
出处
《上海电力学院学报》
CAS
2011年第2期149-152,159,共5页
Journal of Shanghai University of Electric Power
关键词
建筑能耗
异常点检测
分类与回归树算法
building energy consumption
outlier detection
classification and regression treealgorithm
