摘要
在建筑能耗监测中,由于部分建筑的配电支路末端包含多种设备,分项能耗数据无法直接计量获取。为此,在数据协调理论的基础上建立建筑能耗拆分模型;采用内点惩罚函数法将约束优化问题转化为无约束问题,并通过粒子群(PSO)优化算法对拆分模型进行求解。实例研究表明,PSO优化所得的分项逐时电耗相对误差分别为9.60%和4.84%,能反映分项电耗数据的逐时局部特征,可用于建筑电耗的实时在线拆分。
Subentry energy consumption data can not directly measured since there are various devices used in the distribution branch terminals of the existing building. The en- ergy consumption decomposition model of buildings is established based on data reconcilia- tion theory. Internal penalty function method is used in processing constraints problem to transform the constrained optimization into unconstrained optimization. The particle swarm optimization is used to solve the decomposition model. The examples show that the relative errors between the calculated electricity consumption results by the PSO algorithm and measurement results of end A and end B are 9.60% and 4.84% respectively. The parti- cle swarm optimization algorithm can be used to decompose the real-time power consump- tion of building energy consumption monitoring system.
出处
《智能建筑电气技术》
2015年第6期62-65,共4页
Electrical Technology of Intelligent Buildings
关键词
分项计量
数据协调
约束优化
PSO算法
sub-metering
data reconciliation
constrained optimization
particleswarm optimization
