基于聚类-回归算法的制冷机组实际运行特性研究

CLUSTER-REGRESSION ALGORITHM BASED RESEARCH ON PRACTICAL OPERATION CHARACTERISTICS OF REFRIGERATING UNITS

制冷机组部分负荷性能对制冷机组运行能耗影响非常大,本文提出基于聚类–回归算法挖掘制冷机组实际运行特性的方法,即以K–Means聚类算法识别不同负荷率区间的代表性运行数据,以最小二乘法对聚类结果进行回归分析。并基于地源热泵机组的大量实际运行数据,采用提出的聚类–回归算法,研究了地源热泵机组的实际运行特性,拟合了机组COP与部分负荷率PLR的曲线,得到COP与PLR的数学表达式,为开展制冷机组的智能调控奠定基础。

Some bearing capacities of refrigerating unit have very big influence on the operation energy consumption.This paper proposed a cluster-regression algorithm based method to explore the practical operation characteristics of refrigerating unit,namely using K-Means cluster algorithm to determine the representative operation data at different load intervals and then using least square method for regression analysis on the cluster result.Based on the large quantity of practical operation data of ground source heat pump unit,the practical operation characteristics of ground source heat pump unit were researched.COP and PLR curves were fitted and COP and PLR mathematical expressions were determined,providing a solid basis for intelligent control of refrigerating unit.