制冷站整体能耗优化新方法在达实新大楼的应用

A Novel Method of Minimizing Power Consumption for Das New Building's Chiller Plant

在中央空调系统中,制冷主机及其辅助设备的性能和参数决定着制冷主机能耗高低。但在实际应用中很难找到一种合适的模型来准确预测中央空调系统能效及其运行参数设定值。通过对比SL、BQ、MP、GNU、QHP及BP-ANN等数学模型,发现BP-ANN模型可准确预测制冷主机COP。并基于GA遗传算法确定可控变量值,在指定范围内寻找变量最优值。利用达实大厦改扩建项目中的制冷主机的实际运行数据,对比评测了制冷主机COP不同预测模型的性能。同时还确定了影响制冷站能耗的可控变量值,进而实现了整体制冷站能耗最低的目的。

The power consumption of chiller plants depends on its devices＇ performance and setting parameters. However, it is significant difficult to find a suitable mathematical model to predict the energy efficiency of central air-conditioning system. In this paper, several mathematical model, such as SL, BQ, MP, GNU, QHP and BP-ANN, are compared, and BP-ANN is considered having the best accuracy for prediction. The genetic algorithm is used for searching optimum parameters in certain range. In this study, using the real chiller plant operating data of Das new building, the performances of mathematical models for chillers are evaluated. Meanwhile, the effective parameters are determined in order to minimize the power consumption of chiller plants.