基于决策树算法的多联机制冷剂泄漏在线故障诊断

Online Fault Diagnosis for Refrigerant Leakage of Variable Refrigerant Flow System Based on Decision Tree Algorithm

多联机被广泛用于各类建筑的空调系统中,而制冷剂泄漏可以对多联机机组运行带来一系列的影响。因此,本文提出决策树算法对多联机制冷剂泄漏进行在线故障诊断。首先,通过生成决策树和剪枝,建立故障诊断模型;然后,进行实验,获取实时数据,并对实验数据预处理;最后,利用决策树模型和处理后的数据进行故障诊断。结果表明,该决策树模型能够对多联机制冷剂泄漏进行有效故障诊断,且诊断正确率较高。

Variable refrigerant flow (VRF) systems have been used in various buildings, and refrigerant leakage can cause a series of influence on operation of VRF systems. Thus, the decision tree algorithm is proposed to detect refrigerant leakage fault online. Firstly, after the decision tree is generated and pruned, the fault diagnosis model is established. Then, the real time data is got throught the experiment and then preprocessed. Finally, the decision tree model and the processed data are used to diagnose the fault. The results show that the decision tree model can effectively diagnose the fault of refrigerant leakage, and the correct rate is relatively high.