摘要
本文提出了一种基于随机森林的屋顶机空调系统故障诊断方法。首先,在3个不同的屋顶机系统上证明了该方法的有效性。然后,本文研究并得到了训练样本的故障程度与诊断效率的关系,为实际应用中故障等级的实验设计提供了一定的依据。最后,对本文提出的诊断模型应用于不同屋顶机空调系统时的通用性问题进行了验证与分析。实验结果表明,对于膨胀阀类型相同,但制冷剂、压缩机与系统冷量不同的屋顶机空调系统,本文提出的诊断方法具有一定的通用性。本文创新之处在于提出了基于随机森林的故障诊断方法,并就不同故障程度、不同系统两方面进行了研究。
A fault diagnosis method based on random forest for roof top air conditioning system was presented. The method is applied on three different roof top systems, demonstrating the effectiveness of the method. The relationship of fault degrees contained in the training data and the diagnostic efficiency was analyzed, and the results present the reference for the design of the experimental fault level in the practical usage. Finally, the generalized problem of the diagnosis model proposed in this paper is validated and analyzed when it is used in different roof top systems. The result shows the diagnostic method proposed in this paper has a certain degree of versatility, where roof top systems have same type of expansion valve but different refrigerant, compressor and system cooling capacity. The innovations of this paper are that a fault diagnosis method based on random forest is proposed, and two aspects of different fault degrees and different systems are studied.
作者
吴斌
杜鑫
杜志敏
晋欣桥
WU Bin;DU Xin;DU Zhimin;JIN Xinqiao(Institute of Refrigeration and Cryogenics, Shanghai Jiao Tong University, Shanghai 200240, China)
出处
《制冷技术》
2018年第5期46-53,共8页
Chinese Journal of Refrigeration Technology
关键词
屋顶机系统
故障诊断
机器学习
随机森林
Rooftop system
Fault diagnosis
Machine learning
Random forest