摘要
在旧型空调系统中,因系统改造不易导致建筑能源管理和故障诊断困难,本文设计基于嵌入式设备的空调远程软硬件系统,并引入故障数据和实际运行机组,进行验证与分析,研究了制冷剂泄漏的故障诊断,通过反向传播算法的监督学习方法,使用核主元分析法(KPCA)先对特征进行降维,减少计算量,结合多层感知器(MLP)构成KPCA-MLP方法。结果表明:该算法具有95.83%的制冷剂泄漏诊断正确率;建立一套智能软硬件,通过成本较低的嵌入式设备,使用云服务器和数据可视化页面构成远程管理系统,实现对实际空调机组远程监测和故障检测。
In the old air-conditioning system, due to the difficulty of system transformation, users cannot effectively conduct building energy management and fault diagnosis. The remote monitoring software and hardware system based on embedded equipment is developed. Fault data and actual operating unit data are introduced for model verification and data analysis. With the supervised learning method of the back-propagation algorithm, the kernel principal component analysis(KPCA) method is adopted to reduce the dimension, reduce the amount of calculation, combine with the multi-layer perceptron multilayer perceptron(MLP), and form the KPCA-MLP method. The results show that the algorithm has a 95.83% accuracy rate of refrigerant leakage diagnosis. With a set of intelligent software and hardware and lower cost embedded devices, remote monitoring and fault detection of actual air conditioning units can be achieved with the remote management system by cloud servers and data visualization pages.
作者
姜智堯
黄巍
薛扬帆
杜志敏
晋欣桥
JIANG Zhiyao;HUANG Wei;XUE Yangfan;DU Zhimin;JIN Xinqiao(Institute of Refrigeration and Cryogenics,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《制冷技术》
2021年第6期15-20,共6页
Chinese Journal of Refrigeration Technology
关键词
屋顶式空调机组
故障诊断
嵌入式硬件
核主元分析法
多层感知器
Rooftop air conditioning unit
Fault diagnosis
Embedded-system
Kernel principal component analysis
Multilayer perceptron