基于支持向量机和粒子群算法的多联机气液分离器插反故障诊断  

Fault Diagnosis on Opposite-insertion for Gas-liquid Separator in Variable Refrigerant Flow System Based on Support Vector Machine and Particle Swarm Optimization Algorithm

作　　者：郑小海 谭泽汉 郭亚宾[1] 陈焕新[1] ZHENG Xiaohai;TAN Zehan;GUO Yabin;CHEN Huanxin(School of Energy and Power Engineering,Huazhong University of Science and Technology,Wuhan,Hubei 430074,China;State Key Laboratory of Energy Conservation and Operation of Air-Conditioning Equipment and Systems,Zhuhai,Guangdong 517907,China)

机构地区：[1]华中科技大学能源与动力工程学院,湖北武汉430074 [2]空调设备及系统运行节能国家重点实验室,广东珠海517907

出　　处：《制冷技术》2018年第4期16-20,共5页Chinese Journal of Refrigeration Technology

基　　金：空调设备及系统运行节能国家重点实验室开放基金(No.SKLACKF201606);国家自然科学基金(No.51576074)

摘　　要：本文通过多联机实验平台采集数据,选取合适的特征变量,利用支持向量机算法建立了多联机气液分离器插反故障诊断模型。采用粒子群算法优化该模型,通过测试集验证模型的分类准确率,并对两种模型下的故障诊断效果进行对比分析。结果表明,制冷工况下的故障诊断准确率高于制热工况下的准确率。模型优化前两种工况下的故障诊断准确率均高于96%,而优化后的准确率均高于97%,优化后制冷工况下的故障诊断准确率更是高达98.4%,可见优化后的模型性能稳步提升。The data were collected through the variable refrigerant flow(VRF)experimental platform,and the suitable characteristic variables were selected.A VRF gas-liquid separator opposite-insertion fault diagnosis model using support vector machine algorithm was developed.The particle swarm optimization algorithm was used to optimize the model.The test set was used to verify the classification accuracy of the model,and the fault diagnosis effects of the two models were compared and analyzed.The results show that the accuracy of fault diagnosis under refrigeration condition is higher than that under heating condition.The accuracy of fault diagnosis under the two conditions is larger than 96%under support vector machine model.After using the particle swarm optimization algorithm,the accuracy is larger than 97%under these two conditions,and the accuracy of fault diagnosis under refrigeration conditions is up to 98.4%.The performance of the optimized model is steadily increased.

关 键 词：支持向量机 多联式空调系统 气液分离器插反 故障诊断 粒子群算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TB65[自动化与计算机技术—控制科学与工程]