基于优化神经网络的空调系统未知类型故障诊断
|
摘要
目前基于数据驱动方法的制冷系统故障诊断模型,只能对参与建模训练的已知类型故障进行诊断,而对于未参与建模训练的未知类型故障,不能正确地诊断。针对这一问题,本文提出了一种优化神经网络的故障诊断策略。利用已知类型的故障数据建立BP神经网络模型,然后确定一个区分阈值,能够实现对未知类型故障的诊断识别。结果表明:对于包含所有故障类型的测试数据,模型的诊断正确率为88.62%,对于其中的未知类型故障,模型的诊断效果显著,正确率为99.48%。
At present, the fault diagnosis models based on the data driven for refrigeration system can only diagnose the known types of faults which are involved in the modeling process, but they fail to diagnose the unknown types of faults which are not involved in the modeling process. To solve this problem, the strategy based on an optimized neural network is presented. The known types of faults data are used to train the back propagation neural network model. Then, a threshold of distinction is established, which can diagnose the unknown types of faults successfully. The results show that the accuracy rate of the model is 88.62% for the testing data which includes all types of faults. For the unknown types of faults, the performance of the model is significant and the accuracy rate is 99.48%.
At present, the fault diagnosis models based on the data driven for refrigeration system can only diagnose the known types of faults which are involved in the modeling process, but they fail to diagnose the unknown types of faults which are not involved in the modeling process. To solve this problem, the strategy based on an optimized neural network is presented. The known types of faults data are used to train the back propagation neural network model. Then, a threshold of distinction is established, which can diagnose the unknown types of faults successfully. The results show that the accuracy rate of the model is 88.62% for the testing data which includes all types of faults. For the unknown types of faults, the performance of the model is significant and the accuracy rate is 99.48%.
引文
[1]韩玲.基于人工神经网络模型的螺杆式冷水机组故障诊断研究[D].重庆:重庆大学,2005.
[2]任能.制冷系统故障检测?诊断及预测研究[D].上海:上海交通大学,2008.
[3]梁晴晴,韩华,崔晓钰.基于概率神经网络的离心式制冷机故障诊断[J].暖通空调,2015(11):101-107.
[4]郑小海,谭泽汉,郭亚宾,等.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[J].制冷技术,2018,38(4):16-20.
[5]禹法文,陈焕新,李绍斌,等.基于主元分析法的多联机系统压缩机排气温度传感器故障检测与诊断[J].制冷技术,2017,37(4):29-33.
[6]韩琦,魏东,曹勇.暖通空调系统故障检测与诊断技术研究进展[J].暖通空调,2014(3):105-112.
[7]邹冠星,苏阳.制冷剂产品市场分析[J].制冷技术,2018,38(S1):36-47.
[8]杨云雨.基于小波神经网络的空调系统传感器故障诊断[D].上海:上海交通大学,2008.
[9]陈玲,陈志杰,杜志敏,等.基于减法聚类和支持向量机的故障诊断方法研究[J].制冷技术,2018,38(4):1-5.
[10]袁玥,陈焕新,石书彪,等.基于主成分分析和神经网络相结合的制冷剂充注量故障诊断[J].制冷技术,2017,37(6):45-50.
[11]LI G,HU Y,CHEN H,et al.An improved fault detection method for incipient centrifugal chiller faults using the PCA-R-SVDD algorithm[J].Energy and Buildings,2016,116:104-113.
[12]WANG Z,WANG Z,HE S,et al.Fault detection and diagnosis of chillers using Bayesian network merged distance rejection and multi-source non-sensor information[J].Applied Energy,2017,188:200-214.
[13]李如平,朱炼,吴房胜,等.BP神经网络算法改进及应用研究[J].菏泽学院学报.2016,38(2):13-17.
[14]唐忠,谢涛.Matlab神经网络工具NNTool的应用与仿真[J].计算机与现代化,2012(12):44-47.
[15]李志生,张国强,刘建龙,等.基于BP神经网络的制冷机组故障检测与诊断[J].流体机械,2006(9):75-79.
[16]石书彪,陈焕新,李冠男,等.基于改进BP网络的冷水机组故障诊断[J].制冷学报,2015,36(6):34-39.
[17]梁晴晴,韩华,崔晓钰.基于BP神经网络优化的离心式冷水机组故障诊断研究[J].制冷与空调,2016(1):25-30.
[18]周开利,康耀红.神经网络模型及其MATLAB仿真程序设计[M].北京:清华大学出版社,2005:75.
[19]郑华,徐远芳,周旸.基于MATLAB的BP神经网络实现研究[J].微型电脑应用,2006,22(8):41-44.
[20]何正风.MATLAB R2015b神经网络技术[M].北京:清华大学出版社,2016:156-159.
[21]王小川,史峰,郁磊.MATLAB神经网络43个案例分析[M].北京:北京航空航天大学出版社,2013:1-4.
[22]虞和济,陈长征,张省.基于神经网络的智能诊断[J].振动工程学报,2000,13(2):46-53.
[23]HORNIK K.Approximation capabilities of multilayer feedforward networks[J].Neural Networks,1991,4(2):251-257.
[2]任能.制冷系统故障检测?诊断及预测研究[D].上海:上海交通大学,2008.
[3]梁晴晴,韩华,崔晓钰.基于概率神经网络的离心式制冷机故障诊断[J].暖通空调,2015(11):101-107.
[4]郑小海,谭泽汉,郭亚宾,等.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[J].制冷技术,2018,38(4):16-20.
[5]禹法文,陈焕新,李绍斌,等.基于主元分析法的多联机系统压缩机排气温度传感器故障检测与诊断[J].制冷技术,2017,37(4):29-33.
[6]韩琦,魏东,曹勇.暖通空调系统故障检测与诊断技术研究进展[J].暖通空调,2014(3):105-112.
[7]邹冠星,苏阳.制冷剂产品市场分析[J].制冷技术,2018,38(S1):36-47.
[8]杨云雨.基于小波神经网络的空调系统传感器故障诊断[D].上海:上海交通大学,2008.
[9]陈玲,陈志杰,杜志敏,等.基于减法聚类和支持向量机的故障诊断方法研究[J].制冷技术,2018,38(4):1-5.
[10]袁玥,陈焕新,石书彪,等.基于主成分分析和神经网络相结合的制冷剂充注量故障诊断[J].制冷技术,2017,37(6):45-50.
[11]LI G,HU Y,CHEN H,et al.An improved fault detection method for incipient centrifugal chiller faults using the PCA-R-SVDD algorithm[J].Energy and Buildings,2016,116:104-113.
[12]WANG Z,WANG Z,HE S,et al.Fault detection and diagnosis of chillers using Bayesian network merged distance rejection and multi-source non-sensor information[J].Applied Energy,2017,188:200-214.
[13]李如平,朱炼,吴房胜,等.BP神经网络算法改进及应用研究[J].菏泽学院学报.2016,38(2):13-17.
[14]唐忠,谢涛.Matlab神经网络工具NNTool的应用与仿真[J].计算机与现代化,2012(12):44-47.
[15]李志生,张国强,刘建龙,等.基于BP神经网络的制冷机组故障检测与诊断[J].流体机械,2006(9):75-79.
[16]石书彪,陈焕新,李冠男,等.基于改进BP网络的冷水机组故障诊断[J].制冷学报,2015,36(6):34-39.
[17]梁晴晴,韩华,崔晓钰.基于BP神经网络优化的离心式冷水机组故障诊断研究[J].制冷与空调,2016(1):25-30.
[18]周开利,康耀红.神经网络模型及其MATLAB仿真程序设计[M].北京:清华大学出版社,2005:75.
[19]郑华,徐远芳,周旸.基于MATLAB的BP神经网络实现研究[J].微型电脑应用,2006,22(8):41-44.
[20]何正风.MATLAB R2015b神经网络技术[M].北京:清华大学出版社,2016:156-159.
[21]王小川,史峰,郁磊.MATLAB神经网络43个案例分析[M].北京:北京航空航天大学出版社,2013:1-4.
[22]虞和济,陈长征,张省.基于神经网络的智能诊断[J].振动工程学报,2000,13(2):46-53.
[23]HORNIK K.Approximation capabilities of multilayer feedforward networks[J].Neural Networks,1991,4(2):251-257.