天然气深冷燃气轮机故障诊断专家系统研究
|
摘要
由于燃气轮机设备结构复杂,影响其运行的因素较多,且其长期工作在高温、高压状态下,很多部件有可能发生故障。一旦设备出现故障停止运转,则需要投入大量的金钱、精力与时间来进行维修,带来巨大的经济损失,有时甚至会对工作人员的生命安全造成威胁。由于燃气轮机故障诊断能大幅度降低维修成本,同时也能大大提高机组运行的安全性和可靠性,所以对燃气轮机故障诊断系统的研究不仅有重要的理论意义,而且有较好的实用价值。
本文针对TORNADO燃气轮机的故障特点,分析了TORNADO燃气轮机在实际工况下的典型故障,对各种故障的原因和所表现出的特性进行了详尽的研究,并根据TORNADO燃气轮机的特点建立了故障诊断知识库,为故障诊断打下了基础。
传统专家系统处理的知识是显式的、表面的,存在推理能力弱、容易出现知识获取“瓶颈”等问题。文中在传统专家系统的基础上,引入了目前在故障诊断学科较为先进和有效的神经网络理论,构造基于神经网络的TORNADO燃气轮机故障诊断专家系统。这样做能够提高系统的智能水平,改善系统的性能。利用该故障诊断专家系统,即使没有掌握专业的领域知识,也可以直接通过神经网络的输出得到结论。与传统专家系统相比,基于神经网络的专家系统更擅于对数据的处理。
在研究过程中,以组合神经网络为基础建立了TORNADO燃气轮机故障诊断专家系统。根据对相关资料的查询,发现目前尚无学者专门针对TORNADO燃气轮机进行故障诊断研究,本文的研究具有一定的独创性。文章结合TORNADO燃气轮机曾经出现过的各种故障信息,利用现场提供的样本规则,对神经网络进行学习和训练,直到达到所需精度。利用成熟的神经网络专家系统对TORNADO燃气轮机的运行数据进行处理,便得到TORNADO燃气轮机的故障信息。
最后的测试结果表明,该故障诊断专家系统不但能够解决传统专家系统自学习能力差和容易出现知识获取“瓶颈”等问题,故障诊断正确率也是令人十分满意的。
Due to the complex structure, influenced by many factors, and its long-term working in high temperature and high pressure condition, many components of the gas turbine may fail. Once a device malfunction and stop operation, it will cost a lot of money, energy and time to repair and bring huge economic losses, sometimes even threaten to staff's life security. As gas turbine fault diagnosis can greatly reduce maintenance costs, and also can greatly improve the safety and reliability, so gas turbine fault diagnosis system research not only has great theoretical significance, but also have good practical value.
Based on the characteristics of gas turbine failure, we analyzed typical gas turbine failure in the actual conditions and have a detailed study of various failure causes and shown characteristics. According to the characteristics of gas turbine,a gas turbine fault diagnosis knowledge base is established, laying the foundation for fault diagnosis.
The knowledge that the traditional Expert System can deal is the apparent one, so its inductive capacity is weak, and also has problems of the knowledge acquisition. In this paper, we built a new system NNES by introducing neural network which is advanced and effective in the diagnostic domain. This practice can improve the intelligent level of the system and improve the system performance. By this fault diagnosis expert system, even if we do not master the professional domain knowledge, we can also get the output conclusion directly from the neural network. Compared with the traditional one, the expert system based on the neural network is better at the data processing.
Through the study processing, the fault diagnosis expert system of TORNADO turbine based on combined neural network is constructed. With the every situation in detail, the neural network is trained until the desired accuracy is achieved. The leaned network will deal with the diagnosis data and then the diagnosis news will be putout.
The experimental results show that the fault diagnosis expert system improves the shortcoming that the traditional expert system has, and the fault diagnosis efficiency is satisfied.
本文针对TORNADO燃气轮机的故障特点,分析了TORNADO燃气轮机在实际工况下的典型故障,对各种故障的原因和所表现出的特性进行了详尽的研究,并根据TORNADO燃气轮机的特点建立了故障诊断知识库,为故障诊断打下了基础。
传统专家系统处理的知识是显式的、表面的,存在推理能力弱、容易出现知识获取“瓶颈”等问题。文中在传统专家系统的基础上,引入了目前在故障诊断学科较为先进和有效的神经网络理论,构造基于神经网络的TORNADO燃气轮机故障诊断专家系统。这样做能够提高系统的智能水平,改善系统的性能。利用该故障诊断专家系统,即使没有掌握专业的领域知识,也可以直接通过神经网络的输出得到结论。与传统专家系统相比,基于神经网络的专家系统更擅于对数据的处理。
在研究过程中,以组合神经网络为基础建立了TORNADO燃气轮机故障诊断专家系统。根据对相关资料的查询,发现目前尚无学者专门针对TORNADO燃气轮机进行故障诊断研究,本文的研究具有一定的独创性。文章结合TORNADO燃气轮机曾经出现过的各种故障信息,利用现场提供的样本规则,对神经网络进行学习和训练,直到达到所需精度。利用成熟的神经网络专家系统对TORNADO燃气轮机的运行数据进行处理,便得到TORNADO燃气轮机的故障信息。
最后的测试结果表明,该故障诊断专家系统不但能够解决传统专家系统自学习能力差和容易出现知识获取“瓶颈”等问题,故障诊断正确率也是令人十分满意的。
Due to the complex structure, influenced by many factors, and its long-term working in high temperature and high pressure condition, many components of the gas turbine may fail. Once a device malfunction and stop operation, it will cost a lot of money, energy and time to repair and bring huge economic losses, sometimes even threaten to staff's life security. As gas turbine fault diagnosis can greatly reduce maintenance costs, and also can greatly improve the safety and reliability, so gas turbine fault diagnosis system research not only has great theoretical significance, but also have good practical value.
Based on the characteristics of gas turbine failure, we analyzed typical gas turbine failure in the actual conditions and have a detailed study of various failure causes and shown characteristics. According to the characteristics of gas turbine,a gas turbine fault diagnosis knowledge base is established, laying the foundation for fault diagnosis.
The knowledge that the traditional Expert System can deal is the apparent one, so its inductive capacity is weak, and also has problems of the knowledge acquisition. In this paper, we built a new system NNES by introducing neural network which is advanced and effective in the diagnostic domain. This practice can improve the intelligent level of the system and improve the system performance. By this fault diagnosis expert system, even if we do not master the professional domain knowledge, we can also get the output conclusion directly from the neural network. Compared with the traditional one, the expert system based on the neural network is better at the data processing.
Through the study processing, the fault diagnosis expert system of TORNADO turbine based on combined neural network is constructed. With the every situation in detail, the neural network is trained until the desired accuracy is achieved. The leaned network will deal with the diagnosis data and then the diagnosis news will be putout.
The experimental results show that the fault diagnosis expert system improves the shortcoming that the traditional expert system has, and the fault diagnosis efficiency is satisfied.
引文
[1]陈立滇.是历史赋予石油工业的任务——第十九届世界石油大会热点问题概述[J].环球石油瞭望,2008,8:10-19.
[2]单旭昇.神经网络专家系统在凝汽器故障诊断中的应用研究[D].华北电力大学.2009,2.
[3]陈夏夏,陈静杰.基于粗糙集的机载电子设备故障检测专家系统的改进[J].河南科学,2011,01(29):77-79.
[4]熊英.人工智能及其在计算机网络技术中的应用[J].技术与市场,2011,2(18):20.
[5]谢建,何德虎,李良.基于观测器和神经网络的液压系统故障诊断方法研究[J].机床与液压,2011,1(39):135-137.
[6]黄小光,王永泓.燃气轮机混合智能故障诊断专家系统[J].船舶工程,2000,22(6):8-39.
[7]翁史烈,王永泓.基于热力参数的燃气轮机智能故障诊断[J].上海交通大学学报,2002,36(2):165-168.
[8]卜凯旗,任兴民,秦卫阳.燃气轮机发电机组的振动信号监测与分析系统[J].振动与噪声控制,2007,4(2):40-42.
[9]陈旸,肖民,杨朔,蒋磊.模糊神经网络在燃气轮机故障诊断专家系统中的研究与应用[J].科学技术与工程,2008,8(15):4342-4345.
[10]蒋磊等.基于知识的燃气轮机故障诊断系统的研究与开发[J].船舶,2009,20(2):27-30.
[11]夏迪,王永泓.基于热参数的单轴燃气轮机非线性故障诊断方法[J].上海交通大学学报,2008,42(3):381-384.
[12]夏迪,陈矫,王永泓,翁史烈.基于热力学参数的燃气轮机故障诊断系统分析软件[J].上海交通大学学报,2009,43(2).
[13] H.R.Depold, F.D.Gass. The Application of Expert Systems and Neural Networks to Gas Turbine Prognostics and Diagnostics[J].Journal of Engineering for Gas Turbines and Power,2003,121(4): 607-612.
[14] P.Cunningham, J.Carney, S.Jacob. Stability problems with artificial neural networks and the ensemble solution[J].Artificial Intelligence in Medicine,2000,20(3):217-225.
[15] K.K.Botros, G..Kibrya, A.Gtover. A Demonstration of Artificial Neural Networks Based Data Mining for Gas Turbine Driven Compressor Stations[J].Transaction of the ASME,2002,124(2):284-297.
[16] A.J.Volponi, H.Depold, R.Ganguli, C.Daguang. The Use of Kalman Filter and Neural Network Methodologies in Gas Turbine Performance Diagnostics: A ComparativeStudy[J].Journal of Engineering for Gas Turbines and Power,2003,125(4):917-924.
[17] S.O.T.Ogaji, L.Marinai, S.Sampath, R.Singh, S.D.Prober. Gas-turbine Fault Diagnostics: a fuzzy- logic approach[J].Journal of Applied Energy,2005,82(1):81-89.
[18] H.S.Tan. Fourier Neural Networks and Generalized Single Hidden Layer Networks in Aircraft Engine Fault Diagnostics[J].Journal of Engineering for Gas Turbines and Power,2006,128(4):773-782.
[19] Rajeev Verma, Niranjan Roy, Ranjan Ganguli. Gas Turbine Diagnostics Using a Soft Computing Approach[J].Applied Mathematics and Computations,2006,172(2):1342- 1363.
[20] R.Bettocchi, M.Pinelli, P.R.Spina, M.Venturini. Artificial Intelligence for the Diagnostics of Gas Turbines—Part II: Neural-Fuzzy Approach[J].Journal of Engineering for Gas Turbines and Power,2007,129(3):720-729.
[21] Beulens AJM, Vran Nunen JAEE. The Use of Expert System Technology in DSS[J]. Decision Support Systems.2001,(4):421-426,428-431.
[22] S.Edwards, A.W.Lees, M.I.Friswell. Fault Diagnosis of Rotating Machinery[J].Shock and Vibration,2002,30(1):4-13.
[23] Y.Shimshoni, N.Intrator. Classification of seismic signals by integrating ensembles of neural networks[J].IEEE Trans Signal Processing,2003,46(5):1194-1201.
[24]蔡自兴,<美>约翰.德尔金,龚涛.高级专家系统:原理、设计及应用[M].北京:科学出版社.2005.1-2,14-15,18-25.
[25]刘宏伟.具有智能故障诊断的燃气轮机中心数据库系统研究与开发[D].江苏科技大学,2007.
[26]任其智.燃气轮机的检修[M].北京:机械工业出版社.2004.1-175.
[27]张大威.铁路信号故障诊断专家系统探讨[J].铁道通信信号,2011,02(47):14-17.
[28]宋华.模糊故障诊断技术研究[D].北京航空航天大学. 2002.
[29]刘亚军.基于小波分析和神经网络的电机故障诊断研究[D].中北大学.2009,5.
[30]冯志鹏.计算智能在机械设备故障诊断中的应用研究[D].大连理工大学.2003,5.
[31] L.Hansen, P.Salamon. Neural network ensembles[J].IEEE Trans. Pattern Analysis and Machine Intelligence,2000,12:993-1001.
[32]朱四如.组合神经网络在时间序列中的应用[D].武汉理工大学.2007.
[33]李冲祥.神经网络和证据理论集成的数据融合故障诊断方法研究[D]燕山大学.2003,1.
[34]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,2005,75-100.
[35] D.W.Opitz, J.W.Shavlik. Generating accurate and diverse members of a neural networkensemble[J].Advances in Neural Information Processing Systems,2000:535-541.
[36] D.Partridge. Network generalization differences quantified[J].Neural Networks,2002,9: 263-271.
[37] A Lazarevic, Z.Obradovic. Effective pruning of neural network classifier ensembles[J]. International Joint Conference on Neural Networks,2001,2:796-801.
[38] Q.Fu, S.X.Hu, S.Y.Zhao. Clustering-based selective neural network ensemble[J].Journal of Zhejiang university (Science),2005,6(5):387-392.
[39] D.W.Opitz, J.W.Shavlik. Actively searching for an effective neural network ensemble[J]. Connection Science,2003,8(3-4):337-353.
[40] M.P.Perrone, L.N.Cooper. When networks disagree: ensemble method for neura1 network[J].Artificia1 Neural Networks for Speech and Vision,2001:126-142.
[41] P.Sollich, A.Krogh.Learning with ensembles: How over-fitting can be useful[J]. Advances in Neural Information Processing System,2004:190-196.
[42]王万良.人工智能及其应用[M].北京:高等教育出版社,2008:160-187.
[43]王光远,吕大刚等.结构智能选型:理论、方法与应用[M].北京:中国建筑工业出版社,2006:320-364.
[44]李娟莉,王绍进,杨兆建.基于ontology的提升机故障诊断系统研究[J]矿山机械, 2011,1(39):65-67,104.
[45]赵玉红.雷达故障诊断专家系统设计构思[J].雷达与对抗.2010,4(30):18-20.
[46]巴宁,梁玉英,丁国宝等.基于RCM理论的某型雷达故障诊断专家系统[J].信息技术, 2011,02:87-89.
[47]邓日青.基于神经网络专家系统的电喷发动机故障自诊断研究[D].重庆交通大学. 2008.
[48]蒋亚.专家系统在电力系统故障诊断中的应用[J].科技资讯,2011,2:112.
[49]王姣,祁美玲.RBR和CBR在数控机床故障诊断专家系统中的应用[J].组合机床与自动化加工技术,2011,1(1):47-49,53.
[50]王娜.基于神经网络专家系统的高压断路器故障诊断研究[D].河南理工大学.2008.11.
[51]石美,吴谦,李海波等.凝汽器故障诊断方法的研究[J].科技创新导报,2011,1:54.
[52]江志农,王慧,魏中青.基于案例与规则推理的故障诊断专家系统[J].计算机工程,2011,1(37):238-240,243.
[53]黄雷,南建国,隋永华等.基于模糊神经网络的平显设备故障诊断研究[J].计算机应用研究,2011,1(28):179-181.
[54]徐安,陈耀,李锐华,郭其一.微型计算机控制技术[M].北京:电子工业出版社.2008:229-251.
[55]冯定.神经网络专家系统[M].北京:科学出版社.2006:248-275.
[2]单旭昇.神经网络专家系统在凝汽器故障诊断中的应用研究[D].华北电力大学.2009,2.
[3]陈夏夏,陈静杰.基于粗糙集的机载电子设备故障检测专家系统的改进[J].河南科学,2011,01(29):77-79.
[4]熊英.人工智能及其在计算机网络技术中的应用[J].技术与市场,2011,2(18):20.
[5]谢建,何德虎,李良.基于观测器和神经网络的液压系统故障诊断方法研究[J].机床与液压,2011,1(39):135-137.
[6]黄小光,王永泓.燃气轮机混合智能故障诊断专家系统[J].船舶工程,2000,22(6):8-39.
[7]翁史烈,王永泓.基于热力参数的燃气轮机智能故障诊断[J].上海交通大学学报,2002,36(2):165-168.
[8]卜凯旗,任兴民,秦卫阳.燃气轮机发电机组的振动信号监测与分析系统[J].振动与噪声控制,2007,4(2):40-42.
[9]陈旸,肖民,杨朔,蒋磊.模糊神经网络在燃气轮机故障诊断专家系统中的研究与应用[J].科学技术与工程,2008,8(15):4342-4345.
[10]蒋磊等.基于知识的燃气轮机故障诊断系统的研究与开发[J].船舶,2009,20(2):27-30.
[11]夏迪,王永泓.基于热参数的单轴燃气轮机非线性故障诊断方法[J].上海交通大学学报,2008,42(3):381-384.
[12]夏迪,陈矫,王永泓,翁史烈.基于热力学参数的燃气轮机故障诊断系统分析软件[J].上海交通大学学报,2009,43(2).
[13] H.R.Depold, F.D.Gass. The Application of Expert Systems and Neural Networks to Gas Turbine Prognostics and Diagnostics[J].Journal of Engineering for Gas Turbines and Power,2003,121(4): 607-612.
[14] P.Cunningham, J.Carney, S.Jacob. Stability problems with artificial neural networks and the ensemble solution[J].Artificial Intelligence in Medicine,2000,20(3):217-225.
[15] K.K.Botros, G..Kibrya, A.Gtover. A Demonstration of Artificial Neural Networks Based Data Mining for Gas Turbine Driven Compressor Stations[J].Transaction of the ASME,2002,124(2):284-297.
[16] A.J.Volponi, H.Depold, R.Ganguli, C.Daguang. The Use of Kalman Filter and Neural Network Methodologies in Gas Turbine Performance Diagnostics: A ComparativeStudy[J].Journal of Engineering for Gas Turbines and Power,2003,125(4):917-924.
[17] S.O.T.Ogaji, L.Marinai, S.Sampath, R.Singh, S.D.Prober. Gas-turbine Fault Diagnostics: a fuzzy- logic approach[J].Journal of Applied Energy,2005,82(1):81-89.
[18] H.S.Tan. Fourier Neural Networks and Generalized Single Hidden Layer Networks in Aircraft Engine Fault Diagnostics[J].Journal of Engineering for Gas Turbines and Power,2006,128(4):773-782.
[19] Rajeev Verma, Niranjan Roy, Ranjan Ganguli. Gas Turbine Diagnostics Using a Soft Computing Approach[J].Applied Mathematics and Computations,2006,172(2):1342- 1363.
[20] R.Bettocchi, M.Pinelli, P.R.Spina, M.Venturini. Artificial Intelligence for the Diagnostics of Gas Turbines—Part II: Neural-Fuzzy Approach[J].Journal of Engineering for Gas Turbines and Power,2007,129(3):720-729.
[21] Beulens AJM, Vran Nunen JAEE. The Use of Expert System Technology in DSS[J]. Decision Support Systems.2001,(4):421-426,428-431.
[22] S.Edwards, A.W.Lees, M.I.Friswell. Fault Diagnosis of Rotating Machinery[J].Shock and Vibration,2002,30(1):4-13.
[23] Y.Shimshoni, N.Intrator. Classification of seismic signals by integrating ensembles of neural networks[J].IEEE Trans Signal Processing,2003,46(5):1194-1201.
[24]蔡自兴,<美>约翰.德尔金,龚涛.高级专家系统:原理、设计及应用[M].北京:科学出版社.2005.1-2,14-15,18-25.
[25]刘宏伟.具有智能故障诊断的燃气轮机中心数据库系统研究与开发[D].江苏科技大学,2007.
[26]任其智.燃气轮机的检修[M].北京:机械工业出版社.2004.1-175.
[27]张大威.铁路信号故障诊断专家系统探讨[J].铁道通信信号,2011,02(47):14-17.
[28]宋华.模糊故障诊断技术研究[D].北京航空航天大学. 2002.
[29]刘亚军.基于小波分析和神经网络的电机故障诊断研究[D].中北大学.2009,5.
[30]冯志鹏.计算智能在机械设备故障诊断中的应用研究[D].大连理工大学.2003,5.
[31] L.Hansen, P.Salamon. Neural network ensembles[J].IEEE Trans. Pattern Analysis and Machine Intelligence,2000,12:993-1001.
[32]朱四如.组合神经网络在时间序列中的应用[D].武汉理工大学.2007.
[33]李冲祥.神经网络和证据理论集成的数据融合故障诊断方法研究[D]燕山大学.2003,1.
[34]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,2005,75-100.
[35] D.W.Opitz, J.W.Shavlik. Generating accurate and diverse members of a neural networkensemble[J].Advances in Neural Information Processing Systems,2000:535-541.
[36] D.Partridge. Network generalization differences quantified[J].Neural Networks,2002,9: 263-271.
[37] A Lazarevic, Z.Obradovic. Effective pruning of neural network classifier ensembles[J]. International Joint Conference on Neural Networks,2001,2:796-801.
[38] Q.Fu, S.X.Hu, S.Y.Zhao. Clustering-based selective neural network ensemble[J].Journal of Zhejiang university (Science),2005,6(5):387-392.
[39] D.W.Opitz, J.W.Shavlik. Actively searching for an effective neural network ensemble[J]. Connection Science,2003,8(3-4):337-353.
[40] M.P.Perrone, L.N.Cooper. When networks disagree: ensemble method for neura1 network[J].Artificia1 Neural Networks for Speech and Vision,2001:126-142.
[41] P.Sollich, A.Krogh.Learning with ensembles: How over-fitting can be useful[J]. Advances in Neural Information Processing System,2004:190-196.
[42]王万良.人工智能及其应用[M].北京:高等教育出版社,2008:160-187.
[43]王光远,吕大刚等.结构智能选型:理论、方法与应用[M].北京:中国建筑工业出版社,2006:320-364.
[44]李娟莉,王绍进,杨兆建.基于ontology的提升机故障诊断系统研究[J]矿山机械, 2011,1(39):65-67,104.
[45]赵玉红.雷达故障诊断专家系统设计构思[J].雷达与对抗.2010,4(30):18-20.
[46]巴宁,梁玉英,丁国宝等.基于RCM理论的某型雷达故障诊断专家系统[J].信息技术, 2011,02:87-89.
[47]邓日青.基于神经网络专家系统的电喷发动机故障自诊断研究[D].重庆交通大学. 2008.
[48]蒋亚.专家系统在电力系统故障诊断中的应用[J].科技资讯,2011,2:112.
[49]王姣,祁美玲.RBR和CBR在数控机床故障诊断专家系统中的应用[J].组合机床与自动化加工技术,2011,1(1):47-49,53.
[50]王娜.基于神经网络专家系统的高压断路器故障诊断研究[D].河南理工大学.2008.11.
[51]石美,吴谦,李海波等.凝汽器故障诊断方法的研究[J].科技创新导报,2011,1:54.
[52]江志农,王慧,魏中青.基于案例与规则推理的故障诊断专家系统[J].计算机工程,2011,1(37):238-240,243.
[53]黄雷,南建国,隋永华等.基于模糊神经网络的平显设备故障诊断研究[J].计算机应用研究,2011,1(28):179-181.
[54]徐安,陈耀,李锐华,郭其一.微型计算机控制技术[M].北京:电子工业出版社.2008:229-251.
[55]冯定.神经网络专家系统[M].北京:科学出版社.2006:248-275.