基于最小二乘支持向量机的磨损预测
|
摘要
针对机器设备磨损产生的因素多,而且磨损量的多少与产生的因素具有高度非线性,磨损难以预测的问题,同时考虑到监测得到的数据为小样本事件也是磨损难以预测的原因,在齿轮箱实验数据的基础上,利用最小二乘支持向量机,给出预测步骤,提出一种以载荷、温度、振动信号特征、速度和时间为输入量,机器设备的磨损量为输出量的预测方法。用齿轮箱的实验数据验证了所提出的方法的有效性。
The wear loss of the machine is affected by many factors and the relationship between wear loss and factors is a kind of typical nonlinear,so how to forecast the wear loss is very difficult.In the same time,small monitoring samples data is one of the reasons that the wear loss can't be predicted accurately.In order to predict wear loss accurately,least-square-support-vector machine was chose to build up a new method to predict the wear loss of the gear box.Load,the features of vibration signal,temperature,velocity and time of the gear box were chose as inputs,the wear loss was chose as output.The detail steps of the method was shown and discussed.An experiment about a gear box was introduced to verify the method and the result shows the method is effective.
The wear loss of the machine is affected by many factors and the relationship between wear loss and factors is a kind of typical nonlinear,so how to forecast the wear loss is very difficult.In the same time,small monitoring samples data is one of the reasons that the wear loss can't be predicted accurately.In order to predict wear loss accurately,least-square-support-vector machine was chose to build up a new method to predict the wear loss of the gear box.Load,the features of vibration signal,temperature,velocity and time of the gear box were chose as inputs,the wear loss was chose as output.The detail steps of the method was shown and discussed.An experiment about a gear box was introduced to verify the method and the result shows the method is effective.
引文
【1】马智峰,李晓峰.柴油机磨损趋势预测[J].润滑与密封,2000(1):54-56.Ma Zhifeng,Li Xiaofeng.The forecast for the wear trend of thediesel engine[J].LubricationEngineering,2000(1):54-56.
【2】郑建明,李言,肖继明,等.基于神经网络的多特征融合刀具磨损量识别[J].机械科学与技术,2002(1):111-113.ZHENG Jianming,LI Yan,XIAO Jiming,et al.OnIdentification of Tool Wear by Multiple Features Fusion Basedon Artificial Neural Network[J].Mechanical Science andTechnology,2002(1):111-113.
【3】Lee J H,KimD E,Lee S J.Application of neural networks toflank wear prediction[J].Mechanical Systems&SignalProcessing,1996,10(3):265-276.
【4】Stolarski TA.System for wear prediction in lubricated slidingcontacts[J].Lubrication Science,1996,8(4):315-351.
【5】Zhou G Y,Leu M C,Blackmore D.Fractal geometry modelfor wear prediction[J].Wear,1993,170(1):1-14.
【6】Cherkassky V,Mulier F.Learning from Data:Concepts,Theory and Methods[M].John Wiley and Sons Inc,1998.
【7】Vapnik V.The Nature of Statistical Learning Theory[M].Springer-Verlag,1995.
【8】郭崇慧,陆玉昌.预测型数据挖掘中的优化方法[J].工程数学学报,2005,22(1):25-29.GUO Chonghui,LU Yuchang.Optimization Methods inPredictive Data Mining[J].Chinese Journal of EngineeringMathematics,2005,22(1):25-29.
【9】Cortes C,Vapnik V.Support-vector networks[J].MachineLearning,1995,20:273-297.
【10】Suykens J A K,Vandewalle J.Least squares support vectormachine classifiers[J].Neural Processing Letters,1999(3):293-300.
【11】Burges J C.Atutorial on support vector machines for patternrecognition[J].Data Mining and Knowledge Discovery,1998,2(2):121-167.
【12】温诗铸,黄平.摩擦学原理[M].2版.北京:清华大学出版社,2002:301.
【13】金龙,陈小宏,王守东.基于支持向量机与信息融合的地震油气预测方法[J].石油地球物理勘探,2006,41(1):76-80.Jin Long,Chen Xiaohong,Wang Shoudong.Oil/gas-predicted method by seismology based on supporting vectorcomputer and information merging[J].Oil GeophysicalProspecting,2006,41(1):76-80.
【14】张铃.基于核函数的SVM机与三层前向神经网络的关系[J].计算机学报,2002,25(7):696-700.ZHANG Ling.The Relationship between Kernel FunctionsBased SVM and Three-Layer Feed-forward Neural Networks[J].Chinese Journal of Computers,2002,25(7):696-700.
【2】郑建明,李言,肖继明,等.基于神经网络的多特征融合刀具磨损量识别[J].机械科学与技术,2002(1):111-113.ZHENG Jianming,LI Yan,XIAO Jiming,et al.OnIdentification of Tool Wear by Multiple Features Fusion Basedon Artificial Neural Network[J].Mechanical Science andTechnology,2002(1):111-113.
【3】Lee J H,KimD E,Lee S J.Application of neural networks toflank wear prediction[J].Mechanical Systems&SignalProcessing,1996,10(3):265-276.
【4】Stolarski TA.System for wear prediction in lubricated slidingcontacts[J].Lubrication Science,1996,8(4):315-351.
【5】Zhou G Y,Leu M C,Blackmore D.Fractal geometry modelfor wear prediction[J].Wear,1993,170(1):1-14.
【6】Cherkassky V,Mulier F.Learning from Data:Concepts,Theory and Methods[M].John Wiley and Sons Inc,1998.
【7】Vapnik V.The Nature of Statistical Learning Theory[M].Springer-Verlag,1995.
【8】郭崇慧,陆玉昌.预测型数据挖掘中的优化方法[J].工程数学学报,2005,22(1):25-29.GUO Chonghui,LU Yuchang.Optimization Methods inPredictive Data Mining[J].Chinese Journal of EngineeringMathematics,2005,22(1):25-29.
【9】Cortes C,Vapnik V.Support-vector networks[J].MachineLearning,1995,20:273-297.
【10】Suykens J A K,Vandewalle J.Least squares support vectormachine classifiers[J].Neural Processing Letters,1999(3):293-300.
【11】Burges J C.Atutorial on support vector machines for patternrecognition[J].Data Mining and Knowledge Discovery,1998,2(2):121-167.
【12】温诗铸,黄平.摩擦学原理[M].2版.北京:清华大学出版社,2002:301.
【13】金龙,陈小宏,王守东.基于支持向量机与信息融合的地震油气预测方法[J].石油地球物理勘探,2006,41(1):76-80.Jin Long,Chen Xiaohong,Wang Shoudong.Oil/gas-predicted method by seismology based on supporting vectorcomputer and information merging[J].Oil GeophysicalProspecting,2006,41(1):76-80.
【14】张铃.基于核函数的SVM机与三层前向神经网络的关系[J].计算机学报,2002,25(7):696-700.ZHANG Ling.The Relationship between Kernel FunctionsBased SVM and Three-Layer Feed-forward Neural Networks[J].Chinese Journal of Computers,2002,25(7):696-700.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心