基于振动能量趋势预测和K均值聚类的水电机组故障预警方法研究
|
摘要
针对当前常用的基于基线值和3σ方法的固定阈值预警方法存在预警不及时问题,本文提出了一种基于振动信号Teager能量的趋势预测和K均值聚类的预警方法。该方法通过预测值和历史能量序列的聚类完成信号能量"跃迁"的判断,进而识别机组运行状态的变化以实现故障预警。经实际数据验证,本文提出的方法可有效对机组突变故障进行预警。
Because current fixed threshold warning method based on baseline value and 3σ method for hydropower unit faults exist a problem that the warning is not timely,an early warning method based on the trend of vibration signal Teager energy and the K-means clustering is proposed. The method completes the judgment of signal energy"jump"by clustering predicted value and historical energy sequence,and then identifies the change of unit operating state to complete the fault warning. The actual data proves that the proposed method can effectively predict the abrupt fault of unit.
Because current fixed threshold warning method based on baseline value and 3σ method for hydropower unit faults exist a problem that the warning is not timely,an early warning method based on the trend of vibration signal Teager energy and the K-means clustering is proposed. The method completes the judgment of signal energy"jump"by clustering predicted value and historical energy sequence,and then identifies the change of unit operating state to complete the fault warning. The actual data proves that the proposed method can effectively predict the abrupt fault of unit.
引文
[1]陈安华,钟钟掘.关于机械系统突发性故障的若干思考[J].华中理工大学学报,1998(1):98-100.
[2]刘树英,宋雪萍,闻邦椿.转子系统突变故障研究[J].振动工程学报,2004,17(s1):159-162.
[3]孙慧芳,桂中华,李林扬,等.基于时间序列组合模型的水电机组状态趋势预测[J].水力发电学报,2016,35(1):79-86.
[4]安学利,潘罗平,张飞.基于三维曲面的抽水蓄能电站机组故障预警模型[J].水力发电,2013,39(1):71-74.
[5]孙慧芳,潘罗平.基于健康样本的水电机组故障诊断研究[C]∥中国水力发电工程学会信息化专委会、水电控制设备专委会2013年学术交流会论文集.北京:中国水力发电工程学会,2013.
[6]武和雷,朱善安,林瑞仲,等.基于能量算子解调法的滚动轴承故障诊断[J].农业机械学报,2003,34(1):118-120.
[7]王天金,冯志鹏,郝如江,等.基于Teager能量算子的滚动轴承故障诊断研究[J].振动与冲击,2012,31(2):1-5.
[8]KAISER J F.On a simple algorithm to calculate the'energy'of a signal[C]∥International Conference on Acoustics,Speech and Signal Processing,1990(1):381-384.
[9]王洪波,朱启兵.基于EMD和LS-SVM的非平稳振动信号趋势预测[J].计算机工程与应用,2008,44(16):157-159.
[10]孟庆芳.非线性动力系统时间序列分析方法及其应用研究[D].济南:山东大学,2008.
[11]周志华.机器学习:Machine Learning[M].北京:清华大学出版社,2016.
[12]张胜,刘红星,高敦堂,等.ANN非线性时间序列预测模型输入延时τ的确定[J].东南大学学报(自然科学版),2002,32(6):905-908.
[13]GOODFELLOW I,BENGIO Y,COURVILLE A.Deep Learning[M].Massachusetts:MIT Press,2016.
[14]赖玉霞,刘建平,杨国兴.基于遗传算法的K均值聚类分析[J].计算机工程,2008,34(20):200-202.
[2]刘树英,宋雪萍,闻邦椿.转子系统突变故障研究[J].振动工程学报,2004,17(s1):159-162.
[3]孙慧芳,桂中华,李林扬,等.基于时间序列组合模型的水电机组状态趋势预测[J].水力发电学报,2016,35(1):79-86.
[4]安学利,潘罗平,张飞.基于三维曲面的抽水蓄能电站机组故障预警模型[J].水力发电,2013,39(1):71-74.
[5]孙慧芳,潘罗平.基于健康样本的水电机组故障诊断研究[C]∥中国水力发电工程学会信息化专委会、水电控制设备专委会2013年学术交流会论文集.北京:中国水力发电工程学会,2013.
[6]武和雷,朱善安,林瑞仲,等.基于能量算子解调法的滚动轴承故障诊断[J].农业机械学报,2003,34(1):118-120.
[7]王天金,冯志鹏,郝如江,等.基于Teager能量算子的滚动轴承故障诊断研究[J].振动与冲击,2012,31(2):1-5.
[8]KAISER J F.On a simple algorithm to calculate the'energy'of a signal[C]∥International Conference on Acoustics,Speech and Signal Processing,1990(1):381-384.
[9]王洪波,朱启兵.基于EMD和LS-SVM的非平稳振动信号趋势预测[J].计算机工程与应用,2008,44(16):157-159.
[10]孟庆芳.非线性动力系统时间序列分析方法及其应用研究[D].济南:山东大学,2008.
[11]周志华.机器学习:Machine Learning[M].北京:清华大学出版社,2016.
[12]张胜,刘红星,高敦堂,等.ANN非线性时间序列预测模型输入延时τ的确定[J].东南大学学报(自然科学版),2002,32(6):905-908.
[13]GOODFELLOW I,BENGIO Y,COURVILLE A.Deep Learning[M].Massachusetts:MIT Press,2016.
[14]赖玉霞,刘建平,杨国兴.基于遗传算法的K均值聚类分析[J].计算机工程,2008,34(20):200-202.