王卫玉;侯凯;何葵东;莫凡;金艳;陈启卷;

• 1:五凌电力有限公司
• 2:国家电力投资集团水电产业创新中心
• 3:武汉大学动力与机械学院

摘要(Abstract)：

提出了基于引力搜索算法（gravitational search algorithm, GSA）优化变分模态分解（variational mode decomposition, VMD）并结合深度置信网络（deep belief network, DBN）的水电机组故障诊断方法。首先，以变分模态分解能量误差最小化为目标，利用GSA并行优化各样本下VMD的2个典型分解参数（分解层数K和惩罚因子α）；然后，对分解降噪后的信号进行重构，并对重构后的信号构造由能量熵、奇异谱熵以及峭度组成的特征向量；最后，将构建的特征向量输入DBN构建的水电机组故障诊断模型。通过与已有方法比较可知，所提模型可以有效地提取水电机组的故障特征，且故障识别准确率更高。

关键词(KeyWords)： 水电机组;故障诊断;深度置信网络;变分模态分解;引力搜索

Abstract:

Keywords:

基金项目(Foundation): 国家电力投资集团统筹科研项目（编号：TC2020SD01）

作者(Authors): 王卫玉;侯凯;何葵东;莫凡;金艳;陈启卷;

DOI: 10.14188/j.1671-8844.2023-02-015

参考文献(References)：

• [1]周宁刚,徐剑峰,江书樵,等.基于概率统计的水电机组状态监测参数波动预警策略[J].水电与抽水蓄能,2021, 7(4):102-106.Zhou Ninggang, Xu Jianfeng, Jiang Shuqiao, et al.Fluctuating alarm for monitoring parameters of hydropower unit based on probability statistics[J]. Hydropower and Pumped Storage, 2021, 7(4):102-106.
• [2]陈珊珊,杨耿杰.水电机组振动故障诊断方法综述[J].电气技术, 2019,20(6):1-5.Chen Shanshan, Yang Gengjie. Review of vibration fault diagnosis methods for hydraulic turbines[J]. Electrical Engineering, 2019,20(6):1-5.
• [3]付杰.基于经验模态分解和神经网络的水电机组振动故障诊断[D].武汉:武汉大学, 2018.Fu Jie. Fault diagnosis of vibration for hydro-power unit based on empirical mode decomposition and neural network[D]. Wuhan:Wuhan University, 2018.
• [4] Wu Z H, Huang N E. Ensemble empirical mode decomposition:a noise-assisted data analysis method[J]. Advances in Adaptive Data Analysis, 2009, 1(1):1-41.
• [5]李辉.基于经验模态分解的水电机组振动信号分析与故障诊断研究[D].西安:西安理工大学, 2017.Li Hui. Research on vibration signal analysis and fault diagnosis of hydroelectric generating units based on empirical mode decomposition[D]. Xi′an:Xi′an University of Technology, 2017.
• [6]唐拥军,刘东,肖志怀,等.基于卷积神经网络与奇异值分解的水电机组故障诊断方法研究[J].中国农村水利水电, 2021,(2):175-181.Tang Yongjun,Liu Dong,Xiao Zhihuai,et al. Fault diagnosis of hydropower units based on convolutional neural network and singular value decomposition[J]. China Rural Water and Hydropower, 2021,(2):175-181.
• [7]何葵东,陈伽,金艳,等. EEMD多尺度熵和ELM在水电机组振动信号特征提取中的应用[J].中国农村水利水电, 2021,(5):176-182,187.He Kuidong, Chen Jia, Jin Yan, et al. Application of EEMD multi-scale entropy and ELM in feature extraction of vibration signal of hydropower unit[J]. China Rural Water and Hydropower, 2021,(5):176-182.
• [8]孙慧影,林中鹏,刘银丽,等.基于布谷鸟搜索算法优化BP模糊Petri网的水电机组故障诊断[J].水电能源科学, 2017,35(7):179-181,171.Sun Huiying, Lin Zhongpeng, Liu Yinli, et al. Fault diagnosis of hydroelectric set based on cuckoo search algorithm evolving BP fuzzy Petri nets[J]. Water Resources and Power, 2017,35(7):179-181,171.
• [9]李辉,王毅,杨晓萍,等.基于多小波和PSO-RBF神经网络的水电机组振动故障诊断[J].西北农林科技大学学报(自然科学版), 2017, 45(2):227-234.Li Hui, Wang Yi, Yang Xiaoping, et al. Vibration fault diagnosis for hydropower generating unit based on multiwavelet and PSO-RBF neural network[J]. Journal of Northwest A&F University(Natural Science Edition),2017, 45(2):227-234.
• [10] Sutskever I, Hinton G E. Deep, narrow Sigmoid belief networks are universal approximators[J]. Neural Computation, 2008, 20(11):2629-2636.
• [11] Mohamed A R, Dahl G E, Hinton G. Acoustic modeling using deep belief networks[J]. IEEE Transactions on Audio, Speech, and Language Processing, 2012, 20(1):14-22.
• [12] Shao J K, Niu Y F, Xue C Q, et al. Single-channel SEMG using wavelet deep belief networks for upper limb motion recognition[J]. International Journal of Industrial Ergonomics, 2020, 76:102905.
• [13] Dragomiretskiy K, Zosso D. Variational mode decomposition[J]. IEEE Transactions on Signal Processing,2014, 62(3):531-544.
• [14] Li H, Liu T, Wu X, et al. An optimized VMD method and its applications in bearing fault diagnosis[J].Measurement, 2020, 166:108185.
• [15] Rashedi E, Nezamabadi-pour H. Feature subset selection using improved binary gravitational search algorithm[J]. Journal of Intelligent&Fuzzy Systems,2014, 26(3):1211-1221.
• [16]侯鑫烨,董增寿,刘鑫,等.基于能量熵和CL-LSTM的故障诊断模型[J].机床与液压, 2021, 49(16):180-184.Hou Xinhua, Dong Zengshou, Liu Xin, et al. Fault diagnosis model based on energy entropy and CLLSTM[J]. Machine Tool&Hydraulics, 2021, 49(16):180-184.
• [17]张淑清,陈荣飞,张立国,等.基于多尺度高阶奇异谱熵的信号特征提取方法[J].计量学报, 2019, 40(5):848-854.Zhang Shuqing, Chen Rongfei, Zhang Liguo, et al. Signal feature extraction method based on multi-scale highorder singular spectrum entropy[J]. Acta Metrologica Sinica, 2019, 40(5):848-854.
• [18]张萍,张文海,赵新贺,等. WOA-VMD算法在轴承故障诊断中的应用[J].噪声与振动控制, 2021, 41(4):86-93,275.Zhang Ping, Zhang Wenhai, Zhao Xinhe, et al. Application of WOA-VMD Algorithm in bearing fault diagnosis[J]. Noise and Vibration Control, 2021, 41(4):86-93,275.