FPCA-RBF-ELM-Based Gearbox Fault Detection Method
-
摘要: 为了克服在数据处理中出现的信息缺失和冗余以及在故障检测上准确率较低等缺陷,利用函数型主成分所具有的鲁棒性和稳定性强的优点来弥补极限学习机在稳定性方面的不足,结合径向基极限学习机,提出了一种基于FPCA(函数型主成分分析)-RBF(径向基函数)-ELM(极限学习机)的齿轮箱故障检测方法。首先用基函数对原始数据进行预处理,然后应用FPCA提取特征信息建立RBF-ELM齿轮诊断模型,最后利用行星齿轮箱实验数据验证故障检测性能,并与FPCA、FPCA-SVDD和PCA-RBF-ELM的行星齿轮箱故障检测结果对比。结果表明:FPCA-RBF-ELM检测率最高且检测效率快,可用于行星齿轮箱的故障检测,此方法具有可行性和有效性。Abstract: In order to overcome the lack of information in the data processing and redundancy and low defect on the fault detection accuracy, using functional principal component with the advantages of strong robustness and stability to fill the extreme learning machine in the lack of stability, combined with the RBF-ELM, a new gear fault detection model was proposed based on FPCA (principal component analysis)-RBF(radial basis function)-ELM(extreme learning machine) method. First, the basis function was used to preprocess the original gear vibration data, and then FPCA was used to extract the characteristic information as the training set to establish the ELM gear fault diagnosis model. Finally, the planetary gear box experimental data were used to verify the fault detection performance, and the fault detection results of the planetary gear box were compared with those of FPCA, FPCA-SVDD and PAC-RBF-ELM. The results show that the FPCA-RBF-ELM method has the highest detection rate and fastest detection efficiency, which can be used for fault detection of planetary gear box. This method is effective and feasible.
-
表 1 故障类型及编号
故障类型 编号 正常 1 磨损 2 缺齿 3 齿根裂纹 4 表 2 主成分累计贡献率
主成分个数 累计贡献值/% 垂直方向 水平径向 轴向 1 50.78 47.95 45.94 2 73.42 73.83 72.34 3 84.48 87.63 85.16 4 92.31 94.89 91.88 表 3 4种法类比
检测方法 训练集
准确率/%测试集
准确率/%反应
时间/sFPCA 95 90 9.233 FPCA-SVDD 90 97 12.632 PCA-RBF-ELM 93 95 6.526 FPCA-RBF-ELM 100 100 6.188 -
[1] 周伟. 基于多元统计与信息熵的系统监测方法研究[D]. 杭州: 浙江理工大学, 2019Zhou W. Research on system monitoring method based on multivariate statistics and information entropy[D]. Hangzhou: Zhejiang Sci-Tech University, 2019 (in Chinese) [2] Yang Y H, Lu N Y, Wang F L, et al. Statistical process monitoring using multiple PCA models[C]//Proceedings of the 2002 American Control Conference. Anchorage: IEEE, 2002: 5072-5073 [3] Alcala C F, Qin S J. Analysis and generalization of fault diagnosis methods for process monitoring[J]. Journal of Process Control, 2011, 21(3): 322-330 [4] Yang F, Xiao D Y, Shah S L. Signed directed graph-based hierarchical modelling and fault propagation analysis for large-scale systems[J]. IET Control Theory & Applications, 2013, 7(4): 537-550 [5] 张珂, 宋文丽, 石怀涛, 等. 基于改进核主元分析的故障检测方法研究[J]. 控制工程, 2017, 24(2): 418-424Zhang K, Song W L, Shi H T, et al. Fault detection based on improved kernel principal component analysis[J]. Control Engineering of China, 2017, 24(2): 418-424 (in Chinese) [6] 靳刘蕊. 函数性数据分析方法及应用研究[D]. 厦门: 厦门大学, 2008Jin L R. The study on the methods of functional data analysis and their application[D]. Xiamen: Xiamen University, 2008 (in Chinese) [7] 马峻, 赵飞乐, 徐潇, 等. MRA-PCA-PSO组合优化BP神经网络模拟电路故障诊断研究[J]. 电子测量与仪器学报, 2018, 32(3): 73-79Ma J, Zhao F L, Xu X, et al. Fault diagnosis of analog circuit based on optimized BP neural network with MRA-PCA-PSO technology[J]. Journal of Electronic Measurement and Instrumentation, 2018, 32(3): 73-79 (in Chinese) [8] 赵璐, 马野. 基于一维卷积神经网络的齿轮箱故障诊断研究[J]. 测试技术学报, 2019, 33(4): 302-306 doi: 10.3969/j.issn.1671-7449.2019.04.006Zhao L, Ma Y. Fault diagnosis of gear box based on one-dimensional convolutional neural networks[J]. Journal of Test and Measurement Technology, 2019, 33(4): 302-306 (in Chinese) doi: 10.3969/j.issn.1671-7449.2019.04.006 [9] 张淑清, 胡永涛, 姜安琦, 等. 基于双树复小波和深度信念网络的轴承故障诊断[J]. 中国机械工程, 2017, 28(5): 532-536, 543 doi: 10.3969/j.issn.1004-132X.2017.05.005Zhang S Q, Hu Y T, Jiang A Q, et al. Bearing fault diagnosis based on DTCWT and DBN[J]. China Mechanical Engineering, 2017, 28(5): 532-536, 543 (in Chinese) doi: 10.3969/j.issn.1004-132X.2017.05.005 [10] 林桐, 陈果, 滕春禹, 等. 基于超球优化支持向量数据描述的滚动轴承故障检测[J]. 振动与冲击, 2019, 38(2): 204-210, 225Lin T, Chen G, Teng C Y, et al. Rolling bearing fault detection based on the hypersphere optimization support vector data description[J]. Journal of Vibration and Shock, 2019, 38(2): 204-210, 225 (in Chinese) [11] 王博林. 基于极限学习机的分类问题研究与应用[D]. 辽宁大连: 辽宁师范大学, 2018Wang B L. Research and application of classification problem based on extreme learning machine[D]. Liaoning Dalian: Liaoning Normal University, 2018 (in Chinese) [12] 梁银双. 基于函数型数据分析的京津冀空气污染问题研究[D]. 北京: 首都经济贸易大学, 2017Liang Y S. Research on air pollution in Beijing-Tianjin-Hebei based on functional data analysis[D]. Beijing: Capital University of Economics and Business, 2017 (in Chinese) [13] 翟俊海, 张素芳, 胡文祥, 等. 核心集径向基函数极限学习机[J]. 山东大学学报, 2015, 46(2): 1-5, 13 doi: 10.6040/j.issn.1672-3961.1.2014.095Zhai J H, Zhang S F, Hu W X, et al. Radial basis function extreme learning machine based on core sets[J]. Journal of Shandong University , 2015, 46(2): 1-5, 13 (in Chinese) doi: 10.6040/j.issn.1672-3961.1.2014.095 [14] Yildirim H, Özkale M R. The performance of ELM based ridge regression via the regularization parameters[J]. Expert Systems with Applications, 2019, 134: 225-233 [15] Wu Y Q, Boyle L N, Mcgehee D V. Evaluating variability in foot to pedal movements using functional principal components analysis[J]. Accident Analysis & Prevention, 2018, 118: 146-153