基于四维张量特征分解的风电机组轴承故障缺失数据恢复方法研究

doi:10.3969/j.issn.1000-1158.2024.05.18

摘要
图/表
参考文献(21)
相关文章 (15)

全文: PDF (670 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要针对风电机组轴承故障信息采集过程中数据缺失导致故障类型无法识别问题,提出了一种基于四维张量模型特征分解恢复缺失数据的方法。首先,基于转速、时窗、经验模态分解和时间4个维度构建四维张量;其次,通过加权优化算法实现张量填充,修补故障数据的缺失值;然后,对张量进行Tucker分解得到核心张量及因子矩阵;最后,基于梯度优化算法进行迭代优化得到最终核心张量及因子矩阵,并利用二者对四维张量进行重构得到恢复数据。采用实验数据和实际数据来验证提出方法的有效性和可靠性。结果表明:两组恢复数据的RMSE值分别为0.3169和0.0291,远小于4种对比方法的RMSE值。利用双稳态随机共振对2组恢复数据进行故障特征提取,信噪比显著提高,分别为-13.2647和-15.5212,进一步验证提出方法的准确性。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	时培明
	孙航璇
	许学方
	韩东颖

关键词 ：信息采集, 数据恢复;轴承故障诊断, 张量分解, 缺失值数据, 特征提取, 振动测量;风电机组

Abstract：To solve the problem that the fault type cannot be identified due to missing data in the process of collecting fault information of wind turbine bearing, a method based on four-dimensional tensor model feature decomposition is proposed to recover missing data. Firstly, the fourth order tensor of vibration fault data is constructed based on the four dimensions of rotational speed, time-domain window, empirical mode decomposition and time.Secondly, the factor matrix is obtained by Tucker decomposition of the tensor. Then, tensor filling is realized by weighted optimization algorithm to recover the missing values of bearing fault data.Finally, the core tensor and factor matrix are obtained iteratively based on the gradient optimization algorithm, and the four-dimensional tensor is reconstructed to get the recovered data. The validity and reliability of the proposed method are verified by experimental and practical data. The results show that the RMSE values of the two groups of recovered data are 0.3169 and 0.0291, which are much lower than the RMSE values of the four comparison methods. By using BSR to extract fault features from two groups of recovered data, the signal-to-noise ratio is significantly improved to -13.2647 and -15.5212, respectively, which further verifies the accuracy of the proposed method.

Key words： information collection data restoration fault diagnosis of bearing tensor decomposition missing value data characteristic decomposition vibration measurement;wind turbine generator

收稿日期: 2023-07-13 发布日期: 2024-05-23

PACS:

TB973

基金资助:河北省自然科学基金(E2020203147,E2022203093);中央引导地方科技发展资金(216Z4301G)

作者简介: 时培明(1979-),男,黑龙江延寿人,主要研究方向为智能信息处理和设备监测诊断。Email: spm@ysu.edu.cn

引用本文:

时培明,孙航璇,许学方,韩东颖. 基于四维张量特征分解的风电机组轴承故障缺失数据恢复方法研究[J]. 计量学报, 2024, 45(5): 738-746.
SHI Peiming,SUN Hangxuan,XU Xuefang,HAN Dongying. A Method for Recovering Missing Data of Bearing Faults of Wind Turbine Generator Based on Four-dimensional Tensor Model Characteristic Decomposition. Acta Metrologica Sinica, 2024, 45(5): 738-746.

链接本文:

http://jlxb.china-csm.org:81/Jwk_jlxb/CN/10.3969/j.issn.1000-1158.2024.05.18 或 http://jlxb.china-csm.org:81/Jwk_jlxb/CN/Y2024/V45/I5/738

［6］	ZHANG Y, THORBURN P. A dual-head attention model for time series data imputation［J］. Computers and Electronics in Agriculture, 2021, 189: 106377.
［3］	余晓霞, 汤宝平, 王伟影, 等. 复杂工况条件下多头注意力双向长短时记忆网络的风电机组缺失数据修复方法研究［J］. 机械工程学报, 2023, 59(14): 1-9.
［9］	LI D, LI L H, LI X H, et al. A spatio-temporal deep model for multiple time-series missing imputation［J］. Neurocomputing, 2020, 411: 351-363.
［1］	董兴辉, 马晓双, 程友星, 等. 风电机组轴承健康劣化趋势建模与仿真［J］. 系统仿真学报, 2019, 31(1): 151-158.
［2］	彭进, 王维庆, 王海云, 等. 基于EEMD峭度-相关系数准则的多特征量风电机组轴承故障诊断［J］. 可再生能源, 2016 (10): 1481-1490.
［4］	史海鹏, 陈家兑, 吴永明, 等. 数据缺失条件下基于ANFIS与k-means的轴承故障分析［J］. 组合机床与自动化加工技术, 2020, 559(9): 33-36.
［5］	雷亚国, 许学方, 蔡潇, 等. 面向机械装备健康监测的数据质量保障方法研究［J］. 机械工程学报, 2021, 57(4): 1-9.
	DONG X H, MA X S, CHENG Y X, et al. Modeling and Simulation of Bearing Health Deterioration Trend of Wind Turbine ［J］. Journal of System Simulation, 2019, 31(1): 151-158.
	PENG J, WANG W Q, WANG H Y, et al. Bearing Fault Diagnosis of multi-feature wind turbine based on EEMD kurtosis Correlation coefficient criterion ［J］. Renewable Energy, 2016 (10): 1481-1490.
	YU X X, TANG B P, WANG W Y, et al. Research on missing data repair method of wind turbine based on multi-head attention bidirectional short-time memory network under complex working conditions ［J］. Journal of Mechanical, 2023, 59(14): 1-9.
［7］	张晟斐, 李天梅, 胡昌华, 等. 基于深度卷积生成对抗网络的缺失数据生成方法及其在剩余寿命预测中的应用［J］. 航空学报, 2022, 43(8): 225708.
［10］	唐波, 陈慎慎, 郭必奔, 等. 基于特征参数迁移的滚动轴承故障诊断［J］. 计量学报, 2022, 43(3): 386-391.
	SHI H P, CHEN J D, WU Y M, et al. Bearing fault Analysis based on ANFIS and k-means with missing data ［J］. Combined Machine Tool and Automatic Processing Technology, 2020, 559(9): 33-36.
	LEI Y G, XU X F, CAI X, et al. Research on Data Quality Assurance Method for Mechanical Equipment Health Monitoring ［J］. Chinese Journal of Mechanical Engineering, 2021, 57(4): 1-9.
［8］	AWAN S E, BENNAMOUN M, SOHEL F, et al. Imputation of missing data with class imbalance using conditional generative adversarial networks［J］. Neurocomputing, 2021, 453: 164-171.
［11］	DRAGOMIRETSKIY K , ZOSSO D . Variational Mode Decomposition［J］. IEEE Transactions on Signal Processing, 2014, 62(3): 531-544.
	SHI P M, ZHANG H C, YI S Y, et al. An improved adaptive multivariate variational mode decomposition method for bearing fault signal feature extraction ［J］. Acta Metrologica Sinica, 2022, 43(10): 1326-1334.
［14］	竺佐, 郑永军, 罗哉. 分数阶双稳系统随机共振现象研究及FPGA实现［J］. 计量学报, 2022, 43(3): 318-324.
	HU C F, WANG Y X. Research on multi-channel signal denoising method for rolling bearing Compound fault Based on tensor decomposition ［J］. Journal of Mechanical Engineering, 2019, 55(12): 50-57.
［16］	邢婷婷, 关阳, 刘子涵, 等. 基于变分模态分解和奇异值分解的频率相近信号分离方法［J］. 计量学报, 2020, 41(11): 1404-1409.
［17］	SMITH W A, RANDALL R B. Rolling element bearing diagnostics using the Case Western Reserve University data: A benchmark study［J］. Mechanical Systems and Signal Processing, 2015, 64-65: 100-131.
	XIONG Z J, WANG X J, YANG J M, et al. Multi-objective optimization algorithm based on particle swarm and clustering ［J］. Acta Metrologica Sinica, 2023, 44(2): 252-257.
	ZHANG S F, LI T M, HU C H, et al. Missing data generation method based on deep Convolutional generation adversarial network and its application in residual lifetime prediction ［J］. Acta Aeronautica et Astronautica Sinica, 2022, 43(8): 225708.
	TANG B, CHEN S S, GUO B B, et al. Fault diagnosis of Rolling Bearing based on Migration of Characteristic Parameters ［J］. Acta Metrologica Sinica, 2022, 43(3): 386-391.
［12］	时培明, 张慧超, 伊思颖, 等. 一种改进的自适应多元变分模态分解轴承故障信号特征提取方法［J］. 计量学报, 2022, 43(10): 1326-1334.
	ZHU Z, ZHENG Y J, LUO Z. Research on stochastic Resonance Phenomena of Fractional-order Bistable systems and its FPGA implementation ［J］. Acta Metrologica Sinica, 2022, 43(3): 318-324.
	XING T T, GUAN Y, LIU Z H, et al. Signal separation method with similar frequency based on variational mode decomposition and singular value decomposition ［J］. Acta Metrology Sinica, 2019, 41(11): 1404-1409.
［18］	熊志坚, 王晓晶, 杨景明, 等. 基于粒子群与聚类的多目标优化算法［J］. 计量学报, 2023, 44(2): 252-257.
［20］	ZHANG J , FRIDMAN E . L-2-gain analysis via time-delay approach to periodic averaging with stochastic extension［J］. Automatica, 2022, 137: 110126.
［13］	WANG J Y, MO Z L, ZHANG H, et al. A deep learning method for bearing fault diagnosis Based on timefrequency image［J］. IEEE Access, 2019, 7: 42373-42383.
［19］	陈剑, 刘圆圆, 黄凯旋, 等. 基于奇异值分解和独立分量分析的滚动轴承故障诊断方法［J］. 计量学报, 2022, 43(6): 777-785.
［15］	胡超凡, 王衍学. 基于张量分解的滚动轴承复合故障多通道信号降噪方法研究［J］. 机械工程学报, 2019, 55(12): 50-57.
［21］	BELHACHMI Z, JACUMIN T. Optimal interpolation data for PDE-based compression of images with noise［J］. Communications in nonlinear science and numerical simulation, 2022, 109: 106278.
	CHEN J, LIU Y Y, HUANG K X, et al. Rolling Bearing fault diagnosis Method based on Singular Value decomposition and Independent Component Analysis ［J］. Acta Metrologica Sinica, 2022, 43(6): 777-785.