作者：林涛, 严寒, 赵丹阳
作者单位：河北工业大学人工智能与数据科学学院, 天津 300130
关键词：风机主轴轴承;轻梯度提升机;鲸鱼优化算法;故障预警;残差分析
摘要：
为实现风机主轴轴承的早期故障预警，提出一种基于改进鲸鱼算法优化轻梯度提升机（iwoa-lightgbm）的故障预警方法。首先，利用皮尔逊相关系数法选取与主轴轴承温度相关的建模变量；其次，在普通鲸鱼算法的基础上，通过改进收敛因子和增加惯性因子的方式提高算法寻优能力；再次，利用改进鲸鱼算法优化lightgbm超参数并建立主轴轴承温度预测模型；最后，通过对温度残差进行分析计算出主轴轴承故障报警阈值，并利用滑动窗口法消除异常点的影响，实现对主轴轴承故障的有效预警。实验研究表明，该方法建模速度快、预测精度高，且能够提前2.5 h预测主轴轴承故障，具有广阔的工程实用前景。
research on fault early warning of spindle bearing based on iwoa-lightgbm
lin tao, yan han, zhao danyang
college of artificial intelligence and data science, hebei university of technology, tianjin 300130, china
abstract: in order to achieve the early fault warning of fan spindle bearing, a fault warning method based on improvement whale optimization algorithm (iwoa) to optimize light gradient boosting machine (lightgbm) is proposed. firstly, the pearson correlation coefficient method is used to select the modeling variables related to the spindle bearing temperature. secondly, on the basis of the common whale optimization algorithm, the optimization ability of the algorithm is improved by improving the convergence factor and increasing the inertia factor. then, the improvement whale optimization algorithm is used to optimize the lightgbm super parameters and establish the spindle temperature prediction model. finally, through the analysis of temperature residual, the alarm threshold of spindle bearing fault is calculated, and the sliding window method is used to avoid the influence of abnormal points, so as to realize the effective early warning of spindle bearing fault. the experimental results show that the model has fast modeling speed and high prediction accuracy, and can predict the spindle bearing fault 2.5 h in advance, which has a broad engineering application prospect.
keywords: fan spindle bearing;lightgbm;iwoa;fault warning;residual analysis
2023, 49(5):82-88 收稿日期: 2021-08-18;收到修改稿日期: 2021-10-14
基金项目: 河北省科技厅重点研发计划项目（20314501d，19214501d）
作者简介: 林涛(1970-),男,天津市人,教授,博士,研究方向为系统工程、故障诊断技术等
参考文献
[1] jacek b, wadim s, alena f, et al. renewable energy and eu 2020 target for energy efficiency in the czech republic and slovakia[j]. energies, 2020, 13(4): 13040965.
[2] 尹诗, 侯国莲, 胡晓东, 等. 风力发电机组发电机前轴承故障预警及辨识[j]. 仪器仪表学报, 2020, 41(5): 242-251
[3] 张哲文. 基于相关向量机的两种电力设备故障预警研究[d]. 济南：山东大学, 2020.
[4] 邬春明, 银海燕. 改进阴性选择算法的风机振动故障诊断方法[j]. 中国机械工程, 2016, 27(4): 479-482
[5] 张俊, 张建群, 钟敏, 等. 基于pso-vmd-mckd方法的风机轴承微弱故障诊断[j]. 振动. 测试与诊断, 2020, 40(02): 287-296 418
[6] 丁佳煜, 许昌, 葛立超, 等. 基于轴承温度模型的风电机组故障预测研究[j]. 可再生能源, 2018, 36(2): 276-282
[7] jiang g, he h, yan j, et al. multiscale convolutional neural networks for fault diagnosis of wind turbine gearbox[j]. industrial electronics, ieee transactions on, 2019, 66(4): 3196-3207
[8] 赵洪山, 刘辉海. 基于深度学习网络的风电机组主轴承故障检测[j]. 太阳能学报, 2018, 39(3): 588-595
[9] 王桂兰, 赵洪山, 米增强. xgboost算法在风机主轴承故障预测中的应用[j]. 电力自动化设备, 2019, 39(1): 73-77 83
[10] ke g, meng q, finley t, et al. lightgbm: a highly efficient gradient boosting decision tree[c].advances in neural information processing systems, 2017.
[11] 刘帅, 刘长良, 曾华清. 基于核极限学习机的风电机组齿轮箱故障预警研究[j]. 中国测试, 2019, 45(2): 121-127
[12] seyedali m, andrew l. the whale optimization algorithm[j]. advances in engineering software, 2016, 95(5): 51-67
[13] 王东林, 吕丽霞, 王梓齐, 等. 基于gmm工况辨识和dae的风机齿轮箱状态监测[j]. 中国测试, 2021, 47(4): 89-95
