Fault diagnosis of marine centrifugal fans based on WOA-RF
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
万方数据
针对船用风机典型故障诊断问题,利用鲸鱼优化算法优化随机森林(random forest optimized by whale optimization algorithm,WOA-RF)对故障进行诊断,并通过实验验证该方法的准确性.实验模拟包括正常工况和5种异常工况在内的6种工况.采集所有不同工况下的振动信号,分别提取时域、频域下的特征参数构建第一特征向量.通过传统随机森林筛选得到具有更优分类效果的第二特征向量,再输入WOA-RF中完成故障识别.实验结果表明,本文提出的方法能够有效识别故障模式,平均预测准确率超99％.与其他算法对比,这种基于信息融合和WOA-RF的船用风机故障诊断方法准确性更高.
Aiming to the typical fault diagnosis issue of marine centrifugal fans,a diagnosis method based on the random forest optimized by the whale optimization algorithm(WOA-RF for short)is proposed,and its accuracy is verified by experiments.The experiments simulate six working conditions including a normal condition and five faulty conditions.Vibration signals are collected under all different conditions,and feature parameters are extracted in both time and frequency domains to construct the first feature vector.The second feature vector with better classification performance is obtained through traditional random forest screening and input into the WOA-RF for fault identification.The experimental results show that the proposed method can effectively identify fault modes with an average prediction accuracy of over 99％.Compared with other algorithms,this fault diagnosis method for marine centrifugal fans based on information fusion and WOA-RF is of higher accuracy.
random forestwhale optimization algorithm(WO A)marine centrifugal fanfault diagnosis
NETL
NSTL
万方数据
