准确预测滚动轴承剩余使用寿命(Remaining Useful Life，RUL)对维护建筑机械设备稳定运行、保障生产安全具有重要的现实需求和应用价值。为提升滚动轴承RUL预测准确率，提出一种基于归一化最小均方(Normalized Least Mean Square，NLMS)自适应滤波器和Autoformer长序列预测模型的滚动轴承RUL预测新方法。使用NLMS自适应滤波器对滚动轴承原始振动信号进行降噪，从降噪振动信号中分段提取初始时域特征，采用Spearman相关系数进行特征筛选，经归一化后形成多维特征集;利用Autoformer模型中序列分解模块与自相关机制建立多维特征集与滚动轴承RUL之间的分段非线性映射，实现滚动轴承RUL预测;在PHM 2012 数据集与XJTU-SY数据集上进行对比实验，结果表明该方法与已有方法相比可取得最低预测误差，均方根误差(Root Mean Squared Error，RMSE)与平均绝对误差(Mean Absolute Error，MAE)分别提升24。4%与47。2%，证明了该方法在滚动轴承RUL预测的有效性。
RUL Prediction of Rolling Bearing Based on NLMS and Autoformer
For maintaining the stable operation of construction machinery equipment and ensuring production safety,accurately predicting Remaining Useful Life(RUL)of rolling bearing has significant practical needs and application value.Based on Normalized Least Mean Square(NLMS)adaptive filter and long-term series forecasting model Autoformer,a new method was proposed in order to improve the RUL prediction accuracy of rolling bearing.The NLMS adaptive filter was used to denoise the original vibration signal of rolling bearing.The Spearman correlation coefficient was used for feature selection from the denoised signal,and the multidimensional feature set was created.In order to forecast the rolling bearing RUL,the series decomposition block and auto-correlation mechanism in the Autoformer model were used to establish the piecewise mapping relationship between the multi-dimensional feature sets and the rolling bearing RUL.Experimental results on the PHM 2012 dataset and XJTU-SY dataset showed that the proposed method could achieve the lowest prediction error compared with the existing methods,and the Root Mean Squared Error(RMSE)and Mean Absolute Error(MAE)performance were improved by24.4%and47.2%respectively,which proved the effectiveness of the proposed method in the field of rolling bearing RUL prediction.
万方数据
维普
