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...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。展开更多
Reference Evapotranspiration(ETo)iswidely used to assess totalwater loss between land and atmosphere due to its importance in maintaining the atmospheric water balance,especially in agricultural and environmental mana...Reference Evapotranspiration(ETo)iswidely used to assess totalwater loss between land and atmosphere due to its importance in maintaining the atmospheric water balance,especially in agricultural and environmental management.Accurate estimation of ETo is challenging due to its dependency onmultiple climatic variables,including temperature,humidity,and solar radiation,making it a complexmultivariate time-series problem.Traditional machine learning and deep learning models have been applied to forecast ETo,achieving moderate success.However,the introduction of transformer-based architectures in time-series forecasting has opened new possibilities formore precise ETo predictions.In this study,a novel algorithm for ETo forecasting is proposed,focusing on four transformer-based models:Vanilla Transformer,Informer,Autoformer,and FEDformer(Frequency Enhanced Decomposed Transformer),applied to an ETo dataset from the Andalusian region.The novelty of the proposed algorithm lies in determining optimized window sizes based on seasonal trends and variations,which were then used with each model to enhance prediction accuracy.This custom window-sizing method allows the models to capture ETo’s unique seasonal patterns more effectively.Finally,results demonstrate that the Informer model outperformed other transformer-based models,achievingmean square error(MSE)values of 0.1404 and 0.1445 for forecast windows(15,7)and(30,15),respectively.The Vanilla Transformer also showed strong performance,closely following the Informermodel.These findings suggest that the proposed optimized window-sizing approach,combined with transformer-based architectures,is highly effective for ETo modelling.This novel strategy has the potential to be adapted in othermultivariate time-series forecasting tasks that require seasonality-sensitive approaches.展开更多
文摘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。
基金funded by Princess Nourah bint Abdulrahman University and Researchers Supporting Project number(PNURSP2024R136),Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘Reference Evapotranspiration(ETo)iswidely used to assess totalwater loss between land and atmosphere due to its importance in maintaining the atmospheric water balance,especially in agricultural and environmental management.Accurate estimation of ETo is challenging due to its dependency onmultiple climatic variables,including temperature,humidity,and solar radiation,making it a complexmultivariate time-series problem.Traditional machine learning and deep learning models have been applied to forecast ETo,achieving moderate success.However,the introduction of transformer-based architectures in time-series forecasting has opened new possibilities formore precise ETo predictions.In this study,a novel algorithm for ETo forecasting is proposed,focusing on four transformer-based models:Vanilla Transformer,Informer,Autoformer,and FEDformer(Frequency Enhanced Decomposed Transformer),applied to an ETo dataset from the Andalusian region.The novelty of the proposed algorithm lies in determining optimized window sizes based on seasonal trends and variations,which were then used with each model to enhance prediction accuracy.This custom window-sizing method allows the models to capture ETo’s unique seasonal patterns more effectively.Finally,results demonstrate that the Informer model outperformed other transformer-based models,achievingmean square error(MSE)values of 0.1404 and 0.1445 for forecast windows(15,7)and(30,15),respectively.The Vanilla Transformer also showed strong performance,closely following the Informermodel.These findings suggest that the proposed optimized window-sizing approach,combined with transformer-based architectures,is highly effective for ETo modelling.This novel strategy has the potential to be adapted in othermultivariate time-series forecasting tasks that require seasonality-sensitive approaches.
