Factorized axis convolutional gated recurrent unit with dynamic adaptive pooling for remaining useful life prediction of rolling bearings

2026-09-25

AI Summary

This study addresses the limitations of traditional CNNs, where isotropic convolution kernels struggle to capture directional structures in time-frequency representations and global average pooling obscures salient activation locations. To overcome these issues, this work proposes a diagnostic model integrating factorized axial convolutional GRUs with dynamic adaptive pooling. The method employs multi-scale anisotropic convolutions to enhance directional feature extraction and utilizes a dual-axis attention module for precise time-frequency information aggregation. Furthermore, temporal modeling and uncertainty estimation are achieved through GRUs combined with Monte Carlo Dropout. Experiments on two public bearing datasets demonstrate that the proposed model outperforms existing methods. Ablation studies confirm that the anisotropic design yields lower errors than isotropic kernels, effectively improving the extraction of degradation features.

Abstract

Convolutional neural networks (CNN) are widely used to predict the remaining useful life (RUL) of rolling bearings from time-frequency representations (TFRs) of vibration signals. However, during degradation, characteristic structures in TFRs align predominantly along the frequency or time axis, making it challenging for conventional CNN isotropic kernels to capture directional structure. Furthermore, global average pooling (GAP) averages across axes, potentially obscuring the locations and concentrations of salient activations. This study introduces a factorized-axis convolutional gated recurrent unit (GRU) that employs multiscale anisotropic convolution and dual-axis convolution block attention module to enhance directional features and highlight salient time-frequency regions. Dynamic adaptive pooling (DAP) adaptively aggregates the time-frequency-axis information from the extracted feature maps, whereas a GRU captures temporal dynamics in the latent representations and Monte Carlo dropout enables predictive uncertainty estimation. Experiments on two public bearing datasets demonstrate that the proposed model outperforms existing RUL prediction methods across operating conditions. Ablation experiments demonstrate that the factorized axis-wise design achieves lower mean errors than convolutional isotropic kernels. DAP yields clear improvements on one dataset while matching GAP on the other, highlighting the importance of anisotropic feature extraction and adaptive feature aggregation for TFR-based RUL prediction.

Problem

• Remaining Useful Life Prediction
• Rolling Bearings
• Time-Frequency Representations
• Anisotropic Feature Extraction
• Adaptive Pooling

Innovation

• Factorized axis convolution
• Dynamic adaptive pooling
• Remaining useful life prediction
• Gated recurrent unit
• Anisotropic feature extraction