一种风电机组轴承使用寿命预测方法
By improving the whale optimization algorithm to adaptively optimize VMD parameters and the self-attention mechanism to dynamically learn health indicators, and combining it with the CNN-BiLSTM model, the parameter dependence and robustness problems in wind turbine bearing life prediction were solved, and higher accuracy life prediction was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHENGDU UNIV
- Filing Date
- 2026-06-16
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for predicting the lifespan of wind turbine bearings suffer from problems such as strong parameter dependence, poor robustness of health indicators, and low prediction accuracy, especially in complex signal decomposition, health indicator construction, and prediction modeling.
An improved whale optimization algorithm (IWOA) is used to adaptively optimize variational mode decomposition (VMD) parameters, and health indicators are dynamically learned by combining a self-attention mechanism. The remaining service life of bearings is then predicted using a CNN-BiLSTM regression model.
It improves the stability and physical interpretability of signal decomposition, constructs health indicators with monotonicity and trend, enhances sensitivity to different degradation stages, strengthens the ability to capture local features and long-term dependencies, and improves the accuracy and stability of prediction.
Smart Images
Figure CN122413383A_ABST