基于DRC-CNN模型的风力发电机叶片故障诊断分析OA

During long-term operation,wind turbine blades are prone to damage caused by fatigue accumulation.The damage of Glass Fiber Reinforced Polymer(GFRP)composite material used will affect the life of wind turbine blades and even cause them to break.In response to the problem,this study uses acoustic emission acquisition equipment to collect the acoustic emission signals of wind turbine blade materials in the three-point bending experiment,and the morphological characteristics of the fracture of specimen are photographed by a Scanning Electron Microscope(SEM)to observe the microscopic changes in material damage.Through the confirmation between the acoustic emission signal feature extraction and the microscopic change of material,the failure process is divided into micro crack stage,local damage stage and material failure stage.Then,based on the dual requirements of real-time and performance of deep learning algorithms for wind turbine blade material damage stage identification,a DRC-CNN model is proposed to realize automatic identification of different damage stages of materials.The research results show that the DRC-CNN model has improved by 17.4%,17.5%and 15.4%in the three key indicators of accuracy,recall and F1-score,respectively,compared with the baseline model.