Human–Machine Collaborative Health Evaluation for Wind Turbine Gearbox via Versatile Parallel Network and Group Confidence Decision

The health evaluation of wind turbine gearbox (WTG) is crucial to ensure its operation efficiency and safety. Herein, a novel human-machine collaborative health evaluation framework is proposed based on versatile parallel network (VPN) and group confidence decision (GCD). First, the VPN is established for the gearbox compound fault diagnosis, and the fault probability distribution is derived and translated into the status membership function to quantify the fault statue membership. Then, a group decision-making method named GCD is designed to estimate the impact of different fault types on the gearbox behavior, and it can make a tradeoff among multiple expert perspectives to output optimal estimation results for the fault risk weights. Meanwhile, a health index is defined by the extend ratio system to reflect the overall health condition of WTG. Finally, a WTG compound fault simulation platform is built to validate the method effectiveness, and results indicate that the proposed framework can make full use of plentiful monitoring data and expert perspectives for the precise and comprehensive health evaluation.