应用QPSO寻优改进级联森林算法的齿轮箱故障信号诊断OA

Gearboxes,as key equipment for wind turbine loads,often need to operate efficiently and stably under extreme conditions.To improve the recognition efficiency of gearbox fault states,a gearbox fault signal diagnosis method based on Quantum Particle Swarm Optimization and an improved deep forest algorithm is designed.A gradient penalty function is employed to preprocess the weighted neural network,enabling stronger collaboration among particles and significantly reducing signal noise.The research results show that,whether under extreme conditions or with different sample proportions,the proposed method outperforms other approaches such as Convolutional Neural Networks in diagnosing gearbox fault signals.Under the extreme environment of 10 dB strong noise and with a small sample size,the accuracy of the proposed method can exceed 85％,demonstrating high reliability in achieving a high diagnosis rate and ideal recognition performance.This study contributes to enhancing the capability of gearbox fault signal recognition and can be extended to other transmission fields.