旋翼作为直升机的升力面和操作面，其健康状态对直升机的安全至关重要。旋翼故障诊断技术仍是直升机健康与使用监测系统(Health and usage monitoring system，HUMS)领域的薄弱环节，开发旋翼故障诊断技术具有重要价值。基于信息融合技术，首先分析了旋翼故障的诊断机理，建立了旋翼故障模型，通过流固耦合仿真获取了不同故障下桨叶、轮毂和机身的故障特征信息，生成数据集进行网络训练和验证。然后，利用遗传算法反向传播(Genetic algorithm-backpropagation，GA-BP)优化神经网络诊断3种类型的直升机旋翼故障，即后缘调整片误调、变距拉杆误调和桨叶质量不平衡。3个逐级神经网络分别对旋翼故障类型、故障位置和故障程度进行了诊断识别。最后采用加权的Dempster-Shafer(D-S)证据理论对旋翼故障进行诊断和分析。结果证明基于改进D-S证据理论的旋翼故障诊断方法能够成功应用到旋翼故障诊断中，并具有良好的识别效果。
Rotor Fault Diagnosis Based on Weighted D-S Evidence Theory
The main rotor is the lift surface and control surface of a helicopter,and its normal health is crucial for the safety of the helicopter.The rotor fault diagnosis technology is still a weak link in the field of helicopter health and usage monitoring system(HUMS),and the development of rotor fault diagnosis technology is of great value.Based on information fusion technology,the mechanism of rotor failure is analyzed,the rotor failure model is established,and the fault feature information of blades,hub and airframe under different faults are obtained by fluid structure coupled simulation,thus generating data sets for network training and verification.Then genetic algorithm-backpropagation(GA-BP)neural network is used to diagnose three types of helicopter rotor faults,namely,misadjusted trim-tab,misadjusted pitch control rod and imbalanced mass.Three cascaded levels of networks are used to identify fault classification,location and severity,respectively.Finally,the rotor faults are diagnosed and analyzed by the weighted Dempster-Shafer(D-S)evidence theory.The results demonstrate that the rotor blade fault diagnosis method based on the improved D-S evidence theory can be successfully applied to rotor blade fault diagnosis with good identification results.
rotor systemfault diagnosisGA-BP neural networkinformation fusion technologyD-S evidence theory
高亚东、张传壮
展开 >
南京航空航天大学航空学院,南京 210016,中国
旋翼系统 故障诊断 GA-BP神经网络 信息融合技术 D-S证据理论
National Key Laboratory of Helicopter Aeromechanics
国家科技期刊平台
NSTL
万方数据
