基于多体动力学仿真和HMM的直升机传动系统早期故障诊断与预测维护

In the helicopter transmission system, due to complex mechanical structure, high speed operation, large power transmission and variability of working condition resulting from dramatic changing airflow and flight attitude adjustment, failure of the key components and parts in the helicopter main transmission system become the main reason for the helicopter accident. This project takes the helicopter transmission system as the research object, considering the failure characteristics of key components, establishes the multi-body dynamics fault model establishment of the helicopter transmission system. By means of model simulation and experimental study, these vibration signals of several typical faults for the helicopter transmission system are obtained under the complex drive load and multi-source excitation. Then, using un-decimated wavelet transform (UDWT) and pattern recognition techniques to extract the feature vectors of key component typical fault, the GMM-HMM-based fault diagnosis and predictive models are built for these key components. By sample training, model learning, data mining and statistical analysis, those typical faults of the key components of helicopter transmission system are predicted and evaluated, and the method of early fault diagnosis and state assessment is presented for the key components based on Gaussian Mixture Model and Hidden Markov Model （GMM-HMM）.Finally, according to the results of fault prediction and condition assessment for the key parts, maintenance optimization decision-making model is established for the key components and parts by means of the proportion hazard function. Considering the repair costs and availability, multi-objective optimization theory is applied to optimize maintenance cycle and maintenance strategy. Research results of the project will provide a theoretical basis and decision making support to improve safety and reliability of the helicopter transmission system and to reduce maintenance costs, also enrich and develop the theory and method of failure prediction and health maintenance for the key components of electromechanical equipment .

直升机传动系统的机械装置结构复杂,运转的转速高,传递的动力大,气流变化剧烈及飞行姿态调整造成的工况多变性, 直升机主传动系统故障成为直升机事故的最主要原因。项目以直升机传动系统为研究对象，考虑关键零部件的失效特征，建立直升机传动系统多体动力学故障模型，通过模型仿真分析和实验研究，得到直升机传动系统在复杂载荷作用下典型故障的振动信号；然后应用小波变换提取关键零部件典型故障的特征向量，建立基于GMM-HMM的关键零部件故障诊断与预测模型，提出基于HMM的关键零部件早期故障诊断与状态评估方法。最后，根据关键零部件故障预测和状态评估结果，建立基于比例风险模型的直升机传动系统优化维修决策模型，综合考虑维修成本及可用度，应用多目标优化理论优化维修周期和维修决策。项目研究为提高直升机传动系统安全可靠运行及降低维护成本提供理论支撑和决策依据，丰富发展机电设备关键零部件故障预测和健康维护的理论与方法。

直升机传动系统的机械装置结构复杂,运转的转速高,传递的动力大,气流变化剧烈及飞行姿态调整造成的工况多变性, 直升机主传动系统故障成为直升机事故的最主要原因。项目以直升机传动系统为研究对象，考虑关键零部件的失效特征，建立直升机传动系统多体动力学故障模型，通过模型仿真分析和实验研究，得到直升机传动系统在复杂载荷作用下典型故障的振动信号；然后应用小波变换提取关键零部件典型故障的特征向量，建立基于GMM-HMM的关键零部件故障诊断与预测模型，提出基于HMM的关键零部件早期故障诊断与状态评估方法。最后，根据关键零部件故障预测和状态评估结果，建立基于比例风险模型的直升机传动系统优化维修决策模型，综合考虑维修成本及可用度，应用多目标优化理论优化维修周期和维修决策。项目研究为提高直升机传动系统安全可靠运行及降低维护成本提供理论支撑和决策依据，丰富发展机电设备关键零部件故障预测和健康维护的理论与方法。