摘要
针对复杂装备系统故障模糊性强的特点,以及目前基于模型和数据驱动的故障诊断大都局限于决策层融合的问题,提出了1种利用诊断模型作为数据驱动方法的初始条件来辅助网络模型构建和学习的方法。首先,通过TS故障树理论分析,建立系统各故障模式之间的逻辑关系和描述规则;然后,根据T-S故障树模型,将诊断模型映射为模糊神经网络(Fuzzy Neural Networks,FNN)模型,并利用误差反向传播算法对网络参数进行学习,进而提出1种模糊规则自动更新机制;最后,以某组合导航系统为实验对象进行仿真实验。结果表明:提出的方法能够准确地诊断出故障,且具有较快的收敛速度和较好的泛化能力。
Aiming at the characteristics of complex equipment system with strong fault fuzziness and the problem that the current model-based and data-driven fault diagnosis is mostly limited to the fusion of decision layers,a method that uses the diagnostic model as the initial condition of data-driven method to assist network model construction and learning is proposed.Firstly,the logic relation and description rules of each fault mode are established by T-S fault tree theory analysis.Then,according to the T-S fault tree model,the diagnosis model is mapped to the fuzzy neural network model,and the network parameters are learned by using the error back propagation algorithm,and an automatic updating mechanism of fuzzy rules is proposed.Finally,an integrated navigation system is taken as the experimental object,and the simulation results show that the proposed method can effectively identify the fault diagnosis with fast convergency and good generalization capacity.
作者
刘路
史贤俊
LIU Lu;SHI Xianjun(Naval Aviation University,Yantai Shandong 264001,China)
出处
《海军航空大学学报》
2023年第4期368-374,共7页
Journal of Naval Aviation University
