鲸鱼优化算法优化卷积神经网络与双向长短期记忆网络的飞机液压系统优先阀故障诊断  

Fault Diagnosis of Priority Valve in Aircraft Hydraulic System Based on Whale Optimization Algorithm Optimizating Convolutional Neural Network and Bi-directional Long-short Term Memory

作　　者：孙琦 任杰锶 王伟杰 葛磊 熊晓燕 权龙 SUN Qi;REN Jiesi;WANG Weijie;GE Lei;XIONG Xiaoyan;QUAN Long(School of Mechanical and Transportation Engineering,Taiyuan University of Technology,Taiyuan,Shanxi 030024)

机构地区：[1]太原理工大学机械与运载工程学院,山西太原030024

出　　处：《液压与气动》2024年第10期107-116,共10页Chinese Hydraulics & Pneumatics

基　　金：国家重点研发计划(2022YFB3403001);工业和信息化部民用飞机专项科研项目(MJZ3-1N22)。

摘　　要：优先阀作为飞机液压系统的一种重要控制元件,一般用于在液压系统供油压力较低时,优先保证主飞控系统的供油,其稳定性是飞行安全的重要保证。然而,飞机液压系统中仅存在压力传感器,传统诊断方法难以应用。为此,提出了一种用鲸鱼优化算法优化卷积神经网络与双向长短期记忆网络的故障诊断算法。以优先阀进出口压差信号为依据,利用卷积神经网络与双向长短期记忆网络融合算法提取故障特征,并通过鲸鱼优化算法优化超参数,避免所提算法陷入局部最优。通过AMESim与MATLAB多学科联合仿真,结果表明:该方法在复杂工况下的平均诊断准确率超95%;与现存方法相比,其在诊断精度和稳定性上有显著提升。The priority valve,a critical component of the aircraft hydraulic system,is responsible for prioritizing oil supply to the main flight control system when hydraulic system pressure is low.Maintaining its stability is essential for ensuring flight safety.However,the aircraft hydraulic system only has pressure sensors that makes traditional diagnostic methods difficult to implement.To overcome this challenge,a fault diagnosis algorithm using a whale optimization algorithm to optimization convolutional neural network bi-directional long short-term memory networks is proposed.Based on the differential pressure signals between the inlet and outlet of the priority valve,the fault feature is extracted by using the fusion algorithm of convolutional neural network and bi-directional long-short term memory and the super parameters are optimized by whale optimization algorithm to avoid the proposed algorithm falling into local optimization.Results from a multidisciplinary joint simulation using AMESim and MATLAB show that the proposed method achieves an average diagnostic accuracy of over 95%in complex working conditions.Compared to existing methods,this approach demonstrates significant improvements in diagnostic accuracy and stability.

关 键 词：鲸鱼优化算法 优先阀 液压系统 AMESIM 故障诊断 

分 类 号：TH137[机械工程—机械制造及自动化]