Research On Fault Diagnosis Of Civil Electro-hydraulic Joint Steering Gear Based On Data-driven Approaches

As the core actuator of civil aircraft flight control system,the reliability of civil electro-hydraulic joint steering gear(hereinafter referred to as steering gear)directly affects the safety of the flight control system and even the whole aircraft.The failure of the steering gear will lead to the deterioration of the control effect of the aircraft,which will affect the flight quality and safety of the aircraft,and even cause major flight accidents.Therefore,it is of significant theoretical and engineering value to study the fault diagnosis of the steering gear.At the same time,considering that the operation of the steering gear conforms to the intermittent characteristics,accurate sub-phase division and fine modeling of the operating data of the steering gear are helpful to significantly improve the overall performance of the fault diagnosis of the steering gear.In this paper,a fault diagnosis method based on moving window multidimensional information increment matrix(MWMIIM)is proposed for the first time,and this paper mainly completed the following four aspects of work:Firstly,based on the description of the basic theoretical knowledge of information increment matrix(IIM),this paper points out the shortcomings of the current principal component analysis(PCA)algorithm in practical application,and analyzes the advantages of the information increment matrix algorithm,namely more efficient computation,more intuitive physical interpretation and more reasonable data interpretation.Secondly,according to the engineering application demand,the problems of uneven length and unsynchronized running track of steering gear batch data are deeply studied.To solve the problems,the multidimensional information increment matrix algorithm combined with sliding window technique is proposed,and the single batch data of the steering gear are divided into several sub-phases directly only by the variation of the correlation of variables in the historical running data.Besides,according to the different data characteristics of different regions,fine modeling is carried out in each region to obtain accurate monitoring model.Then,in order to improve the accuracy of online monitoring,different monitoring strategies are adopted in different regions after fine modeling for each region.Besides,theproposed MWMIIM algorithm is utilized to conduct simulation experiments of online monitoring and fault diagnosis of the steering gear by the statistical analysis of the information gain index.Compared with the comparison algorithm,the simulation results prove the superiority of the proposed algorithm in fault monitoring and fault diagnosis of steering gear.Finally,the proposed MWMIIM algorithm is verified based on the actual electro-hydraulic joint steering gear test platform and gearbox test platform.Specifically,the working principle and available data variables of the experimental platform of steering gear and gearbox are introduced respectively.Subsequently,the effectiveness of the MWMIIM algorithm is proved by comparison with different algorithms under the same experimental conditions.Experimental results show that the MWMIIM algorithm can better distinguish the contribution rate of each variable to fault,owns better real-time performance for fault diagnosis and lower error fault alarm rate,which means the proposed algorithm has certain advantages in fault monitoring and fault diagnosis.