Research On Fault Diagnosis Method Of Hydraulic System Of An Ammunition Coordinator

The future of artillery ammunition loading system will move towards automation,informatization and intelligent direction.Ammunition loading system goal is to achieve any shot angle or arbitrary azimuth angle under the automatic ammunition loading,to achieve automatic monitoring of work status and fault diagnosis.As a key component of the ammunition device,the coordinator is driven by hydraulic system.Hydraulic system is one of the subsystems of the artillery failure.Because of the closeness and diversity of failure modes and effects,it is difficult to locate faults quickly and accurately and to remedy them.Therefore,it is very important to study the fault diagnosis of the hydraulic system of the coordinator.Combining the characteristics of Function principal component analysis(FPCA)in the feature extraction process and neural network in fault diagnosis,a method of fault diagnosis based on function principal component analysis and BP neural network is studied,which is applied to fault diagnosis of hydraulic system of coordinator.This paper mainly completed the following work:(1)This paper analyze the structure and working principle of the coordinator and its hydraulic system,study the failure mechanism and failure mode,sum up the common fault of the hydraulic system and the individual fault of the coordinator hydraulic system,and establish the FMEA form.(2)In the ADAMS and AMESim,the co-simulation model of the hydraulic system of the coordinator is established,and the failure state information of the hydraulic system is determined which is compared with the experimental data of the real equipment.Simultaneously,the simulation data of the co-simulation model are compared with the data of the hydraulic model in MATLAB to further verify the correctness of the model.(3)The typical fault parameters of the hydraulic system of the coordinator are selected and simulated,which can provide fault data for the fault diagnosis system.(4)Using function principal component analysis,the sample data are functionized and the feature parameters are extracted.The mapping between the characteristic parameters and the fault parameters is trained by BP neural network to verify its feasibility.Finally,the fault diagnosis method is completed.