Research On PHM Architecture And Key Technology Of Military Electronic Equipment

With the continuous improvement of the weaponry and equipment system integration and informationization,real-time monitoring of military electronic equipment operating status and health status assessment,timely adjustment of maintenance and maintenance strategies,change "after repair" to "repair according to the situation",can effectively improve Equipment maintenance efficiency and protection level.The research on the failure prediction and health management of military electronic equipment is the fundamental requirement for improving the level of support and combat effectiveness of existing weapons and equipment systems.PHM is a technology for system health management and comprehensive fault detection,isolation and prediction.Based on the actual development of military equipment,the main research work of the dissertation includes the following sections:First,construct the PHM system structure of military equipment.This section mainly analyzes the existing PHM open process,combines the actual needs of military electronic equipment fault prediction and diagnosis,learns from the open system architecture for condition based maintenance(OSA-CMM),highlights the military features,and builds a military service-based PHM service architecture based on cloud services..Second,implemented nonlinear manifold learning and hybrid HMM combined with fault diagnosis techniques.Using the advantage of nonlinear manifold learning method in dealing with high-dimensional and nonlinear data and the advantage of HMM model in pattern recognition process,a hybrid HMM model is proposed and established for analog circuit fault diagnosis.Through actual circuit modeling and simulation analysis,compared with the conventional diagnostic methods,this method can effectively identify early fault features and has a high fault recognition rate.Third,based on Bayesian network of military electronic equipment module level fault diagnosis research.The existing military electronic equipment has high antiinterference requirements.It adopts modularization and hierarchical design.The structure is complex and the degree of integration is high.In addition,when the fault occurs in a complex electromagnetic environment,the fault propagation mechanism is not easy to grasp and maintenance is difficult.The Bayesian network method is used to diagnose this type of failure with obvious results.Fault modeling and positioning can be achieved through structural learning and parameter learning,respectively,and the diagnostic efficiency is much higher than the general method.This paper takes the receiver of a certain type of radar system as an example,established a Bayesian network model is established according to the topological structure of the circuit system to realize the diagnosis and positioning of the faulty module.Fourth,research based on the particle swarm search algorithm and improved gray model of electronic equipment system state prediction technology research.The trend forecasting of military electronic equipment for faults is studied.The short-term status prediction of electronic systems based on the improved grey model is mainly studied.Taking the radar voltage of the high-voltage power supply of a certain type of radar transmitter,the body of the gate-controlled traveling-wave tube,and the historical data of the collector current as the research basis,the most suitable gray model GM(1,1)is selected,and the optimized particle swarm search using online update is used.The algorithm finds the best dimension parameters to achieve good convergence of the algorithm,improves the accuracy of the improved grey model,and predicts the state of the radar transmitter system.