Research On High-speed Automata Health Management System

As the core component of automatic weapon equipment,high-speed automata plays an important role in realizing cyclic actions such as feeding,transporting,firing,and ejecting shells.Complicated structure and failure modes affect the combat performance of high-speed automata to a certain extent.Traditional after-the-fact maintenance methods are not only costly and low fault tolerance,but also cannot eliminate hidden troubles in time.Therefore,developing a fast and efficient high-speed automatic machine failure prediction and health management system to realize the intelligent development from after-the-fact maintenance and regular maintenance to health monitoring,evaluation,prediction and decision-making,thereby extending the service life cycle and upgrading repair workpiece ratio of weaponry installations.This paper takes a certain type of 30 mm caliber revolving automata as the research object,and devises a suit of high-speed automata soundness supervision system.The effectiveness and advanced of the health management system proposed in this paper are demonstrated by the verification of the actual operation data and working state of the automata.The concrete disquisition content subsumes the following respects:(1)Monitoring of state parameters of automata.Taking high-speed automata as the research object,construct a monitoring framework for the temperature,pressure,displacement and speed of key components such as equipment revolving body and sliding plate,study the intelligent monitoring of automata state parameters in complex environments,and aim at the randomness of the monitoring process.For ambiguity and uncertainty,it is proposed to use state monitoring as the basis,and adopt wavelet algorithm to restrict the rang of threshold value,and actualize the intention of removing interfering signal impact,and explore the impact of time-domain character on the status of device.(2)Research on health status assessment methods.The health status of the automata system is affected by the superimposition of complex components.Aiming at the problems of diverse failure modes,difficulty in defining the health status and dynamic deterioration in the maintenance and support process,an assessment method found on the fuzzy neuronic network prototype is propounded,and the valuation indicatrix is blurred.In the cause of continuously updating the prototype parameters and retrofit on this fundament,by introducing a health index,integrating multiple characteristic values and establishing an early warning threshold,the dynamic effects of equipment aging are eliminated,and the valuation consequence are more precise and dependable.(3)Health trend and failure prediction.Aiming at the problem of equipment accumulating with working time,component performance degradation and failure rate increase,a forecast algorithm found on wavelet nervous network prototype is studied.By modifying the network weights and wavelet coefficients,the prototype constringency velocity and the local signal mutation learning ability are enhanced.By contrast the performance of the BP neuronic network,the predicted value is closer to the actual output performance of the equipment,the forecast precision is higher,and the speed of approaching the limit optimal target is faster,and the requirements for predicting the degree of automata performance degradation are achieved.(4)Analysis of residual longevity of device.In connection with the issue of cracks and ablation induced by the ablation of elevated temperature gunpowder on the key mechanisms of the automata system,the linear relationship between the failure development factor and the residual life of the equipment was studied,and the residual longevity of the faulty component was simulated by the minimum square technique,and this article establishes a residual longevity prototype combined with wavelet nervous network and gives a simulation prediction example.The maximum prediction error is 0.6%,which verifies the validity of the model.