Consistency Analysis System Of Solid Small Rocket Based On Bayesian Network

In order to meet the requirement of production consistency management and deviation source diagnosis for small batch,multi-batch and different types of small solid rockets in a pyrotechnics factory,the design and implementation of a consistency analysis platform for small solid rockets based on Bayesian network are carried out.There are many problems in the production of solid small rocket,such as multi-source information,small samples and uncertainties.At present,the mainstream method of consistency analysis is to invite experts to make qualitative evaluation of the relevant factors affecting the important performance indicators of small solid rockets.It relies too much on the technical level of experts and has too much subjectivity.Therefore,this paper introduces Bayesian network,combines production data with expert knowledge,constructs consistency evaluation index of solid small rocket,studies quantitative analysis method of deviation source and consistency evaluation index,builds product quality consistency analysis model based on Bayesian network,designs and implements corresponding product consistency analysis platform.The main work of this paper is embodied in the following aspects:(1)Through the analysis of the actual processing flow of a small solid rocket and the failure mechanism of the small solid rocket,the consistency evaluation index of the small solid rocket is summarized from three aspects of structure,production process and performance.Through the analysis of the production process,the deviation mapping model from the deviation source node of the small solid rocket to the measuring point is established.(2)Transforming qualitative expert knowledge into Bayesian network,combining with consistency performance indicators and starting from multi-source data,the measuring points which have great influence on the performance of small solid rocket are expressed as evidence nodes,and a diagnostic model of two-stage Bayesian network is established;uncertain knowledge is measured by probability,and the implicit causality between father and son nodes is expressed as expert knowledge to solve the deviation sources of small solid rocket.The problem of small sample modeling in diagnosis.Then,through Netica simulation,the state of the measurement points is set,the posterior probability of the deviation source nodes is observed,and the model is adjusted according to the diagnosis results,so that the diagnosis results of the final model are consistent with the qualitative reasoning results.(3)based on the above model,the consistency analysis platform of solid small rocket products based on Bayesian network is designed.The application software platform is designed from three aspects: requirement analysis,function module design,interface design and business logic.The expert system of Bayesian network deviation source diagnosis is realized and applied to practical engineering.This paper makes full use of the ability of Bayesian network to express and mine knowledge,and provides a set of very effective solutions for the informationization of solid rocket consistency analysis and the improvement of production and processing level.