Research On Key Technologies For Reliability Assessment Of Complex Product Simulation System

With the increasing complexity of simulation systems and the increasing requirements of users for simulation applications,the credibility evaluation of simulation systems is facing more and more challenges.The credibility evaluation refers to a series of complicated processes that analyze,calculate and evaluate the credibility of the simulation system and finally judge whether it is credible.How to correctly carry out simulation credibility assessment is the hot and difficult point of current research,and also the focus of this paper.The concept of acceptability criteria(AC)is introduced in credibility evaluation abroad,and the satisfaction of specific expected applications of simulation systems is confirmed by summarizing appropriate acceptability standards from simulation needs or intended uses.Based on the credibility assessment of the guidance simulation system,this paper analyzes the theory of acceptability standards and proposes a method for automatic construction of specific domain AC based on supervised learning.First,based on the dictionary and conditional random field(CRF)method,three types of entities: parts,models,and index items are extracted from the simulation requirements text,on this basis,the support vector machine(SVM)is used to extract the relationships between entities to complete the extraction of qualitative AC;At the same time,an attribute extraction method based on rules and SVM is proposed to extract the attributes of component entities and their attribute values from the simulation requirement text,thereby completing the quantitative AC extraction.Aiming at the lag problem of the credibility evaluation process of the current simulation system,an AC-based simulation online evaluation method is proposed.First,map the AC to the indicator system and the indicator system to the evaluation process to complete the decomposition of the indicator;Secondly,a seven-tuple model of the simulation online evaluation process was designed,which expanded the two tuples of the index system and evaluation function,and realized the closed-loop structure of online evaluation,and through the finite automaton model,the completeness verification of the system state transition is carried out,and the automatic deduction of the evaluation process from the credibility of the index,the credibility of the evaluation link to the credibility of the simulation is completed.In order to be able to calculate the comprehensive credibility of the link in each evaluation link,in this paper,the credibility evaluation method of simulation system based on the optimal combination weighting TOPSIS method is constructed,and the effectiveness of the credibility evaluation method is verified by specific evaluation examples.This paper designs and develops a credibility evaluation tool for complex product simulation systems,and verifies the practicability and effectiveness of the online simulation evaluation method based on the credibility evaluation process of specific examples.