Research On Risk Assessment Method Based On Belief Rule Base

Risk assessment is one of the common methods used in many areas to reduce accidents and avoid risk,and the scope of application has been concerned about domestic and foreign scholars.In many important fields such as industrial manufacturing,energy,transportation,and national defense,risk assessment is used as an important method to improve operational reliability and safety and reduce accident rates.Taking the oil and gas industry as an example,due to the risk of major production safety accidents in this industry,it attaches great importance to production safety related issues.Risk assessment is highly concerned as an effective means of preventing major safety accidents,and has been widely used in various links such as oil and natural gas mining,processing,storage,transportation,etc.Risk assessment methods mainly include quantitative risk assessment methods,qualitative risk assessment methods and semi-quantitative risk assessment methods.The semi-quantitative risk assessment method combines quantitative analysis and qualitative analysis,making the assessment results with higher accuracy.However,the inexplicability of existing semi-quantitative risk assessment methods makes the reliability of such methods questionable and affects their popularization;Some semi-quantitative risk assessment methods are issued in a particular industry,and there is a problem that the accuracy of the model is lowered by excessive indicator or redundancy.To solve these problems,this article has deeply studied the semi-quantitative risk assessment method of the belief rule base.The main contributions are reflected in the following aspects:(1)A risk assessment model based on principal component regression and hierarchical belief rule base(PCR-HBRB)is proposed.This model solves the inexplicability of existing semi-quantitative risk assessment methods and the problem of model accuracy reduction caused by excessive indicators or redundant indicators in actual risk assessment.In the model,PCR is used to screen the indicators in the risk assessment to obtain the indicators that have the greatest impact on the assessment results,and then the selected indicators are evaluated through the hierarchical belief rule base(HBRB)to establish a risk assessment model,finally,the projection covariance matrix adaptive evolution strategy(P-CMA-ES)is used to optimize the model through parameter adjustment.(2)As BRB has problems in risk assessment such as difficulty in obtaining expert knowledge,unclear risk indicators,and difficulty in establishing logical relationships between various risk indicators,it has some negative impacts on the results of risk assessment.This paper expands the BRB knowledge base through fault tree mechanism analysis,and proposes a risk assessment model based on fault tree analysis and belief rule base(FBRB).This model also solves the inexplicability of risk assessment methods.In this model,the conversion mechanism of fault tree analysis(FTA)and belief rule base(BRB)is first analyzed,and the conversion rules of FTA and BRB are established through the conversion mechanism.According to the conversion rules,the knowledge base of BRB can be expanded,then the improved BRB is used to establish a risk assessment model,and finally the model is optimized through P-CMA-ES.(3)The FBRB model is used to evaluate the risk of the flywheel system of the spacecraft,which verifies that the method in this paper is feasible and has high accuracy.In the realization of the prototype,firstly,the fault tree is established by taking the friction torque increase fault in the flywheel system as an example,and then the BRB knowledge base is expanded through the fault tree and the evaluation model is established,and finally the model is further optimized through the optimization algorithm.The research work of this article on risk assessment has solved the unexplainable problems of existing semi-quantitative risk assessment methods;the proposed risk assessment based on PCR-HBRB focuses on the problem of reduced model accuracy caused by too many indicators or redundant indicators;The risk assessment model based on FBRB focuses on solving problems in risk assessment,such as difficulty in obtaining expert knowledge,unclear risk indicators,and difficulty in establishing logical relationships between various risk indicators.FBRB-based flywheel system risk assessment prototypes further verify the feasibility of FBRB-based risk assessment model.