Study On Risk Assessment Of Sudden Water Pollution Accident Based On Fuzzy Bayesian Networks

In recent years,with the continuous growth of China's population,the acceleration of industrialization and urbanization,the pressure on the water environment is constantly increasing.Water is the source of life.At present,the shortage of water resources and frequent accidents of water pollution have seriously threatened people's health and safety of life and property.The contradictions between development,pollution and health have gradually evolved into major issues that need to be coordinated and solved urgently.Therefore,in the process of further promoting the healthy development of urbanization,it is very necessary to set up the technical system of risk assessment of sudden water pollution accidents,strengthen the risk management of water environment,control or reduce the incidence of sudden water pollution accidents and ensure the safety of drinking water for residents.The article takes the important risk sources and risk receptors of sudden water pollution accidents as research objects based on the research project of "Study on Risk Management of Sudden Water Pollution in Zones of Watersheds ".By analyzing and summarizing the characteristics and the scope of application,advantages and disadvantages of the existing methods of the sudden water pollution accident,this paper presents a new risk assessment method of Fuzzy Bayesian Network which is suitable for the uncertainty of water pollution accidents and the fuzziness of risk factors,and makes a case for verification.It provides a scientific method for emergency decision-making and risk management of sudden water pollution accidents.First of all,this paper deeply studies the risk formation mechanism of sudden water pollution accidents,determines the composition of the risk system,and conducts risk investigation and analyzes key factors in the process of risk occurrence.Through the use of field research,literature research and questionnaires and other means of identification,classification and screening of risk factors,the following results are obtained: Three major risk factors of sudden water pollution accidents and their specific risk factors are identified;Network nodes are introduced to identify networks Nodes,including 1 target node,9 intermediate nodes and 18 evidence nodes(parent nodes);The Bayesian network topology of sudden water pollution accident risk is established by Ge NIe software.Secondly,on the basis of Bayesian Network structure,a risk assessment model based on Fuzzy Bayesian Network for sudden water pollution accidents is constructed and the basic steps of model realization are proposed: Establishing a set of basic factors;Setting up evaluation set;Establishing membership matrix;Convert membership degree into probability;Calculating the fuzzy conditional probability;Reasoning and analyzing of Bayesian Network.In this study,a set of basic factors including 18 risk factors is established;The evaluation set of the four levels is constructed,namely,the risk of particularly significant sudden water pollution accidents(Grade ?),the risk of major sudden water pollution accidents(Grade ?),the risk of larger sudden water pollution accidents(Grade ?)and the risk of general sudden water pollution accidents(Grade ?),the degree of risk in turn described as extremely high,high,medium and low;it is determined that the membership degree is calculated by using the half-trapezoidal distribution function,and the membership matrix is constructed;the conversion formula of membership degree to probability and the formula of fuzzy conditional probability are determined;it is determined that Ge NIe is used to make inferences about the risk of sudden water pollution accidents.Finally,the risk assessment model of sudden water pollution accidents is applied to Lanzhou section of the Yellow River.According to the diagnostic reasoning calculation and sensitivity analysis of the Fuzzy Bayesian Network,the risk of sudden water pollution accidents in Lanzhou section of the Yellow River is “Medium”;the order of importance of the sources of risk,risk control and risk receptors of sudden water pollution accidents in Lanzhou section of the Yellow River is as follows: the factors of risk source,risk receptors,and risk control status;the factors of risk management should be paid attention to in turn as follows: the level of production technology and equipment,the discharge of wastewater,the discharge of wastewater,the condition of running roads,the rate of development and utilization of water resources and the emergency response mechanism.The results show that the combination of Bayesian Network and Fuzzy Theory can make up for the deficiency that the traditional Bayesian network can not deal with continuous variables and rely on large sample data to calculate the probability of accident of basic factors,and enhance the explanatory and scientific of the reasoning results of the Bayesian Network,the evaluation results are in line with the actual situation of the Yellow River Lanzhou section.The application of the risk evaluation model of the Fuzzy Bayesian Network not only provides a scientific basis for decision-making on risk management of sudden water pollution accidents in Lanzhou section of the Yellow River,but also provides reference and guidance for risk management of unexpected water pollution accidents in other sections.