Uncertainty Analysis Of Main Steam Line Break Accident Transient Model Based On Surrogate Model

The Main Steam Line Break(MSLB)is an important accident that threatens the integrity and safety of the AP1000 containment.When MSLB accident occurs,the break of main steam pipeline will cause a large amount of steam to be injected into the containment,resulting in a sharp increase in temperature and pressure in the containment.In order to study the key influence parameters of transient response of AP1000 containment in MSLB accident and quantify the comprehensive influence of these uncertain input parameters on key safety parameters,various of uncertainty analysis algorithms are used to analyze the sensitivity and uncertainty of input parameters and critical safety output in MSLB accident in this paper,which provides data for the Best Estimate Plus Uncertainty analysis(BEPU).Traditional sensitivity and uncertainty analysis is usually based on sampling methods,which need a large number of samples.The calculation time is long if a complicated thermal hydraulic model is used.If the traditional sampling method is used for sensitivity and uncertainty analysis,the calculation efficiency is low,and the cost of computer resources is high.More importantly,sometimes it is even difficult to implement.Therefore,it is urgent to develop more efficient sensitivity and uncertainty analysis methods to meet the accuracy requirements to quickly obtain the sensitivity indices of key impact parameters and the uncertainty distribution of key safety parameters.In this paper,the more efficient surrogate model algorithms are studied,including Polynomial Chaos Expansion(PCE)and Stochastic Collocation(SC).By theoretical derivation and program development,a more advanced high order Arbitrary Polynomial Chaos Expansion method(APC)without subjective judgment distortion is studied,which solves the problem effectively and subjective interference of traditional sampling algorithm.In this paper,the accuracy of the surrogate model is verified by comparing with the traditional large-scale sampling model.In this paper,the transient analysis model of the AP1000 containment is established.At the same time,the sensitivity indices of key input parameters and the uncertainty distribution of key safety parameters are obtained by coupling the sensitivity and uncertainty analysis model with the transient analysis model of AP1000 containment.This paper analyzes the sensitivity indices of the key influence parameters such as atmospheric temperature and atmospheric pressure.The results show that the influence of atmospheric temperature and atmospheric pressure on the peak pressure and temperature in containment is very important under MSLB accident.There is a strong interaction between atmospheric temperature and atmospheric pressure.At the same time,the uncertainty probability distribution of key safety parameters such as peak pressure and peak temperature of containment under MSLB accident is obtained.This paper shows that the surrogate model algorithm is more efficient than the sampling method,and can greatly reduce the computational cost and resources.When the accuracy is satisfied,the surrogate models algorithm can greatly reduce the calculation cost,and save the calculation resources.The efficiency of sensitivity and uncertainty analysis is improved by hundreds of times.The research results of this paper can provide a more efficient sensitivity and uncertainty analysis method to meet the accuracy requirements for the application of optimum analysis and uncertainty analysis in complicated nuclear reactor safety analysis,which has important academic value and engineering application value.