基于自适应代理模型的非能动系统可靠性分析  

Passive system reliability assessment based on an adaptive surrogate model

作　　者：王晨阳 夏庚磊 彭敏俊 徐青蓝 陈果 WANG Chenyang;XIA Genglei;PENG Minjun;XU Qinglan;CHEN Guo(Fundamental Science on Nuclear Safety and Simulation Technology Laboratory,Harbin Engineering University,Harbin 150001,China;Science and Technology on Reactor System Design Technology Laboratory,Nuclear Power Institute of China,Chengdu 610213,China)

机构地区：[1]哈尔滨工程大学核安全与仿真技术国防重点学科实验室,黑龙江哈尔滨150001 [2]中国核动力研究设计院核反应堆系统设计技术重点实验室,四川成都610213

出　　处：《哈尔滨工程大学学报》2024年第2期383-389,共7页Journal of Harbin Engineering University

摘　　要：许多先进反应堆采用非能动设计来提高安全性,开展可靠性分析研究是推动其广泛应用的重要环节。由于非能动安全系统依赖于自然循环等物理法则实现系统功能,更容易受到不确定性因素的影响。热工水力过程失效是导致非能动系统失效的主要原因,而传统的故障树等方法无法分析这类失效概率。本文通过热工水力结果训练代理模型,并基于自适应抽样策略有效减少热工水力程序调用次数,通过高度非线性测试函数验证了算法的计算效率,并应用于某一体化压水堆的非能动余热排出系统。计算结果表明:采用自适应克里金模型相比于传统蒙特卡罗与传统克里金模型方法具有更高的计算效率。Passive design is adopted by many advanced reactors to enhance safety,and reliability assessment is im-portant for its wide application.A passive safety system is more susceptible to uncertainties because its work de-pends on physical laws such as natural circulation.Thermal hydraulics process failure is the main contribution to passive system failure;however,traditional methods such as fault tree analysis are unavailable for analyzing this type of failure possibility.In this study,a surrogate model is trained with thermal-hydraulic simulation results.Fur-thermore,the adaptive sampling strategy is applied to effectively reduce the call times of the thermal hydraulics pro-gram.The algorithm is validated using a highly nonlinear test function and then applied to a passive residual heat removal system of an integral-type pressurized water reactor.The calculation results indicate that the adaptive Krig-ing model has higher computational efficiency than traditional Monte Carlo and Kriging model methods.

关 键 词：非能动安全系统 可靠性 代理模型 概率安全分析 一体化压水堆 RELAP5 自适应抽样 

分 类 号：TL33[核科学技术—核技术及应用]