Application Research For Polynomial Chaos Expansion In Uncertainty Quantification Of Reactor Calculation

In the real physical world of nuclear reactors,there will be some natural uncertainty,including geometric uncertainty,initial conditions and boundary conditions uncertainty,physical chaotic state uncertainty,etc.This uncertainty may make the decision according to simulation results based on deterministic assumptions risk of error.At present,in the field of nuclear engineering,the "best estimation plus uncertainty analysis" method is recommended to replace the traditional conservative safety analysis methods and a series of uncertainty analysis methods for system codes are created.With the continuous development of computer technology and requirements of nuclear safety analysis for high-fidelity physical models and effective accurate codes,high-precision simulation techniques,such as CFD technology,are increasingly used in the field of nuclear power plant design and safety analysis.The cost of the simulation accuracy improvement is the consumption of a large amount of computing resources,which makes many uncertainty quantification methods based on the results of multiple codes calculations become difficult to implement.Therefore,it is of great significance to carry out the application research of the uncertainty quantification method for reactor thermal-hydraulic calculations based on high-precision simulation codes.In order to reduce the CPU cost required for uncertainty quantification and make full use of mature commercial CFD calculation software,based on the non-intrusive polynomial chaos expansion,this paper presents an uncertainties propagation analysis scheme suitable for the reactor CFD calculation.Taking the CFD model of the 5×5 coolant channel as the research object,uncertainty quantification and sensitivity analysis of output parameters caused by the uncertainties of boundary conditions are performed.At the same time,the uncertainty quantification of the above model based on the Monte Carlo method is also carried out.to verify the computational accuracy and economy of the polynomial chaos expansion method.Then,the heat transfer model between fuel rods and coolant is introduced into the CFD model,and the uncertainty quantification of the coolant channel under different core power conditions considering the heat transfer process is performed.There are some complex geometrical areas in the core,and these special geometric structures may affect the uncertainty of output parameters.Based on this consideration,this paper also takes the CFD model of the coolant channel with grid spacers and mixing vanes as the research object,and carries out the uncertainty quantification based on the polynomial chaos expansion method.The characteristics of the uncertainty in the complex structure area of the core are obtained and sensitivity analysis results are also given.Finally,this paper compares the uncertainty quantification results between using Gaussian distribution and uniform distribution as input parameters distribution,and discusses the influence of different input parameter probability distribution types on the uncertainty analysis results.