Research On Modeling,simulation And Model Reduction Of Primary And Secondary Side Fluids In Steam Generators

The rapid development of digital twin technology will make the nuclear power system gradually transform to Digital transformation,which will further promote the digitalization,safety and convenience of nuclear energy applications.The digital twin implementation of nuclear power plants requires very high computational accuracy and efficiency of digital models.For the more mature traditional numerical calculation methods,the computational accuracy and efficiency are often inversely proportional,so it cannot meet the requirements of high accuracy and efficiency in digital twin modeling.However,model reduction methods can greatly reduce the computational time of the original model and improve computational efficiency while ensuring high accuracy.As one of the complex equipment in the nuclear power system,the steam generator is an important research topic for quickly and accurately predicting the flow field distribution of the primary and secondary side fluids after a certain physical process occurs.This article conducts research on the simulation and model reduction of the primary and secondary fluid models of steam generators,which will contribute to the implementation of digital twins in nuclear power equipment and systems in the future.The first part is the establishment and validation of finite volume models for the primary and secondary side fluids of steam generators.Firstly,the fluid flow regions in the primary and secondary sides of the steam generator were analyzed,and the geometric modeling of the heat transfer tubes and the primary and secondary sides of the steam generator were established.Then,the geometric modeling was reasonably meshed,and the finite volume models of the primary and secondary sides of the steam generator were established after setting their material properties and boundary conditions.Then,the established model is used to calculate the flow field distribution of the primary and secondary sides of the steam generator under steady-state full load conditions,analyze its thermal hydraulic characteristics,and compare it with actual operating parameters to verify the accuracy of the model.Finally,the universality of the finite volume models for steam generator primary and secondary side fluids under different operating conditions is verified by calculating and analyzing the thermal parameter changes of the flow field under different steady-state loads using the established model.The second part is the validation study of reduced order model methods.In order to verify the feasibility of the model reduction method,model reduction studies were conducted using two cases: a liquid cooled plate with a small number of grids and a finite volume model of the fluid in the main pipeline of a nuclear power plant.Firstly,the input and output parameters of the model are specified and parameterized.Then,experimental design(DOE)is used to design multiple sets of different operating conditions,and the finite volume model is used to calculate the training data of the flow field under different operating conditions.Then,singular value decomposition(SVD)and response surface methodology(RSM)are used to reduce the training data and generate a reduced order model(ROM).The computational efficiency and error of the reduced order model and the full order model are compared,Verify the effectiveness of the reduced order model method.The third part is to conduct model reduction and error analysis research on the established finite volume models of the primary and secondary side fluids of the steam generator.Firstly,training datasets for flow fields under different operating conditions are generated using finite volume models of the first and second side fluids of the steam generator.Then,the SVD reduction method and modal coefficient interpolation method are used to reduce the order of the training dataset to generate temperature reduction models for the first and second side fluids of the steam generator.The calculation errors and time of the reduced model and the full order model are compared to verify the accuracy of the first and second side fluids of the steam generator The high-precision and high-efficiency performance of the second-order fluid reduction model is analyzed,and the computational errors of the reduced model are studied to investigate the effects of the number of modes and the number of snapshots in the learning set in the SVD reduction method on the computational accuracy of the reduced model.