Study Of Dynamic Point Core Mdel Combined With The Neural Network For A PWR

The Japanese tsunami-affected nuclear spill aftershocks at the stage is not over yet, but as the global energy crisis coming, nuclear energy’s research and development will enter a new era because of its contribute to the environment and energy. The incident also reflects the safety of nuclear reactors will be the focus of future research for the study and control of pressurized water reactor core will be more strict and cautious.This article is the research of dynamic PWR core combined with ANN. For the calculation of the reactor core neutron flux, the traditional methods is introduced first, including Hermite interpolation, Gear, Prompt jump approximate and so on. This article uses Euler-difference algorithm based on the classic method of Prompt jump approximate to calculate it’s numerical value and by a simple reaction experiments verify the feasibility and accuracy. The numerical solution method is simple, less time-consuming, it can meet a certain engineering requirements within the error allowed range.BP neural network has a strong operability and a simple structure, its learning algorithm has a better effect for curve-fitting. It can get a good effect by NN modeling to follow the setting parameter of the original in the considering of influencing factor of the reactor by power factor and the changing of control rod. The negative reactivity of the poison is calculated by numerical method, its produce and change is analyzed too.After coupling the various influencing factor BP neural network is used to calculate the reactivity. In a certain condition, the fuel latent reactivity is supposed a definite number, the neutron density can get a stable number via the reactivity is controlled.Neural network plays a good role in studying and controlling the reactivity, it saves time and easy to operate. It is used to real-time tracking control the influencing factor of reactivity.