A method for optimizing parameters of a nuclear power process twin system simulation model based on a PSO-TD3 hybrid framework

By optimizing the parameters of nuclear power simulation models using the PSO-TD3 hybrid framework, the problems of slow convergence speed and easy getting trapped in local optima in the optimization of model parameters in nuclear power systems are solved, and efficient and stable parameter optimization and dynamic tracking are achieved.

CN122263644APending Publication Date: 2026-06-23CHINA NUCLEAR POWER OPERATION TECH CORP +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHINA NUCLEAR POWER OPERATION TECH CORP
Filing Date
2026-03-25
Publication Date
2026-06-23

AI Technical Summary

Technical Problem

Existing nuclear power simulation model parameter optimization methods have slow convergence speed in high-dimensional, nonlinear systems, are prone to getting trapped in local optima, and are difficult to meet the real-time adjustment requirements of nuclear power systems.

Method used

A PSO-TD3 hybrid framework-based approach is adopted, which combines particle swarm optimization (PSO) algorithm and dual-delay deep deterministic policy gradient (TD3) reinforcement learning to dynamically adjust key parameters, construct a nuclear power process twin system simulation model, and optimize the nuclear power simulation model parameters.

Benefits of technology

It improves the dynamic tracking capability and parameter optimization efficiency of nuclear power simulation models, avoids local optima, and achieves fast convergence and high-precision model optimization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122263644A_ABST
    Figure CN122263644A_ABST
Patent Text Reader

Abstract

The present application relates to a kind of nuclear power process twin system simulation model parameter optimization method based on PSO-TD3 hybrid framework, comprising: based on demand, construct nuclear power process twin system simulation model, determine the optimization parameter contained in nuclear power process twin system simulation model;Loss function for measuring the combination performance of current each parameter of nuclear power process twin system simulation model is constructed, with the value of loss function minimization as optimization goal;With particle, each parameter of nuclear power process twin system simulation model is represented, based on particle swarm optimization algorithm, the PSO update function for updating the speed and position of particle is constructed;With the state parameter of PSO update function, TD3 state vector is constructed, and the output action of TD3 is mapped as the weight parameter of PSO update function;Based on optimal particle group, the parameter of nuclear power process twin system simulation model is updated and optimized.For solving the problem that nuclear power simulation model parameter space dimension is high, nonlinear is strong, traditional static optimization algorithm is easy to fall into local optimum.
Need to check novelty before this filing date? Find Prior Art