Study On Design Optimization Of Nuclear Power Plant Equipment

The performance of nuclear power plant equipment can be optimized with designparameters variation and overall performance can be calculated by optimizing the weight orvolume of the equipment. Computer codes based on the optimization theory are developedand used to analyze the optimization of nuclear power plant equipment.Based on the characteristic of nuclear power plant equipment design, Different Evolution(DE) optimization algorithm is used as main algorithm due to its simplicity and effectiveness.A series of improvements of DE are discussed.1) A novel adaptive normalization constrainthandling method is proposed in order to improve the performance of constraint handling inDE optimization algorithms.2) A hybrid optimization algorithm SADENM is developed byincorporating the local optimizer and self adaptive control parameters means. The use of localoptimizer helps to speed up the convergence process of the standard DE and self adaptivecontrol parameters helps to avoid the neediness of adjusting control parameters for differentkind of problems. The comparison on a suite of widely used constraint benchmark problemswith several state of the art evolution algorithms is performed. The results show thatSADENM outperforms the other algorithms with faster convergence and more precise andaccurate optimization.3) In order to solve the multi-objective nuclear power plant equipmentdesign problems, SADENM is updated with the ability to handle multi-objective constraintoptimization problems.The optimization process of nuclear power plant equipment design with the improvedoptimization algorithm is carried out for steam turbine as example equipment. The steamturbine evaluation programming code is developed on the basis of steam turbine designmathematical model and validation is carried out. The steam turbine design parameters areselected as optimal variables with unchanged external parameters. The weight, volume andefficiency of steam turbine are optimized as single-objective optimization problem withSADENM algorithm. Further More, multi-objective optimization of steam turbine is solvedusing SADENM algorithm.The traditional optimization algorithm and equipment evaluation program couplingapproach generally takes long runtimes hence has high computational cost. So the application of surrogate model in optimization of nuclear power plant equipment design is discussed. Aspecial version of SADENM algorithm with online-updating surrogate model is developedwhich shows superior convergence rate than the original SADENM for weight optimizationof steam turbine. The feasibility of using global surrogate model for optimization of nuclearpower plant equipment is verified for the natural circulation steam generator volume. Theresults show that optimization of nuclear power plant equipment using global surrogate modelsignificantly reduces the overall runtime maintaining high accuracy.