Multi-objective optimization design of steam generators based on a novel dual population genetic algorithm

Aiming at the limitations of poor convergence and low optimization accuracy of genetic algorithm, a novel dual-population genetic algorithm (DPGA) is proposed and used to optimize the weight and volume of a vertical U-tube natural circulation steam generator (SG) in this study. Individual differences and population diversity is used to interfere with population crossover, mutation, and information exchange between populations. Quantum behavior operations and Kmeans clustering methods are integrated to further improve global search capability. The test results demonstrate that the algorithm proposed exhibits a high optimization rate. Furthermore, the multi-objective optimization variables are determined with the utilization of sensitivity analysis. Under geometric and engineering constraints, the DPGA achieves a 22.2245% reduction in weight and an 11.2309% reduction in volume compared to the baseline SG design, demonstrating superior convergence and optimization efficiency.