Multi-objective optimization of a steam generator

Lei CHEN,Changqi YAN,Jianjun WANG

DOI: https://doi.org/10.3969/j.issn.1006-7043.201212099

2014-01-01

Abstract:To receive higher precision from the non-dominated sorting genetic algorithm , an improved non-domina-ted solution sorting genetic algorithm was developed .The hyper-mutation strategy was employed to enhance the local searching in the improved algorithm , thus the precision of the non-dominated solution was increased .Based on the data from the Qinshan I nuclear power plant , an SG mathematical model was proposed; the optimization variables were chosen by the sensitivity analyses; the primary flow rate and weight of the SG were optimized by adjusting these variables through the multi-objective optimization theory .The results indicate that the non-dominated solutions obtained by the improved algorithm show better convergence and are more continuous than the solutions of the origi -nal algorithm .The primary flow rate reduction and SG weight increments are caused by the decreasing of the reactor inlet temperature , when the SG weight is less than 190 t.The primary flow rate reduction and SG weight increments are caused by the coolant speed decreasing in the U-tube, when the SG weight is greater than 225 t.The weight and flow rate can be optimized by 17.66%and 17.58%at most, respectively through the optimization design .