Application of the hybrid distance algorithm to the optimal design of a steam generator
Yan Chang-q · Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University · 2014
The non-dominated solutions obtained by the traditional distance algorithm in solving nonlinear constraint optimization problems are usually narrowly-distributed,discontinued and not precise. In order to solve these issues, a hybrid distance algorithm has been proposed. With this algorithm,the reflection operation in the complex algorithm was introduced to carry out further searches for the newly-found solutions; in addition,the hyper-mutation strategy for the immune clone algorithm was employed to further seek out the optimal non-dominated solutions. By comparing the standards,the qualities of non-dominated solutions were tested. The results showed that the nondominated solutions obtained by the hybrid distance algorithm are widely distributed and more continual,even more convergent than with the distance algorithm. The application of the hybrid distance algorithm realizes the multi-objective optimized design of the steam generator with the weight and primary side flow as the objective. The optimization results showed that,in the front of the non-dominated solutions,there is a discontinued area,and the reactor inlet temperature and the coolant flow speed in the U-tube influence the distribution of the non-dominated solutions.