1. Improved adaptive inertia weight PSO algorithm and its application in nuclear power pipeline layout optimization
- Author
-
Yan LIN, Dengyue XIN, Xuanyi BIAN, Qiaoyu ZHANG, and Tieli LI
- Subjects
marine nuclear power ,primary loop system ,particle swarm optimization (pso) algorithm ,nonlinear inertia weights ,adaptive ,linear learning factor ,Naval architecture. Shipbuilding. Marine engineering ,VM1-989 - Abstract
ObjectivesThis study explores the use of a nonlinear adaptive inertia weight particle swarm optimization (PSO) algorithm to realize the optimal design of the path and arrangement of pipelines in the nuclear power primary loop systems of ships. MethodsAccording to the pipeline layout design characteristics, the constraints, evaluation functions and spatial model of the primary loop system are established. Based on the number of pipeline nodes, a new fixed-length coding method for the PSO algorithm is proposed, along with a direction guidance mechanism. As the standard PSO algorithm has such shortcomings as a slow convergence speed and susceptibility to falling into the local optimal solution, an improved nonlinear adaptive inertia weight PSO algorithm supplemented by a linearly changing learning factor is proposed. The improved PSO algorithm is combined with a co-evolutionary algorithm to form a co-evolutionary PSO algorithm for solving branch pipeline problems. The improved algorithm is then applied to the pipeline layout optimization problem of the nuclear power primary loop systems of ships. ResultsThe simulation results show that the convergence speed of the proposed algorithm is increased by 40% –50% compared with that of the standard algorithm. The improved algorithm can not only obtain higher quality pipeline layouts, but also solve the problem in which the standard PSO algorithm can easily fall into the local optimal solution. Conclusions The results of this study can provide useful references for the pipeline layout optimization of the nuclear power primary loop systems of ships.
- Published
- 2023
- Full Text
- View/download PDF