Multi-objective Optimization of Spiral Plate Heat Exchanger Based on NSGA-Ⅱand BP Neural Network
In this paper,a multi-objective optimization scheme for structural parameters of spiral plate heat exchanger with spoilers is proposed.This project mainly uses NSGA-Ⅱ,BP neural network and CFD orthogonal test to find the optimal structural parameters of spoilers in the given environment.First,the nonlinear mapping relationship between spoiler structure parameters and optimization variables can be obtained according to the analysis on the data from 16 groups of models set up for orthogonal simulating test.Then,take convective heat transfer coefficient and pressure drop as objective function,using NSGA-Ⅱalgorithm to optimize the extremum of the optimized BP neural network to obtain the Pareto solution set under the optimization objective.Finally,the multi-objective optimization result about the spoiler structure can be acquired through the utilization of TOPSIS evaluation method.The simulation results show that the predicted value error of the optimized structure and the multi-objective optimization algorithm is less than 15%,and the convective heat transfer coefficient h is increased by 31.4%compared with the traditional spiral plate heat exchanger.It indicates that the optimization method suggested in this paper is well practicable for the heat exchangers'structure improvement and engineering application to providea worthwhile reference.