华中科技大学学报（自然科学版）2025，Vol.53Issue(7)：115-121,7.DOI:10.13245/j.hust.250405
基于DBO-BP神经网络的活动导叶磨蚀预测模型
Active guide vane abrasion prediction model based on DBO-BP neural network
摘要
Abstract
In order to efficiently predict the abrasion of active guide vane of Francis hydro turbine,the abrasion model of composite resin mortar coating material was fitted based on the high-speed Gaza experimental data.The abrasion model was compiled using user defined function(UDF)on the Fluent platform to achieve simulation analysis of erosion of guide vanes under different operating conditions.A new and efficient erosion prediction model was proposed using a BP neural network optimized by the dung beetle optimization algorithm.This model can predict the future abrasion with the parameters of the turbine flow rate,particle concentration in the flow path and the current abrasion amount.The results show that the dung beetle optimization algorithm reduces the root mean square error of BP neural network by more than 40%,and the average absolute error by 60%,which improves the computational accuracy of the BP neural network.关键词
磨蚀模型/蜣螂优化算法/BP神经网络/活动导叶/复合树脂砂浆涂层Key words
erosion model/dung beetle optimization algorithm/BP neural network/guide vanes/composite resin mortar coating分类
能源与动力引用本文复制引用
陈小翠,姬中瑞,郑源,陈文杰..基于DBO-BP神经网络的活动导叶磨蚀预测模型[J].华中科技大学学报（自然科学版）,2025,53(7):115-121,7.基金项目
国家自然科学基金面上项目(52271275),中央高校基本科研业务费专项资金资助项目(2018B05614). （52271275）
