| 基于MPSO-ANN的火炮零部件硬度无损检测研究 |
| |
| 引用本文： | 陆军仁,艾东,柏逢明.基于MPSO-ANN的火炮零部件硬度无损检测研究[J].计算机测量与控制,2009,17(3). |
| |
| 作者姓名： | 陆军仁 艾东 柏逢明 |
| |
| 作者单位： | 1. 长春理工大学电子信息工程学院,吉林,长春,130022
2. 燕山大学,计算机科学与信息工程学院,河北,秦皇岛,066004 |
| |
| 基金项目： | 中国兵器科学院资助项目 |
| |
| 摘 要： | 依据音频无损检测原理,针对某火炮零部件(凸轮轴)硬度与音频参数的非线性映射问题,将ANN网络应用到音频检测中,同时采用改进粒子群算法(MPSO)优化BP神经网络结构和初始权值;MPSO算法把PSO算法的单向搜索变为多向搜索,提高了搜索精度,平衡了局部和全局搜索能力,较好地收敛到最优解,克服了BP网络结构难以确定和易于陷入局部极小值的缺点,实现了不同零部件硬度检测中的样本训练与预测,分别对内推、外推样本进行比较分析;结果表明其适应度逐渐趋于稳定并迅速收敛,精度满足要求.
|
| 关 键 词： | 音频无损检测 火炮零部件 MPSO算法 BP网络 硬度检测 |
Nondestructive Testing for Hardness of Artillery's Parts Using MPSO-ANN |
| |
| Lu Junren,Ai Dong,Bai Fengming.Nondestructive Testing for Hardness of Artillery's Parts Using MPSO-ANN[J].Computer Measurement & Control,2009,17(3). |
| |
| Authors: | Lu Junren Ai Dong Bai Fengming |
| |
| Affiliation: | 1.College of Electric Information Engineering;Changchun University of Science and Technology;Changchun 130022;China;2.Computer Science&Information Engineering college;Yanshan University;Qinhuangdao 066004;China |
| |
| Abstract: | Aimed at nonlinear mapped relation between the hardness of artillery's parts and sonic parameters,basis on sonic nondestructive testing theory,this paper applied ANN to sonic testing,and put to use modified particle swarm optimization(MPSO),to optimize the framework of MPSO-BP Neural Networks and initialize Weight.MPSO algorithm translated from singleness into multitudinous different from PSO,so the precision of the algorithm was improved effectively and it balanced part and whole search and could converge ... |
| |
| Keywords: | sonic nondestructive testing parts of artillery MPSO algorithm BP network hardness detection |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
