| 基于遗传神经网络的镍基高温合金激光熔覆层形貌质量预测 |
| |
| 引用本文： | 杨友文,田宗军,潘浒,王东生,沈理达.基于遗传神经网络的镍基高温合金激光熔覆层形貌质量预测[J].焊接学报,2013(11):78-82. |
| |
| 作者姓名： | 杨友文 田宗军 潘浒 王东生 沈理达 |
| |
| 作者单位： | 南京航空航天大学机电学院, 南京 210016,南京航空航天大学机电学院, 南京 210016,南京航空航天大学机电学院, 南京 210016,南京航空航天大学机电学院, 南京 210016;铜陵学院机械工程系, 铜陵 244000,南京航空航天大学机电学院, 南京 210016 |
| |
| 基金项目： | 国家自然科学基金资助项目(51205198);中国博士后科学基金面上资助项目(2012M 511266);江苏省博士后科研计划资助项目(1102052C) |
| |
| 摘 要： | 采用反向传播(back propagation,BP)人工神经网络(artificial aeural network,ANN)和遗传算法建立了激光熔覆层形貌质量(熔覆层高度、宽度及稀释率)与激光功率、送粉速率和扫描速率之间的遗传神经网络预测模型.设计正交试验得到预测模型训练样本数据,并在正交试验的基础上,用极差分析法分析了各加工参数对熔覆层形貌质量各个指标的影响规律.经过试验验证,遗传神经网络模型预测值与试验实测值误差不大于4.6%.结果表明,运用该模型可以为准确的选择镍基高温合金激光熔覆参数提供一定参考,从而有利于提高镍基高温合金激光熔覆层形貌质量.
|
| 关 键 词： | 激光熔覆 形貌质量 人工神经网络 遗传算法 极差分析 |
| 收稿时间： | 2012/7/1 0:00:00 |
Geometry quality prediction of Ni-based superalloy coating by laser cladding based on neural network and genetic algorithm |
| |
| YANG Youwen,TIAN Zongjun,PAN Hu,WANG Dongsheng and SHEN Lida.Geometry quality prediction of Ni-based superalloy coating by laser cladding based on neural network and genetic algorithm[J].Transactions of The China Welding Institution,2013(11):78-82. |
| |
| Authors: | YANG Youwen TIAN Zongjun PAN Hu WANG Dongsheng and SHEN Lida |
| |
| Affiliation: | College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China,College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China,College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China,College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China;Department of Mechanical Engineering, Tongling College, Tongling 244000, China and College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China |
| |
| Abstract: | Combination of back-propagation (BP) artificial neural network (ANN) and genetic algorithm was used to set up genetic neural network model to predict the quality of laser cladding layer according to the laser power,powder feed rate and scan rate.An orthogonal test was designed to obtain the training data of prediction model,and then the influence of different process parameters on the cladding layer geometry quality was analyzed by the method of range analysis.The validation results show that the relative error between predicted values and experimental data is less than 4.6%,which indicated that the use of the model can accurately select cladding parameters to improve the geometry quality of the laser cladding layer of nickel-based superalloy. |
| |
| Keywords: | laser cladding geometry quality artificial neural networks genetic algorithms range analysis |
|
| 点击此处可从《焊接学报》浏览原始摘要信息 |
|
点击此处可从《焊接学报》下载全文 |
|
