针对神经网络在预测复合镀层性能方面的应用情况,以及传统的BP神经网络存在缺陷;通过对RBF神经网络的基本原理和特点的研究,建立了利用RBF神经网络对Ni-TiN纳米复合镀层显微硬度进行预测的模型。通过实验数据验证了所建立的RBF神经网络模型具有很高的精确度,其最小相对误差可达0.62%,而且所建立的预测模型具有优化工艺参数的功能,对复合镀层的其它性能进行预测具有指导意义。
Against the application of the neural networks in forecasting performance of the composite coating and considering the defects of the traditional BP neural network, a new model of using the RBF neural network to predict the microhardness of the Ni-TiN nanocomposite coating was established by the study of the basic principles and the characteristics of RBF neural network. Model of RBF neural network which has high accuracy was confirmed by experimental data, the minimum relative error is 0.62%, and that the model has the function of optimization process parameters, it has important significance for using the neural network to predict the other capability of composite coatings.
参考文献
| [1] | 沈彬,刘磊,邓意达,胡文彬.电铸制备SiC颗粒增强镍基复合材料的工艺与性能[J].机械工程材料,2009(01):76-79. |
| [2] | 夏法锋,贾振元,吴蒙华,王帆,霍峰.用人工神经网络优化Ni-纳米TiN复合镀层的超声-电沉积工艺[J].稀有金属材料与工程,2008(08):1479-1482. |
| [3] | 仵亚婷 .自润滑Ni-P化学复合镀层的制备及其摩擦磨损行为研究[D].上海交通大学,2005. |
| [4] | 方莉俐,张兵临,禹建丽,姚宁.用人工神经网络预测电铸自支撑金刚石-镍复合膜沉积结果[J].稀有金属材料与工程,2006(04):638-641. |
| [5] | 谌爱文 .基于BP和RBF神经网络的数据预测方法研究[D].中南大学,2007. |
| [6] | 董会丽,郑世杰.基于RBF神经网络的复合材料层合壳荷载识别[J].工程力学,2008(03):64-67. |
| [7] | 李仁军,刘宏昭,原大宁.基于RBF神经网络的混合输入机构自适应控制[J].农业机械学报,2010(04):204-208. |
| [8] | 张冬青,宁宣熙,刘雪妮.基于RBF神经网络的非线性时间序列在线预测[J].控制理论与应用,2009(02):151-155. |
| [9] | 孙大洪,王发展,刘强,张顾钟,原思聪.基于BP和RBF神经网络的滚动轴承故障诊断比较[J].轴承,2010(02):53-56. |
| [10] | 白洁.基于模糊神经网络的往复压缩机的性能预测研究[J].煤矿机械,2010(11):68-69. |
| [11] | 赵雪松,疏达,肖平.基于神经网络的脉冲电化学抛光建模方法研究[J].机械制造,2009(06):11-13. |
| [12] | Z. Ji;Y. Li.The Application of RBF Neural Network on Construction Cost Forecasting[A].Moscow,Russia,2009:102-107. |
| [13] | Y. LIU;S. L. ZHAO;C. YI.The Forecast for Corrosion of Reinforcing Steel Based on RBF Neural Network[A].Baoding,China,2009:72-76. |
| [14] | 朱大奇;史慧.人工神经网络原理及应用[M].北京:科学出版社,2006 |
| [15] | 杨旭 .基于RBF神经网络的工业过程建模与优化研究[D].哈尔滨理工大学,2009. |
- 下载量()
- 访问量()
- 您的评分：
-
10%
-
20%
-
30%
-
40%
-
50%