Neural Network Modeling to Control Process of Induction Soldering
Neural Network Modeling to Control Process of Induction Soldering
复制标题
神经网络建模控制感应焊接过程
DOI：
10.1109/icieam.2019.8743031
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
A. Murygin
中科院分区：
文献类型：
--
作者：
A. Milov;V. Tynchenko;A. Murygin
航天器波导路的感应焊接工艺过程非常耗时。这种过程控制的质量受各种外部因素的影响，主要与测量仪器的误差有关。经典方法不能提供对快速流动过程的参数的充分控制。提出了一种基于人工神经网络的控制方法。本文对该方法进行了分析，提出了一种基于人工神经网络的感应焊接工艺过程控制方法，并对不同结构的人工神经网络的有效性进行了比较分析，选择了最有效的人工神经网络。采用基于人工神经网络的航天器薄壁铝波导轨感应焊接工艺控制方法，可以提高该工艺过程的控制质量，减少测量工具的规则性误差和非标性误差的影响，减少人为因素的影响，从而提高产品质量。
The technological process of induction soldering of spacecraft wave guide paths is very time-consuming. The quality of such process control is influenced by various external factors, mainly related to errors in measuring instruments. Classical methods do not provide sufficient control over the parameters of a fast-flowing process. This paper proposed using a control method based on artificial neural networks. The article analyzes the proposed method, develops a method for controlling the technological process of induction soldering based on artificial neural networks, and also performs a comparative analysis of the effectiveness of artificial neural networks with different structures and the selection of the most effective one. Using the method of technological control of the induction soldering of spacecraft thin-walled aluminum wave guide paths based on artificial neural networks can improve the quality of control of this technological process, reduce the impact of regulatory and non-standard errors of measurement tools, reduce the impact of human factors, and, consequently, improve the quality of products.