Neural Network Modeling to Control Process of Induction Soldering

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.