焊接工艺参数的神经网络智能设计

Intelligent Design of Welding Procedure Parameters Based on Neural Networks

关键词:  人工神经网络  焊接工艺  智能设计

Key words:  artificial neural network  welding procedure  intelligent design

摘要:

将人工神经网络技术引入焊接专家系统,建立了多种焊接规范参数设计网络模型。以实际焊接工艺数据对所建网络模型进行训练,经测试表明,网络具有很强的自学习自适应能力、容错与协调能力以及显著的联想功能。本文有效解决了焊接工艺设计专家系统的关键困难——焊接规范参数设计,为焊接专家系统研究开辟了新的方向。

Abstract:

The artificial neural network was introduced into welding expert system and a series of neural network models were set up by the authors for the design of welding parameters.These models were trained with actual welding procedure data. Testing results of these models show that they possess strong abilities in self-learning, self-adapting, error tolerating, coordinating and associating. The research of this paper has solved the most difficult problem of welding procedure design expert systems-the design of welding parameters. This opens up a new direction for the study of welding expert systems.