基于均匀设计的神经网络预测焊条熔化特性

THE MELTING CHARACTERISTICS PREDICTION OF THE ELECTRODE COVER OF ANN BASED ON THE UNIFORM DESIGN

利用VB语言开发了均匀设计软件并进行配方均匀设计试验,对神经网络模型的建立进行了详细的研究,用均匀设计的样本对建立的预测纤维素型焊条熔化特性的人工神经网络模型进行训练,试验结果表明:该模型可根据药皮成分对焊条熔化特性进行较为准确的预测。

The uniform designed software was developed using the Visual Basic language and the uniform designed experiments were carried out in this paper. The development of the neural network model was comprehensively analyzed. The artificial neural netmodel for predicting the melting characteristics of the cellulose electrode was trained using the uniform design samples. The resuits show that the model can exactly predict the melting characteristics according to the composition of the dectrode coating.