Simultaneous Electrochemical Determination of Glue and Thiourea in Copper Refining by Using Artificial Neural Network

摘要

In this research an artificial neural network was invented for simultaneous determination of both glue and thiourea in industrial electrolytes. First, a network was designed, which electrochemical data was its input. This network consisted of three layers of nodes, and back propagation rule was used for training the network. Sigmoid transfer function was applied in the hidden and output layers. The precision and accuracy of model was tested. The resulting MSE (Mean Square Error) for prediction of glue and thiourea were 0.0011 and 0.0008 respectively.