Application of artificial neural networks for the prediction of the effects of batch variations on the properties of rubber mouldings

In order to describe and predict the effect of the different process steps on the part quality, the comprehensive inspection of the whole rubber manuf acturing process chain becomes necessary. For the description of the comple x connections between the different process steps artificial neural network s and regression models are capable tools. In this paper the possibilities and limitations of artificial neural networks describing the effect of batc h and storage variations on the properties of rubber mouldings compared to regression models are discussed.