On the prediction of parameters for the glass-rubber model for polyethylene terephthalate (PET) based on observed data in the injection stretch blow moulding (ISBM) process

In this publication a method is presented, that provides a shortcut to find material model parameters for rubber-like materials. Here it is shown exemplarily for PET and especially rPET types. This Method uses a simulation approach to create a dataset consisting of a novel process image representation for the blow moulding process and the corresponding material model parameters. This dataset is used to train an artificial neural network to identify the model parameters in the process image representations. Further, it is proposed to use the same process image representation, created from experimental blowing processes, to identify key material model parameters for experimentally tested materials. The suitability of the method is shown by using an artificial neural network trained to predict key material parameters of the Glass-Rubber model. The validity of the neural network and its prediction is exemplarily shown on the example of the inextensibility factor, a key material parameter of the Glass-Rubber-Model.