Hybrid model and method for determining the mechanical properties and processing properties of an injection moulded article

A model has a first database (12) for storing correspondences between a trade product and its respective components,units for input of consituent parameters, preferably trade product names, which specify the composition of an injection molded part, units for accessing the first database for determining the component parameters from the list of constiutents and a neural network (NN1-NNK) for inputting component parameters and output of predictive values. An Independent claim is made for a method for determining the mechanical and manufacturing properties of an injection molded component using a hybrid model comprised of a number of neural networks and rigorous models (R1-R4). The rigorous models can be described with partial models that are described with mathematical formulae. The neural part model serves to image processes that are only connected via data.