Research And Development Of Polymer Rheological Parameters Prediction System

The rheological property of high polymer is an indispensable factor for designing mould, choosing processing equipment, deciding craft of processing correctly and filling mould, flow analysis and simulation of the melting body of high polymer. But rheological property of high polymer is affected largely by shear rate, temperature and pressure etc, and the factors take on complicated non-linear relations each other, so it is difficult to describe this exactly with equation of rheology. Further more, high polymer is various and the same kind of production produced by different manufactory has quite different rheological property due to the difference of distribution of their relative molecular weight, which cause the inconvenience for designing mould and ascertaining processing craft of the processing enterprise of high polymer. Therefore, it has become an urgent problem to be solved immediately to measure accurately the rheological property data of high polymer under different condition. However it is impractical and unnecessary that the rheological property data are obtained by tests under all conditions, the effective method is only using limited tests to predict the rheological properties data of high polymer only. With wide use of non-linear science and technology in various fields, such as the artificial neural network and so on, has offered the new way for solving the problem because of its superior advantages. Integrating the artificial neural network and classical rheology theory on the basis of forefathers' experience, author chose the BP network that is used extensively to solve the problem of approaching function to predict the rheological property of high polymer and took some useful measures to improve the structure of BP network. After summarizing the method and principle of setting up the neural network, integrating the advantage of VB and MATLAB, author has developed an integrative internet application that can be used to predict the rheological property of high polymer by method of hybrid programming. Validated by the experiments, predicting result of the software is more satisfying and it can be used for instructing project applications and production practice.