Neural Networks Approach for Characterization of Non-Isothermal Thermoplastic Membrane

AbstractRecent developments in computer-aided polymer processing have brought with them the need for an accurate description of the behaviour of industrial thermoplastic membranes under the combined effect of applied stress and temperature. In order to serve this purpose, we consider a non-isothermal approach to characterize the ABS (Acrylonitrile-Butadiene) membrane under biaxial deformation using the bubble inflation technique. Thereafter, Rivlin's theory of hyper-elasticity is employed to define the constitutive model of flat circular membranes. The nonlinear equilibrium equations of the inflation process are solved using finite difference method with deferred corrections. For the final step, a neuronal algorithm (ANN model) is employed to minimize the difference between calculated and measured parameters to determine material constants. The effect of experimental temperature (between -30 and 80 oC) on behaviour is considered in this work