Construction of a Prediction Model for an Acceptable T-shape Tube Product in Magnesium Alloy Hydro-forming

The finite element method in conjunction with abductive neural network is applied to predict an acceptable T-shape product of which the minimum wall thickness and the protrusion height being fulfilled the industrial demand in the magnesium alloy hydro-forming. AZ31 magnesium alloy circular tube is used as the billet material for hydro-forming with hydraulic pressure as the main forming power combined with the mechanically auxiliary force from the punch to fabricate a qualified T-shape tubing product. Finite element software is adopted to investigate the forming characteristics of T-shape tube firming, by changing process parameters such as punch velocity, hydraulic pressure, fillet radius of the die and tool-workpiece interface friction etc. to investigate the material flow of tube forming, and the variations of wall thickness and protrusion height. And the neural network method in turn is applied to synthesize the data sets obtained from the numerical simulations and a prediction model for an acceptable T-shape tube product in magnesium alloy hydro-forming is thus constructed. From the prediction model, a suitable range of the process parameters variation fir producing those qualified T-shape tubes that may be accepted in industrial applications can thus be identified.