Determination of the extended Drucker–Prager parameters using the surrogate-based optimization method for polypropylene nanocomposites

Abstract

In this article, a new method is proposed to identify the constants of the extended Drucker–Prager yield surface for polypropylene nanocomposites. The method is based on optimizing the difference between the numerical and the experimental results of a three-point bending test. The test specimens are made of polypropylene/nanoclay and polypropylene/nano-calcium carbonate nanocomposites with different nanoparticles content. Moreover, the effect of composite filler content on the extended Drucker–Prager constants of polypropylene, as the composite matrix, is investigated. Inasmuch as numerical simulation is usually very time-consuming and highly nonlinear, a surrogate-based model with radial basis function is used to approximate cost function. For calibration of the surrogate model, sample points are selected using the Latin hypercube sampling technique which enjoys good filling-space properties. The results indicate that the cohesion strength and elastic modulus of the neat polymer are improved by adding nanoparticles. The improvement, however, is more significant for the polypropylene/nanoclay with respect to that for the polypropylene/nano-calcium carbonate. The nanoparticles were found not to be considerably effective on the dilation angle of the neat polymer. In addition, it was found that the non-associative plastic flow assumption can describe plastic deformation of polypropylene composites more properly. The surrogate-based optimization method was validated by identifying some of the extended Drucker–Prager constants directly from tension and compression tests.