Modeling the mechanical properties of biopolymers for automotive applications

The automotive industry is constantly looking for alternative solutions to reduce manufacturing cost and use renewable materials. Implementing agro-fibres as polymer fillers in thermoplastic matrix will satisfy the automotive criteria without sacrificing the mechanical properties currently set by the conventional fillers such as glass fibre, talc, or mica. This paper proposes the use of wheat straw as filler in polypropylene for automotive industry and investigates models for determining compositions of the materials to correspond to mechanical properties. Data collection is performed by varying weight percentages of wheat straw and polypropylene to create the biocomposites through an extrusion process. The end products are molded into proper shapes for mechanical testing. Different modeling approaches that include polynomial regression, artificial neural networks and support vector machines are investigated to prepare predictive models for the biocomposite properties. A comparison between the methods shows that support vector machines produced the best model, followed by artificial neural networks, and then polynomial regression.