Electronic Thesis and Dissertation Repository


Doctor of Philosophy


Chemical and Biochemical Engineering


Dr. Andrew Hrymak


Long Fiber Thermoplastics (LFT) are promising new materials with high physical properties and low density. These high properties are obtained by embedding very long fibers (~100 mm) into a thermoplastic matrix. Such a high fiber length dictates the use of a compression molding process for manufacturing as the length of discontinuous fibers in injection molding is limited by pellet length.

LFT composites are of great interest for the automotive industry. These materials are already used in some interior and exterior car parts such as bumpers, seat structures, door module etc. This research is inspired by the desire to manufacture load carrying parts for vehicles such as wheel rims which would dramatically reduce vehicle weight and subsequently save fuel. This, however, requires a much better understanding of long fiber orientation and distribution during compression molding.

Current orientation models were developed for short fibers(several millimetres). Recently the models are being extended even for the LFT-D fibers which can reach up to 80 mm. Since several of the governing assumptions for short fiber models are not suitable for long fibers, the models can not provide accurate results for long fibers. Due to this limitation long fibers require independent treatments.

This thesis presents a new model which is specifically designed for long flexible fibers. This model is confirmed by comparing results obtained for simple shear flow to results found in the literature. The model was implemented in a rheometric squeeze flow, which is defined as flow between two approaching to each other parallel plates, and provided results previously not seen in the literature. Interactions were implemented into the model and tested for rheometric squeeze flow and simple shear flow cases. In addition to providing insight into fiber orientation and deformation in rheometric squeeze flow, which was not previously studied in the literature, the proposed model shows more predictive results than previously found in the literature.