PSI - Issue 19

Marc J.W. Kanters et al. / Procedia Structural Integrity 19 (2019) 698–710 Marc Kanters et al. / Structural Integrity Procedia 00 (2019) 000–000

699

2

1. Introduction Plastics are increasingly used in a large variety of applications. With typical benefits, like freedom in design, ease of processing, reduction in cost, and high specific strength and stiffness the material can offer weight reduction via metal to plastic conversion. Composites of continuous or short fibers embedded in a thermoplastic matrix are the material of choice for metal replacement [Weiss2017]. In automotive, the continuous strive towards lighter vehicles [Winter2016, Winter2017], to reduce emissions and compensate for a weight introduced by the increasing number of sensors and safety features in a car, renders the conversion of metal covers and housings insufficient to meet the needs. As a result, the upcoming trend is to design load bearing structural parts in reinforced plastics [Weiss2017, Elringklinger2017]. Predictability is key when designing load bearing components, to be able to reduce development time, design first time right, and ensure part performance in service. This involves modelling of all relevant properties such as part stiffness, strength, durability, impact, noise, vibration and harshness (NVH), creep deformation, and chemical- and thermal stability. This work focusses on durability, i.e. fatigue lifetime, predictions of injection moulded short glass fiber reinforced plastics (SFRP). The first challenge in accurately predicting lifetime is to properly understand and model the mechanisms that lead to failure. In general, the long-term performance of thermoplastics is limited by three failure mechanisms: I) plasticity controlled failure, related to accumulation of plastic strain, II) slow crack growth, controlled by crack propagation, and III) molecular degradation [Gedde1994, Lang1997]. All three have a different origin and are affected differently by loading conditions, such as load magnitude, load amplitude or load ratio, and frequency [Janssen2008, Kanters2016, Kanters2018]. Here the focus is on the first two failure mechanisms. The second challenge is that, due to processing, the local glass fiber orientation makes the material behavior of these types of materials highly anisotropic. Depending on the local arrangement of fibers, variations in stiffness and strength of a factor of 1.5 to 2 are no exception [DeMonte2010]. Therefore, to successfully model durability, one needs to be able to accurately capture the local glass fiber orientation after processing and incorporate the resulting anisotropic material behavior in the stress calculation. In combination with the effect of fiber orientation on the failure mechanisms one can make a basic lifetime estimation. Unfortunately, predicting part performance based on input data from standard test specimens will show that capturing the anisotropy is not enough. To improve accuracy, one needs for example to use local stresses for the coupon data [Sonsino2008], compensate for local stress concentrations using a fatigue notch factor [Mortazavian2016, Primetzhofer2019], and bear in mind that the local stress amplitude is not necessarily the same as the global force amplitude applied. This combined complexity limits the accuracy of the durability predictions of glass fiber reinforced plastics, providing little comfort to the design engineer when designing load-bearing parts. Therefore, to improve the accuracy of fatigue life predictions, a step-by-step approach is taken where the complexity in loading conditions, anisotropy, and stress state is increased systematically. This allows validation of methods, identification of potential gaps, and provides direct leads for improvement. This work presents the practical framework that resulted from this exercise. The framework combines engineering tools that allow design engineers to predict fatigue life of engineering plastics applications, including anisotropy, in high detail. While step-by-step highlighting the key features and assumptions of the framework, the consistency and accuracy that one can expect is displayed. Subsequently the framework is validated on a representative demonstration part.