Fast prediction of the Wöhler curve from heat build-up measurements on Short Fiber Reinforced Plastic

The fatigue design of short fiber reinforced plastic (SFRP) components for structural applications in the automotive industry requires an accurate knowledge of the numerous factors affecting the fatigue lifetime (process, loading type, environment, etc.). It is therefore crucial to reduce the characterization duration, which is even higher than for metallic materials as the test frequency has to be limited in order to reduce the influence of the temperature rise. The goal of this paper is to apply a heat build-up protocol on a SFRP used for structural automotive applications (PA66GF50, a polyamide filled with 50% of glass fiber in mass) and to predict the fatigue properties from the temperature measurements. A first approach based on the temperature measurements only is proposed, providing a good correlation to the fatigue data obtained classically. In order to provide a relation between the full heat build-up curve and the Wöhler curve, the dissipated energy is then evaluated from the thermal measurements and the parameters of an energetic fatigue criterion are identified from the results obtained from one single sample. The fatigue curve predicted from the heat build-up curve is compared to the one obtained from a classical fatigue campaign and the agreement is very good. The energy based criterion as well as the identification of the dissipation sources from the temperature measurements are finally challenged by an experimental campaign on smaller samples, conditioned at the same humidity ratio of 35% and tested for the same load ratio. The agreement obtained is there again very good and clearly calls for further investigation on the ability of the approach to predict the influence of other factors (humidity, load ratio, fibers orientation, etc.).

► We used temperature measurements to predict the fatigue properties of SFRP.
► The dissipated energy is identified from the temperature measurements.
► The parameters of an energy base fatigue criterion are identified from one sample within 2 days.
► The deterministic fatigue curve predicted exhibits a very good correlation to the fatigue tests performed on the other samples.
► The same approach is applied and validated on another sample geometry.