Industrial design of Short Fiber Reinforced Composites (SFRC) structures is subject to several compounding and processing steps of optimization. Moreover, these structures are often submitted to fatigue loading. Therefore, SN curves have to be established for each new composite formulation and for several type of microstructure involved in the real component due to processing. While these preliminary characterizations are time and money consuming, this paper propose a new hybrid methodology for fast fatigue life prediction. Moreover, both monotonic and fatigue behavior of SMC composites is essentially determined by local damage propagation. Therefore, the key idea of the proposed approach is to use a Mori and Tanaka based micromechanical model in order to establish an equation of state relating local damage rate to macroscopic residual stiffness rate. The generalization of this relation to fatigue damage multi-scale description leads to the SN curve fast determination of each considered microstructure. Very limited experimental characterization is required in such a way that SN curve could be established in just one day. Comparison between experimental and simulated Whöler curves highlights a very good agreement for several microstructure configurations in the case of a SMC composite material.

The majority of fatigue life prediction models which have been proposed for the Short Fiber Reinforced Composite (SFRC) materials have been developed for constant temperature. However, in real situations, SFRC structures are subjected to variable temperature. This study focus on the response of SFRC composites subjected to different sequences (or blocks) under variable temperature conditions. Experientially, this kind of study requires a lot of investment from the point of view of cost and time. In this paper, the results coming from modelling the fatigue life and residual stiffness of short fiber reinforced composites subjected to thermomechanical loadings are reported. In fact, we propose to use a hybrid micromechanical-phenomenological model to quantify the evolution of the local damage rate during each loading block. Indeed, damage accumulation is calculated and cumulated step by step through the calculation of the evolution of a local damage ratio which describes the evolution of micro-cracks density until failure. Life prediction for specimens submitted to a variable temperature loading found to give acceptable results compared to experiments.

Because of the high variability of SMC microstructure due to material flow during thermoforming, fatigue life prediction in real automotive structure represents a huge challenge. In this paper, we present a two-step microstructure selection involving an original ultrasonic method which is briefly presented. Then, on the basis of four selected microstructure configurations, an accurate experimental damage analysis is performed including both monotonic and cyclic loading. The high microstructure dependence of the obtained Whöler curves is demonstrated. Moreover, an experimental link between monotonic damage and fatigue life is emphasized. Then, a new fatigue life prediction methodology based on the later is proposed. This methodology also uses a micromechanical damage model in which a local damage criterion is involved for monotonic loading damage prediction. A very good agreement between experimental and predicted Whöler curves is demonstrated for all studied microstructures and three working temperatures. Finally, the model allows building a microstructure dependent Whöler curve abacus which may be very useful for SMC structures design.

To reinforce the environmental standards, we need to strengthen the lightening of vehicles and to generalize new composite materials in order to reduce weight. To use these innovative composite materials in the mass production of automotive parts, it is essential to propose a predictive approach of the S-N curves, which must be established for each new composite formulation and for several types of microstructure within real components. Although these preliminary characterizations consume time and money, this paper proposes two hybrid methodologies to predict the fatigue life during the fatigue test. Both methodologies are based on micromechanical modeling which is developed under monotonous loading with fatigue effects under different amplitudes. The suggested methodology is based on an experimental analysis of monotonic behavior under fatigue loading and on multi-scale modeling of damage. In the results, the proposed model and the used approaches are in good agreement with the experimental results.

No abstract