Predictive framework of fatigue crack growth in metal matrix composites with mixed reinforcements

Lightweight, corrosion-resistant and electrically conductive aluminum and its composites are perfect for electronics, automotive, marine and aerospace applications. Pure aluminum, however lacks the fatigue resistance and strength required for structural applications. To solve these problems, Aluminum hollow glass microsphere (HGM) Metal matrix composites are developed, which combine the advantages of aluminum with improved mechanical properties from reinforcements. These syntactic foams improve stiffness, fatigue/corrosion resistance and strength to weight ratios, which makes them appropriate for harsh conditions. This research introduces a new deep learning framework that uses a modified mixed attention mechanism with deep bidirectional long-short term memory (M2AM-Deep BiLSTM) network to forecast the propagation of fatigue cracks in aluminum-based Metal Matrix Composites, especially aluminum HGM and aluminum silicon carbide. To increase prediction accuracy, the model uses hybrid optimization using V-shaped and S-shaped binary divine religious algorithms (bDRA). The proposed model should achieve error less than 5% with the best value than other methods. The performance of the proposed model is executed on a Python platform and is compared to established approaches. The results demonstrate a high predictive capability when compared to traditional methods, enabling better experimental planning and material selection.