Multi-Modal Sensor Fusion for Fatigue Life Prediction in Additively Manufactured Ti-6Al-4V Alloys Under Variable Amplitude Loading

**Abstract:** This paper introduces a novel methodology for fatigue life prediction in additively manufactured (AM) Ti-6Al-4V alloys exposed to variable amplitude loading. The approach combines acoustic emission (AE) data, strain gauge measurements, and high-resolution optical microscopy images to develop a comprehensive multi-modal prediction model. Utilizing a hyper-dimensional vector representation of the fused sensor data, combined with a Bayesian neural network trained via reinforcement learning, enables significantly improved fatigue life predictions compared to traditional methods. The framework targets immediate commercialization within the aerospace and biomedical industries, providing a more reliable and efficient decision-making tool for component design and lifespan assessment.