Orthotropic failure criteria based on machine learning and micro-mechanical matrix adapting coefficient

چکیده مقاله

Several studies have examined the use of fibrous and non-fibrous anisotropic materials in various industries and proposed failure criteria for consideration. These criteria can be categorized into theories that address specific failure modes and those that do not. However, these theories often fail to address both the representation of inherent material properties and the combination of different loading modes. To address this, a micromechanical approach called micromechanical modeling of laminates (“MACML”) has been developed, which gradually transitions from anisotropic to isotropic materials by incorporating reinforcement and adapting coefficients. MACML combines machine learning (ML) techniques and introduces a new method to identify safe and failure-prone regions in composites. This approach allows for independence from the influence of fibers by representing them as an active stress proportion of the isotropic matrix. Additionally, standard experimental procedures can be applied based on the isotropic matrix, and experimental results have been conducted to validate the established failure criteria in the MACML method