Predicting Static Stress in One-Dimensional Bar Elements Using Artificial Neural Networks coupled with DAI5 Framework-by-Tanveer Khan

Hello, Welcome to my second page. Stay connected with this page this is all about my final project. Thank you

Introduction

Artificial Neural Networks (ANNs) have emerged as powerful tools for structural stress analysis, offering efficient alternatives to traditional methods like finite element analysis (FEA). These models leverage machine learning techniques to predict stress distributions and responses under various loading conditions, significantly reducing computational costs and time, Machine learning (ML) has shown significant promise in the field of structural stress analysis, offering innovative solutions to traditional challenges, ML techniques have been employed to expedite the multiscale analysis of composite materials. By learning the relationships between different scales, ML models can significantly reduce computation times while maintaining high accuracy in stress predictions[1]. Artificial Neural Networks (ANNs) have been applied to improve stress recovery in FEA. Traditional FEA often provides accurate stress values only at specific points (super-convergent points. ANNs can map the stress field across the entire domain, enhancing accuracy and efficiency[2]. ML algorithms have been used to estimate axial stress in CWR by analyzing vibration data. By training ML models with data from finite element models, researchers can predict stress from vibration frequencies, demonstrating ML's viability in real-world applications[3]. Various ML models, including Gaussian Process Regression (GPR), Neural Networks (NN), and Boosted Trees (BST), have been used to predict stress-strain curves for materials like aluminum alloys. These models have shown high accuracy, with NNs achieving a coefficient of determination (R²) of 0.998[4]. ML models can provide highly accurate predictions, often surpassing traditional empirical models. For example, gradient boosting models have shown superior prediction accuracy for flow stress in high entropy alloys[5]. An Artificial Neural Network (ANN) is utilized to establish the relationship between selected feature elements and correlation elements for structural stress analysis. This approach enhances accuracy in measuring global stress in marine structures, achieving 93.6% accuracy in real-scale model tests[6].