Architectural Image Modeling as a Predictive Tool for Civil Engineering Stress and Load Distribution Analysis

Abstract

Accurate prediction of structural stress and moment responses is essential for ensuring the safety, stability, and reliability of civil engineering systems under varying loading conditions. Conventional simulation-based approaches, such as Finite Element Analysis (FEA), provide reliable structural assessment; however, they often require high computational resources and extended processing time. This study proposes a machine learning-based surrogate modeling framework for predicting structural stress and moment responses using numerical simulation datasets. The research utilized three merged simulation datasets containing structural geometry, material properties, and loading parameters, including eccentricity, axial load, soil modulus, and concrete modulus. Several machine learning techniques were employed to model the relationships between structural parameters and response variables. Correlation analysis revealed that eccentricity and axial load were the most influential factors affecting tensile and compressive moment behavior. The developed predictive framework demonstrated strong capability in estimating structural responses while significantly reducing computational complexity compared with conventional simulation approaches. The findings indicate that machine learning models can effectively support rapid structural assessment, predictive analytics, and intelligent infrastructure applications. Furthermore, the proposed framework contributes to the advancement of AI-assisted structural engineering by providing a scalable and computationally efficient alternative for stress and load distribution prediction in civil engineering systems.