Multi-Scale AI Modeling And Decision Support Systems For Mechanical Behavior In Bio-Based Building Composites: An Environmental And Computational Perspective

Abstract

There has been an increasing interest in constructing in a more sustainable way and this is why the bio-based building composites are getting a lot of use but they have complicated mechanical behavior which bring problems in terms of modeling when subjected to varied conditions. In this paper the multi-scale mechanical performance of five bio-composite materials- Hemp-Lime, Bamboo-Cement, Bagasse Ash Concrete, Flax Epoxy and Cornstarch-Biofoam are studied using an holistic approach based on a combination of finite element models, environmental model and artificial intelligence (AI)-based predictive modeling. The data on the simulation of micro, meso, and macro levels were inserted into the supervised machine learning models to predict the tensile strength and elasticity using the factors of material composition, fiber architecture, and climatic activities. These models reported excellent predictive capability (R 2 = 0.93), and fiber volume fraction and type of matrix were found to be critical variables of mechanical properties using SHAP analysis. A decision support system (DSS) in the form of a web application has been created to convert model results to actionable guidance to develop sustainable design, or to make in-the-field material selections in response to environmental conditions. Moisture-sensitive composites such as hemp-lime and cornstarch were demonstrated to be highly susceptible to environmental variability and especially humidity which act as strong deterrents to their strength. The combination of AI and multi-scale modelling in the study lends itself to scalability, interpretability, and computation efficiency in evaluating, and optimising bio-composites in smart city infrastructure. The findings present meaningful implications to the engineers and material scientists and policymakers who seek resilient and low-carbon approaches to buildings.