AI-Driven Optimization of Reinforced Concrete Structures Considering Embodied Carbon Using Physics-Informed Machine Learning

AI-Driven Optimization of Reinforced Concrete Structures Considering Embodied Carbon Using Physics-Informed Machine Learning

Author:

Kypros Pilakoutas

Faculty of Engineering The University of Sheffield Sheffield, United KingdomPhD, University of London, Imperial College of Science, Technology and Medicine (1990)

Zholaman Shashpan

L.N. Gumilyov Eurasian National University

Astana, Kazakhstan

Corresponding author:

Kypros Pilakoutas

k.pilakoutas@sheffield.ac.uk

Introduction

The construction sector accounts for approximately 40% of global CO₂ emissions. Reinforced concrete structures significantly contribute due to cement and steel production. Traditional design methods separate structural analysis and environmental assessment, limiting optimization potential. AI enables simultaneous optimization of both aspects.

Methodology

Architecture of the Proposed Framework

The framework consists of three main modules:

The interaction between these modules allows the exploration of a large design space while ensuring compliance with structural safety constraints.

Figure 1 — Architecture of the proposed AI-based structural optimization framework

2.2 Optimization Objective

The objective function is defined as:

where

C — embodied carbon of the structure

R_eq — equilibrium residuals derived from structural mechanics equations

R_LS — penalty term associated with limit-state violations.

2.3 Optimization Algorithm

Figure 2 — Structural optimization algorithm

The optimization procedure involves the following steps:

3. Numerical Experiments

The proposed framework was tested on three typical reinforced concrete structural elements.

3.1 Reinforced Concrete Beam

Figure 3 — Reinforced concrete beam reinforcement layout

Parameters:

Span: 6 m

Uniform load: 20 kN/m

Optimization resulted in a 12% reduction in reinforcement mass, leading to a significant reduction in embodied carbon

3.2 Reinforced Concrete Column

Figure 4 — Reinforced concrete column reinforcement

Parameters:

Height: 3 m

Axial load: 1500 kN

Embodied carbon reduction: ≈19%

3.3 Reinforced Concrete Slab

Parameters:

Span: 5 m

Thickness: 200 mm

Embodied carbon reduction: ≈21%

Figure 5 — Reinforced concrete slab reinforcement layout

4. Statistical Analysis of Results

Figure 5 — Reinforced concrete slab reinforcement layout

5. Comparison with Existing Optimization Methods

6. Discussion

The results demonstrate that artificial intelligence can significantly improve the environmental performance of reinforced concrete structures.

However, several challenges remain:

Future research should focus on combining the proposed framework with digital twin technologies and BIM-based design platforms.

7. Conclusion

This study presents a physics-informed machine learning approach for optimizing reinforced concrete structures with respect to embodied carbon.

The proposed methodology:

The proposed approach can serve as the foundation for next-generation AI-based design systems in sustainable construction.

List of references