Explainable artificial intelligence-based compressive strength optimization and Life-Cycle Assessment of eco-friendly...

Area of Science:

• Materials Science
• Civil Engineering
• Sustainable Construction

Background:

• Growing demand for sustainable building materials.
• Need for eco-friendly alternatives to traditional cement.
• Utilizing industrial byproducts like sugarcane bagasse ash (SCBA) in construction.

Purpose of the Study:

• To investigate the potential of SCBA as a supplementary cementitious material (SCM) in concrete.
• To develop interpretable models for predicting the compressive strength of SCBA concrete using explainable AI (XAI).
• To optimize SCBA concrete mixes and assess their environmental impact through Life-Cycle Assessment (LCA).

Main Methods:

• Dataset compilation of 2616 SCBA concrete data points.
• Application of black box machine learning models for compressive strength prediction.
• Utilizing model-agnostic XAI for local and global interpretation of influencing parameters.
• Development of a nonlinear model equation based on feature influence.
• Conducting Life-Cycle Assessment (LCA).

Main Results:

• Achieved a compressive strength of 40 MPa with SCBA concrete.
• Optimized water-to-binder content (0.42-0.48) with reduced cement content.
• Identified key parameters influencing compressive strength through XAI.
• Developed a nonlinear model equation for SCBA concrete strength prediction.

Conclusions:

• SCBA concrete demonstrates potential as a viable, eco-friendly building material.
• XAI provides valuable insights into the behavior of SCBA concrete.
• Optimized SCBA concrete mixes can achieve high compressive strength with reduced environmental impact.