BOxCrete

Interactive Concrete Strength Explorer
Cite

About This Explorer

This tool predicts concrete compressive strength using a Gaussian Process model trained on laboratory mortar and concrete specimens. Predictions reflect controlled lab conditions and may differ from field performance due to batching variability, curing environment, and aggregate properties.

How to Use

Uncertainty

Shaded bands represent ±2σ model uncertainty. Wider bands indicate regions with less training data — predictions there are less reliable.

Papers

Baten, Iqbal, Ament, Kusuma, & Garg (2026) — BOxCrete: A Bayesian Optimization Open-Source AI Model for Concrete Strength Forecasting and Mix Optimization.
Ament, Witte, Garg, & Kusuma (2023) — Sustainable Concrete via Bayesian Optimization.

Composition (kg/m³)

Adjust sliders or click a point in the scatter plot to explore predicted strength curves with uncertainty.

Tap a point to explore its properties. Tap X/Y toggles to switch axes.

Performance Tradeoffs

Click a point to explore its properties. Click X/Y toggles to switch axes.

X: GWP Y: 28-day strength Filter

Mix Insight

Pareto-optimal

Click a data point to see mix analysis.

Ingredient Insight

Click an ingredient name in the Composition panel to learn more.

⚠ Far from data

Predicted Strength Curve

GWP: – kg CO₂/m³
Cost: – $/m³
W/B: –

References

BOxCrete: A Bayesian Optimization Open-Source AI Model for Concrete Strength Forecasting and Mix Optimization

Baten, Iqbal, Ament, Kusuma, Garg (2026)

Extends the framework to comprehensive concrete mix analysis across different curing temperatures and material sources. The resulting model was used to design a low-carbon mix deployed at a Meta data center.

arXiv
Cite
Sustainable Concrete via Bayesian Optimization

Ament, Witte, Garg, Kusuma (2023)

Introduces the custom Gaussian process model and multi-objective Bayesian optimization formulation for concrete mix design, with initial results demonstrating efficient discovery of low-carbon, high-strength mortar formulations.

arXiv
Cite