Geometry‑Harmonized Fly‑Ash based Alkali‑Activated Concrete Dataset for Data‑Driven Modelling

Overview

This repository contains a harmonized compilation of about 800 fly‑ash alkali‑activated concrete (FA‑AAC) mix designs extracted from 67 peer-reviewed publications & the open-access parent dataset created by Torres et al. (1 630 mixes, ref. [70-71] in the bibliography).

• Single precursor & activator: Only fly ash as binder precursor & NaOH(aq) as activator.
• Geometry-normalized strengths: All compressive strengths have been converted to the equivalent 150 × 300 mm cylinder value using ACI 318‑19 & CEB‑FIP factors; raw values are preserved for traceability.
• Rich feature set: Binder oxide chemistry, activator properties, mix proportions, fresh & hardened properties, curing regime, & CO₂ footprint, besides a statistical overview.
• Bibliographic traceability: Complete bibliography of the 67 papers included.
• Deliverables: Tidy CSV & README—ready for ML modeling, meta‑analysis, & low‑carbon mix benchmarking.

Value of the Data

• Provides the first openly shared, geometry-harmonized collection of FA-AAC mixes, eliminating specimen-geometry bias via size normalization.
• Serves researchers developing machine‑learning or optimization models by pairing full mix parameters with cylinder‑equivalent compressive strengths. Supply a ready-to-model table of mix designs & properties.
• Offers industry professionals an objective benchmark for screening low‑carbon, FA‑only AAC formulations.
• Fulfills FAIR principles by supplying a tidy CSV file & a complete bibliography—facilitating reproducibility, reuse, & future extensions.

Data description (#Features)

• Statistical overview (#7): Number of data, minimum & maximum value, 10^th & 90^th percentile, average value, & standard deviation.
• Identification features (3 features): number of the mixture, references, & the mixture code.
• Fly ash properties (#18): the dosage, specific density, & the chemical composition including the sum of the pozzolanic oxides & the Si/Al ratio. 
• Aggregate properties (#7): Content & density of fine & coarse aggregates, the total aggregate, the fine-to-total-aggregate ratio, & the binder-to-total-aggregate ratio.
• Alkali-activator content & concentration (#18): Properties of Na₂SiO₃ & NaOH including content, composition, concentration & specific density. Besides, some significant ratios, such as Na₂O/SiO₂ ratio, liquid/solid ratio, activator-to-binder ratio, etc.
• Workability features (#8): including those related to mix design such as water, superplasticizer, water-to-binder ratio; slump flow, flow-table, & setting time.
• Curing features (#4): including initial & final curing time, initial & final curing temperature. 
• Hardened properties (#39): This includes the sample dimensions, the original mechanical strengths for different cube & cylinder sizes, then the converted strengths to the equivalent 150 × 300 mm cylinder, the young modulus & porosity.
• CO₂ footprint feature: in kg emission per 1m³ of samples produced & cured based on ref. [68-69] in the bibliography.

Limitations

• Literature coverage stops at December 2023; newer FA‑AAC studies are not yet included.
• Size-effect factors are derived from OPC concrete; validation for very high‑Ca or ultra‑high‑strength mixes is advised.
• Non‑strength properties (workability, durability) appear only when reported by the original authors, resulting in sparse coverage.

Ethics & competing interests  

The authors declare no competing financial interests.