Abstract
Rock aggregates have been extensively exploited in construction, and the associated engineering features play a critical role in their application. The main aim of this research is to assess the impact of petrographic characteristics on the engineering properties of carbonate rocks. A total of 45 carbonate rock samples from different geological formations within the Salt Range (Western Himalayan Ranges, Pakistan) were subjected to comprehensive petrographic analyses and standard aggregate quality control tests. The engineering characteristics encompassed Los Angeles abrasion value, aggregate crushing value, aggregate impact value, specific gravity, water absorption, and unconfined compressive strength, whereas petrographic examination of thin sections quantified the mineralogical composition. Statistical methods and machine learning models have been applied to elucidate the relationships and establish potential predictive relationships between the petrographic and engineering features of the aggregates. The analysis identified clay, calcite, feldspar, and dolomite as the primary determinants influencing the engineering behavior of carbonate aggregates. Although multiple regression analyses produced R² values exceeding 0.84, the multiple regression equations did not provide substantial insights into the impact of all petrographic parameters on engineering properties. To enhance predictive accuracy, advanced machine learning models, including Random Forest, Gradient Boosting, Multi-Layer Perceptron, and Categorical Boosting, were applied. Among these, the Gradient Boosting model demonstrated superior predictive capability, overcoming both traditional regression methods and other machine learning algorithms validated through Taylor diagram and ranking system which included (correlation coefficient (r )=0.998, R²=997, Root mean square error=0.075, Variance Accounted For (VAF)%=99.50, Mean Absolute Percentage Error (MAPE) %= 0.385, Alpha 20 Index (α20-index)= 100, and performance index (PI))= 0.975. These results highlight the ability of machine learning techniques in providing a more effective and reliable prediction of aggregate engineering properties based on petrographic data. This approach offers significant advantages in the preliminary assessment of aggregate suitability, contributing to more efficient resource allocation in construction projects.
Full Text
Preamble
Impact of Petrographic Characteristics on the Engineering Properties of Carbonate Aggregates: A Machine Learning Approach
Javid Hussain¹,²,³,⁴, Jian Chen¹,²,³,⁴*
¹State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan 430071, China
²University of Chinese Academy of Sciences, Beijing 100049, China
³China-Pakistan Joint Research Center on Earth Sciences, Islamabad 45320, Pakistan
⁴Hubei Key Laboratory of Geo-Environmental Engineering, Wuhan 430071, China
Corresponding author: chenjiaan@whrsm.ac.cn
ABSTRACT
Rock aggregates are extensively used in construction, where their engineering properties critically influence application performance. This research investigates the impact of petrographic characteristics on the engineering properties of carbonate rocks. Forty-five carbonate rock samples from various geological formations within the Salt Range (Western Himalayan Ranges, Pakistan) underwent comprehensive petrographic analysis and standard aggregate quality control tests. Engineering characterization included Los Angeles abrasion value, aggregate crushing value, aggregate impact value, specific gravity, water absorption, and unconfined compressive strength, while petrographic examination of thin sections quantified mineralogical composition. Statistical methods and machine learning models were applied to elucidate relationships and establish predictive frameworks linking petrographic and engineering features.
The analysis identified clay, calcite, feldspar, and dolomite as the primary determinants influencing the engineering behavior of carbonate aggregates. Although multiple regression analyses produced R² values exceeding 0.84, these equations failed to provide substantial insights into how all petrographic parameters affect engineering properties.
To enhance predictive accuracy, advanced machine learning models—including Random Forest, Gradient Boosting, Multi-Layer Perceptron, and Categorical Boosting—were implemented. The Gradient Boosting model demonstrated superior performance, surpassing both traditional regression and other machine learning algorithms. Validation via Taylor diagram and ranking system yielded exceptional metrics: correlation coefficient (r) = 0.998, R² = 0.997, root mean square error = 0.075, Variance Accounted For (VAF) = 99.50%, Mean Absolute Percentage Error (MAPE) = 0.385%, Alpha-20 Index (α20-index) = 100, and performance index (PI) = 0.975. These results underscore the effectiveness of machine learning techniques for reliable prediction of aggregate engineering properties from petrographic data. This approach offers significant advantages for preliminary assessment of aggregate suitability, contributing to more efficient resource allocation in construction projects.
Keywords: Construction projects; engineering properties; Gradient Boosting; petrographic characteristics; statistical analyses; Salt Range