Central Asian Research Index

Science and Technology of Kazakhstan, 2026, № 3,2026, с. 435-445

PREDICTING OPTIMAL COPPER PRECIPITATION PARAMETERS IN THE CYANIDATION PROCESS FOR GOLD-COPPER ORES: FROM EXPERIMENTAL DESIGN TO ADABOOST MODELING

R. V. Sapinov, D. B. Kassymova

Аннотация · Abstract

This article discusses ways to improve the technology for precipitating copper from productive solutions obtained during the processing of refractory gold-copper ores in Kazakhstan. The presence of copper in such ores critically complicates the selective extraction of precious metals and leads to excessive reagent consumption. The authors conducted a comparative analysis of the effectiveness of classical mathematical approaches (V. P. Malyshev’s method) and predictive analytics tools based on the AdaBoost algorithm. The modeling was based on data from 50 experimental cycles, in which the pH, copper concentration, and dosage of the sulfidizing agent (Na2 S) were varied.It was established that ensemble learning demonstrates exceptional accuracy (R² = 0.9933), significantly minimizing prediction error compared to the traditional Protodyakonov equation. A key result of the study was the identification of a critical inversion point at pH ≈ 3.7, which determines the rationality of using Na2 S. The feature importance analysis conducted allows not only for the stabilization of gold extraction but also for significant cyanide savings through the precise real-time control of copper-removal parameters. Keywords: gold-copper ore, optimal parameters, cyanide leaching, mathematical approaches, AdaBoost algorithm, precipitating.

Сведения · Details