Data-Smart Hydrometallurgy: Active Learning, Bayesian Optimization, and MLOps for Heapleaching

Data-Smart Hydrometallurgy is the first book to unite hydrometallurgical science, surrogate modeling, and modern machine-learning workflows into a single, coherent framework. Designed for metallurgists, process engineers, data scientists, and graduate researchers, this book (1 &2) provides a practical and powerful roadmap for predicting, optimizing, and operating complex leaching systems—from bottle rolls to full-scale heaps.With increasing ore variability, tighter reagent budgets, and rising environmental constraints, traditional empirical models are no longer enough. This book introduces a new generation of tools: Gaussian processes for uncertainty-aware prediction; multi-fidelity models that connect lab, column, and heap scales; physics-informed Neural ODEs; stochastic kinetics using SDEs; RTD-based heap hydrodynamics; cellular-automaton front propagation models; and real-world Bayesian optimization strategies for acid allocation, irrigation scheduling, and recovery maximization.Each chapter blends metallurgical fundamentals with modern computation, showing how to embed Arrhenius kinetics, monotonicity constraints, diffusion limits, transport physics, and operational boundaries directly into learning systems. The result is a robust surrogate modeling workflow capable of safe extrapolation, risk-aware forecasting, and data-driven experimentation.Complete with hands-on Python notebooks, animated visualizations, validation workflows, and MLOps deployment templates, this book equips professionals with everything required to design smarter tests, build more reliable models, and confidently navigate the uncertainty inherent in Heap leaching.Whether modernizing R&D pipelines, guiding pilot-heap design, or implementing predictive dashboards for plant operations, Data-Smart Hydrometallurgy gives you the computational edge the metallurgical industry now demands.Francis Dakubo delivers a groundbreaking multidisciplinary text—one that will shape the future of metallurgical modeling, digital twins, and data-driven leaching optimization." Read more