Evolution of Drying Process Modeling: From Mathematical Foundations to Machine Learning‐Driven Prediction and Control

Source abstract

ABSTRACT Modeling the drying process is essential for optimizing efficiency, reducing energy consumption, and improving product quality in food and agricultural systems. Over time, modeling approaches have progressed from empirical and semi‐empirical correlations to theoretical and diffusion‐based models, which offer insight into transport mechanisms but are often limited in adaptability and real‐time prediction. Computational and multiphysics frameworks have enhanced accuracy by coupling heat, mass, and momentum transfer phenomena; however, their high computational cost restricts widespread industrial use. The advent of machine learning (ML) has transformed drying analysis by enabling data‐driven, adaptive, and real‐time modeling. Techniques such as artificial neural networks (ANN), support vector machines (SVM), random forest (RF), adaptive neuro‐fuzzy inference systems (ANFIS), and hybrid models provide superior flexibility in handling complex, nonlinear, and multivariate drying data. This review presents a comprehensive overview of the evolution of drying process modeling—spanning empirical, theoretical, computational, and ML‐based approaches—across four critical domains: (i) drying kinetics and moisture prediction, (ii) product quality evaluation, (iii) energy and exergy analysis, and (iv) process control and automation. The synthesis highlights the transition toward intelligent, data‐driven drying systems with improved prediction accuracy, optimized energy use, and enhanced product quality.