Computational modelling, neural networks and experiments for biomass drying process prediction: A review

• Review of lignocellulosic biomass drying using models, CFD, experiments and ANN. • Systematic screening and classification of 170 predictive biomass drying studies. • Comparative analysis of mechanistic, CFD and ANN approaches for drying kinetics. • Identification of research gaps in biomass characterisation and model validation. • Proposal of integrated CFD–experiment–ANN frameworks for industrial dryer design. Biomass drying is a critical step in bioenergy and bioproduct chains, yet predicting drying behaviour for diverse lignocellulosic feedstocks and dryer configurations remains challenging. This review examines how mechanistic models, computational fluid dynamics (CFD) simulations, experimental measurements and artificial neural networks (ANNs) have been used over the past 15 years to predict biomass drying kinetics and process performance. A systematic screening of 335 articles yielded 170 studies that met predefined inclusion criteria regarding biomass type, dryer configuration, modelling approach and availability of quantitative performance indicators. These were grouped into theoretical developments, experimental investigations, CFD-based models and ANN/ANFIS applications. The review analyses how drying rates, effective moisture diffusivity and related transport coefficients are determined, and how these parameters are used in physics-based and data-driven models for convective, fluidised-bed, rotary and solar-assisted dryers. Reported predictive accuracies are typically high (e.g., R 2 values above 0.95 for the best-performing models), but depend strongly on biomass characterisation, experimental design and model structure. Particular attention is given to ANN and ANFIS predictors, discussing their data requirements, training strategies and ability to complement or surpass conventional semi-empirical models in terms of prediction error and flexibility. The review identifies persistent gaps related to diffusive dryers, standardised characterisation of lignocellulosic biomass, and the external validation of predictive models. On this basis, it proposes integrated CFD–experiment–ANN frameworks and outlines design-oriented research directions for industrial biomass dryers, supporting more reliable, energy-efficient and low-emission drying systems for bioenergy applications.