Frequency domain analysis for identifying dominant segregation units in a chain of material handling processes: A cellular automaton framework

Degradation and material segregation are persistent challenges in bulk solids handling. During storage and transport through silos, domes, ship holds, and hoppers, mechanical stress generates fines and dust, exacerbating segregation and reducing process efficiency. This can cause fines spikes during silo discharge, increasing the risk of fire and dust explosions, raising energy consumption, and elevating maintenance needs. Moreover, variability in fines content in material stream negatively impacts downstream processes such as milling, combustion, and emissions control. To address the challenge of modelling segregation in multi-stage wood pellet handling systems, a Cellular Automaton (CA) model was previously developed to simulate segregation at individual transfer points. However, full-chain simulations remain computationally intensive. This study introduces a novel Frequency Domain Analysis (FDA) method to identify the most influential segregation stages, allowing simplification of the modelling scope. In this framework, each storage unit is conceptualised as having a “forcing function” that introduces its own segregation pattern, and a “damping function” that attenuates upstream effects. Applied to large-scale systems, FDA enables quantitative assessment of how each unit contributes to downstream segregation. The analysis revealed that inflow signals with frequencies below a corner frequency of 0.5 pass through the system and induce segregation, while higher frequencies are increasingly attenuated. This approach supports the development of targeted, efficient handling strategies by isolating critical stages for detailed simulation. The proposed approach enables focused mitigation strategies, directing attention to handling stages with the greatest influence by maximising both process efficiency and operational safety. • A novel Frequency Domain Analysis (FDA) approach identifies dominant segregation units in pellets handling chains, based on validated Cellular Automata (CA) models. • FDA analysis indicates that segregation in smaller silos (2,500 tonnes) is negligible when materials are transferred to larger storage units (90,000 tonnes). • The FDA method reduces simulation complexity by eliminating the need for full-chain modelling, providing a novel computationally efficient alternative for predicting segregation.