Moment-Equivalent Adhesion Statistics Inferred From Particle Resuspension Kinetics

Particle resuspension from surfaces is a fundamental process in many chemical engineering systems, including aerosol dispersion, powder processing, filtration, contamination control, and environmental dispersion. Accurate prediction of resuspension requires knowledge of particle-surface adhesion statistics, yet direct adhesion measurements are often technically demanding and strongly system-dependent. Here, a data-driven framework based on an augmented Rock'n'Roll (RnR) model is developed to reconstruct adhesion-force statistics from resuspension measurements. The present approach integrates a flexible bounded beta distribution into the RnR framework, allowing adhesion heterogeneity to be reconstructed without restricting the distribution to a lognormal form. It is shown here that this enables a robust, percentile-resolved reconstruction of adhesion statistics that remain mechanically consistent with rolling detachment physics and aerodynamic forcing. Implementing this approach across multiple datasets and particle systems, the reconstructed adhesion percentiles were found to exhibit a systematic aerodynamic scaling when normalized by the rolling lever-arm ratio. This reflects the linear contribution of rolling in the moment balance and serves as an internal consistency check of the inverse reconstruction. These results indicate that resuspension percentiles encode quantitative information about particle-surface adhesion interactions. By isolating the mechanically relevant resistance to rolling detachment, the proposed method infers adhesion statistics without direct adhesion experiments.