Discrete Element Simulations and Machine Learning for Improving the Performance of Dry Catalyst Continuous Impregnation Processes

TL;DR

This paper combines discrete element simulations with machine learning to analyze and optimize dry catalyst impregnation processes, identifying particle bed regimes and predicting process variability.

Contribution

It introduces a novel integrated approach using DEM and machine learning to understand particle behavior and improve process control in dry impregnation.

Findings

Two distinct particle bed regimes identified via DEM

A general function for RSD as a function of process parameters

Uneven spraying reduces variability at low RPM and angles

Abstract

In this work, discrete element method (DEM) simulations coupled with machine learning are used to study the process of dry impregnation. Our results show that the particle bed contains two regimes. Regime 1 shows smaller inclination angles and a larger mass hold-up which implies more forces restricting the particle movement. Regime 2 reveals larger inclination angles and rotational speeds and a smaller mass holdup, which indicates a smaller bed height. Using Machine learning, we found a general function for the Relative Standard Deviation (RSD) as a function of time, angle of inclination and speed of rotation for both even and uneven flow rates for a full range of the parameters fed in the LASSO algorithm. Machine learning gives insight on both regimes and reveals that for low RPM and low angles, uneven spraying gives a lower RSD which is consistent with our observations of the DEM studies.