Semi-analytical modelling of a multi-shaft mill
Carlos Les (a),∗, James M. Finley (a), Boaz Friedland (b)
a) Element Digital Engineering, United Kingdom
b) Energy and Densification Systems (EDS), Johannesburg, South Africa
*This paper was presented at MEi Communution 23 Conference held 17-20 April 2023 in Cape Town, South Africa.
Abstract
Comminution is the process of breaking rock ores to extract valuable mineralsand materials. The machines used to process rocks often are complex and feature many parameters which are difficult and costly to optimize. Comminution process modelling is therefore an attractive means to reduce uncertainty in comminution performance, improve energy efficiency, optimize maintenance cycles,and ultimately to deliver cost savings to manufacturers and consumers. One way to model comminution processes is via the discrete element method (DEM), a technology which simulates the time-dependent motion of particles. However, as rocks break down into smaller pieces, the exponential growth in particle collisions can become computationally intractable. Using DEM to modelfull-sized comminution devices at scale is for that reason often unfeasible. To overcome such limitation, we have developed a reduced-order mathematical model that simulates comminution processes in just minutes. The model combines physics-based breakage and mass transport models along with surrogate model strained on impact data generated from DEM simulations. This paper describes themodel architecture and compares predictions of product particle size distributionswith those from experiments using a multi-shaft vertical mill as a case study.
Keywords: machine learning, reduced-order modelling, discrete element method,comminution, vertical mill
∗Corresponding author
Email addresses: This email address is being protected from spambots. You need JavaScript enabled to view it. (Carlos Les),This email address is being protected from spambots. You need JavaScript enabled to view it. (James M. Finley), This email address is being protected from spambots. You need JavaScript enabled to view it. (Boaz Friedland)
URL: www.element.com/digital-