Online prediction of pressing iron ore concentrates in an industrial HPGR. Part 1: Modeling approach

Abstract

High-pressure grinding rolls reached great popularity for pressing iron ore concentrates since its first application in the 1990s. For this particular application, mathematical models describing HPGR performance on the basis of operating conditions and feed characteristics have successfully been used by the authors to map industrial-scale operations under controlled conditions. Despite these important advances, this modeling approach has only been used so far offline and under steady-state conditions. The present work applies the Modified Torres and Casali model proposed by the authors as a novel online model coupled with real-time information from an industrial-scale HPGR pressing iron ore concentrates. The model is demonstrated to be able to predict throughput, power and product size distribution on the basis of data available online from the process. Results showed capability of the model to map the operation giving a realistic description of the process. A new method was proposed to circumvent the model limitations when describing HPGR operating with worn rolls, and bench-scale data was used to improve description of the size reduction in industrial-scale.