Digital Twin HPGR Software Sensor Proof Of Concept

Description

The project consisted of the generation of a mixed model based on a phenomenological model and a particle filter for a HPGR (High Pressure Gridding Roll) that allows estimating the throughput and the ore re-circulation rate in the circuit in which the equipment is integrated. The Bayesian scheme used allowed a fast convergence of the model and a correct estimation of the response variables, which were treated as latent variables in the model.


Principal Researchers: Alvaro Egaña, Gonzalo Díaz, Alejandro Ehrenfeld. Participating Researchers: Felipe Navarro.

Fund/Partner: WEIR