Applications of Dynamic Modeling in Crushing Plants

Doctoral thesis, 2021

Modeling is a tool to describe phenomena in a simplified way, and the models can then be used to simulate these phenomena. Models of equipment used in the mining and aggregate industries can be used for process simulations of the processes in those industries to improve the operations. To study processes and the operation of processes, time dynamic models are a great tool. This thesis focuses on applications of time dynamic modeling in crushing plants. The time dynamic models predict the output of the equipment as a function of time. The work presented within this thesis focuses on three areas; Unit modeling, process modeling, and control modeling.

Unit modeling refers to developing models of single processing units, which could be a comminution unit, classification unit, or materials handling unit. The new models presented in this thesis are for jaw crushers, high pressure grinding rolls (HPGR), and storage units (e.g., bin, silo, or stockpile). The developed models are based on the fundamental insight of the physics that happens within the unit. The validity of the models is aimed to be broad and cover many operating points and uses. The models are intended for high fidelity process simulation applications.

Process modeling refers to the modeling of many interconnected units, and the modeling presented in this thesis has been done with both high-fidelity unit models and with simplified models. Both high fidelity and simple simulations are demonstrated within the thesis. The simpler models are used to try new concepts of plant design or control and study plant robustness or ability to handle variations. Meanwhile, the high-fidelity models can be used to study topics such as particle size distribution, debottlenecking and specific control issues.

Control modeling refers to developing controller models to control plants like those modeled within the process modeling section. Optimal control, such as model predictive control (MPC), relies on models to steer processes optimally relative to some objective. The models within those controllers have been discussed in this thesis. Additionally, being able to move between the various fidelity domains of models is beneficial for this application.

In this thesis, multiple new models and methods are presented, along with how they can be applied within the minerals processing and aggregate industry, ultimately improving the efficiency and performance of the industries.