- Title
- Modelling of high-pressure grinding rolls using discrete element method
- Creator
- Hassanzadeh, Vahid
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2021
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- In this research work, an in-house DEM code was developed to improve the implementation of the particle breakage in DEM simulations and advance the simulation accuracy of the grinding process in High-Pressure Grinding Rolls (HPGR). The DEM code used a Particle Replacement Model (PRM) to simulate particle breakage. The PRM-based breakage model incorporated a calibrated breakage function obtained from experimental compression tests to precisely predict the progeny particle size distribution after each breakage event. A new approach was employed to resolve the increase in the number of particles in DEM simulations resulting from the particle breakage process that has been a major issue in the simulation of large systems. The approach was based on establishing a minimum size for breakage and compensating for the mass loss when the progeny particle size distribution is truncated. The DEM simulations of comminution processes in a piston press and a lab-scale HPGR proved the accuracy and capabilities of the developed model in the prediction of the devices’ performance including the product particle size distribution and the throughput. To achieve the overall aim of this thesis, first, a DEM code was developed in Fortran 77. As a means to validate the code, the effects of rolling friction torque, cohesion force, and particle size on the angle of repose were compared with published literature in the absence and presence of cohesion force. Also, a statistical analysis of the obtained angles of repose in various conditions resulted in a new dimensionless group that contained the bond number (the ratio between the cohesion force and the particle weight) and the rolling friction coefficient to the powers of 0.086 and 0.177, respectively. Additionally, to characterise the microscopic behaviour of the system, a thorough analysis of the combined influence of rolling friction and the bond number on the coordination number was carried out. This study demonstrated that by increasing the bond number the coordination number increased but by increasing the rolling friction the coordination number decreased. In contrast, the angle of repose increased with both bond number and rolling friction. The differences in the influence of these two parameters on the coordination number and the angle of repose suggested a difference in the functional mechanisms of the two parameters in the pile formation process. After the successful development of the preliminary DEM computer code, a Particle Replacement Model (PRM) was implemented into the code as a primary step towards the simulation of breakage processes. PRM works by substituting a given parent particle by a progeny when a minimum required force for breakage was met. The breakage function model proposed by Austin and Lukie (1972) was fitted to the experimental data of single-particle and particle bed breakage tests and incorporated into the PRM aiming to predict the progeny particle size distribution. In the developed PRM, a relaxation factor was employed to reduce the large contact forces between the replaced progeny particles. Also, a truncation size and a minimum size for breakage parameters were used to eliminate the fine progeny particles and compensate for the breakage impossibility of small particles in a compression test, respectively. Firstly, the calibration of the relaxation factor resulted in the value of 0.250 for the factor that led to the lowest possible particle velocity and breakage events in the simulation of a piston press unit. Then, DEM simulations of particle bed breakage processes in the piston press unit were used to obtain the best values of 0.93 mm and 3.0 mm for the truncation size and the minimum size for breakage, respectively, by comparing the simulation results with experimental data. In the next phase of this study, the geometry of a laboratory-scale HPGR was implemented in the DEM code to simulate a continuous grinding process. A spring-damper model was employed to simulate the hydraulic system connected to the floating roll of an HPGR. Then, the reactions of the spring-damper model to the change in the size and stiffness of unbreakable particles were examined. By increasing the size and stiffness of feed particles, the predicted working gap and the compressive force increased. These increases were in good agreement with the published literature. Since the mass loss resulting from the elimination of fine particles caused a significant decrease in the volume of the compressed particle bed, the model was modified to compensate for the mass loss without affecting the final PSD. In the modified model, the lost mass was distributed to the progeny particles smaller than the minimum size for breakage those were not allowed to break, so, the final PSD was not affected by the compensation. Simulations of the HPGR in the absence of mass loss showed a significant improvement concerning the case in the presence of mass loss. The observed improvement was in the trends of the throughput, energy consumption, working gap, and final PSD as a function of the initial working gap, initial hydraulic force, and roll speed. These promising results proved the potential of the model for further employments in DEM simulations of grinding processes in future works.
- Subject
- discrete element method (DEM); high-pressure grinding rolls (HPGR); particle replacement model (PRM); breakage function; JKR cohesion force model; angle of repose; modelling and simulation; non-linear contact model
- Identifier
- http://hdl.handle.net/1959.13/1509546
- Identifier
- uon:56257
- Rights
- Copyright 2021 Vahid Hassanzadeh
- Language
- eng
- Full Text
- Hits: 9
- Visitors: 10
- Downloads: 3
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Thesis | 10 MB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Abstract | 497 KB | Adobe Acrobat PDF | View Details Download |