Improved dust dispersion modelling for surface quarries: an optimized RANS k − ε approach

Joseph, Genora M.D. (2016) Improved dust dispersion modelling for surface quarries: an optimized RANS k − ε approach. PhD thesis, University of Nottingham.

[img] PDF (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (12MB)


Conventional dust dispersion models which employ Gaussian concentration distributions are routinely applied to predict the dispersion and deposition of fugitive dust arising from quarrying activity. However, these models are known to over-predict the long range transport of particulates beyond the confines of the quarry pit, because their complex terrain algorithms are unable to accommodate the steep gradients which are typically encountered in quarry excavations. They therefore cannot account for the internal flow regimes which contribute to the removal of suspended particulates from the air stream.

Consequently, a need arises within the extractive industries, for improved dust dispersion models that are not constrained by quarry topography. This research project attempts to address this deficiency in the modelling of dust emissions from quarry blast events, by presenting an optimized RANS k−ε approach which can adequately simulate the flow-field in which particulates are entrained under a range of meteorological conditions.

The stages involved in the incremental development of the numerical model are documented in the thesis, and commence with characterization of the atmospheric boundary layer. In particular, the Monin-Obukhuv Similarity Theory is applied to parametrize the atmospheric surface layer, which exerts the most influence on the dispersion and deposition of near-ground particulate emissions.

Modifications to the standard k − ε model coefficients and the inclusion of buoyancy source terms have been adopted in this work, in accordance with previous studies by Alinot and Masson (2005). These modifications ensure that the turbulence closure equations are compatible with Monin-Obukhuv Similarity scaling of the atmospheric surface layer. The Businger-Dyer flux profile functions have been employed to introduce stability modifications to the logarithmic velocity, temperature and turbulence profiles which have been defined at the inlet boundaries of the computational domain to enable numerical representation of both adiabatic and diabatic atmospheric conditions. Furthermore, the Lagrangian Discrete Phase Model has been coupled with Eulerian solution of the flow field to provide a robust means of replicating fugitive dust dispersion through the stochastic tracking of injected particulates. The project also presents a consolidated post-processing methodology to incorporate wind direction variability due to mesoscale atmospheric effects into the CFD model. This methodology use the Moore (1976) equation to parametrize the standard deviation of wind direction variability and proposes a novel, Gaussian probability weighted averaging procedure to arrive at a resultant plume which accounts for the influence of mesoscale wind variability on particulate trajectory and improves the k −ε predictions of lateral spreading of the dust plume.

Simulations of flow and dispersion over a series of idealized cosine depressions of varying aspect ratio have been used to assess model predictions of the flow regime and the corresponding plume attenuation within artificial valleys. To allow the model to accommodate negative terrain elevations and to produce profiles of the flow variables which conform to wall topography, a wall-distance scalar has been introduced to ensure consistency of the inlet profile with the flow solution within the domain. These simulations have demonstrated that the proposed model surpasses UK-ADMS in terms of its ability to resolve strong recirculation regimes in deep depressions.

The project culminates in a case study of the Old Moor Quarry in Buxton, Derbyshire. This case study tests the viability of the proposed k−ε model and validates the model predictions of dust dispersion with field measurements obtained over the course of a monitoring campaign of approximately one month duration. Meteorological pre-processing steps in accordance with the findings of Holtslag and Van Ulden (1983) have been employed to derive atmospheric surface layer input parameters from routine meteorological data measured at the quarry site, eliminating the need for sophisticated meteorological measurements. Dust dispersion predictions obtained using the conventional dust dispersion model UK-ADMS, have been compared to the CFD model results to demonstrate the improved prediction accuracy of the proposed k −ε approach. Notably, the CFD model is shown to account for the various flow regimes which arise due to the combined effects of the site meteorology and the complex terrain of the quarry excavation.

Importantly, the statistical Performance metrics, FAC2, MG, FB and NMSE recommended by Hanna et al. (2004) for the evaluation of dispersion model performance, have been used to assess the accuracy of fugitive dust deposition predictions obtained from the proposed model. The performance evaluation exercise indicates that the buoyancy modified k − ε model outperforms UK-ADMS for all of the metric tests. The incorporation of the wind variability weighted averaging procedure in the case study simulations is seen to reduce uncertainty due to random error, quantified by NMSE. This due to the fact that the wind variability averaging procedure evens out outlying predictions which may be due to the inherent stochasticity of the DPM model.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Lowndes, I.S.
Hargreaves, D.M.
Keywords: CFD, particulates, dispersion, quarries, complex terrain, fugitive dust
Subjects: T Technology > TD Environmental technology. Sanitary engineering
Faculties/Schools: UK Campuses > Faculty of Engineering
Item ID: 33767
Depositing User: Joseph, Genora
Date Deposited: 16 Aug 2016 12:39
Last Modified: 19 Oct 2017 16:21

Actions (Archive Staff Only)

Edit View Edit View