The flow of dense, particle-laden suspensions in planar channels contains myriad unexplored fundamental fluid mechanics problems. To investigate these, a high-fidelity numerical framework has been developed, which uses the lattice Boltzmann method (LBM) and discrete element method. It is based on the open-source TCLB code base [1], which facilitates efficient calculations on GPU architectures and parallelisation on HPC clusters, allowing computations at significant length and time scales or repeated, small-scale simulations for stochastic modelling. The framework has been applied to two fundamental and distinct problems associated with particle laden flows in planar channels. By directly simulating the shear-induced migration of polydisperse suspensions for the first time, it was discovered that the smallest particles preferentially form the plugs in the plugging regime, causing the largest particles to segregate to the channel walls [2]. Then, the hydrodynamic clogging of micro-particles was studied using a stochastic methodology based on hundreds of high-fidelity simulations, demonstrating dependence on the system dynamics and electrostatics via detailed parametric maps [3]. Most current modelling assumes smooth channel walls, however in reality all channels, particularly at the micron scale, exhibit some measure of surface roughness. Most recently, the above modelling framework has been applied to planar fractures with textured surfaces, to answer questions such as where does the cross-over from clogging to jamming occur, what is the associated variation in effective permeability with solid volume fraction, and how do fracture and particle parameters affect these? Results suggest that, if boundary effects are eliminated by making the fracture geometries large enough, then effective permeability smoothly decreases with solid volume fraction, as opposed to a step-decrease due to clogging of the entire system, which could fundamentally change the way that dense suspension transport in narrow geometries is perceived. [1] Łaniewski-Wołłk, Ł. and Rokicki, J., 2016, Adjoint lattice Boltzmann for topology optimization on multi-GPU architecture, Computers & Mathematics with Applications, 71(3), 833–848, 2016. [2] Di Vaira, N.J., Łaniewski-Wołłk, Ł., Johnson Jr., R.L., Aminossadati, S.M. and Leonardi, C.R., Influence of particle polydispersity on bulk migration and size segregation in channel flows, Journal of Fluid Mechanics, 939, A30, 2022. [3] Di Vaira,