Entrained flow gasification: mathematical modelling based on RANS for design and scale-up

Production of synthetic fuels and chemicals in a closed carbon cycle economy is one of the main challenges concerning climate change. In order to convert biomass residues to high-density fuels, the bioliq® process [1] has been realised at Karlsruhe Institute of Technology (KIT). Between 2005 and 2013, all stages of the process were developed and commissioned including the bioliq entrained flow gasifier (bioliq EFG) [2]. During the design phase in 2008/2009, a basic numerical model of the bioliq EFG was already developed [3, 4] and its major limitations were identified. These include simplified sub-models for atomisation, for conversion of char, pyrolysis oil and suspensions thereof, for gas and particle radiation and for slagging at high-pressure and high-temperature conditions. Since 2013, many experimental bioliq EFG campaigns have been carried out and reliable experimental results have been obtained. In parallel, the work on the numerical model of the bioliq EFG has been resumed. This paper focusses on the status quo of the numerical model, on the accuracy of recent numerical results and on the challenges faced in the development of a CFD based tool for design and scale-up. The numerical model is based on the RANS and the Euler-Lagrange approaches and assumes a steady-state. The rates of the gas-phase reactions are described either using the eddy-dissipation-concept model in combination with a reaction mechanism or using a relaxation-to-chemical-equilibrium model. Gas radiation is calculated using the discrete ordinates model and a weighted-sum-of-grey-gas model. The latter assumes six gases (five grey gases and one clear gas) and has been obtained from accurate line-by-line calculations. Slagging is implemented as boundary condition assuming both the downwards axial slag flow and the radial heat flow through the slag layer as one-dimensional. Temperature profiles in the liquid and the solid slag layer and in the wall are predefined as linear. The numerical results are reported for the bioliq EFG experiment V82.1 in which a model fuel (96 % ethylene glycol + 4 % A-glass) was applied. The results show that both the total heat extracted from the bioliq EFG and the gas-phase composition at outlet can be predicted with good accuracy. Due to uncertainties in the prediction of particle deposition, some deviations are observed concerning both the local heat extracted by the six cooling segments and the slag thickness. Sensitivity studies show that simplifications of the gas-phase reactions are acceptable.