Radical tar removal : numerical modeling of tar conversion in a partial combustion reactor

In our current society, society is more than ever aware that fossil fuel stocks are finite. As a result, the interest in alternative sustainable energy sources has increased significantly. A promising renewable conversion technique is biomass gasification. Gasification can be defined as thermal degradation (devolatilisation) in the presence of an externally supplied oxidizing agent. The resulting gas mixture is called producer gas or syngas, and is itself a fuel. The gas consists mainly of CO, CO2, H2O, H2, CH4, and other hydrocarbons [van Loo and Koppejan, 2000]. Syngas can be directly used for the production of heat and electricity. During the process of gasification, tars are formed which exit the gasifier in vapor phase. On, or in cold pipes, e.g. sampling pipes of biomass pyrolysis or gasification plants, tars tend to condense, and then gradually carbonize or polymerize [Vreugdenhil and Zwart, 2009]. Tar condensation creates problems like fouling and plugging of after-treatment, conversion and end-use equipment. Tar formation during the thermal decomposition of biomass is not avoidable. For the wide-spread application of biomass gasification it is of great importance to convert or remove the tars at high temperature before condensation takes place. The research presented in this thesis focuses on an after-treatment technique which is known as partial combustion or partial oxidation. During ex-situ partial combustion producer gas is partially combusted at a low air factor, reducing the tar content by tar cracking. Former experimental research has demonstrated the possibilities and difficulties of tar conversion by partial combustion, leaving questions unanswered. The current status of technology and the promising advantages led to the inspiration to uncover more of the fundamentals of tar conversion in a partial combustion reactor. A modeling approach has been chosen as the main research tool. Modeling of combustion applications very often requires the use of detailed chemistry models in two or even three dimensional simulations. To account for tar conversion, complex reaction mechanisms involving many chemical species and reactions have to be used. It is well-known that the use of such detailed mechanisms results in high computational costs, and that reduction methods are needed, such as chemical reduction techniques, to decrease the computational burden. The reduction technique regarded here is the flamelet-generated manifold (FGM) approach [van Oijen, 2000]. The purpose of this research is twofold. First, the bigger picture is concerned with gathering additional knowledge of the physical and chemical mechanisms behind tar conversion by partial combustion. This is required to optimize the process and the reactor geometry for the partial oxidation of tar. Secondly, to achieve this, an investment has been made into the available reduction technique FGM. This combined approach has led to the development of a modeling tool which has been used to gain fundamental knowledge of tar conversion in a partial combustion reactor. The development of FGM, which is extensively discussed in this work, is also applicable to other multidimensional combustion systems. Within the partial combustion reactor several laminar diffusion flames are created to convert the tars present in the producer gas. To study the complex combination of physical and chemical processes taking place in the reactor, an extensive validation study has been executed of the application of FGMs to laminar diffusion flames. Firstly, a well-documented diffusion flame is modeled. The solution of a FGM is compared to the solution of the full set of transport equations. Special attention has been given to preferential diffusion. Polycyclic Aromatic Hydrocarbon (tar is a collection of PAHs) chemistry is not taken into account at this stage, due to the limitation of the number of species that can be involved in solving the full set of transport equations. By including a tabulated Lewis number for one of the controlling variables, the progress variable Y, large improvements are observed in the FGM results compared to the full chemistry solution. It can be concluded that FGM is an efficient dimension reduction technique that has great potential for accurate simulations of the laminar non-premixed flames studied in this thesis. Next, PAH chemistry is included, and it is demonstrated that the results achieved with FGM in one-dimensional flamelet calculations, both with and without the inclusion of preferential diffusion effects, are in good agreement with the full chemistry solution. The number of species that can be involved in solving the full set of transport equations in a multi-dimensional geometry is limited, and it was therefore not possible to numerically validate the FGM results in a multi-dimensional environment. As a result, a FGM is applied within a two-dimensional environment, and compared to the experimental results of qualitative measurements of the concentration of PAHs, and OH. Planar Laser Induced Fluorescence, and Laser Induced Incandescence measurements have been executed on a laminar diffusion flame, where the fuel flow is seeded, at three different doping rates of tars, here represented by benzene and toluene. The OH profiles of the FGM solution and the experiment are in good agreement. Phenomena like ring-growth and an increase in PAH formation with increasing dopant concentration, are observed in both the numerical simulations and the experiments. Finally, all gained knowledge has been combined to model the three-dimensional partial combustion reactor, including detailed PAH chemistry and preferential diffusion using FGM. The results show that it is unlikely that naphthalene (tar modeling component) can be converted up to 95%. The numerical results showed a decrease in the concentration of naphthalene of only 5%. It is likely that the remaining 95 % will subsequently lead to soot. So, based on the observations in this thesis work, it appears that applying a partial combustion reactor to convert tars in producer gas, is not able to convert the tars, and the remaining tar will most likely lead to soot formation.