Development of an environmentally friendly use of biomass and solid recovered fuels (SRF) for energy production is an important task of the Institute of Technical Chemistry at KIT. This task is being realized with help of a so called 3-step characterization concept for a detailed description of these fuels. In the first step a chemical composition of a biomass or an SRF is determined. The second step is focused on the combustion specific characterization of a fuel in the laboratory scale fixed batch reactor “KLEAA”, in which the combustion behaviour is described with local temperatures in the fuel bed, air ratios, gas composition above the bed and characteristic numbers: ignition time, reaction front velocity, mass conversion rate and specific heat release. In the third step a transfer of the results form a fixed bed onto a moving bed is performed. This step is possible when the mixing in the moving bed is limited, i.e. on a travel or a forward acting grate and for fuels with a low ash content. Furthermore, local primary air flows have to be equal for both beds. The resulting combustion behaviour on the moving grate can be validated in the pilot scale facility “TAMARA” equipped with a forward acting grate.
Experimental estimation of the combustion behaviour can be additionally reinforced with a numerical model “KLEAA Code”, which was initially developed for woody biomass. The model is based on a one dimensional cascade of ideally stirred batch reactors. The fuel is considered as a porous bulk, through which the gas passes through, therefore the mass, element and energy balance is formulated for both phases. The heat transfer occurs between the phases through convection and radiation, and between solid particles through conduction and radiation. Mass transfer is allowed to occur in the model in one direction – from the solid phase into the gas. The solid fuel is considered in the pyrolysis model to consist of cellulose, hemicellulose and lignin, for which heterogeneous, thermal decomposition reactions are formulated. Additionally, the cellulose decomposes partially into tar, which decomposes further into gas. The pyrolysis model consists therefore of five heterogeneous reactions.
The “KLEAA Code” was initially developed, to allow a simulation of the combustion of wood spheres (model fuel) in the KLEAA reactor, but it can be extended to also allow conducting simulations for the combustion of technical fuels such as wood chip and SRFs on a moving grate. Extending the model has to include a mathematical description of the differences between the physical, chemical and geometrical features of said fuels.
Defining these differences makes up the core part of this dissertation and was realized by characterization of three fuels according to the 3-step characterization concept – wood chips and two solid recovered fuels “BIOBS” and “SBS®1”, both produced by REMONDIS Rheinland GmbH. The physio-chemical analysis in the first step was complemented by a detailed sorting analysis and a near-infrared detection of the main components in the SRFs. Both analyses have proven that the SRFs can be summarized as a mixture of a woody biomass, inert matter (glass, stones, sand) and a number of various plastics. On the basis of these results the “SBS®1” was chosen as a representative solid recovered fuel and a thermogravimetric analysis was performed. It allowed to summarize an SRF as a mixture of wood, polyethylene and ash. In the results, the development of the pyrolysis model included an additional reaction for the thermal decomposition of polyethylene. A necessary characterization of the physical and geometrical properties of the chosen fuels was realized using two parameters important in the models of heat exchange, pyrolysis and char burnout in the “KLEAA Code”: density of the fuel and the specific surface area of its particles. A new measurement methodology for the determination of both parameters was developed. The density was measured with a pycnometer built especially with SRFs in mind. In contrast to other pycnometers in which a liquid or a gas is used, a very fine powder with liquid-like kinematic properties was applied. This allowed to avoid a measurement inaccuracy resulting from a liquid penetrating the pores of the fuel particles. The method was validated for materials with a known geometry and density, and showed a high reproducibility and plausibility of the results. Next the density of the investigated fuels was measured. As expected the wood chips are characterized by a density close to the density of the hard wood, whereas the density of the solid recovered fuel is noticeably higher, regardless of a considerable content of light, spongy particles.
Determination of the specific surface area of the fuel particles was carried out indirectly, by measuring the pressure drop in a cold, unreacted fuel bulk. The surface area was then calculated with three different formulas describing the relation between the pressure drop and the surface area of the particles. In order to validate this method several measurements were performed on bulks consisting of particles with a known geometry, such as spheres, cubes and plates, whereas the best results were obtained for spheres. Furthermore a formula showing the lowest sensibility towards the gas velocity through the fuel bulk was chosen. The measured specific surface areas for the wood chips and the SRF were considerably higher than the surface areas calculated directly from the Sauter mean diameter of their particles.
Characterization of the combustion behaviour of both solid recovered fuels and of the wood chips was realized in the fixed bed reactor „KLEAA” (second step) and in the pilot scale grate firing “TAMARA”. It was described with temporal or local temperatures in the fuel bed, air ratios, gas composition above the bed and with characteristic numbers: ignition time, reaction front velocity, mass conversion rate and specific heat release. The combustion behaviour of the wood chips and of the “BIOBS” can be described as representative. The quasi-stationary main combustion zone and the following char burnout were observed for the wood chips as expected. “BIOBS” combusted on the other hand without a distinct char burnout, which is a typical combustion behaviour for very wet, biogenic fuels. The results obtained for both fuels in the fixed bed were then transferred onto a moving bed and were validated in the pilot scale plant “TAMARA”. The second solid recovered fuel, the “SBS®1” combusted in the fixed bed differently from typical, biogenic fuels – combustion of volatiles coming from plastic particles laying on the bottom of the reactor was observed also during the char burnout.
The last part of the dissertation discusses the combustion behaviour of the wood chips and the „SBS®1“ simulated with the extended “KLEAA Code”. The measured fuel densities, the specific surface areas and the composition of the solid recovered fuel (assumed as a mixture of wood and polyethylene) were applied as input in the model. The simulated combustion behaviour for the wood chips was accurate in the main combustion zone, but the char burnout was too short comparing to the measurement. The simulated concentration of carbon monoxide and hydrogen above the fuel bed in the main combustion zone were at the same time higher than measured, which indicates that the carbon gasification with water occurs too fast in the model. The combustion behaviour of the “SBS®1” was simulated with a limited accuracy at the beginning of the main combustion zone. Neither the distinct char burnout, nor the combustion of plastics at its end were observed, which may be explained by the missing mass transport of the solid phase between the fuel layers (ideally stirred reactors) in the model. Furthermore the description of a solid recovered fuel with only average surface areas of its particles and with a simplified composition is most probably not sufficient.
The validation of the “KLEAA Code” for the combustion on the moving grate was performed for the wood chips and in a very limited extent for the „SBS®1“. The moving grate of “TAMARA” consists of four separate primary air zones, but the combustion of all three fuels occurred only in the first one due to a high, unavoidable false air input. The mass conversion was simulated correctly for the wood chips and with a moderate accuracy for the solid recovered fuel. Measuring the gas composition in the first primary air zone was however possible only at one location. The gas components measured at these positions were compared with the simulated ones and showed a similar accuracy as for the fixed bed.
Future work should include further development of the measuring methodology for the fuel density and for the specific surface area. An extensive database of the measured values would help to determine the plausibility of the measurements. The development of the “KLEAA Code” should include incorporating the heat losses and refining the pyrolysis model. Furthermore a more complex description of solid recovered fuels can be recommended.