On hydrogen addition effects in turbulent combustion using the Flamelet Generated Manifold technique

The idea of reducing emissions and pollution in turbo-machinery technology is growing significantly in the last decades. In order to reach these standards and to guarantee, at the same time, efficient combustion systems, new configurations for burners are required. Classical approaches such as experimental techniques require demanding configuration setups and high costs. The H2- IGCC project has been started in order to provide and demonstrate technical solutions for highly efficient and reliable gas turbines in the next generation of Integrated Gasification Combined Cycle (IGCC) plants. Inside this project, a CFD combustion analysis for gas turbine applications has been carried out. Thereby, a combustion model for numerical calculation is used in order to reach a reliable design approach. Among different combustion models, a reduction chemistry method called Flamelet Generated Manifold (FGM) is adopted. This technique becomes an answer to the problem of the huge computational effort required by the solution of a whole reactive system, where all species equations need to be solved. In FGM, chemistry is modeled by using the solution of one-dimensional flames called flamelets. In this way, the whole reactive partial differential equations system can be replaced by a small number of controlling variable equations. A typical controlling variable, in addition to the reaction progress variable, is for example the enthalpy, to take the heat loss effects into account. The key properties of the flame, such as density, diffusivity, temperature, are stored in the FGM database called manifold. During the CFD simulations, these properties are retrieved from the manifold. In turbulent combustion, a presumed beta-PDF approach can be assumed as a reasonable choice for the probability distribution of the sub-grid chemical terms. An algebraic model for variance is used and, therefore, variance of the progress variable becomes an extra controlling variable of the FGM system. The approach described above is suitable for relatively simple gases such as methane, for which there is a balance between molecular and thermal diffusion. In case of hydrogen addition, difficulties increase due to the instability of the gas. The high mobility of its molecules, which is much larger than the diffusion of heat, is known as preferential diffusion, resulting in a Lewis number lower than unity. For gases such as methane there were no such effects. From the physics point of view, the consequence of this problem is that the flame front brakes up into cellular structures. Moreover, super-adiabaticity phenomena appear in hydrogen flames, attributed also to preferential diffusion effects. In order to model hydrogen addition in the fuel, preferential diffusion effects in the equation system have to be accounted, taking care on the dependencies between enthalpy and element mass fraction in order to obtain good predictions for the burning rate and emissions. In order to analyze the turbulent structure of hydrogen flames, DNS data have been scrutinized and compared with laminar flame structure