In this paper the development and the application of a numerical code suited for the simulation of gas-turbine combustion chambers is presented. In order to obtain an accurate and flexible framework, a finite-rate chemistry model is implemented, and transport equations for all species and enthalpy are solved. An assumed PDF approach takes effects of temperature and species turbulent fluctuations on the chemistry source term into account. In order to increase code stability and to overcome numerical stiffness due to the large-varying chemical kinetics timescales, an implicit and fully-coupled treatment of the species transport equations is chosen. Low-Mach number flow equations and k-ε turbulence model complete the framework, and make the code able to describe the most important physical phenomena which take place in gas-turbine combustion chambers. In order to validate the numerical simulations, experimental measurements are carried out on a generic non-premixed swirl-flame combustor, fuelled with syngas-air mixtures and studied using optical diagnostic techniques. The combustor is operated at atmospheric and high-pressure conditions with simulated syngas mixtures consisting of H2, N2, CH4, CO. The combustor is housed in an optically-accessible combustion chamber to facilitate the application of chemiluminescence imaging of OH* and planar laser-induced fluorescence (PLIF) of the OH-radical. To investigate the velocity field, particle image velocimetry (PIV) is used. The OH* chemiluminescence imaging is used to visualise the shape of the flame zone and the region of heat release. The OH-PLIF is used to identify reaction zones and regions of burnt gas. The fuel composition is modelled after a hydrogen-rich synthesis gas, which can result after gasification of lignite followed by a CO shift reaction and a sequestration of CO2. Actual gas compositions and boundary conditions are chosen so that it is possible to outline differences and similarities among fuels, and at the same time conclusions about flame stability and combustion efficiency can be drawn. A comparison between experimental and numerical data is presented, and main strengths and deficiencies of the numerical modelling are discussed.

This content is only available via PDF.
You do not currently have access to this content.