Modern concepts for reducing thermal NO emissions require the use of very lean fuel/air mixtures. Therefore a problem of lean quench should be solved during design process of gas turbine combustor and it’s operational development. Since maintenance of flame stability for wide range of gas turbine engine operational modes is essential, therefore there is a great demand for models which are able to predict lean blow out limits of turbulent, premixed and partially premixed, aerodynamically stabilized flames. In this paper a model describing flame destabilization process is presented. This model takes into account various physical processes, which lead to flame destabilization. The model is based on equation for reaction progress variable. An expression of source term of this equation contains turbulent flame speed, which is calculated with the use of Zimont’s formula modification, proposed by authors. The results of simulation were compared with test results for our lean premixed combustor. Fuel mass flow rate of pilot zone was decreased during test until heat release of pilot flame front became insufficient and couldn’t support a combustion process in a lean premixed zone. Our simulation with modified model allows to get prediction of lean blowout limit.

Most of the modern combustor’s designs use staged concepts for reducing thermal NO emissions. Usually, a combustion process takes place inside the main zone, which uses very lean premixed fuel/air mixtures. A diffusion pilot zone supports combustion process inside a lean main zone. Thermal NO formation process takes place predominantly inside hot diffusion flame. So, operation modes of pilot and main zones must be arranged to provide low NO emissions of pilot zone and maintain flame stability inside the main zone simultaneously. In this paper, a new turbulent combustion model is presented. This model allows to model diffusion and premixed flames and takes into account various physical processes, which lead to flame destabilization. The model uses an equation for reaction progress variable. Within the considered approach this equation has two source terms. These terms are responsible for different conditions of combustion process: diffusion flames and premixed flames, and distributed reacting zones. This paper studies the problem, concerning modeling of lean blowout process of diffusion flame front. To test the proposed combustion model we have simulated lean blowout process inside combustion zone of a gas turbine combustor. Good predictions of lean blowout limits were obtained.

Most of the modern combustor’s designs use staged concepts for reducing thermal NO emissions. Usually, a combustion process takes place inside the main zone, which uses very lean premixed fuel/air mixtures. A diffusion pilot zone supports combustion process inside a lean main zone. Thermal NO formation process takes place predominantly inside hot diffusion flame. So, operation modes of pilot and main zones must be arranged to provide low NO emissions of pilot zone and maintain flame stability inside the main zone simultaneously. In this paper a concept of new turbulent model combustion model is presented. This model allows to model diffusion and premixed flames and takes into account various physical processes, which lead to flame destabilization. The model uses an equation for reaction progress variable. In the frameworks of considered approach this equation has three source terms. These terms are responsible for different conditions of combustion process: diffusion flames, premixed flames and distributed reaction zones. A proposed model was widely validated for different types of combustion chambers such as: 1) Bluff-body flameholder (lean premixed combustion: modeling of lean blow out); 2) Conventional diffusion regime of combustion chamber of gas turbine engine (modeling of flame stabilization and NO emissions); 3) Combined combustion regime of combustion chamber: burning process is inside pilot diffusion and main premixed zones (NO emissions and lean blow out limits for several operational modes). These tests had shown a good agreement of experimentally obtained data with results of simulations.

No abstract