Chemical reactor network modelling of a 1000 HP turboshaft engine reverse flow combustor

Turbulence and combustion are two important phenomena's that are very difficult to solve for scientists and engineers. These two phenomena are playing the biggest role in the all fossil-fuelled systems which have been using for over 150 years. By means of the rapid computer technology developments, it has become possible to solve the turbulent combustion flow problems with computational fluid dynamics (CFD) simulations and to predict the physical results such as fluid flow and temperature, as well the major chemical species with high accuracy. However, by using the same method, it is still a difficult and time consuming process to calculate the minor chemical species such as NOx and CO which are released during combustion process. For this reason, the prediction of this pollutants is very important in the design phase of gas turbine engines. On the other hand, due to environmental factors and legal limitations, development efforts to reduce pollutants such as NOx and CO produced by fossil-fuelled systems such as gas turbine engines have become a necessity. The realization of these efforts in conceptual and detailed design phase's before production is very important in terms of reducing the time and costs of the projects. In this thesis study, a new methodology was developed and applied to predict the emissions of a reverse flow combustor of a 1000 horsepower turbo-shaft helicopter engine in the design process. This methodology includes the following respectively: (1) Three dimensional CFD-combustion simulation for 1 / 13th part of the combustion chamber by using FLUENT v.14 software and obtaining temperature values and velocity vectors at each solution point. (2) Converting the three dimensional CFD results and transferring to a two dimensional plane by using two and a half dimension (2.5 D) approach. (3) Developing a MATLAB based tool to create Chemical Reactor Network (CRN) model on two dimensional plane. (4) Solving the CRN model with using n-heptane fuel mechanism by using CHEMKIN PRO software and obtaining emission results. In this methodology, the temperature and mass flow values are obtained by CFD simulations and the CRN model is constructed by using these values. In this way, complex chemical mechanisms can be calculated quickly and small chemical species can be obtained precisely. The temperature, which is the most important parameter affecting the amount of NOx and CO generated as a result of the combustion reaction, is determinative at the creation stage of the CRN model in this study. Most of the previous similar studies were made for combustion chambers in the form of axially symmetric and simple geometric tubes. Therefore, CFD simulations for a two dimensional plane passing through the centre of the injector, and the creation of a chemical reactor network from these simulation results, are become easier than in this type of combustors. In case of the geometry in xxi this study is an annular and non-symmetrical combustion chamber which are consisting of a sequential fuel injector system. Building the reactor network model using the three dimensional CFD results and the methods used in the previous studies is quite difficult, complicated and long process for the geometry in this work. To overcome this challenge, the 2.5 D approach, which is generally using in image processing studies, has been utilized in this thesis. In this method, nine sections were taken at equal intervals over the three dimensional volume, and the average of CFD results in these sections were transferred to a created two dimensional plane and the CRN model was constructed on this plane. A MATLAB-based tool has been developed to determine reactor volumes and mass fluxes between these volumes. With this tool, the chemical reactor network for three different operating conditions of the combustion chamber (idle, cruise, take-off) was modelled. These models were transferred to the CHEMKIN PRO software and detailed chemical reaction analysis was carried out by using reduced n-heptane fuel mechanism which consist of 159 chemical species and 770 reaction steps. In this study, reactors are assumed as ideal and adiabatic.