A highly accurate model is presented that predicts boundary layer flashback (BLF) for lean premixed hydrogen combustion. In contrast to existing models, that fail for lean hydrogen mixtures, the present model includes the effect of flame instabilities. The model is applied to study the effect of burner geometry on flashback tendency. A diverging geometry results in a higher flashback tendency due to an adverse pressure gradient. These observations are in line with experimental results. It shows that the BLF model is a tool that could significantly contribute to the development of lean premixed hydrogen burners. This is a great step towards carbon-free industrial gas turbines.

The worldwide gas emissions regulations are getting more stringent. Together with the fluctuations of the finite fuel resources, it is expected that the use of unconventional and renewable fuels will increase. Hydrogen fuel is a zero-emission, renewable energy source. It has a great potential as one of the principal sustainable energy sources of the future. A gradual conversion to a hydrogen based energy system, if the necessary government framework is provided, is far more likely to happen. This opportunity can be found in the natural gas grid of the Netherlands. This natural gas infrastructure is accessible for both the industry and Dutch households. The IEA HGH RD programme has conducted research about this gradual transition, with promising results. It is currently therefore a key agenda point for the Dutch government. In order to safely make a transition to natural gas with hydrogen addition, it is of importance to know the fundamental combustion properties of these fuel mixtures. These properties are needed for the prediction of and modeling of combustion systems. Traditionally, the quantification of the combustion process is done by focusing on measuring properties such as laminar burning velocity, ignition and extinction characteristics, flammability limits and emissions. The laminar burning velocity is an important property of combustion mixtures. It contains fundamental information on reactivity, exothermicity and diffusivity.

One of the challenges in gas turbines is in the reduction of NOx and CO emissions due to simultaneous
dependence on stringent emission norms and subsequent demand for change in operating conditions. To cater the volatile demand of power supply, most of the gas turbines have to be flexible to changing loads. Closer the load is to full load, more is the impact on gas turbine emissions. In order to mitigate the effects of emission rise, a shift towards emission reduction techniques are being adopted. Of the available options, Exhaust Gas Recirculation (EGR) is a promising option for good part load performance and emission reduction. EGR is a technique to reduce the flame temperature by recirculating the products of combustion into air intake and replacing part of the oxygen content with inert constituents such as CO2. This leads to reduced NOx emissions and increased CO2 content at the exhaust for effective carbon capture. Despite these advantages, the introduction of an oxygen-depleted oxidizer leads to changes in combustion behaviour giving rise to flame stability issues. In this thesis, a chemical kinetic model to predict emissions in a lean premixed combustor coupled with EGR, is built using chemical reactor networks (CRN). A CRN model was developed by splitting the whole combustor volume approximately in half before and after the recirculation zone. The first half was further partitioned into the flame reactor and the recirculation reactor. The flame reactor volume
is set based on the chemical time scale approach. This approach quantifies the flame zone volume based on the difference in reaction kinetics for different oxidizers. The model was also validated with results from the literature for various flame cases with similar flow fields and was in good agreement with the data. Various chemical mechanisms and their emission prediction ability are investigated. GRI-Mech 3.0 is found to accurately predict the emissions for a given set of operating conditions. The effect of change in oxidizer composition with different amounts of CO2 is investigated. As a consequence of O2 starvation in the oxidizer, NOx formation is inhibited and CO levels are escalated.
At the same flame temperature, NOx levels reduced by a maximum of 40%, whereas the CO levels rose simultaneously as hgh as 50%. A parametric study on varying the combustor pressure indicated a rise in NOx levels and drop in CO levels with increase in pressure. The pressure rise augmented the NOx formed through thermal pathway and lead to the rise in NOx beyond 1800 K. On the other hand, CO2 dissociation to CO is suppressed by the rise of pressure, hence decreasing its magnitude. The impact of using wet EGR is investigated and compared with results of dry EGR. It is observed that NOx reduction is enhanced by 5-10%. Whereas the CO levels is escalated by 10-20%. The combined kinetic effect of CO2 and thermal effect of H2O is found to be reason for the change in emissions. This indicates an enhancement in combustion reactivity. Hydrogen injection raised the NOx levels by 5-10% and the CO levels dropped by a maximum of 12.5%. CO level suppression was a clear indication of the augmented flame stability and was found to be achieved with a slight compromise on NOx levels. Additionally, the effect of premixer efficiency is tested by discretizing the volume as parallel reactors with varying equivalence ratio inputs. Due to decreased premixing, unmixedness leads to various rich and lean pockets inside the combustor. This is a reason evident rise in both NOx and CO levels with decreasing premixer efficiency.