Coherent structure impacts on blowoff using various syngases

摘要

Swirl stabilized combustion is one of the most successful technologies for flame and nitrogen oxides control in gas turbines. However, complex fluid dynamics and lean conditions pose a problem for stabilization of the flame. The problem is even more acute when alternative fuels are used for flexible operation. Although there is active research on the topic, there are still various gaps in the understanding of how interaction of large coherent structures during the process affect flame stabilization and related phenomena. Thus, this paper approaches the phenomenon of lean premixed swirl combustion of CH4/H2/CO blends to understand the impacts of these fuels on flame blowoff. An atmospheric pressure generic swirl burner was operated at ambient inlet conditions. Different exhaust nozzles were used to alter the Central Recirculation Zone and observe the impacts caused by various fuel blends on the structure and the blowoff phenomenon. Methane content in the fuel was decreased from 50% to 10% (by volume) with the remaining amount split equally between carbon monoxide and hydrogen. Experimental trials were performed using Phase Locked PIV. The Central Recirculation Zone and its velocity profiles were measured and correlated providing details of the structure close to blowoff. The results show how the strength and size of the recirculation zone are highly influenced by the fuel blend, changing stability based on the carbon-hydrogen ratios. Nozzle effects on the shear flow and Re numbers were also observed. Modelling was carried out using the k-ω SST CFD model which provided more information about the impact of the CRZ and the flame nature close to blowoff limit. It was observed that the model under-predicts coherent structure interactions at high methane fuel content, with an over-prediction of pressure decay at low methane content when correlated to the experimental results. Thus, complex interactions between structures need to be included for adequate power prediction when using very fast/slow syngas blends under lean conditions.