Experimental and Numerical Advancement of the MGT Combustor Towards Higher Hydrogen Capabilities

Abstract

Blending of natural gas with hydrogen is a viable pathway for the decarbonization of industrial gas turbines for combined heat and power applications. Very high blending ratios of hydrogen are needed to achieve significant CO2 emission reductions. However, burning high hydrogen contents in the gas turbine is challenging in terms of NOx emissions and the mitigation of flashback risks as well as suppressing thermoacoustic instabilities. This paper illustrates a design modification to improve the hydrogen capabilities of the Advanced Can Combustion (ACC) system and its ultra-low emission industrial swirl burner for the MGT6000 gas turbine that was originally designed for pure natural gas combustion. A flow conditioner is installed upstream of the swirler aiming to decrease the fuel amount close to the combustor walls and thereby increase the flashback resistance of the burner. High pressure (≈14bar) full power (≈4MWth) single can combustion tests and atmospheric burner tests are used for the assessment of the hydrogen capabilities for the original and the retrofitted burner. Different levels of hydrogen blending of up to 45 vol-% at high pressure and 93 vol-% at atmospheric conditions as well as different gas turbine relevant flame temperatures are assessed in terms of emissions, flame flashback and thermoacoustic stability. Low speed thermocouple measurements at the burner walls are identified as a good precursor for hydrogen induced flame flashback at the walls. The amplitude of the thermocouple fluctuation is observed to be similar for atmospheric and elevated pressure. Moreover, it is shown that the increase in NOx emissions associated to hydrogen blending can be transferred from atmospheric conditions to elevated pressure. The experimental dataset is used for the calibration of Computational Fluid Dynamics (CFD) calculations to allow for the assessment at different operating conditions and future modifications. The CFD is focused on the prediction of flashback resistance for different blends of hydrogen and natural gas at high pressure conditions.