Session: 04-43 Combustion dynamics - modeling I

Paper Number: 123589

123589 - LES and POD/DMD Analysis of Thermoacoustic Oscillation of a Hydrogen Micromix Flame Pair 

In September 2023, Kawasaki launched the world's first DLE gas turbine for Hydrogen. Its combustor is a new design based on the micromix (MMX) combustion principle. This combustion concept is inherently safe towards flashback, and the NOx emissions are suppressed by short residence time in the high temperature zone. The underlying gas turbine is the M1A-17 with a power output of 1.7 MWe. Natural gas co-firing of 50-100vol.% H₂ is possible for the entire load range. Utilization of hydrogen as gas turbine fuel drastically changes the combustion characteristics and can promote combustion instabilities. In a visualization test combustor with a line-of-sight of flame arrangement, a low-frequency oscillation with longitudinal mode at a frequency of 875 Hz has been found beyond the rated power by OH chemiluminescence imaging.

Numerical simulation of thermoacoustic instabilities in a combustor is a promising method for the understanding of the phenomena of combustion induced thermoacoustic oscillations. The objective of this article is the investigation of dynamic effects of the micromix flames located inside a combustion chamber with emphasis on the self-exciting feedback loop of heat release and pressure fluctuations.

The thermoacoustic properties of a micromix flame pair is modelled with Simcenter STAR-CCM+ from Siemens Digital Industries Software. For the turbulent modelling, the Large Eddy Simulation (LES) method is applied. The inherent unsteady regime is feasible for resolving temporal periodical phenomena like thermoacoustic instabilities. The computational domain comprises two flames with opposed injection direction. The boundary conditions are carefully prescribed to replicate the experimental results that have been acquired at a low-pressure test facility. The LES results are validated against experimental results. The snapshots of the unsteady fields are captured and processed by Proper Orthogonal Decomposition and Dynamic Mode Decomposition.

A dominant oscillation pattern could be reproduced by the application of LES under the given boundary conditions. The POD clearly depicts the dominant modes of the unsteady fields. The 0^th mode corresponds to the mean, followed by the higher mode associated with the higher energies. DMD, on the other hand, decompose the unsteady fields into the spatial modes associated with the temporal frequencies. Both POD and DMD are capable of extracting distinct features from the complex flow field.

Presenting Author: Daniel Kroniger Kawasaki Heavy Industries, Ltd.

Presenting Author Biography: 2004-2012: Mechanical engineering (Diploma) and Energy Economics (Diploma) at RWTH Aachen University, Germany
2012-2018: Scientific Researcher at RWTH Aachen University, Germany (Doctor)
2019-today: Researcher and Team Leader Gas Turbine at Corporate Technology Division, Kawasaki Heavy Industries, Ltd., Japan

Authors:

Daniel Kroniger Kawasaki Heavy Industries, Ltd.
Hiromu Kamiya Kawasaki Heavy Industries, Ltd.
Atsushi Horikawa Kawasaki Heavy Industries, Ltd.
Ryuta Suzuki Siemens K.K.
Erik Munktell Siemens Industry Software AB
Rene Braun Siemens Digital Industries Software