DNS of acoustic Instabilities in low Emission Combustion Systems
MetadataVis full innførsel
In a state of the art gas turbine using Lean Pre-Mixed fuels one of the main challenges is efficient and reliable control of sound generated during combustion. Knowledge of sound generation in gas turbine combustion chambers has to be enhanced in order to develop a reliable model with predictive capabilites.In this thesis Direct Numerical Simulations of two dimensional laminar imploding circular flame fronts have been performed. One dimensional simulations of laminar opposing flame fronts have been performed to establish modeling conditions for the two dimensional simulations, evaluate the boundary influence on the simulations and provide comparable simulation results. In addition a pre-study for three dimensional simulation of an inwards burning sphere of fuel has been done. The motivation of this thesis is to enhance the knowledge related to generation of acoustic waves in the combustion chamber of a gas turbine.The S3D code (a parallel DNS code for solving reactive flows) was modified to include two dimensional circular imploding flame fronts. A thorough investigation to validate the boundary influence on the annihilation event is recommended. This is due to simulations which indicate that the boundary conditions may influence especially the pressure drop after time of impact.In two dimensions DNS have been performed at pressure 1atm for fuel equivalence ratios 0.3, 0.5, 0.8 and 1.5 with detailed chemistry representation. Care has been taken to ensure adequat resolution of the flame (a minimum of ten points over the flame). The results from the simulations were used to measure key parameters (as the pressure drop after impact, the laminar flame speed, etc.). The following trends were found for the two dimensional simulations with increasing fuel equivalence ratio (in the given range): The flame thickness, unburned and burned gas density decrease, while the fluid expansion velocity, laminar flame speed, propagation speed of the pressure wave, pressure difference before the flame fronts meet, pressure drop after impact and burned temperature increase. This coincides with the the trends in the one dimensional simulations, and is consistent with the given theory.The results were compared with analytical relations developed in the candidate's project work of fall 2012. It was found that the three relations gave a poor impression of the measured values.It is indicated that to fully understand the annihilation process a number of simulations have to be run. The propagation speed of the pressure wave, the fluid expansion velocity and the pressure drop after time of impact require special attention.