ED 26-Modeling Physics

Modeling of Thermoacoustic Combustion Instabilities

ED - Mechanical Engineering

Thermo-Fluid Dynamics (TFD-group)

Modeling / Simulation of thermocacoustic instabilities

Aerospace / GeodesyComputer ScienceFurther disciplinesMathematicsMechanical EngineeringPhysics

Applied Mathematics, Thermodynamics, Fluidmechanics

Participation also possible online only:

Planned project location:

TUM Boltzmannstr. 15, D-85747 Garching

M.Sc.

Gregor

Döhner

gregor.doehner@tum.de

+49 (89) 289 16242

Thermoacoustic combustion instabilities result from constructive feedback between fluctuations of heat release rate and acoustic perturbations of velocity and pressure. Such instabilities may occur in modern gas turbines and rocket motors as well as in domestic or industrial burners. Possible consequences are increased emissions of noise or pollutants, limited range of operability or severe mechanical damage to a combustor. Thermoacoustic instabilities have hindered the development of low-emission, reliable and flexible combustion systems for power generation and propulsion. To analyze and control these instabilities, fluid mechanics, acoustics and combustion science are combined in an interdisciplinary approach with methods of system identification and control theory.

The ongoing projects in the TFD group contribute to several topics related to the field of thermoacoustic instabilities. Promising approaches to handle the multi-scale and multi-physics nature of thermoacoustic instabilities range from the development of reduced order models and machine learning techniques to the application of high fidelity CFD simulations.

Reduced Order Models: In our research group, we develop low-order acoustic network models as well as global linear stability tools for reactive flows. Combined with adjoint sensitivity approaches, these models provide insight into the physical mechanisms driving thermoacoustic instabilities and efficient optimization strategies for combustors with respect to stable combustion.

Machine Learning Techniques: Machine learning offers a data-driven approach to predict, analyze, and control thermoacoustic instabilities. By harnessing vast amounts of simulation or experimental data, machine learning models can uncover complex relationships, potentially leading to novel insights and more efficient control strategies. In our research group, we mainly focus on physics augmented machine learning and probabilistic machine learning.

CFD Simulations: High fidelity Computational Fluid Dynamics (CFD) simulations provide detailed insights into flow fields and interaction mechanisms that drive thermoacoustic instabilities. While computationally intensive, these simulations provide a comprehensive picture of the intricate dynamics and are crucial for the design of combustors, as well as the design and validation of reduced order models. System identification techniques are applied to post process CFD data to gain insight into the acoustic transfer behavior of flames.

In this project, research can be conducted on the multi-physics phenomenon of thermoacoustic instabilities, using any of the interdisciplinary approaches. The project-specific details can be agreed upon based on the student’s own interests and background.

40

3 years of bachelor studies completed

Advanced Mathematics;
Beneficial: Thermodynamics, Fluidmechanics, Control Theory, Machine Learning; Experience in Matlab / Python / Openfoam