Mathias Lemke

Model reduction of reactive processes

Interpolation based model reduction is applied to a reactive process. A zero-dimensional, perfectly stirred, constant pressure reactor with complex chemistry, modeled by the GRI3.0 scheme, is considered. The aim of this work is to analyze how interpolatory model reduction performs for combustion processes, where the solution is very sensitive to the choice of input parameters. In this study the initial temperature is chosen as varying parameter.

Adjoint-based analysis of thermoacoustic coupling

We derive adjoint equations for reactive multi-species, compressible flows. Examples for homogeneous and non-homogeneous flows are calculated. To validate the adjoint approach for this highly non-linear problem a comparison between the adjoint solution and a finite difference expression for the objective function is done. We find, that the adjoint approach works well for reactive flows in spite of the strong non-linearity.

Adjoint-based sensitivity analysis of quantities of interest of complex combustion models

An adjoint-based approach for the sensitivity analysis of complex reaction mechanisms is presented. It builds purely on the evaluation of the governing equations. No adjoint equations have to be derived explicitly. Instead, the required adjoint operator is constructed numerically. The approach can be utilised for various kinetic models and in existing codes with minimal implementation effort. All dependencies on the state and on model parameters are fully evaluated without simplifications. Sensitivities are calculated more efficiently and more robustly compared with the often-used brute-force method. The approach is demonstrated for a homogeneous (zero-dimensional) reactor with different complex reaction mechanisms including several reaction types.

Pressure estimation from PIV like data of compressible flows by boundary driven adjoint data assimilation

Particle image velocimetry (PIV) is one of the major tools to measure velocity fields in experiments. However, other flow properties like density or pressure are often of vital interest, but usually cannot be measured non-intrusively. There are many approaches to overcome this problem, but none is fully satisfactory. Here the computational method of an adjoint based data assimilation for this purpose is discussed. A numerical simulation of a flow is adapted to given velocity data. After successful adaption, previously unknown quantities can be taken from the – necessarily complete – simulation data. The main focus of this work is the efficient implementation of this approach by boundary driven optimisation. Synthetic test cases are presented to allow an assessment of the method.

Adjoint-based reconstruction of an entropy source by discrete temperature measurements

The heat release rate is a significant quantity of lean premixed flames in technical flows. The heat release is commonly determined by OH* chemiluminescence measurements which are not fully reliable for turbulent flames owing to a superposition of desired emissions and broadband noise. The present study applies an adjoint-based data assimilation technique in order to reconstruct the heat release rate of a flame from pointwise temperature measurements. As a simplified test case, the method is applied to a two-dimensional channel flow and the combustion process is modelled by an artificial entropy source distributed in space and time.

Reconstruction of an Entropy Source by Temperature Measurements at Discrete Points with Adjoint Methods

The link between experiments and numerical analysis is a matter of particular concern. Numerical analysis can help to investigate complex and not well measureable flow fields. To this purpose the reconstruction of a flame in a complex channel flow is investigated. The fluidmechanical variables of the flame will be assimilated by means of adjoint methods using temperature measurements at discrete points downstream the flame.