Ingo Roehle

Investigation of Entropy Noise in Aero-Engine Combustors

Strong evidence is presented that entropy noise is the major source of external noise in aero-engine combustion. Entropy noise is generated in the outlet nozzles of combustors. Low-frequency entropy noise, which was predicted earlier in theory and numerical simulations, was successfully detected in a generic aero-engine combustion chamber. It is shown that entropy noise dominates even in the case of thermo-acoustic resonances. In addition to this, a different noise generating mechanism was discovered that is presumably of even higher relevance to jet engines: There is strong evidence of broad band entropy noise at higher frequencies (I to 3 kHz in the reported tests). This unexpected effect can be explained by the interaction of small scale entropy perturbations (hot spots) with the strong pressure gradient in the outlet nozzle. The direct combustion noise of the flame zone seems to be of minor importance for the noise emission to the ambiance. The combustion experiments were supplemented by experiments with electrical heating. Two different methods for generating entropy waves were used, a pulse excitation and a sinusoidal excitation. In addition, high-frequency entropy noise was generated by steady electrical heating.

Fundamental Mechanism of Entropy Noise in Aero-Engines: Experimental Investigation

Entropy noise caused by combustors increases rapidly with rising Mach number in the nozzle downstream of the combustion chamber. This is experimentally shown with a dedicated test facility, in which entropy waves are generated in a controlled way by unsteady electrical heating of fine platinum wires immersed in the flow. Downstream of the heating module called entropy wave generator (EWG), the pipe flow is accelerated through a convergent-divergent nozzle with a maximum Mach number of 1.2 downstream of the nozzle throat. Parameters like mass flux of the flow, nozzle Mach number, amount of heating energy, excitation mode (periodic, pulsed, or continuously), and propagation length between EWG and nozzle have been varied for the analysis of the generated entropy noise. The results are compared with the results of a one-dimensional theory found in early literature.

Acoustic PIV: Measurement of the acoustic particle velocity using synchronized PIV‐technique

This paper outlines a technique for measuring the acoustic particle velocity and the flow field simultaneously by applying synchronized particle image velocimetry (PIV). As test set‐up a squared acrylic glass chamber was chosen. One side of the test section is connected to a loudspeaker, which allows a sinusoidal excitation of the chamber. To point out constrains of this method the investigation includes an analysis of excitation amplitude and frequency as well as the effect of the mean flow magnitude. Therefore a small PC fan can be mounted inside the test section to produce an adjustable mean flow. It can be shown that for a low number of averaged images (80) reasonable results can be achieved up to a certain level of fan rotation speed. Beyond this level the turbulence sensitivity increases and more images are necessary for the calculations. However, the acoustic particle velocity can be computed in the presence of turbulent flow. The presented method called acoustic PIV is a non intrusive technique, a...

Indirect Combustion Noise: Investigations of Noise Generated by the Acceleration of Flow Inhomogeneities

The entropy noise mechanism was experimentally investigated under clearly defined flow and boundary conditions on a dedicated test setup. Previous experimental research on the topic of entropy noise could draw only indirect conclusions on the existence of entropy noise due to the complexity of the physical mechanism. In order to reduce this complexity, a reference test rig has been set up within this work. In this test rig well controlled entropy waves were generated by electrical heating. The noise emission of the entropy waves accelerated in an adjacent nozzle flow was measured accurately and therewith a qualitative and quantitative determination of the entropy noise source mechanism could be experimentally accomplished. In addition to this, a parametric study on the quantities relevant for entropy noise was conducted. The results were compared to a one-dimensional theory by Marble & Candel. In a next step investigations on a combustor test rig showed a broadband noise generation mechanism in the frequency range between 1 and 3.2 kHz. The combustor rig was set up with a similar outlet-nozzle geometry like the reference test rig (EWG) and provided therefore outlet-boundary conditions like in real-scale aero-engines (outlet Mach number = 1.0). It was found that this broadband noise has a strong dependency on the nozzle Mach number in the combustor outlet. The summed-up broadband sound pressure level increases exponentially with the nozzle Mach number. However, investigations of comparable cold flow conditions did not show this behavior. Since the results of the reference experiment with artificially generated entropy waves did not show this exponential increase with the nozzle Mach number, this leaves the conclusion, that this additional noise is generated by the interaction of small-scale fluctuations, e.g. in entropy or vorticity, with the turbulent nozzle flow in the combustion chamber outlet nozzle.

Impact of Defects and Damage in Aircraft Engines on the Exhaust Jet

The inspection of aero engines is a complex and timeconsuming process, often requiring the disassembling of the engine or boroscopic examinations. The development of a method to locate and characterize defects and damage at an early stage, without disassembling the engine would accelerate the inspection process. For that purpose, the spatial density distribution pattern of the exhaust jet of aircraft engines may be measured with the Background Oriented Schlieren method (BOS). The hypothesis is that defects in the hot gas path have a noticeable impact on the density pattern of the exhaust jet. To establish the connection between defects and measurable patterns, in the present paper numerical simulations of an aero engine are performed including three potential defects. Non-uniformities resulting from a burner malfunction, the increase of the radial gap between blade tip and casing as well as burned trailing edges are propagated with only small degree of dispersion through the turbine and reach the engine exit. The paper shows that each considered defect results in a different exhaust density pattern.

Optical measurement of acoustic pressure amplitudes—at the sensitivity limits of Rayleigh scattering

Rayleigh scattering is a measurement technique applicable for the determination of density distributions in various technical or natural flows. The current sensitivity limits of the Rayleigh scattering technique were investigated experimentally. It is shown that it is possible to measure density oscillations caused by acoustic pressure oscillations noninvasively and directly. Acoustical standing waves in a rectangular duct were investigated using Rayleigh scattering and compared to microphone measurements. The comparison showed a sensitivity of the Rayleigh scattering technique of 75 Pa (7·10(-4) kg/m(3)) and a precision of 14 Pa (1·10(-4) kg/m(3)). Therefore, it was also shown that Rayleigh scattering is applicable for acoustic measurements.

Numerical Study of Incidence Angles and Gap Heights in Turbine Cascades and Rotors on Tip Clearance Losses

The design and optimization of turbines demands the use of fast low-fidelity tools. To obtain adequate results, loss correlations simplifying the complex turbine throughflow are implemented. Accounting for modern turbine designs and flow conditions, revisions of profile and secondary loss correlations were primarily focused upon, while improvements of the tip clearance loss correlations are difficult to achieve. Realistic enginelike conditions concerning variations of the tip clearance, blade loading and solidity are time- and cost-intensive to investigate. This paper is focused on an extensive numerical study, intending to support experiments on tip clearance loss correlations. The losses of a high pressure axial turbine rotor are analyzed for different tip clearance gap heights and incidence angles at cruise condition. The results are contrasted with a cascade having comparable tip profile and gap heights. The cascade’s flow is comparable to the rotor, but with respect to experimental restrictions concerning inlet and outlet conditions. Steady 3D calculations in the stationary and rotating frame were performed applying DLR’s turbomachinery CFD code TRACE using Menter’s SST k-w turbulence model. The tip clearance loss coefficients were extracted from the flow field by post-processing data of an outlet plane and as massflow averaged global values. The findings are discussed referencing previous publications about the leakage flow system and tip clearance loss. Finally, a comparison to results from tip clearance loss correlations of Ainley-Mathieson and Dunham-Came is presented.

Experimental Investigations on Cooling Air Ejection at a Straight Turbine Cascade Using PIV and QLS

Due to the high turbine inlet temperatures in modern aircraft engines the adoption of several cooling techniques in the first turbine blade rows is state of the art. For this reason the influence of cooling air ejection on the main flow is in the interest of scientists. In this paper experimental and numerical investigations on the trailing edge cooling air ejection at a stator profile are presented. All measurements are performed at the Straight Cascade Wind tunnel Gottingen. To verify the influence of the cooling air flow on the flow field, the velocity field is measured by Particle Image Velocimetry (PIV). The development of the cooling air concentration is analyzed by utilizing the Quantitative Light Sheet (QLS) technique. For validation purposes the QLS results are compared to CO2 concentration measurements. Both measurement techniques are in good agreement with each other. One of the most important advantages of PIV and QLS is the possibility of combining them at the same test bed due to the identical experimental setup. The experimental investigations are supported by numerical simulations based on the numerical code TRACE. Both the numerical results as well as the experimental results prove the reduction of the trailing edge shock by increasing the coolant mass flow ratio.Copyright

Wavelet Analysis of Vortical Structures in Turbomachinery Applied to PIV Data

In this paper the unsteady flow field of a highly loaded turbine rotor blade at different transonic Mach numbers by means of Particle Image Velocimetry (PIV) is investigated. The experiments are performed using the linear cascade wind tunnel at DLR Gottingen at isentropic exit Mach numbers 0.7, 0.9, and 1.1. The process of vortex shedding at the trailing edge of the blade and its proceeding downstream in the wake area is analyzed using a wavelet based vortex identification algorithm. The wavelet algorithm automatically detects vortices and extracts e.g. their position, their vorticity, and their convection velocity. This procedure enables a statistical analysis on the basis of numerous PIV recordings. The vortex production, the downstream transport and the vortex decay are observed successfully. The Strouhal number and the shedding frequency are directly derived from the vortex position and confirm previous investigations. The PIV combined with a wavelet based analysis constitutes a remarkable insight to time dependent flow structures.Copyright