Report

Teaser, summary, work performed and final results

Periodic Reporting for period 2 - INSPiRE (Industrialisation of Jet Noise Prediction Methods)

Teaser

Jet noise is a significant component of overall aircraft noise and has a detrimental societal impact as air transport continues to grow. Although pure jet noise has successfully been reduced by increasing engine diameters, this has unfortunately led to a strong increase in the...

Summary

Jet noise is a significant component of overall aircraft noise and has a detrimental societal impact as air transport continues to grow. Although pure jet noise has successfully been reduced by increasing engine diameters, this has unfortunately led to a strong increase in the interaction noise caused by the jet\'s proximity to the wing and flap. The development of countermeasures is currently held back by a lack of understanding of the interaction mechanism, which could be delivered by high-fidelity turbulence-resolving simulations. The industrial exploitation of such simulation technologies is however prevented by their high computational expense and the lack of industrialised simulation processes for such complex problems.

The INSPiRE project contributes to the industrialisation of jet noise prediction technology by driving forward the development of scale-resolving methods. The focus is on a family of turbulence-resolving approaches known as Detached-Eddy Simulation (DES), in which the jet plume region is resolved whereas the attached boundary layers are captured more efficiently with lower-order models. The propagation of acoustic signals to far-field observers is accomplished in a hybrid approach for maximum efficiency.

The specific objectives of the INSPiRE project are:

- To enable the simulation of higher frequencies at minimal additional computational overhead through the extension of an innovative meshing strategy
- To improve the flexibility and reliability of far-field noise integration by pioneering the automatic detection of optimal FWH data surface placement
- To define and validate procedures to automatically demarcate the initial transient and statistically-steady states of the solution and to define statistical error bars on flow and far-field noise quantities using innovative in-house algorithms
- To achieve at least a 20% reduction of computational effort for jet noise simulations via accelerated initialisation techniques
- To validate the enhanced DES/FWH process chain against a database of existing measurements including complex jet-flap configurations with multiple flap angles
- To establish, document and verify best practice guidelines corresponding to the newly-developed methodologies
- To ensure direct exploitation of the developed technologies by conducting validation simulations and implementing applicable improved methods in an industrial CFD solver

Work performed

\"Development of enhanced methods was carried out in WP2, in which the methods were validated for a simplied static, single-stream round jet test case. WP2 was organised into four main tasks:

1. \"\"Hybrid structured/unstructured meshing strategy\"\":
An enhanced hybrid meshing strategy achieved a factor four increase in the azimuthal resolution near the nozzle with only 30% increase in CPU cost per iteration. Clear improvement in the acoustic prediction was achieved, with spectra at the peak observer angle predicted to within 1dB up to a Strouhal number of 3.

2. \"\"Efficient and flexible far-field noise integration\"\":
A novel sensor function was developed to automatically determine the optimal FWH data surfaces location. The approach makes a significant contribution to industrialisation, since it maximises computational efficiency whilst minimising user burden.

3. \"\"Statistical evaluation of simulation progress\"\":
A procedure has been validated for the automatic statistical evaluation of DES/FWH simulations of jet noise. Novel statistical algorithms were found to give reliable detection of the \"\"initial transient\"\" phase of the simulation. The method enables the optimisation of computational resources and reduction of user burden. The definition of statistical error bars on far-field directivity plots is an important additional benefit, e.g. to discriminate true noise differences from statistical error when judging competing designs.

4. \"\"Initial transient acceleration techniques\"\":
Two separate methodologies were investigated to reduce computational wastage in the initial transient computation. A more efficient simulation process involving lower-fidelity computation of the initial transient combined with optimised computational settings for the productive statistical portion resulted in a factor 7.6 reduction of computational expense compared to previous best practice settings. This greatly exceeds the project objective of a 20% efficiency gain.

In the second main work package of the project, WP3, the industrial feasibility of the enhanced methods has been assessed by applying them to a complex configuration consisting of a short cowl, coaxial jet mounted to a wing/flap/fuselage model via a pylon. Two different flap angles and two different flight Mach numbers were simulated. A time-saving approach was pioneered whereby a single mesh was designed for the simulation of multiple flight speeds. The simulations were carried out in parallel on a cluster of 288 CPU cores, which is considered representative of current industrial resources. In comparison of the simulation results with measurement data, the effects of flap angle and flight speed were correctly captured and the level of absolute agreement was encouraging.

Dissemination of the project results to the scientific and industrial communities was carried out via the publication of two scientific papers and the organisation of a mid-term workshop. Exploitation of the project results has been ensured by the implemention of methods directly in an industrial solver and the delivery of accompanying best practice guidelines.\"

Final results

For each of the four method enhancements developed, the state of the art before and after the INSPiRE project is compared:

1. Hybrid structured/unstructured meshing:
- BEFORE: In-house methods pioneered, including techniques for smooth structured/unstructured interfacing
- AFTER: Extension of meshing strategy to include local azimuthal refinement in shear layers

2. Efficient & flexible far-field noise integration:
- BEFORE: Parallel FWH implementation and validated method to (manually) define data surface locations irrespective of mesh
- AFTER: Sensor to automatically detect optimal FWH placement for arbitrarily complex cases

3. Statistical evaluation of simulation progress:
- BEFORE: In-house methods for transient detection and statistical error quantification; Efficient implementation of algorithms in software tool meancalc
- AFTER: Validation of meancalc specifically for jet flows and derivation of associated best practice (monitor signals and quantities to evaluate)

4. Initial transient acceleration techniques:
- BEFORE: Methods to increase efficiency of transient simulation not yet attempted
- AFTER: Validation of optimal low-fidelity initialisation approach in industrial process, achieving factor 7.6 reduction in overall computational expense

These approaches were firstly validated for a simple single-stream jet case. In the second half of the project, the approaches were successfully applied to complex installed coaxial jet/wing interaction cases with a variety of flight conditions and flap settings.

By making a tangible contribution to the industrialisation of high-fidelity jet noise prediction methods, the INSPiRE project has therefore delivered a significant contribution to the challenging goals set by the renewed ACARE Strategic Research and Innovation Agenda, namely a 65% reduction in perceived noise by 2050 compared to 2000 levels.

Website & more info

More info: http://www.cfd-berlin.com/.