Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - ADAPT (ADvanced Aeroacoustic Processing Techniques)

Teaser

Airbus is conducting various experimental studies, dealing with installing microphones on the fuselage, on the wall of a wind tunnel test section or on specific microphone supports in the flow, in order to record the sounds emitted by each aircraft component (i.e. engine, wing...

Summary

Airbus is conducting various experimental studies, dealing with installing microphones on the fuselage, on the wall of a wind tunnel test section or on specific microphone supports in the flow, in order to record the sounds emitted by each aircraft component (i.e. engine, wing etc.). In either case, microphone measurements are polluted by hydrodynamic noise, which hampers the correct assessment of acoustic signals emitted from aircraft components, thus leading to very low signal-to-noise ratio.
A few methods exist today that can eliminate hydrodynamic noise from acoustic signals. However, the efficacy of these methods strongly depends on various parameters, such as signal-to-noise-ratio, number of microphones, number of incoherent sources etc. No systematic study has been performed so far to assess the parameter boundaries within which acceptable results are obtained.
The ADAPT project aims at improving and optimizing the ability of identifying sources emitted by aircraft components by using the most effective techniques out of the three proposed approaches in the call for proposal :
• aeroacoustic source separation,
• de-noising techniques based on cyclostationarity
• aeroacoustic sources localization.
Through these three approaches, the ADAPT project aims at delivering to AIRBUS tools dedicated to separating airframe and engine noise components (tonal, broadband cyclostationary components in particular) from hydrodynamic pressure noise and identifying these sources in space, satisfying various technical and economical requirements that has been discussed out with AIRBUS at the beginning of the project.

Work performed

The first part of the work has been conducted by PSA3 aiming at providing expert and reference revue in the field of the extraction of an acoustic component from hydrodynamic component. Thus, PSA3 made an adapted version of CLEAN-SC available to the partners as a baseline solution for separating hydrodynamic noise from acoustic signals. For testing purposes, a range of benchmark solutions was created, featuring synthesized microphone array data at various signal-to-noise ratios. The acoustic signals were obtained with 5 point sources and with a line source, at various distances from the array. These benchmark solutions were also made available to the partners. Using the benchmark solutions, the merits of CLEAN-SC beamforming were assessed and compared with recently published methods for CSM reconstruction, including INSA’s new “Probabilistic Factor Analysis” (PFA) method. It was demonstrated that PFA renders high-quality denoised data, when the number of sources is not too high. Beamforming can yield even better results, but then the source positions need to be known. This was done in co-operation with INSA. A paper has been presented at the 25th AIAA/CEAS Aeroacoustics Conference (Delft, The Netherlands, 20-23 May 2019).

Tests in aeroacousitcs wind tunel was conducted in the main subsonic wind tunnel at Ecole Centrale de Lyon, LMFA. Different test configurations have been realized, for varying flow speeds from 10 to 50m/s, and for different acoustic source configurations. The consortium gets now the necessary database for academic validation of developed methods.

The case of two contra-rotating rotors is complicated in terms of tonal noise generation process (many interferential phenomena). INSA-LVA developed specifically a method to extract such tonal noise and applied this method to CROR data provided by Airbus. The performance of the method regarding this complex tonal noise extraction has been assessed. Results obtained on Airbus data (CROR) are quite good and Airbus can use these methods.

MicrodB starts working on denoising and source coherence length estimation using the bayesian focusing framework. The preliminary obtained results are encouraging.

Final results

The key point of the project, or its phylosphy, is to integrate apriori knowledge on aeroacoustics phenomena into models and signal processing techniques. Thus the development and the validation of approaches to de-noise and separate acoustic sources will help to refine aeroacoustic source identification and make the aircraft noise reduction a more efficient task for flight and/or wind tunnel test. These techniques will be integrated in innovative acoustic measurement products and applied through industrial services. ADAPT will finally impact a large range of industries at several levels (from OEM to suppliers) and will contribute to maintaining European competitiveness on a topic that is directly related to human comfort.

Website & more info

More info: https://cs2adapt754881.frama.site/.