CORDIS - EU research results

Research on Core Noise Reduction

Final Report Summary - RECORD (Research on Core Noise Reduction)

Executive Summary:
The RECORD project (Research on Core Noise Reduction) was investigating the core noise generation mechanisms in aero-engines which are caused by the combustor and the combustor-turbine interaction.

RECORD integrated contributions from internationally renowned organisations from across Europe. These included leading engine manufacturers, with expertise in combustor design and manufacture, leading research establishments and universities. Care had been taken to choose the best test benches from across Europe in order to carry out the best, uncompromising validation work. This teaming of partners and resources brought together the multi-disciplinary mix of expertise that was required for successful development, validation and exploitation of methods for combustion noise prediction and control. This approach yield improved working practices and data exchange across the European research and industrial supply chain and reduce future product development lead times. Industrial partners also focused their efforts on further exploitation of existing experimental results from previous EU-funded projects (as SILENCE(R), Friendcopter, TEENI, ...). These results had increased relevance in the context of the new findings of RECORD, maximising and bringing forward the return on investment made on core noise as a research topic.

Within RECORD an integrated approach was pursued in order to thoroughly understand direct and indirect noise sources from the combustor and their interaction with the turbine. Various numerical models from state-of-the-art LES simulations to low-order physical models were applied to the inhomogeneous reacting flow in a combustion chamber in order to identify sources for direct and indirect noise. Most of these methods were developed by combustion specialists, and some of them were used with a focus on acoustics for the first time. Dedicated experiments were performed in order to validate the results of these numerical tools and enable further development of the tools based on the new data. Finally, the assessment of the achievements of RECORD was made by the comparison of these full-scale combustion noise calculations with existing real engine data of the engine manufactures obtained outside the proposed project.

RECORD was a step towards the necessary supplementary noise reductions. Through better modelling of this very complex noise source, a better understanding of important parameters now allows relevant components – combustor, NGV, turbine – to be designed for reduced noise. RECORD provided tools that allow trade-offs between noise and emissions to be calculated and understood, and allow low-noise technologies to be developed. Working this way will ensure that meeting ACARE objectives on noise does not compromise ACARE objectives on emissions.

Project Context and Objectives:
Project Overview
In order to achieve the greening of the European air transport with the deployment of low emission and low noise propulsion systems the reduction of core noise, the noise related to the combustion system, plays an important role. The ability to design low core noise aero-engines requires the development of reliable prediction tools. This development demands extensive research with dedicated experimental test cases and sophisticated numerical and analytical modelling work to broaden the physical understanding of core noise generation mechanisms.

One major challenge for the greening of air transport is the currently observed noise penalty of low pollutant emission combustion systems in aero-engines. All sources of noise in an engine contribute to the overall emission and combustion noise was not addressed for a long time, since it was not a dominant source. However, due to reductions in fan, turbine and jet noise in modern turbofan engines, combustion noise became a significant contributor now. Already back in the late 1970’s it was predicted, that indirect and direct combustion noise becomes important, if highly loaded turbines or lean combustion systems are used. Both are currently key technologies for next generations of green low-emission and ultra-high-bypass ratio turbofan engines.
For the smaller aero-engines in small and mid range aircraft, the cross-over of combustion and jet noise to the overall engine noise contribution has already been reached without “green” technologies. In small engines, the core noise at low frequencies and jet noise at very low frequency are the main sources in the sideward arc of the engine. In larger engines the low frequency core noise is dominant at sideward and backward arc. It is hardly attenuated in air and by structures and thus affects cabin comfort and a large public as well. With “green” technologies, reduced jet noise and increased core change the importance order for these sources in the rearward arc as well.
Aside from airplane engines, core noise effect is even more dramatic for helicopter engine. Turboshaft engines transform the remaining energy in the hot-gas stream into mechanical power, leading to much lower exhaust speeds. Little - if any - jet noise is generated and turbine noise is generally very high-frequency, so most of the exhaust noise in the audible range is linked with the combustion process. The Friendcopter PCRD6 IP project has demonstrated in flight on an EC-135 helicopter up to 1.3EPNdB noise reduction at take-off conditions while using treated ejectors mounted on both ARRIUS 2B2 engines. This sensitivity to engine exhaust noise underlines the importance of this noise source for take-off, with current technology rotors. The relative importance of engine exhaust noise will even rise, as helicopter manufacturers are doing their best to reduce helicopter rotors noise. Should noise regulation become more stringent for helicopter take-off condition, and, to a lesser extent, cruise condition, the engine core noise would be the dominant noise source to be reduced in priority. Engine manufacturers have to anticipate this trend by proposing the most adapted liners, or favourable engine architectures. But existing design tools are not really adapted to accomplish this challenge.
Because it took time for core noise to become a concern for Turbojet engines, and because of the complexity of the physical mechanisms involved, combustion-related noise was far less studied than fan or jet noise for the last decades. Now that this noise source becomes critical, and thanks to the huge progress which have been made on numerical methods, a project such as RECORD becomes feasible and meaningful.
Therefore, research is necessary to understand the core noise generation mechanism. Based on this physical understanding innovative noise reduction technologies can be developed, that address the noise at the source. This will allow minimising the associated impact on cost and weight. These investigations will therefore help to reduce the fuel consumption of green, low-pollutant emission engines as well. The correct prediction, understanding and reduction of combustion noise generated directly and indirectly from the engine core is consequently an important research topic within the next few years to maintain the competitiveness of European products with the introduction of low pollutant emission combustion systems. Finally, this research will also provide the ability of engine manufactures to meet new noise regulations. Currently, most of the knowledge is situated in research institutions and universities. With this knowledge from basic research and partners from the industries in need of such knowledge, a technical breakthrough is achievable. RECORD was designed to allow this breakthrough for the European industries through continuous research funding in all stages.

For the reduction of core noise in prospective aero-engine generations the RECORD project pursued two main objectives:

1. The improvement and validation of core noise modelling and prediction methods through carefully specified experimental means concerning:
- the generation of direct combustion noise,
- the generation of indirect combustion noise, and
- the transmission of combustion noise through a turbine.

2. The development and testing of core noise reduction methods.

In summary, RECORD enhanced the understanding of noise generating mechanism and its propagation taking the interaction of combustor and turbine into account. The importance of direct and indirect noise was analysed. Through carefully designed experiments and extensive numerical calculations based on recent high-fidelity combustion solvers, the numerical methods and assumptions were validated and extended. Low-order fast models were also be developed to provide a quick approach for the ”noise design” of combustors and subsequent turbine stages while the more time-consuming and expensive LES calculation provided a more detailed picture of the flow physics. This new knowledge will lead to more efficient, and probably innovative, ways to reduce combustion and core noise.

Project Results:
See attachment "RECORD_Main_S_and_T_results-foregrounds_v3_final.pdf".

Potential Impact:
The experimental investigations in the RECORD project delivered valuable information, needed to understand the generation and transmission of combustor-turbine interaction noise.
The numerical noise simulation and prediction techniques enhanced in RECORD will help to support the aircraft industry, in particular aero engine manufacturers, with respect to the development of a new generation of quiet aero-engines.
In RECORD also the potential of improved noise reduction concepts with respect to combustor/turbine interaction noise could be identified and evaluated.
Last not least RECORD provides DLR with a very comprehensive data base with respect to core noise generation and transmission for the purposes of current and future validation.

The RECORD project represented a good opportunity to deeply understand and quantify the core noise in modern aero-engines. The numerical methodologies adopted for the assessment of direct and indirect combustion noise (low frequency acoustic disturbances and noise generation due to entropy /vorticity perturbations interacting with the turbine) provided good agreements with the project experimental data and confirmed the possibility to be further developed and exploited for the prediction of the combustor-turbine acoustic interaction in future programs. Finally, the experimental tests within the project enabled the extension to transonic HPT configurations of the turbine tone noise prediction capabilities, i.e. LRANS and URANS-based approaches.

During RECORD, CERFACS has developed a complete simulation chain called CONOCHAIN. CONOCHAIN is a fully deterministic simulation tool for combustion noise in gas turbines. It does not use any correlation and has been shown to produce reasonable sound levels, especially for the engine case of RECORD (the TEENI engine). This opens exploitation possibilities since predicting combustion noise is becoming a major issue for many companies worldwide. In addition, the publications associated to the work have contributed to increasing the visibility of CERFACS’s work in collaboration with EM2C, Turbomeca, and DLR.

The experimental activities carried out at CNRS/EM2C have been successfully completed and several results have been achieved. The CESAM-HP test rig suitable for combustion noise measurements is now available and running. CNRS/EM2C carried out the experiments and supplied to all the partners the database for combustor noise source determination, i.e. direct and indirect noise (entropy noise). Measurements of temperature (steady & unsteady) in the combustion chamber were made possible by innovative methods and using complementary computational data. Post-processing for dynamical analysis could be performed to successfully extract direct and indirect noise contributions. It was shown that the contribution from indirect combustion noise existed inside the combustor, and could be estimated, but the levels were found not to be the dominant noise source. All these results can be made available to researchers working on acoustic propagation in reactive media for CFD as well as analytical codes validation. The potential impact of this activity is wide and can drive further research in the field of combustion noise. Apart from the PhD grant funded in the framework of RECORD, one post-doctoral researcher (18 month) and one PhD student could be hired on this subject as a follow-up of RECORD. A total of 6 persons (3 permanent and 3 non-permanent researchers) are now working on the combustion noise at CNRS/EM2C. Finally, several journal/conference papers and presentations have been completed or are under completion, in cooperation with other partners from Work Package 2, in particular for the next ASME Turbo Expo Conference and International Symposium on Combustion, the two major conferences in the field.

Free Field Technologies (FFT):
Actran is an acoustic solver used by several engine manufacturers in the analysis and design of aircraft engine noise. Typically the excitation consists of a set of imposed duct modes, injected in the engine exhaust after the last turbine stage, but it is difficult to predict the intensity of these imposed modes. In RECORD, a new excitation mechanism has been introduced in Actran, based on low order models describing the conversion of entropy into acoustic waves. Although simplified, the new excitation can be derived from known entropy fluctuation and thus help predict the absolute value of engine exhaust noise. These models have been tested and validated in the framework of RECORD by comparison with very detailed and high quality measurements. The validated models are going to be released in 2016 and probably improved by additional physics in the next releases, depending on inputs from Actran’s users. The models are extending the capabilities of Actran to combustion noise mechanisms. Additional relevant discoveries in RECORD by other partners (flame transfer function, technology maturity level) are also indicating possible future improvements.

Industria de Turbo Propulsores (ITP):
The participation in the RECORD Program has been very useful for ITP. On the one hand, the validation of Mu2s2t-L for self-tonal noise prediction in turbomachinery has been broadened with a stator-rotor interaction of a transonic HPT. The results obtained are consistent with the measurements as well as with the other partners’ results. On the other hand, it is the first time this code has been used with the imposition of perturbations at very low frequency. In future, this will be an essential step towards the combustion noise modeling via CFD. It has been demonstrated that the code is able to work at this very low frequency, with the imposition of perturbations of vorticity and entropy in the HPT that generate indirect noise consistent with the measurements taken at PoliMi. Together with the CFD verification for HPT self-noise and direct/indirect combustion noise, this Program has generated a valuable representative experimental database for future research programmes. Future work will be related to both CFD and experimental capabilities. Regarding CFD, some issues have been found when running at very low frequency, and these have been partially addressed by the partners during the RECORD Program, although more exploration could be recommended in the future. With regard to the experimental measurements, some issues have been also identified. For instance, similar levels of incoming and outgoing waves in the microphone rings have been measured at certain conditions, suggesting the presence of acoustic reflections in the rig. The use of acoustic lining in the installation could improve the quality and reliability of the measurements.

Polytechnic University of Madrid (UPM):
The noise computational results (in particular those on WP3) from RECORD will be used to set-up and check new and existing codes for noise evaluation in rotor/stator configurations. In particular, the simplified methodology for low frequency noise propagation developed by UPM (Task 3.11) will be also useful for improving existing noise codes in the low frequency regime. The whole RECORD results (both, experimental and numeric) constitute an extremely valuable dataset that will be used as realistic examples for turbomachinery aeroacoustics lectures. The RECORD project increased the possibility of potential future collaborations with RECORD partners on new research projects.
The dissemination and exploitation list include the following
• Two UPM graduate students worked in the RECORD project, and benefited from the training in turbomachinery aeroacoustics.
• Two contributions to the ASME TurboExpo conferences of 2015 and 2016: GT2015-42376 and GT2016-57788.
• A description of the RECORD project results will appear in the official UPM research web page.
• The RECORD project will be presented to the students in graduate and undergraduate lectures.
• RECORD results will be used in future scientific publications, and in MSc and PhD level courses.

ONERA contributed to WP1 and WP2. Despite note being officially involved in WP1, ONERA showed a particular interest in the experimental results provided in the Work Package. During the recent years, ONERA developed an analytical model to evaluate the noise generated by the acceleration of acoustic or entropy perturbations through a nozzle. The HAT experimental results are the first data used to validate the model with comparisons to experiments. Both temperature and Mach number effects are recovered analytically with a very good agreement for acoustic forcing. Limitations of the experiments with entropy forcing have been highlighted and additional efforts are planned to be performed in 2016 to validate the model for entropy-generated noise. Concerning WP2, ONERA performed compressible HFLES computations which were found in good agreement with experimental data. The simulation results were also used to analyze and optimize the CESAM-HP experiment operation which demonstrated the maturity of the code. From the acoustics point of view, the work performed in WP2 helped us validate the post-processing procedure to extract the noise from the CFD simulation, determine analytically the noise generated through the nozzle and the relative contributions of direct and indirect noise. Validations have been performed with comparisons to experimental data. This post-processing methodology is the basis of the acoustic works planned in the future for the analytical modelling of a full engine, including the turbine stages. The RECORD project provided good visibility of ONERA tools and capability in addressing the combustion noise issues for full engine. It also favored cooperation with research teams on the subject, with one common article being prepared and submitted. This should open the possibility of future works on collaborative projects.

Polytechnic University of Milan (PoliMi):
The activities foreseen at PoliMi have been successfully completed and several results have been achieved. First, a test rig suitable for acoustic tests is now available and running with prescribed stage inlet disturbances. Second, new measurement techniques have been developed and now available for future experimental campaign and EU-financed projects. Finally, a huge database on aerodynamic measurements, upstream of the stage with and without disturbances as well as downstream of the stator and rotor, has been made available to the consortium; moreover, due to the nature of the project, the results can be made available to all researchers on aero-acoustics for CFD and analytical codes validation. The potential impact of the results is wide and can drive further researches beyond the state-of-the-art in the field of aero-acoustics and aerodynamics. In fact, the test rig can be run, with minor modifications, for studying flow conditions and inlet disturbances representative of real engine flows. The same rig can be exploited for research projects on the performance of a turbine stage operated under highly unsteady inlet flow conditions. Modern and high level expertise in unsteady temperature and pressure measurements make the PoliMi team a reliable partner for future investigation. The free of restriction stage geometry and database will allow for a wide dissemination of the benchmark gained in the RECORD project. The availability of the database would also allow for validating new codes and for further developing existing ones, including optimization codes for new and quieter generation of HP stages. According to the mission of the University, master students have been involved in the research project, thus fulfilling a valuable step of the dissemination process: results have been presented in master thesis committee and made available to other professors at PoliMi. Furthermore, in cooperation with other partner, a number of journal/conference papers and presentations have been completed or are under completion.

Methodologies and methods have been developed within the work-package 2.3 and these will be introduced into the RRD design process of prospective combustors. The CFD/CAA methods developed in RECORD have increased the internal prediction capability for noise emissions and noise generation in aero-engines, and will be further optimized, leading to a faster design-iteration process. As WP2.3 is based on academic issues with strong industry implication, the results of the research will be published in national and international conferences (e.g. ASME/IGTI Conference –paper in preparation) and in archival journals (also paper in preparation) on combustion and acoustics. In addition, more publications are still in preparation; the dissemination of these within the consortium has been handled through regular meetings and reports. As TUD is a part of the RR-UTC (University Technology Centre), there was thus a clear path for particular dissemination and implementation by RRD, even though there was still a long way for these recommended numerical simulations to be included in the company’s design process for noise reductions. The cooperation with industry partners also results in a fortunate situation where researchers apply state-of-the-art techniques to state-of-the-art technology and help benefit upcoming projects with the latest understanding.

The RECORD project has contributed significantly to the understanding of combustion noise generation and propagation through the turbines, and the relative importance of the two sources of combustion noise in the total combustion noise radiated from the engine. The various methodologies to predict the heat release rate (produce direct combustion noise) in a model combustor was very informative with regards to predicting the levels in a real combustor. In particular the empirical method of Hirch’s et. al. for heat release rate in combustor has provided a quick methodology for industrial use. The various numerical approaches, i.e. RANS/LES/HFLES, carried out in RECORD for predicting combustion noise sources gave insight into the limitations of these approaches in resolving the various features in the combustor flow field. The analytical tool developed as a part of WP4 by UCAM has provided RR PLC a validated tool to predict combustion noise generated by large civil turbofan engines, and to discriminate the different sources of combustion noise. As RRPLC’s product ranges from Large to Small engines, the insight provided by the deliverables of WP4 helped understand the relative importance of the different sources (i.e. direct and indirect) of combustion noise for various types of aero-engines. The RECORD project also helped RRPLC understand the various expertises within Europe for combustion noise research, and this information may be helpful for future collaborative projects.

SMCPFA designed a new type of Entropy Wave Generator (EWG) based on fast alternation of hot-cold spots. This EWG installed on new designed PoliMi test rig permitted creating of a huge database on aerodynamic measurements, upstream and downstream of a turbine stage with and without disturbances which has been made available to the consortium. These results are available to all researchers on aero-acoustics for CFD and analytical codes validation. On the other hand, the EWG drawings and manuals are available for all European research institutes and universities working in aeroacoustic research. EWG is still used now for various experiments at PoliMi and will be available for future experiments inside other EU-financed projects. It is possible as a more powerful EWG based on the same principle to be designed and manufactured in future for other advanced research. The potential impact of EWG working on this new principle is wide and can drive further researches beyond the state-of-the-art in the field of aero-acoustics and aerodynamics. This equipment is very useful for studying of influence of inlet disturbances on real flow in turbines. High level expertise gained by SMCPFA in creating of controlled unsteady temperature and pressure perturbations makes SMCPFA a trustable partner for participation in future research projects where such equipment is necessary. It is possible as EWG principle and technological experience accumulated during manufacturing to be used in other technical fields where high frequency temperature and pressure perturbations are necessary.

Although combustion noise is usually not the main issue in aircraft acoustic certification process, it is however the main noise source during idle or aircraft taxiing operations. As a consequence, it is a significant contributor to the overall airport noise. Moreover, the current tendency to increase the by-pass ratio and the efforts on fan design will drastically decrease the jet and the fan noise during in flight aircraft operation. Therefore combustion noise has to be looked carefully not to become a major noise contributor of the next generation of engines. Collaboration between academic and industrial research is needed to understand clearly the mechanisms responsible for combustion noise in turbofans. Numerical tools development allowing engineers to design low combustion noise engine is another challenge of this collaboration. The RECORD project has permitted a better understanding of the combustion noise mechanisms. However, the dominant mechanism (direct or indirect noise) is not yet clearly identified in a real turbofan engine. The lab experiments performed within RECORD have provided a large amount of experimental database for code validation. The numerical tools already used in Snecma combustion noise calculation chain have been compared for the first time to canonical experiments and therefore clearly validated. Moreover, this project was a great opportunity for Snecma to have an exhaustive overview of combustion noise, and to promote new collaborative opportunities.

The RECORD project has provided a new insight into the generation and propagation of combustion noise of Turboshaft engines, for which combustion noise is of prime importance. In particular, the work developed in WP3 and 4 will enable us to modify our 1D methods to take into account RECORD’s findings. This should prove valuable in improving the prediction accuracy of such methods, which are used to provide guarantees to our clients before the acoustic certification process. The CONOCHAIN methodology put together and validated in this project will be used for noise evaluation purpose of new engines, as it is based on the tools currently used for combustor design. The results from CONOCHAIN methodology will be compared to the new 1D methods output. Development of tools and know-how acquired in WP3, on CFD methods used for acoustics, will also be used as a basis for analysis and will be generalized for the silent compressor design project as well. The full analysis of the calculations and the possible comparisons to experimental data are not fully complete and will go on as a short term activity. This activity will complement our lessons learnt collection on aero-acoustic practices. Concerning combustion noise research, Turbomeca will actively involve in preparing experimental and theoretical work that will contribute to address the questions left open by RECORD.

Technische Universität München (TUM):
In the RECORD project TUM successfully demonstrated the capabilities of the hybrid RANS/LNSE methodology for the fast prediction of combustion noise of both premixed as well as non-premixed flames. Due to the low computational requirements the presented methodology is highly suitable for predesign tools for academic up to industrial applications, which is in particular relevant for GE AVIO. Hence, the TUM contributions have a substantial accelerating effect on the build-up of competence and computational methods as well as an impact on the reduction of the noise emission of future aero-engines. The numerical results of the indirect noise generation were presented and discussed on international conferences such as the AIAA/CEAS Aeroacoustics Conference and Internoise. The potential impact for reducing the computational efforts was highlighted. Furthermore, the main purposes and framework of the RECORD project was presented at Internoise in the dedicated session “EU projects on aircraft noise”. In addition, a broad set of RANS and acoustic data were generated with RECORD, and these can be used for further investigation and validation activities.

University of Cambridge (UCAM):
The RECORD project enabled UCAM to continue the development of their low-order codes LOTAN and LINEARB for linear waves within the combustor and turbomachinery blade-rows respectively. It has provided experimental data and comparison with high order methods (e.g. LES) that have enabled validation of these codes. The low-order methods have proved to be surprisingly accurate, performing even better than expected. The calculations are very quick, enabling multiple cases to be performed in seconds and so are appropriate for use in preliminary design. The RECORD project has identified that the modelling of the attenuation of entropy waves at engine conditions in the combustor and the first stages of the turbine needs to be improved. Further funding has been won in the EU Clean Sky programme which will enable data to be collected at engine-representative conditions and used for this further enhancement of the codes.

University of Florence (UniFi):
The RECORD project has improved the knowledge about aero-acoustic phenomena on direct and indirect noise analysis. The development and the extension of numerical solver could be achieved through validation against experimental data. The results from the RECORD project were disseminated in scientific papers, masters and PhD theses, and courses for students.

List of Websites: