Final Report Summary - NAA-CROR (Numerical aero-acoustic assessment of installed Counter Rotating Open Rotor (CROR) power plant)
Executive Summary:
During the course of this project, several improvements of the CFD/CAA system were pursued to ensure a professional, user-friendly and quality assured software environment applied to open rotors. The Non-Linear Harmonic (NLH) method, implemented in NUMECA software, is applied, providing a highly efficient method for a simultaneous prediction of the unsteady aerodynamics and the near-field acoustics of Counter-Rotating Open Rotors (CROR). Far-field acoustic propagation method (computed with a Ffowcs-Williams and Hawking solver) is also applied to NLH results. Several aspects of the software have been enhanced with respect to the meshing tool (AutoGrid5™), the CFD solver in FINE™/Turbo, the acoustic solver FINE™/Acoustics and its integration with FINE™/Turbo.
The numerical CFD/CAA approach has then been extensively applied to several CROR configurations (cruise, take-off, with/out incidence, with/out pylon, and with/out imposed transition). Both pre-test and post-test simulations have been run. When available, the results have been compared to those of other partners or to wind tunnel test data. The performed calculations prove that the NLH method is capable of accurately predicting CROR configurations at a very low computational cost (when compared to full unsteady calculations).
Project Context and Objectives:
The inherently high propulsive efficiency of advanced propellers and Counter-Rotating Open Rotors (CROR) have a great potential for fuel savings, but the level of noise emitted by the open blades represents a major obstacle to their environmental acceptance. Consequently, the design of a low noise, fuel efficient open rotor powerplant is one of the major objectives in the Clean Sky JTI. Within the virtual prototype design environment, the availability of highly efficient calculation procedures of noise sources and their propagation are essential to achieve the set objectives. While the CAA approach for the acoustic far-field noise propagation is well established, the critical issue remains the delivery of fast and accurate unsteady CFD-solutions for prediction of the noise sources. The present NAA-CROR project responds to this objective, through an advanced new approach for the CFD determination of the noise sources. The NAA-CROR project relies on the nonlinear harmonic method (NLH) which allows a gain in CPU performance for CROR’s compared to current CFD sliding grid or Chimera methodologies, of two to three orders of magnitude. This method, defined in the frequency domain, has been largely validated and successfully applied on multistage turbines and compressors at many companies. Its extension to propeller and CROR configurations has recently been achieved. We further extended this approach with the inclusion of the ability to capture installation effects for CROR configurations. The near-field and far-field noise are evaluated with an acoustic propagation module solving the Ffwocs Williams and Hawkings (FW-H) equations. The acoustic module is fully integrated with the NLH code, allowing a turnaround time for a complete CFD-CAA simulation of a few hours on a low number of processors.
Project Results:
WP 1 covers the development of the integrated CFD/CAA simulation system, capable of fully capturing installation effects. The code now allows for capturing the interaction between the pylon/nacelle and both rotors. The development required the propagation of inlet distortion perturbations through the second rotor and solving additional harmonic equations for the second rotor.
WP 2 deals with the validation of the numerical approach and the numerical simulations for selected CROR configurations. Extensive aero-acoustic calculations were performed, related to pre-test and post-test open rotor geometries, at model scale, for both take-off and cruise conditions. Results are compared with those of other partners or experimental data when available.
During pre-test calculations, NUMECA performed CFD computations for several CROR configurations (cruise, take-off, with/out incidence, with/out pylon, and with/out imposed transition). We investigated the effects of boundary layer transition for selected configurations. The aerodynamics and aeroacoustics of the engine were found to be weakly affected by the delayed transition for the selected configurations. All calculations simulate the flow by means of the non-linear harmonics method or NLH that was developed by NUMECA. The main idea is that the flow perturbations that make the flow unsteady are written about a time-averaged value of the flow and are Fourier decomposed in time. The user controls the accuracy of the unsteady solution through the order of the Fourier series. For the NLH method only one blade channel is required like a steady flow simulation, leading to a considerable gain in computational time compared to full unsteady calculations such as domain scaling techniques. The performed calculations prove that the NLH method is capable of accurately predicting CROR configurations at a low computational cost (when compared to full unsteady calculations). NUMECA performed an extended CAA analysis based on CFD results for the low-speed test cases. Acoustic results were obtained with several methods, among others with the NLH solution and with the FW-H solver. A detailed comparison and assessment of the acoustic results has been provided. We showed that FW-H computations evaluated with porous source surfaces generally provide more accurate results than those obtained with solid source surfaces since additional quadrupole sources are inherently captured within the porous FW-H surface.
Post-test aero-acoustic calculations have also been performed on the model scale counter-rotating open rotor, as tested in the S1 wind tunnel, for cruise conditions. Only one angle of attack is considered, which is 0°. For the CFD calculations, two blades geometries are considered according to the modeling of the root gap present in the experimental set-up or not. Partial hub gaps are easily modeled thanks to the dedicated AutoGrid™ capabilities. Modeling the hub partial gaps does not affect significantly the global performances. Acoustic calculations are performed on the geometry without gaps only, using the FW-H propagation module implemented in FINE™/Acoustics. The boundary conditions are directly coming from the CFD NLH calculation performed previously with FINE™/Turbo. Near-field acoustics are also evaluated using direct CFD results.
With respect to previous pre-test acoustic results obtained for take-off conditions, the FW-H permeable and non-permeable formulations provide quite different results. The predicted noise level computed considering only solid blades as acoustic sources is reduced compared to the use of a permeable surface. Whatever the location of the microphones line, FW-H Solid results show (compared to FW-H Porous results) an underestimation of about 5dB on the SPL peak value of the main rotor tones (1F and 1R) while a difference of 10dB is reached for the first harmonic (2F and 2R tones). The difference between FW-H Solid and FW-H Porous can be explained by the aero-acoustic sources located in the fluid region close to the blades, that are indeed not negligible for cruise conditions.
The analysis also shows that the original mesh used for aerodynamic performances is too coarse for near-field acoustics analysis. The CFD-based results obtained on a mesh that is refined in the radial direction above the blades show higher noise levels. On the first microphones line, the original CFD mesh gives similar peak results for the main rotor tones (1F and 1R), but already underestimates the SPL of about 2dB for the 2F and 2R tones. The difference between meshes even reaches 10dB for interaction tones. When the radius of the microphones line increases, the differences of results between meshes becomes larger. In the refined region, close to mid propeller plane, the noise levels obtained with CFD are quite close to the FW-H porous results. In order to avoid too much dissipation of the acoustic levels, the CFD mesh should therefore be enough refined in the region of interest. Around 15 points per wavelength (for the max. frequency) have to be used when the distance between the acoustic sources and the microphones lines is about 2 wavelengths. More points would be needed in case more wavelengths are considered : microphones located in the upstream and downstream regions. This is even more critical in the upstream direction where the Doppler effect is considerable for this high speed operating point.
Finally, WP3 aims at gathering the user specifications dedicated to the integrated CFD/CAA software system as applied to open rotors, in order to address these requirement in NUMECA software.
A specific wizard is now available in the turbomachinery-oriented structured grid generator AutoGrid5™ to automatically mesh contra-rotating open rotor (CROR) configurations. Both axial and radial far-field limits can easily be controlled. The flow paths and blade-to-blade mesh distribution can be set easily with a high-quality grid assurance. Current developments are also tackling rounded blade tips configuration with the introduction of a specific block topology at blade tips and a local three-dimensional smoothing.
The NLH methodology has been extended to fully account for installation effects of pylons/nacelles on the CROR aerodynamics and noise. The code now allows for capturing the interaction between the pylon/nacelle and both rotors. The development required the propagation of inlet distortion perturbations through the second rotor and solving additional harmonic equations for the second rotor. Before, the base NLH method could model only the effects between the adjacent rows, the rank-1 effects, because the three rows have three different rotation speeds (zero for the pylon and contra-rotating for the rotors). Now, besides the rank-1 effects, the extension makes possible the capture of the rank-2 effects.
The coupling between the CFD solver of FINE™/Turbo and the Ffowcs-Williams and Hawking (FW-H) solver of FINE™/Acoustics has been entirely automated. Acoustic computations are performed as a post-processing step (i.e. after CFD computations), with the CFD data on the FW-H surface used as input for the FW-H solver. The solver self-consistently reconstructs the CFD solution on a 360-degrees surface from the available CFD solution on the sector that has been meshed. The coupling of the FW-H module with the NLH method presents a couple of important advantages over a coupling with a classic unsteady CFD solver. In the retarded time formulation, the entire set of unsteady CFD solutions has to be evaluated and then stored before launching the FW-H solver. The manipulation of large data files between the CFD code and the acoustic propagation code can be quite cumbersome, especially when dealing with unsteady solutions. In the present case, only the steady state solution containing the harmonics on the FW-H surface has to be transmitted to the FW-H module. The method still requires computing the retarded time for each cell of the FW-H surface with an iterative solver. Nevertheless, the retarded-time CFD solution can be evaluated from a closed analytical expression (using the harmonics), without the need for interpolation between the available unsteady solutions.
A new feature has been created in the visualization tool CFView™ to define acoustic radiating surfaces, capable to automatically produce surface meshes made by uniform elements. Its implementation takes advantage of the capability already available to define probes equally distributed on a cylindrical surface. This tool is currently limited to the generation of meshes with cylindrical shape, which can be used to define FW-H porous surfaces. Before, the surface mesh generated purely by cutting surfaces from the CFD was characterized by an excessive refinement in the region of the fictive shroud, close to the blades tip. As a consequence the efficiency of the radiation analyses in FINE™/Acoustics was penalized. This new development allows us to easily perform the sensitivity analysis with respect to the FW-H surface location and its refinement level.
Finally, the graphical user interface of FINE™/Acoustics has been enhanced recently. Acoustic computations using solid or porous FW-H surfaces can be done in parallel. Lines of microphones, whose specifications were given by the SFWA-ITD partners, can easily be specified to perform the acoustic propagation computations. Moreover, the FINE™/Acoustics work-flow can now be launched automatically through Python scripts. All these features have greatly facilitated the latest computations performed for the NAA-CROR project and contributed to an improved CFD-Acoustic integrated computational chain.
Potential Impact:
The main objectives of this project are oriented at a contribution to the reduction of the environmental impact of future air transport, by supporting the development of a new generation of open rotor engines, promising a fuel consumption reduction of the order of 20 to 30%, compared to current engines. Since noise is one of the main concern of this new technology, all efforts made at reducing the produced noise will contribute to the economic development of European sustainable air transport.
In view of this objective of noise reduction, the availability of reliable and efficient simulation tools, will allow the introduction of these tools in advanced design optimization systems, in order to improve the blade design towards minimum noise generation. The present NAA-CROR project, based on a powerful new CFD approach that provides a gain in CPU time of two to three orders of magnitudes compared to current CFD methods for unsteady flow simulations, represents a major technological breakthrough. The systematic applications of the advanced CAA/CFD integrated software system to be applied and extended towards installation effects, ensures a significant progress towards the objective of designing open rotors for noise reduction. By reducing the cost and the time for a CFD/CAA simulation, a significant gain in productivity can be achieved. Hereby, many more design options can be analyzed, increasing the possibilities for identifying optimum configurations, within a always limited design time framework.
There is no doubt that this will have a major impact on the objectives of the SFWA-ITD sustainable aircrafts and support the global Clean Sky objectives. In addition, as several other subprograms, within Clean Sky and current EU projects, are focused on the open rotor technology and noise reduction, we can expect that the results of the NAA-CROR project will benefit other tasks having similar objectives, such as the GRA-ITD program.
List of Websites:
Name of the co-ordinating person: Charles Hirsch
Co-ordinating organization: NUMECA int.
Co-ordinator email: charles.hirsch@numeca.be
Co-ordinator fax: +322 6479398
During the course of this project, several improvements of the CFD/CAA system were pursued to ensure a professional, user-friendly and quality assured software environment applied to open rotors. The Non-Linear Harmonic (NLH) method, implemented in NUMECA software, is applied, providing a highly efficient method for a simultaneous prediction of the unsteady aerodynamics and the near-field acoustics of Counter-Rotating Open Rotors (CROR). Far-field acoustic propagation method (computed with a Ffowcs-Williams and Hawking solver) is also applied to NLH results. Several aspects of the software have been enhanced with respect to the meshing tool (AutoGrid5™), the CFD solver in FINE™/Turbo, the acoustic solver FINE™/Acoustics and its integration with FINE™/Turbo.
The numerical CFD/CAA approach has then been extensively applied to several CROR configurations (cruise, take-off, with/out incidence, with/out pylon, and with/out imposed transition). Both pre-test and post-test simulations have been run. When available, the results have been compared to those of other partners or to wind tunnel test data. The performed calculations prove that the NLH method is capable of accurately predicting CROR configurations at a very low computational cost (when compared to full unsteady calculations).
Project Context and Objectives:
The inherently high propulsive efficiency of advanced propellers and Counter-Rotating Open Rotors (CROR) have a great potential for fuel savings, but the level of noise emitted by the open blades represents a major obstacle to their environmental acceptance. Consequently, the design of a low noise, fuel efficient open rotor powerplant is one of the major objectives in the Clean Sky JTI. Within the virtual prototype design environment, the availability of highly efficient calculation procedures of noise sources and their propagation are essential to achieve the set objectives. While the CAA approach for the acoustic far-field noise propagation is well established, the critical issue remains the delivery of fast and accurate unsteady CFD-solutions for prediction of the noise sources. The present NAA-CROR project responds to this objective, through an advanced new approach for the CFD determination of the noise sources. The NAA-CROR project relies on the nonlinear harmonic method (NLH) which allows a gain in CPU performance for CROR’s compared to current CFD sliding grid or Chimera methodologies, of two to three orders of magnitude. This method, defined in the frequency domain, has been largely validated and successfully applied on multistage turbines and compressors at many companies. Its extension to propeller and CROR configurations has recently been achieved. We further extended this approach with the inclusion of the ability to capture installation effects for CROR configurations. The near-field and far-field noise are evaluated with an acoustic propagation module solving the Ffwocs Williams and Hawkings (FW-H) equations. The acoustic module is fully integrated with the NLH code, allowing a turnaround time for a complete CFD-CAA simulation of a few hours on a low number of processors.
Project Results:
WP 1 covers the development of the integrated CFD/CAA simulation system, capable of fully capturing installation effects. The code now allows for capturing the interaction between the pylon/nacelle and both rotors. The development required the propagation of inlet distortion perturbations through the second rotor and solving additional harmonic equations for the second rotor.
WP 2 deals with the validation of the numerical approach and the numerical simulations for selected CROR configurations. Extensive aero-acoustic calculations were performed, related to pre-test and post-test open rotor geometries, at model scale, for both take-off and cruise conditions. Results are compared with those of other partners or experimental data when available.
During pre-test calculations, NUMECA performed CFD computations for several CROR configurations (cruise, take-off, with/out incidence, with/out pylon, and with/out imposed transition). We investigated the effects of boundary layer transition for selected configurations. The aerodynamics and aeroacoustics of the engine were found to be weakly affected by the delayed transition for the selected configurations. All calculations simulate the flow by means of the non-linear harmonics method or NLH that was developed by NUMECA. The main idea is that the flow perturbations that make the flow unsteady are written about a time-averaged value of the flow and are Fourier decomposed in time. The user controls the accuracy of the unsteady solution through the order of the Fourier series. For the NLH method only one blade channel is required like a steady flow simulation, leading to a considerable gain in computational time compared to full unsteady calculations such as domain scaling techniques. The performed calculations prove that the NLH method is capable of accurately predicting CROR configurations at a low computational cost (when compared to full unsteady calculations). NUMECA performed an extended CAA analysis based on CFD results for the low-speed test cases. Acoustic results were obtained with several methods, among others with the NLH solution and with the FW-H solver. A detailed comparison and assessment of the acoustic results has been provided. We showed that FW-H computations evaluated with porous source surfaces generally provide more accurate results than those obtained with solid source surfaces since additional quadrupole sources are inherently captured within the porous FW-H surface.
Post-test aero-acoustic calculations have also been performed on the model scale counter-rotating open rotor, as tested in the S1 wind tunnel, for cruise conditions. Only one angle of attack is considered, which is 0°. For the CFD calculations, two blades geometries are considered according to the modeling of the root gap present in the experimental set-up or not. Partial hub gaps are easily modeled thanks to the dedicated AutoGrid™ capabilities. Modeling the hub partial gaps does not affect significantly the global performances. Acoustic calculations are performed on the geometry without gaps only, using the FW-H propagation module implemented in FINE™/Acoustics. The boundary conditions are directly coming from the CFD NLH calculation performed previously with FINE™/Turbo. Near-field acoustics are also evaluated using direct CFD results.
With respect to previous pre-test acoustic results obtained for take-off conditions, the FW-H permeable and non-permeable formulations provide quite different results. The predicted noise level computed considering only solid blades as acoustic sources is reduced compared to the use of a permeable surface. Whatever the location of the microphones line, FW-H Solid results show (compared to FW-H Porous results) an underestimation of about 5dB on the SPL peak value of the main rotor tones (1F and 1R) while a difference of 10dB is reached for the first harmonic (2F and 2R tones). The difference between FW-H Solid and FW-H Porous can be explained by the aero-acoustic sources located in the fluid region close to the blades, that are indeed not negligible for cruise conditions.
The analysis also shows that the original mesh used for aerodynamic performances is too coarse for near-field acoustics analysis. The CFD-based results obtained on a mesh that is refined in the radial direction above the blades show higher noise levels. On the first microphones line, the original CFD mesh gives similar peak results for the main rotor tones (1F and 1R), but already underestimates the SPL of about 2dB for the 2F and 2R tones. The difference between meshes even reaches 10dB for interaction tones. When the radius of the microphones line increases, the differences of results between meshes becomes larger. In the refined region, close to mid propeller plane, the noise levels obtained with CFD are quite close to the FW-H porous results. In order to avoid too much dissipation of the acoustic levels, the CFD mesh should therefore be enough refined in the region of interest. Around 15 points per wavelength (for the max. frequency) have to be used when the distance between the acoustic sources and the microphones lines is about 2 wavelengths. More points would be needed in case more wavelengths are considered : microphones located in the upstream and downstream regions. This is even more critical in the upstream direction where the Doppler effect is considerable for this high speed operating point.
Finally, WP3 aims at gathering the user specifications dedicated to the integrated CFD/CAA software system as applied to open rotors, in order to address these requirement in NUMECA software.
A specific wizard is now available in the turbomachinery-oriented structured grid generator AutoGrid5™ to automatically mesh contra-rotating open rotor (CROR) configurations. Both axial and radial far-field limits can easily be controlled. The flow paths and blade-to-blade mesh distribution can be set easily with a high-quality grid assurance. Current developments are also tackling rounded blade tips configuration with the introduction of a specific block topology at blade tips and a local three-dimensional smoothing.
The NLH methodology has been extended to fully account for installation effects of pylons/nacelles on the CROR aerodynamics and noise. The code now allows for capturing the interaction between the pylon/nacelle and both rotors. The development required the propagation of inlet distortion perturbations through the second rotor and solving additional harmonic equations for the second rotor. Before, the base NLH method could model only the effects between the adjacent rows, the rank-1 effects, because the three rows have three different rotation speeds (zero for the pylon and contra-rotating for the rotors). Now, besides the rank-1 effects, the extension makes possible the capture of the rank-2 effects.
The coupling between the CFD solver of FINE™/Turbo and the Ffowcs-Williams and Hawking (FW-H) solver of FINE™/Acoustics has been entirely automated. Acoustic computations are performed as a post-processing step (i.e. after CFD computations), with the CFD data on the FW-H surface used as input for the FW-H solver. The solver self-consistently reconstructs the CFD solution on a 360-degrees surface from the available CFD solution on the sector that has been meshed. The coupling of the FW-H module with the NLH method presents a couple of important advantages over a coupling with a classic unsteady CFD solver. In the retarded time formulation, the entire set of unsteady CFD solutions has to be evaluated and then stored before launching the FW-H solver. The manipulation of large data files between the CFD code and the acoustic propagation code can be quite cumbersome, especially when dealing with unsteady solutions. In the present case, only the steady state solution containing the harmonics on the FW-H surface has to be transmitted to the FW-H module. The method still requires computing the retarded time for each cell of the FW-H surface with an iterative solver. Nevertheless, the retarded-time CFD solution can be evaluated from a closed analytical expression (using the harmonics), without the need for interpolation between the available unsteady solutions.
A new feature has been created in the visualization tool CFView™ to define acoustic radiating surfaces, capable to automatically produce surface meshes made by uniform elements. Its implementation takes advantage of the capability already available to define probes equally distributed on a cylindrical surface. This tool is currently limited to the generation of meshes with cylindrical shape, which can be used to define FW-H porous surfaces. Before, the surface mesh generated purely by cutting surfaces from the CFD was characterized by an excessive refinement in the region of the fictive shroud, close to the blades tip. As a consequence the efficiency of the radiation analyses in FINE™/Acoustics was penalized. This new development allows us to easily perform the sensitivity analysis with respect to the FW-H surface location and its refinement level.
Finally, the graphical user interface of FINE™/Acoustics has been enhanced recently. Acoustic computations using solid or porous FW-H surfaces can be done in parallel. Lines of microphones, whose specifications were given by the SFWA-ITD partners, can easily be specified to perform the acoustic propagation computations. Moreover, the FINE™/Acoustics work-flow can now be launched automatically through Python scripts. All these features have greatly facilitated the latest computations performed for the NAA-CROR project and contributed to an improved CFD-Acoustic integrated computational chain.
Potential Impact:
The main objectives of this project are oriented at a contribution to the reduction of the environmental impact of future air transport, by supporting the development of a new generation of open rotor engines, promising a fuel consumption reduction of the order of 20 to 30%, compared to current engines. Since noise is one of the main concern of this new technology, all efforts made at reducing the produced noise will contribute to the economic development of European sustainable air transport.
In view of this objective of noise reduction, the availability of reliable and efficient simulation tools, will allow the introduction of these tools in advanced design optimization systems, in order to improve the blade design towards minimum noise generation. The present NAA-CROR project, based on a powerful new CFD approach that provides a gain in CPU time of two to three orders of magnitudes compared to current CFD methods for unsteady flow simulations, represents a major technological breakthrough. The systematic applications of the advanced CAA/CFD integrated software system to be applied and extended towards installation effects, ensures a significant progress towards the objective of designing open rotors for noise reduction. By reducing the cost and the time for a CFD/CAA simulation, a significant gain in productivity can be achieved. Hereby, many more design options can be analyzed, increasing the possibilities for identifying optimum configurations, within a always limited design time framework.
There is no doubt that this will have a major impact on the objectives of the SFWA-ITD sustainable aircrafts and support the global Clean Sky objectives. In addition, as several other subprograms, within Clean Sky and current EU projects, are focused on the open rotor technology and noise reduction, we can expect that the results of the NAA-CROR project will benefit other tasks having similar objectives, such as the GRA-ITD program.
List of Websites:
Name of the co-ordinating person: Charles Hirsch
Co-ordinating organization: NUMECA int.
Co-ordinator email: charles.hirsch@numeca.be
Co-ordinator fax: +322 6479398