Final Report Summary - GO4HYBRID (Grey Area Mitigation for Hybrid RANS-LES Methods)
Executive Summary:
Computational Fluid Dynamics (CFD) has become a key technology in the rapid and cost-effective design of green aircraft with reduced fuel consumption and aero-acoustic noise emissions. In general, this led to a reversed order on simulations, putting now more weight to the numerical simulations compared to wind tunnel investigations – although the latter are not less important.
The accurate and efficient prediction of turbulent flow, however, represents one of the central limitations of CFD, with precise methods requiring unfeasible computational resources and more efficient methods introducing approximations and inaccuracy. A new family of hybrid Reynolds Average Navier-Stokes - Large Eddy simulation (RANS-LES) methods have recently emerged, which offer a significant increase in accuracy whilst limiting expense to levels that are affordable with current and near-future computational capacity.
Despite excellent results by hybrid RANS-LES methods, a fundamental issue remained to be addressed, known as the ‘Grey-Area Problem'. This aspect concerns the 'transition region' between the RANS and LES modes of such hybrid methods. As the grey-area problem has a particularly detrimental impact on flows featuring shallow regions of boundary layer separation and re-attachment, the accuracy of hybrid RANS-LES predictions was downgraded, and applications with respect to aerodynamic and aero-acoustic flows, such as wings near the borders of the flight envelope and jet noise, tended to suffer from grey-area issues.
A range of approaches to reducing the grey-area severity have been developed and evaluated. The evaluation took into account not only the predictive accuracy of the improved methods but also practical issues, such as computational expense and user-friendliness. A balance was struck between simple academic test cases (for reduced computational expense and more “pure” evaluation) and complex application test cases (for demonstrating applicability).
Significant improvements to hybrid RANS-LES methods with respect to grey-area mitigation for non-zonal methods and improved embedded approaches were achieved, from which a significant contribution to increased industrial confidence in CFD for challenging flows has been made.
Project Context and Objectives:
Based on the mentioned academic fundamental and complex application test cases, technically, a two-pronged strategy was adopted. On the one hand, grey-area mitigation strategies for non-zonal hybrid approaches have been pursued. These methods are inherently more flexible and applicable to complex industrial geometries; but they suffered most strongly from grey-area effects. On the other hand, zonal or embedded strategies have the potential to eliminate the grey-area problem entirely. However, these zonal methods are inherently more complicated to set up and are most readily applied only to simple configurations or a limited class of industrial problems. They are hence far from being a generalised tool in a complex industrial environment. Improvements to the flexibility and applicability of embedded hybrid strategies therefore constituted the second aim of Go4Hybrids efforts.
All the development work focussed on just two academic test cases. The direct comparability and ranking of the methods was facilitated by common grids and an innovative Common Assessment Platform (CAP).
The latter was used as a possibility to circumvent the use of different codes by the Go4Hybrid partners, codes that are different with respect to numerical issues. Hence, the CAP served as a basic software tool where the different methods have been employed and tested to ensure a direct comparison. Considerable effort was undertaken to achieve comparable results, but more time and resources would have been needed to ensure an additional more elaborated comparison with methods in use by the Go4Hybrid partners.
Besides the fundamental test cases, the research and development of proof-of-concept test cases was enabled by a range of complex industrial demonstration applications, including delta wing flows as well as aero-acoustic jet noise investigations, flow around a three-element airfoil, a shallow recirculating flow and the flow around a complex helicopter fuselage.
Project Results:
In Go4Hybrid a dual strategy is pursued, reflecting the most promising avenues to resolving the Grey Area issue:
• Development of techniques for Grey Area Mitigation (GAM) in non-zonal methods, and
• Improvement of flexibility and fidelity of embedded approaches
The term non-zonal method here refers to those in which the model, not the user, defines where RANS and LES is applied. Such methods are generally more flexible and suitable for complex applications, however they suffer strongly from the Grey Area problem.
Embedded approaches on the other hand are those in which explicit treatment is used to translate modelled (RANS) turbulence into resolved (LES) eddies at a user-defined interface. This in principle (i.e. assuming a perfect such interface condition) would solve the Grey Area issue, however the disadvantages are a less flexible approach that can be applied less widely to complex applications.
To address these issues, key technical activities aimed at Grey Area mitigation for non-zonal approaches (WP2) and improved embedded approaches (WP3) were conceived. Finally, a dedicated work package 4 serves as the “hub” of the project, in which overarching conclusions are drawn, best practice guidelines are collated and the methods assessed via the innovative “Common Assessment Platform” (CAP).
WP2: Grey Area Mitigation for non-zonal methods
GAM improvements for non-zonal hybrid RANS-LES methods were developed and tested by numerous partners (CFDB, FOI, NLR, NTS, ONERA). The corresponding mandatory, fundamental test case of a planar shear layer downstream of a flat splitter plate was chosen, since this gives a very direct characterisation of the underlying Grey Area issue.
Significant acceleration of the transition from modelled to resolved turbulence was observed for all proposed grey-area mitigation approaches, which contrast strongly with the sluggish development of resolved turbulence downstream of the splitter plate trailing edge from baseline "pre-Go4Hybrid" methods. Importantly, these improvements could also be demonstrated for complex applications in Task 2.2.
WP3: Improved embedded approaches
Enhanced embedded approaches were developed and tested by the four involved partners (DLR, ONERA, NTS and UniMan). The developments principally focussed on synthetic turbulence formulations for the interface between RANS and LES zones in the stream-wise direction. For this activity (Task 3.1) the mandatory test case of a flat plate turbulent boundary layer was specified. Generally, all methods performed very well, achieving recovery of the flow within 5-10 boundary layer thicknesses of the synthetic turbulent interface.
In the associated Task 3.2 complex demonstration activities were conducted for numerous test cases, including a 3-element airfoil and a 2D wall-mounted hump. An effective translation from the smooth RANS upstream flow to a viable resolved LES flow downstream of the interface was achieved. Impressive quantitative comparison with skin friction for the 2D wall-mounted hump case was seen, which was neither achieved by RANS nor with pre-Go4Hybrid non-zonal hybrid RANS-LES methods.
WP4: Common Assessment Platform (CAP)
The majority of approaches proposed by the partner have been implemented in the open-source CFD solver OpenFOAM for the purposes of cross comparison within a common platform. Two test cases have been used for model evaluation; namely a spatially developing boundary layer (for the embedded approaches), and the planar shear layer (for non-zonal approaches). In both cases, the highly sensitive and challenging flows chosen enabled a ranking of the performance of the various methods with identical mesh and underlying numerics. The fact that this ranking did not apply to the results produced by each partner with their own codes highlights the high sensitivity of the test cases to numerical differences.
Potential Impact:
The Go4Hybrid project has gathered groups that are active in the further development of hybrid RANS-LES methods resulting in novel developments and increased experience and knowledge related to the grey area problem.
Based on this experience and on the needs of engineers and researchers without in-depth, “developer” knowledge, the need for generalisation and automation is obvious. Here, the Go4Hybrid project findings can be an important support for software vendors and engineering development departments.
Summarising impact:
The Go4Hybrid project addressed the aeronautics priorities by means of a consortium with extremely skilled partners together with companies endorsing the project and working in the fields of aeronautics, turbo-machinery, helicopters and ground (automotive and rail) transportation, who guided the proposed technical work by cross-fertilisation and broad knowledge dissemination. On this basis, the innovative Grey Area Mitigation methods have been thoroughly assessed and will likely become a standard in European CFD tools. Moreover, the hybrid RANS-LES approaches involved (by the different partners) will both serve and guide industry by using these methods in a less ambiguous way where less user directives are necessary – the latter somewhat still a bottleneck for day-to-day use so far.
Together with the “group of observers” with their duty to assess/audit the project, other industries were involved as “associate partners”. Hence, it can be finally said that the Go4Hybrid project very well complied with the structure of the Aeronautics and Space Thematic Priority about Research for strengthening the competitiveness by innovation of all partners involved (in particular the aeronautics industry) in the global market by providing new/advanced numerical concepts and technologies. Generally speaking, the industrial awareness of the hybrid RANS-LES methods was be pushed up to the next step of the Technology Readiness Level.
To conclude, the Go4Hybrid project - incorporating a large number of European industrial partners as endorsers together with the project partners - offered a broad spectrum of synergy for increasing every partner’s technical capabilities and competitiveness founded on successful dissemination and exploitation of results and expertise. Thus the formation of trans-national teams by calling upon the best European specialists independent of their location served as a source for setting up European standards and did contributed to the establishment of a more harmonised aeronautics environment in Europe, being increasingly “favourable to innovation”.
List of Websites:
http://go4hybrid.mace.manchester.ac.uk
Point of contact:
C. Mockett
CFD Software E+F GmbH (CFDB)
Wolzogenstr. 4 14163 Berlin
Tel. +49-(0)30-590083320
E-mail: charles.mockett@cfd-berlin.com
Computational Fluid Dynamics (CFD) has become a key technology in the rapid and cost-effective design of green aircraft with reduced fuel consumption and aero-acoustic noise emissions. In general, this led to a reversed order on simulations, putting now more weight to the numerical simulations compared to wind tunnel investigations – although the latter are not less important.
The accurate and efficient prediction of turbulent flow, however, represents one of the central limitations of CFD, with precise methods requiring unfeasible computational resources and more efficient methods introducing approximations and inaccuracy. A new family of hybrid Reynolds Average Navier-Stokes - Large Eddy simulation (RANS-LES) methods have recently emerged, which offer a significant increase in accuracy whilst limiting expense to levels that are affordable with current and near-future computational capacity.
Despite excellent results by hybrid RANS-LES methods, a fundamental issue remained to be addressed, known as the ‘Grey-Area Problem'. This aspect concerns the 'transition region' between the RANS and LES modes of such hybrid methods. As the grey-area problem has a particularly detrimental impact on flows featuring shallow regions of boundary layer separation and re-attachment, the accuracy of hybrid RANS-LES predictions was downgraded, and applications with respect to aerodynamic and aero-acoustic flows, such as wings near the borders of the flight envelope and jet noise, tended to suffer from grey-area issues.
A range of approaches to reducing the grey-area severity have been developed and evaluated. The evaluation took into account not only the predictive accuracy of the improved methods but also practical issues, such as computational expense and user-friendliness. A balance was struck between simple academic test cases (for reduced computational expense and more “pure” evaluation) and complex application test cases (for demonstrating applicability).
Significant improvements to hybrid RANS-LES methods with respect to grey-area mitigation for non-zonal methods and improved embedded approaches were achieved, from which a significant contribution to increased industrial confidence in CFD for challenging flows has been made.
Project Context and Objectives:
Based on the mentioned academic fundamental and complex application test cases, technically, a two-pronged strategy was adopted. On the one hand, grey-area mitigation strategies for non-zonal hybrid approaches have been pursued. These methods are inherently more flexible and applicable to complex industrial geometries; but they suffered most strongly from grey-area effects. On the other hand, zonal or embedded strategies have the potential to eliminate the grey-area problem entirely. However, these zonal methods are inherently more complicated to set up and are most readily applied only to simple configurations or a limited class of industrial problems. They are hence far from being a generalised tool in a complex industrial environment. Improvements to the flexibility and applicability of embedded hybrid strategies therefore constituted the second aim of Go4Hybrids efforts.
All the development work focussed on just two academic test cases. The direct comparability and ranking of the methods was facilitated by common grids and an innovative Common Assessment Platform (CAP).
The latter was used as a possibility to circumvent the use of different codes by the Go4Hybrid partners, codes that are different with respect to numerical issues. Hence, the CAP served as a basic software tool where the different methods have been employed and tested to ensure a direct comparison. Considerable effort was undertaken to achieve comparable results, but more time and resources would have been needed to ensure an additional more elaborated comparison with methods in use by the Go4Hybrid partners.
Besides the fundamental test cases, the research and development of proof-of-concept test cases was enabled by a range of complex industrial demonstration applications, including delta wing flows as well as aero-acoustic jet noise investigations, flow around a three-element airfoil, a shallow recirculating flow and the flow around a complex helicopter fuselage.
Project Results:
In Go4Hybrid a dual strategy is pursued, reflecting the most promising avenues to resolving the Grey Area issue:
• Development of techniques for Grey Area Mitigation (GAM) in non-zonal methods, and
• Improvement of flexibility and fidelity of embedded approaches
The term non-zonal method here refers to those in which the model, not the user, defines where RANS and LES is applied. Such methods are generally more flexible and suitable for complex applications, however they suffer strongly from the Grey Area problem.
Embedded approaches on the other hand are those in which explicit treatment is used to translate modelled (RANS) turbulence into resolved (LES) eddies at a user-defined interface. This in principle (i.e. assuming a perfect such interface condition) would solve the Grey Area issue, however the disadvantages are a less flexible approach that can be applied less widely to complex applications.
To address these issues, key technical activities aimed at Grey Area mitigation for non-zonal approaches (WP2) and improved embedded approaches (WP3) were conceived. Finally, a dedicated work package 4 serves as the “hub” of the project, in which overarching conclusions are drawn, best practice guidelines are collated and the methods assessed via the innovative “Common Assessment Platform” (CAP).
WP2: Grey Area Mitigation for non-zonal methods
GAM improvements for non-zonal hybrid RANS-LES methods were developed and tested by numerous partners (CFDB, FOI, NLR, NTS, ONERA). The corresponding mandatory, fundamental test case of a planar shear layer downstream of a flat splitter plate was chosen, since this gives a very direct characterisation of the underlying Grey Area issue.
Significant acceleration of the transition from modelled to resolved turbulence was observed for all proposed grey-area mitigation approaches, which contrast strongly with the sluggish development of resolved turbulence downstream of the splitter plate trailing edge from baseline "pre-Go4Hybrid" methods. Importantly, these improvements could also be demonstrated for complex applications in Task 2.2.
WP3: Improved embedded approaches
Enhanced embedded approaches were developed and tested by the four involved partners (DLR, ONERA, NTS and UniMan). The developments principally focussed on synthetic turbulence formulations for the interface between RANS and LES zones in the stream-wise direction. For this activity (Task 3.1) the mandatory test case of a flat plate turbulent boundary layer was specified. Generally, all methods performed very well, achieving recovery of the flow within 5-10 boundary layer thicknesses of the synthetic turbulent interface.
In the associated Task 3.2 complex demonstration activities were conducted for numerous test cases, including a 3-element airfoil and a 2D wall-mounted hump. An effective translation from the smooth RANS upstream flow to a viable resolved LES flow downstream of the interface was achieved. Impressive quantitative comparison with skin friction for the 2D wall-mounted hump case was seen, which was neither achieved by RANS nor with pre-Go4Hybrid non-zonal hybrid RANS-LES methods.
WP4: Common Assessment Platform (CAP)
The majority of approaches proposed by the partner have been implemented in the open-source CFD solver OpenFOAM for the purposes of cross comparison within a common platform. Two test cases have been used for model evaluation; namely a spatially developing boundary layer (for the embedded approaches), and the planar shear layer (for non-zonal approaches). In both cases, the highly sensitive and challenging flows chosen enabled a ranking of the performance of the various methods with identical mesh and underlying numerics. The fact that this ranking did not apply to the results produced by each partner with their own codes highlights the high sensitivity of the test cases to numerical differences.
Potential Impact:
The Go4Hybrid project has gathered groups that are active in the further development of hybrid RANS-LES methods resulting in novel developments and increased experience and knowledge related to the grey area problem.
Based on this experience and on the needs of engineers and researchers without in-depth, “developer” knowledge, the need for generalisation and automation is obvious. Here, the Go4Hybrid project findings can be an important support for software vendors and engineering development departments.
Summarising impact:
The Go4Hybrid project addressed the aeronautics priorities by means of a consortium with extremely skilled partners together with companies endorsing the project and working in the fields of aeronautics, turbo-machinery, helicopters and ground (automotive and rail) transportation, who guided the proposed technical work by cross-fertilisation and broad knowledge dissemination. On this basis, the innovative Grey Area Mitigation methods have been thoroughly assessed and will likely become a standard in European CFD tools. Moreover, the hybrid RANS-LES approaches involved (by the different partners) will both serve and guide industry by using these methods in a less ambiguous way where less user directives are necessary – the latter somewhat still a bottleneck for day-to-day use so far.
Together with the “group of observers” with their duty to assess/audit the project, other industries were involved as “associate partners”. Hence, it can be finally said that the Go4Hybrid project very well complied with the structure of the Aeronautics and Space Thematic Priority about Research for strengthening the competitiveness by innovation of all partners involved (in particular the aeronautics industry) in the global market by providing new/advanced numerical concepts and technologies. Generally speaking, the industrial awareness of the hybrid RANS-LES methods was be pushed up to the next step of the Technology Readiness Level.
To conclude, the Go4Hybrid project - incorporating a large number of European industrial partners as endorsers together with the project partners - offered a broad spectrum of synergy for increasing every partner’s technical capabilities and competitiveness founded on successful dissemination and exploitation of results and expertise. Thus the formation of trans-national teams by calling upon the best European specialists independent of their location served as a source for setting up European standards and did contributed to the establishment of a more harmonised aeronautics environment in Europe, being increasingly “favourable to innovation”.
List of Websites:
http://go4hybrid.mace.manchester.ac.uk
Point of contact:
C. Mockett
CFD Software E+F GmbH (CFDB)
Wolzogenstr. 4 14163 Berlin
Tel. +49-(0)30-590083320
E-mail: charles.mockett@cfd-berlin.com