Evaluation of Engine Inlet Vortices Using CFD

Transcription

1 50th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition January 2012, Nashville, Tennessee AIAA Evaluation of Engine Inlet Vortices Using CFD Luis Gustavo Trapp 1 EMBRAER, São José dos Campos, , Brazil Roberto da Motta Girardi 2 Instituto Tecnológico de Aeronáutica ITA, São José dos Campos, , Brazil The effects of inlet vortices on engines operating near the ground have been studied for more than 50 years, but the phenomenon is not yet completely understood. Computational Fluid Dynamics (CFD) is used, on a DLR-F6 modified nacelle, to allow a more detailed look into the inlet vortices phenomenon. An updated correlation for the prediction of the ground vortex strength is shown and compared to CFD results. Additionally is presented a new technique that helps the understanding of phenomena that are highly influenced by vorticity generated on the wall. This technique is based the CFD ability to mix wall boundary conditions with and without velocity slip at the wall. Four regions on the nacelle were identified that contribute differently to the ground vortex intensity. Nomenclature A = Area CR = Contraction ratio = A h /A t D = Diameter DR = Diffusion ratio = A f /A t H = Distance from the engine axis to the ground L = Length L i = Interference length M = Mach number MFR = Mass Flow Ratio = U t /U 0 = A 0 /A t n r = Surface normal vector U = Time average velocity SDR = Streamtube diameter ratio = D 0 /2H W = Engine mass flow rate Z = Distance from the inlet lip to the ground Γ = Circulation ω = Vorticity Subscripts 0 = Freestream h = Inlet highlight station t = Inlet throat station I. Introduction ET engines are adversely affected when operating near the ground in crosswind conditions, especially at high J power. A high power static engine is affected by crosswind through in at least three different ways: the leeward engine ingests the vorticity shed by the fuselage; it may suffer inlet flow separation; and a inlet vortex can form between the ground and the engine. All these phenomena produce a non-uniform airflow that is ingested by the inlet and can affects the engine operation, being able even to affect its physical integrity. This non-uniform flow ingested 1 Propulsion Systems Engineer, Center of Competence, Av.Faria Lima Professor, Aeronautical and Mechanical Engineering Department, Praça Marechal Eduardo Gomes, Copyright 2012 by the, Inc. All rights reserved.

2 by the engine is generally referred to as inlet distortion or distortion, which is usually assessed through standard procedures, e.g. Ref.1. The ground vortex forms between the engine inlet and the ground, at very low aircraft speeds and is created by the interaction between the inlet flow and the crosswind, being able to suck considerable large objects into the engine. This vortex can also affect engine operability, when it is ingested by the engine core and it is able to lift particles from the ground which are much larger than the ones carried by the average inlet flow, due to the vortex higher flow speed and lower pressure compared with the neighboring flow. Over the last 50 years many authors have studied inlet vortices, either through experiments 2-9 or through numerical means The first author to propose the minimum requirements for a ground vortex to exist was Klein 3, in 1959, conditioning it to the existence of: 1. A stagnation point on the ground (or other fixed structure), where an omni-directional air flow parallel to the surface converges, in a manner similar to a sink; 2. An updraft from the stagnation point to the inlet; 3. The presence of the ambient vorticity. The evolution of Computational Fluid Dynamics (CFD) tools and computer hardware has allowed the study of nacelles on crosswind conditions near the ground with increased accuracy, greater stability and faster turn around. Compared to wind tunnel testing, CFD allows a more detailed look into the physical phenomenon and a greater control than the one allowed in wind tunnels or in actual size nacelles. The objective of this work is to present a new technique that allows the evaluation of phenomena that are highly influenced by vorticity generated on the walls. This technique is based on using the CFD ability to mix wall boundary conditions with and without slip. Fig. 1 Ground vortex in a wind tunnel model (Ref. [8]) II. CFD Analysis The inlet vortices simulations were performed using the DLR F6 nacelle 16, a wind tunnel model, long duct scaled nacelle model, from the wing-body-pylon-nacelle configuration used in the AIAA Drag Prediction Workshops. The original nacelle is a hollow through-flow nacelle (TFN). In order to use this nacelle model in an actual crosswind CFD simulation, the internal geometry was modified to more closely represent the effects of a real engine, therefore it was included engine outlet and inlet boundaries, but no effort was made to add a spinner at the fan plane. Typical dimensions for the DLR-F6 wind tunnel model nacelle together with its design parameters are given on Ref. 15. There are two major inlet design parameters that impact the smoothness of the flow that will reach the engine: the inlet diffuser ratio and the inlet contraction ratio. The inlet diffusion ratio is the ratio between the fan and the inlet throat areas, and it influences the air flow speed on the fan blades. The inlet contraction ratio is defined 2

3 as the ratio between the inlet highlight and the inlet throat areas. It regulates the flexibility of the inlet to ingest flows that are not aligned with the inlet plane as occurs, for example, when the inlet is at high angles of attack or at static crosswind conditions. The inlet lip cross-section shape also influences the inlet flow separation margins, an elliptical cross-section, rather than the original circular one, was used to avoid separation under crosswinds. Additionally the contraction ratio was increased from 1.24 to 1.34, also to improve the inlet crosswind capability. In order to assess the nacelle behavior during crosswind operation, it was chosen, the following baseline condition: sea level, aircraft static, ISA day and a 90 crosswind, with wind intensities varying between 5 and 30 knots. The nacelle was positioned at a distance from the engine centerline to the ground equal to the highlight diameter. The computational domain was hemispherical, with a diameter 50 times the nacelle maximum external diameter. The CFD analysis were performed using CFD++ 17, a commercial CFD software based on a finite-volume formulation, which can deal with arbitrary mesh types. The Reynolds Averaged-Navier Stokes three-dimensional equations are solved for the compressible flow using implicit, second order interpolation, nodal based polynomials and pre conditioned relaxation. The turbulence model used was the realizable k-ε. A hybrid tetra-prism mesh containing approximately six million elements was used. Volumetric mesh densities were used to increase the number of elements in the inlet, ground vortex and nacelle wake regions. The prism layer contained 25 elements across its height, with a growth ratio of The parameter y+, which is important in evaluating the ability of the grid to capture near-wall effects, was of the order of one in all cases. This mesh can be seen in Figure 2. At the fan face boundary a mass flow rate boundary condition was used. The mass flow rate imposed was typical of an engine at high thrust condition, in this case 0.34 kg/s. The nacelle exhaust was not modeled and was treated as a wall, therefore the engine jet is not present in the solution. The domain initial conditions were identical to the far field boundary condition, at which a characteristic based velocity inflow/outflow was prescribed, imposing wind speed, temperature, turbulent intensity and turbulent length. Figure 2 View of the nacelle and the ground surface meshes On a previous work, Ref. 15, it was proposed a non-dimensional parameter, called the streamtube diameter ratio (SDR), to evaluate the relationship between the inlet streamtube interference with the ground and its effect on the inlet vortex strength. The SDR, as shown in Figure 4, is the ratio between the ideal streamtube diameter (D 0 ) and the diameter of an imaginary cylinder of radius equal to the distance from the engine axis to the ground (H), Eq. 1. The ideal streamtube is the axis-symmetrical streamtube of an isolated nacelle, undisturbed by the ground. In this 3

4 definition, the ideal streamtube diameter can be calculated taking into account the ambient conditions and the engine mass flow rate, as shown in the expanded form of Eq.1. The SDR definition implies that when it is greater than one the inlet streamtube encompasses the ground, which is one of the conditions for the inlet vortex to exist. Ideal streamtube Ideal streamtube D 0 D 0 H H Figure 3 Parameters involved on the definition of the Streamtube Diameter Ratio (SDR). D0 1 W SDR = = (1) 2H H πρ U Different cases were run to evaluate how the vortex intensity varied not only as a function of the SDR, but also as a function of the inlet mass flow rate, the crosswind intensity and the distance from the inlet to the ground. Originally, on Ref. 15, the SDR was compared against the peak suction induced on the ground, however a more appropriate relationship is against the non-dimensional circulation. The vortex circulation can be calculated using the Stokes theorem: Γ = r ω n r da (2) A Ground Where Γ is the circulation, ω r is the vorticity vector, A is the integration area and n r is the vector normal to the integration area. The circulation can be made non-dimensional (Γ * ) by dividing it by the inlet throat diameter (D t ) and the inlet throat speed (U i ): 0 Γ Γ * = (3) D t U i Figure 5 shows results of non-dimensional circulation as a function of SDR for the following series of cases: (a) 20 knots crosswind, an inlet height ratio of 1.03 and different engine mass flow rates; (b) 20 knots crosswind, inlet height ratio of 1.13 and different engine mass flow rates; (c) inlet mass flow rate of 0.40 kg/s, height ratio of 1.13 and different crosswind intensities; (d) inlet mass flow rate of 0.34 kg/s, height ratio of 1.13 and different crosswind intensities. This Figure 5 plot shows three different ground vortex regimes exist: a weak vortex regime between SDR 1.1 and 1.4 with minimum ground pressures close to the ambient pressure; a transient regime between 1.4 and 1.6, where the ground vortex strength increases suddenly; and strong vortex regime at SDR greater than 1.6, starting with the vortex peak strength, which then decays smoothly. Figure 5 also shows that, for the same nacelle distance to the ground, different wind and inlet mass combinations can produce the same vortex strength when plotted against SDR. Additionally it can be seen that the factors that influence the maximum vortex strength, for a given nacelle geometry, is the nacelle distance to the ground, the closer the nacelle to the ground, the highest the peak, as it was the case with an inlet height ratio equal to

5 0.50 Ground vortex circulation Non-Dimensional Circulation Streamtube diameter ratio (SDR) H/D 1.13; 20 knots; varying W H/D 1.13; 0.40 kg/s; varying Uo H/D 1.03; 20 knots; varying W H/D 1.13; 0.34 kg/s; varying Uo Figure 4 Ground vortex non-dimensional circulation as function of the SDR, with different scenarios. The ground vortex can contain vorticity created on the ground and on the nacelle walls. However it is not clear how the vorticity from these surfaces influence the vortex strength. Additionally, in case the nacelle walls influence the vortex intensity, it would be useful to understand whether this influence is caused by the whole nacelle surface, or some parts of the nacelle have greater influence than others. The methodology employed to improve the understanding of the ground vortex phenomenon was evaluating the influence of different vorticity sources, through selectively mixing wall boundary conditions of the non-slipping and slipping types. The slipping boundary condition, typically used to model solid surfaces in an inviscid flow simulation, imposes a symmetry condition on the flow in which all scalar and vector quantities are considered to be mirror-imaged in the planes defined by the surface-boundary faces. In this case the flow velocity experiences a free slip at the boundary, becoming tangential to the surface of the boundary. When using this boundary condition, vorticity is not produced on the walls and therefore it is possible, by alternating selected surfaces between slipping and non-slipping boundary conditions, to assess their influence on the ground vortex. Initially four different cases were evaluated, all with 10 knots crosswind, 0.34 kg/s of inlet mass flow and a nondimensional distance to the ground of 1.13: (a) the standard case, where all walls are non-slipping; (b) the slipping nacelle walls case; (c) the slipping ground case; and (d) the case with all walls slipping. Figure 6 presents results for these cases, showing isocountours of static pressure on the ground with the pallete range set between the minimum and maximum ground pressures. It can be seen that the standard case (a) resulted in a concentrated ground vortex, with a small low pressure core, as it is usually seen on actual engines under crosswind. The slipping nacelle wall case (b) resulted in a vortex of similar diameter on the ground as case (a), but with a lower ground suction peak, showing that the vorticity produced on the nacelle surface influences the ground vortex strength. The slipping ground case (c) contains a vortex with similar core pressure, compared to case (b), however it spreads through a much larger area than the other vortices. Case (d), which has all walls slipping, on the other hand, did not produce a vortex, showing that for the ground vortex to exist, a source of vorticity must be present in the flow. When a non-slipping condition is set on either the ground or the nacelle wall, this vorticity source exists and a ground vortex will form. 5

6 (a) All walls non-slipping (b) Slipping nacelle walls (c) Slipping ground wall (d) All walls slipping Figure 5 Ground pressures for different wall boundary conditions. Figure 6 shows vortex streamlines colored by vorticity intensity for the (a) nacelle wall slipping, (b) ground slipping and (c) all walls slipping cases. The stream lines are released from a vorticity isocurve created at a plane situated at 40% of the distance between the ground and the highlight, with 90% of the peak vorticity in that plane. Figure 6(a) shows a wider vortex compared to the one with a non-slipping ground (Figure 6(b)). This is probably due to the lack of the non-slip condition which does not restrict the flow and allows the vortex to spread without being dissipated by the ground shear; in this case the vortex is large and shallow on the ground, rather than concentrated as with the non-slipping ground. The remaining case, where all the walls are slipping (Figure 6 (c)) shows as expected a convergence of streamlines into the inlet (with a stagnation line in its center), however there is negligible vorticity levels in the flow, which are not enough to induce the streamlines to rotate and form a ground vortex. Additional analysis were performed using results of previous analysis as initial conditions. When results of the all-walls slipping case, were used as initial conditions for the ground slipping case, a ground vortex appeared. In the opposite case, when using results of the ground slipping case as initial condition for the nacelle walls slipping case, the ground vortex disappeared. This showed that the ground vortex needs an independent source of vorticity to exist, not being self-sufficient. 6

7 Wind (a) (b) (c) Figure 6 Vortex streamlines colored by vorticity intensity for the (a) ground slipping; (b) nacelle walls slipping; (c) all walls slipping. A comparison of the vortex core vorticity for the different cases, at a non-dimensional height to the ground of 20% (Figure 7) shows that the no slip vortex being the case with the highest vorticity peak. On the other hand, the slipping ground case has the largest core, but its maximum vorticity is much lower than the other cases. If one integrates the vorticity through the whole radius, given that its area is much greater than the other cases, it is the slip ground vortex that has the highest circulation. Both slipping nacelle and no slip cases have vortex cores of about the same diameter, showing that is the ground shear the major factor that limits the spread of the vortex core radially. Ground vortex vorticity (/s) Vorticity Distance to vortex core (m) No slip Slip Ground Slip nacelle Figure 7 Vorticity distribution relative to the vortex center for the different wall boundary condition combinations. Another aspect of the ground vortex that can be evaluated using different slip and non-slip boundary conditions is how the vortex strength varies between the ground and the inlet. Figure 8 shows the ground vortex circulation at different ground heights (h), adimensionalized by the engine diameter D, for the different cases. The ground level is at 0%, and the 35% position is approximately the bottom of the inlet lip. The first conclusion from this plot is that the ground vortex looses intensity as it approaches the inlet, the closer it is to the inlet, the higher the difference. This can have two reasons, one is that the vortex stretching near the inlet, which decreases the vortex radius, while increasing the core velocity gradient, increases vortex dissipation. The other reason may be that the ground vortex is interacting more strongly near the ground and increasing its intensity as extends away from the inlet. The case where the nacelle walls are slipping and all vorticity comes from the ground has an inlet vortex with the lowest circulation, even though a lot of vorticity is being produced on the ground, the ground shear layer is at the same time dissipating it. The case where all walls are no-slip, have the second greatest vortex intensity, showing that 7

8 a major responsible for the inlet vortex strength are the nacelle walls. Additionally, as already observed on the previous figures, the case where all the walls are slipping, produce no inlet vortex, and this is reflected on the plot below, where a very small circulation is seen on the all slip case. 2.0 Ground vortex circulation 1.6 Circulation (m2/s) % 5% 10% 15% 20% 25% 30% 35% 40% h/d Nacelle slip No slip Ground slip All slip Figure 8 - Ground vortex circulation as a function of measurement plane height An additional use of the CFD capability of mixing slipping and non-slipping wall boundary conditions can be to apply it to different parts of the same surface. Therefore the nacelle surface was initially divided into two parts: one with the outer nacelle surfaces up to the highlight, other one with the surfaces internal to the inlet. In all cases the ground was kept as a slipping wall. In the case where the internal nacelle wall was turn into slipping conditions, nothing happened the ground vortex was kept with the same intensity. However when the internal surfaces were kept non-slipping, but the external surfaces were turn into slipping made the ground vortex did not form, leading to the conclusion that the nacelle external walls were the ones responsible for the ground vortex vorticity. Further on, the external surface was split into more surfaces: the front of the nacelle, i.e. the region from the highlight to the maximum nacelle diameter, was split into 4 equal longitudinal length sectors; then the rest of the nacelle was split into slices with the same length as the front (Figure 9); finally the last sector (12 th ) was merged with the rest of the 13 th sector, which was smaller than the reference length. CFD runs were performed with these sectors with different combinations of slipping and non-slipping boundary conditions. This methodology allowed to pin point the major features that influence the inlet vortex and to study the role that the nacelle surface plays on strengthening the inlet vortex. Figure 9 Nacelle with splitted external walls. Two series of cases were ran with a 10 knots crosswind: the "increasing slip" and the "increasing non-slip". On the "increasing slip" the nacelle slices were turned successively from non-slipping to slipping boundary condition, from the nacelle highlight to the nacelle trailing edge (i.e. a case was run with the first slice slipping and the remaining slices non-slipping; the next case had the first and second non-slipping, the remaining non-slipping, and so on successively.). The "increasing non-slip" followed the opposite process: on a non-slipping nacelle the slices boundary conditions were successively turned into non-slipping. All other walls were kept slipping, including the ground, the nacelle inlet and the back of the nacelle, meaning that the nacelle external surface was the only surface generating vorticity in the CFD domain. 8

9 Results for these cases are shown on Figure 10, where the ground vortex circulation is measured close to the ground and non dimensionalized by the throat diameter and velocity. It can be seen that, when the non-slipping walls are increasingly turned into slipping, the ground vortex looses circulation slowly and progressively, disappearing only when the last nacelle slice (#12) is turned into slipping. When the opposite process is employed, initially a ground vortex does not form; when the 6 th slice start producing vorticity finally a weak ground vortex forms; as more slices are turned into non-slipping there is a small decay on circulation; this changes when the 9 th slice is turned into non-slipping, producing a sudden increase on circulation up to the level of a completely nonslipping nacelle. 7.0 Ground Vortex Circulation Non-dimensional circulation % 20% 40% 60% 80% 100% Non-slipping nacelle wall length (%) Increasing non-slip Increasing slip Figure 10 - Ground vortex non-dimensional circulation variation with different external wall combinations of slipping and non-slipping boundary conditions. Figure 11 shows results for the case where the first three sectors have slipping walls, while the other sectors are non-slipping. It contains isocurves of constant vorticity on planes passing to the center of each sector. The external vorticity downstream of the nacelle, like in a cylinder, accumulates in a wake and it is progressively sucked by the inlet. While it is being sucked, the wake gets further concentrated and the vorticity is intensified by vortex stretching. The isocurves depict the location of the wake behind the nacelle and how this wake is further concentrated on the forward portion of the inlet. The location of the center of the top wake, which forms the trailing vortex, moves to the back of the nacelle, as the flow approach the inlet lip. A more detailed look on the concentrated vorticity region in the front of the nacelle lip shows that two vortices coexist, a bigger ground vortex and on its side a smaller vortex, which we have called the feeding vortex. This vortex seems to be responsible for inducing the stagnation line to rotate and producing the first step in circulation that occurs in the increasing non-slip case with 6 slices non-slipping. 9

10 Trailing vortex Ground vortex Feeding vortex Figure 11 Splitted nacelle with the three first sectors with wall slipping boundary conditions. The isocurves of constant vorticity on planes are on the center of each sector. Figure 12 shows, for the same case, two isosurfaces of constant longitudinal vorticity, with values equal to and 1000 s -1, colored by longitudinal vorticity intensity. The nacelle surfaces are colored using two shades of gray, the lighter gray shows extend of the slipping sector, whereas the darker one (inside the inlet; outside the nacelle away from the highlight) shows the non-slipping surfaces. Vorticity is being produced on the non-slipping sectors of the nacelle, the flow that is going through the top of the nacelle is producing negative vorticity, the bottom is producing positive longitudinal vorticity. This positive longitudinal, as is being sucked by the engine and reaches the slipping wall region, ceases to produce vorticity. At the same time the wind component flowing around the nacelle circumferentially, forced longitudinal vorticity to accumulate in a narrow channel that is sucked by the inlet this vorticity channel contain the feeding vortex. Figure 12 Isosurfaces of constant longitudinal vorticity, colored by longitudinal vorticity intensity. Case with the first three sectors slipping (lighter gray shows extend of the slipping sector). Based on the results for the different combinations of wall slip and non-slip conditions it was possible to identify four different nacelle regions that influence the ground vortex: film region, intermediate region, wake region and shielding region; as shown on Figure 13. It shows a top view of the nacelle, with results for a case where the first 9 slices have non-slipping wall boundary conditions; it shows a isosurface of Q criteria equal to -1000, colored by vorticity intensity. The film region is located close to the highlight, this region is subjected to a strong suction which induces a longitudinal velocity much greater than the one induced by the crosswind flow, therefore creating a film of vorticity that is predominantly in the circumferential direction. The small circumferential velocity component not only creates a small amount of longitudinal vorticity on the walls, but also does not allow this vorticity to accumulate. The longitudinal vorticity is ingested by the inlet as a vorticity film, with no concetrated vorticity. This is the reason why no ground vortex forms, on the increasing non-slip case of Figure 10, until the 6 th slice is turned into non-slip, there 10

11 is only a vorticity film. Concentrated vorticity that is produced elsewhere, when passing through this region, is stretched and subject to increased dissipation. Further aft on the nacelle starts the intermediate region where the circumferential velocity component is of the same order of magnitude as the longitudinal region. In this region there is an increased amount of longitudinal vorticity that is created and accumulates, creating a small separated region both on top and below the nacelle. This separation region is not large and its flow usually reattaches as it nears the inlet, due to the strong suction forces. Vorticity generated on the wind- and leeward sides of the nacelle is captured by the inlet. The concentrated vorticity on the bottom (the feeding vortex) interacts with the inlet stagnation line and starts the ground vortex, producing on the increasing non-slip case of Figure 10 the first increase in vortex circulation. The wake region is the part of the nacelle that has a large separation on the leeward side, with the vorticity accumulating. It contains two contra-rotating separated regions, a larger one in the top, a smaller one in the bottom. The bottom region accumulates vorticity that will be channeled to the inlet further increasing the feeding vortex. The top separated region, that contains the trailing vortex that will be explained in more details below, will be carried downstream, part of it eventually being captured back by the inlet suction, into the ingested trailing vortex; part of this wake being carried downstream as the trailed trailing vortex. Vorticity from this region does not enter the inlet directly, only after separating. Eventually the trailing vortex is wide enough to interact and further induce the rotation of the ground vortex, provoking the second circulation on the increasing non-slip case of Figure 10. Finally, on the nacelle end is the shielding region, an extensively separated region from where the wall generated vorticity is completely carried downstream. Because this region is not influenced by the inlet suction and is close to the nacelle back, it separates more easily and creates a large separated region that provides shielding for the neighboring wake region to extend and increase its size. Figure 13 - Top view of the nacelle with a an isosurface of Q criteria colored by vorticity intensity. Another vortex present in the flow around an engine close to the ground is the trailing vortex. This vortex forms in the back of the inlet and was first observed experimentally by De Siervi (Ref. 5). Using CFD a more a detailed analysis of this vortex can performed. Releasing particles traces adequately, at different locations of the domain, it can be shown that the trailing vortex consists of two different vortices, one that trails with the downstream flow, other that is ingested by the engine inlet. Figure 14 shows both components of the trailing vortex: the ingested trailing vortex is shown by the blue particle traces which, after flowing around the nacelle, accumulate vorticity and flow downstream in a cylinder wake fashion, the particles eventually loose speed and get captured back to the inlet in a more concentrated vortex that flows through the core of the trailing vortex. The trailed trailing vortex (gray particle traces on Figure 14), on the other hand, consists of the externally portion of the trailing vortex that the inlet suction is not capable of capturing; it keeps flowing downstream, eventually occupying the core of the vortex, downstream of the ingested trailing vortex.. 11

12 Wind Figure 14 - Particle traces that form the trailing vortex (blue: ingested vortex; gray: trailed vortex) and ground vortex (white trace); the ground shows isocontours of pressure; the plane inside the inlet shows vorticity in the longitudinal direction. The ground vortex and the trailing vortices share a continuous boundary, that starts on the ground and connects the ground vortex and the trailed trailing vortex; close to the inlet and inside the inlet the interaction is between the ground and ingested trailing vortex. This interaction produces a direct relation between the trailing vortex and ground vortex intensities. Additionally, the interaction shall also be taken into account when measuring the vortices intensities. For example, Ref. 7 measured the trailing vortex three diameters downstream of the inlet and mentions that the circulation of the inlet and trailing vortices are approximately equal and opposite; ref.19, on the other hand, has measured the trailing vortex circulation inside the inlet and found the ground vortex much more intense than the trailing vortex. Our analysis have found that both authors are correct. III. CONCLUSIONS A comparison of the stream tube diameter ratio (SDR) against the ground vortex non-dimensional circulation was presented and it seems to take into account the major parameters that contribute to the ground vortex strength, namely the inlet distance to the ground, the engine inlet flow and the crosswind speed. It points to a value of 1.6 as the condition where the highest strength vortex exists. It was shown that it is possible to use different types of wall boundary conditions, slipping and non-slipping wall, simultaneously to identify how the local production of vorticity influences the overall flow. The use of these boundary conditions side-by-side has not influenced the solver behavior and did not cause any convergence or instability problems. Moreover they have allowed to pin point the most important regions that contribute for the existence of a ground vortex. For the ground vortex to exist, a source of vorticity must be present in the flow. If the non-slipping condition is set on either the ground or the nacelle wall, this vorticity source exists and a ground vortex will form; however if only slipping walls boundary conditions are used, no ground vortex will form. The vorticity created on the nacelle external wall contribute more for the ground vortex intensity than the ground. The subdivision of the nacelle external walls into many sectors has allowed the identification a region of concentrated vorticity, that is sucked by the inlet and merges into the ground vortex. This concentrated vorticity region has been called feeding vortex, as it seems to be one source of vorticity for the ground vortex. Another source seems to be the interaction between the trailing and ground vortex. Once the feeding and trailing vortices do not exist, as when the nacelle external surface has slipping boundary conditions, the ground vortex ceases to exist. 12

13 The trailing vortex was shown to be composed of two different vortices: ingested and trailed, the ingested is in the center of the trailing vortex and is captured by the engine inlet, whereas the trailed vortex flows externally. Finally, with the results for the different combinations of wall slip and non-slip conditions it was possible to identify four different nacelle regions that influence the ground vortex: film region, intermediate region, wake region and shielding region. It is important to remark that the engine jet was not modeled, but preliminary analysis taking the jet into account have shown that it does affect the inlet vortex existence. IV. REFERENCES 1. SAE, Inlet Total-Pressure-Distortion Considerations for Gas-Turbine Engines, Society of Automobile Engineers Aerospace Information Report 1419, Rev. A., Rodert, L.A., and Garrett, F.B., Ingestion of foreign objects into turbine engines by vortices, NACA TN-3330, Washington, USA, Klein, H. J., Douglas Aircraft Company, US Patent for Vortex Inhibitor for Aircraft Jet Engines, No. 2,915,262. B64D33/02, filed 01 Dec Glenny, D.E., and Pyestock, N.G.T.E., Ingestion of debris into intakes by vortex action, Aeronautical Research Council, London, Current Papers No. 1114, 50 pp, De Siervi, F., A flow visualization study of the inlet vortex phenomenon, MSc Dissertation, Massachusetts Institute of Technology, USA, Liu, W., Greitzer, E.M., and Tan, C.S., Surface static pressures in an inlet vortex flow field, Journal of Engineering for Gas Turbines and Power, vol. 107, p , Shin, H.W., Greitzer, E.M., Cheng, W.K., Tan, C.S. and Shippee, C.L., Circulation measurements and vortical structure in an inlet-vortex flow field, Journal of Fluid Mechanics, Vol. 162, pp , Brix, S., Neuwerth Untersuchungen zur Entstehung und Stärke von Triebwerkseinlaufwirbeln, Dissertation, Fakultät für Maschinenwesen der Rheinisch-Westfälischen Technischen Hochschule Aachen, Deutschland, Murphy, J.P. and MacManus, D.G., Ground vortex aerodynamics under crosswind conditions, Experiments in Fluids, No.50, pp , doi: /s Tourrette, L., Navier-Stokes simulations of air-intakes in crosswind using local preconditioning, AIAA Paper No , June Yadlin, Y., and Shmilovich, A., Simulation of Vortex Flows for Airplanes in Ground Operations, AIAA Paper No , Jan Trapp, L.G., Argentieri, H.G., Souza, F.J. and Girardi, R.M., Wind effects on isolated nacelles near the ground, Proceedings of the 11th Brazilian Congress of Thermal Sciences and Engineering, Curitiba, Brazil, Rehby, L. Jet engine ground vortex studies, MSc Dissertation, Cranfield University, Cranfield, United Kingdom, Nakayama, A. and Jones, J.R., Correlation for Formation of Inlet Vortex, AIAA Journal, Vol. 37, No.4, pp , 1999.doi: / Trapp, L.G., Girardi, R.M., An engine nacelle in crosswind: the inlet vortex, Journal of Aircraft, vol.47, No.2, pp , 2010.doi: / DLR F6 Geometry. 2nd AIAA Drag Prediction Workshop, Orlando, USA, URL: [retrieved 15 December 2011]. 17. Metacomp Technologies, CFD++ v , User Manual, Metacomp Technologies, USA, Ensight, User s Manual for Version 9.2.2, Computational Engineering International, Inc., Apex, USA, Brix, S., Neuwerth, G., and Jacob, D., The inlet-vortex system of jet engines operating near the ground, AIAA Paper , Jun