ICAS 2002 CONGRESS

COPING WITH WAKE VORTEX

Klaus-Uwe Hahn Institute of Flight Research, German Aerospace Center (DLR) Lilienthalplatz 7, 38108 Braunschweig, Germany

Keywords: wake vortex, encounter, control, safety, separation distance

Abstract MTOW maximum take-off weight nm nautical mile The well known phenomenon of wake vortices P2P probabilistic two phase model behind a lift producing can adversely af- r radius / radial co-ordinate fect flight safety if encountered by trailing air- ROT runway occupation time craft. The strength of the vortices increases with s spanwise load factor SD separation distance the weight of the vortex generating . t time co-ordinate Therefore, weight dependent separation dis- T thrust, time constant tances have been established for approach and v lateral velocity component landing to avoid dangerous wake vortex en- V airspeed (no index), velocity counters. These proven separation distances w vertical velocity component Wweight have to be investigated carefully with the aim to α discover possible margins to be used to solve β angle of sideslip the current and future capacity problems at air- χ flight path azimuth ports. This demand forms the need for more γ flight path angle flexible separation procedures taking into ac- η deflection count the actual weather situation and the pa- ν effective viscosity ρ air density rameters of the individual aircraft pairing. ξ Separation reductions can be achieved by wake deflection ζ deflection vortex avoidance using prediction systems for ∆ difference the wake vortex development and movement. Φ bank angle The hazardous areas around a wake vortex can Θ pitch angle be approximated by the simple geometry of a ψ azimuth rectangle or an ellipse according to the aircraft indices pairing, the actual weather dependent decay ccore F follower aircraft and the available control power for vortex com- FB feedback pensation. Using specific controllers vortices FF feed-forward can be passed through even if the required g geodetic control power temporarily exceeds the available K kinetic (refers to flight path) capacity. Aircraft equipped with such a control- L leading/generator aircraft req required ler might follow another aircraft closer than sink indicates downwards motion authorized by the current separation distances. t tangential Wwind WV wake vortex List of Symbols and Abbreviations WVL wake vortex line AVOSS Aircraft Wake VOrtex Spacing System • time derivative bwing span ' characteristic time scale DoF degrees of freedom * denotes normalized parameters DLR German Aerospace Center 0 initial value G aircraft weight 1 diffusion phase IMC instrumental meteorological conditions 2 rapid decay phase 732.1 K.-U. Hahn

1 Introduction Presuming approaches on a single runway ψ ≈ ψ ≈ ψ As a reaction to the generation of lift, two where small encounter angles L WVL counter rotating vortices arise at the wing tips of can be assumed, normally the situation of a par- an aircraft. This phenomenon is well known as allel-like encounter will prevail. The investi- wake vortex and can adversely affect flight gated scenario is sketched in Fig.1. safety of trailing aircraft. This is particularly true if a light aircraft follows a heavy one since 2 Wake Vortex Spacing of Aircraft the strength of the generated vortices depend on the required lift correspondingly to the aircraft 2.1 Fixed Aircraft Separation Distances weight. To avoid dangerous vortex encounters To guarantee safe flight operation, currently the aircraft staggering for approach and landing fixed mass dependent separation distances have is restricted. The allowed minimum separation been established for aircraft lined up for is a limiting factor for airport capacities. The approach and landing. Various matrices of current and the expected capacity problems at minimum separation distances from different airports with a high volume of traffic together organizations are available. Tab.1 gives the with future large and very heavy aircraft form mass classification of aircraft and in Tab.2 the the need for a revision of the existing rigid sepa- corresponding IMC vortex separations after ration concepts. This demand has led to a new FAA are listed. Such fixed matrices can be interest and comprehensive investigations into easily applied to flight operation. But they the nature of the wake vortex phenomenon. suffer from the lack of flexibility to actual The following investigation and the pre- ambient weather situation. sented simulation results will focus on the com- 2.2 Dynamic Aircraft Separation bination of a small aircraft following behind a To reduce the current rigid separations research large one. For this encounter situation real flight activities [2, 3, 4] have been initiated for the test data have been available for validation. The development of concepts for dynamic wake drawn conclusions are of general importance vortex spacing systems. The most mature since the underlying physics are true for any system is the AVOSS for which already per- aircraft combinations. formance validations are carried out [5]. All ac- The effects of wake vortices on aircraft can tivities finally aim for the increase of the aircraft be very different. Encounters perpendicular to throughput to improve airport capacity. The the vortex axis will lead to substantial vertical principle of dynamic spacing is based on the load factor variations by inducing rapid and consideration of all relevant influences: large angle of attack variations all over the wing - weather forecast improved and updated by [1] comparable to strong gusts. Another situa- - sensing the ambient weather conditions used tion is present if the flight path of the encoun- together with tering aircraft runs parallel to the vortex axis. - data of aircraft pairing to Then the vortex flow will induce varying angles - predict the vortex behavior to provide a of attack along the wing producing strong roll- - non-hazardous vortex or vortex free space ing moments which can create significant roll rates resulting in large bank angles. Between along the approach path. To overcome the un- these two cases a broad variety of encounter certainties in the predictions adequate margins situations exists. In any case the wake vortex have to be introduced. The required time to produces an adverse effect on aircraft motion guarantee a safe approach for a following air- which is the reason for the today’s encounter craft (due to vortex demise or vortex drifting avoidance policy. The pilots position concern- away) defines the minimum separation distance. ing wake vortex is formulated by the statement To minimize the required separation actions can that no aircraft should encounter a wake vortex be taken to push the distances to their (safe) by intention whatever its strength is. limits:

732.2 COPING WITH WAKE VORTEX

- measures to reduce vortex strength have been performed in the frame of DLR’s - passive or active devices to speed up the “Wirbelschleppe” project [3] by means of a vortex decay realistic and detailed simulation program. - more precise prediction of vortex behavior - minimization of the non-hazardous vortex or 3 Data Gathering from Flight Tests vortex free space - active aircraft encounter control For the validation of a reliable wake vortex en- counter simulation real flight test data have been Contributions have to come from aerodynamics, used. The respective flight tests have been per- and meteorology aiming for low formed in the frame of the European S-Wake vortex design, accelerated decay, weather and project within flight experiments especially de- vortex behavior prediction. Another potential is signed for wake vortex encounter analyses [9]. available coming from aircraft control. In the following a brief description of the tests The use of an will improve the for the data gathering will be given. accuracy of flight path tracking [6]. A precise As a vortex generating aircraft DLR’s twin flight path tracking of the vortex generating air- engine jet VFW614 ATTAS (Advanced Tech- craft reduces the scatter in position deviations nologies Testing Aircraft System) was used around the nominal approach path and thus re- (Fig. 2). Having a MTOW of about 20 tons this duces the extension of the space in which the aircraft has to be classified as “LARGE” follow- generation of vortices have to be expected (→ ing the classification of Tab.1. The test aircraft reduced uncertainty of initial vortex position). A was equipped with a smoke generator on its left more precise flight path tracking of the follow- wing (Fig.3) to mark the vortex at the tip. This ing aircraft shrinks the extension of the space made it easy for the follower aircraft to en- which will be probably penetrated due to excur- counter the vortex by intention. sions from nominal path (→ reduced space of The encountering aircraft was the Do 128 interest of vortex contamination). Therefore, it test aircraft of the Technical University of can be stated that the application of automatic Braunschweig powered by two turbo-prop en- approach will help to minimize two sources of gines (Fig.4). With a maximum MTOW of probabilistic uncertainties. about 4.35 tons the aircraft has to be classified Automatic control can also improve the as “SMALL”. This aircraft was ideally suited aircraft response to a wake vortex encounter for the encounter trials providing very high ac- which might happen unintended. Accepting that curacy air data probes at 4 different stations al- the hazard of a wake vortex penetration results lowing multipoint scans when passing the wake from the very dynamic response of the aircraft vortex. Numerous scans of encounters have then the situation can be treated like a flight in been performed at various distances (between strong turbulence which can be improved by a about 0.5nm and 1.5nm) during flight test cam- gust load alleviation system [7, 8]. The use of a paigns in late summer 2001 and in early 2002. vortex controller (of course controlling the air- The recorded data have been analyzed by means craft response and not the encountered vortex of a parameter identification method to validate itself) will shift the limits for a non-hazardous the mathematical models representing the wake vortex situation to a higher level of acceptable vortex flow field and the encounter behavior. vortex strength. Consequently, regarding the aircraft actual pairing and the available control power provided by the penetrating aircraft the 4 Modeling and Simulation separation distance can be further reduced. In The simulation system used for the investigation any case the wake vortex encounter has to be consists of various sub-models. The available safe in terms of acceptable flight path excur- modules and the data flow are shown in Fig.5. sions and flight state deviations. The investiga- In the following the used main modules will be tions into controlled wake vortex encounters roughly introduced. For the choice of models

732.3 K.-U. Hahn

the rule “as complex as necessary but as simple function of the time co-ordinate is the most im- as possible” was applied. portant factor. Manifold models for the mathe- matical description of the vortex decay are 4.1 Aircraft Modeling available [14, 15, 16, 17] showing considerable different results. 4.1.1 Vortex Generating Aircraft A new approach for the wake vortex decay The vortex generating aircraft was simply is proposed by HOLZÄPFEL [18]. This model modeled by the wake vortex produced by itself. is ideally suited for the use in flight mechanics The common approach for the calculation of the simulations. This P2P model is a probabilistic strength of a fully developed single vortex is parametrical model considering two different W phases of vortex decay (Fig.6). First the diffu- Γ = L (1) sion phase followed by the second phase 0 2 ⋅ ρ ⋅V ⋅ b ⋅ s L L showing a rapid decay. The P2P model consid- The term ers the relevant aircraft and atmospheric data = ⋅ b0 bL s (2) determining the wake vortex behavior. The model delivers the time dependent strength of is the distance between the two vortex lines the vortices and their positions. Accepting that it since the position of the trailing vortices are not will not be possible to get exact results from any exactly located at the wing tips. It can be as- model the P2P model uses a probabilistic com- sumed that s = π/4 applies if no specific data are ponent which gives uncertainty bounds of the available. These relations are well accepted in spatial and temporal variations covering the the community of wake vortex research and expected vortex behavior. The model is checked have been confirmed by the flight test data. against real data sets, e.g. the Memphis data 4.1.2 Encountering Aircraft [19] showing good correspondence (see Fig.7). The encountering aircraft was modeled being a For more details about this model it is referred rigid body with full 6 DoF. The rigid body to [18]. aerodynamics and the engine model were pro- 4.2.2 Vortex Velocity Field vided by the Institute of Flight Guidance from From the circulation of the vortex the cor- the Tech. Univ. of Braunschweig [10, 11]. responding velocity field can be calculated. The effect of the wake vortex flow induced Again there are numerous approaches [20-24] supplementary angle of attack was considered available transforming the circulation Γ of a following the approach of the ONERA strip L completely developed vortex into a radial de- model [12]. This model calculates the forces pendent velocity distribution. In [13] it is shown and moments from the local angle of attack that the best fit of a continuous model compared variations computed at different sections, so with real flight test data is obtained by the called strips along the wing and along the hori- approach of BURNHAM and HALLOCK [22] zontal . Γ 2 The encounter model was validated by a L r (3) V = ⋅ . t ⋅π ⋅ 2 + 2 parameter identification method applied to the 2 bL rc r flight test data [13]. The results show that the model can be regarded to be a good approach in This approach is able to produce a good match terms of realistic encounter representation. of the overall radial velocity shape. Corre- sponding to the two phase vortex decay the growth of the core radius rc expressed by the 4.2 Wake Vortex Modeling * normalized core radius rc with 4.2.1 Vortex Decay and Movement = * ⋅ rc rc b0 (4) For the velocity field behind a vortex generating is considered to have also two phases [25]: aircraft the strength of the circulation ΓL as a

732.4 COPING WITH WAKE VORTEX

* ≤ * * = * pilot is based on a model following concept t T2 : rc rc0 (5) [26]. This provides that the aircraft follows a * > * * = * + t T2 : rc rc0 (smooth) predefined angular state command ν * ⋅ ()t* +T * − 2 ⋅ν * ⋅T * model representing a realistic aircraft behavior. 2 2 2 2 The main control is performed by feed-forward ν * based on an inverse aircraft model. Unmodeled 2 is the normalized effective viscosity and * real world effects in the applied inverse vehicle applicable for the rapid decay phase, T2 is the * model cause model following errors that have to normalized time when this phase starts and t is be compensated by feedback. The block dia- the normalized time coordinate calculated by gram of the autopilot is shown in Fig.10. The * = ' t t / t (6) autopilot was designed to have a good perform- ance in strong turbulence and gusts. based on the characteristic time scale ' = ()⋅π ⋅ 2 Γ 4.3.2 t 2 b0 0 For the thrust control a simple autothrottle The initial core radius for the fully (Fig.10) is used based on the total energy man- developed vortex is rc0. Also from parameter agement of the aircraft [27, 28]. The required identification it was found out [13] that the core energy state for the steady flight is described by radius rc0 needs to be chosen smaller than given the airspeed V and the flight path angle γ. Any ≈ by the common estimate rc 0.05 bL. As a good deviations from this reference state formulate approach it can be assumed that the required thrust variation to reestablish the ≈ ⋅ nominal situation. It applies rc0 0.035 bL (7) ∆ = − () +∆γ ⋅ applies. The maximum tangential velocity at the T V g W (8) core radius was identified to match that one where W is the aircraft’s weight. From the coming from the LAMB-OSEEN model in demand for thrust the lever position for combination with rc ≈ 0.05. This outcome fits the aircraft control is calculated. perfectly to the factor of 0.035 used in Eq.(7) affecting the radial distribution of the velocity 4.2.3 Vortex Controller field after Eq.(3) and thus the maximum veloc- To cope with a vortex encounter a specific ity at core radius. The time dependent develop- controller has been designed acting against the ment of the core radius described by Eqs.(5) and vortex induced moments before noticeable flight (7) is shown in Fig.8. state deviations occur. In the following this controller is named “vortex controller” and it 4.2.2 Wake Vortex Flow should be clear that the controller is controlling For the complete wake vortex flow behind the the encountering aircraft and not the vortex of leading aircraft a pair of counter rotating the leading aircraft. The vortex controller con- vortices one for each are superimposed. sists of a feed-forward (FF) and a feedback (FB) The distance between the two vortex lines b0 is part (Fig.11). defined by Eq.(2). Looking into the direction of Considering the experience from the design flight the left wing vortex is clockwise rotating. of gust load alleviation systems and wind shear Fig.9 illustrates the result of this wake vortex controllers [7, 8, 28] it is known that atmos- model compared with data from in-flight pheric disturbances can be treated best using FF measurement. control. This has the additional advantage of leaving the original aircraft behavior un- 4.3 Encounter Control changed. Knowing the aerodynamics (including the control surface efficiency) of the respective 4.3.1 Autopilot aircraft to be controlled the necessary com- To keep the aircraft on the desired approach mands for compensation of the induced mo- path it needs to be controlled. The used auto- ments can be calculated. In the ideal case the FF 732.5 K.-U. Hahn

controller will fully compensate for the flow motion. The worst case occurs when the wake disturbances. But in a real system an additional vortex runs exactly parallel to the approach path FB controller is needed to cover real world and the following aircraft is permanent exposed effects and uncertainties. to the vortex flow field in a quasi stationary flight. The areas to be avoided within a wake 4.4 Vortex Flow Measurement vortex flow field then can be calculated using the required control power normalized by the For the presented investigation the measurement maximum available control power of the of the wake vortex flow was not addressed to be encountering aircraft expressed in terms of an issue. Nevertheless two different principles aileron deflection. The wake vortex induced roll should be mentioned. moment can be compensated if 1. By means of LIDAR or LASER meas- ξ * = ξ ξ < urement systems placed in the nose of the req max 1 (10) aircraft the area on both sides in front of the applies. Depending on the combination of wake can be scanned. The velocity field of vortex generating and encountering aircraft ar- the flow disturbances around the aircraft eas of various threats can be defined by the nor- can be calculated by    malized aileron deflection ξ*. = − VW VK V . (9) The encountering aircraft is the Do128 (MTOW = 4.35t). As vortex generator two The measurement ahead of the wing offers aircraft have been considered: the advantage to compensate for system 1) VFW614/ATTAS: MTOW = 20t delays [7], e.g. computation delays. Γ 2 2. Another kind of measurement can be initial circulation: L0 = 158m /s performed by using an array of pressure 2) generic aircraft: MTOW = 79t 2 ports along the wing span (minimum 4 per initial circulation: ΓL0 = 342m /s. wing). The pressure variations induced by Both pairings present a SMALL aircraft behind the vortex flow can be interpreted in terms a LARGE one (see Tab.1). Aircraft 1) is at the of changed local lift coefficients from lower and aircraft 2) is at the upper bound of the which the required control commands can classification of “LARGE” aircraft. In principle be computed. both pairings show the same results concerning Whatever the principle of flow disturbance the encounter behavior except the fact that due measurements is, it is important to know that the to the lower initial circulation of combination 1) effects of interest show higher dynamics com- the threat vanishes earlier in this case. There- pared to the low dynamics of the normal aircraft fore, this paper will focus on the latter combina- motion. This offers the opportunity to filter out tion since it is the more demanding case and the relevant signal spectrum [7]. covers the combination 2). Fig.12 illustrates the flow field behind the generic aircraft in terms of 5 Aircraft Control in a Wake Vortex the amount of normalized aileron deflection ξ* at different separation distances. Assum- There are different standards for separation ing an approach speed of V = 60m/s for the distances established. In this paper the US IMC F follower aircraft the corresponding ages of the Separation Standards given in Tab.2 will be wake vortex are also listed in the table below: used. For the proposal of reduced but safe aircraft separations the underlaying physics SD [nm] 1.62 2.5 4.0 5.0 6.0 have to be understood. SD [km] 3.0 4.6 7.4 9.3 11.1 vortex age [s] 50 76.7 123.3 155.0 185.0 5.1 Steady Flight in a Wake Vortex SD/ bL [-] 88 135 217 273 326 It is obvious that for parallel-like encounters the In addition to the standard SDs the distance aircraft’s roll response is the dominating equivalent to the minimum runway occupation 732.6 COPING WITH WAKE VORTEX

time ROT = 50s is given which represents the can stay in the approach area. The shorter the absolute limit for aircraft separation. SD the higher the risk of penetrating a In Fig.12 the red areas indicate that the re- hazardous area due to uncertainties in wake quired control power exceeds the available per- vortex behavior. For the acceptance of reduced formance. In such a situation the aircraft cannot SD it must be guaranteed that even unforeseen perform a steady flight. The displayed situations situations can be passed safely. This will only are true for neutral atmosphere and considering be possible by the support of automatic control. only aircraft self-induced turbulence (no atmos- pheric turbulence is present). These conditions 5.2 Passing Through a Wake Vortex can be regarded to be the worst case considering the vortex decay. In Figs. a) to c) it can be seen In the real world a wake vortex encounter will that the conditions do not change very much be only a temporary but very dynamically event. since only the diffusion phase has taken place A typical example for an unintended wake and the rapid decay has not yet really devel- vortex encounter is represented by the following oped. The required SD for the relevant aircraft situation. The encountering SMALL aircraft is combination is 4nm/7.4km. Even in this nomi- approaching the runway on a 3° glide slope and nal situation red areas are still existent. The rea- will penetrate the left vortex of a LARGE sons why normally no wake vortex encounters aircraft. The vortex line has a horizontal are reported are: First, after the vortices came orientation and crosses exactly the glide slope. into existence they will start to sink down (ini- The SD is chosen to be 2.5 nm which is about tial sink rate for the vortices of the generic air- 63% of the current standard. Fig.14 shows the encounter with only the craft is wsink ≈ 2 m/s) and thus they will clear the approach path for the follower aircraft. Second- autopilot (designed for strong turbulence) ac- ly, cross winds will shift the vortices away into tive. When the vortex is penetrated the autopilot lateral direction. Both effects normally prevent a tries to keep the nominal flight path. Although full hit of the vortex core of a following aircraft. the aileron deflection acts against the induced A pilot may interpret the penetration of the rolling momentum the aircraft experiences bank Φ − outer regions of the wake vortex as normal at- angles of about +7° > > 13°. In the begin- mospheric turbulence. ning the aircraft banks to the right. Together Dependent on the accepted aileron deflec- with the lateral vortex velocity coming from the tion for vortex compensation it will take only left it moves more than yg > 10m to the right. some seconds until the situation becomes non- There it is exposed to the strong downdraft hazardous due to the wake vortex movement. between the two vortices and dives below the Assuming that a control demand of ξ* < 0.3 is glide slope. After passing the vortex core (about tolerable this will happen in less than 30s. 10s) the aircraft is recovered close to its nominal flight path. The hazardous area of ξ* > 0.3 can be In Fig.14 also the performance of the vor- roughly approximated by an ellipse or a rectan- tex controller is illustrated. It can be seen that gle as illustrated in Fig.12. The representative Φ − parameters versus SD of the approximated the aircraft bank reaction 0° > > 2° can be boundaries are given in Fig.13. For a SD of neglected although stronger vortex velocities are * encountered. But significant flight path de- 6nm or more no hazardous areas of ξ > 0.3 do viations occur since the vortex controller is a exist anymore. Together with the prediction of state controller and not designed for flight path the wake vortex movement these simplified tracking. To overcome this deficiency it has to boundaries can be used to generate SDs as a be combined with the autiopilot. function of tolerable normalized aileron Fig.15 displays the results of the encounter deflections ξ*. using the combined autopilot and vortex con- For specific meteorological conditions the troller. The results show little bank angle varia- vortices will not move away fast enough or even tions and satisfactory flight path deviations. But

732.7 K.-U. Hahn

to fulfill the task of flight path tracking and state cepted deflection for vortex compensation, these control the required control power exceeds the areas can be approximated by the simple ge- available aileron capacity. If the applicable ai- ometry of a rectangle or an ellipse. With regard leron deflection is limited to its realistic maxi- to the vortex motion these simple bounds can be mum (ξ* ≤ 1) the resulting aircraft behavior used to calculate SDs which provide non-haz- is deteriorated but still acceptable (Fig.15). ardous conditions within the relevant space Even the situation corresponding to the ROT around the nominal approach path. (SD = 1.62nm) which leads to a SD of about Vortices can be passed through even if the 40% of the current nominal SD can be passed required control power temporarily exceeds the by automatic control (Fig.16) available capacity. This is possible since the It is clear that flying through the core of a extension of the cells of very high rotational vortex is the worst case that can happen and has velocities are limited to small areas. Therefore, to be avoided. As a result of the motion of the the corresponding time of exposure is normally vortices such an extreme vortex encounter will short. The necessary aileron deflection rates are be a very rare situation. But if this case occurs it high. A human pilot surprised by an unintended can be coped by automatic control especially penetration will have problems to apply the cor- designed for this event. The exhibited aircraft rect control input amplitudes without phase de- behavior in terms of load factor might be un- lays [29]. This situation can be improved by an comfortable but not hazardous. Much more automatic controller especially designed to dangerous are the high sink rates which occur counteract the effects of vortices. The “vortex when the down-wash between the vortex pair is controller” introduced in this paper is well penetrated. Especially close to ground such rates suited to cope with wake vortices in combina- cannot be accepted. To make sure that no vortex tion with an autopilot. It is mainly based on a encounter will happen in the vicinity of the disturbance feed-forward concept. This has got ground an observation system is needed for the advantage of leaving the aircraft control dy- sensing the space around the glide slope. From namics unchanged. Using forward looking this information a prediction of the vortex mo- information would allow to take control actions tion in front of the runway can be derived. before the vortex penetration. This provides the More likely than hitting a vortex core will potential for further controller improvements. be penetrations of the outer vortex region. It is important to understand that the vortex Fig.17 illustrates a flight a few meters left from controller is not designed to substitute a human the vortex line of the left vortex. The passed pilot. The intention is to support the pilot and to area (see Fig.12: yellow area) requires a nor- relieve him from a high dynamic task. The malized aileron deflection of 0.5 <ξ*< 1. The vortex controller will compensate for the vortex combined autopilot/vortex controller has no induced disturbances and the pilot still has to problem to keep the flight path and the aircraft control the aircraft. The situation can be states with only small variations although the compared with an active gust load alleviation maximum required normalized aileron deflec- system smoothing the aircraft’s gust response tion is about ξ* ≈ −0.8. Even close to the ground without interfering the pilot’s task. this aircraft behavior is acceptable. The vortex controller is not thought to al- low the disregard of the wake vortex hazard but it could be an element in the puzzle of reducing 6 Summary and Conclusions the SDs. An aircraft equipped with such a de- The hazard of the velocity field of a wake vice should be able to follow another aircraft in vortex can be expressed by areas of normalized a shorter SD since it is able to cope with unin- aileron deflections ξ* required for permanent tended encounters which of course have to be an roll-momentum compensation in a quasi station- exception. It is proposed to reduce the currently ary flow field. According to the aircraft pairing, applicable SDs in small steps to gather opera- the actual weather dependent decay and the ac- tional experience. In any case the resulting sepa- 732.8 COPING WITH WAKE VORTEX

ration distances have to be safe in terms of ac- [14] Greene GC. An Approximate Model of Vortex Decay ceptable flight path excursions and flight state in the Atmosphere. Journal of Aircraft, Vol. 23, No. deviations. 7, pp 566-573, 1986. [15] Han et al. Large Eddy Simulation of Aircraft Wake Vortices in a Homogeneous Atmospheric Turbu- 7 References lence: Vortex Decay and Decent. AIAA 99-0756, 37th AIAA Aerospace Sciences Meeting and Exhibit,, [1] König R. Aircraft Response and Pilot Behaviour Reno, 1999. During a Wake Vortex Encounter Perpendicular to [16] Sarpkaya T. A New Model for Vortex Decay in the the Vortex Axis. Flight in Adverse Environmental Atmosphere. AIAA 99-0761, 37th AIAA Aerospace Conditions, Gol, Norway, paper 17, pp 17/1-17/18, Sciences Meeting and Exhibit, Reno, 1999. 1989. [17] Donladson C snd Bilanin AJ. Vortex Wakes of [2] Hinton D A. Description of Selected Algorithms and Conventional Aircraft. AGARDograph No. 204, Implementation Details of a Cocept-Demonstration AGARD-AG-204, Princeton, 1975. Aircraft VOrtex Spacing System (AVOSS). NASA/TM-2001-211027, Hampton, Virginia, 2001. [18] Holzäpfel F. A Probabilistic Two-Phase Wake Vortex Decay and Transport Model. Report No. 163, [3] Gerz T. Wirbelschleppe – Minimierung, Vorhersage, German Aerospace Center, Institut für Physik der Beobachtung und Erkennung von Wirbelschleppen Atmosphäre, Oberpfaffenhofen, 2001. ziviler Verkehrsflugzeuge. Projektplan, German Aerospace Center, 2001. [19] Campbell S D, Dasey T J, Freehart R E, Heinrichs R M, Matthews M P, Perras G H and Rowe G S. Wake [4] de Bruin A et al. Assessment of Wake Vortex Safety Vortex Field Measurement Program at Memphis, TN, –S-WAKE, ANNEX 1 “Description of Work”, Euro- Data Guide. Project Report NASA/L-2, Lincoln Lab., pean Commission, Proj.No: G4RD-1999-0099, 1999. MIT, Cambridge, MA, 1997. [5] Ruitshauser D K and Cornelius J O’C. Aircraft Wake [20] Oseen C W. Arkiv för Mat., Astron. och Fys., Nr. 14, Vortex Spacing System (AVOSS) Performance 1911. Update and Validation Stud. NASA/TM-2001- 211240, Hampton, Virginia, 2001. [21] Lamb H. Hydrodynamics. Cambridge University Press, pp 590-592, Cambridge, 1932. [6] Frauenkron H et al. FLIP – Flight Performance Using Frankfurt ILS. DFS, German Air Navigation [22] Burnham D C and Hallock J N. Chicago Monostatic Services, Air Traffic Management Devision, 2001. Acoustic Vortex Sensing System. Vol. 4, Wake Vortex Decay, Springfield, VA, National Information [7] König R and Hahn K U. Load Alleviation and Ride Service, 1982. Smoothing Investigations Using ATTAS. 17th Con- gress of the International Council of the Aeronautical [23] Proctor F H. The NASA-Langley Wake Vortex Sciences, Stockholm, Sweden, pp 1379-1393, 1990. Modelling Effort in Support of an Operational Air- craft Spacing System. 36th AIAA Aerospace Sciences [8] Hahn K U and König R. ATTAS Flight Test and Meeting and Exhibit, AIAA 98-0589, Reno, 1998. Simulation Results of the Advanced Gust Manage- ment System LARS. Atmospheric Flight Mechanics [24] Winckelmans G. Effect of non-uniform windshear Conference, Hilton Head Island, SC, pp 58-70, 1992. onto vortex wakes: parametric models for operational systems and comparison with CFD studies. 4th [9] Krag B. S-WAKE Flight Test Report. Technical WakeNet Workshop “Wake Vortex Encounter”, Note SWAKE-D-221_2_1, IB 111-2001/40, German Amsterdam, 2000. Aerospace Center, Institute of Flight Research, Braunschweig, 2001. [25] Holzäpfel F. Authors personnel correspondence on realistic wake vortex modeling, 2001. [10] Sucipto T. Erweiterung eines Flugsimulationspro- gramms auf einer UNIX-Workstation. Konstruktiver [26] Gockel W. Design of an Autoland System for UAV Entwurf, TU Braunschweig, Institut für Flugführung, Application Using ATTAS as Demonstrator. Institute Braunschweig, 1993. of Flight Research, Report IB 111-2000/13, 2000. [11] Dunkel W. An improved Model of a Propeller- [27] Lambregts A A. Integrated System Design for Flight Driven Type Aircraft. 1. Mathmod Vienna, Wien, pp and Propulsion Control Using Total Energy 918-921, 1993. Principles. AIAA Paper 86-2143CP, 1986. [12] Escande B and Aureche Y. Trailing Vortices and [28] Hahn K-U. Effect of Wind Shear on Flight Safety. Safety. CEAS/AAAF Forum “Research for Safety in Progress in Aerospace Sciences, Vol. 26, pp 225- Civil Aviation”, Paris, 1999. 259, 1989. [13] Fischenberg D. Bestimmung der Wirbelschleppen- [29] Höhne G. A Model for the Pilot Behavior During Charakteristik aus Flugmessdaten. Deutscher Luft- Wake Vortex Encounters. Report No. IB 111- und Raumfahrtkongress, Stuttgart, Germany, 2002. 2001/41, German Aerospace Center, Institute of Flight Research, Braunschweig, 2001.

732.9 K.-U. Hahn

8 Figures and Tables

ground track leading a/c x g −χ RWY L heading leading a/c −ψ −ψ L= WVL

y WVL ground track trailing a/c heading trailing a/c

Fig.2: DLR’s Advanced Technologies Testing ILS track Aircraft System VFW614 ATTAS s e p a r a t io n

d is ψ t a n c e lane χ e p wak

x WVL right vortex line left vortex line Fig.1: Encounter Scenario

CLASS MASS HEAVY > 116 000 kg B757 - Fig.3: Smoke Generator Mounted on the Left LARGE > 18 600 kg Wing of ATTAS SMALL ≤ 18 600 kg Tab.1: Aircraft Classification after FAA

Generator A/C Follower A/C Separation Minima HEAVY HEAVY 4 NM / 7.4 km LARGE 5 NM / 9.3 km SMALL 6 NM / 11.1 km B757 HEAVY 4 NM / 7.4 km LARGE 4 NM / 7.4 km SMALL 5 NM / 9.3 km LARGE HEAVY 2.5 NM / 4.6 km LARGE 2.5 NM / 4.6 km SMALL 4 NM / / 7.4 km Fig.4: Do128 Test Aircraft of the University SMALL HEAVY 2.5 NM / 4.6 km of Braunschweig LARGE 2.5 NM / 4.6 km (with courtesy of Institute of Flight SMALL 2.5 NM / 4.6 km Guidance of TU Braunschweig) Tab.2: US Wake Vortex Separation for IMC 732.10 COPING WITH WAKE VORTEX

Leading A/C Initial Wake Vortex Core Radius Position - Light

- Medium - Rc0 = s b - Geo. Coordinates - Heavy ν* * - Origin: RWY - Rc0 = f( 1eff , T 1 )

Initial Circulation Core Radius Encounter Model . G x, x Γ = - Rc = const. - Aerodynamics 0 b .. ρ V 0TAS Γ - Geometry - Rcc00 = f(t, R , ) - R = f(t, ν* , T* ) - Inertia c1eff1 - . . .

tage

Atmosphere Vortex Decay Vortex Velocity Field Encounter Control - Lamb-Oseen ICAO Standard - Donaldson - FF-Moment Control & Bilanin - Burnham & Hallock Atmosphere - Greene - Proctor - FB-Disturbance Alleviation - Robins & Delisi - Modif. Lamb-Oseen - Auto-Pilot - Han et al. - Winckelmans - Sarpkaya - Holzäpfel

Fig.5: Block Diagram of nonlinear 6 DoF Encounter Simulation

1

] only diffusion −

[ no turbulence light turbulence 0 neutral

Γ / 0.8 − Γ higher strati fication z [m]

0.6 y [m]

0.4

0.2

t [s] t [s] 0 0 5 10 15 t* [−] Fig.6: Circulation Strength Calculated by the [m²/s] 5-15 P2P Model after HOLZÄPFEL [18] Γ

2.4 t [s] ] − [ 2.2 Fig.7: Uncertainty Bounds of the P2P Model c0 /r c 2 Compared with Memphis Data 1.8 (from [18])

1.6 T* 1.4 2

1.2 r = const. c 1 normalized core radius, r 0.8 0 5 10 15 normalized time, t* [−] Fig.8: Development of Vortex Core Radius

732.11 K.-U. Hahn

χ 15 RWY lateral velocity m/s Nominal Flight Path Angular State FF Trajectory Control Command Model Control δη δξ δζ

χ, ψ Position H,wKg, az FB measured Estimation Control Θ, Φ model output H,V ∆GLS, ∆LOC p, q, r GND a,y V -20 signal measurements V, H, V δ 10 GND Total Energy throttle vertical velocity Management γ m/s Vref ref Fig.10: Block Diagram of Autopilot and measured Autothrottle model output 0 1 24356 aircraft time, s data V(b) VK Flow Field WV FF Control Fig.9: Comparison of Measured Velocity Determination (inverse aerodynamics) V δη Field and Model Output (from [13]) signal measurements δξ δζ p, q, r FB Φ,, Θ Ψ Control

Φ Θ Ψ ref, ref, ref Fig.11: Block Diagram of Vortex Controller

ξ ξ* > 0.1 * > 0.1 ξ* > 0.1 ξ 30 ξ* > 0.2 30 * > 0.2 30 ξ* > 0.2 a) ξ* > 0.3 b) ξ* > 0.3 c) ξ* > 0.3 ξ* > 0.5 ξ* > 0.5 ξ* > 0.5 ξ ξ* > 1.0 ξ 20 * > 1.0 20 20 * > 1.0

10 10 10

0 0 0 dH in [m] dH in [m] −10 −10 dH in [m] −10

−20 −20 −20

−30 t* = 0 −30 t* = 3.8 −30 t* = 5.8

−50 0 50 −50 0 50 −50 0 50 dy in [m] dy in [m] dy in [m] ξ ξ* > 0.1 ξ* > 0.1 * > 0.1 ξ ξ* > 0.2 ξ* > 0.2 30 * > 0.2 30 30 ξ d) ξ* > 0.3 e) ξ* > 0.3 f) * > 0.3 ξ ξ* > 0.5 ξ* > 0.5 * > 0.5 ξ ξ ξ* > 1.0 20 * > 1.0 20 * > 1.0 20

10 10 10

0 0 0 dH in [m] dH in [m] dH in [m] −10 −10 −10

−20 −20 −20

* −30 t* = 9.4 −30 t* = 11.7 −30 t = 11.7

−50 0 50 −50 0 50 −50 0 50 dy in [m] dy in [m] dy in [m]

Fig.12: Required Aileron Deflection Behind a Vortex Generation Aircraft (SMALL {4.35t} behind LARGE {79t})

732.12 COPING WITH WAKE VORTEX

35

30 B Fig.13: Parameters of the Hazard Boundary * A Approximations for ξ ≤0.3 25 (SMALL {4.35t} behind LARGE {79t}) b 20 a dimensions of the rectangle: A, B 15 semi-axes of the ellipse: a, b

10 aircarft combination: 1) a 2) a

Parameters A, B, a, b in [m] 5 b b A A B B 0 0 1 2 3 4 5 6 Separation Distance SD in [nm]

66 4 8 in [m/s] 64 in [m/s] 6 WV

WV 2 62 4 0 60 2

airspeed V in [m/s] 58 −2 0

0 20 40 60 lat. vortex vel. v 0 20 40 60 0 20 40 60 vert. vortex vel. w ]

− time t in [s] time t in [s] time t in [s] 500 vortex line 30 1 * in [ ILS track right vortex line ξ

450 in [m] 20 g 0.5 400 10 0 350 0

altitude H in [m] left vortex line −0.5 300 lat. distance y −10

norm. aileron def. 0 20 40 60 −10000 −8000 −6000 −10000 −8000 −6000 distance to THR x in [m] distance to THR x in [m] time t in [s] g g

10 10 ] 1.5 −

in [m/s] 5

8 in [ Kg z in [Grad]

Φ 0 6 1 −5 4 −10 load factor n

2 bank angle −15 0.5

vertical speed w 0 20 40 60 0 20 40 60 0 20 40 60 time t in [s] time t in [s] time t in [s] Fig.14: Wake Vortex Encounter Using Different Controllers (SMALL {4.35t} behind LARGE {79t}) - - - - autopilot  vortex controller  vortex lines - - - - ILS

732.13 K.-U. Hahn

(SMALL {4.35t} behind LARGE {79t}) 64 4 8 in [m/s] in [m/s] 6 WV

62 WV 2 4 60 0 2

airspeed V in [m/s] 58 −2 0

0 20 40 60 lat. vortex vel. v 0 20 40 60 0 20 40 60 vert. vortex vel. w ]

− time t in [s] time t in [s] time t in [s] 500 vortex line 30

* in [ ILS track right vortex line ξ

1 450 in [m] 20 g

0.5 400 10

0 350 0

altitude H in [m] left vortex line

300 lat. distance y −10

norm. aileron def. 0 20 40 60 −10000 −8000 −6000 −10000 −8000 −6000 distance to THR x in [m] distance to THR x in [m] time t in [s] g g

6 ] 1.4

5 − in [m/s]

5 in [ Kg z

in [Grad] 1.2 4 Φ 0 1 3

−5 load factor n

2 bank angle 0.8

vertical speed w 0 20 40 60 0 20 40 60 0 20 40 60 time t in [s] time t in [s] time t in [s] Fig.15: Wake Vortex Encounter Using Combined Autopilot and Vortex Controller (SMALL {4.35t} behind LARGE {79t}) ] 10 ] 1 500 − − 1 vortex line

* in [ ILS track * in [ in [m/s]

ξ 0.5 ξ

WV 0.5 450 0 5 0 400

−0.5 altitude H in [m] −0.5 norm. aileron def.

norm. aileron def. −1 350 vert. vortex vel. w 0 −1 0 50 −10000 −8000 −6000 0 50 0 50 time t in [s] distance to THR x in [m] time t in [s] time t in [s] g 5 1.05 500 5 vortex line ] −

ILS track in [ z in [Grad]

450 Φ in [Grad] 0 1 Φ 0 400 load factor n altitude H in [m] bank angle

bank angle −5 0.95 350 −5 0 50 0 50 −10000 −8000 −6000 0 50 time t in [s] time t in [s] distance to THR x in [m] g time t in [s] Fig.17: Wake Vortex Encounter Left of the Fig.16: Encounter of a Wake Vortex with an Left Vortex Line Age of 50 Seconds (ROT) (Wake Vortex Age: 50s)

732.14