24TH INTERNATIONAL CONGRESS OF THE AERONAUTICAL SCIENCES

A SIMPLIFIED HAZARD AREA PREDICTION (SHAPE) MODEL FOR WAKE VORTEX ENCOUNTER AVOIDANCE

Klaus-Uwe Hahn, Carsten Schwarz, Holger Friehmelt German Aerospace Center, Institute of Flight Systems Lilienthalplatz 7, 38108 Braunschweig, Germany

Keywords: Safety, Wake Vortex, Encounter, Hazard Area, Separation Distance

Abstract List of Symbols and Abbreviations This paper gives details from offline simula- a/c tions, full-flight simulator studies and the ATTAS Advanced Technologies Testing Aircraft preparation of a flight test experiment using the System b span In-Flight Simulation capability of DLR’s (Ger- cg center of gravity man Aerospace Center) test aircraft ATTAS DoF degrees of freedom (Advanced Technologies Testing Aircraft) to DLR German Aerospace Center validate safe boundaries around a wake vortex IFS in-flight simulation for encounter avoidance. The so developed haz- IMC instrumental meteorological conditions ard zone is an element in the Wake Vortex Pre- MTOW maximum take-off weight diction and Observation System which is under NLR National Aerospace Laboratory development in the frame of the DLR “Wake P2P probabilistic two phase model * Vortex II” project to reduce aircraft approach RCR roll control ratio (= x ) separation and to improve flight safety. t time co-ordinate In any case the separation distance of ap- SHA simplified hazard area proaching aircraft has to be determined in such SHAPe simplified hazard area prediction v lateral velocity component a way that the approach corridor is free of haz- V airspeed (no index), velocity ardous vortex flows for the trailing aircraft. The VMC visual meteorological conditions question to be answered is what is the meaning w vertical velocity component of “hazard free”? The presented approach for ZFB Zentrum für Flugsimulation Berlin solving this fundamental problem follows the G circulation idea that it is possible to define an area around h deflection a wake vortex outside which the vortex flow is x deflection definitely not hazardous to an aircraft. The rea- z deflection son for this “inverse” definition is the fact that D difference a clear criterion of what is hazardous is difficult F bank angle to set up (especially if a pilot is in the loop). The Q pitch angle so defined boundaries of the hazard zone can y azimuth easily be determined by the developed Simpli- indices fied Hazard Area Prediction (SHAPe) Model. It g geodetic allows the computation of hazard zone dimen- K kinetic (refers to flight path) sions for any aircraft pairing (real or non exist- max maximum ing aircraft). This potential of a universal appli- nom nominal cation is based on the parameterization of the W wind relevant aircraft wake vortex characteristics WV wake vortex which are fitted by simple functions. WVL wake vortex line * denotes normalized parameters 1 K.-U. Hahn, C. Schwarz, H. Friehmelt

1 Introduction - wind (e.g. cross wind, wind shear) atmospheric turbulence The well known phenomenon of wake vortices - VMC, IMC (ceiling, visibility) behind a lift producing wing can adversely af- - fect flight safety if encountered by trailing air- · Aircraft Control craft. The strength of the vortices increases with - individual pilot behavior (skill) the weight of the vortex generating a/c. There- - performance of automatic controllers fore, weight dependent separation distances The list above claims not to be complete. But it have been established for approach and landing underlines the difficulties (especially if a pilot is to avoid dangerous wake vortex encounters. in the loop) to set up a clear criterion of what is These proven separation distances have to be hazardous (in terms of constraints leading un- investigated to discover possible margins which questionably into an unsafe situation) as many could be used to solve the current and future ca- attempts illustrate. pacity problems at airports. This demand forms Approaching the problem from a safety the need for more flexible separation procedures point of view leads to the idea that it is possible taking into account the actual weather situation to define an area around a wake vortex outside and the parameters of the individual a/c pairing. which the vortex flow is definitely not hazard- Within this scope a Wake Vortex Prediction and ous to an aircraft. This does not imply that any Observation System is under development in the encounter (slight penetration) of the so defined frame of the DLR “Wake Vortex II” project to Hazard Zone must result in a threatening situa- reduce aircraft approach separation and to im- tion. But the area outside the Hazard Zone has prove flight safety [1]. to be an absolute Safety Zone. Following this approach it is necessary to identify safe boundaries around a wake vortex 2 Safety and Hazard defining this Hazard Zone. The today policy to In any case the separation distance of approach- avoid hazardous situations is formulated by the ing aircraft has to be determined in such a way obligation that “no vortex should be encoun- that the approach corridor is free of hazardous tered by intention [2]”. But what means “no vor- vortex flows for the trailing a/c. The question to tex encounter”. A qualitative statement can eas- be answered is what is the meaning of “hazard ily be made: the more distant the encountering free”? For solving this fundamental problem a/c flies away from the vortex cores, the less are two different methodologies are applicable. One the noticeable effects of an encounter. From this approach can be to define the constraints for declaration at least the following two conditions which an encounter becomes hazardous. The for a “no vortex encounter” situation can be de- difficulty of this procedure is the great variety rived: of possible encounter situations and the numer- · no vortex core encounter ous parameters of influence: · the a/c behaviour should be not abnormal compared to a flight in normal atmospheric · Encounter Scenario disturbances (natural gusts and turbulence) - state of the a/c - orientation of flight path and wake vortex The latter statement is the more demanding one (encounter angles) which has to be regarded for the development of - relative positions of wake vortex lines safe boundaries around a Hazard Area. - encounter altitude · Aircraft Pairing - vortex generator (a/c dimension, vortex 3 Simplified Hazard Area (SHA) strength, shape of velocity distribution) For parallel-like encounters the aircraft’s roll - follower (mass, a/c dimension, airspeed, response is the dominating motion. The worst aerodynamics, cg, control power) case occurs when the wake vortex is parallel to · Meteorological Conditions the approach path and the following a/c is per- 2 A SIMPLIFIED HAZARD AREA PREDICTION (SHAPE) MODEL FOR WAKE VORTEX ENCOUNTER AVOIDANCE

manently exposed to the vortex flow field in a 4.1 Numerical Offline Simulations quasi stationary flight. The effect on roll within The numerical offline simulation is the nowa- a wake vortex flow field then can be calculated days universal state-of-the-art analysis tool. The using the required control power normalized by reliability of the simulation results depends on the maximum available control power of the the quality of the applied models. The choice of encountering a/c expressed in terms of aileron models was based on the rule “as complex as deflection [3]. The wake vortex induced roll necessary but as simple as possible”. Neverthe- moment can be compensated if equation (1) ap- less, great store was set by realistic modeling. plies. * x = x x < 1 4.1.1 Modeling and Validation max (1) A complex simulation system [3] was devel- Areas of various roll control demands can be oped including the main modules: defined by this normalized aileron deflection x* · Vortex Generating Aircraft which is equal to the roll control ration (RCR) (® flow field of a wake vortex) known from other publications. Thus, x* (re- · Encountering Aircraft spectively RCR) is a suitable measure to quan- (® 6 DoF simulation and wake vortex/ tify the effect of a wake vortex flow field on an aerodynamic interaction model) aircraft. Fig. 1 illustrates the flow field behind a · Encounter Control generic vortex generating a/c (MTOW=79 to) in (® ILS model and / ) * terms of normalized aileron deflection x for a While the basic simulation system is based on light aircraft (MTOW=4.35 to). Different col- standard modules and models particular effort ored regions represent different levels of re- was dedicated to include the special effect of quired normalized aileron deflections. wake vortices. In the frame of the S-Wake pro- The complex shape of the areas for a spe- * ject [4] sponsored by the European Commission cific nominal x nom can be conservatively ap- real wake vortex encounter flight experiments proximated by a simple rectangle which covers were executed to gather a valuable high quality the respective region (Fig. 1). SHA includes the data base from real flight tests. For the flight closer regions around the vortex lines. As a re- tests DLR’s Advanced Technologies Testing * sult, for small values of x nom all other encounter Aircraft System VFW614/ATTAS was used as parameters of interest (e.g. roll acceleration, an- vortex generator. The encountering aircraft were gle of attack and load factor, lateral accelera- the NLR’s Cessna Citation and the Dornier tion) are covered and remain within acceptable Do128 test aircraft of the University of Braun- limits in terms of a/c controllability (which will schweig [Fig. 2]. The latter is ideally suited for be shown later). So if the term encounter is used in-situ wake vortex measurements since it has 4 in this paper the respective situation has to be different stations with flow probes (nose boom, understood to be rather a passing by than a hit wing tips, fin). The collected data base was used of a vortex. The dimension of this SHA depends to validate different approaches of wake vortex * on the amount of x nom which is tolerable with- flow field and aerodynamic interaction models out any adverse effect on safety. The appropri- by means of parameter identification and flight ate value has to be established and validated for path reconstruction techniques [5, 6]. At the end a reliable sizing of the SHA. of this process reliable models for wake vortex encounter simulation were available. Fig. 3 shows the comparison between 4 Validation of Safe SHA Boundaries simulation and flight test results of the Do128 The validation of safe boundaries for SHA has encountering the wake vortex of ATTAS [5]. been planned to be performed in 3 steps with The black curves give the a/c response meas- increasing level of reality. ured during the encounter from real fight tests. The red curves represent the output from the

3 K.-U. Hahn, C. Schwarz, H. Friehmelt

numerical simulation. Especially the roll and ters were conducted in a simulator study using a vertical axes illustrate the very good fit due to full flight simulator. This simulator (see Fig. 6) the high quality of the simulation system. is operated by the “Zentrum für Flugsimulation Berlin” (ZFB). It is used for airline pilot training 4.1.2 Simulation Results as well as for scientific research of the Univer- Accepting that the rectangular SHA has to be sity of Berlin. For the latter application a special avoided the worst case situation for an aircraft is simulation computer (Scientific Research Facil- a flight along its boundaries. Numerous offline ity, SFR) can be connected to the simulator. On simulations have been performed using the de- this simulation computer flight mechanics mod- veloped simulation tool. As an example the els are available representing different aircraft. automatic controlled flight along the upper * The SFR generates the data to control the mov- boundary of an SHA belonging to x nom = 0.3 is ing cockpit. The cockpit itself can be chosen to illustrated in Fig. 4. The wake vortex lines are represent an A330 or A340 heavy transport type positioned in such a way that the nominal ILS aircraft. flight path passes exactly the upper boundary of On the simulator described above the flight SHA with an encounter azimuth angle of DyWVL mechanics model of a small twin engine turbo- = 5° between approach path and vortex lines. prop a/c similar to a Do228 (MTOW= 5.4 t, see The respective time histories of the aircraft re- Fig. 7) was implemented. It is essential to note sponse are given in Fig. 5. that the dynamics of a commuter a/c have been The SHA vanishes if the flow field pro- controlled from a large transport a/c cockpit. duces only aileron deflection demands below * * The most important difference is the control of the acceptable threshold of x < x nom = 0.3 (due the roll and pitch axes via side-stick in the simu- to the vortex decay). When the SHA no longer lator, while the real Do228 a/c is controlled by a exists even a flight through the vortex core is no conventional control column and wheel. The more a problem. From the simulations the fol- pilots with a Do228 type rating stated that the lowing main conclusions were drawn: real a/c is easier to handle than the simulator · As expected the encounter situations down due to the described mismatch. Therefore, it can the left and right vertical boundary show mir- be assumed that the results coming from the ror symmetry. Flights along the upper and simulator experiments were not overoptimistic lower boundary have to be differentiated. at all. Consequently, the derived safe boundaries · Since flight path deviations from the nominal of the SHA can be considered to be conserva- flight path occur the actual normalized ai- tive. * leron deflection x respectively RCR during 4.2.2 Experiment Design and Pilot Task the wake vortex flow penetration can become For the encounter investigation the standard slightly higher than the nominal value for situation of a 3° ILS approach was chosen. The which the SHA is designed. Nevertheless, the initial conditions were defined at a distance of normalized aileron deflection is a well suited 6nm to the threshold, the a/c was established on measure for hazard avoidance. the nominal path of the ILS. The gear was up · For automatic control (AP/AT) a normalized * and the flaps were in approach configuration. aileron deflection of x nom = 0.3 seems to The reported wind was 15 kt coming 30° from provide safe margins to hazardous situations the left. During the approach the a/c was ex- in terms of flight state and path deviations. posed to different levels of turbulence in the range between no and light turbulence. The 4.2 Flight Simulator Study visibility in VMC was 50km. Additionally en- counters in CAT I conditions with a decision 4.2.1 The Full Flight Simulator height of DH = 222ft were investigated. The For the next step of validation of the SHA IMC was defined having a visibility of 4km and boundaries manual flown wake vortex encoun- a ceiling from 450ft down to 225ft. Different

4 A SIMPLIFIED HAZARD AREA PREDICTION (SHAPE) MODEL FOR WAKE VORTEX ENCOUNTER AVOIDANCE

encounter scenarios were defined by the pa- Pilot License Flight Hours Do228 Rating rameters A CPL-IFR 300 no · max nominal normalized aileron deflection B ATPL 6500 yes * C ATPL 4000 yes resp. RCR (x nom = RCR nom = 0.3, 0.25, 0.2) Tab. 1. Experimental Pilots · flights along upper, lower, left, and right boundaries of SHA parameter acceptable limits · encounter angles glide slope dev. -0.5 DOT = DGS = 1 DOT (vertical: 3° , horizontal: 0° / ±5° / ±10°) localizer dev. DLOC = ±1 DOT bank angle dev. · encounter altitude DF = ±20° (3m < H < 250m) indicated airspeed -5 kt = DVIAS = +15 kt descent speed dzg/dt = 1000 ft/min The pilots’ task was to track the ILS signal roll rate dev. Dp = ±15°/s to perform the approach and landing. They had load factor 0.6 = n = 1.6 to follow the common crew procedures (call z Tab. 2. Limits for Acceptable Conditions outs) and to establish the aircraft’s landing con- figuration (flaps, gear down). An approach se- Similar to the off-line simulations the quence was completed after touch down and complete SHA border (upper, lower, left, and roll-out or in case of a go-around when a safe right boundaries) was a matter of investigation. climb phase was established. And again the wake vortex is parallel to the After each single run the pilots had to horizontal plane. make their ratings on controllability, pilot de- A typical example from [8] for an approach * mand, aircraft excursions, and over all hazard. with a max nominal x nom =RCR nom = 0.2 shows The rating scales are based on [7] with some Fig. 9. These are the results of a flight along the modifications (see Fig. 8). E.g., only four levels right border of the SHA. The encounter position of ratings are established. It was known from is indicated by a red line. It takes place in IMC other experiments that the pilots were unhappy at a height of 78m. The effect of the vortex flow with more levels of gradations for which they is not significant compared to other flight path found it really difficult to differentiate smaller deviations occurring during this approach. The nuances. For the determination of acceptable main effect can be observed in the roll-axis but SHA boundaries the subjective pilot ratings us- the max bank angle does not exceed |F|<10°. ing Fig. 8 have to be better than 4. The pilot roll input is about 25% of the max 4.2.3 Man-in-the-Loop Simulation Results available roll control power. So the design value With the above experimental setup 3 simulator of RCRnom = 0.2 is exceeded which was already campaigns were executed with a total of 82 ap- observed from the off-line simulations. The nu- proaches carried out by 3 different pilots (see merical analysis shows only very small penetra- Tab. 1). Two of the pilots have a type rating for tion of the limits defined in Tab. 2. The pilot the a/c implemented on the simulator and one ratings coming from Fig. 8 are 2 for all catego- pilot was still a beginner. The respective data ries. Thus, the a/c behavior during encounter analysis is still going on but first results are can be considered to be acceptable. available [8]. Fig. 10 shows the average ratings of the pi- lot for the investigated RCRs. The min and max During the simulator experiments manifold ratings for each category are indicated by mar- parameters were recorded including pilot inputs, * a/c states and trajectory. For the objective gins (black lines). It can be seen that for x nom = analysis of numerical data recorded the limits RCR = 0.2 the average of each rating is 2 or between acceptable and unacceptable encounter even better. But the more important message is situations (see Tab.2) were derived carefully that no ratings worse than 3 occur. This implies based on information from different sources [9, that all encounters were at least acceptable. The 10, 11, 12, 13]. numerical analysis of the flights shows only ex-

5 K.-U. Hahn, C. Schwarz, H. Friehmelt

tremely rare situations of very slight penetration has already been successfully demonstrated of the limits defined in Tab. 2 and only for neg- [16]. The flight tests for the validation of the ligible periods. From the simulator results the SHA boundaries are currently in preparation. following main conclusions can be drawn [14]: The first tests will be executed end of this year. · The roll control inputs applied by the pilots can become higher than the reference value for the SHA design case. Nevertheless, the 5 The SHAPe Concept normalized aileron deflection/RCR is a well The dimension of the SHA can be calculated for suited measure for hazard avoidance. each a/c pairing as described in chapter 3 if the * · For manual control a max nominal normal- tolerable x nom is known. But there are different * ized aileron deflection of x nom =0.2 seems levels of abstraction for the prediction of these to provide acceptable wake vortex encounters dimensions. For this prediction the Simplified covering all the other encounter parameters Hazard Area Prediction (SHAPe) Model has of interest (e.g. roll acceleration, angle of at- been developed. tack and load factor, vertical speed and lat- eral acceleration) which stay in between ac- 5.1 Levels of Abstraction ceptable limits. SHAPe has 2 levels of abstraction for the vortex generating a/c and 3 levels for the encountering 4.3 In-Flight Simulation a/c (see Fig. 12) [17]. The final validation of the SHA boundaries will · Vortex Generating A/C be done using the most realistic simulation tool - The 1st level (highest quality) uses data of which is the In-Flight Simulation. The IFS is the real a/c (wing span, and relevant data DLR’s standard method to investigate pilot-in- for circulation calculation) the-loop behavior in real flight conditions [15]. - For the 2nd (lowest) level the information DLR’s full fly-by-wire/light testbed ATTAS is needed are derived from parameterized a/c especially designed for this application. data (see Fig. 13) using the MTOW of the The principle of the IFS is roughly respective a/c sketched in Fig. 11. The complete simulation · Encountering A/C system to be simulated in real flight (a/c model - The 1st level (highest quality) uses the to be simulated, aerodynamic interaction model, comprehensive data sets of aerodynamics wake vortex model) has to be implemented on and geometry of the actual a/c the computers onboard the host aircraft ATTAS. - For the 2nd (lower) level only the wing The experimental pilot sitting on the left cockpit aerodynamics and geometry of the real a/c side of ATTAS is flying the simulated a/c using is considered (no effects of or tail his real controls. These inputs are fed into the are taken into account) onboard computers stimulating the model a/c - For the 3rd (lowest) level the information which reacts on his inputs and on the effects needed for are derived from parameterized coming from the virtual wake vortex flow. The data (see Fig. 13) using the MTOW of the resulting model a/c states are fed into the model respective a/c following control system. The model following For the parameterization all information from controller calculates the control surface deflec- existing a/c available to the author has been tions of the host a/c which are necessary to used. Then the a/c data relevant for wake vortex make the host a/c behave like the simulated a/c. encounters have been correlated with the So the flight states of the host a/c experienced MTOW (Fig. 13). Fitting curves have been ap- by the experimental pilot are in coincidence proximated which allow the determination of with the flight states of the simulated a/c. wake vortex relevant parameters only knowing That the concept described above can be the MTOW. Due to the uncertainty of this applied to wake vortex encounter experiments method only the fits representing the worst case 6 A SIMPLIFIED HAZARD AREA PREDICTION (SHAPE) MODEL FOR WAKE VORTEX ENCOUNTER AVOIDANCE

situation have to be used for the SHA calcula- wake vortices can be assumed. The ellipse can tion. be conservatively simplified by a rectangle. As an example: for the determination of the DLR’s Probabilistic Two Phase Wake Vortex initial circulation of a vortex generating a/c with development and decay model (P2P) [18] then a specific MTOW the red curve of the respec- predicts an area of probable vortex locations tive diagram in Fig. 13 has to be used. This will after a period Dt. It is assumed that no vortex produce the max possible circulation for the re- will leave this P2P area. At the corners of this spective MTOW. If we are looking for the max area the hazard zone computed by SHAPe is available roll control power Cl(xmax) of an en- superimposed. The envelope around the indi- countering a/c having a specific MTOW, the vidual hazard zones is the overall hazard zone min amount of this parameter coming from the which is not allowed to be penetrated. For the respective diagrams in Fig. 13 has to be chosen determination of safe separation distances for an to consider the worst case situation. The re- individual aircraft pairing the period Dt of two quired value is specified by the red curve. consecutive a/c has to be determined in such a If the above described generic approach us- way that no overlapping of approach corridor ing the parameterized a/c data is applied the re- and over all hazard zone exists. This calculation sulting dimensions of the SHA are always big- is done at gates along the approach path (Fig. ger (see Fig. 14) than using actual data of spe- 17). cific a/c pairings. So this approach can be con- In Fig, 16 it is assumed that SHA always sidered to be conservative for the prediction of covers a vortex pair. This is not the case if the 2 SHA. The parameterized approach has the ad- vortices of a wake split and move separately. vantage that any generic a/c (being in line with Investigations show that SHAPe can also work existing aircraft) can be treated using SHAPe. with hazard areas of a single vortex without any changes of the presented methodology [19]. 5.2 Computation Process All levels of abstraction of SHAPe can be used 7 Summary and Conclusion to calculate the SHA dimensions as sketched in Fig. 12 to perform an on-line process. Another The presented approach to solve the fundamen- possibility offers the pre-calculation of SHA tal problem of the definition of a “hazard free” dimensions and their storage in multi dimen- area follows the idea that it is much easier to sional tables (see Fig. 15). This can be done for define an area around a wake vortex outside specific real a/c pairings (if respective data are which the vortex flow is definitely not hazard- available) as well as using the parameterized ous to an aircraft rather than looking for en- approach for any generic a/c pairing. For the counter boundaries and constraints leading un- determination of the SHA the actual (time de- questionably into a hazardous situation. The reason for this “inverse” definition is the fact pendent) circulation G of the vortex generating that a clear criterion of what is hazardous to an a/c must be known. Respective models are aircraft is difficult to set up (especially if a pilot available, e.g. ref. [18]. is in the loop) as many attempts illustrate. For not too close vortex fly-bys the threat 6 Dynamic Separation Distances can easily be assessed by the required control power expressed in terms of normalized aileron The SHAPe model is an important element in * deflection x = |x/xmax| also known as RCR. A DLR’s Wake Vortex Prediction and Observa- certain fraction of this expression is identified to tion System. Its application is sketched in Fig. present a safe limit in terms of wake vortex en- 16 which shows the cross section of an ILS ap- counter depth. An rectangular Simplified Haz- proach. The vortex generating aircraft shows ard Area is defined covering this region around some flight path deviations forming the ellipti- a wake vortex. From offline simulations and cal approach corridor where the generation of 7 K.-U. Hahn, C. Schwarz, H. Friehmelt

from simulator experiments it was found that for [8] Hahn K-U, Schwarz C and Kloidt S. Full-Flight Si- * mulatorstudie zur Verifizierung von Wirbelschlep- automatic approaches nominal values of x nom = * pengefährdungsgrenzen. DLR Technical Report, IB 0.3 and for manual approaches x nom = 0.2 seem 111-2004/42, DLR Institute of Flight Systems, to provide safe encounter conditions. The latter Braunschweig, Germany, 2004. value will finally be validated by real flight ex- [9] Sammonds W A, Stinnet jr G W and Larsen W E. periments using DLR’s In-Flight Simulator Wake Vortex Encounter Hazard Criteria for Two ATTAS. Aircraft Classes. NASA TM-X-73113, Ames Re- The SHA dimensions can easily be deter- search Center, Moffett Field, 1976, (FAA RD-75- 206, 1976). mined by the Simplified Hazard Area Prediction [10] N.N. Operations Manual – Flight Operations Proce- (SHAPe) Model. It allows the computation of dures. Flight Operations Manager, Deutsche Luf- hazard zone for any aircraft pairing (real or non thansa AG, Frankfurt/Main, Germany, 2003. existing aircraft). This potential of a universal [11] Luckner R.. Requirements for the Flight Control application is based on the parameterization of Laws of the EFCS Demonstrator Aircraft (VFW614, the relevant aircraft wake vortex characteristics G15). Technical Note, TN-EF-012/96, Daimler-Benz Aerospace – Airbus, Hamburg, 1996. which are fitted by functions. [12] Magni J-F, Bennani A and Terlouw J. Robust Flight The SHAPe model is a important element Control – A Design Challenge. Lecture Notes in Con- in DLR’s Wake Vortex Prediction and Observa- trol and Information Sciences 224, Springer-Verlag tion System. Using the wake vortex develop- Berlin Heidelberg New York, 1997. ment and decay output of DLR’s P2P model [13] Held C. Analyse eines Flugregelungssystems zur SHAPe can determine safe separation distances Reduzierung des Gefahrenpotentials von Wirbel- by predicting the conditions when no longer an schleppen währen der Landephase. Diplomarbeit, In- stitut für Luft- und Raumfahrttechnik, Technische overlapping of approach corridor and over all Universität Berlin, 2003. hazard zone exists. [14] Hahn K-U and Schwarz C. Hazard Avoidance Crite- rion - Investigation of WVE Hazard Zones on the 8 References A330 / A340 Simulator at TU Berlin. Presentation on the 1st Workshop of WG5, WakeNet2-Europe, Airbus [1] Gerz T. Wirbelschleppe – Minimierung, Vorhersage, Deutschland, Hamburg, Germany, 2004. Beobachtung und Erkennung von Wirbelschleppen [15] Hanke D. In-Flight Simulation – An Advanced Simu- ziviler Verkehrsflugzeuge. Projektplan, German Ae- lation Technique. 7th Annual International Sympo- rospace Center, Germany, 2001. sium "Aviation Technologies of the XXI Century: [2] NN. IFALPA Wake Vortex Policy. IFALPA News, New Aircraft Concepts and Flight Simulation", ILA International Federation of Air Line Pilots Associa- 2002, Berlin, 7-8 May 2002, TsAGI Central Aerohy- tion News Letters, 1998. drodynamic Institute, 2002. [3] Hahn K-U. Coping with Wake Vortex. Proc of the [16] Reinke A, Bauschat M and Leißling D. Simulation 23rd ICAS Congress, Toronto, Canada, paper no 732, des Einfluges in Wirbelschleppen mit dem ATTAS pp 732.1-732.14, 2002. Flugsimulator. Proc of the Deutscher Luft- und [4] de Bruin A et al. S-WAKE - Assessment of Wake Raumfahrtkongress, DGLR-JT2003-245, München, Vortex Safety, Publishable Summary Report, Euro- Germany, 2004. pean Commission, Proj.No: G4RD-1999-0099, NLR- [17] Hahn K-U and Schwarz C. The Wake Vortex En- TP-2003-243, NLR, Amsterdam, 2003. counter Avoidance-Computation of Safe Aircraft [5] Fischenberg D. Bestimmung der Wirbelschleppen- Separations using the SHAPe Concept. Presentation Charakteristik aus Flugmessdaten. Proc of the Deut- on the 1st Workshop WakeNet2-Europe, Heathrow scher Luft- und Raumfahrtkongress, DGLR-JT2002- Control Tower, London, United Kingdom, 2003. 170, Stuttgart, Germany, 2002. [18] Holzäpfel F. A Probabilistic Two-Phase Wake Vor- [6] Fischenberg D. Wake encounter flight tests per- tex Decay and Transport Model. Report No. 163, formed within S-Wake project. Presentation on the German Aerospace Center, Institut für Physik der 1st Workshop WakeNet2-Europe, Heathrow Control Atmosphäre, Oberpfaffenhofen, 2001. Tower, London, United Kingdom, 2003. [19] Schwarz C and Hahn K-U. Gefährdung beim Einflie- [7] Sammonds R I and Stinnet jr. G W. Hazard Criteria gen in Wirbelschleppen. Proc of the Deutscher Luft- for Wake Vortex Encounters. NASA Technical und Raumfahrtkongress, Jahrbuch 2003, DGLR- Memorandum, NASA-TM-X-62473, Ames Research JT2003-242, München, Germany, 2004. Center, Moffett Field, California, 1975.

8 A SIMPLIFIED HAZARD AREA PREDICTION (SHAPE) MODEL FOR WAKE VORTEX ENCOUNTER AVOIDANCE

9 Figures and Tables Dornier Do128 VFW614/ATTAS

0.1 < x* £ 0.2

0.2 < x* £ 0.3

0.3 < x* £ 0.5 Cessna Citation

0.5 < x* £ 1.0

1.0 < x*

encounter a/c ATTAS leading a/c with smoke generator for vortex indication Fig. 1. Required Aileron Deflection in a Wake Fig. 2. Test Aircraft for Wake Vortex Encounter Vortex and Simplified Hazard Area Flight Experiments

Fig. 3. Wake Vortex Encounter Simulation (Small A/C Behind Large) [5]

cross

section

vertical

plane vortex lines ILS track

vortex lines helicopter ILS track Fig. 4. Flight Along the Upper Boundary of the view * Simplified Hazard Area (x nom = 0.3) (flight path angle g = -3°, encounter azimuth Dy = 5°) WVL 9 K.-U. Hahn, C. Schwarz, H. Friehmelt

Fig. 6. ZFB Full Flight Simulator

Fig. 5. Time Histories of a Flight Along the Upper Boundary * of the Simplified Hazard Area (x nom = 0.3)

Fig. 7. Do228 Aircraft (MTOW=5.4t)

table p acce

table p

Fig. 8. Pilot Rating Scales nacce u

10 A SIMPLIFIED HAZARD AREA PREDICTION (SHAPE) MODEL FOR WAKE VORTEX ENCOUNTER AVOIDANCE

* Fig.9. Trajectory and Time Histories of a Vortex Encounter (x nom = 0.2) from Simulator Experiment

Max. Normalized Aileron Deflection

ing * t (x nom = RCRnom)

* x nom = 0.30 pilot ra

xnom = 0.25

xnom = 0.20

Fig.10. Pilot Ratings of Wake Vortex Encounters with Different Strength 11 K.-U. Hahn, C. Schwarz, H. Friehmelt

Fig.11. Principle of the In-Flight Simulation for Wake Vortex Encounter Experiments

existing aircraft

Õ

Fig.13. Parameterization of Aircraft Data as a Function of MTOW

Fig.12. Levels of Abstraction of the Simplified Hazard Area Prediction (SHAPe) Model Fig.14. Comparison of Computed SHA Dimensions 12 A SIMPLIFIED HAZARD AREA PREDICTION (SHAPE) MODEL FOR WAKE VORTEX ENCOUNTER AVOIDANCE

Fig.15. Application of the SHAPe Concept for the Generation of Dynamic Separation Distances

Fig.16. Application of the SHAPe Concept for the Generation of Dynamic Separation Distances

Fig.17. Integration of SHAPe in DLR’s Prediction and Observation System

13