ICA 2000 CONGRE CONFIGURATION OF AN UNMANNED GROUND EFFECT VEHICLE M. Millar, L. mrek Department of Aerospae Engineering University of Glasgow Keywords: UAV, Ground Effet, Experimental Testing, Applied Aerodynamis Abstrat Tis paper looks at te proesses and onsiderations involved in using an unmanned aerial veile (UAV) as a ground effet test veile. IT looks at te possible resear avenues and suggests any problems wi are foreseen in developing su a veile. Tere is te added issue wit te fat tat te UAV is not a speifially designed wing-in-ground (WIG) veile, so problems are seen tere and te neessary modifiations made. Nomenlature α = Angle of attak AR = Aspet ratio b = pan D = Drag fore = Mean aerodynami ord L Lα L mα m H H m θ C = Lift oeffiient = Piting moment oeffiient = Lift urve slope = Inrement of lift oeffiient wit eigt = Moment urve slope = Inrement of piting moment wit eigt = Equivalent boundary layer eigt at te slot = Boundary leyer eigt at te slot = Ground plate learane equal to te ontration setion inlet eigt = Contration setion eigt at te slot T = Trust fore U U U θ vˆ = = Freestream veloity = Veloity troug te ontratio under te slot = Equivalent boundary layer veloity Non - dimensional speed W = Weigt 1 Introdution Te priniple of wing-in-ground-effet or (WIG) as been known for a long time. Ground effet is a natural penomenon tat is experiened as an aerofoil approaes te ground. Upon wi its lifting ability inreases and te drag redues, in simple terms. Te onsequene of tis is tat te lift to drag ratio inreases as te pattern of irulation around te wing anges and tus te overall effiieny inreases. Tis pattern of air around te wing auses a usion on te underside of te wing to develop. Figure 1. demonstrates te surrounding air flows of a wing in free fligt and in ground effet. Operating a raft in ground effet as te same effet as aving a mu larger wing area but witout te atual pysial struture, weigt or drag assoiated wit it. 133.1
M. Millar, L. mrek ~0.8 m Figure 2. Figure 1. As a rule of tumb, te definition of inground is aepted as being te altitude up to one and a alf times te wing span, expressed as eigt above surfae, altoug WIG veiles generally operate at eigts less tan 20% of te span, i.e. b/ < 0.2. However, at eigts greater tan alf te ord lengt (ie / > 0.5) any veile tends to be less effiient. From a regulatory point of view, te maximum ground effet eigt is seen as te servie eiling (te eigt above wi a veile will not sustain fligt under its own power). If te veile an operate full time above tis eigt, it legally beomes an aeroplane and must meet all of te assoiated regulations. Currently oused at te University of Glasgow s pener treet worksop is an unmanned aerial veile (UAV). Originally, tis was a military raft altoug modifiations are under way in order to make meet te speifiations of te Department of Aerospae Engineering. One omplete, te UAV will be a fully operational flying laboratory. However, te first step in tis proess is to turn te UAV into a ground effet veile apable of analysing aerofoil performane. Figure 2 sows te side plan of te UAV wit te distane referring to te wing eigt above ground. Te way in wi te pressure data around an aerofoil will be gained is by te use of a wing glove, te envisaged UAV omplete wit wing gloves is depited in figure 3. Tis is a devie tat an simply be slid over te existing wing and attaed. Tis wing glove an be manufatured to any aerofoil profile tat is desired to be tested. Numerous pressure tappings exist around te irumferene of te profile and tese are ea measured individually and in turn. Tis enables a pressure distribution around te aerofoil to be formed, wi an ten be ompared wit existing data from bot freestream and ground effet tests. Tis will elp provide a guide as to te auray of any aquired data and also te auray of using te UAV in su a way. It sould be noted at tis stage owever, tat veiles tat utilise ground effet ave been speifially designed wit low altitude flying in mind, tis UAV as not. In onverting an airborne veile into a ground effet veile some problems are foreseen and a onsiderable amount of analysis and testing is required before te UAV is used for ground effet testing. Figure 3. Te need to analyse te ground effet during take off and landing is twofold. Firstly, obviously it is relevant to te urrent ground 133.2
CONFIGURATION OF AN UNMANNED GROUND EFFECT VEHICLE effet studying enompassing te UAV projet and seondly, beause it is neessary to understand ow te UAV will andle under tese irumstanes. Tese are simply unknown quantities due to te fat tat beforeand te UAV was roket launed and paraute reovered. Te modifiations arried out tot te UAV mean tat te onventional landing will now beome te norm. It is important from an operational as well as operational analytial point of view. Te ontroller must be informed as to ow te UAV will andle and be prepared, te more information te better. Tis is vitally important as te anes of aidents are inreased ere and te safety meanisms will be virtually ineffetive at su low altitudes. Airraft only experiene ground effet for relatively sort stages of take off and landing. It is ere tat te proximity of te ground in relation to te airraft dimensions is omparably small. Publised in 1921, a simple metod based on te Prandtl lifting line teory was direted towards te performane aspet of ground effet. Of late, referene [1] as disussed te influene of ground effet on sort take off and landing airraft. Tey explored te more important airraft design parameters and added perturbation terms to te equations of motion to investigate te fligt arateristis lose to te ground. Ground effet is of ourse of vital importane to WIG veiles and te stability of su raft as been studied extensively and orrespondingly doumented. In partiular, linearised approaes ave been utilised for tis very purpose. From tis it is deduible tat tat ground effet will most definitely play a role during te take off and landing pases of a onventional airraft s fligt regime, Te figure sows te relationsip between tis manoeuvre and te arateristis of longitudinal stability in ground effet. Analysis of tis situation may ontribute to a more realisti simulation and te risk of an aident during tese two pases redued. 2 Modifiations to te UAV Te UAV purased by te Department of Aerospae Engineering is of military origin. Te department basially bougt te airframe wit a view to modifying for a speifi use, and sine ten te department as been gradually modifying te said airframe and turning into a flying laboratory. everal modifiations ave been made to date, in partiular: Provision of an underarriage Wing Extensions Wing Glove (optional fitting) Provision of rudder Fligt Control ystem Data Aquisition ystem ine te original airframe was intended for military usage, te above modifiations were neessary. Te implementation of te underarriage was vital as te original UAV was launed by a ballisti devie and landed via deployment of a paraute. Bot metods required igly skilled and experiened operatives to undertake speifi tasks, resoures simply not available to te department. Te underarriage made onventional take off and landing possible and so te need for an expensive launer and suitably skilled staff was dispensed wit. Te underarriage was designed in te department and fitted around te present strutural features of te existing airframe. Te wing extensions serve several advantageous purposes. Namely, inreasing te range, dereasing te take off distane, landing distane and stalling speeds. Additionally, as will be explained furter, tey also elp to prevent te wing glove being adversely affeted by wing tip vortex effets, wi ould be detrimental to te data reeived from te glove. Wit te onventional take off and landing, additional yaw ontrol was required. Te initial design did not ave tese wit te entire yaw being produed by flaperons. To provide tis 133.3
M. Millar, L. mrek additional yaw, rudders were implemented into te tail fin design. In addition to tese basi modifiations to te airframe it was neessary to equip te UAV for low level flying. Essentially tis only requires a eigt sensor and adequate autopilot to ensure proper andling at su low eigts. At tese eigts it is simply not aeptable to be able to expet a uman ontroller to maintain a onstant low altitude wilst maintaining safe fligt, it is too mu of a burden. Certain safety measures will also ave to be introdued. Tis will most likely take te form of one or two braking parautes to be ejeted out of te tailbooms. 3 Experimental Teniques for Testing Ground Effet in Wind Tunnels Aurate representation of te flow fields for ground effet testing in a wind tunnel is diffiult due to te presene of boundary layers on te tunnel surfaes. Generally, no boundary layer is present on te ground for motion at very low altitudes, tus any tests attempted would ultimately affet te ram wing effet. In referene [2] it is te disovered tat tis boundary layer as a fairly drasti effet on te on te measured lift oeffiient. For a wing wit aspet ratio 6, te lift oeffiient was altered by 33% at a eigt of 20% of te span, wi equates to 120% of te mean ord. Considering tat te stati and dynami stability is fundamentally based on su derivatives tese effets are of vital onsideration, along wit te inidene and eigt. It furter attempted to investigate te areas wit wi te boundary layer affeted te lift oeffiient by using a flat plate wit various aspet ratios and a belt. Te preise eigt to span ratio were disrepanies appeared sowed linear orrelation wit te lift oeffiient aording to, b C L < 0.5, AR = 6 (1) ullivan [3], also analysed tis penomena and arrived at a similar onlusion, b C L 1 < AR π (2) Power augmented ram (PAR) is a penomenon tat many WIG veiles ave utilised reently. Tis is wen te engines are plaed forward of te lift produing wing and te exaust is direted under te wing to provide additional lift. Tis is benefiial during te takeoff and landing pases as te speeds an be redued tus inreasing te safety aspet. Tougt te influene of PAR on te boundary layer during experiments is unlear. A metod proposed by Turner involves removing te boundary layer wit sution and aving a moving belt running at rougly te same speed as te freestream veloity. A relative insensitivity to te belt speed means tat preise speed ontrol is not required. A semati of su a metod an be seen in te figure. Te leading edge of te ground plate as a smooted leading edge spanning te widt of te wind tunnel and a 45 o slot ut aross. Tis was osen as te optimum slot angle as a ig angle would ave a sort lengt over wi te sution would work and a low angle would ave a sallow rotation of te boundary layer. A ontration and expansion setion is fitted under te plate wit te position of maximum ontration aligned wit te slot in te plate. Tis results in a low pressure tat suks of te boundary layer on te upper surfae of te plate. ubsequent testing of te metod suggested in referene [4] revealed tat te boundary layers were redued by approximately 50% Teoretial analysis of te situation by owdon and Hori (1996) yields te relationsip, U U θ θ H = + H C (3) U U 133.4
CONFIGURATION OF AN UNMANNED GROUND EFFECT VEHICLE Te advantages of tis type of metod mean tat several parameters an be altered in order to obtain te desired onditions. Tese involve te plate leading edge sape, te level of sution, slot widt and slot sape. Unfortunately, altoug in lose proximity to te wind tunnel floor te metod for verifying te wing glove involved used not wall boundary layer ontrol and so tis may well ave ad an effet on te results. 4 Wing Glove Te wing glove is te fundamental key to te UAV being used as flying laboratory as it is tis devie wi ollet all te aerofoil data. Te wing glove is merely a wing mok up wi is plaed over te UAV. Two gloves are neessary to even out any uneven, detrimental aerodynami effets indued by te gloves. Full details of te wing glove design and testing an be seen in referenes [5], [6] and [7]. 4.1 Aerodynami Considerations In implementing a wing glove, two important aerodynami penomena must be onsidered tat of te downwas effet and te rossflow effet. Tese are typially tree-dimensional ourrenes assoiated wit finite wings and are not present in infinite wing analysis. On te UAV, two major soures of vortiity are likely to our one te wing glove is installed and operational. Vortiity will almost definitely arise from te wing tips and also at te glove edges. Wing tip vorties are unavoidable wit a finite wing and are responsible for downwas, wi ultimately redues te effetive angle of attak of te wing. Tey also introdue spanwise flows, or rossflows, wi affet boundary layer development. Fortunately owever, te UAV is intended as te 2D analysis tool and tese penomena sould be redued dramatially or even eradiated so tat tere effet on any results an be virtually ignored wit te introdution of endplates on te wing glove. 4.2 Experimental Glove Positioning & Geometry Te wing glove is entrally loated on te wing, tis is in order to keep te unwanted effets of wing tips and wing/fuselage interations from effeting adversely te flow over te wing glove. Due to te wings being retangular wit no sweep, tis is also te sape tat was adopted for te wing glove. Te attament of te wing glove means tat te inner ailerons are disabled, sine tey are overed, and te ailerons on te wing extensions used to indue rolling moments. Te fat tat tese wing extensions are easily removable te wing glove an easily be slid over te original wing attaed and ten te wing extensions reattaed. Te minimum ord of te glove is dependent on a ombination of fators. Obviously, te ord of te wing plays an important roll and also its profile sape along wit te intended profile of te aerofoil wising to be tested. Finally, te tikness of te material and te need for spaing for te pressure sensing equipment ditates te glove size. Te prototype glove was a NACA 0012, as tere is plenty of data in existene on tis aerofoil setion providing a more tan adequate omparison. Te design, one finalised, possessed a wing ord of 1.05 m. A series of pressure tappings, sequentially spaed around te glove, are present in te middle of te love; 30 on te upper surfae and 30 on te lower. Tese lead to 2 pressure sanners, via tubing, situated in te RPV fuselage. It is important to realise at tis stage tat te wing glove does not ontain any instrumentation, tis allows a ertain amount of flexibility wen it omes to testing a number of profiles were te tubing an be disonneted and new tubing attaed wit relatively little fuss. Two wing gloves are likely to be used, altoug only one will atually be used for measuring, te oter will simply be a dummy in order to balane up te airraft and prevent any unwanted rolling or yawing moment wi may be aused by aving only te one wing glove. 133.5
M. Millar, L. mrek 4.3 Numerial Modelling & Wind Tunnel Tests Te purpose of numeri modelling was to establis te quantitative effets of endplates on te wing glove. Te proposed designs traed te aerofoil sape and ranged from 5 m to 50 m in eigt (see Fig. 4). Figure. 4 Te analysis ose to primarily look at te downwas effet, resulting from te wing tips and also te rossflow effet and te result of altering te size of end plates. Te onlusion to tis was tat te downwas ould not be redued, ene tis will ave to be onsidered wen results are olleted and an effetive angle of attak obtained. It also found tat te larger te end plates te less te rossflow, as may well be expeted. Bearing in mind tat a balane between size and performane must be found it was disovered tat 5 m end plates would do a suffiient job in reduing ross flow. Te speial endplates also proved to be fairly effiient in reduing rossflows aross te wing glove. Te wind tunnel tests were intended as an extension to te numeri modelling, to establis an effetive size and design of end plate, in addition to testing te pressure sensing equipment. Te wing, omplete wit wing glove, was plaed in te wind tunnel wit te elp of a speially onstruted rig. Various onfigurations and angles of attak looked at, su as wit and witout wing extension, to establis downwas effet and wit and witout endplates, to establis te rossflow effet. A ertain amount of agreement wit te numerial analysis was disovered at ertain angles of attak and also tat te downwas ould be altered. Te tests also demonstrated te need for endplates in order to redue rossflow, altoug te wind tunnel tests suggested tat te best design to employ was te speial one. Te instrumentation involved in te testing performed better tan expeted and produed some good results, demonstrating tat te system works well and will do under proper fligt onditions. 5 Wing-in-Ground Effet tability Criterion Here, referene [7] extends upon te work of referene [8] stating tat te riterion for stati stability is F > 0, were F an be fatorised into tree parts: m F α vˆ, ( T D/ W) vˆ m,n= 1 n (4) ( /) vˆ, m Te first term is te stati piting moment stability, known as stati stability. Te seond term gives te minimum drag speed at te stability boundary. Te tird term is a new term ontaining just te influene of te ground and is alled te stati eigt stability. n vˆ = onst. < 0 / m 0 = (5) If te equilibrium in eigt is disturbed, te above equation gives a restoring fore. Te stati eigt stability an be expressed as a funtion of te aerodynami derivatives: n / Generally, L vˆ,m = vˆ 2 0. L 1 m mε. Lα < 0 (6). L is negative; tus, te ondition for stati eigt stability an be expressed in te form: 133.6
CONFIGURATION OF AN UNMANNED GROUND EFFECT VEHICLE F m m < ( m )( L ) α. Lα 1 (7) Wen a wing approaes te ground, te lift oeffiient inreases, ( L ) > 0, and also a nose down piting moment ours, m > 0. Terefore, in order to possess a ertain margin of stati eigt stability, a ig value of must be ounterated by ig values of m ( L ), and te stati piting stability ( mα ). Te first term is primarily influened by te aerofoil arateristis of te wing. Te seond term an be inreased by, witout inreasing m as well, by a ig orizontal tail working out of ground effet. Te effet of te entre of gravity on stati eigt stability is minimal. Te dynami longitudinal stability of a veile in ground effet is determined by te roots of te fift order arateristi equation. As 5 + Bs 4 + Cs 3 + Ds 2 + Es + F = 0 Out of ground effet te F term disappears and tis beomes a quadrati equation wit te sort period mode and te pugoid mode as solutions. Flying in lose proximity to te ground, an aperiodi mode is obtained in addition to a pair of omplex roots. In ground effet, te aerodynami oeffiients depend strongly on eigt and angle of attak, a fat wi leads to non-linear motions espeially in limbs and desents, as well as in response to ontrol inputs. Te nonlinear beaviour is partiularly observed wen te fligt pat rosses te transition eigt regime, 0.5 < / < 2. Witin tis region a stable trimmed fligt is not possible. At tese eigts te veile will limit-yle osillate bot in eigt and pit motion. Anoter very important non-linear arateristi of ground effet veiles is an automati flare wit fixed ontrol. I tis ase te inrease of lift oeffiient gives a spring type fore during approa to te ground, wereas te vertial kineti energy is dissipated in several yles of eigt and piting motions. In priniple, elevator and trust inputs ould be used for eigt and speed ontrol. Appliation of te elevator mainly indues disturbanes in bot angle of attak and eigt. Te danger of touing te ground or exiting eavy osillations is a disadvantage of applying te elevator as a primary longitudinal ontrol. In omparison to te elevator, te trust ontrol is a very favourable means of eigt ontrol. Good steady state anges in eigt an be obtained wit small transients in angle of attak and eigt. 6 Conluding Remarks At some time in te future it may be te ase tat te UAV is no longer used as a ground effet veile and tat tis projet is merely a stepping stone to a fully operational free fligt laboratory. In tis senario a rig would be built to be plaed on te front of a ar or van and utilise te wing and wing glove of te UAV for testing. Tis ould be used to aquire simple ground effet data using te wing glove or even to aquire take off and landing data by mimiking te motion of te wing under tese irumstanes and building up a take off and landing profile for te given wing Tis paper as addressed te problems involved in produing a reliable UAV to operate in ground effet. It as assessed te potential pitfalls and offers possible solutions to ommon issues. Tere is a risk assoiated wit tis, in using an UAV wi is not designed for te intended purpose, toug it is seen as valuable stepping stone in evaluating te system as a wole in preparation for iger altitude free fligt. Te low eigts will ensure a relative amount of safety and opefully minimal damage sould any emergeny our. Referenes [1] Gratzer, L. B. & Maal, A.., Ground Effets of TOL Operation, Journal of Airraft, Vol. 9, 1972, p236-242 [2] Turner, T. R., Endless-Belt Tenique for Ground imulation, NAA P-116, 1966 [3] owdon, A. & Hori, T., An Experimental Tenique for Aurate imulation of te Flow Field for Wing- 133.7
M. Millar, L. mrek in-urfae-effet Craft, Aeronautial Journal, June/July 1996, p215-222 [4] ullivan, M. C., Flow Breakdown for Wings in Ground Effet, Journal of Airraft, 1978, p859. [5] Darida, M., mrek, L. & Coton, F., Development of an RPV Laboratory for In-Fligt Aerofoil Testing [6] Darida, M., mrek, L., Coldbek, D. P., Feasibility tudy on an RPV for te Utilisation as a Flying Laboratory [7] Darida, M., mrek, L., Wind Tunnel Investigations on a RPV Wing Glove Configuration, ICA 98 [8] Kumar, P.E., ome tability Problems of Ground Effet Wing Veiles in Forward Motion, Aeronautial Quarterly, February 1972, p41-52 [9] taufenbiel, R. W., ome Nonlinear Effets in tability and Control of Wing-in-Ground Effet Veiles, Journal of Airraft, Vol. 15, No. 8, August 1978. 133.8