MINISTRY OF TECHNOLOGY - PDF document

C.P. No. 1091 . MINISTRY OF TECHNOLOGY AERONAUTICAL RESEARCH COUNClL CURRENT PAPERS Atmospheric Gusts - A Review of the Results of some Recent R.A. E. Research bY /. Burnhom Aerodynamrcs Dept., R.A.E , Bedford LONDON. HER MAJESTY’S STATIONERY OFFICE I970 PRICE 13s k/ [65p] NET 1 l U.D.C. 551.551 : 551.54 C.P. No. October 1968 ATMOSPHERIC GUSTS -AP.EVIEW OF THE REWS OF SCMEIfECENTR.A.E.RESEARCH J.Bumhsm Aerodynamics Dept., R.A.E., Bedford ,* l I been investigated. l Replaces B.A.E. Technical Report 68246 - A.R.C. 31113. CONrENTS 1 INTRODUCTION 2 SOURCES OF SEVERE GUSTS 3 W5TS IN THE STRATOSPHERF, 4 GUSTS AT ALTITUDESUSEDBY PRESENTTRANSPORT lLIRCF!A.FT 5 GUSTS IX RELATION TO TAKE-OFF AND LANDING 6 THEDESCRIITION OF GUSl'S 6.1 Gust spectra 6.2 Gust probabilities 6.3 Theoretical models of gusts 7 THE EFFECTS OF PILOT CONTROL ACTIONS ONFLIGRT. THRCUGH GUSTS 8 CONCLUDINGRFMARKS *PP- Accidents to civil aircraft involving turbulence in which major structural failures occurred in flight References Illustrations Detachable abstract cards m 3 4 11 13 18 19 Figures 1-33 3 . 1 INTRODUCTION This Report gives a brief account of what R.A.E. has learned from research on atmospheric gusts smce Zbrozek's review' of 1964. The period frcm that tme to the present has seen* the consolldatmn of several trends which were then apparent. 2 SOURCES OF SFVERE GUSTS A survey of catastrophic accidents to civil transport aircraft in which encounters with gusts played a sl&lficant part 1950. Of these 20, 17 are clearly ltied with thunderstorms, as probably are 2 others. near Mount FuJi 6 The remaining one is the accident to a Boeing 707 , where structural failure occurred due to 6.8 millwn nm of the progrsnme, 35% of the cases were noted by the crew as being associated wdh storms and further 1% were 36% 'unknown'. Many of the latter, however, occurred in areas of the world at times of the year when thunderstorms are c~rmon. Of the encounters m relatively storm free parts of the world, about half of those over Italy, the U.K., and the eastern seaboard of the U.S.A., were due to thunderstorms. Taking the encounters as a whole, 54% occurred &n-q climb or descent and. 46% during cruJ.se. The dumtion of the turbulence patches in which the encounters took place tended to be short IO,11 , have recently been reviewed 12 . Gusts large enough for their avoidance to be desrrable by clvll aircraft have been famd, m the R.A.E. work, in clear arr near the tops of thunder- 10000 ft vertically frm the of the cloud), and in associatmn with mountainwaves. One patch of severe turbu- lence of the latter type, found at 46000 ft, contalned a horizontal gust whxh reduced the indicated airspeed of the aircraft by 50 knots in 1 5 the storm tops I3914 (Fig.6). There is some evidence that slgnlficant gusts cast, at around storm top level ,5 , at rather greater &stances fr~nthe storms than is the case at altitudes between 25ooO ft end ft nlrmber.¶ to be assigned to these relationships. However, it is clear that many thunderstorms, 21 and 22 show similar stable layers. Little is bown, fman a climstological point of view, about the existence of these stable layers but it appears that they are uncamnon. Their identification with severe wave disturbances in the stratosphere is by no means certam and complete, nor 6 At the present tune, the more enllghtened au-lines avoid flight through areas ~tl which significant mountain wave activity is forecast in the troposphere (e.g. Ref.23 and 24). In all tiown cases of encounters in the stratosphere with 0 This connnent, and simzlsr one about thunderstorms, assume that SSTs will not be much different to current aircraft ~fl the way 111 which they respond to atmospheric disturbances. This assumption 1s reasonable for the types of SST which we have so far considered. 7 /c The R.A.E.-N.S.S.L. work has allowed the relatlonship bekveen gust intensity and properties of the radar echoes of the storms to be firmly establxhed16 over an altitude range frm 23000 ft to 35000 ft. A clear relationship exmts between the probability that the largest 25 , m which derived and true gust velocltyO statxstux were compared and found to be in good agree- ment (Fig.10) suggest 1% of occasions. Questions of the applxabillty of these results to storms in other parts of the world snd of their interpretation in terns of aIrborne weather radars are consIdered at length m Ref.16. Suffice * True gust velocities are actual cunpanents of airmtmc. Derived gust velocities are obtamed directly frm measurements of mcremental noma: acceleration, berg the size of gust of' a particular shape whxh wmld have produced that acceleration (see Appendix to Ref.16). 8 Reference has already been made to the meteorologist's present ability to forecast regions of the troposphere likely to be effectedby mountain and lee waves and associated distwbances. The latter can, of course, be severe and have been responsible for at least one catastrophic 27 , m which an rnstrumented air- craft was flown innnediately in the lee of a high mountarn ridge in strong wind conditions have been analysed and will be referred to in 6. Although convective clouds and waves are the situations in which, opera- tionally, dangerous gusts are likely to be met, and R.A.E. research has there- fore concentrated on them, the much more common but less intense CAT is a significant nuisance to airlines and their passengers. The current situation regarding the forecasting of 28 : "The present position is, therefore, that we know a great deal about the possible locations of CAT, sufficient to be able to predict general areas where the chance of encountering turbulence is greater thsn elsewhere. This is possible mainly by identifying the position of the . I 5 GUSTS IN RELATION TO TAKELQFF AND LANDING In recent years there has been considerable interest in the U.K. in gust problems in relation to take-off' and landing 31 , particularly in connec- tion with the manual lading af large or their forecasting be facdltatedby some other means. The importance of windshear effects can cope with the average level of gustiness likely in a 30-l@ lmot wind is not likely to be greatly troubled by windshear, except perhaps for the human pilot who is taken by surprise. However, signi.ficant shears may be associated with large temperature stratiPication and important 15 n0vmb.2 1966 0-a. The aircraft touched down with a large sideways velocity and the runway not beenwet, allowing the aircraft 14, are avadable from the Meteorological Office but their ttiscale is such that the rapidity of the fluctuations cannot be resolved to anything like the accuracy needed to determine the effects theywculd have on aircraft. In an attempt to obtain further tiomatian, continuous records of win&peed at heights of 33 100 ft are being obtained cm an expanded timescale using an instrumented tmer at R.A.E. Bedford. no large squalls have yet occurred but a record of a relatively small one is given as Fig.15. Although the recording speed used 1s some 12 times that the standard Meteorological . 11 Office lnstrwnent, It 1s still inadequate to resolve the more rapid fluctua- tions. A further recording system, giving an addltiond spsedmg-up by a factor of 30, has therefore been introduced in the last few An interesting record 1 - cosine functmn . 'Spectral methods' were mtroduced to gust studies in the IVLQ's, having been used m the description of turbulence, from a fluid mechanics pout of view, for the previous 20 years. The gusts are here conceived as exsmples from a ran&an process with a d&e rmrnate spctral density(average variation of energy mth frequency or wavelength). Know- ledge of 12 turbulence was not hcmogeneous - for example, if energy was being fed into it m a region where measurements were being made, but due to the mean flav the decay partly took place scenewhere else - a slope steeper than 1750 ft to 7800 ft have been obtained from examples of the vertical 12 cqxexnt of turbulence measured in clear air near storm tops . Corresponding values lnslde the thunderstorms range from 750 ft to Z3CO ft, these for the qectra shown in Fig.17. Saane rather unexpected . * The autocorrelation and spectral density are Fourier transforms of each other. I3 The spectral densities described above show how, on the average, the energy of the turbulence varies mth wavelength. Spectral methods also provide a valuable tool for c-ring measured aircraft loads and motxms wdh theoretical predxtlons and provide agood insight into OCCU?.-. The spctral density, alone, will not tell him this. 6.2 Gust probabilities If . is 14 Gaussian. For a given patch of moderate or severe turbulence, the probability that transitions in excess a? given will scour tends to be exponential for values more than about twice the nns in all the examples so far examined 51,52 , and sanetimes ~ig.25. In canparing the behaviour of different sizes of arip craft, it is sometisms convenient 40 to consider the behaviour of lines of equal value of the product of the number of gusts exceeding a given size and the gust length and such curves are . 6.3 Theoretical models of' msts The above doubts about the valxllty of Gaussian process representatlom of turbulence had led to a search for theoretical models which, while utdlsrng spectral ideas UI conslderlng the average distribution af energy with atmos- phere. A number of concepts are being studied., rncluding those of the transi- tion function mentioned earlier snd of self-sxnilar intermittent random pro- cesses and several prrcrmising ideas have emerged 53,54* This work has been of also been constiered. Mention was mde earlier of a series of gust accxdents and incidents hewn as the Jet upsets' and stilar accdents 55,56 appear to have happened to propeller &-lven aircraft. A feature of the jet upsets was to be closely related to elevator deflection. Fig.31 shows the variation of the ratio of nns derived gust velocity to zms true gust velocity with the latter quantity for a series of flights with a fighter aircraft. through oon- vective cloud. a CONCLUDING REXARKS The past few years has seen an increasw concern, on the part of those involved in aircraft operations, about the effects of atmospheric gusts and the means ofavoiding particularly the more severe ones. The great concern expressed in scme quarters . understanding of the physicalmechanlsms responsible for severe gusts. There are, however, 18 Apoendix ACCIDBNlTSTOCIVILAI~ INVC&VINGTLlRBLU.ENCE INWHICH MAJOR STRUCTURALF-S OC'XJRFXC INFLIGHT Date and place Jun '9 1 Cb.@ E6A LearstAr 18 x3 viscount B26C C@ Viscount DC6 dovm frm hills". Cold front wth thunderstorms. Cold front with scattered thunderstorms. Thunderstonm. Squall lme BAC ill Extensive linked thunderstoms. _! d/Y . ._ 1” No -- 1 4 5 6 ‘. 7 8 9 ,. IO Author(s) J.K. Zbrozek WA. Crooks REFERENCES Title, etc Atmospheric (1965) Gust research - present status and future plans. Unpublished R.A.E. paper (1968) 65192 (A.R.C. 27450) (1965) Aeronautical aspects of atmospheric turbulence. Unpublished Meteorological Office paper Some aspects 707 GAPE at the foot aC Mount 5 March 1966. CAP 286 H.M.S.O. London The civil aircraft airworthiness data recarding programme. R.A.E. Technical Report 641104 (A.R.C. 26490) Rapports presente a la sous-cmmission de meteorologic de la camission pe-ente du transport supersanique. Direction de la Meteorologic Nationde Paris (1965) Civil amcraft airworthiness 974 High altitude clear air turbulence. AFFDL Technical Report 65-l& (1965) . 20 No. II 12 13 W.T. Roach 15 18 19 20 Author(s) w.m. crooks FA Hoblet -c.Es (conta) Title, etc An investigation of 30 (1966) A study of turbulence xntensity 68141 (A.R.C. 31013) (1968) Weather radar - a handbook for pilots. H.M.S.O. London (1968) Thunderstorms. FAA Advisory Circular CO-24 (1968) High altitude turbulence andsupersonic aircraft. Paper given in Section IO of the 1964 Annual Meeting of the British Association for the 1967. R.A.E. Technical Report (in preparation) . 21 . &. 21 23 24 25 . 26 27 Author(sj T.L. 30 , 31 J. Burnhsm PEFXRENCES (conta) Title, 75000 ft for several areas of the Northern Hemisphere. NASA Technical Note D Ozone measurements for diagnostic studies of atmos- pheric circulation. AIAA pBper 65-462 (1965) The mountain wave. NAsA ~3 315 (1965) Mountain wave 37 (1967) High intensity gust investigation. Boeing-Co. Dot D-13273-333.4-2 (1964) Clear air turbulence. (1967) Report an clear air turbulence meeting 26-27 April, 1965. U.S.A.F. un-numbered document Proceedings of National air meeting on clear air turbulence 1966. Inst. Nav. and SAE note on turbulence problm associated with take- off and lsnding. R.A.E. Technical Report 672&O (A.R.C. 29520) (1967) 22 a. Author( 4 32 W.J.G. Pinker 33 J- Hall 34 J. 35 J. Bumban M.J. Calmer 37 ' H.E. Whittinghsm 39 &A. Giblett 40 J.Burnham 41 F. Pssquill 42. G. Batohellaf 43 C.G.B. Mitchell RFzmmNms (conta) Title, etc SW observations on the dynmics of large slender aircraft. AGARD 17 Part 1 (1966) The effeotaflow altitudegusts onautanatic lanains. A&SD C.P. 17 Part 2 (1967) c&&s aa&apprmchat Gibraltar airfield. 27345 (1965) ICAO Report of the Air Navigation Conference Onlargeandrapidwin3fluctuations~ocour when the wind had previously been relatively light. RILE. Technical Report in d’hfet, Bulletinti The th7mderstonn. U.S. Dept. of Cmrneke (1949) The structure of w5nd mer level country. 54 (1932) An analysis of some measurements of severe atmospheric gusts. � R.A.E. Technical-Report in preparation Atmospheric diffusion. vsnNorstrad Homogeneous 66275(A.R-c- 28633) (1966) , 23 REXWENCES (Contd) Title. e tc Calculatum of the response of a transport aircraft to contmuous turbulence and discrete gusts and 68083 (1968) Assessment of the accuracy of gust response calcmh- tions by cmpanson with experments. AIAA Paper 68-892 (1968) (1963) Estunates of probability distribution of root-mem- square gust velocity of atmospheric turbulence frcm 3362 (1956) Aircraft response to turbulent air. Paper 10 m atmospheric turbulence and its relatlcm to aircraft. H.M.S.O. London The mathematical analysis of randam noise. Bell System Tech. Journal Vols. 23 ana 24, reprmted in Noise and Stochastic Processes, Dover, N.Y., (1954) An expermental check (1965) a. Author(s) 4J+ C.G.B. Mitchell 45 C.G.B. Mitchell 16 N.I. Bullen 47 H. Press LT. Meadows I. Hadlook LB J. Burnhem 49 S.O. Rice 50 J. Bunhem 51 M.J. Calmer An snalys~s of wind. fluctuations before the passage of a cold front m Oklahoma. R.A.E. Techmcal Report m preparation &. Author(sj 52 J.G. Jones I. Pace -s (Corm) Title. etc Transition probability d~tributiom and spectra 12 Idcur, 1959. CAB Aircraft Accident . 57 J.A. Stran -66~ high altitude gust survey technical analysis. T.G. Weatheman ASD-TDR-63-145 (1963) . i . Places where moderate Or severe ’ turbuknce was encountered Fig.1 Positions where moderate or severe turbulence was _ encountered during worldwtde CIVII aircraft operations and average numbers of days per year on which thunderstorms occur Percentage of patches of turbulence contominq occeieratton increments exceeding O-69, with durations exceedmg given values 0 In ;; 0 . 20 + In mountain wove5 u Near thunderstorms + .L 0 + 0 5 t 00 + 0 .5 + 0 + 0 0 I O-05 0.1 o-5 I-0 Distance which change took place, nautical miles Fig.3 Examples of large and rapid changes in air temperature encountered In the stratosphere -3o- ,-Position of severe horizontal qust Positrons OF I---severe -0 turbulence I * . IO 20 o&once : nautical miles Fig, 4 Time history of air temperature encountered during flight through a mountain wave at an altitude of 46 ft \ IO I 0 Oiston;;. +$J-gusonds 0 4 8 IO Distance; thousands of Flg.5 Examples of gusts measured In clear air near thtinderstorm tops Air temperature -65 -e---- A\ 0 IO I Approx horizontol scale: nauttcal miles Flq.6 Tentative model of a quasi- steady storm top (from ref 14) l . i L I IO 0 IO 20 SO D~stanco nautical m~los Fig.7 Streamlines of flow in mountain waves over the western U S A, based on constant potential temperature surfaces Stable layers -60 -IO 0 ’ AII- temperature deg C Fig. 8 Examples of upper air temperatures showing the exqtance df stable layers In the stratosphere IO - za max is maximum radar reflectivity OF storm mm6 /m3 Fig. 9 Effect of the maximum radar reflectivity of the storm on the maximum derived gust velocity encountered on thunderstorm penetrations 0 papaam JJD sanp u+16 q+p uo sunJ jo a6Dpamd I 1 0 20 Missmg storm ‘core by more than 5 miles Maximum dewed gust velocity : lsec EAS . Fig.11 Effect of passing through or missing the storm core, on maxlmum gust velocities encountered on thunderstorm penetrations (at he:lqhts between 23000ft and 28000ft through storms with maximum reflectivitles between IO4 and I 05*5 ) I 16 I1 18 ‘20 Time hours Tennont Creek (Australlo) I2 Dee 61 Fig.12 Anemogrom of a large squall Wind drerkion wtnd speed 1600 BedFord 15 NOV 66 1800 Bedford 31 March 62 Flq.13 Anemoqrams of squalls recorded at Bedford . Wind direction 0900 0800 . Leemmg Church Fenton 2 July 68 July 68 Fig 14 Anemogroms of squalls showing morked wind changes . “I 20 ft 2 z IO Y m -0 ti 20 50 ft t? IO -0 : 20 33ft IO i_, Fig.15 Time hlstory of wind speed during a squall at Bedford soft --I. de.- 1 . -*.. -- 'a.. ---- . ...*.. / s ooo ’ IO 20 1326 GMT Time: seconds Bedford 20 Aug 68 Rg 16 Time history of o rapid change of wmd speed . Actual spectra 7 I I Log frequency Flg.18 Shape of ‘theoretical’ gust spectra compared with possible shape of actual spectra at long wavelengths 1 . . i . Autocorrelotlon 0 IO Time: set Woveknqth : Fig.19 Experimental gust autocorrelation function, fitted theoretical curve and corresponding spectral densities i 0 Zero crossrng frequency; N, ; per nautical mile normal Fig 20 Per centage of runs on which zero zrossmgs of true gust velocity and incremental normal acceleration exceed given values, obtained on fhghts through convective clouds IOO- Convective clouds (ref 25) II 0 I 2 Ftg 21 Percentage of runs through convective clouds and thunderstorms on which the maxlmum normal acceleration exceeds n times the rms Gust length, H ft 70 . 0 \, WH- '4 Fig.24 Number of gusts of length H which are bigger than W, for data from rvf 27 . 0 gusts exceeding W . IO 100 +20 x 50 q - /&o-o Lo */ 9 ’ +/+ +\ ;s;,-to-0 0 *-. . + / X’ x/x\x/x, 2 0’ O-O\ X o-o \’ +-+ X 0 / X Thunderstorms I 200 0 Ref 27 1 I I 400 Gust length, H ft I 0 W 20 40 Fig.25 Variation with gust length of number of gusts of a given size encountered Oatnfrom fcf 27 I20 N is number of gusts OF length H which exceedw 100 W 80 Thunderstorms 60 4 ’ \ p \ 1’ \ &’ \ $‘/ oo,‘-\ O 4’ y’o,,\ / q+ / ,/‘,/ 1, O/ ,‘,+ ‘, ! .*‘,’ ? +’ C$@~‘\, /xQQ ‘4 0 ,o% “, / 3’ 00 ‘\, ‘x 0 \ 0’ \ \o Gust lenght, 14 ft. Fig.2 6 Variation of gust size with gust length for constant values of the product of gust length and the number of gusts of length exceeding W in size Number of ysts ’ x 1000 I 7.:: . 200 ‘“fl$ii:, ) i 1000 Gust Ieyht l-j fd’“: , ,* ” 3’ ’ L . Fig.27 Variation with gust length of number of gusts of a given size for self-similar gusts with a spectral density which varies as the square of wavelength and exponential probability distribution 120 r 100, 20 N = Number of gusts of lenght H which exceed W / \0= -._- / WC 000 Gust length, H. ft Fig. 28 Voriation of gust size with gust length for constant .a values of the product of gust length and the number of gusts of length which QxCQQd W in size, for self - similar gusts with a spectral density which varies as the square of wavelength and exponential probability distribution . Vertical gust velocity Normal acceleration Elevator deflection Pitch attitude s I 0 25 50 Tune: seconds Fig.29 Time -histories of true vertical gust velocity, normal acceleration elevator deflection and pitch attitude, measured during flight through a convective cloud , .o s? G a 2 c Symbols denote A di+%ent pilots +A + 0 I _ OA $t A + 00 ;AAo o A 0 0 1,'$ 12 rms truu vPrl$al’ gust velocity' ft/sec 0 Fig31 Ratio of rms derived gust velocity to rms true gust velocity vs rms true gust velocity, for flights by different pilots t hrouq h convective clouds . c Square OF modulus of apparent aircraft Frequency response (from ratio of measured normal acceleration ond gust spectraldensit@ 0 . b ui 0 . A 3 Symbols denote diFFerent pi lots A 0 0, I A o+ a + A 4 ( 0 15 Maximum true vertical gust velocity : ftlsec - Fig.33 Ratio of maximum derived gust velocity to maximum true n gust velocity vs maxlmum true gust veloclty for flights by different pilots through convective clouds ~nnted tn E.qland fo- Aer Majesty;s Stationery Offwe oy the Royal Akcruft Estaollshwnt, Fambosou#h. Dd.lY8915. i(.Y. DETACHABLE ABSTRACTBARD . ..R.C. C.P. NO. 1Lyl October !968 Km. J. ATPlCsPHERlC GUSTS - A REVIEW OF TIE REsbzTs OF Published by To be purchased from 49 High Holborn. London w c.1 13a Castle Street, Edinburgh EH 2 3AR 109 St. Mary Street, CarddTcxl 11s’ Brazennose Street, Manchester 2 50 Fa,rfax Street, Bristol BSI 3DE 258 Broad Street, B~mungham 1 Lincnhall Street, Belfast BTZ 8.0’ or throu& any bookseller C.P. No. 1091 , 1\ A-- d.P. No. 1091 SBN 11