U.S. DEPARTMENT OF COMMERCENational Iechn-cal Information Service

AD-A030 882

Evaluation of Param etersAffecting HorizontalStability of 'Landing MatsIrm EnierWtrwy xeien tto icsugMs

t Sep 76

294061

TECHNICAL REPORT S-76-10

O EVALUATION OF PARAMETERS

AFFECTING HORIZONTALSTABILITY OF LANDING MATS

by

Yu T. Chou, Walter R. Barer, William P. Dawkdns

Soils and Pavements LaboratoryU. S. Army Engineer Watmwos Exprmnnt Station

P. 0. Box 631, Vidcsburg, Miss. 39180

September 1976

Faal Report

Aepwd .~ ~ac 9166om; 6Obafto unow"

. . - . .

Pm O~ce, Ch• of Enginers, U.* S. D D,,sb,.qti, D. C. 20314 D

Undtr Project ,4A762719AT31 {I~o 19-Task 02, Work Unit 001 r, 0TI 196

NATIONAL TECHNICAL Q

M oi l

INFORMATION SERVICE F),sPRiN ILLaD. VA. 220'1I

U~nclass!ifiedSECUR.TY CLASSIFICATION~ Of THIS PAGE fU?..e Data Foised!

REPORT DOCUMENTATION PAGE READ INTRUCTONSI. REPORT NUMBER 2.GOVT ACCESSION NO. 3. RECIPIENT'S CATALOG NUMBER

Technical Report S-76-10

14. TI T LE (wd Subtitle S. TYV'E OF REPORT A PERIOD COVERED

INAIIJA!IOX OF l'AR:NMFIlERS AFFECr-IN, Fia rpr

IIORIZO.N IAl. SI.BII I.iY OF L..MNING NIAIS 6 PERFORMSING ORG. REPORT NUMBER

7. AUTNQR4.j ll. CON.TR ACT OR GRANT NUMBENSS)

YuI 1. Chou. W~alter R. Barker. William 1P. 1)a%%i~n%

9. PERFOnMING ORGANIZATION NAME AND ADDRESS M) PROGRAM ELEMEN T. PROJECT. TASK1'. S. Arn% tEnginsecr Watcrua%% Expe~riment Stat ion AREA 4 WORK UNIT NUMBERS

Stvil and P'a~cnwntN. I a3uratior% 'rtijc~t 4A762719A D31.1P. 0. K)\ 631. VicL%hurg. \Ix%%%%~irpi 39I1() r1ai. 02. Wurk U'nit (XI1

11. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT OATW

Office. Chief tit Finginecr.. 1'. S. Arm eptember 197ks\%a'hingti'n. :1 C. 20314 11 NUMBER OF PAGES

14. MONITORING AGENCY NAMNE k A.DORESS~tf differnmt tres Con.trolin Office) IS. CUNITY CLASS. (of this ,tp.fl)

RS.. OECL AS3FICATION, DOWNGRADINGSCHEDOULE

IS. DISTRIBUTION .TATEISENT (of able RpA )

IT. DISTRIBUTION STATEMENT (at th. eabtract 1.entrdi Block 20. #1 Affoetme hef Report)

1S, SUPPLEMENTARY NOTYES

It. KEY IWORDS (Continu, o M~t..*e side # necessar ad IdWefltt by block mbor)

MG AF4"RACr (Vowemoru o,. reoe~ 06D se em ad fdontifr bw War&h #w~a~b)'I he palrarn~eter" that .xlk-ct the htiri.-tntal -%tablilty oftairttl~d ,%urtacing mat% %cre 'tudied h\ coilnutctilg a

Mcritc' ot %tatic lull-scale buckling te%t% in the~ alhi'rato'n u%rnng %aini1U% Iflt% anti ~i% patternl%. A\12. X\I IX. anidN\\ I 11 nlaft-% %ith 'dmiulattd ua:terprrtotfing uere u'cd in li the te.ing. I he mut ie't %melion% ranged Irom tine tol ti~cpan,+-' %idc '%ith \%idth% All tip it, R%' It. It %a fitind that the* tinom predo~minant Iacti'rattecting the hutkling load\%t, the initial Lextcorieitt charaeiterictie t Ith nut %\tm Other rficttr% %~ere Ipanel Aidth, nut unit %c; .fit..%lid toirration 'uaidth. I he magnitude ot i~cking angek 'tad no' elketc tin the hit -kLung load hut did ailedi 0.'pro'tile ofi the hucklcd m'aw. I he cv.~tcnx olf re'.ilier': till.:r in the ti'int% tit the \waterprquitin; mat-- reduceed thelou~kmg~ angle hut did no~t iinercaw the buclikng load.

(Co'ntginued)

DO 73 103 EDITIOn OF N OVl GS ISOBSOLETE I'nclav'ilieujSECURITY CLASSIFICATION OF THIS PAGE (Who. Vora rnistod)

I PICS U~ EL U CMHGE

UnclassifiedSECURITY CLASSIFICATION OF THIS PAOE(nmPalo DE. ntand)

20, ABSTRACT (Continued).

A series or buckling tests of model AM,^ mats or variouts widths were conducted in the laboratory thermatswere obtained from Utah State University. The results enabled the extrapolation of full-sealc mat tests fromshort widths in the laborator/ to airfield widths.

Attempts were made to develop a mathematical model to compare wvith the buc'kling bechavior of theprototype mats. It was round that the horizontal load at which buckling wvould be initiated depends aimostexclusively on thec vertic~al eccentricities existing in the mnat at the time the load is applied; because of the randomrnature of initial irregularities in the real system, the initial buckling load is an unreliable measure of 'he load-carrying capacity of the system. Therefore, it %kas concluded that little bcnfit could he gained from mcort.elaborate mathematical ana?.yscs directed toward a more exact determination of initial buckling load or (ifsustained posthuckling resistance. Analyses were thus made by a simplified articulated system composed ofstraight rigid bars and mnovement-free joints. Only qualitative comparisons between the idealized and real[systemis were made because (a) the joints of the real mat are not miovement-free before locking. (b) there areshifts in the joints caused by horizontal load transfer through the joints ditring buckling.

Mehods by which the stability can be increased included several alternative lay patterns. These patterns,which enhance the postbuckling behavior and may increase the initial buckling lexid, are sugg~ted and

explained.

stc awlt S401aa D J J~,

i UJ4

OT19 1916::UTY M3,

at .- 1 04,0" boom2LLW00 4

-0)

I THE CONTENTS OFTUIS, REPORT ARE NOTTO BE UiSED FOR ADVERTISING.PUJBLICATION. OR PROMOTIONALPURPO)SES. CITATION OF TRADE NAMESDOES NOTI CONSTITUTE AN OFFICIAL.ENDORSEMENT OR APPROVAL OF TillUSE OF SUCH G(OMNIERCIAL PRODUCTS.

PREFACE

The study reported herein was conducted at the U. S. Army Engineer Waterways Experiment

Station (WES) under Department of the Army Project 4A762719AT31. Research for Lines ofCommunication Facilities in Theater of Operations. Military Engineering RI)TE Program. Task 02."Air Iinc of Ccrmmunication Facilities." Work Unit OttI, "Evaluation of IParameters Affecting theHiori/ontal Stability of Landing Mats," under the sponsorship and gaidance of the Office, Chief ofEngineers. U. S. Army.

The study %as conducted by Drs. Y. 1. Chou and W. R. Ba.kcr during the period July 1972-December 1974 1 ider the gentral supervision u, Mr. J. P. Sale. Chief. Soils and Pavements Laboratory.Professor W. P. Dawkins of Oklahoma State University was zmployed by WVES as a technicalconsultant. This report was prepared by' Drs. Chou and Ikrkcr. Professor Dawkins prepared Part VI."Mathematical Analysis of Landing Mats." and made suggestions and comments on the report.

BG E, 1). Peixotto. CE. and COL Gi. it, Hilt, CF. were lDirectors of WES during the conduct of thisinvestigation ,ud the preparation of this report. Mr. F. R. Brown was the T"chnical D 'ector.

II

i . -. *. •

CONTENTS

Page

PREFACE ......................................................................... 2

CONVERSION FACTORS. U. S. CUSTOMARY 1T0 METRIC (SI)UNITS OF MEASUREMENT . 4

PART 1: INTRODUC'TION ........................................... 5

Background .................................................................... 5Purpose and Scope ............................................................ . 6

PART II: LABORATORY FULL-SCALE BUCKLING TESTSOF LANDING MATS . ........................................ 8

Equipment ..................................................................... 81vst M uthod ................................ .................................. 8

PART III: LAB01ATORY BUCKLING TESTS OF MODEL AM2 MATS ............. It

PART IV: ANALYSIS AND PRESENTATION OF RESULTS ........................ 13

(ChI racteristics of Mats Tested ......... ......................................... 13 jM od l AN12 M at% ............................................................... 13:Prototype Buckling Tests ............................................ 14"Comparisom of Buckling Loads of Different Mats ............ ................. 29Simarlatcd Watcrproof XM1S Mt ..................................... 29

....................................................................... 31

PART V: FACTORS AFFECTING PERFORMANCE OF MATSDWRING LANDING OPERATIONS ................................. 32

PART VI: MA IMUMATICA! . ANALYSIS OF LANDING MATS 35

PART VII: CONVLUSIONS AND RECOMMEODA ••NS .............. 45

.................................................................... ....... 45

S..........................................lABI -gE •(S 4'?

APPI)NM A., AM2 IýA• .XO ,%MAi P! RV0AAUt& I 11NOWR 0-5A,. IRAPI1C Atl DOESS AVIM tU.AS

4...1-

'C,

CONVERSION FACTORS, U. S. CUSTOMARY TO METRIC (SI)UNITS OF MEASUREMENT

U. S. custoinary unit of measwc•inz used in thi report can be converted to metric (I) units asWfoows:

Multiply By TO Obtain

inches 2.54 centimetrC3feet 0.3048 metressquare inches 6.4516 square centiirjaoespounds (mass) 0.4535924 kilogramspounds (force) 4.448222 ncwionskips (force) 4,1t2fl kilonewtonspounds (force) per square inch 6.894757 kilopascalspaunds (mas) per square foot .4.882428 kilograms per square mercewich-pounds (force) 0. 1 12984a newtoi-metresfoot-polmnds (force) 1.355818 newton-metresinches per minute 2.54 centinltres per m .knots (internatimon) 0.5144$ 4 mectres/!Ccond

dwgees (an4&) 0.1452 adian

'-,-S

>-. -I:.:,- -- . .

tA

i .

,. - • • ,. . ..

EVALUATION OF PARAMETERSAFFECTING HORIZONTAL

STABILITY OF LANDING MATS

PART t: INTRODUCTION

SACKGROUND

SI, n August 1970. an experiment on the ievrations of the C-SA aircraft ont the landing mat test' t 4 was conducted at Vyen Air Force Base (AFB). Texas. A failure of an ANM2 mar runway system

was observed. indicatnig an insta bility of mat r unways t o loads, appd 4by aireraft braking forces duringlanding operations. A detailed description of the observations of'the incident is reported in Reference I.To facilitate this presentation. the observation is briefly described in the following paragraphs

2. Thc landing mra test facility was 6000 ft long and 96 to98 ft wide. and consisted of XMIS.XM 19. and AM2 Landing mats The C-SA used in thi operation had an initial $rossweight of 480.0001bwith 82-psI tire pessure on the main gears and 108 psi on the nos gear. Prior to thO landing whichcaused the failure incident, Ohe test facility haW alre3adCye rincd fourtake-offs and three bluian s. Inthe fourth landing, th aircraft made at, itiliat main gear touchdown on XMI9 mat at a spNWd ofapproximately 140 to 150 kots. Only brakes and spoilers were used in stopp ng the aircraft. The initialdamag$ to the mat ofiginated at the tranxit4in betwcn tft XM 419 and AM2 mams which isaboul 1200 ft

* . ~~~~from the location where braking Started. The center Portion of t1W NI ha shfted in h adndircction approxirnatelv 11 in. the uncrlap-fcnutctransiionl-ptLr bad be bent aWn dttrtednd

* . ,cmieof the conncttniqw were sheared off. The da-ag#d mat wjý prin ,rily Mtit(M~awere p~ced inrhoara, behin the ANI nM bPdaue A4mtdbwcndiwn wcteAWOnV30 . OwENIimd joinus due tpjoiAt (Adore. End joint Weld &vIuorc* were cvsdem0 wLith p4,nd disortion Of alltype Nwnrriaosu '4 joknts were toga 0 owt Ind sefeAl po6ocs of pWnel Were dri6n inO the ground.

WaS CAW lofa S4 av "toIUS Ieca d4ý a-ftr r'Sc apedm at UWct AES I ockhcd- pcewaMl 4Wn &'f4octso I wo

p4c-nn i *4aW41 t- k ie r W I t. MWtass p M Wcn.c k at t % 64i *Cti, 1Tho w.5 b (ii brk441 ct i4fAt% K to 'M t afat U ijth h u.f( vtt, kt Ititwumý h t -. %twCtf. htCuiM

poablF b-c 100 puctat Wain 4t the tirto sac the arstWa dcsw tnu vwag 44t

SJ.!

I* fotWa th waw dtspc 1An4 .

* .. itIfs~h sk.' fotces sadoed 4- the, oA d ii te A bta)64 s ug u 4*rsS dwe thepotai"W Wof%"Jog" Wib te a dS04*&L

tA Wok 66 "A Wain "*ew vu t Q) v saanw -s Wi mesaSg Saw 1514 nia e~ 1~ w6 r

4. After the failure incident at Dyess A120, a wcrits or tests and studies were conduitt at the U$. S.Army Engineer Waterways Expcriment Station (WES) and at Wait State t1nivers~ty to .udy' thebuckling characteristics of landing mats subjected. to horizontal hiads for better understai...,ng anddefinition of the problems associated with the C-Saicraft and landing mat. Thesestdiesa tesLbaftWalely dewcribcd in the following paragraphs.

a. Imm.Wiately after the incident. tension anti compresdun tests and tests to del~rnine th..coefficient af friction berevath the na, were conduwted at WEiS to obtain additionalinfotroation concerning the behav'ior of the ANI2 mat. I t~ itest were comtpttcd in

L-tober 1910. Valuabk resulls Were obtained and warranted further full-scale laboiratorybuckli;-g tests. 17W material presented in th6 report is a continuation of the 1970)

int~aton o frthr ud~rtand the factors affecting the hori/ontal ~iia o adingmats, The results of 14h- 1970 study were reported in wwwworadwa t'orn and are includcd inAppendix A.

b. .Ik'ltowingIth 1970 tests. extensive skid tests were conduicied at WETS on XNI 11. XN1 M 1.and11landing mat% placed irn contact with soil and placed tin vaembrane ou "oL The resul6

werc reporttd in the form (if a nwnatoraradum for record..c. A mnid;: stUdy of C-5A landings on AM N2landing mats sponsored by the Air Force Weapons

Laboratory IAITWI.) was initiated at tVtah Staw Unsvefsity in 1974. A rmathematical 11aodeltio silluatet the buckling response oif thte tria'i to th;: hora/ttntal ltAds way, atso elp.

1 h reuls o th sudyarereortd a Rfernce3.A metig ~ hld n 972tocvaluatc(1tah State's work and to gta recommnictdations for WFSs prqvumr to cvaltuaie Lndingmat bhchioitr under C.SA loadings. 'I he review ing board consited- of engi neers frout WIFTS.FýOtecs~r F~. I.- wilson of the vjiviersity ot California. Profcessor M1. V. Mirr of PurdueUniversity, Atnd Profkssom IT. .1, liroNbrg of the Ujniversity of Illinois. I1 hC ggneralco'sn.-su--iof th4! reting %a% that l'ecausc of the evtr 4; cop~tV Of the C tn.a *cot practical it) atterlipt An Cbo~rat.ý itechanistiic analytical so~lution; rather. a s~iimpkermodl dealinig with statkic em lmyeaaW~so be tried, itierctorc. Ntotcý.o!W. P. PIAt-kin of the Oklahorna State kuzim.ersity woas Cuplo0yed by WITS to %tudly andoatiatvio the resvlt of the bwUckirt tes*t% repsrti hqrmnl alu to Oxwmitv the tmmt o

etvopoig an 4naytw.tal c.esoi to siUn uM ce o wt z sa I'Mtsrepor 4essers tN. work 01 MIc*Woe tuwýOvs.

Jý At OwCMPkW of tmothoh she model %tuds cut ANUtnat*. AN Swah mughvetok ra w6t~'

t Ire h [o wAU f*;s id4 V tu4*4 e' o t tt~w- ut * t rusva -ý-pt c, w t m at %t orxtlt -,v 4 . wa wt

triAfli wtie 4-14 amih ~toraror Nwbuacgn !Ns,4 on, pfý*AlP tie$* With WmA14 sto~ thw

* 04 L yJo prt oktýa tof~st~ tko''kO *tht ksto rto & tsvww vutt 169w's#-tN PO&WtN4a 40M ONOuct LWI'd1of at-tt~~~h&atw to piedm &niv !'.wLts %khwfi * bovoth I ,a~dag k, 4

fis N ,.th ij tý*Aph W.;. V 1 taa-%;* Ptoo. a-tft Th 4 hwo

1wttu WON~ _M~4,v %;tftWOO4 to fkPL tei e qslhftes

1. fi Iftcs %WO werec qu'ivik-u to~ 6 to '96 ft in prowuivji wale. The Purpose ~ tit Ithc k 11140 .Iing

~r~Y( .4:i.A :IIICr4'6.; Wr~d 46 wv ,-duechcd to situdy ilw b".1fiol; b~twhavicr ol 11w tluzt symcol

resuts wrc UU ýOwfAff~~h x~vrllwtal ita

PART 11 LW8ORATORY FULL-SCALE BUICKLING TESTSOF LANDING MATS

MOM'QENT

The nulgnitwke of 1114 Iuad arid its eurrcpoudsitg Lui uwcuct wgrc vavwrdcd ctctruuicdtv%wcrc ptoucd Oil a separate X*Y. rceordcr. L,%QJ speed va's abut 0 tH.'ifill.

A

KQW 44WQ uww

ItatW4

*911

Fipn* ~~sM ,tgu% kw

ru'. 1 ~: 4..' I'

!7[

F-

Figure 2. Overal view o t ma t Iormation ritledbok

+. .. .. ... .

AI

7 77 'MR

Figure 4. Mat formation with concrete blocks during buckling

was approximately 2000 lb each and that of a concrete block approximately 450 Ib; therefore, the

concrett 'ock more loosely approximated the weight of the extra width of mat.

10. O .e shapes of the mat system during the buckling process were measured optically using a level* !set. The elevations of each mat at the center line w the entire mat system in the longitudinal direction (in

the direction of the applied force) were men,-ured before and during the buckling process.

I. Since the mat runways in an actual airfield are never placed on a perfectly flat subgrade surface,

ecceiaric.ties inevitably exist in the mat system during braking operations. To study the effect of"'A ,eccentricity on the buckling behavior of the mat, laboratory tests were conducted with a mat system

having various magnitudes of eccentricity. Eccentricity was introduced by placing a steP! eod across thecenter line of the mat system in the direction perpendicular to the application of the load.

12. Each mat formation was subjected to repeated buckling tests (generally four or five

repetitions). Results of repetitive tests of a formation were found to be reproducible if the mat was pulledto a full extended position after each loading. Otherwise. for a subsequent loading the buckling load was

smaller than that of the previous test.

10

. : .. ..

~YI

PART III: LABORATORY BUCKLING TESTS OFMODEL AM2 MATS

13. The model AM2 mats were obtained from the Utah State University. The mats were fabricated ~to a 1/ 7 scale with respect to the width, length, and weight of the prototype, but were ohrienot in

acodnewith teengineering siiiue apticularly with repc othe geometry adsifesothe joints. Figure 5 shows two model mats connected along the side joint. A detailed description of themats can be found in Reference 3.

Figufe M Aodel AA*2 mat

14. Tests wi~re condlicted on various widths ranging from one to eight panels. The width of the

eight-panel formation corresponds to the 96-ft width oif the prototype mat runway at M~ess AEB.

Figure 6 shows the setup of the model buckling test, the load was applied to the mat by a field CDRmachine through an arrangement to simulate the hotriiontal force applied by two C-5A 12-whcel gcar

assemblies. Tests were also conducted by pushing thentats witL the CHR piston. wbich hasa dialnetea.of2 in.

15. Because the joitits (if the model mat'i were very weak, difficulty was expericened in assembhling

thle 111ts. In some cases. the mats had to be pried open by a knife for aswimbly purposes. Unlike thleprototypc buckling tests. reproducible results could niot be obtained in the moidel mat testing lthe

buckling load alwayi decreased as the test was repecated. E~vidently. the joints. were hent once the nmtswere buckled. the subsequent buckling loiad deecrased because of' the damaged joints even though 1K,mats were returned to a fully extended position before lthe test was repeated.

Alm

ttul~Sttr~ al moafl AL92 bn *teng t-.st

UM

PART -IV:-'ý ANALYSIS AND PRESENTATION OF RESULT$

CHARACTERISTICS OF MATS TESTED

16. General descriptions of AM2, XM 18, and XM 19 mats can 6-, round in References 5. 6, and 7,respectively, and are not repeated in this report. Waterproof XM 18 mats were not 3vailable during thetests~ however, tests were conducted on XM 18 mats wiith a heavy-duty rubber hose and plastpc wires'inserted along the joints. It is believed that the mats with the inserts simulated prototype waterproofmats insofar as bucklin~g tests are concerned.

MODEL AM2 MATES

17. Figure 7 shows the relationships between mat width and buickling load for model ANI2 mats.Curve A shows the results for the mats pushed by the 3-sq-in.- COR piston. It can be seen that thebuckling load increased almost linearly with increasing mat wdth; az; the number of mats increasedbeyone three panels (o7 .36 ft in prototype) the buckling load rate of increase began to slow. As thenumb4,4 of the mats was 4icreased beyond five panels (or 60ft in prototype). the rate of increase droppedsharply. Curvc 53 shows the r-ults for the mats pushed by the two simulated C-SA 12-wheel Searassemblies.

184, As explained earlier, the purpose of the laboratory tcsts was to obtain inkormation toextrapolate laboratory results of buckling tests to mat formation of greater width. I ig ire 7 provides thisinformation. It should be noted, however. that since the model mats werc not manufactured accordingto engineering similitude laws, and the joints were very weak, the informatkmr provided in Figure 7should be viewcd as a crude approxunation.

'.4. ~ Yuw~4.

'it.Al

--

PROTOTYPE BUCKLING TESTS

"19. The joint configurations at both sides of the mat in compressed and buckled up positions forXM 19, AM2, and X M 18 mats are shown in Figure 8. The contact points between two pieces of mat are

shiown in the figures. The dimensions are very similar between AM2 and XN4 18 mats: therefore, somedimensions of the two mats are presented. It should be noted that the dimensions of landing mats varyfrom one to the other; the information presented in Figure 8 is measured from the specific mats. Forinstance. the contact points may be more than those shown in Figure.~ & as the buckling process continues.

XM19 Mats

20. The honeycomb X M 19 mat is a 4- by 4-ft square mat with a urit weight o04. I psf. The tests were-conducted with mat formation widths, of one, three, and five panels, and with the connector bars

o perpendicular and parallel to the !:aad application. The tests were also performed with the square matsstaggered with the connector bars in the position perpencicular to the load, which is the normal method

.'. The i..king angle of the longittdiral joint (which is parallel to the landing direction in theactual laying pattern) is larger ilian that of t te transverse joint (which is perpendicular to the landingdirection and has the connector bar); the locking ai.gles arc approximately 10 and 4 deg. respectively.However, these angles vary grc.atly with diffcrent mats. The locking angle is defined as the maximumangle at the joint between two adjacent panels at which free rotation can exist. Tables I and 2 show theresults of buckling tests of the XM 19 mats with th. connector bars perpendicular and parallel to theload. respectively. The mat formation had a width of 20 ft (five panels). The vetrical miovemnent values ateach of the 12 panels delineate the shapes of the buckled mats at different horizontal movemncnts; theyare shown in Figures 9 and 10. respectively, corresponding to Tables I and 2. It isseers that although themaximum buckling loads were nearly the same. the shapes of'the buckled mats were distinctly different.When the joint with connector bars (transverse joint) is perpendihular to the load. the shape of the matsunder the buckling load was similar to a continuous plate. because the locking angle along the transversejoint is small (about 4 deg). the maximum hucklng lhad was obtained durin!g the first buckling and thoload* decrcased as the height of the buckled mat eontiiiucd to increase. It is believed that more pasel.swould be huckied attd the height of the buckled wats would mut~nuc to tnumras if the length of the matsystem had been increased during the test.

L., ikcaute the hloking atl•. along thie lougivtdintal joint ituintv 'itfttql connect1or pin,, is lUrge(,about 10 degli fewer paneln were displa•cd v•etically during huckling than when tile tlranVCrs jointsuere P!rt~ftdi4ular to the l-ad. It ii helieved that it the buckhng ploc"t w"C otined. the mat WOuWdbrcak along the joint* at locatt.to A at.d 9 m Figure 10,

11, At the top o is mgure If. the shapes of the bukckld tnatsa loca-tions 7-I10 of higureo lfduving thelavt tw (:neeuttVe loding %Vogt* are presented, ifte figures were drawti to the s-atic %scale sin -sertial ard horsilonal direCttorr. It 1% ween that the angle at ltwaltq~n wat I cpegat the end oflictetestWhic was S dCg gyeater tl•tn tho lockitig aigle of the Joints It should he p*4ntd out that at loiato.s 7

7 are ptc-l-nte, The angles at loittitors 8 aint 9 wvre H &-p. %whhh %s riuch lUrger ihlia-i t lhe 4-deeftasitd locking angk., A* nofttd z4attot the f rlwu redN-l4 king angIte %4,6"s geatly with rdie•vt ntAN1

th.- meAsured 4-&-g angle rnta ke less3 thin the true iravge vaitto4; hi to'Ws Ytudy. it was hý-pohcsaxd that the blwfeIang k IWJ would iutawaca d the loo-king aigle 4i

'

COMPRESSED POSITION INITIAL BUCKLED POSITION

~~-CONTACT POWNS

I -p

TRANSVERS, JOINT PERIENDICULAR TO LANDING

LONGITUDINAL JOINT PARALLEL TO LANDING

XMI9

COMPRESSED POSITION BUCKLED P"'*TItON

.- O.1147*

VITAN3VERWS JOINT PERPENOICULAR T0 LAN0ING

AM.

0.01"5

an~e

TR~ANSVERSE JOINT PERPENDICULAR 10 LANDING

i*ul 40 of teult o mol in ctvW u k*be 0141

F TEST NO. 19R-20-0MAT XU19, 4-FT X 4*PT

ZERO ECCENTRICITYWIT420T EJGH48F

NOTE NUMBERS SL LINES .REHORIZONTAL FORC~E (LEFT)AND HORIZONTAL MOVEMENT,RIGH~T)

~3O

z

to2

-Ij

42

0 gg 4a101"OMOVLLOA11#4,N /tPNLftm61 h~wofjkW# 1suw at ih owepm p"I~ t he6o9 ~

61"oto

the joint could be reduced. The results shown in Figures 9-11 indicate that the lotking angle affects theshape of the buckled mat surface but not the buckling load. The influence of the locking angle will bediscussed subsequently. Although the maximum buckling loads are nearly the same, the advantage ofhaving a continuous and smooth buckled mat surface as shown in Figure 9 may be rewarding; the bowwave appeared to be so flexible that the chance for the mat to break is believed to be smaller than whenthe mats have an abrupt surface as shown.i in Figure 10.

25. Table 3 shows the test results of staggered XMI9 square mats with connector bars

14-TEST NO. 19-20-0MAT XM19, 4-FT X 4-FTWIDTH 20-FT, LENGTH 48-FTZERO ECCENTRICITY

S-NG NUMBERS BYLINES AREHORIZONTAL FORCE (LEFT)AND z4ORtZONTA'. MOVEMENT

to-

17.65 Kips /543 iv

*0

4..1

I. I . . . . . .. . . . .I. . . .i .. . . L ..

W- PiZ•ONTAt LOCfIPONS , NO Or PANiIN

tOSao fICW4t ~AVfal ~ N WJWp~W1 h W &

perpendicular to the load. The maximum buckling load was nearly dnubled as compared with the mat: " ~system when joints perpendicular to the loid were not staggered (Figure 9and Table I1). Figure 12 shows

the shapes of the buckled mats at three consecutive ioading stages. It can be noted that when compared,at the same horizontal movement, the heights of the buckled mats w m. about the same whether the matsw'.re staggered (Figure 12) or not (Figure 9). The shapes of the staggered mats during the two lastloading stages are shown at the bottom of Figure II.

26. Figure 13 shows the effects of eccentricity on the nmaximum buckling l3adsof X1M19 nms. The

maximum eccentricity used in the tests was I in. It is seen that for mat formations of one-mat width, i.e.,4 ft, and ! 2 ft with joints not staggered. eccentricity did not affect the buckling loads. The. test results

were rather inconsisieit for mats with staggered joints. This may be due to the fact that the mats were notlaid exactly flat at the zero-eccentricity position. i.e.. the mats bad initial ecctntricities in some areas

because of roughness and irregularity in the joints. Consequently when the 0.5-in. eccentricity wasinduced, the mat system became flatter and the buckling loads were thus increased. In most cases, thebuckling loads were greatly deceamsd when the eccnticity was increased to I in.

27. The aluminum AMM2 landing mat is 12 ft long and 2 ft wide with a unit weight of 6.1 psI. Testswere conducted with widths of 6. 12, and 24 ft, and were conducted only with the connector bars parallel

to the load application which is the actual laying pattern in aidrfeid runways. The measured locking angleof the•AM2 mat is about 10 deg.

28. Table 4 shows the test results of AM2 mat complex with Z width of 24 ft. Figure 14 depicts thr.

shapes of buckled mats, Because the locking angle is relatively large, the surface of buckled AM2 matwas -ov so smooth and continuous as that of the XM 19 mat with connector pins perpendicular to theload (se Figure 9); however, it was smoother than that of the XNM 19 mat with connector pins parallel to

the load (se- Figure 10). - --

29. Fiure 15 shows the effect* ofeccntricity on buckling loads of AM2 maIs. In the tests with oftwide half panel, a concrete block was placed near the load ýipplication to prevent the premature buckling

"of the mat at th panel. Figure 15 indicates tW the buchiagloadii were vey sensiive toe ntrý,ity.

30. The oluminum XMIS mat is geometrically similar to AM2 mat. i.e.. 12 ft long and . ft waid.

"The unit weight of XMI18 mat 4• 4 "psI, which i tbatantally lct than that Of the Alt It (6.1 pI).The locking angle of the XMI8 mat t approximtely the sme a fthat of the AMr miat (aboukt l des),

Tests tre condoled with widths of 6, 12. 24. and .6 ft. and with theconnwtor harm only Pavolll to thediwtiom of the appli load, which i* the aotual laying pattern in aitfieW runwav*. Table 5 show% the tell

to ults of 24-ft-wie XMI8 tmats In Figute 16, the %hapt of buckled X M1I matA are otwtrawtcd,31, The effets of eccntri•ity andi mat width on buckling loads of XMIO ato are aiwan in

Figure 17, As It the tcts of other matu, the Iuckling to&a• dectetse with itwreae codmxntficoy. Al•o•,the

buckling koadt itwtease with increasng width ot theritats it all threei etx iiis32. S&aust the maximum widtu #4 mat qytem teted in the labommtovy w•a M t 0t i the tmat

runway used in the O ia AFI hid a -idt5 of a ft. 46(t. retc blocks *t plUcd Wlon$ t1W edPte o the364-t mat s•,•cm to sýmutae the adde wteight ts the additasiwl widlh k4 ma, Te•t results itlh atndwithwst the, bloks amk shown la Figuce I S for induae ceuetc f005 rd I Ow. As *ucxrWctd4the budding loads isie raewith the addition of blocs. atiough iadueod cowanricy &Wcus to have

:.S

.. ... . -...... -••-.. ... . . . ..... . . . . ,3• • .. . . .. . . l • • • .

j 4

!A-

i.2

II

g--

4 - A19'

.4 ,

TEST NO. 19S-20-0MAT XM19, 4.FT X e-F T

I' WIDTH 48-FT. LENGTH 20-FT* ~ZERO ECCE#JTRICITY

24 NOTE! NUMBERS BY LINES AREHORIZONTAL FORCE (LEFT)AND HOt ZONTAL MOVEMFENT

cm IRIGHTj

20-4

a-e kLCAIti 0 0"C1

tv0tv mw -ý 0*O A wo*C~ o*ka ,1

tal fojIh o2so . odA u% f 0 1 U CU- oA"qhau'

aIt b ~~%40WO-t t ,t * tW '%4

WJark I-b *ath tU .AI Cjk

aw-d od ~~tt~v(oc U h -k*f ,Aa~ t0 c~ h ~i34 S "IIso-ti wamWwcllbik MI ýc*wf Q&bw

a0 toGMU 20TACE0 s

yKA -I A

I a ' :9 ~ 3 4 I0%

NU -NcsO 0cohA on bcvo k" of 2419 a

coI $ccMC iue2 h%* ebkkIdsifxwhe aiur fth 4 l nikwFguc2

shous dowup viewo a. o P-rIssCOIMi uy tbd"w a

to,41 oeWw t pm

V"uik 01

OF -NCLW LON W 4MTMA

but wwlk, ttu F ro 'AM Efats au eccmn'y *4 O tucb-AntoQ 1 044501 i (daw'SU bw4*5

coaonret No1s Figrc the s teh tcdsuw ae 00414the "%iur of the oits nicpc ue2

ahwiows tK c owc upII&5O tode joit s 1kwoa hc tet odadt hasdy h ukigpca

;i-U Thd actstheX a 1dMUpt. I bc scnbktatgW toad W4 m4ta~ wisdt-h ('wd4rak t thbtt kAt 4tado

the Vt Ma "at becr Kto meg~cd the tc'A "e %L.dtc wt pnmto a5 bthd ma""ithe

21

IEST NO D-1OAT AM2 2-FT X 12-FT

I"4 KIPS ".*94 W. WIDTH 24-FT, LENGTH 22-FTZERO ECCENTRICITY

*J.So DtIPS 'iji I. 40TE NUh.8ERS Uy LiNES AREHOQRIZONTAL. rORCC ILEMFAND Honizo~dAL MOV1UEsiT

zt4 94n t$ .COW.~

WK0UIs

6 ki

AFgw Ni *;oq *I

b~k-44W-a w W vwt ii l-ft oi

ý ow to fkfxL Lhf Ow "L--~ Oe tht-.-aid %w 0#kco kmau~"kV~-%mt v~wtac omwm evted( imo

6 FT, WITH BLOCKz1.5-

zwU

U .0l~j

0.5-

24FT, NO BLOCK

12 FT, NO BLOCK

0 5 I0 15 20 25 30* 10MAXIMUM BUCKLING LOAD , KIPS

Figure 15. Effects of eccentricity on buckling loads of AM2 mats

37. The joints of AM2 and XM18 mats are very similar, as shown in Figure 8. However, thedimensions shown in te figure may vary greatly among different mat sections.

SIMULATED WATERPROOF XM18 MATS

38. The purpose of the tests of simulated waterproof mats was to determine the effect of the fillersinserted along the joints on the buckling loads. Prior to the tests, it was considered that the fillerinsertion would reduce the locking angle and thus increase the buckling loads. As previously mentioned,waterproof XM 18 mats were not available during the tests; thus a 0.5-in. (outside diameter) heavy-dutyrubber hose was used as shown n Figure 25. Because of difficulty in inserting the hose along the fulllength of the transverse joint, 6-in. pieces of hose were placed at two sides and the center of each panel. Itis believed that this arrangement was an adequate simulation of the waterproof XM 18 mats. Tests werealso performed with a 5/ 16-in.-diameter, plastic-coated utility wire inserted along the joint as alsoshown in Figure 25.

39. Test results for a 6-fL-wide mat showed no increase of buckling load with the insertions alongthe joint. Tests were also conducted with the center of the rubber hose filled with plastic wire to increase

r" iits stiffness, but no increase of buckling load was observed. However, it was found that the surface of thebuckled mat with insertions was much smoother and more continuous than that of the buckled matswithout insertions. This is particularly true when plastic-coated wires were placed at the upper part ofthe joint. Tests with mat width greater than 6 ft were not conducted.

23

27TEST NO. G-1

MAT XM18, 2-FT X 12-FTWIDTH 24-FTZERO ECCENTRICITY

24 NOTE: NUMBERS BY LINES AREHORiZONTAL FORCE (LEFT)AND HORIZONTAL MOVEMENT(RIGHT)

0 12

00

z-3

0 2 4 6a 10 1

HOIOTLLCTOS O FPNL

Fiur 16whpso uke M8mt,2 twd

ýj --,7 77 7 7

1.0 A

U

0 5 10 is 20 Z5

36-

~24-

2 12 103 103 ZERO ECCENTRICITY4000 S-iN. ECCENTRICITY

01 A 1.0-IN. ECCENTRICITYA

0 II0 5 10 15 20 25

MAXIMUM BUCKLING LOAD, KIPS

Figure 17. Effects of eccentricity and mat width on buckling loads of XM1E8 mats

1.-

U

U J,.

zz.XgZI J

0tua7.Efcso ~ocnlwl ntiihut ~d tdlEfn cetitV DDDnt

p . . .0

71/

s405-10

Figure 1.9. Surf ace of 36-It-wide XAU8 mats after irst buckling

iuf&V&OW wW ofE $01U18 Ow~a 36 ft wide

"&A l

Ape,

RV

""P wa

-r��-�nrrr� � ____________

I'

4

*iIupC12

I

"Co0�

*2�(3SCo

$a

:1 a

I $

0CU

Cob

U

2W

* * 'I

*v A

tvt

A.M

C N Figure 23. Close-ups of failed joint

du to buckling XAE18 mats

4,12;4-r; VI

45j

40-

35- ~1

'30-L

A A,

lo-NOECNRCIYINUE

A ONCO PN AALE OLA

20 1 03 05W& ITH N

Fow 4Rltoshpo ukigWd n a it frdlUilol

47

N-) FILLER

* /5/IS-IN. WIRE INSERTED 1/2-IoI. RUBBER HOSEZERO LOCKINC ANGLE ZERO LOCKING ANGLE

Figure 21. Configurations of XA,418 mats with and without

tillers during buckling

BUCKLING SHAPE

40. As noted earlier in this report there is a dk'fferencc in the buckled shape depending on the type ofmat, the lay pattern, or the joint locking angle. In the tests it was also observed that the stability of thebuckled form relative to its horizontal position along the mat depended on the buckled shape of the mat.The stability of the buckled shape has significance wvith respect to the ability of the buckling wave tomove ahead of a rolling aircraft. All the evidence~ .' dicates the wave which f'ormed ahead of the C-SAa.Dycss AFB was a stable type which would not move horizontally ahead of the aircraft. As can be notedin the movies and photos which were t~aken of the failure. complete disintegration of the mat occurredwhen the aircraft ran onto the buckled mat. Disintegration was not caused by the horizontal brakingofthe aircraft.

41. The ability of the wave formed by buckling to move ahead of an aircraft depends not only on

the shape of the wave but also oni the speed of the aircraIf and the height of the wave. T'hus- any statement4 ~regarding the ability of an unstablie wave. such as the wave formied by the buckled XMN 119, to run ahead of

a landing aircraft would be pure conjiecture, About the only statement that can heý made at this time is4 - that a wave such as that formed by the buckling ot ANI2 o'r XMI18 will not nmove ahead ol'an aircraft.

However, it is possihie that thle wave formed by the buckling of the XM 19 mat, or by other mnats withInsertions. may move ahecad of the aircraft and roll without causing a mnat lailure.

PART V: FACTORS AFFECTING PERFORMANCE OF MATSDURING LANDING OPERATIONS

42. The forces imposed on a mat runway due to aircraft landing and braking are indicated inFigure 26. These forces are:

a. A normal component W due to the weight of the aircraft.

b. A friction force Fn between the tire and the mat which is proportional to the frictioncoefficient between the tire and the mat pt , the aircraft weight W . and thedeceleration a

c. A friction force F. between the mat and the subgrade which is a'lso proportional toaircraft weight as well as the coefficient of fricton between the mat and subgrade p

d. A tensile force FTV in the mat behind the aircraft.

e. A compressive force F, in the mat ahead of the aircraft.

The maximum horizontal reaction of the mat system occurs when all joints behind the aircraft arefully

extended while all joints ahead are compressed. To reduce the compressive force ahead of the aircraft.which contributes to buckling, it is desirable that as much tensile re-i&tance be mobilized as possible.This can be accomplished if the joints of the system are maintained ii a fully extended position. Theinsertions shown in Figure 25 would assist in keeping the joints in a state capable of resisting tension

without requiring la:ge horizontal displacements. In addition. the surface of the buckled mats issmoother and more continuous with the insertions because the locking angles of the mat are reduced.

This would result in a configuration which would be more likely to move ahead of an aircraft withoutcausing mat failure.

__________AA&W01 ORECTION

-4iAWRAFr' TARE

,mm t-dowr,w

Fogure FW03oc~ on mata Wayd

43. i'est results indicate that the huckling hlid is scisitive twth ocentrlnitý ti the imnat l hectlo

of induced initial eccentricity in most cass •has faarl• pronounccd or a narrow sNysitm but wia ohsured

a, tlte sy,,tecm width increased, It is easier lor a 4- or (-ft sysitm to be laid flat anid Nitraghtl[ a wtdcr % ielwould have a lUrger number Oti natu•al irreput:.rittie% 0'e.. Warped liancl... vrkokLd or ditltgcd lotnlt%ý.

ot.I which t'Auld o(ershadow the Mlfcchit ! l, indutcd eicettivcit. Neqerthekis. test rcsults nidiate

that the runlway SuIsgtade Olould he a' slmoth as pomI'.tblet Morw lfa~i ngllt44, 11 die ect oh pante width on buckling lWd i% eident iront t c,'ul'. ol tic, test' \t l1nut

Uith jolntis not .tapggred. lo use the avate AM2 ,nd XM II h•1•, t!w tlie~tate pattern '.hown In

hqgUre 27 w. iuU increase Ihe huckhing hud saiuce t he hoatgiitudl~I jotlis Ihe 1w M it l tautt are conutaltb

• 32

-7 7-7r -77X 7

TRAFFIC

4X4

Figure 27. Lay pattern for AM2 or XM18 mat to increase elfectwe panel width with XM19 panels

with those of the AM2 and XM 18. This arrangement would make the AM2 and XM 18 panels behavemore like the XM 19 with continuous transverse joints and with connector bars parallcl I Figure: 9) to

traffic. However. the increase of the buckling load is believed to be not very significant.45. The lay pattern shown in Figure 28 for AM2 and XM 18 mats can increase the buckling load.

Special rods and panels placed at the center of the runway will havr to b manulactuzed. Because the

mats are placed at 45-dcg angle:; to the center rods, the component force parallel to the connector pins isreduced, The other component force will not cause tihe mats to shift b4cause t11 mats are kidreduced.cal cawcwmthc ,symmetrically with rcpct to the center rods.

•- -'

gF

Pq* *!j kfft o W*dMNIfttL-ow WIwbv4.4

TR~f"7TRAFFC

Figure 22 ODagonaJ lay Palitemn lor XA#19 mats

"46. The lay patterns for XMI9 mat shown in Figure 29 nmy increase the buckling load to SoimeI

extent. The checkerboard pattern is more continuous but predominant migration is unidirectional andparallel to connector pins. In the diamond pattern the migration problem may be less severe but thc

system has a predominant weak axis parallel to the connector bars.

47. Test results indicate that cdge restraints, either wicghts o'r anchors. are dffective in increasingthe Nckling load when the load is applied near the edge. Hlowe•er, it is bilievcd tlht this clctliveners:would diminish very rapiidly with distance front thc edge and would not pwovWuk much bNawft ko thecenter poction ot the mats.

]1V

PART VI: MATHEMATICAL ANALYSIS OF LANDING MATS

48. The results (f static tes ts of various, n•at systemsw have b n replutted ir -Fiure s -32. In

Figure -40 the buckling ioad has been normalized with resia t•I 14•t wd idth asRI t wll syg 4 *eiht,.

700- LEGEND

-- -O AM2 - NO ECCENTRICITYO.----4. AM2 - WtTH ECCENTRICITY

- XMI8 - NO ECCENTRICITY

9.----, XMIS - WITH ECCENTRICITY600- ......... 0 XMI9 - JOINTS STAGGERED

- ,---0-O XMI9 - JOINTS NOT STAGGERED

.00-

300-

43

•. I.. PINS PARALJ.EL TO LO4L•

zX

• C, PINS P•RPNDLCCJLA TO 4.OAD

Q300

100

0

0 so s e0]o 40 soaMAI wooy, ry

i ,hq

) .• .

].""l T i- ' ' , ," " ' : " : . . . ... '...., , . .. . , , , , , . . '

LEGEND

250C AvZ - %0 ECC%'TQ1`-rv

----- O"] AV - wiTH ECCETA,C-Tý

xvial - No ECCE'4TWC.Tý

~.-.-e ~18- 0. T" ECCEN !N.T*D - *.----0 m ti . JO N TS %O ! JTA 4,:G FP rD

- 200x

40

. \

Zi.

0za to \0304

'-4

WAT W0" FT

50

-zIt

-

0] A I0 20 440

MAT *10Th.rN I•

Fi~ure 37. Normalized buchh.'*q load versus mat width

SA&aKa. 1 .

-Z.

0 Q2 C14 04 0 0 10

MAVIIuuM VE[RTICAL DEFLECTION

DIVIDED BY PANEL WIDT". 1IN

Figure 32. Normalized horizontal load versus normalized vertical deflection

- Co-Y

This plot indicates the beneficial effect of panel width (4 ft for XMI9, 2 ft for AM2 and XMI8) on

buckling load. In Figure 31 the buckling load has been further normalized with respect to panel width.The curves of Figure 31 indicate that the nondimensional bucklitig load is relatively independent of Nii system width within the range of widths tested. -1

49. The variation of the nondimensional horizontal load on the mat with vertical deflection isillustrated in Figure 32 for sev ral mat systems. These curves indicate that the maximum horizontal

reaction of the mat is atlainec. at low vertical displacements and as soon as vertical displacement isinitiated, the horizontal load required to induce further deflection diminishes to essentially a constant 4value and additional imposed horizontal deformation results only in increasing the number of panels"undergoing vertical displacement during the buckling process.

50. Except for the XMI9 mat with staggered joints, the behavior of all mats tested may be

demonstrated by a simplified awliculated system composed of straight rigid bars and moment-free(frictionless) joints as shown in Figure 33a. Once buckling of the real mat system was initiated, thebuckled shape progressed through the series of configurations shown in Figure 33b-f.

51. For the configuration shown in Figure 33, the nondimensional horizontal load, Q , requiredto sustain each of the buckled shapes may be expressed as

3 (-I sx 33- (-1)

~~j. Q ~ ~cot[+ L O->: '4 Q- 44~t 01

and the maximum nondimensicnal vertical displacement A (maximum displacement divided by

panel width) is

2

-=sin [ "(14 + Ji'sin ar( tan 4i+1 (-44 4 J,...---::3 (- l)n•.;., -tan - -

where0 = argle stiown in Figure 33

n = number of patmcs in buckled shapesJi = 0 for n < 4

. • = i for n>4

"J2= 0 for n <6

,J2= I for n 6

T:. •i Plotsof Q versus A areshownin Figure34forvaluesoftheangle 0 Figure 33,upto ldeg.Nondimensiosal potential energy of the system weight (weight potential energy divided by total systemr

weight) as a function of maximum vcrtical displacement is shown in Figure 35."52. The straight mat shown in Figure 33a is capable of transmitting any level of horiitntal toad.

37

- .•,.

K~RIGID 8AR

"P-IRICTIONLESS JOINVTo. INITIAL STRAIGHT MAT

6. TWO PANEL SUCKLED SHAPE

c. THREE PANEL BUCKLED SHAPE

d. FOUR PANEL BUCKLED SHAPE

e FIVE PANEL BUCKLED SHAPE

f. SIX PANEL 9UCKI-ED SHAPE

Figure 33. Buekled conf~iguations of idealized mat

Ho'uemr. :1 atis initial %.'rtiecal displacement kConsistent w~ith otie of 1t1v displac'ed sNiapes of F-igure 33b-lis promu then tile horilornal load riccewry to sustain tle Nyst-con 'it equiahrium decreas"~ rapidl% wsthle deflect ion it~ereawcs.

53. 1 hie joints of thle real mnat are refwa~cly mnoment free, as assumed in the ideahld sv-Jelm ouf>%* up to a liiting value f the an le I Ilv. hlmiting watic. reterred tit is the l u rI"aoe

* dcpvnd%. on tile typv. of inat and thle condition of tle Itiolill and ha% been obs~erved lto ranlge froilappromirtately t o 10) dcg. Wheni 0 reaches tile Iot"mIng ang~le at a Joint. theat that1 joimim 1% tIk lon1ger

moment free and the horizontal load required to induce further vertical deflection increases rapidlydepending on the stiffness of the joint. This is illustrated by the dashed vertical lines in Figure 34 foT alocking angle of 5 deg. Up to this value of displacement, the total potential energy of the system isessentially equal to the potential energy of the system weight. When the locking angle is reached anyadditional work done on the system is stored as strain energy in the locked joint and the total energy ofthe system increases with little change in the weight potential energy. This incrcasc is illustrated by thearrows in Figure 35 for a locking angle of 5 deg.

54. It is now possible to obtain a qualitative visualization of the buckling phenomena observed intests of the real ;nat. Although the real mat was initially flat before loading, there existed myriad naturalirregularities in the system. These irregularities provide the vertical displacements required to beginbuckling. As horizontal load wa, 3pplied to the system, the total potential energy of the system increasedwith the work done oa the system being stored as axial strain energy due to the axial flexibility of the real

"mat panels. As the load and energy increased above the values consistent with those of the initial state,the initial state became unstable and the system underwent a sudden transition to a new configuration ata higher vertical displacement and an attendant reduction of load necessary to sustain equifibrium in thenew shape. When additional horizontal loading effort was applied to the system, the vertical deflection

-increased, ztcompanied by decreasing horizontal load and increasing weight potential energy until thelocking angle at a joint was again ieached. When joint locking occurred, the load required for further

displacement increased due to the moment rcsistance developed in the locked joint and the totalpotential energy of the system increased. The increased energy in the system was stored as strain energyin the locked joint. When the horizontal load an.4 total energy in the current state reached stfficientlV:high levels, the system again became unstable and progressed to a new configuratiin with anaccompanying decrease in axial load. increase in vertical deflection. and increase in the number ofipanels comprising the buckled shape. This process was repeated as the buckled system movedprogressively through the shapes iv,,wn in Figure 33. Illustrative va riat. ins in loads, displacement, and

energy during the buckling process are shown in Figures 36and 37 by the heavy curves. The similaritybetween the speculative load-displacement behavior shown in Figure 36 and the observed behavior ofthe real system. Figure 32, is apparent.

55. Because the joints of lhe real mat are not moment tree before locking, and due to shifts in thepoints of horizontal load transfer through the joint during buckling iscc Figure 8), only qualitativecomparisons between the idealized and real systems are available. Iowever, it is possible to draw severalconclusions which apply to an assessil:ent ot tlie c•-ýillpressive load transter capabilities of the real mat.

a. The horizontal load at which buckling will begin depends almost V\clusiely on the skft.lcaleccentricities existing it) the mat at tile ti,•le tile lokd is applied Blecause of tile randomnature ot' initial irregularities of the real systenm lc.g., %atilvd panels, damaged joints.uneven subgrade) the initial buckling load is an unrecable nmeasurc ot the load-carryingcapacit of tile system.

b. Once bucklingfl his begun. 1he horieontal remsstance of the miat dlnnml•,h%,e% rapidl\ andattailn.s it rclativel% Constant %alt eln though the s,,tevnl Colitlntlnui esto dpla;e ýCrjtallywith progrussiwiy larger ,unuberw, tl paelsl imol~lwd in the huckled shap.,

c. I he rteistance: of the eal tit horlilintal lo;ds alter bluklil depend redonnnantll on thelhwking angle at the iomt. I he IipthtuCklintg re'~i'ta(.3C cap.4bllbt Ol the nuit luWieal¢

* railpidly Ulth decriesed ltok-ullln anglelIe 1t• Pow t4)W. i e. nuill r oi panelV. Inloled in the hbuMkld sihlar unrewi•.s, t redlucd h•l kg lin anglei"L's-n though the sprtwicl &-pUace lntWrease.k 'Alth ite n hnber of •niawl. t1w prtilt. 14

.........................

ca

40

0z

CL Li. u

0 0

0tv*

0 b

1.4'

z .0

LI'

0.1

- o3

w ONLII ScLOiAo

pL

Iwo%.--

*0WWtw"O A41doa

aansj 33) 93 -0 130N

N In 0

I 0

tu

MN:

o;'U

w

w

bf

CJ.

00 w

-lowN

ofN

I1.4

0

Z 4

0.8z

0

N

0 BEHAVIOR OF MAT FOR LOCKINGANGLE OF 5

0O44

0.2

2 PANELS IN BUC'E$.D SHAPE

0 0.0 0,2 0.3 0,4 0.5NONDILIEPSIONALIULI M.AX~IM~UM~ VERTICAL UISPLACMEPAJy. A

FaWOW sho 0Q~?1 4*a:d W vi*f 0~I~? LW WI* vy etow

- ---'s'. -• - • . :•z '-U,•

the buckled shape presented to a landing aircraft is longer and smoother than that for a lownumber of panels and offers the possibility that the buckled shatp will move ahead of theaircraft.

e. Because of the random nature of irregularities in the real system, there appears to be littlebenefit to be gained from more elaborate mathematical analyses directed toward a moreexact determination of initial buckling load or of sustained postbuckling resistance.

*1

I I I " I : I i " I I I I ! - I , - .. . ..

PART VII: CONCLUSIONS AND RECOMMENDATIONS

CONCLUSIONS

Laboratory Buckling Tests

56. Because lthe available data are limited, it is only possible to present the following qualitativeconclusions:

a. The initial buckling load of the mat increases with panel w..th, mat unit weight, and systemwidth.

b. The initial buckling load of the mat decreases rapidly with increased initial eccentricities inthe system.

c. The initial buckling load is independent of' the locking angle of the Joints.

d. The postbuckling resistance of the mat is substantially lower than the initial buckling loadand increases rapidly with decreased locking angle.

e. The profile of the buckled wave becomes smoother and more continuous as the lockingangle is reduced.

j:Insertion of filler rods in the joints nerpendicular to the loading direction reduces thelocking angle and in addition maintains the joints in a fully extended position to permitmaximum mobilization of the tensile resistance of lthe mat behind a landing aircraft.

g. Any revision such as resilient filler insertions or alternative lay patterns which cause themat to tend toward a continuous system or which increase the effective panel width willenhance the postbuckling behavior and may increase the initial buckling load.

Theoretical Analysis57. As previously mentioned, because of the random nature of initial irregularities of the real

system (e~g.. warped panels, damaged joints. uneven subgrade) lthe initial buckling load is an unreliablemeasure of the load-carrying capacity of the system, more elaborate deterministic mathematicalanalyses directed toward a more exact determination of initial buckling load or of' sustainedpostbuckling resistance are unwarranted.

RECOMMENDATIONS

58. 'Ill investigate the potential enhancement tit nia? beia% ior without 1-Ciigetnsv revisionof panel design. the following expecrintental investigations are recommended:

a. The smoother profile tif a buckled mat with filler ld netdinthe mi'mnts nma periit lthebuckled watc to propagate: along the miat wheni it i., encounteredl by lthe 41heel. Thtiýpo~sSibiilty *1hould bet examined~ 411 u ingthe C..SA wheel ah*9-41blý With 2 reLtitely lurge matlaout.

h. TIhe alternative mat la4i'ut showil tin tigure 27s oarelatitcly --%pcdienltneaw oil iticreasinpthe llet iepael ~dtIt.'[he belviuikr to 011%~ Lyout tw~d eexaiilwd Iina Wencs toftem%

W.,adar tot those reprteid above.

'Ith diagL a alarttcrn for \%14trial Na hown Iin 1, gure _1 ntav ii reacae till: lhetoo'pallel width whileo reucuing the ojiinpoteiltt4 lti~hnfointAl load [erivildiulir to theprethittititit Aeak toitit tsuikhoul kwlcncltor rintio 1 NI hebehot oft theC Alternuilse%4141"14 We chillaiwed expearlu~at'ly koth %titl and w114ithot fil- let iwauolt0 Ill 11W %eAk

REFERENCES

I. Green, H. L., -Observation of C-5A Operations on Landing Mat Test Facility, Dyess Air Force Base,Texas," Miscellaneous Paper S-72-10, Mar 1972, U, S. Army Engineer Waterways ExperimentStation, CE, Vicksburg, Miss.

2. Carr. G. L, -Skid Tests on XM 18. XM 19, and I'l I Landing Mat;4 Placed in Contact with Soil andPlaced on Membrane on Soil," Memorandum for Record, 27 Feb 1973, U. S. Army EngineerWaterways Experiment Station, CE, Vicksburg, Miss.

3. Kiefer. F. W., Blotter. P. T., and Christianscii. V. T., "Model Study of C-5A Landings on AM2Landing Mat." Technical Report No. AFWL-T-R-72-210, Aug 1973. Air Force WeaponsLaboratory, Kirtland Air Force Base, N. Mex.

4. . -Model Study of C-5A Landings on D)ow Truss Web Landing Mat," Technit..:Report S-75-3, 1975. Air Force Special Weapons Center (PM RB), Kirtland Air Force Base. N. Mex.

5. Brabston. W. N.. -Evaluation of Three-Piece AM2 Aluminum Landing Mat," Miscellaneous PaperNo. 4-886, Apr 1967, U. S. Army Engineer Waterways' Ex periment Station, CE. Vicksburg, Miss.

6. White. D. W. and Gerard. C. J.. "Evaluation of Dow Chemical Extruded Aluminum Landing Mat(M X 18 E I)." Miscellaneous Paper S-69-5 I. DIec 1969. U. S. Armny Engineer Waterways Experiment

Station. CE. Vicksburg. Miss.

7. teller, L. W., "Kaiser Landing Mat Failure Study (MX 19)." Miscellaneous Paper No. 4-881. Mar1967. U. S. Army Engineer Waterways Experiment Station. CE, Vicksburg. Miss.

S+ + ;. .: - ... . .• - , • .

oLL-j 0 o~

ol CO a C QCO N \Q 4-)

Ht *ý .- .n - ,

C9 ll CC)

ff 00C

C)

~~ ~ ~g o cuC) U^ 00O\N' co aD

14) 0H o m C C) o C

0 LOA4- +) CWCl~l'

0.3 oH AnAC -

ý4 Cd Q z9 C SNId dO.L33NNO3 C

A43ý 4-3 0 0 0.- *' 1L\

+31

Hi' .. -0-'~

A4 -4 JP4.

100000 0-1

H0

0I 0

IoI o5eaI C!7

01 0jOJc

ci rl f C00 Q~ --

C) o

W -4 C404

U CIO

~) 0

f0 n 0 7

~-4~ ~ <71

0 0 O 0 0-4

(y))

r 4 ck

4)U pt0) r4 cli

0 'ý 4- r4* *

4E -

I. . N t

o- F 0 r-

C-n UA k.ý C

0 Q LZ

f5 00

4-'4

r--N

'o'. o o Q 0

0 0 04 r-4 t- CN n \O ,O 0

0M 0X Cýl O. \0 L- 4 -t

H I 0i GuLC'\D\O C0 0 0 0 r-4 CU \ -

Cnj\ OOHOD -t t~C\ 0 4D

00 0 C~j -4-0 Hý Cý -4 LL L89Cpz

H N ic

4-5] \o \ H -:I H~ CC) C100 C\J-. 0C m Ho.~I. '.0 '.04 N Mc 0 c

Nit 0

\D\O00 HDO 0 H 0")Y

Cd 4-)

aCC) 0 0 00 0C

1 N-

E- 0. '- Q . '-.'c

CMj

0- \D0 C; o-,-0

'.0~ 4 H HO-O

EHH

0 I 0 NC 0 0 04'

H4 \H H \'o

'C /4- )

APPENOWXA

AU2 L.ANDING MAT PERFOR4MA14CE UNDER-C-SA TRAFFICAT DYES.S APS. TEXAS

DEPARTME-NT OF T14E ARMYWATERWAYS OCIPERIMENT STATION. CORPS OF ENGINEERS

P. 0. B9X 631VICKSBURG, MISSISSCPPI 39160

oi li.v "PYan Ty ESSF 3 November 1970

f, MEMORANDUM FOR RECORD

SUBJECT. AM2 Landing Mat Pefgormance under C-5A Traffic at Dyess AFB, Texas

I. In a meeting with Mr. Ronald L. Hutchinson on 21 September 1970. Dr. Yu-Tang Chou andDr. Walter R. Barker were requested to study the available irformation concerning the AM2 matfailure at Dyess AFB. Texas. The failure in question occurred on 24 August 1970 during a landingAperation of a C-5A aircraft.

2. The available information on the events of the failure consisted in its emirety of the report byML. H. L. Green and a film taken by Mr. R. H. Ledbetter (selected frames are shown in Figure Al).

3. Information on the A M2 mat is available from reports of simulated-traffic investigations conductedby the Waterways Experiment Station (WES). The report covering the evaluation of May two-pieceAM2 mat is given in Miscellaneous Paper S-68-1 I. One WES report (Miscellaneous Paper S-69-50)covers the reconstruction of the landing mat test facility and its performance during C-141A flight tests.Another report by the U. S. Army Test and Evaluation Command covers ith pflormance of the mat atDyess AFB from 15 August 1966 to 18 October 1967.

4. hI order to obtain additional information concerning the behavior of the A 12 mat, several tests werecondtmcd at W\,S by pcrsoitel of the Flexible Pavenieii Branch. I lihe tests. which are described onpages A8 through AI2, consisted of a tension test of the longitudinal joint (side jointlk thrce bucklingtests, and I test to determine thcv.vefficient ol friction beneath the mat. ivrcs A2, A3. A4-AN (which

.Q. are fromo the ledhetter tilm). A9.A 13. and I ab• Al illustrate thes tests. Additional tests are presentil

beingveonducted by persontIWl of the Mat Section. --xpedict Surfaces BIawih,

5. I he horifontai thrust applied to the mat runwar h. a braking airvralt is a tunctin ot normnal lorte.the ctwiticivnt of friction Netwoen the tlfe and ;iat. and the tivi'icien of trit•tto betwcen the owt andsubgratde. t he re:atlotnhp •cay bh expnscd in equattion turn as

C, v the owl 4;~(kI1 4 I4sMtwviJ t-ouilC, Ism '~t lj ttwtssral

• ~~~~ ~W J;|l 4'tUb orgtttrlthn

WESSI, 3 November 1970SUIJJECU: AM2 Landing Mat Pikrfroritwr undcr C-3A Trafici at Dycsi AiII. Icas

a. First distress noted. (Dot is reference point at edge of b. Bow wave starting to formrunway)

C Bow wave completely bLocliing light beneath plane d Mat Start iag to conw apartI

a*

3f" M-41

i t~t II?1t tIDNIuNnn :ol f iiw * r4v Žt t u

toot'

4 11W

WE~S j 3 November 1970)SUMELT AN42 Latnding Mat l'erformainve under C-5A Trafflic at lDycs AFHi. 'le~as

4 CAD BL OCKS

Z000-4B BLOCKN

dACK(

0-8DOER

~JAWN LB

r CFO Xý00 Lb TO 40.000 Lb AT yU30 VT

t)~o ANU (tl~~ ~crtiwa hcii vom w,"Ated o mmutio %;kWdti-tv tei~th ffoktb KatAt 0K lkfi 4tuCttlxo4 A m1 hskz.k OkW full(Amfl ftAklimct nwrtwintbe ovj ari twbriCthuat k-l~t I'ctv roI I be~e~tk fol. At Ilnafi

4 ltwk tj Il fv6(wiltcit ?UW

to Atomi~o av t'n'Iintiw 4"t~ dut4 h aidzjn le att;fj Ih~t* AIt if %fivu a fk'A-&%-4ht tg C4 the etatw I14Cwemfl in ltwiofi 'IN,1 Am:Tlt, mat"i lJ~nit A Noiaw mrn-i'listig

~u~to cight pahkt, awl 0M.Uhlugil tw,~h of t-4 (I okfl soi havý Nxuv(0j ef~I~j~~~J aan. A ridi,;attcJ that a b ut.Nmtaat hruo 4-4 apotommatcIF 00.0w to Wow t* lbW hki oma

AA

WEiSSE 3 Novembir 1970SIJIJECI: AM2 Lauding Mat Pecrfurmatwe under C-5A haflic at Dyess Ali.T exas

voqSAFoju.% vtu utlixh~% o ikucdhytc oii citt w'lf:o tkt6Wth hAN~w 4 £'O 13 oitdti w XAcwtodhwIstlwahrtinilt1"m rie w 1 i

I t.4o-w lI iclIl% m kj j !I ~)J I-11ý atittl tt1~i 0I411iltK af,4tA.1viu o'k hm fmv % a. ao wikoit tcf m - itdtw u t N ol'oath ottM10.4

Itttk lm 14 t Omwm1Wvv iotl~~o:I h:tIt l!tmmjttvw vilNwl

OX111Atog o k w:Vgf ; ~~vlý

j4V'ct~t ki4- *, it9 :

th tgu ole 44 Pr'n P' L4m (Mmsog~Qie~mnblitwti,6WNg Itwgill fvuvomten 'AWu Oloil Kt3n tttio

ta~~ttt~~t~~buehtmchkn% Oec immduced by the IlN'aset 1,04 1hw -Aircrat tot the d'vApun to i %Iw Oh-tAwloht~ ni~u I tlkt& %NU twt towtwivhw khadti I1,tok~t aptiv It'4( Jkthe r t In ttU-i ot th V,1rt1,l-ttt, Al~i

J the nuat otii ub.Ad tiktdd 11 he Jrit ArwcI Ii: rtthouMAW 0htQ:\ e 1` igAlitelt \I hie it~a.Ai i~AMtd

~ to t mihe tt4I ott v4 Jint tiwall\1 4tiW 14tttt ittititJ -the om : MO-tiapab IKt Jtý.t1%iti t tj:etztad tth t~c tellf Oakt %I' N 0100 Willur NitOt gtmdw1atcd110wN' eWf~iblt WAtl one ituvctoWN 1*wlawo t the drow~.tJ jub ~tle itmtht4 thw munwin ti, jairmmag--%rateti NW) IN Iltt th :m#.atj md Naleto.- tit tt

14k I' t tq ntwl tito It~thtj~~ i t,.~j 4 ~~t 4ittlitthca~t te ~tna. nd nold"w I

WESSF3 November 1970SUBJECT: AM2 la rding Mat Performance under C-5A Traffic at Dyess AFB, Texas

a.~ Fou Paes nbcln-hp

a For pnel inbuc~lig sapeb. Fifth panel has popped up in buckling shape

Ow MAI I..

k; Si~lh P~Wpa~1iug un 0 LUI in fin4J bukising twin

Fig"s AS Sw*# 01 wwEsf ftme "inckJ ft j D ictm ts 0ift n CA4IKIO cie M OucM igw SANO

SAtter -.tud~Jktn Ow iwLthteit tiUforntttaiOd- h..* 111W (OROIVIV WWWA114cn Oil O3W %abje e

~It tow the~ falueat 1,v% Ali. b"uAWtan did mvur I~ kw ~the"tawai by 1wn.M40 W Ilia uIuni1tt

ItPti11 ItW tttt Otm t;ij folUI 1*t#~~A4- 4'nttuu ttVbt 11w v ilto 0 di to w the pttutarjIC114 ie ii h0.41I noi~tW k#On Ilk AtINII.#as

WESSF3 November 1970ISUBJECT: AM2 Landing Mat Performance wider C-SA Traffic at Dyens AFB, Texas

4 14

F AA

IIM

* #%usfl Alp system" r W 4tv A onts M tW ma

WESSF 3 Novcnmber 1970SUBJECT: AM2 Landing Mat Performance under (M3A Traffic at Dyess AFHI Texas

ý.P.

Figu'o 48. Ovoyt viw a$ #Wt sctson stW anchtrage system

A4a~t W~f SAP't W'0CC 0#' OW ALI thS

ttut (W %h~~*4v *6 & e-

4* tt thwt mwr Non :wtat Ihvui ti tAhrn h,ý k4tw4tLt m%vtasuti Ilk, fv%0n4hn.v t; Itlakswtd"04awn.tit b) "(Csna04 th.. itta4uwwv .4 VfikumkAP k iu the nut .A#4 m44ra&l

Vin lo't IN. the W41 6N* ?ntwk-j at i&m57wul~Lht NiW .lb aUt Nscitn L.Jdotct .iut

mo the I" buctwe4 mode, be foula uupmek Aa% a% 4kmuan Ito a~ At

WVESSF 3 Novembr u3 970ISUBJECtl: AM2!Landing Mait Pc.rormanccunderC fAA Traffic at DyenAFIS, l'feW

Fogu A a Com Won I m shoooKvW adovR" c~IAT2ALKWMTF*1,0 .

-j asZOW

Fo;;" AIt ~bu*4- to A4 *-M -*;AA# oi o"I~wjlý To w -k oftw-tw~b

NOW-9 ros Loio.

4 4

P (PLL, OL) ATOf~iUSE0 000 L

'%o 9:A:toA0a%~o w 44A-mS6Mg %

WiiSSi 3 Novemnber 1970SUBJECT: AM" Landing Mait Pterforw~nwc unsk4r C-SA Traffic at Dycm AIIJ. Texas

416* p

F~~ jui~

A?3

N~~w~

l~btC N

36t'b~2AN c6. pi4A W"A4 A

l~~~iwj~~~ LL#i~Ii to__ b Il~~~~~~~.P L (to)ttkti.I tk 4

i~~~~~~~~_O FICAC~~t~m~tc It ~ 4

a ~ it~ kj~th t~ ~ r~ttiCaEFF145c~ Of r ~~ ~4i~cION wa

Iiquoa Au AW~t FetN

Ii~ ~ ~ ~ ~~a Al~;~t.~iT' ,

Codiw~~~~col. t, hto frTo)cc M

wEiSiF 3 Noveintbcr 1974)SUBJLCl AN12 Laning1w M~at l'crlorniuccuntkr C-SA 'lraffiat l1yess AVli.1Icias

a. 1w ~rcru:h f :w jont n bnding is froin AM0 to 454) in.-lb per in. ofi tat. sir troul .36.000 t

SOW00 ft-lb per I 2-at panel.

b. the failure of the male rail cj'uwdJ ttwJoint failure. I5 ltv rsi failure was in the top skin of temalerail; then t11w lower Akin failed.

v. I his mat failure waa. IN. bame tviv of failure observed at 1)hebb AFIJ.

12. '11 tcindings fromt lest No. 3 (buiwl~ing ot 6P-it niat) arc:

a. The strength of thAe joint was 2800 in.-lb per In. or 33.000 alt-lb per 12-0i pancl.

b. Failure of th- top !,kin of the mnal rail eauscd joint tailurc.

e.1wan&l a '.ais approximately 24.5 tleg.

d. dAfter nuekfing be-gin, only abo~ut a 20 percent uwicaw in tore' %hat, required to fail thienuta.

13. 11 he indmV3' from lest% No. 4 (tens10on test) are:

*a. I he stten 'ah tit a n undalmAgedJ toait %&.voapproxoznuicly ISM) blb pr m. ti1 patwl or apptomnutel'(Y 116.00U l eth 5W2-It panel,

hi. I N. mat tiked in tunmaon bv tviuatag of Ow C-rail1 and che Joint bccomiiiitj woupk~d.

i,; I ti upper part of tiw IC-rail eN-tt up%;Ard 0.1lin.

iJ Wiath the C-rAil tkm!. the. nut had to roltat mili abouit 12 iJkP to kiceoltc Utiiouipled onl.

imi U91dainjixJ mnAt. tIW r4tm~ll reqjUtie tor the nl1it tit beeoine1 tiii iuipi m' ;About ~41 dcg-

a4 1he l tWttig.Irk I" tri~iivl S etU~en the btton i011 t4 4 14clW~tI .iu tlV.tion A~I rc~~ ~itA

I Kh,~ tekWdWit of tr..muot lvet et the N4.lktt %it0 Oe "nut anJ tuhgio~a~w~J t~tbI hS

otwt'~e ý.ivftcll~tkt-eI~ 'Am atswltk-J W~ii ith IN- IIIAAI "UI U.. WetV znJ!if" vbr~t th% mut

Ul'o the ~ fitit 1\tt li umo t~im v.4*"d to* ' ra~i ttv0%a toiatiihA4dt.W

It o 'L b~ thAi ul i~~~t~ ;n 1 a14..atoti)?0vamk woom'ý' Now t t4 sfitools VR-1

d t LA he Nku AT o.- 64M c~t%*b.0AFQ .A-wia"t4 k fuA4mw-JQtv o

WINSS. 3 Nownblr i974)SI) IJ~iI:.,I" ANI2 L,'dw Mat l~rlornlilu• uniigr C.-A l"r'JIIh at l-y.•¢b AFBt. tLia•.

c. For the de•,,lopntwi ol a furrc sufficient to buckl t1K i"lat. the ditfftuer in tlh ci oliwrat (ifriction imwe• c the mat and tire and htwtcen the mit arnd suhgrade uould have had to h.. at least 0.?.I fie test contducted indicated a difference ol only 0. !. It is most likdev that the test conduiions for tiedetermination of t4c cockicut of friction .•tk:cCu 14. real and nwnalrag did not dupitaze 11w field

• !ditww 3

I,