analysis of the wingbox with spliced skin and estimation · pdf fileanalysis of the wingbox...

154

Int. J. Mech. Eng. & Rob. Res. 2013 S Sarath et al., 2013

ANALYSIS OF THE WINGBOX WITH SPLICEDSKIN AND ESTIMATION OF THE FATIGUE LIFE

FOR THE WINGBOX

S Sarath1*, Jason Cherian Issac1 and K E Girish2

*Corresponding Author: S Sarath, [email protected]

The attachment joints are inevitable in any large structure like an airframe. Splicing is normallyused to retain a clean aerodynamic surface of the wing skin. The wings are the most importantlift-producing part of the aircraft. Wings vary in design depending upon the aircraft type and itspurpose. The wing box has two crucial joints, the skin splice joint and spar splice joint. Top andbottom skins of inboard and outboard portions are joined together by means of skin splicing.Front and rear spars of inboard and outboard are joined together by means of spar splicing. Theskins resist much of the bending moment in the wing and the spars resist the shear force. In thisstudy the chord-wise splicing of wing skin is considered for a detailed analysis. The splicing isconsidered as a multi row riveted joint under the action of tensile in plane load due to wingbending. Stress analysis of the joint is carried out to compute the stresses at rivet holes due toby-pass load and bearing load. The stresses are estimated using the finite element approachwith the help of PATRAN/NASTRAN. In a structure like airframe, a fatigue crack will appear atthe location of high tensile stress. Further these locations are invariably the sites of high stressconcentration. Life prediction requires a model for fatigue damage accumulation, constantamplitude S-N (stress life) data for various stress ratios and local stress history at the stressconcentration The response of the splice joint will be evaluated. The splice joint is one of thecritical locations for fatigue crack to initiate. In this study prediction of fatigue life for crack initiationwill be carried out at maximum stress location.

Keywords: Splice joint, Finite element analysis, Wingbox, Fatigue, Stress analysis, Lifeprediction

INTRODUCTIONThe wingbox is a structural component in anaircraft designed to provide support and

ISSN 2278 – 0149 www.ijmerr.comVol. 2, No. 2, April 2013

© 2013 IJMERR. All Rights Reserved

Int. J. Mech. Eng. & Rob. Res. 2013

1 SAINTGITS College of Engineering, Kottayam 686002, Kerala, India.2 Banglore Aircraft Industries, Ltd. (P), Bangalore, India.

rigidity to the wings. Designs vary, dependingon the size and function of an aircraft, butgenerally this is the strongest section of the

Research Paper

155


fuselage and it can include a number ofsupportive spars, as well as chambersdesigned to isolate impacts. Usually, thiscomponent is not readily visible, although wecan assume it lies between the wing roots, theparts of the plane where the wings attach.Aircraft in flight experience concentrated shearstresses on their wings. Without adequatesupport, the wings would eventually fold upagainst the side of the plane. The wingboxabsorbs some of this stress and distributes itacross a supportive framework, preventing thewings from wobbling or bending. In addition toholding the wings in place, it helps absorbimpacts sustained during events liketurbulence to keep the plane in the air. In awingbox, most of cases the stringers areattached to the skin through rivets. These jointswill help in to transmit forces mainly along therelength. Forces parallel to the skin and directedat right angles to stringers will be limited bytorsional flexibilty of these members. Forcesnormal to the skin will be limited in magnitudeby the small bending strength of the skin andstringers. Splicing is normally used to retain aclean aerodynamic surface of the wing skin.The splicing is considered as a multi rowriveted joint under the action of tensile in planeload due to wing bending. They are prone tocrack due to fatigue (Michael, 1988).

Fatigue is a phenomenon associated withvariable loading or more precisely to cyclicstressing or straining of a material. Just as wehuman beings get fatigue when a specific taskis repeatedly performed, in a similar mannermetallic components subjected to variableloading get fatigue, which leads to theirpremature failure under specific conditions.Fatigue cracks are most frequently initiated at

sections in a structural member where changesin geometry, e.g., holes, notches or suddenchanges in section, cause stressconcentration (Jaap, 2004).

LITERATURE REVIEWAdarsh et al. (2012) has studied about thestress analysis and fatigue life prediction forsplice joint in an aircraft fuselage through anFEM approach. Aluminium alloy 2024-T351material is considered for all the structuralelements of the panel for fabrication of theaircraft body. Force due to cabin pressurizationis considered as one of the critical load casesfor the fuselage structure. Fuselageexperiences constant amplitude load cyclesdue to pressurization. Splice joints are usedfor the fuselage structure. Typical splice jointpanel consisting of skin plates, doubler plateand a longitudinal stiffener is considered forthe study. The project includes the stressanalysis of a splice joint in a transport aircraft.A two-dimensional finite element-analysis iscarried out on the splice joint panel. Distributionof fasteners loads and local stress field at rivetlocations are studied using finite elementanalysis. The work involved the analysis of thesplice joint using software’s MSC/NASTRANand MSC/PATRAN. A two-dimensional finiteelement-analysis is carried out on the splicejoint panel. Distribution of fasteners loads andlocal stress field at rivet locations is studiedfrom finite element analysis. The work alsoinvolves the modifications required to correctthe boundary effects of the panel. The globalfinite element analysis of a segment of typicalfuselage will be carried out. This global finiteelement analysis results will be bench markfor comparing the results from the splice jointpanel analysis. Repeated finite element

156


analysis is carried out to get the response ofthe parent structure (fuselage) at the jointlocation. The response of the splice joint isevaluated. The splice joint is one of the criticallocations for fatigue crack to initiate. Henceprediction of fatigue life for crack initiation iscarried out at maximum stress location.

New concepts were proposed related tostructural design, materials, productiontechniques, inspection procedures and loadspectra by Jaap (2009). The paper presentsa personal impression of evaluatingexperience, design aspects, predictions andexperiments. It explains that predictions offatigue properties and experimentalverifications are most important tools.

Xiong and Bedair (1999) mentioned aboutmodeling procedures for the stress analysisof riveted lap joints in aircraft structure.Analytical methods have been developedbased on the complex variational approach forlap joints with single or multiple rivet holes. Thejoined plates can be either metallic orcomposite materials. The stresses in the twojoined plates and the rivet loads is determined.Finite element analyses are conducted usingthe commercial packages MSC/Patran andMSC/Nastran.

Amarendra (2006) conducted a study wherethe main objective of the research was toestablish a link between critical rivetingprocess parameters on the potential of fatiguedamage in the joint. Aircraft fuselage splicesare fatigue critical structures and the damageassociated with these structures has beenwidely recognized as a safety issue that needsto be addressed in the structural integrity ofaging aircraft. An effective means for structuralevaluations of airworthiness of aging aircraft

and obtaining essential data for evaluation ofsuch type of fatigue cracking is airframeteardown inspections and laboratory fatiguetesting of lap joint. The Federal AviationAdministration and Delta Airlines teamed upin such an effort to conduct destructiveevaluation, inspection and extended fatiguetesting of a retired Boeing 727-232 (B727)passenger aircraft near its design service goal.Preliminary visual inspection revealed a largenumber of cracks in the aircraft fuselage lapjoint emanating from the rivet/skin interface.Most of these cracks were observed in thelower skin such that they could not be detectedunder an operator’s routine maintenance. Thepresence of these cracks was attributed to thesharp gradients of stress arising from contactbetween the installed rivet and rivet holes. Theresidual stress field generated during the rivetinstallation has a strong impact on thenucleation and propagation of fatigue cracksat and around the rivet/skin interface. Floyd(2010) determined the applicability of anapproach to predict the number of cycles offatigue loading of a structure to failure.

PROBLEM DEFINITIONIn this study the chord-wise splicing of wingskin is considered for a detailed analysis. Thesplicing is considered as a multi row rivetedjoint under the action of tensile in plane loaddue to wing bending. Stress analysis of the jointis carried out to compute the stresses at rivetholes due to by-pass load and bearing load.The main objective are:

• Global and local stress analysis of the splicejoint in an aircraft wingbox to compute thestresses at rivet holes due to tension withthe help of MSC PATRAN and MSCNASTRAN.

157


• Fatigue life prediction for crack initiation atthe splice joint region by Miner’s Rule.

Al 2024-T351 is used in current wingboxdue to high strength and fatigue resistanceproperties. The ultimate tensile strength of thismaterial is 68 ksi (470 MPa) and yield strengthis 41 ksi (280 MPa) and it has an elongationof 19% (Michael, 1993).

Geometrical Configuration

Wingbox modeled in CATIA was been shownin Figure 1. It consists of different structures.Wingbox used here consists of five ribsincluding a middle rib, stiffners, bottom and topskins, spars. Each part is modeled in CATIAsoftware and assembled to form wingbox.

maximum load is acted nearer to the wingroots and minimum load is acted at the tip of awing box.

Load calculation for the wingbox

Weight of the aircraft: 19620 N

Design load factor: 3"g”

Factor of safety: 1.5

Therefore,

Total design load on the aircraft will be:88290 N

As we mentioned earlier, total lift load onthe aircraft is distributed as 80% and 20% onwing and fuselage respectively,

Hence total load acting on the wing = 70632N

Therefore total load acting on the each wing= 35316 N

But we know the resultant load is acting atthe distance 2000 mm from the wing root asshown in Figure 2.

Bending moment at the root of the wing canbe calculated as 70.632 106 Nmm

Figure 1: Wingbox Modeled in CATIA V5

Loads on the Wingbox

Lift load is considered as important criteriawhile designing an aircraft. Fuselage andwings are the two main regions where lift loadacting in an aircraft. Here 80% of the lift loadis acted on the wings (i.e., maximum lift loadis acted on the wings) and remaining 20% inacted on the fuselage. Therefore in wings the

Figure 2: Load Case for Current Wing

158


The load required at section C-C tostimulate the bending moment is P =21361.57 N

Total length of C-C cross section of wingboxincluding all the parts = 6006 mm

Load distributed on the cross section =3.5567 N

Finite Element Package

In this project MSC PATRAN software is usedas the preprocessor and postprocessor. Thepreprocessing task includes building thegeometric model by importing it from CATIAsolid model of wingbox and extractinggeometry, building the finite element model,giving these elements the correct properties,setting the boundary conditions and loadingconditions and finally, assembling theseelements into a connected structure foranalysis. Analysis is done in MSC NASTRAN.The analysis stage simply solves for theunknown degrees of freedom, as well asreactions and stresses. In the post processingstage, the results are evaluated and displayed.The accuracy of these results is postulatedduring this post processing task. The PATRANand NASTRAN software together perform all3 of the principle tasks of a finite elementanalysis (MSC Software, 2010).

ANALYSIS OF WING BOXThe geometric model of the wingbox done inCATIA V5 software package is imported toMSC PATRAN for preprocessing. Each partsare extracted by its points and lines to getgeometrical accuracy to the model. In ourwingbox we do have five ribs including onemiddle rib. Other than middle rib all other fourribs are identical. Each rib consist of two

flanges, two sides and a web. This has to bemeshed separately creating different groups.Special care is to be taken for meshing thesemicircular region in the web or the stiffnercut-out. Middle rib have more wider flange andribs as it has to support the skin splice joint.Finite element properties are provided to twospars used here are C-section structures.While meshing top skin as well as bottom skinof the wingbox it is taken care that mesh seedsare provided for the all the riveting positionsfor the later simplicity. As we know we are usingL-stiffners on bottom part and Z-stiffners on toppart of the wingbox, these stiffners are spacedat a certain gap and running parallel to spar.Special cut out are provided in the ribs toprovide a fine run along the skin. We have threeL-stiffners and four Z-stiffners in our currentwingbox. Rivets are made by using onedimensional beam element. We had alreadyprovided mesh seeds to create node for rivetswherever needed in all other parts. All the jointsare riveted in wingbox, hence a large numberof rivets are used. All the elements of the partsare verif ied for boundaries, normals,connectivity and duplicates. All the elementaland material properties (Aluminium 2024T351) are provided for analysis. Figure 3

Figure 3: Assembled Wingbox with FiniteElement Properties

159


shows the whole finite element meshgeneration.

The stress distribution for the given loadshave been observed and that reveals thestress is distributed uniformly but maximumstresses are developed nearer to spliced jointexactly at the rivets which connects spar andbottom skin as shown in Figure 4. Themagnitude of maximum principal stressdeveloped here is 193 N/mm2. We had alsoconducted an analysis keeping all the rivetrotation constrained in rotating direction (xaxis). Here also the maximum stress isdeveloped on the same location and samerivet but the stress magnitude is decreasedconsiderably to 145 N/mm2. Since themaximum stress we getting is at the samelocation we do local analysis on the specific

location. It includes skin spliced region whichwe can take as two sheet plates and the lowerspar region which we can consider as a Lsection. Corresponding rivets holes areprovided at the exact location. The geometryis as shown in Figure 5. The exact location ofmaximum principal stress from global analysisis created in PATRAN for local analysis withthe help of geometrical tools. It is then meshedseparately forming different groups. Final solidmodeling including corresponding parts forlocal analysis has been shown in Figure 6.

Figure 4: Maximum Principal StressContour of Wingbox

Figure 5: Geometry for Local Analysis

Figure 6: Whole Finite Element Modelfor Local Analysis

160


Average Stress value at a distance 85 mmto 90 mm from the spar tip is taken wheremaximum stress is identified in global analysisfor calculation of load in local analysis andwhich is found to be 8.5 N/mm2.

Hence total load = Stress area = 1700 N

Distributed load on cross section for localanalysis = 17 N

For the load case, one end of the section iscompletely constrained and the other end isapplied with the calculated load of 17 N/mmfrom calculation in order to create tension atthe other tip. Hence one end is constrained inall its six degree of freedom, i.e., translationand rotation.

As in the case of global analysis, theparticular area considered for local analysisundergoes tension in bottom skin. In order tocreate the same surrounding, we constrain anyone or two translation direction, hence we tooktwo cases.

Case 1: With z translation constraint

Case 2: With x and z translation constraint

During different iteration the maximumprincipal stress is found out to be 503 N/mm2

Figure 7: Maximum Principal StressContour for Case 1

Figure 8: Maximum Principal StressContour for Case 2

in case 1 and 250 N/mm2 in case 2 as shownin Figures 7 and 8 respectively.

The structure is safe because the stressmagnitude which was obtained from theanalysis is less than the yield strength of thestructural material. Once wing box is safe fromlinear static analysis, next step is the fatiguelife prediction of the wing box.

FATIGUE LIFECALCULATIONSFrom the global and local stress analysis ofthe wingbox the maximum tensile stresslocation is identified. A fatigue crack will alwaysinitiate from the location of maximum tensilestress. From the stress analysis it is found thatsuch a location is at a rivet hole. Normallyaircraft wing experiences variable spectrumloading during the flight. A typical transportaircraft flight load spectrum is considered forthe fatigue analysis of the wingbox. Calculationof fatigue life is carried out by using Miner‘sRule. For the fatigue calculation the variablespectrum loading is simplified as blockloading. Each block consists of load cyclescorresponding to 100 flights. Damagecalculation is carried out for the completeservice life of the aircraft. The load factor g, is

161


defined as the ratio of the lift of an aircraft toits weight and represents a global measure ofthe load to which the structure of the aircraft issubjected.

The maximum principal stress in localanalysis is 503 N/mm2 is taken for thecalculation, as this is the most safest value topredict the life. As we know the maximumstress value obtained from the analysis iscorresponding to 4.5 g condition. Thereforethe stress value corresponding to 0.5 gcondition is obtained as 55.89 N/mm2.

Correction factors for fatigue lifecalculations of wingbox is considered. Hencemaximum stress with correction factor iscalculated (Jaap, 2004).

• Surface roughness correction factor = 0.8

• Type of loading = 1

• Correction factor for reliability in design =0.897

Maximum stress with correction factor forall the other conditions are calculated andshown in Table 1. When the alternating ormaximum stress is plotted versus the numberof cycles to failure (fatigue life) for a givenmaterial, the curve is known as S-N curve

0.50 55.889 77.88

0.75 83.834 116.83

1.00 111.778 155.77

1.25 139.722 194.70

1.50 167.667 233.65

1.75 195.611 272.60

2.00 223.556 311.50

Table 1: Stress Values at Various “g”Conditions with Correction Factors

“g”Condition

MaximumStress (N/mm2)

Maximum Stress withCorrection Factor (N/mm2)

(Michael, 1988). Using the maximum stressesvalue at different g conditions, correspondingnumber of cycles to failure is obtained fromS-N curve of Aluminium 2024 T351 as shownin Figure 9 (Serrano et al., 2010).

Figure 9: Typical S-N Diagram for FatigueBehavior of Aluminium 2024 T351

The simplest and most practical techniquefor predicting fatigue performance is thePalmgren-Miner hypothesis. The hypothesiscontends that fatigue damage incurred at agiven stress level is proportional to thenumber of cycles applied at that stress leveldivided by the total number of cycles requiredto cause failure at the same level. If therepeated loads are continued at the samelevel unit failure occurs, the cycles ratio willbe equal to one.

From Miner’s equation (Jadav et al., 2012),

ni/N

f = C

where, ni = Applied number of cycles

Nf = number of cycles to failure

Table 2 shows damage D, accumulated oneach range of load condition.

162


Total damage accumulated for all load caseis given by D

a = D

1 + D

2 + D

3 + D

4 + D

5 + D

6 +

D7 = 0.155735

Total damage accumulated is 0.155735,which is less than 1.

Therefore a crack will not get initiated fromthe location of maximum stress in the wingboxfor given load spectrum.

Hence total damage is 0.155735 for 1 blockof loading or for 100 flights.

One flight is considered 10 flying hourswhich eventually means 100 flights as 1000flying hours.

For damage to become critical (D = 1), thenumber of blocks required is 6.421 blocks or6421 hours.

Hence it is advised to meet the wingbox partmaintenance atleast by this required time.

CONCLUSIONStress analysis of the wingbox is carried outand maximum tensile stress is identified at one

of the rivet holes near splice joints which isfound out to be lower than yield strength of thematerial. Local analysis is conducted for thespecific region for maximum principle stress.By local analysis it is validated that themaximum stress is at the same rivet holeduring global analysis. Maximum tensile stressof 503 N/mm2 is observed in the wingbox.Several iterations are carried out to obtain amesh independent value for the maximumstress. A fatigue crack normally initiates fromthe location maximum tensile stress in thestructure. The fatigue calculation is carried outfor an estimation of life to crack initiation. Sincethe damage accumulated is less than thecritical damage the location in the wingbox issafe from fatigue considerations. Life of theparticular region in wingbox is predicted tobecome critical and found out to be 6421 flyinghours or 6.421 blocks, hence advised toconduct the maintenance without fail during thisperiod.

Fatigue crack growth analysis can becarried out in the other parts of the wingboxas future work. As well as damage toleranceevaluation and structural testing of thewingboxcan also be carried out for thecomplete validation of all theoreticalcalculations.

ACKNOWLEDGMENTI express my sincere gratitude to all theengineers in BAIL especially to Mr K E Girishand Dr P K Dash for their support and guidanceto carrying out the study in a successful manner.I stand grateful to my guide Prof Dr JasonCherian Issac, for his guidance, motivation andvaluable advice that he had given methroughout the work.

0.5 “g” to0.75 “g” 48000 3 x 107 1.6 x 10–4

0.75 “g” to1.0 “g” 33000 3 x 106 0.011

1.0 “g” to1.25 “g” 26000 6 x 105 0.0433

1.25 “g” to1.50 “g” 22000 2 x 105 0.1

1.75 “g” 45 6 x 104 7.5 x 10–4

2 “g” 1 4 x 104 2.5 x 10–5

-0.5 “g” to1.5 “g” 100 2 x 105 5 x 10–4

Table 2: Damage Accumulatedfrom Miner’s Formula

Range of“g”

Applied No.of Cycles

(ni)

DamageAccumulated

(ni/N

i)

No. of Cyclesto Failure

(Ni)

163


REFERENCES1. Adarsh Adeppa, Patil M S and Girish K E

(2012), “Stress Analysis and Fatigue LifePrediction for Splice Joint in an AircraftFuselage Through an FEM Approach”,International Journal of Engineering andInnovative Technology (IJEIT), Vol. 1,No. 4, pp. 142-144.

2. Amarendra Atre (2006), “A Finite Elementand Experimental Investigation on theFatigue of Riveted Lap Joints in AircraftApplications”, May, Dissertation, GeorgiaInstitute of Technology.

3. Floyd Caiado (2010), “Fatigue LifeEstimation of Notched Aluminum SheetSpecimens Subjected to Periodic TensileOverloads”, Proceedings of the 6th AnnualGRASP Symposium, pp. 17-21, WichitaState University.

4. Jaap Schijve (2004), “Fatigue ofStructures and Materials”, KluwerAcademic Publishers.

5. Jaap Schijve (2009), “Fatigue Damagein Aircraft Structures Not Wanted butTolerated”, International Journal ofFatigue, Vol. 31, No. 6, pp. 998-1011.

6. Jadav Chetan S, Panchal Khushbu C andPatel Fajalhusen (2012), “Review of the

Fatigue Analysis of an AutomobileFrames”, International Journal ofAdvanced Computer Research, Vol. 2,No. 4, pp. 103-107.

7. Michael Bauccio (1993), ASM MetalsReference Book, 3rd Edition, ASMInternational, Materials Park, OH.

8. Michael C Y Niu (1988), “AirframeStructural Design”, Lockheed InternationalSystem Company, Conmilit Press Ltd.

9. MSC Software (2010), “CompanyOverview Brochure”, Expanding theHorizon of Engineering Simulation, MSCPress Release.

10. Serrano A S, Infante V I M N and MaradoB S D (2010), “Fatigue Life TimePrediction of PoAF Epsilon TB-30 Aircraft– Part I: Implementation of Different CycleCounting Methods to Predict theAccumulated Damage”, CIFIE 2010,Iberian Conference on Fracture andStructural Integrity, March 17-19, FEUP,Porto, Portugal.

11. Xiong Y and Bedair O K (1999),“Analytical and Finite Element Modelingof Riveted Lap Joints in Aircraft Structure”,The American Institute of Aeronauticsand Astronautics (AIAA) Journal, Vol. 37,No. 1, pp. 93-99.

analysis of the wingbox with spliced skin and estimation · pdf fileanalysis of the wingbox...

Documents