Tools › Structures › Joints › Fastened Repair Doubler


Fastened Repair Doubler

A fastened doubler over a cutout or a blended-out damage in a skin, as a repair engineer sizes it: a one-dimensional strip one fastener pitch wide, the skin and the doubler as bars and each fastener as a shear spring whose flexibility comes from the Huth, Swift, Tate and Rosenfeld, Grumman or Boeing relation or is entered. The spring-bar system is solved for the load each fastener row transfers and for the load left in the skin and built up in the doubler at each row, with the skin cut (everything passes through the doubler) or continuous (the load divides by stiffness). Margins of safety follow for fastener shear, bearing in the skin and the doubler, net-section tension, gross-section yield and inter-rivet buckling of the doubler in compression, at ultimate load for strength and at limit load for yield. The layout is compared with the minimum edge distance, spacing, patch gauge and rivet count of FAA AC 43.13-1B, each limit editable, and the first fastener row is described by a fatigue severity factor relative to an open hole. The default is an illustrative 0.040 in 2024-T3 skin with a 0.050 in doubler and three rows of 5/32 in rivets.

  • Displacement compatibility of skin, doubler and fastener springs (the method of NACA TN 1051), n = 1 to 10 rows on each side; the report's five-bolt example and the first-fastener loads of Swift's repair paper are reproduced

  • Fastener flexibility by Huth, Swift (aluminum rivets or steel fasteners, FAA-AIR-90-01), Tate and Rosenfeld, Grumman or Boeing, or an entered stiffness; skin cut out, or continuous with a blended region

  • Fastener shear, bearing with e/D, net section, gross yield, blended section, inter-rivet buckling (NACA TN 3785); strength margins at ultimate, yield margins at limit × FSyld

  • FAA AC 43.13-1B paragraph 4-57 and 4-58 layout rules: edge distance 2 D, spacing 3 D, patch of the same or next heavier gauge, rows for 75 % of the sheet strength

  • Fatigue severity of the first row: Ktb × bearing stress + Ktg × bypass stress against the open hole (Ktb from NASA TP-1862), with Swift's bearing to gross stress ratio and effective stress; a description, not a life

Repair Doubler


Unit System: /

/

Repair and Load










Skin









Doubler













Fasteners









Fatigue Severity




Repair Rules (AC 43.13-1B)






Load Level and Factors




×

Fastened Repair Doubler Method

Fastened Repair Doubler: Load Transfer by Row, Margins, Layout Rules and Fatigue Severity

The repair and the question

A damaged area of skin is cut out, or dressed out to a reduced thickness, and a doubler (patch plate) is fastened over it with several rows of rivets or bolts on each side. The static question is whether the fasteners, the doubler and the skin around the holes carry the load that the removed skin carried. The second question, which governs the life of the repair, is how hard the outermost fastener row loads the skin, because that row sits in skin that still carries nearly its full stress. Both answers depend on how the load divides among the rows, and the rows do not share it equally.

The strip model

The tool follows one strip of the repair, one fastener pitch w wide, along the load. The strip contains one fastener of each row. The skin and the doubler are bars with axial stiffness E w t; between two rows a distance s apart each bar is a spring E w t / s. A fastener is a shear spring kf = 1 / C between the two bars. The plane through the center of the repair is a plane of symmetry, so only one half is solved, with n rows numbered from the end of the doubler (row 1) to the row next to the cutout (row n).

The remote skin stress f puts the load Pw = f w ts into the strip. Row i transfers Fi from the skin to the doubler. Inboard of row i the skin carries Pw − (F1 + … + Fi) and the doubler F1 + … + Fi. The Fi follow from compatibility: the slip of two neighboring fasteners differs by the difference in stretch of the two sheets between them,

C (Fi − Fi+1) = (s / (Es w ts)) × (skin load between the rows) − (s / (Ed w td)) × (doubler load between the rows).

This is the method of Tate and Rosenfeld (NACA TN 1051), written there for a butt joint, and the displacement compatibility analysis that Swift applies to repair doublers and lap splices in Repairs to Damage Tolerant Aircraft (FAA-AIR-90-01, 1990): a strip with each rivet an elastic spring and each length of skin and doubler a bar. The tool solves the same equations as a stiffness matrix of bars and springs. Swift's example case (0.04 in skin, 0.05 in doubler, 3/16 in aluminum rivets at 1 in, 15 ksi: 187.2 lb in the first rivet), his table of first-rivet load against doubler thickness and his charts of first-fastener load for doublers and for lap splices of three to five rows are reproduced by the library. The five-bolt example of the report's Appendix B (bolt loads 0.256, 0.185, 0.159, 0.173 and 0.228 of the joint load) is reproduced to the printed digits.

Two limits make the behavior clear. Very flexible fasteners share the load equally, Pw / n each. Rigid fasteners put all of the transfer into the two end rows, in proportion to the flexibility of the sheets, and nothing into the rows between. Real joints lie between: the end rows carry more than the average, and the first row carries the most when the doubler is stiffer than the skin.

Swift's central point is that a doubler loads the skin at its first row even when the skin under it is undamaged and nothing needs to be transferred: the two sheets are fastened together and must strain together. In his example the first rivet puts a bearing stress of 1.64 times the gross stress into a skin hole that had none, and the open-hole fatigue life he reads for it drops from 160,000 to 53,800 cycles; a thicker doubler makes it worse. A repair sized only for equal or better static strength does not show this.

Cutout or continuous skin

Cutout. The skin carries nothing between the two innermost rows, so the rows of each side transfer the whole strip load: F1 + … + Fn = Pw. For two rows the solution is F1 = Pw (C + a) / (2 C + a + b) with a = s / (Es w ts) and b = s / (Ed w td).

Damage blended out. The skin stays continuous, with a region of length Lb and reduced thickness tb at the center. The skin and the doubler between the innermost rows are two more springs to the plane of symmetry, and the doubler picks up only what its stiffness attracts. With one row the result is F = Pw / (1 + ksc (1 / kf + 1 / kdc)), where ksc and kdc are the stiffness of the skin and of the doubler over half the distance Lc between the innermost rows. The fasteners carry less than in the cutout case, but the blended section of the skin has to be checked, and the first row still loads a hole in fully stressed skin.

Fastener flexibility

The flexibility C of a fastener in single shear is taken from one of the published relations, the same ones as in the Fastened Joint Stiffness tool: Huth (riveted or bolted metallic joints), Swift, the single-shear form of Tate and Rosenfeld, Grumman and Boeing; or a tested stiffness is entered. Swift's relation is printed in his repair paper as δ = F [A + B (D / td + D / ts)] / (E D) with E the sheet modulus, A = 5.0 and B = 0.8 for aluminum rivets and A = 1.666 and B = 0.86 for steel fasteners. The aluminum option writes the first term with the fastener modulus, 5 / (D Ef), which is the same thing when rivet and sheet are both aluminum; the steel option uses the constants and the sheet modulus as printed. They are empirical fits to different fastener types and can differ by a factor of two. The page prints the first-row and last-row load for all of them side by side, which shows how much the answer depends on the choice. A stiffer fastener raises the end-row loads.

Shear

An entered panel shear flow q is handled simply. With a cutout the doubler carries all of it across the opening; with a continuous skin it carries the share Ed td / (Es ts + Ed td). The n fasteners of one column share the shear of one pitch equally, V = qd w / n, and V is added at right angles to the load each row transfers. The sheet margins do not include the shear stress.

Margins of safety

The radio states whether the entered load is ultimate or limit. Strength checks are written at ultimate load (limit × FS) and yield checks at limit load × FSyld:

fastener shear, Psu against the resultant load of the most loaded row; bearing in the skin and in the doubler, Fbru and Fbry against R / (D t), the allowable taken at the e / D of the repair (the 2.0 value at e / D of 2.0 or more, a straight line to the 1.5 value below that); net-section tension of the skin at row 1, Pw / ((w − D) ts), and of the doubler at row n, where each carries its largest load through a line of holes; gross-section yield of both; the blended section of a continuous skin; and, for a compression load, inter-rivet buckling of the doubler between rows as a wide column, Fir = c π² E / (12 (1 − ν²)) (td / s)² with the end-fixity coefficient of the fastener head and the plasticity correction of NACA TN 3785.

Layout rules of AC 43.13-1B

FAA Advisory Circular 43.13-1B, chapter 4, gives minimums for a riveted sheet repair where the manufacturer's data are not available. The tool compares the layout with them and lets each limit be changed:

edge distance not less than 2 rivet diameters and spacing not less than 3 diameters (paragraph 4-57c(1); Figure 4-5 shows 4 D for a two-row pattern and 3 D between the rows of a multiple-row pattern); a patch plate of the same material as the original and of the same or the next heavier gauge (Figure 4-16); a rivet diameter of about three times the thickness of the thicker sheet (paragraph 4-57g(3)); and enough rivets to develop the sheet. Tables 4-9 to 4-11 of the circular list rivets per inch of width, computed for a strip at its tensile allowable with the rivet shear and sheet bearing strengths stated in their notes, and allow 75 percent of that number for single-lap sheet joints. The tool applies the same count with the entered allowables: rows required = 0.75 Ftu ts w / min(Psu, Fbru D ts). The circular gives no limit on the stiffness of the doubler relative to the skin; the ratio Ed td / (Es ts) is reported and can be given a limit of your own. Swift gives no number either. He calls a doubler one gauge thicker than the skin the usual convention, shows that the fatigue life of the skin is extremely sensitive to doubler thickness, and recommends lowering the first-row load instead of limiting a ratio: taper the doubler between the first and second rows (tapering only outboard of the first row does nothing), laminate it from two thin sheets with the inner one extended one row, or finger the inner sheet.

Fatigue severity of the first row

At the first row the hole in the skin sees two loads at once: the fastener bears on it with F1, and the rest of the strip load passes around it. The elastic peak stress at the edge of the hole is estimated by adding the two effects,

σpeak = Ktb F1 / (D ts) + Ktg (Pw − F1) / (w ts),

and divided by the peak stress Ktg f of an open hole in the same skin at the same stress. The ratio is the severity factor: 1.0 means the row is no worse than an open hole, 1.3 means a peak stress 30 percent higher. Ktb comes from the finite-element results of NASA TP-1862 for a pin-loaded hole in a strip (2.20, 1.56, 1.22, 1.02, 0.94 and 0.87 for w / D = 2, 3, 5, 10, 20 and infinity, interpolated in D / w) and Ktg from the fit to Howland's solution for a hole in a strip, Ktg = 0.284 + 2 / (1 − λ) − 0.600 (1 − λ) + 1.32 (1 − λ)², λ = D / w. Both can be replaced by entered values. The idea of ranking joints by such a factor is Jarfall's stress severity factor; this page uses only its elastic stress part, without his factors for hole quality and fastener fit.

Two more numbers are printed for use with fatigue test curves, both from Swift's paper. The first is the ratio of bearing stress to gross stress, σbr / σgr, the parameter of his S-N curves for open and loaded loose-fit holes, which he advises for field repairs because their riveting is often not of production quality. The second is his effective stress for good-quality riveting, reduced from Jarfall's work: σeff = (1.3 Ktbr σbr + Ktbp σbp) / Ktbp with Ktbp = 3.2 and Ktbr = 1.4, that is σeff = 0.569 σbr + σbp, printed here divided by the gross stress. It weighs the fastener load more heavily than the elastic factor above (1.3 × 1.4 = 1.82 against a Ktb near 1.2 at usual pitches), so it is the more severe of the two. The tool ships none of the paper's fatigue curves and gives no life.

Swift also shows what the factor cannot: a crack at the first row grows hidden under an external doubler and is seen from outside only when it has passed the doubler edge (in his 9 in doubler example the visually inspectable crack-growth life falls from 29,600 to 1,200 cycles). A repair to damage-tolerant structure needs its own inspection threshold and interval; the paper gives the method, not a number, and neither does this page.

The factor describes the repair; it is not a life. It answers a design question: which layout loads the first row least. More rows, a thinner or tapered end of the doubler and softer fasteners in the first row all lower it.

What it assumes

Linear elastic sheets and fasteners; the same fastener, pitch and row spacing in every row; rows straight across the load and not staggered; one doubler on one side of the skin (single shear); load along the strip only, uniform across the width, so that every column behaves alike; no friction between the sheets; holes filled, with no clearance. The hole diameter equals the fastener diameter and a countersink is not deducted from the net section. The doubler is as wide as the pattern: the skin beside the doubler and the load that flows around a narrow repair are not modeled.

When it misleads

The strip is one-dimensional. A real doubler of finite width attracts load from the skin beside it, and the corner fasteners of the first row carry more than this model gives; a two-dimensional analysis or a test is needed for a repair that is not wide compared with its length. The single-shear joint bends as it is loaded (secondary bending), which adds a surface stress at the first row that the severity factor leaves out. At ultimate load the fasteners and holes yield and the row loads even out, so the elastic first-row load is conservative for static strength but is the right one for fatigue. The flexibility relations scatter widely. In compression only the buckling of the doubler between rows is checked: the doubler spanning a cutout, and the panel as a whole, can buckle first. Bearing allowables below e / D = 1.5 are outside the handbook's range.

Not covered

A second doubler, a filler in the cutout, a tapered or fingered doubler with different fasteners per row, staggered rows, bonded doublers, pressure bending of a fuselage skin, and the damage-tolerance evaluation (crack growth, residual strength and inspection interval) that a repair to transport-category primary structure requires. AC 43.13-1B applies only where the manufacturer's repair data do not, and to the aircraft it names.

Material values

The default skin and doubler are 2024-T3 bare sheet, MIL-HDBK-5J Table 3.2.3.0(b), A basis, longitudinal: Ftu 64 ksi, Fty 47 ksi, Fcy 39 ksi, Fbru 104 and 129 ksi and Fbry 73 and 88 ksi at e / D = 1.5 and 2.0, E 10.5 Msi. The default fastener allowable, 575 lb, is a 5/32 in 2117-T rivet at the 30,000 psi shear strength of AC 43.13-1B Table 4-8 on the nominal shank area. Swift's factors Ktbp = 3.2 and Ktbr = 1.4 are fixed in the second fatigue row. The default fastener modulus (10.3 Msi) and the Ramberg-Osgood shape parameter (15) were typed from memory and are illustrative: enter the values of your fastener and sheet.




References

Federal Aviation Administration (1998). Acceptable Methods, Techniques, and Practices: Aircraft Inspection and Repair. Advisory Circular AC 43.13-1B with Change 1 (2001), U.S. Department of Transportation, Washington, DC.

Tate, M. B., and Rosenfeld, S. J. (1946). Preliminary Investigation of the Loads Carried by Individual Bolts in Bolted Joints. NACA Technical Note 1051, National Advisory Committee for Aeronautics, Washington, D.C.

Huth, H. (1986). "Influence of Fastener Flexibility on the Prediction of Load Transfer and Fatigue Life for Multiple-Row Joints." In Fatigue in Mechanically Fastened Composite and Metallic Joints, ASTM STP 927, American Society for Testing and Materials, Philadelphia, pp. 221–250.

Swift, T. (1971). "Development of the Fail-Safe Design Features of the DC-10." In Damage Tolerance in Aircraft Structures, ASTM STP 486, American Society for Testing and Materials, Philadelphia, pp. 164–214.

Swift, T. (1990). Repairs to Damage Tolerant Aircraft. FAA-AIR-90-01, Federal Aviation Administration, U.S. Department of Transportation, Washington, DC.

de Rijck, J. J. M. (2005). Stress Analysis of Fatigue Cracks in Mechanically Fastened Joints: An Analytical and Experimental Investigation. PhD thesis, Delft University of Technology, Delft.

Crews, J. H., Hong, C. S., and Raju, I. S. (1981). Stress-Concentration Factors for Finite Orthotropic Laminates with a Pin-Loaded Hole. NASA Technical Paper 1862, Langley Research Center, Hampton, VA.

Howland, R. C. J. (1930). "On the Stresses in the Neighbourhood of a Circular Hole in a Strip under Tension." Philosophical Transactions of the Royal Society of London, Series A, Vol. 229, pp. 49–86.

Jarfall, L. E. (1967). "Optimum Design of Joints: The Stress Severity Factor Concept." In Aircraft Fatigue: Proceedings of the 5th ICAF Symposium, Melbourne.

Gerard, G. (1957). Handbook of Structural Stability, Part V: Compressive Strength of Flat Stiffened Panels. NACA Technical Note 3785, National Advisory Committee for Aeronautics, Washington, DC.

U.S. Department of Defense (2003). Metallic Materials and Elements for Aerospace Vehicle Structures. MIL-HDBK-5J, 31 January 2003.

Worked Examples

Published examples this tool reproduces; each page lists the inputs to enter and the printed answers.

All worked examples


DISCLAIMER: The methods and calculation on aNOWysis.com are provided as means to educate the reader. There is no warranty, expessed or implied, as to the relevance or accuracy of the methods or calculations. All calculations should be checked by a qualified analyst for appropriateness and accuracy. We will not be held liable for any damages, injuries or deaths caused by using methods or calculations from this site.