Tools › Structures › Thin-Walled Structures › FEA Batch Margins


FEA Batch Margins

Paste the element forces of a finite element run and get one margin-of-safety table for every panel or fastener and every load case, sorted with the lowest margin first, with the governing load case of each element. Panels: membrane forces Nx, Ny, Nxy of shell elements checked for compression buckling in both directions, shear buckling and their interaction with plasticity (NACA TN 3781 coefficients, Stowell's plasticity factors, the exact biaxial expression for a simply supported plate, Rc + Rs² = 1 with shear), for tension and shear strength at ultimate load and for yield at limit load. Fasteners: the three forces of bush (spring) elements resolved into axial and resultant shear and checked for shear, tension, sheet bearing and the combined interaction Rsa + Rtb = 1. The paste box reads plain rows (ID, case and three numbers) and the QUAD4, TRIA3 and CBUSH force blocks of a Nastran .f06 file with their SUBCASE headings. The default numbers are an illustrative set made for this page.

  • Paste plain rows (ID, case, Nx, Ny, Nxy or ID, case, F1, F2, F3) or the FORCES IN QUADRILATERAL ELEMENTS (QUAD4), FORCES IN TRIANGULAR ELEMENTS (TRIA3) and FORCES IN BUSH ELEMENTS (CBUSH) blocks of a Nastran .f06 file; up to 5,000 rows; properties by element ID range

  • Panels: uniaxial and biaxial compression buckling, shear buckling and their interaction, simply supported or clamped edges, plasticity by Stowell's factors on a Ramberg-Osgood curve; principal tension against Ftu and maximum shear against Fsu at ultimate load; von Mises yield at limit load

  • Fasteners: axial and resultant shear from the three bush forces; shear, tension and sheet-bearing margins and the interaction Rsa + Rtb = 1 solved for the proportional load factor; fitting factor

  • One table sorted by lowest margin with a row limit and a negative-margins-only option, the envelope of each element across its load cases, a chart of the lowest margin by element, and the worked check of the governing element and case

FEA Batch Margins


Unit System: /

Analysis and Element Forces





Panel Properties by Element ID Range

1
2

Panel Material, Edges and Options












Fastener Groups by Element ID Range

1
2

Fastener Axes and Interaction





Load Level and Factors of Safety




Output



×

FEA Batch Margins

FEA Batch Margins: Method, Recognized .f06 Layouts, Assumptions and Limits

What the tool does

A finite element run prints element forces for thousands of elements and many load cases; the stress report needs one margin of safety for each structural item and the load case that governs it. This tool takes the pasted force output, applies the entered properties by element ID range, checks every element and case with hand-analysis methods, sorts the result with the lowest margin first and prints the envelope of each element across its load cases. Two kinds of elements are read: shell elements, checked as flat panels with their membrane forces, and bush (spring) elements, checked as single fasteners. The numbers of the governing row are then written out step by step so that the method can be followed and one row can be checked by hand.

What can be pasted

Plain rows. One element and load case per line: an integer element ID, a load case label, and three numbers, separated by commas, tabs or spaces. A line whose fields are not numbers is taken as a header and skipped; a semicolon ends a row like a line end. With four fields (ID and three numbers) the load case is 1. For panels the three numbers are Nx, Ny and Nxy, forces per unit length with tension positive; for fasteners they are the three forces of the element in its own axes.

Nastran .f06 pages. The tool reads the following blocks, in the layouts printed by MSC Nastran and NX Nastran sample output files (the printed layouts were taken from three sample files of the open-source pyNastran project; the order and meaning of the columns were checked against the MSC.Nastran 2004 Quick Reference Guide):

  • F O R C E S  I N  Q U A D R I L A T E R A L  E L E M E N T S  ( Q U A D 4 ), also QUAD8 and QUADR, and F O R C E S  I N  T R I A N G U L A R  E L E M E N T S  ( T R I A 3 ), also TRIA6 and TRIAR. Centroid layout: element ID followed by FX, FY, FXY, MX, MY, MXY, QX, QY. Corner layout (OPTION = BILIN and the higher-order elements): only the rows marked CEN/4, CEN/3, CEN/8 or CEN/6 are read; the grid-point rows under them are ignored. FX, FY and FXY are used; the moments and transverse shears are not.
  • F O R C E S  I N  B U S H  E L E M E N T S  ( C B U S H ): element ID followed by FORCE-X, FORCE-Y, FORCE-Z and three moments. The three forces are used.

What the Nastran columns mean, and what to check in the model. The statements below were checked against the MSC.Nastran 2004 Quick Reference Guide (element force item codes; the CQUAD4, CTRIA3 and CBUSH entries; the FORCE command) and the MSC.Nastran 2004 Reference Manual (shell element forces).

  • FX, FY and FXY are the membrane forces on the x and y faces and the in-plane shear force, each per unit length of the element, with tension positive. The tool therefore treats a negative FX or FY as compression. The next five numbers of a row are the two bending moments, the twisting moment and the two transverse shears, also per unit length; they are not used.
  • The forces are in each element's own coordinate system, not in the basic system and not in a common panel system. The x axis of a four-node quadrilateral bisects the angle between its two diagonals, which makes it parallel to the side from the first to the second grid point only when the element is a rectangle; the x axis of a triangle runs along the side from its first to its second grid point. Two neighboring elements numbered in a different order have x axes that differ by 90 or 180 degrees, and Nx and Ny then change places. Before pasting, confirm that the element x axis of every element in a property row lies along panel side a, or renumber or re-orient the elements.
  • A material orientation angle or material coordinate system on the element does not turn the force output by itself. It does when the run asks for output in the material coordinate system (a parameter of the program); the page of the .f06 then carries the line RESULTS IN MATERIAL COORD SYSTEM, the tool says so in a note, and x is the material x axis. That is the convenient way to give every element of a panel the same axes.
  • With center output each element has one row. With corner output the row marked CEN/ holds the center values and the rows under it hold the values at the grid points; only the center row is read, so a corner peak is not seen.
  • A bush element between two separate grid points and without a coordinate system of its own has its x axis along the line from the first grid point to the second, and reports elongation as a positive x force: FORCE-X is then the axial force and positive is tension, provided the two grid points lie along the fastener. With a coordinate system named on the element, which is required when the two grid points coincide, the element axes are the axes of that system and the force is the stiffness times the displacement of the second grid point less that of the first; the axial component is the axis that lies along the fastener and the sign of tension depends on the order of the grid points. Choose the axis and the sign on the form and confirm them on one fastener.

The load case is the number in the last SUBCASE heading above the block (1 when the file has none). Any other block of the file (displacements, stresses, strains, bar, beam and rod forces) is skipped; other force blocks are named in a note. A page heading ends a block, and the block title that Nastran repeats on the next page starts it again. Rows of an element ID and load case already read are ignored and counted. Complex, random and transient output, composite ply output and the punch and OP2 formats are not read. Other programs' output can be pasted as plain rows.

The pasted numbers carry no unit: choose the unit they were printed in (lb/in and lb, N/mm and N, N/m and N, or kN/m and kN). The limit is 5,000 rows per run.

Panels

Each property row gives a rectangular flat panel: side a along the element x axis (the direction of Nx), side b along the element y axis, thickness t, with all four edges simply supported or all four clamped. The panel is the bay between stiffeners, frames or ribs, not the element; every element in the row's ID range is checked as if its membrane forces acted uniformly on the whole panel. Alternatively the forces of all elements of a property row are averaged for each load case and the row is checked once as one panel.

Buckling of each load alone. Fcr = η k π² E / (12 (1 − ν²)) (t / width)², from the Local Buckling Coefficients tool of this site: compression along x with the width b, compression along y with the width a, and shear with the shorter side. For simply supported edges the compression coefficient is the exact value minimized over the number of half-waves and the shear coefficient comes from a converged energy solution; for clamped edges the coefficients are those of NACA TN 3781. The plasticity factor η is Stowell's (NACA Report 898; TN 3781 Table 2) with the secant and tangent moduli of a Ramberg-Osgood curve through Fcy, or, as an option, the elastic value cut off at Fcy.

Compression in both directions, simply supported. The plate buckles into m by n half-waves when Nx (m/a)² + Ny (n/b)² = π² D [(m/a)² + (n/b)²]² with compression positive and D = E t³ / (12 (1 − ν²)). The tool finds the smallest proportional factor on Nx and Ny over all m and n. The expression is exact for simply supported edges and it credits a tension in one direction against a compression in the other. Its reciprocal is the direct-stress ratio Rb.

Compression in both directions, clamped. No closed form exists. The tool adds the two uniaxial ratios, Rb = Rcx + Rcy. The set of load combinations that a plate carries without buckling is convex, so the straight line between the two uniaxial buckling points lies inside it and the rule is on the safe side; TN 3781 gives the same line (its equation (54)) as exact for plates that buckle in square waves. A tension is not credited.

With shear. The proportional factor λ is the root of λ Rb + (λ Rs)² = 1, the parabola Rc + Rs² = 1 of TN 3781 (equation (55) and Table 5) with the direct-stress ratio in place of the compression ratio. TN 3781 states the parabola for long plates in longitudinal compression (or tension) and shear and reports that it also fits finite plates; for transverse compression with shear on long plates it is conservative. For biaxial compression with shear it is a rule of this tool: in 240 seeded cases with side ratios from 0.5 to 4 it stayed within about 1.5 percent above an energy (Ritz) solution of the combined loading and was otherwise below it. A tension that would stabilize the panel against shear is not credited, which can make the margin of a tension and shear case several times too low.

Plasticity under combined loads. The ratios above are elastic. The stress intensity of the applied stresses is σi = √(fx² − fx fy + fy² + 3 τ²). The intensity at buckling s solves s = η(s) λel σi, where η is the geometric blend of the compression factor and the shear factor Es/E with the weights λel Rb and (λel Rs)², which add to one. For one load alone this returns the single-load critical stress above; when both factors are Es/E it is Stowell's combined equation, TN 3781 equation (59). The blend is a construction of this tool, not a published formula. The margin is λ − 1 with the forces at ultimate load.

Strength. At ultimate load the largest principal stress is compared with Ftu and the largest in-plane shear stress with Fsu. At limit load times the yield factor the von Mises stress is compared with Fty, or with Fcy when the principal stress of larger magnitude is compressive.

Fasteners

One of the three bush axes is named as the fastener axis. The force along it is the axial force; the other two are combined into the resultant shear V = √(Fa² + Fb²). The sign of a bush force depends on the order of the element's two grid points and on its orientation, so the sign that means tension is chosen on the form, or either sign is treated as tension. The loads are multiplied by the ultimate factor (when limit forces are pasted) and by the fitting factor. Margins: shear against the entered shear allowable, tension against the entered tension allowable, sheet bearing against Fbru D t, and the interaction Rsa + Rtb = 1 with exponents 2 and 2, or 3 on shear and 2 on tension. The interaction margin is the proportional factor λ that brings (λ Rs)a + (λ Rt)b to 1, less one, so it is a true margin on load and not one minus the sum. Axial compression is taken to be carried by contact of the sheets and is not checked.

Load level and factors of safety

The form states whether the pasted forces are ultimate or limit. Buckling, tension, shear, bearing and fastener checks are strength checks at ultimate load: limit forces are multiplied by the ultimate factor, ultimate forces are used as pasted. The yield check is at limit load times the yield factor: limit forces are multiplied by the yield factor, ultimate forces are divided by the ultimate factor first. The margin boxes show the level of the allowable (ULT or LIM). Where the skin is allowed to buckle below ultimate load, buckling can be reported without entering the lowest margin.

What it assumes

Flat, rectangular, isotropic panels of constant thickness with uniform membrane forces and the same support on all four edges; element x along panel side a; membrane forces only. One material for all panels of a run. Linear static results from the finite element model, so that a load factor on the forces is a load factor on the stresses. A fastener allowable that already contains everything it should (shear planes, head type, sheet material, knock-downs). Bush elements whose three forces are in the fastener's own axes.

When it misleads

When the element axes are not aligned with the panel sides, Nx and Ny are applied to the wrong sides without warning: check the material or element coordinate system of the output. When a panel is modeled with several elements and the forces vary across it, the most loaded element taken as uniform over the bay is conservative and the average can be unconservative for a load that peaks at one edge. Element bending moments and pressure are ignored, so a load case that is mostly bending shows as lightly loaded. Curved panels buckle at higher stresses than the flat-panel values, and stiffened panels can fail by crippling or column buckling of the stiffeners, which this tool does not see. Edge support between simple and clamped is not offered; simple support is the safe choice. After buckling the forces in a real structure redistribute and the linear forces pasted here no longer apply. Fastener forces from a model depend on the bush stiffnesses and on the mesh around the fastener; peaks at the first fastener of a row are sensitive to both. Prying, fastener bending, shim and gap effects and fatigue are not considered. The default allowables are examples, not design values.

Checks behind the tool

The buckling coefficients and plasticity factors are those of the Local Buckling Coefficients tool, itself checked against NACA TN 3781. The biaxial factor, the interaction roots, the plasticity solution and every margin of the default case are compared with an independent numpy and scipy computation (a double-sine energy solution for shear and for combined loading, a search over the half-wave numbers, root finding for the plasticity and interaction equations). The .f06 reader is tested on three sample output files printed by MSC Nastran and NX Nastran. The clamped-edge coefficients and the combined-load plasticity blend have no independent published check in this tool.




References

Gerard, G., and Becker, H. (1957). Handbook of Structural Stability. Part I: Buckling of Flat Plates. NACA Technical Note 3781, National Advisory Committee for Aeronautics, Washington, DC.

Stowell, E. Z. (1948). A Unified Theory of Plastic Buckling of Columns and Plates. NACA Report 898, National Advisory Committee for Aeronautics, Washington, DC.

Stowell, E. Z. (1949). Plastic Buckling of a Long Flat Plate Under Combined Shear and Longitudinal Compression. NACA Technical Note 1990, National Advisory Committee for Aeronautics, Washington, DC.

Stowell, E. Z., and Schwartz, E. B. (1943). Critical Stress for an Infinitely Long Flat Plate With Elastically Restrained Edges Under Combined Shear and Direct Stress. NACA Wartime Report L-340, National Advisory Committee for Aeronautics, Washington, DC.

Batdorf, S. B., and Stein, M. (1947). Critical Combinations of Shear and Direct Stress for Simply Supported Rectangular Flat Plates. NACA Technical Note 1223, National Advisory Committee for Aeronautics, Washington, DC.

Libove, C., and Stein, M. (1946). Charts for Critical Combinations of Longitudinal and Transverse Direct Stress for Flat Rectangular Plates. NACA Wartime Report L-224, National Advisory Committee for Aeronautics, Washington, DC.

Timoshenko, S. P., and Gere, J. M. (1961). Theory of Elastic Stability, 2nd ed. McGraw-Hill, New York.

Bruhn, E. F. (1973). Analysis and Design of Flight Vehicle Structures. Jacobs Publishing, Carmel, IN.

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

MSC.Software Corporation (2003). MSC.Nastran 2004 Quick Reference Guide. MSC.Software Corporation, Santa Ana, CA.

MSC.Software Corporation (2003). MSC.Nastran 2004 Reference Manual. MSC.Software Corporation, Santa Ana, CA.

Doyle, S. (2026). pyNastran: A Python-Based Interface Tool for Nastran's File Formats. Open-source software.

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.