Artur Czarnecki

2025-01

Systematic validation of hand calculation tool library against FEM benchmarks — 35 calculation types, 5 FEM verification cases each, documented accuracy envelope and applicability limits.

Challenge: Some hand calc tools in the library had been developed 8-10 years ago with no documented validation. Discovering errors in legacy tools required re-deriving the formula from first principles to confirm whether the implementation or the formula was wrong.

Learned: Legacy calculation tools accumulate undocumented assumptions and errors over time. Systematic validation with FEM benchmarks is the only reliable way to establish confidence in a tool library. Allocate at least 6 months for a comprehensive validation campaign.

2024-10

Stiffened skin panel compression analysis — Needham's method for crippling of stiffener sections, diagonal tension field theory for stiffened shear panels, interaction of crippling and shear buckling.

Challenge: Needham crippling method requires dividing the stiffener cross-section into flat sub-elements. Non-standard cross-sections (hat stiffener with variable flange width) required careful sub-element identification and weighting by area fraction.

Learned: Needham's method is sensitive to how the cross-section is divided into sub-elements. For non-symmetric or non-standard sections, document each sub-element explicitly with area and Fcc (crippling stress). Ambiguous decomposition leads to non-reproducible results.

2024-04

Lug and pin analysis per MIL-HDBK-5 for engine mount clevis fittings — bearing stress, net section tension, shear-out, and pin bending under combined axial, transverse, and oblique loading.

Challenge: MIL-HDBK-5 lug analysis assumes standard geometry (circular hole, straight lug sides). One lug had a slotted hole — not covered by the standard. Applied FEM-derived bearing stress concentration factor for the slotted hole geometry as a modification to the standard formula.

Learned: MIL-HDBK-5 lug analysis is well-validated for standard geometries but requires engineering judgment for non-standard features. Documenting the deviation from standard and the technical basis for any modifications is mandatory for certification.

2023-11

Thermal stress in bimetallic engine bracket — bonded aluminum-titanium section, closed-form solution for differential thermal expansion stress as function of temperature change. Compared against ANSYS FEM for validation.

Challenge: Closed-form bimetallic beam theory assumes perfect bond and no end effects. The FEM showed significant stress concentration at the bond edge (3x higher than the uniform mid-section stress) — the closed-form completely misses this. Added a stress concentration factor from FEM as a post-processing correction to the hand calc.

Learned: Bimetallic strip formulas give accurate mid-section stress but are wrong at edges and interfaces. For bonded dissimilar material structures, FEM is necessary to assess the bond edge stress where failure initiates.

2023-07

Plate buckling analysis under combined in-plane loading — Roark's formulas for rectangular plates under compression, shear, and combined loading. Buckling interaction equation (Rsc + Rs^2 = 1) for combined loads.

Challenge: Roark's tables for plate buckling cover specific boundary condition combinations. The actual boundary conditions (partially clamped along two edges, simply supported on others) were not directly tabulated. Used conservative simply-supported assumption and noted the conservatism in the calculation record.

Learned: When exact boundary conditions are not tabulated in Roark, the appropriate conservative assumption must be identified and documented. Simply supported gives lower buckling load than fixed — appropriate for conservative assessment. Never interpolate between tabulated BCs without understanding the trend.

2023-02

Shear flow analysis in thin-walled closed and open cross-sections for engine mount beam — Bredt-Batho theory for closed sections, direct shear flow for open sections. Comparison of torsional stiffness for candidate cross-section shapes.

Challenge: A multi-cell closed cross-section requires solving a system of equations for shear flow — the number of unknowns equals the number of cells. For a 3-cell cross-section, setting up the compatibility equations in Excel without errors required careful cell referencing.

Learned: Multi-cell closed section shear flow analysis is straightforward in concept but error-prone in spreadsheet implementation. Python implementation with matrix algebra is more robust and easier to validate than nested Excel formulas.