Przemyslaw Olszewski

2024-10

Comparison study: hand calc predictions vs detailed ANSYS ACP FEM for 12 test configurations. Documented where hand calc methods are conservative/unconservative and by how much.

Challenge: Setting up the comparison required consistent definitions of 'same loading' between hand calc and FEM — different boundary condition assumptions were the main source of discrepancy, not the structural analysis methods themselves.

Learned: Hand calc vs FEM discrepancies are often due to boundary condition differences, not formula errors. Documenting the boundary condition assumptions explicitly in each hand calc is essential for valid comparison and later reuse.

2024-04

First-ply failure screening tool in Excel using CLT + Tsai-Wu criterion — engineers enter layup, load resultants (N, M), and environmental conditions; tool outputs failure index for each ply and identifies critical ply.

Challenge: The tool needed to handle non-standard fiber orientations (not just 0/45/90) from optimization results. Implementing the general CLT rotation transformation in Excel required careful array formula use — VBA was more maintainable for this complexity.

Learned: Excel VBA provides much better maintainability than complex array formulas for CLT implementations. The structured programming approach catches errors that complex nested array formulas hide.

2023-11

Closed-form solutions for composite plate deflection and stress under uniform pressure — Timoshenko plate theory with orthotropic bending stiffness matrix (D matrix). Compared with ANSYS ACP results for validation.

Challenge: Timoshenko plate theory for composite plates assumes constant D matrix — not valid for plates with non-uniform stiffness due to ply drops. The closed-form solution agreed within 8% for uniform laminates but diverged significantly for tapered laminates.

Learned: Closed-form composite plate solutions are valid only for laminates with constant thickness and stacking sequence. For tapered or stepped laminates, FEM is necessary. Knowing the limits of analytical methods is as important as knowing the methods themselves.

2023-07

Core shear failure and face wrinkling analysis for acoustic liner sandwich panel — CFRP face sheets, Nomex honeycomb core, under acoustic pressure loading and concentrated point loads at attachment fittings.

Challenge: Face wrinkling load varies significantly with skin-to-core adhesive bond quality — the wrinkling formula assumes perfect adhesion. For degraded adhesive (post-service), effective wrinkling load can drop by 30-50%. Applied environmental knockdown factors from test data.

Learned: Sandwich panel wrinkling is sensitive to adhesive bond quality and environmental degradation. Standard wrinkling formulas for pristine bonds are non-conservative for service-aged panels. Test data at end-of-life moisture and temperature states is necessary.

2023-02

Margin of safety spreadsheets for composite sandwich panel under combined in-plane compression, shear, and bending. Failure modes: fibre compression, fibre tension, matrix cracking, core shear, face wrinkling, intracell buckling.

Challenge: Intracell buckling formula in the existing spreadsheet had an error in the cell size term — it was using cell diameter instead of cell apothem for hexagonal honeycomb. MOS was non-conservative by 18%. Caught during cross-check with Hexcel data sheets.

Learned: Honeycomb core cell geometry definitions vary between references — some use cell diameter, some use cell width (apothem), some use wall length. Always trace the formula to its original source and verify the geometric definition used.