Solver strategy
FatDyn-X works from an inertia-relief / static run (SOL 101) or from a normal-modes run (SOL 103). Inertia-relief / static remains important for stiff structures and when the loads are preliminary. When dynamics matter, a single normal-modes run gives the quasi-static and the modal-transient solutions, without the FORCE, TLOAD and TABLED decks of SOL 112.
The established route: unit load cases from the solver, combined with the load histories. FatDyn-X adds node, element-force and mass-force histories to the fatigue result.
The quasi-static response is assembled from the modes. Loads are applied inside FatDyn-X, straight from the load-channel files.
The same modes solved as a transient, with damping per mode, accurate from t = 0 with no start-up oscillation.
A duty cycle that mixes quasi-static and dynamic events uses one set of modal stresses throughout.
| Solver | Solver result used | Damage per repeat |
|---|---|---|
| Inertia relief | Static inertia-relief solve, one subcase per load channel | 1.08E-06 |
| Inertia relief via modes | The normal-modes file only | 1.08E-06 |
Unified
Fatigue and dynamic results come from one input file, one executable and one solver result. There are no separate tools to keep in step and no results to stitch together.
| Solver | What it shows | Min life, location A | Min life, location B |
|---|---|---|---|
| Inertia relief / static | Baseline | 1,945 | 2,169 |
| Inertia relief via modes | Same answer from the modes alone | 1,945 | 2,169 |
| Modal transient | Effect of the dynamics | 240 | 2 |
| Modal resonance | Diagnostic of dynamic amplification, mode by mode; not a life prediction by itself | ||
Screening
A durability program holds many kinds of events: rough road, braking, jumps, parking, curb strikes. Some are quasi-static and others excite structural modes. Running dynamics on everything is expensive and can mislead. FatDyn-X classifies every event and every mode.
Time histories and PSDs with natural-frequency overlays for every event and channel, and a summary that classifies each event.
Loads are projected to modal space, and the modal-load PSD tags each mode as modal-transient or inertia-relief via modes.
Modes with a low ratio are solved quasi-statically, which prevents unrealistic resonance from modes that only matter for NVH.
Insight
Which load channel and which mode drive the damage at an element or a weld? FatDyn-X ranks every channel and every mode, so design and test priorities follow the data.
For each element: damage from the channel alone and damage without the channel. Output is contribution, importance and change in damage per channel.
Mode-wise damage contribution per element. Shows the mode that drives the fatigue, and where damping or stiffening will pay.
Ratio of dynamic to static modal-coordinate range at the chosen damping. A high value means a static analysis underestimates that mode.
| Mode | Contribution % | Importance | Dynamic range | Static range | Amplification |
|---|---|---|---|---|---|
| 7 | 3.69 | 0.997 | 27.514 | 17.043 | 61.4 % |
| 8 | 0.00 | -0.003 | 5.368 | 4.285 | 25.3 % |
| 21 | 0.004 | 0.800 | 0.668 | 0.668 | 0.0 % |
Mode 7 is the dynamically dominant mode here; removing it drops the damage by 99.7 %. The same ranking is available per spot weld and per seam weld.
What's new
Generally limited to stress and fatigue analysis. The full dynamic information sitting inside the inertia-relief run goes unused.
The same inertia-relief / static result plus the load histories give a full suite of dynamic calculations, in addition to stress and fatigue.
Mass-element forces size the mounts of EV battery packs, motors, payloads and avionics. The acceleration contours combine the inertia-relief accelerations with a trusted mass, inertia and CG, for example from the crash model or corner weights.
Fatigue
Give a point by element id or by XYZ coordinates with a plane normal. FatDyn-X finds the host element, transforms the normal and reports the life on that plane next to the absolute-maximum-principal life. Useful for weld lines, build planes of printed parts and postulated crack paths.
Testing
Electronics, battery packs, motors and engine components are proven on shaker tests. FatDyn-X combines the PSDs of many field events into a single test PSD, so days of duty cycle are compressed into a test of a few hours.
Each event's acceleration or load history is converted to a PSD.
The event PSDs are combined, with their repeats, into a single PSD of the chosen test duration.
FatDyn-X runs the duty cycle of the event PSDs and the single test PSD, and compares the damage.
Joint design
Slip factor and margin of safety for every bolted joint, from the force histories of the CBAR / CBEAM bolt elements, with any of the solvers.
Applied on an EV program with about 1,000 bolted joints under long time-history loading.
Graphical front end
Studio writes the input deck as you work through eleven steps, checks it before the run, starts the solver and the weld and crack-growth tools, and plots the results. The deck text can still be edited directly.
Material and property ids, sets and mode frequencies come from the bulk data. Load channels are imported from CSV, and RSP files are converted.
Pre-flight checks list what is missing or inconsistent. Double-click a message to go to the step that needs fixing.
Tables and curve plots for S-N, ε-N, brittle, crack-growth and weld data, and a browser for the open material database.
Fits your process
PUNCH, OP2 and H3D results from Nastran and OptiStruct are read directly. Load channels come from CSV or RSP files, selected by column or by name.
Modal coordinates from one normal-modes run are written in PCH and CSV formats, for use in an established fatigue process.
When a conventional solver run is still wanted, FatDyn-X writes the static, inertia-relief and transient load cases from the load histories.
Results are written as CSV tables and as contour files for HyperView and other post-processors. Parallel processing has no per-core fees.