From the Experience 2026 workshop: gearbox thermal chains, DoE on materials, automatic post-processing and Blender renders, all driven from a script. Plus the SDK route to extend the solver itself.
Six things Particleworks can do when it is scripted
All shown on real runs from the workshop, none needs custom training data.
01
Gearbox, end to end
Churning run, heat-transfer map and steady thermal model chained in one command. Every step is gated before it spends GPU time.
9.2 h of GPU condensed into a 20 s thermal solve
02
DoE on materials
16 fluids across the rheology map, Newtonian to Bingham, shear-thinning and shear-thickening, against 9 obstacle shapes.
144 cases from one study definition
03
Post-processing on autopilot
Wetting maps, peak forces, torque, HTC and the report, produced the same way for every run.
A plausibility gate rejects bad fields before you see them
04
Blender in the loop
Particles to surface to cinematic frames. Colour and translucency can follow each particle's own history.
Beauty renders without leaving the pipeline
05
Complex workflows
The omelette: pour, swirl, cooking history integrated per particle, then compared against the no-swirl case.
Swirl cuts thickness variation 0.31 → 0.22
06
Beyond the GUI
Scripted motions, custom viscosity laws, and the SDK to hook into the solver and add physics that did not exist.
Viscosity as a function of temperature, per particle
Agenda · 45 min
Five blocks, one live Q&A
15:007 min
From script to agent
Four levels of control: your own Python, AI chat with the API in context, agents that run the machine, and the SDK.
15:0710 min
Gearbox, end to end
Churning loss, HTC map, steady CHT. What was automated, what was machine-checked, what the temperatures really mean.
15:179 min
DoE on materials, plus post-processing
16 fluids × 9 shapes, wetting and force maps, and the Blender renders generated from the same script.
15:269 min
Complex workflows
The omelette: temperature-driven viscosity, per-particle history, and what it takes to prepare such a case.
15:355 min
Engineering insights
The checks that separate a plausible result from a correct one.
15:405 min
Live Q&A
Bring your own case.
An omelette, simulated and rendered end to end from a script. Viscosity follows temperature.A study array launched, monitored and compared from one script.
Engineering insights
Three numbers from the gearbox report
2.1×
ISO over-predicts the churning loss
118 W by ISO/TR 14179-2 against 55 W simulated. ISO has no term for gears fed by thrown oil, which is most of this box.
3–6×
Gears cool harder than the housing
Mean HTC 214 W/m²K on a gear against 36 on the housing. Relative velocity drives it, not how wet the surface is.
0
HTC in an adiabatic run
Zero non-zero values over 670,920 triangles. The thermal map needed a real re-run, not a post-process.
Demonstration gearbox, Particleworks 8.2, 375 k particles, 2 s run. No customer data.
Free · live online
Thu 26 November, 15:00 CET. Bring a case.
45 minutes, presented by Michele Merelli, Particleworks Europe / EnginSoft. Prefer e-mail? info@particleworks-europe.com
