Technology

Our Technology

Three physics engines, one idea: serious simulation that runs on the GPU in your own browser. OTOM AFP Web models laser-assisted fibre placement from robot path to crystallinity, OTOM CFD solves incompressible and supersonic flow in real time, and OTOM SPH follows fluids and polymer melts particle by particle.

Thank you for reading this post, don't forget to subscribe!

OTOMcomposite started with OTOM, our optical-thermal model for laser-assisted tape placement and winding. The same in-house physics now drives a family of web-native solvers, and we use them in our engineering services, on site or remotely.

What every OTOM engine shares

GPU compute with WebGPU

Solvers are written as WebGPU compute shaders, so a laptop graphics card does work that used to need a workstation, often in real time.

Nothing to install, data stays local

Open a link and start. The simulation runs on your own machine, so your geometry, G-code and process settings are not uploaded to a solver farm.

AI assistants built in

Local or cloud language models can set up, run and explain a simulation from a plain request. See AI agents.

OTOM AFP Web: laser-assisted fibre placement

Process digital twin

From robot path to crystallinity

AFP Web follows the whole chain of laser-assisted AFP: the machine moves, the laser heats, the tape bonds, and the polymer crystallises. Each step feeds the next.

  • Kinematics: loads the tool (STL/STEP) and the machine programme (G-code, XML, RAPID), projects courses onto the surface and builds the lay-up ply by ply.
  • Optics: GPU ray tracing of the laser bundle against tool, tape and roller, with Fresnel absorption and reflections.
  • Thermal: incoming tape (2D finite differences), placed tape (3D finite elements) and substrate, cumulative over the whole lay-up.
  • Crystallinity, optimisation and IR camera: crystallisation kinetics for PEEK, PEKK and PPS, laser power control to a target temperature, and direct comparison with thermal-camera recordings.
~6 sfull thermal run on the Funnel example
5×faster than the CPU solvers, same temperatures

OTOM CFD: real-time fluid dynamics

Grid-based solver

Low-speed flow to Mach 5, live

A grid-based Navier-Stokes solver that runs fast enough to interact with: move an obstacle, change the inlet or the angle of attack, and the flow responds immediately.

  • Incompressible engine: Chorin projection method with a pressure Poisson solve, Smagorinsky LES turbulence and vorticity confinement.
  • Compressible engine: finite-volume Euler solver with the HLLC Riemann solver, which resolves bow shocks, oblique shocks and expansion fans.
  • Geometry: obstacles drawn by hand or imported, handled by Brinkman penalisation. Grids up to 4096 cells across in 2D, plus full 3D with wings and Ahmed bodies.
  • Engineering output: live drag and lift coefficients, Reynolds number, probes, Schlieren, streamlines and two-fluid flow.

OTOM SPH: particle-based flow

Mesh-free solver

Free surfaces, mixing and melts

Smoothed-particle hydrodynamics needs no mesh, which suits free surfaces, splashing and mixing. It is also how we study molten thermoplastic flowing around and between fibre bundles during impregnation and consolidation.

  • GPU SPH: density, pressure and viscosity in WebGPU compute shaders, with prefix-sum spatial sorting for tens of thousands of particles at 60 FPS.
  • Turbulence and walls: SPS-LES, vorticity confinement and a log-law wall model.
  • CAD in the loop: swap in obstacles, containers and mixers as STL or 3MF files.
  • Flow insight: streamlines, pathlines and drag and lift forces on bodies.

Which engine for which question

EngineMethodTypical questions
OTOM AFP WebRay tracing + FDM/FEM thermal + crystallisation kineticsIs the nip-point temperature in the process window? Where does the laser shadow? What laser power profile gives uniform bonding?
OTOM CFDEulerian grid: projection method (incompressible) or HLLC (compressible)What are the drag and lift? Where does the flow separate? How should a greenhouse or a room be vented?
OTOM SPHLagrangian particles, mesh-freeHow does a melt or liquid mix, splash and fill? How does material discharge from a hopper? What happens at a free surface?

Challenges we help you resolve

  • Temperature localisation and process deviations
  • Insufficient or non-homogeneous bonding quality
  • Flow separation, drag and poor ventilation
  • Poor mixing, voids and incomplete wet-out
  • Long engineering lead times and costly trials
  • High final cost and reliability issues

Have a process to simulate?

We offer the software, tailored engineering studies and production-line assessments, remotely or at your site.