Technology
Robotic welding sits where geometry, motion, sensing and metallurgy meet. Our technology work is organised around the decisions a welding cell has to get right.
Programs generated from engineering data, not taught point by point.
The problem
Teaching or manually programming every new part does not scale to high-mix production, and the people who can do it are scarce.
Our approach
We generate robot programs from CAD assemblies and process rules: joint detection, trajectory planning, welding procedures and controller output in one continuous chain, with engineering review at the points where judgement matters.
Reachable, collision-free motion for robots and external axes.
The problem
A weld seam is a line on a part. Turning it into motion means respecting robot kinematics, torch orientation, external axes, cell obstacles and weld sequence at the same time.
Our approach
Our planning technology, neXXus®, computes feasible trajectories under geometric, kinematic and operational constraints, and coordinates the robot with positioners and linear tracks.
Measuring the real part, inside the welding cell.
The problem
Real parts deviate from their models. Without measurement, offline programs inherit every tolerance of cutting, fit-up and fixturing.
Our approach
We use robot-mounted 3D vision and tactile sensing to measure parts in the cell and to express the result as corrections that a welding program can apply.
From point clouds to meshes to engineering CAD.
The problem
Many parts and structures have no usable 3D model, which blocks automatic programming entirely.
Our approach
EagleEye® reconstruction fuses multi-view 3D data into coherent geometry and interprets it as engineering CAD models exported as STEP to the Lumen® pipeline.
A cell model that matches the physical cell.
The problem
Offline programs are verified against a model of the cell. If the model is wrong, simulation gives false confidence.
Our approach
We build and maintain cell models for simulation and planning, calibrate them against the physical installation and reconstruct real scenes where models are missing.
Parameters, pass plans and procedures encoded in software.
The problem
Welding quality depends on parameters and pass layouts that are often held in the heads of experienced welders and engineers.
Our approach
We encode welding procedures, multipass scheduling and parameter recommendation into software, and use recorded production data to improve them.
Measuring the process and adapting while welding.
The problem
Heavy joints change as they are welded. A fixed program cannot follow a groove that moves and fills pass after pass.
Our approach
We use laser profile sensing to measure the joint during welding and adapt torch position with strategies operators can select and understand.
Geometry measured, results recorded, quality demonstrable.
The problem
Quality that is not measured and recorded cannot be demonstrated to customers, auditors or certification bodies.
Our approach
We reconstruct and measure weld geometry, compare it with defined tolerances and record process and inspection data per part, per seam and per pass.
Software that can be commissioned, diagnosed and supported.
The problem
Technology that works in the lab but cannot be deployed, diagnosed and maintained on a factory floor does not create value.
Our approach
We engineer for deployment: controlled releases, diagnostics, audit trails, backup and restore, role-based access and interfaces designed for the people who run the cell.
Next step
Tell us about the parts you weld, the cell you run and the problem you want to solve.
