Cloud-based simulation of a laser welding system

Bollinger & Ohr is an expert for job shop laser welding, which among many other things produces heatsealing tools for packaging machines. The heatsealing tool is made of two parts of high-grade steel, which are welded together. For series production a robot is used, which is programmed to run with a cw (“continuous wave”) laser welding process. By suitable choice of the process parameters an improved weld connection can be found, that means an optimized welding seam depth and reduction of heat impact to minimize welding distortion. However, these studies are very cost-intensive and need lots of specimen.

Lauer & Weiss is a module expert for CAE analysis, among many other things for simulation of welding processes. Their experts have developed a CAE analysis method to optimize welding processes by varying parameters like welding seams depth, welding speed and order of the welding seams (“segmented welding”). Welding distortion can minimized by suitable choice of welding parameter, especially be using segmented welding.

Lauer & Weiss started this experiment to maximize the welding seam depth and minimize the heat impact into the work piece. Aim is to find the best welding process parameters, i.e. focus, welding speed, laser power, best order of weld seams. By using CAE simulation numerous samples can be saved, that (if produced) afterwards would have had to be quality inspected, which again is expensive and time-consuming. However, these analyses require huge computing powers, as the CAE model of the sample has enormous number of DOF. Besides that the material behaviour is non-linear. The standard in-house computing equipment is not sufficient for these simulations.

By using HPC-cloud based clusters you gain time by factors 10 – 20, as many CAE analyses can run in parallel. In addition to that the single runs are much faster and results are available sooner. This experiment will demonstrate the technical feasibility of the HPC-cloud approach for high-resolution CAE simulations replacing cost-intensive prototype and welding experiments. This will give Bollinger & Ohr a competitive advantage from cost and time savings.At Lauer & Weiss the proposed HPC-cloud-based solution will enable them to provide reliable results to their customers with low costs and short time, without purchasing a multiple CPU-cluster.

 

 

 

Project:
Fortissimo

Enterprises:
Bollinger & Ohr UG, Lauer & Weiss GmbH, SICOS BW GmbH, University of Stuttgart, HLRS, University of Stuttgart

Sector
Machinery & equipment

Keywords
High-Perfomance Computing

Read More