As a supplier of mechatronic solutions for the processing of steel sheets, Clecim has recently extended its range of laser welders by offering a new generation machine capable of cutting and welding heavy plates using a 12kW laser source. . In 2019, the laser cutting process was widely studied and the use of Machine Learning techniques made it possible to create a model capable of delivering robust cutting instructions for the entire thickness range. Today, attention is focused on the autogenous laser welding process (without filler wire). Obtaining a good weld requires fulfilling two criteria: having a weld bead free from defects (types of splashes, runs, under-filling, holes, etc.), and having good resistance (this is evaluated via a stamping test Erichsen type, the objective being to be as close as possible to the basic resistance of the material). To achieve this result, many steps were put in place, each requiring the use of a JMP platform:
- "Fit X by Y" and "Distribution" not only to statistically determine the value of the basic resistance of the material, but also to ensure that the studied sheets can be considered as identical and homogeneous
- "Graph Builder" to synthesize the visual observations and obtain a mapping of the welding defects according to the studied parameters, the result allowing to determine an area where the weld bead is free of defects
- And finally, "DoE / Custom Design" and 2 Fit Model "which allow to study and model on the previous zone the resistance of the weld bead in order to determine the optimal parameters. As the study area is irregular in shape, it is entered in the form of a set of candidate points (Covariates)
