Numerical weld pool simulation for the accuracy improvement of inline weld depth measurement based on optical coherence tomography

Abstract

The aluminum alloy AA6082 is often used in high-voltage storage systems due to its favorable electrical and mechanical properties. Laser beam welding is a flexible process for producing the required welds. To protect the sensitive components of the battery cells, the required weld depth must be maintained in order to avoid destruction of the components and to reduce the number of defective parts. Optical coherence tomography (OCT) is a promising method for the inline monitoring of the deep welding process, since the depth of the vapor capillary can be measured without the influence of process emissions such as process light or welding fumes. Due to the robustness of the measurement method, the measured depth of the keyhole can be used as an input variable for a process control. However, there is a geometric difference between the optically detectable keyhole depth and the melt pool depth. This depth is additionally determined by fluid dynamic processes in the melt pool. In order to consider the difference between the melt pool depth and the keyhole depth when evaluating OCT measurement signals, a thermal simulation model was developed. By iteratively adapting the keyhole geometry in the simulation on the basis of genetic algorithms, an optimization of the correlation was achieved. The model was calibrated based on the comparison of metallographic cross sections with the calculated melt pool geometries. The results were used to provide a database for improving the performance of the signal processing algorithms. Based on the calculation results, it was possible to determine and quantify a dependence of the melt film thickness below the keyhole on the welding process parameters. By including the melt thickness below the keyhole, it was possible to increase the accuracy in the interpretation of OCT data with regard to the welding depth. The quantity can thus be used as an input quantity for signal processing algorithms for OCT data evaluation.

A Novel Approach to the Holistic 3D Characterization of Weld Seams—Paving the Way for Deep Learning-Based Process Monitoring

“…The procedure and the potentials are illustrated in Figure 10 using the example of the holistic utilization of OCT keyhole depth measurements. With the knowledge about the weld properties at each point of a measurement signal, the signals of the sensor can be characterized precisely and used for the further development of welding process models [26]. In addition, ray-tracing approaches that correlate the keyhole geometry with measurement signals can be validated [27].…”

“…It can be clearly seen that the area of the OCT signal with the highest density in depth direction was below the real weld depth. This behavior is due to a melt layer surrounding the vapor capillary in deep penetration laser beam welding [26]. In addition to evaluating the capillary depth, the OCT signal can be utilized to assess correlations between the course of the weld depth and the process stability.…”

Inline Weld Depth Evaluation and Control Based on OCT Keyhole Depth Measurement and Fuzzy Control

“…In the case of a different welding optics setup in further research experiments, the same FFNN can be used for interpreting the features. In previous studies, Schmoeller et al [30] evaluated the thickness of the process-parameter-dependent melt layer below the keyhole, which results in the difference between the measured keyhole depth and the depth obtained from metallographic cross-sections. The thickness of the melt layer obtained from a numerical weld pool simulation served as an additional offset for the estimated weld depth from the FFNN.…”

A Novel Approach to the Holistic 3D Characterization of Weld Seams – Paving the Way for Deep Learning-Based Process Monitoring

“…It can be clearly seen that the area of the OCT signal with the highest density in depth direction was below the real weld depth. This behavior is due to a melt layer surrounding the vapor capillary in deep penetration laser beam welding [25]. In addition to the capillary depth, correlations between the weld depth and the process stability could be obtained from the OCT signal.…”