Inline Testing Method Based On Paschen’s Law With Integrated Machine Learning For Hairpin Stator Production

Due to the rapidly increasing global demand for electric vehicles, OEMs and suppliers in the automotive industry are facing new challenges. One of these challenges is the production of novel product technologies - such as the Hairpin stator technology - in large quantities in a time and cost-efficient manner. Compared to conventional stochastically wound stators, Hairpin technology uses solid copper bars and thus has great advantages on both the product and process side. However, due to the novelty of the Hairpin technology in the automotive sector, the necessary technical expertise to produce efficiently and to achieve the high quality standards in the automotive sector is currently still missing. A promising approach is to use suitable testing methods at the right points in the process chain to ensure product quality. One of the most common causes of defects in the production of Hairpin stators is damage to the insulation layer of the Hairpins. Pre-tests of a testing method based on Paschen's law have shown that these insulation faults can be detected by an optical-electrical test. In combination with a Machine Learning concept, the causes of the insulation damage can be quickly identified and assigned to the upstream process steps. In addition, the time and cost of the test scope is significantly lower compared to other testing methods. In summary, this paper explains the structure and application of the testing method in combination with the Machine Learning concept against the background of its use in series production.