Case Study: Continental Corporation Automotive Group achieves automated solder‑joint quality inspection and cuts classification errors to 79 ppm with STATISTICA Data Miner

Continental Corporation Automotive Group - Customer Case Study

Continental Automotive, a global supplier of vehicle powertrain systems, produces millions of piezo-actuator solder connections each year and must test 100% of production for durability-critical parts. Because some quality defects (voids, weld flow, etc.) cannot be reliably assessed by electrical or geometric measurements, the company needed an automated process to replicate manual visual inspection—which was labor intensive and carried a ~2% error rate.

Using STATISTICA Data Miner, Continental converted solder-joint images into numeric matrices and trained Support Vector Machine classifiers (with cross-validation and parameter tuning) to distinguish good from bad joints. The model was integrated on the production line, continually updated with new error patterns, and has delivered significant labor savings while reducing incorrect classifications to about 79 ppm; flagged bad parts are routed to manual reinspection for safety.

Continental Corporation Automotive Group

Dr. Udo Kreißig

Technology Development Manager