An Automated Data-Driven Framework for Pre-Assembly Quality Control and Component Matching in Automotive Manufacturing

In automotive manufacturing, component-level variations can propagate through assembly processes and cause quality defects detected during end-of-line inspection. This is important for backlit interior components, where optical transmittance variations in silicone pads can lead to unacceptable luminance non-uniformity after assembly. This study presents an automated, data-driven pre-assembly quality control framework that identifies optical variations before assembly and matches components to achieve target luminance performance. A PLC-based measurement, classification, matching, and sorting system was integrated into a production line for steering-wheel button assemblies. The system combines vision-based product verification, a FEASA low-light LED analyzer, SQL-based data management, and a SCARA robot-assisted sorting station. Each silicone pad was evaluated using a nominal reference printed circuit board assembly (PCBA). Luminance homogeneity was assessed according to the difference between the maximum and minimum luminance values. For compliant parts, predefined positive and negative offsets were mathematically applied to generate alternative luminance scenarios. The optimum offset was determined by selecting the scenario that provided the closest luminance value to the nominal target, and the corresponding offset category was used to assign each silicone pad to the appropriate PCBA variant. The framework was validated using data from 64,535 silicone pads covering BA, CA, and DA product versions within a serial production line. Compared with the pre-intervention rate of 2.01%, the luminance-related defect rate decreased to 0.27%, corresponding to an 86.6% relative reduction, while the production defect rate decreased from 5.46% to 2.87%. Optimum offset matching enabled recovery of parts that would have been rejected. The SQL-based database provided traceability of measurement results, offset distributions, mold-cavity performance, and production statistics. The proposed framework shifts luminance inspection from end-of-line defect detection to automated pre-assembly process control, thereby improving production efficiency and reducing material waste in the investigated production process. Further validation on other automotive components and production lines is required to assess its broader applicability.