<![CDATA[The reliability of electric vehicle charging connectors depends on correct seal installation, since the seal protects the interface from dust and moisture. This study developed an IoT-enabled cyber-physical assembly system integrating fiber-optic and photoelectric sensors, a programmable logic controller, pneumatic actuation, a Raspberry Pi edge gateway, and a human-machine interface. The system was evaluated on an industrial production line using the same connector model, machine, operating procedure, and inspection criteria as the baseline. The baseline rejection rate, drawn from routine production records before implementation, was 19.18%. After calibration, 600 consecutive connectors were assembled during a continuous five-hour validation run, with manual inspection by experienced quality personnel serving as the reference standard. The post-implementation rejection rate fell to 2.80%, an 85.41% relative reduction (16.38 percentage points), yielding a production yield of 97.17%. The detection system achieved 99.50% accuracy, 88.89% precision, 94.12% recall, 99.66% specificity, and a 91.43% F1-score. Average cycle time was 30 seconds per unit, with a repeatability error of ±0.10 mm across 30 repeated press-depth measurements. These results indicate improved defect prevention, assembly consistency, and traceability under the evaluated conditions, with remaining defects mostly outside the seal-sensing scope.]]>
Copyrights © 2026
