Industrial validation of a poka-yoke system to prevent defects in stamped metal components

Abstract

Defect prevention in stamped components requires mechanisms capable of acting directly at the source without substantially increasing process complexity. This study aimed to design, implement, and evaluate a mechanical Poka-Yoke device to prevent the release of parts missing a critical hole. An applied before-and-after design was employed, integrating production records, root-cause analysis, CAD development, CNC machining, and functional inspection. Baseline data showed 630 nonconforming parts out of approximately 35,000 components produced per month (1.8%). The available post-implementation monitoring period covered five months and 175,000 inspected parts, with no documented escapes of the targeted defect. With zero observed events, the one-sided 95% exact binomial upper confidence limit was approximately 0.0017% (17.1 ppm). Direct timing of 30 positioning, verification, and release cycles using the Poka-Yoke yielded a mean cycle time of 54.91 ± 2.79 s/cycle (95% CI: 53.87–55.95 s; range: 50.11–60.53 s).

The attribute-based R&R validation involved 200 reference parts (100 conforming and 100 with the critical hole intentionally absent), three operators, and two repetitions, resulting in 1,200 evaluations. The assessment yielded 591 true positives, 600 true negatives, nine false negatives, and no false positives, corresponding to an overall sensitivity of 98.50%, specificity of 100.00%, accuracy of 99.25%, false-negative rate of 1.50%, and false-positive rate of 0.00%. The device required an initial investment of approximately MXN 2,000, compared with an estimated annual cost of MXN 100,000 for additional manual inspection. Under the stated assumptions, the first-year simple return on investment (ROI) was 4,900%, with a payback period of 0.02 years.