Aprendizado de máquina para suporte à decisão do nível de abrangência de bloqueio de lotes na inspeção de saída de produtos eletrônicos
DOI:
https://doi.org/10.20397/2177-6652/2026.v26i3.3451Palabras clave:
Manufatura eletrônica; Inspeção de saída; Aprendizado de máquina; Qualidade preditiva; Desbalanceamento de classes.Resumen
Objetivo: Proponer un sistema predictivo basado en aprendizaje automático para apoyar la decisión sobre el alcance del bloqueo de lotes en la inspección de salida de un proceso de manufactura electrónica con líneas SMT, clasificando automáticamente el alcance de contención más adecuado para cada no conformidad detectada.
Metodología: Se siguió una adaptación de CRISP-DM, con comprensión del negocio, análisis de los datos, construcción de la variable objetivo mediante reglas operativas, ingeniería de atributos, modelado y evaluación. Se compararon Random Forest, XGBoost, LightGBM y SVM en dos enfoques: entrenamiento solo con datos reales y con aumento selectivo de la clase minoritaria mediante SMOTENC, usando registros históricos de una fábrica de la Zona Franca de Manaos.
Originalidad/Relevancia: El estudio aborda una brecha práctica poco explorada: la decisión sobre el alcance del bloqueo tras la detección del defecto, y no la detección del defecto en sí. Propone además una estrategia para derivar la variable objetivo cuando la decisión de contención aún no se registra en el sistema.
Principales resultados: Random Forest entrenado solo con datos reales obtuvo el mejor desempeño, con un F1-macro de 0,7937 en el conjunto de prueba. La aplicación de SMOTENC no generó una ganancia consistente en datos reales no vistos, indicando que el principal límite del sistema es el reducido volumen de registros con defecto.
Contribuciones teóricas: El trabajo transforma datos operativos ya existentes en apoyo predictivo a la contención de riesgos en el proceso SMT y demuestra la viabilidad de derivar la variable objetivo mediante reglas operativas, haciendo las decisiones de bloqueo más rastreables, consistentes y orientadas por datos.
Citas
AIAG, & VDA. (2019). Failure mode and effects analysis: FMEA handbook. Automotive Industry Action Group.
Akula, S. K., Salehfar, H., & Behzadirafi, S. (2022, October). Comparison of traditional and fuzzy failure mode and effects analysis for smart grid electrical distribution systems. In 2022 North American Power Symposium (NAPS) (pp. 1–6). IEEE.
ANSI/ASQ. (2008). Sampling procedures and tables for inspection by attributes (ANSI/ASQ Z1.4-2008). American Society for Quality.
Arevalo, F., Piolo, C. A. M., Ibrahim, T., & Schwung, A. (2022). Production assessment using a knowledge transfer framework and evidence theory. arXiv. https://doi.org/10.48550/arXiv.2207.03611
Batista, G. E. A. P. A., Prati, R. C., & Monard, M. C. (2004). A study of the behavior of several methods for balancing machine learning training data. ACM SIGKDD Explorations Newsletter, 6(1), 20–29.
Branco, P., Torgo, L., & Ribeiro, R. P. (2015). A survey of predictive modelling under imbalanced distributions. arXiv. https://doi.org/10.48550/arXiv.1505.01658
Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16, 321–357.
Chen, Z., Deng, S., Chen, X., Li, C., Sanchez, R. V., & Qin, H. (2017). Deep neural networks-based rolling bearing fault diagnosis. Microelectronics Reliability, 75, 327-333.
Cramér, H. (1946). Mathematical methods of statistics. Princeton University Press.
Drury, C. G., & Addison, J. L. (1973). An industrial study of the effects of feedback and fault density on inspection performance. Ergonomics, 16(2), 159–169.
Elkan, C. (2001). The foundations of cost-sensitive learning. In Proceedings of the 17th International Joint Conference on Artificial Intelligence (pp. 973–978).
Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861–874.
Fernández, A., García, S., Galar, M., Prati, R. C., Krawczyk, B., & Herrera, F. (2018). Learning from imbalanced data sets. Springer.
Ford Motor Company. (1987). Team oriented problem solving: 8D process. Ford Motor Company.
Gonzalez, L. F. A., & Yacout, S. (2025). Predictive modeling for quality prediction in multi-stage manufacturing systems using artificial intelligence. Machine Learning with Applications.
Gunraj, H., Guerrier, P., Fernandez, S., & Wong, A. (2022). SolderNet: Towards trustworthy visual inspection of solder joints in electronics manufacturing using explainable artificial intelligence. arXiv. https://doi.org/10.48550/arXiv.2211.10274
Hand, D. J., & Till, R. J. (2001). A simple generalisation of the area under the ROC curve for multiple class classification problems. Machine Learning, 45, 171–186.
Hopp, W. J., & Spearman, M. L. (2011). Factory physics (3rd ed.). Waveland Press.
IPC, & JEDEC. (2015). Moisture/reflow sensitivity classification for nonhermetic surface mount devices (IPC/JEDEC J-STD-020E). IPC.
ISO. (1999). Sampling procedures for inspection by attributes – Part 1: Sampling schemes indexed by acceptance quality limit for lot-by-lot inspection (ISO 2859-1:1999). International Organization for Standardization.
ISO. (2015a). Quality management systems – Fundamentals and vocabulary (ISO 9000:2015). International Organization for Standardization.
ISO. (2015b). Quality management systems – Requirements (ISO 9001:2015). International Organization for Standardization.
Japkowicz, N., & Shah, M. (2011). Evaluating learning algorithms: A classification perspective. Cambridge University Press.
Lobo, A., Sampaio, P., & Novais, P. (2024). Enhancing Quality 4.0 and reducing costs in lot-release process with machine learning-based complaint prediction. The TQM Journal.
Montgomery, D. C. (2016). Introdução ao controle estatístico da qualidade (7ª ed.). LTC.
Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company: How Japanese companies create the dynamics of innovation. Oxford University Press.
Powers, D. M. W. (2011). Evaluation: From precision, recall and F-measure to ROC, informedness, markedness and correlation. Journal of Machine Learning Technologies, 2(1), 37–63.
Prasad, R. P. (1997). Surface mount technology: Principles and practice (2nd ed.). Van Nostrand Reinhold.
Prasad-Rao, J., Heidary, R., & Williams, J. (2023). Detecting manufacturing defects in PCBs via data-centric machine learning on solder paste inspection features. arXiv. https://doi.org/10.48550/arXiv.2309.03113
Qin, W., Zhuang, Z., Guo, L., & Sun, Y. (2022). A hybrid multi-class imbalanced learning method for predicting the quality level of diesel engines. Journal of Manufacturing Systems.
Qiu, F., Wang, A., Wu, L., Nie, H., & Deng, B. (2026). A deep hybrid framework for quality prediction in complex manufacturing processes assisted by polynomial feature engineering. AIP Advances.
Rahman, M., Zhang, R., Gladstone, D. J., Williams, B. B., Chen, E., Dexter, C. A., Thompson, L., Bruza, P., & Pogue, B. W. (2021). Failure modes and effects analysis (FMEA) for experimental use of FLASH on a clinical accelerator. arXiv. https://doi.org/10.48550/arXiv.2112.10338
Sankhye, S., & Hu, G. (2020). Machine learning methods for quality prediction in production. Logistics.
Schwab, K. (2016). The fourth industrial revolution. World Economic Forum.
Slack, N., Brandon-Jones, A., & Johnston, R. (2018). Operations management (8th ed.). Pearson.
Sokolova, M., & Lapalme, G. (2009). A systematic analysis of performance measures for classification tasks. Information Processing & Management, 45(4), 427–437.
Tercan, H., & Meisen, T. (2022). Machine learning and deep learning based predictive quality in manufacturing: A systematic review. Journal of Intelligent Manufacturing, 33(7), 1879–1905.
Ulger, F., Yuksel, S. E., Yilmaz, A., & Gokcen, D. (2022). Fine-grained classification of solder joints with α-skew Jensen-Shannon divergence. arXiv. https://doi.org/10.48550/arXiv.2209.09857
Zhang, Q., Zhang, M., Gamanayake, C., Yuen, C., Geng, Z., Jayasekara, H., Zhang, X., Woo, C.-W., Low, J., & Liu, X. (2020). Deep learning based defect detection for solder joints on industrial X-ray circuit board images. arXiv. https://doi.org/10.48550/arXiv.2008.02604
Descargas
Publicado
Cómo citar
Número
Sección
Licencia
Derechos de autor 2026 Revista Gestão & Tecnologia
Esta obra está bajo una licencia internacional Creative Commons Atribución-NoComercial 4.0.
Os direitos, inclusive os de tradução, são reservados. É permitido citar parte de artigos sem autorização prévia desde que seja identificada a fonte. A reprodução total de artigos é proibida. Em caso de dúvidas, consulte o Editor.
