Low-Parameter Critic-Based Multivariate WGAN Model for Clogging Detection in Drives

Artur Dawid Surówka; Marcin Kosiba; Teemu Mikkelä; Asko Kavala; Marcin Firla

doi:10.2478/pead-2025-0008

Abstract

Efficient detection of anomalies in the cooling system of variable frequency drives (VFDs) is crucial to minimise downtime costs from overheating. Smart condition monitoring tools, especially those using machine/deep learning, have proven effective for failure detection. Recent research has focussed on environmental factors, such as pollution and humidity, affecting VFDs. Clogging is particularly harmful as it can damage power electronics, leading to extended downtimes. This study explores the use of Wasserstein Generative Adversarial Networks (WGANs) for detecting clogging in drives, including inlet/outlet/heatsink clogging and fan blockage. WGANs are adept at recognising complex temporal patterns due to their feedback-driven training. Despite generative AI models being typically large and unsuitable for embedded systems, this work demonstrates the feasibility of a low-parameter WGAN critic-based model for detecting cooling issues in VFDs. Using temperature signals, the model can detect clogging as low as 20%–30% with high performance metrics, achieving up to 90% accuracy and an F1 score above 0.9 for heatsink clogging detection, using a lightweight 26-parameter critic model. This study shows the potential for developing low-parameter WGAN critic-based models for clogging detection in VFDs.

References

Al-Naseem, O. A. and El-Sayed, M. A. (2013). Analysis of electrical and non-electrical causes of variable frequency drive failures. In 2013 9th IEEE International Symposium on Diagnostics for Electric Machines, Power Electronics and Drives (SDEMPED) (pp. 221–226). IEEE.
Search in Google Scholar Back to article
Avor, J. K. and Chang, C.-K. (2019). Reliability Analysis of Application of Variable Frequency Drive on Condensate Pump in Nuclear Power Plant. Journal of International Council on Electrical Engineering, 9(1), pp. 8–14. doi: 10.1080/22348972.2018.1564548
Open DOI Search in Google Scholar Back to article
Bear, J., Prügel-Bennett, A. and Hare, J. (2024). Rethinking Deep Thinking: Stable Learning of Algorithms using Lipschitz Constraints. Proeedings of 38th Conference on Neural Information Processing Systems (NeurIPS 2024), arXiv:2410.23451. https://doi.org/10.48550/arXiv.2410.23451
Search in Google Scholar Back to article
Beattie, A., Mulink, P., Sahoo, S., Christou, I. T., Kalalas, C., Gutierrez-Rojas, D., and Nardelli, P. H. (2022). A Robust and Explainable Data-Driven Anomaly Detection Approach for Power Electronics. In 2022 IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm) (pp. 296–301). IEEE.
Search in Google Scholar Back to article
Bokeh Development Team (BDT) (2018). Bokeh: Python library for interactive visualization. http://www.bokeh.pydata.org.
Search in Google Scholar Back to article
Chandola, V., Banerjee, A. and Kumar, V. (2009). Anomaly Detection: A Survey. ACM Computing Surveys (CSUR), 41(3), pp. 1–58. doi: 10.1145/1541880.1541882
Open DOI Search in Google Scholar Back to article
Chen, Y., Gao, Q. and Wang, X. (2022). Inferential Wasserstein Generative Adversarial Networks. Journal of the Royal Statistical Society Series B: Statistical Methodology, 84(1), pp. 83–113. doi: 10.1111/rssb.12476
Open DOI Search in Google Scholar Back to article
Chollet, F. (2018). Deep Learning with Python. New York, Manning Publications.
Search in Google Scholar Back to article
Ciappa, M. and Fichtner, W. (2000). Lifetime Prediction of IGBT Modules for Traction Applications. IEEE International Reliability Physics Symposium Proceedings, San Jose, CA, USA.
Search in Google Scholar Back to article
Dai, H., Wang, J., Zhong, Q., Chen, T., Liu, H., Zhang, X. and Lu, R. (2024). A GAN-Based Anomaly Detector Using Multi-Feature Fusion and Selection. Scientific Reports, 14, p. 52378. https://doi.org/10.1038/s41598-024-52378-9
Search in Google Scholar Back to article
Flach, P. and Kull, M. (2015). Precision-recall-gain curves: PR analysis done right. In: Proceedings of Advances in Neural Information Processing Systems 28 (NIPS 2015), Montreal, Quebec, Canada.
Search in Google Scholar Back to article
Gómez, P. I., López, G. M. E., Mijatovic, N. and Dragičević, T. (2024). A Self-Commissioning Edge Computing Method for Data-Driven Anomaly Detection in Power Electronic Systems. IEEE Transactions on Industrial Electronics, vol. 71, no. 10, pp. 13319–13330. doi: 10.1109/TIE.2023.3347839
Open DOI Search in Google Scholar Back to article
Goodfellow, I., Bengio, Y. and Courville, A. (2016). Deep Learning. Cambridge, US, MIT Press.
Search in Google Scholar Back to article
Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V., Courville, A. (2017). Improved Training of Wasserstein GANs. arXiv, arXiv:1704.00028, doi: 10.48550/arXiv.1704.00028
Open DOI Search in Google Scholar Back to article
Harris, C., Millman, K., van der Walt, S., Gommers, R., Virtanen, P., Cournapeau, D., Wieser, E., Taylor, J., Berg, S., Smith, N. J., Kern, R., Picus, M., Hoyer, S., van Kerkwijk, M. H., Brett, M., Haldane, A., Del Río, J. F., Wiebe, M., Peterson, P., Gérard-Marchant, P., Sheppard, K., Reddy, T., Weckesser, W., Abbasi, H., Gohlke, C. and Oliphant, T. E. (2020). Array Programming with NumPy. Nature, 585, pp. 357–362. doi: 10.1038/s41586-020-2649-2
Open DOI Search in Google Scholar Back to article
Hunter, J. D. (2007). Matplotlib: A 2D Graphics Environment. Computing in Science & Engineering, 9(3), pp. 90–95. doi: 10.1109/MCSE.2007.55
Open DOI Search in Google Scholar Back to article
Kang, Y., Dang, L., Yang, L., Wang, Z., Meng, Y., Li, S., Sun, Y., Wang, Y., and Dong, H. (2023). Research Progress in Failure Mechanism and Health State Evaluation Index System of Welded IGBT Power Modules. Electronics, 12(15), p. 3248. doi: 10.3390/electronics12153248
Open DOI Search in Google Scholar Back to article
Kennedy, D. (2021). Common Causes of VFD Failure [online]. Available at: https://goemc.com/2021/05/20/common-causes-of-vfd-failure/ [Accessed 12 Feb, 2025].
Search in Google Scholar Back to article
Kumar, R., Carroll, C., Hartikainen, A. and Martin, O. (2019). ArviZ a Unified Library for Exploratory Analysis of Bayesian Models in Python. Journal of Open Source Software, 4(33), p. 1143. doi: 10.21105/joss.01143
Open DOI Search in Google Scholar Back to article
León-López, K. M., Mouret, F., Arguello, H. and Tourneret, J.-Y. (2021). Anomaly Detection and Classification in Multispectral Time Series Based on Hidden Markov Models. IEEE Transactions on Geoscience and Remote Sensing, 60, p. 5402311. doi: 10.1109/TGRS.2021.3101127
Open DOI Search in Google Scholar Back to article
Lüer, F. and Böhm, C. (2021). Anomaly Detection using Generative Adversarial Networks: A Review. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, Singapore.
Search in Google Scholar Back to article
Manaserh, Y. M., Tradat, M. I., Hoang, C. H., Sammakia, B., Ortega, A., Nemati, K., Seymour, M. J. (2021). Degradation of Fan Performance in Cooling Electronics: Experimental Investigation and Evaluating Numerical Techniques. International Journal of Heat and Mass Transfer, 174, p. 121291. doi: 10.1016/j.ijheatmasstransfer.2021.121291
Open DOI Search in Google Scholar Back to article
Otsuki, M., Onozawa, Y., Kanemaru, H., Seki, Y., Matsumoto, T. (2003). A Study on the Short-Circuit Capability of Field-Stop IGBTs. IEEE Transactions on Electron Devices, 50, pp. 1525–1531. doi: 10.1109/TED.2003.813505
Open DOI Search in Google Scholar Back to article
Peterson, D. (2022). Teardown: What’s Inside a Variable Frequency Drive (VFD)? Control Automation, technical article, website: https://control.com/technical-articles/teardown-whats-inside-a-vfd/, accessed on 05-05-2025
Search in Google Scholar Back to article
Qi, S., Chen, J., Chen, P., Wen, P., Shan, W. and Xiong, L. (2023). An Effective WGAN-Based Anomaly Detection Model for IoT Multivariate Time Series. Kyoto, Japan: Pacific-Asia Conference on Knowledge Discovery and Data Mining.
Search in Google Scholar Back to article
Surówka, A., Tan, R., Saberi, A., and Firla, M. (2023). Performance of machine-learning-based algorithms for anomaly detection in variable frequency drives using temperature signals. In: 2023 IEEE 14th International Symposium on Diagnostics for Electrical Machines, Power Electronics and Drives (SDEMPED), Chania, Greece, 28–31 August 2023.
Search in Google Scholar Back to article
Surówka, A., Mikkelä, T., Kavala, A., and Firla, M. (2024a). Dual-mode hidden Markov models for smart detection of clogging in variable frequency drives. In: IEEE 21st International Power Electronics and Motion Control Conference, Pilsen, Czech Republic, 2024.
Search in Google Scholar Back to article
Surówka, A., Tan, R., Saberi, A. and Firla, M. (2024b). Out of Bounds Anomaly Scores in Anomaly Detection in Variable Frequency Drives Using Temperature Signals. IEEE Transactions on Industry Applications, 60(5), pp. 6988–7000. doi: 10.1109/TIA.2024.3427712
Open DOI Search in Google Scholar Back to article
Tang, S., Shi, H., Song, B., Tao, Y., and Tan, S. (2025). Physically-Consistent-WGAN Based Small Sample Fault Diagnosis for Industrial Processes. Chinese Journal of Chemical Engineering, 78, pp. 163–174. doi: 10.1016/j.cjche.2024.10.028
Open DOI Search in Google Scholar Back to article
Tuli, S., Casale, G. and Jennings, N. (2022). TranAD: Deep Transformer Networks for Anomaly Detection in Multivariate Time Series Data. arXiv preprint, p. 2201.07284. doi: 10.48550/arXiv.2201.07284
Open DOI Search in Google Scholar Back to article
Xia, X., Pan, X., Li, N., He, X., Ma, L., Zhang, X., and Ding, N. (2023). GAN-Based Anomaly Detection: A Review. Neurocomputing, 493, pp. 497–535. doi: 10.1016/j.neucom.2021.12.093
Open DOI Search in Google Scholar Back to article
Xu, L., Xu, K., Qin, Y., Li, Y., Huang, X., Lin, Z., and Ji, X. (2022). TGAN-AD: Transformer-Based GAN for Anomaly Detection of Time Series Data. Applied Sciences, 12(16), 8085. doi: 10.3390/app12168085
Open DOI Search in Google Scholar Back to article
Yellamati, D., Arthu, E., James, S., Morris, G., Heydt, T., and Graf, E. (2013). Predictive Reliability Models for Variable Frequency Drives Based on Application Profiles. Orlando, FL: USA.
Search in Google Scholar Back to article
Yinka-Banjo, C., and Ugot, OA. (2020). A review of generative adversarial networks and its application in cybersecurity. Artif Intell Rev 53, 1721–1736. doi: 10.1007/s10462-019-09717-4
Open DOI Search in Google Scholar Back to article
Zhang, C. and Yang, T. (2023). Anomaly Detection for Wind Turbines Using Long Short-Term Memory-Based Variational Autoencoder Wasserstein Generation Adversarial Network under Semi-Supervised Training. Energies, 16(19), p. 7008. doi: 10.3390/en16197008
Open DOI Search in Google Scholar Back to article
Zhang, M., Gómez, P. I., Xu, Q. and Dragicevic, T. (2023). Review of Online Learning for Control and Diagnostics of Power Converters and Drives: Algorithms, Implementations and Applications. Renewable and Sustainable Energy Reviews, 186, p. 113627. doi: 10.1016/j.rser.2023.113627
Open DOI Search in Google Scholar Back to article
Zmrhal, V. and Boháč, J. (2021). Pressure Loss of Flexible Ventilation Ducts for Residential Ventilation: Absolute Roughness and Compression Effect. Journal of Building Engineering, 44, p. 103320. doi: 10.1016/j.jobe.2021.103320
Open DOI Search in Google Scholar Back to article

Low-Parameter Critic-Based Multivariate WGAN Model for Clogging Detection in Drives

Abstract

Paradigm

My account