تشخیص ناهنجاری مبتنی بر رایانش لبه‌ای در ماشین‌های دوار با استفاده از شبکه‌های عصبی مصنوعی

نوع مقاله : مقاله پژوهشی

نویسندگان

دانشکده‌ی مهندسی مکانیک، دانشگاه تهران، تهران، ایران.

چکیده

در پژوهش حاضر، چارچوبی نوآورانه مبتنی بر فناوری رایانش لبه‌ای برای پایش بلادرنگ وضعیت تجهیزات صنعتی ارائه شده است. طراحی سخت‌افزار و ثابت‌افزار به گونه‌ای انجام شده است که تمام وظایف حیاتی از دریافت و پیش‌پردازش داده‌ها تا استخراج ویژگی‌ها و آموزش الگوریتم با وجود محدودیت حافظه و توان پردازشی بر روی گره‌ی لبه‌ای، یعنی میکروکنترلر، ممکن باشد. با دریافت داده‌های ارتعاشی به وسیله‌ی یک شتاب‌سنج سه‌محوره و ذخیره‌سازی داده‌ها برای آموزش در حافظه‌ی فلش میکروکنترلر، یک خودرمزگذار با سه لایه‌ی پنهان برای مدل‌سازی رفتار سالم ماشین بر روی لبه‌‌ آموزش داده شده و خطای بازسازی داده‌ها برای شناسایی ناهنجاری‌ها استفاده شده است. طبق بررسی‌های انجام‌شده، احتمالاً پژوهش حاضر، اولین پژوهشی است که یک شبکه‌ی عصبی مصنوعی را روی میکروکنترلر آموزش داده و کل فرآیند پایش وضعیت را بر روی لبه‌ پیاده‌سازی کرده است. آزمایش چارچوب پیشنهادی روی یک پمپ گریز از مرکز نشان می‌دهد که سیستم پیشنهادی می‌تواند انواع عیوب پمپ را با دقت بالای 9/99% شناسایی کند.

کلیدواژه‌ها

موضوعات

عنوان مقاله [English]

Edge-Computing-Based Anomaly Detection of Rotating Machines Using Artificial Neural Networks

نویسندگان [English]

  • Alireza Mostafavi
  • Aysan Alizadeh
  • Ali Sadighi
School of Mechanical Engineering, University of Tehran, Tehran, Iran
چکیده [English]

Manufacturing companies face issues due to demand for high-quality, affordable products. Since maintenance costs account for 60–70% of production costs, real-time fault detection is vital to lower maintenance expenses and extend equipment life. This article introduces a high-performance anomaly detection framework using edge computing for real-time industrial asset monitoring. Hardware and firmware were designed to perform critical tasks such as data acquisition, preprocessing, feature extraction, and algorithm training on the microcontroller unit (MCU), despite limited processing and memory. Using a 3-axis accelerometer for vibration signals, the MCU stores training data in Flash memory. An autoencoder with three hidden layers is trained on the edge device to model normal operating conditions, and reconstruction error of new data detects anomalies. This study is, to the best of our knowledge, the first to train an artificial neural network (ANN) on an MCU for comprehensive edge-based condition monitoring. achieved over 99.9% accuracy when validated on a centrifugal pump

کلیدواژه‌ها [English]

  • Edge computing
  • online learning
  • anomaly detection
  • autoencoders
  • rotating machines

مراجع

1. Ding, S.X., 2008. Model-based fault diagnosis techniques: design schemes, algorithms, and tools. Springer Science & Business Media. https://doi.org/10.1007/978-3-540-76304-8
2. Khan, S. and Yairi, T., 2018. A review on the application of deep learning in system health management. Mechanical Systems and Signal Processing, 107, pp.241-265.  https://doi.org/10.1016/j.ymssp.2017.11.024
3. Wu, D., Liu, S., Zhang, L., Terpenny, J., Gao, R.X., Kurfess, T. and Guzzo, J.A., 2017. A fog computing-based framework for process monitoring and prognosis in cyber-manufacturing. Journal of Manufacturing Systems, 43, pp.25-34.          https://doi.org/10.1016/j.jmsy.2017.02.011
4. Xenakis, A., Karageorgos, A., Lallas, E., Chis, A.E. and González-Vélez, H., 2019. Towards distributed IoT/cloud based fault detection and maintenance in industrial automation. Procedia Computer Science, 151, pp.683-690.          https://doi.org/10.1016/j.procs.2019.04.091.
5. Park, D., Kim, S., An, Y. and Jung, J.Y., 2018. LiReD: A light-weight real-time fault detection system for edge computing using LSTM recurrent neural networks. Sensors, 18(7), p.2110.https://doi.org/10.3390/s18072110.
6. Shi, W., Cao, J., Zhang, Q., Li, Y. and Xu, L., 2016. Edge computing: Vision and challenges. IEEE internet of things journal, 3(5), pp.637-646. https://doi.org/10.1109/JIOT.2016.2579198.
7. Qian, G., Lu, S., Pan, D., Tang, H., Liu, Y. and Wang, Q., 2019. Edge computing: A promising framework for real-time fault diagnosis and dynamic control of rotating machines using multi-sensor data. IEEE Sensors Journal, 19(11), pp.4211-4220. https://doi.org/10.1109/JSEN.2019.2899396.
8. Huang, H., Yang, L., Wang, Y., Xu, X. and Lu, Y., 2021. Digital twin-driven online anomaly detection for an automation system based on edge intelligence. Journal of Manufacturing Systems, 59, pp.138-150. https://doi.org/10.1016/j.jmsy.2021.02.010.
9. Yang, H., Mathew, J. and Ma, L., 2002. Intelligent diagnosis of rotating machinery faults-a review. InProceedings of the 3rd Asia-Pacific Conference on Systems Integrity and Maintenance, pp. 385-392. Queensland University of Technology Press.
10. Markou, M. and Singh, S., 2003. Novelty detection: a review—part 2:: neural network based approaches. Signal processing, 83(12), pp.2499-2521. https://doi.org/10.1016/j.sigpro.2003.07.019.
11. Maasoum, S.M.H., Mostafavi, A. and Sadighi, A., 2020, December. An autoencoder-based algorithm for fault detection of rotating machines, suitable for online learning and standalone applications. In2020 6th Iranian Conference on Signal Processing and Intelligent Systems (ICSPIS), pp. 1-6. IEEE.       https://doi.org/10.1109/ICSPIS51611.2020.9349574.
12. Ahmad, S., Styp-Rekowski, K., Nedelkoski, S. and Kao, O., 2020, December. Autoencoder-based condition monitoring and anomaly detection method for rotating machines. In 2020 IEEE International Conference on Big Data (Big Data), pp. 4093-4102. IEEE. https://doi.org/10.1109/BigData50022.2020.9378015.
13. A simple, fast and optimized method to deploy Neural Networks on STM32: https://www.st.com/content/st_com/en/about/innovation---technology/artificial-intelligence.html.
14. TensorFlow Lite for Microcontrollers: https://www.tensorflow.org/lite/microcontrollers
15. CMSIS NN software library: https://www.keil.com/pack/doc/CMSIS/NN/html/index.html.
16. CUBE-AI, AI expansion pack for STM32CubeMX: https://www.st.com/en/embedded-software/x-cube-ai.html.
17. Apache TVM: https://tvm.apache.org/.
18. Ren, H., Anicic, D. and Runkler, T.A., 2021, July. Tinyol: Tinyml with online-learning on microcontrollers. In2021 International Joint Conference on Neural Networks (IJCNN), pp. 1-8. IEEE. https://doi.org/10.1109/IJCNN52387.2021.9533927
19. Tandon, N. and Choudhury, A., 1999. A review of vibration and acoustic measurement methods for the detection of defects in rolling element bearings. Tribology International, 32(8), pp.469-480. https://doi.org/10.1016/S0301-679X(99)00077-8.
20. Williams, T., Ribadeneira, X., Billington, S. and Kurfess, T., 2001. Rolling element bearing diagnostics in run-to-failure lifetime testing. Mechanical Systems and Signal Processing, 15(5), pp.979-993.            https://doi.org/10.1006/mssp.2001.1418.
21. Xia, Z., Xia, S., Wan, L. and Cai, S., 2012. Spectral regression based fault feature extraction for bearing accelerometer sensor signals. Sensors, 12(10), pp.13694-13719.  https://doi.org/10.3390/s121013694.
22. Goumas, S.K., Zervakis, M.E. and Stavrakakis, G.S., 2002. Classification of washing machines vibration signals using discrete wavelet analysis for feature extraction. IEEE Transactions on Instrumentation and Measurement, 51(3), pp.497-508. https://doi.org/10.1109/TIM.2002.1017721.
23. CodeBlocks C compiler, https://www.codeblocks.org/.
24. STMicroelectronics, "STM32H742xI/G, STM32H743xI/G: 32-bit Arm® Cortex®-M7 480MHz MCUs, up to 2MB Flash, up to 1MB RAM, 46 com. and analog interfaces," https://www.st.com/resource/en/datasheet/stm32h743zi.pdf

فایل ها

هم رسانی

ارجاع به این مقاله

آمار