Remaining Useful Life Prediction of Servo Gear Trains Using Motor Current Signature Analysis and IoT-Enabled Exponential Degradation Modelling

Paper ID: EIJTEM_2025_12_4_185-198

Author's Name: Kanchi Lakshmi Prasanna, Dr. D. Naga Sudha

Volume: 12

Issue: 4

Year: 2025

Page No: 185-198

Abstract:

Predictive maintenance has become an indispensable aspect of modern industrial systems, facilitating early fault detection and minimizing unplanned downtime. This project presents a comprehensive framework for estimating the Remaining Useful Life (RUL) of servo gear trains, which constitute a critical component of servomotors extensively utilized in robotics, industrial automation, and radio-controlled mechanisms. A servo motor integrates an electric motor, control circuitry, and gearing—where the gear train ensures mechanical precision and reliability. Accurate RUL prediction of the servo gear train aids in preventing unexpected breakdowns, optimizing maintenance schedules, and reducing overall operational costs. The proposed work employs synthetic motor current signal data to visualize and predict the RUL of servo gear trains through Motor Current Signature Analysis (MCSA), a robust and non-intrusive method for fault diagnosis. MCSA analyzes the current signals driving the servo motor to extract frequency-domain (spectral) features from characteristic frequency regions corresponding to motor and gear train faults. This approach effectively detects torque and speed fluctuations without requiring additional sensors such as vibration transducers—making it particularly suitable for inaccessible or enclosed gear train assemblies. Extracted spectral features, including motor current (mA), revolutions per minute (RPM), and fault band metrics, are utilized to develop a Health Indicator (HI) that reflects the progressive degradation of the gear train. The HI is subsequently applied within an Exponential Degradation Model (EDM), a machine learning-based method tailored for automated and adaptive RUL estimation. Key model parameters—Θ (Theta), β (Beta), ρ (Rho), Beta Variance, and Theta Variance—enable dynamic updating of RUL predictions as new feature data is acquired, ensuring high accuracy and responsiveness to system changes. To facilitate real-time monitoring and data management, the project integrates MATLAB with ThingSpeak IoT services for seamless data acquisition, cloud storage, processing, and visualization. This integration enables remote supervision of gear train health and supports predictive decision-making, thereby advancing the reliability and efficiency of servo-based industrial systems.

Keywords: Predictive Maintenance, Remaining Useful Life, Servo Gear Train, Motor Current Signature Analysis, Exponential Degradation Model, Health Indicator, MATLAB, ThingSpeak IoT, Fault Diagnosis, Industrial Automation

References:

[1] A. K. Jardine, D. Lin, and D. Banjevic, “A review on machinery diagnostics and prognostics implementing condition-based maintenance,” Mechanical Systems and Signal Processing, vol. 20, no. 7, pp. 1483–1510, Oct. 2006.
[2] M. E. H. Benbouzid, “A review of induction motors signature analysis as a medium for faults detection,” IEEE Transactions on Industrial Electronics, vol. 47, no. 5, pp. 984–993, Oct. 2000.
[3] MathWorks, “Predictive Maintenance Using MATLAB and ThingSpeak,” MathWorks Technical Documentation and Examples, 2023. [Online]. Available: https://www.mathworks.com/help/thingspeak/examples.html
[4] P. Zhang, Y. Lu, T. W. S. Chow, and M. H. J. Bollen, “Motor current signature analysis based on autoregressive modeling for fault diagnosis,” IEEE Transactions on Industrial Electronics, vol. 60, no. 8, pp. 3480–3488, Aug. 2013.
[5] S. Zheng, K. Ristovski, A. Farahat, and C. Gupta, “Long short-term memory network for remaining useful life estimation,” in Proc. IEEE International Conference on Prognostics and Health Management (PHM), Dallas, TX, USA, 2017, pp. 88–95.
[6] S. Li, Y. Wang, and X. Ding, “Deep ordinal regression networks for remaining useful life estimation,” Neurocomputing, vol. 459, pp. 285–297, Oct. 2021.
[7] Y. Lei, N. Li, L. Guo, N. Li, T. Yan, and J. Lin, “Machinery health prognostics: A systematic review from data acquisition to RUL prediction,” Mechanical Systems and Signal Processing, vol. 104, pp. 799–834, May 2018.
[8] M. Zhao, M. Kang, B. Tang, and M. Pecht, “Multiple wavelet coefficients fusion for gear fault diagnosis using improved Hilbert–Huang transform,” IEEE Transactions on Industrial Electronics, vol. 65, no. 7, pp. 5892–5902, Jul. 2018.
[9] J. C. An, C. B. Choi, and B. K. Lee, “Monotonicity-based health indicator construction for prognostics of rotating machinery,” Reliability Engineering & System Safety, vol. 216, p. 108005, Oct. 2021.
[10] S. Li, X. Y. Zhang, and C. G. Huang, “Nonlinear health indicator construction using kernel principal component analysis for remaining life prediction,” Journal of Intelligent Manufacturing, vol. 33, pp. 1539–1552, Feb. 2022.
[11] D. Wang, K. Tsui, and Q. Zhou, “Remaining useful life estimation of lithium-ion batteries based on Bayesian exponential degradation model,” Applied Energy, vol. 154, pp. 356–367, Sep. 2015.
[12] S. Gebraeel and M. Lawley, “A particle filtering framework for prognostics of discrete degradation signals,” IEEE Transactions on Reliability, vol. 61, no. 1, pp. 216–225, Mar. 2012.
[13] A. Saxena, K. Goebel, D. Simon, and N. Eklund, “Damage propagation modeling for aircraft engine run-to-failure simulation,” in Proc. IEEE International Conference on Prognostics and Health Management (PHM), Denver, CO, USA, 2008, pp. 1–9.
[14] MathWorks, “Develop Predictive Maintenance Algorithms Using MATLAB and ThingSpeak,” MathWorks Technical Documentation, 2023. [Online]. Available: https://www.mathworks.com/help/thingspeak
[15] M. Khadar, C. Smaili, and R. Younes, “Digital twin-driven fault prognosis using synthetic data for industrial gearboxes,” IEEE Access, vol. 11, pp. 44756–44770, Apr. 2023.
[16] P. S. Moghadam, S. T. Ali, and M. K. Lee, “Industrial IoT-based predictive maintenance framework for smart factories,” Sensors, vol. 22, no. 5, p. 1804, Mar. 2022.
[17] Deloitte Insights, “Predictive maintenance and the smart factory,” Deloitte Development LLC Report, 2021. [Online]. Available: https://www.deloitte.com
[18] C. S. Byington, M. Roemer, T. Galie, and J. D. Watson, “Prognostic enhancements to diagnostic systems for improved condition-based maintenance,” IEEE Aerospace Conference Proceedings, Big Sky, MT, USA, 2002, pp. 2815–2824.
[19] A. Bellini, F. Filippetti, C. Tassoni, and G. A. Capolino, “Advances in diagnostic techniques for induction machines,” IEEE Transactions on Industrial Electronics, vol. 55, no. 12, pp. 4109–4126, Dec. 2008.
[20] M. R. P. Kiran, K. V. Suresh, and P. K. Reddy, “Edge–cloud collaboration framework for industrial predictive maintenance using motor current data,” IEEE Internet of Things Journal, vol. 10, no. 3, pp. 2357–2368, Feb. 2023.
[21] J. Hu, H. Li, and X. Zhou, “A Bayesian exponential degradation model for dynamic remaining useful life prediction,” Reliability Engineering & System Safety, vol. 235, p. 109147, Jun. 2023.
[22] A. C. Saha, R. T. Baiju, and P. M. George, “Adaptive health indicator update for online RUL estimation using recursive exponential modeling,” Measurement, vol. 217, p. 113157, Jan. 2024.
[23] N. Kumar, S. Gupta, and R. Prakash, “IoT-enabled predictive maintenance framework for rotating machinery under variable speed conditions,” IEEE Access, vol. 11, pp. 94532–94546, Aug. 2023.
[24] Y. Zhao and M. Pecht, “Deep learning-based IoT framework for real-time fault prognosis of industrial motors,” IEEE Transactions on Industrial Informatics, vol. 19, no. 5, pp. 5876–5885, May 2023.
[25] H. Liang, Y. Tao, and D. Xu, “Cyber-physical integration for intelligent maintenance in Industry 4.0,” Annual Reviews in Control, vol. 55, pp. 129–142, Dec. 2023.

View PDF