Prediction of Time to Failure (TTF) of Power Systems Using a Deep Learning Technique

Muhammad Zohaib Sohail, Taimoor Zafar, Talha Ahmed Khan, Muhammad Asim, Sadique Ahmad, Tariq Mairaj, Mohammed A. ELAffendi

Abstract

Power distribution systems consist of various components, including transformers, generators, bus bars, switch gear, utility electric poles, and power distribution cables. A high reliability of a power distribution system is extremely required for the economic growth of any country. The reliable system ensures continuous power supply to the metropolis and industrial sector. Sudden power failures in residential areas and industries will affect the economy of any country. Similarly, unwanted inspection and maintenance/replacement activities will also increase the cost. The research aimed to forecast the time to failure (TTF) of a power distribution system to ensure condition-based predictive maintenance. This will help the power sector to plan maintenance activities when they are highly required. In this study, a prognostic framework to estimate TTF based on long short-term memory (LSTM) is proposed. Prior information on the TTF will ensure continuous power supply to the metropolis and industrial sector, resulting in improved economic growth of the country. To address the issues of sudden power failure, the authors proposed a prognostic framework based on deep learning Long Short-Term Memory (LSTM) approach to forecast the TTF of power grid stations for the first time. The data sets of TTF of the power grid due to equipment failure and environmental failures are used for training the LSTM algorithm. In this approach, different layers and optimizers of the LSTM scheme are used, including input, output, and hidden layers, and adaptive moment estimation (ADAM) and stochastic gradient descent moment (SGDM) optimizers that control the accuracy of the prediction step. In addition, LSTM can handle long-term dependencies, which is unlikely for other recurrent techniques. The results obtained through the LSTM approach using ADAM are found to be quite satisfactory in terms of the root mean square error (RMSE) between the actual values and predicted (computed) values compared to the SGDM. Accurate results show that the proposed prognostic framework will help power companies to plan maintenance and replacement activities well in time before the time of failure to ensure continuous power supply to the country.

 

Keywords: power systems, reliability, root mean square error.

 

https://doi.org/10.55463/issn.1674-2974.50.12.9


Full Text:

PDF


References


CANIZES B, SOARES J, VALE Z, and KHODR H M. Hybrid fuzzy Monte Carlo technique for reliability assessment in transmission power systems. Energy, 2012, 45(1): 1007-1017. https://doi.org/10.1016/j.energy.2012.06.049

PANAGIOTIS H, TSAROUHAS A, THEODOROS H, VARZAKAS B, et al. Reliability and maintainability analysis of strudel production line with experimental data. A case study. Journal of Food Engineering, 2009, 91(2): 250-259. https://doi.org/10.1016/j.jfoodeng.2008.09.002

TSAROUHAS P H. A comparative study of performance evaluation based on field failure data for food production lines. Journal of Quality in Maintenance Engineering, 2011, 17(1): 26-39. https://doi:10.1108/13552511111116231

SARASWAT S, and YADAVA G S. An overview on reliability, availability, maintainability and supportability (RAMS) engineering. International Journal of Quality & Reliability Management, 2008, 25(3): 330-344. https://doi:10.1108/02656710810854313

WANG H, YANG Z, YU Q, and HONG T. Online Reliability Time Series Prediction via Convolutional Neural Network and Long Short Term Memory for Service-Oriented Systems. Knowledge-Based Systems, 2018, 159: 132-147. https://doi.org/10.1016/j.knosys.2018.07.006

SALEH A R, DRISS M, and ALMOMANI I. CBiLSTM: A hybrid deep learning model for efficient reputation assessment of cloud services. IEEE Access, 2022, 10: 35321-35335. https://doi.org/10.1109/ACCESS.2022.3163246

DURBHAKA GK, SELVARAJ B, MITTAL M, et al. Swarm-LSTM: Condition monitoring of gearbox fault diagnosis based on hybrid LSTM deep neural network optimized by swarm intelligence algorithms. CMC-Computers, Materials and Continua, 2021, 66(2): 2041-2059. https://doi.org/10.32604/cmc.2020.013131

AMIN J, SHARIF M, RAZA M, et al. Brain tumor detection: a long short-term memory (LSTM)-based learning model. Neural Computing and Applications, 2020, 32: 15965-15973. https://doi.org/10.1007/s00521-019-04650-7

RASHID T A, FATTAH P, and AWLA D K. Using Accuracy Measure for Improving the Training of LSTM with Metaheuristic Algorithms. Procedia Computer Science, 2018, 140: 324-233. https://doi.org/10.1016/j.procs.2018.10.307.

https://www.ausgrid.com.au/Industry/Our-Research/Data-to-share/Past-outage-data

MEMON ML, MAHESHWARI MK, SHIN DR, et al. A framework for deep learning–based discontinuous reception in 5G wireless networks. Transactions on Emerging Telecommunications Technologies, 2019, 30(3): 1-13. https://doi.org/10.1002/ett.3579.

LE X-H, HO HV, LEE G, and JUNG S. Application of Long Short-Term Memory (LSTM) Neural Network for Flood Forecasting. Water, 2019, 11(7): 1387. https://doi.org/10.3390/w11071387.

STAUDEMEYER RC, and MORRIS ER. Understanding LSTM--a tutorial into long short-term memory recurrent neural networks, 2019, https://doi.org/10.48550/arXiv.1909.09586


Refbacks

  • There are currently no refbacks.