International Peer-Reviewed JournalOpen AccessISSN 2456-8880
irejournals@gmail.com+91-7433024337

Home / Current Issue / Paper 1707685

1707685PublishedVol 8 · Issue 9

Prediction of Electrical Grid Collapse Using Hybrid Deep Learning Model

Olumide Olayode AJAYI Umar ILIYASU

Subject area: Science,Engineering and Technology  ·  Area of research: Electrical power and Deep Learning

Abstract

Electrical power supply is crucial to the growth and development of every economy. Power system components including generators, transformers, transmission lines, buses and loads are interconnected to form an electrical grid. However, on large grids such as the Wide Area Synchronous Grid (WASG), a problem that occurs in one part of the network would result in disturbance across the entire grid leading to possible collapse of the grid and eventual blackout. Thus, an accurate and timely prediction of the power network instability is of utmost importance. This paper proposes a grid-collapse prediction model developed by hybridizing the One-Dimensional Convolutional Neural Network (1D-CNN) and Long Short-Term Memory Networks (LSTM) deep learning techniques. Historical real-time system frequency data of the national grid in the Nigeria Electricity System Industry (NESI) was used in this study. The system frequency data was divided into training and test datasets. The 1D-CNN was used to extract the features for training the LSTM network. Experimental results showed that the 1D-CNN-LSTM model gave RMSE, MAE and MAPE values of 0.0109 Hz, 0.0008 Hz, and 0.0015%, respectively, whereas using the LSTM model alone gave RMSE, MAE and MAPE values of 0.0194 Hz, 0.0099 Hz, and 0.0197%, respectively. The proposed 1D-CNN-LSTM model gives significant improvement over the LSTM model in prediction accuracy.

Keywords

Electrical Grid Collapse, System Frequency, 1D-CNN, LSTM, Nigeria Electricity System Industry (NESI).

How to cite this paper

Olumide Olayode AJAYI, Umar ILIYASU "Prediction of Electrical Grid Collapse Using Hybrid Deep Learning Model" Iconic Research And Engineering Journals Volume 8 Issue 9 2025 Page 1396-1404
Olumide Olayode AJAYI, Umar ILIYASU "Prediction of Electrical Grid Collapse Using Hybrid Deep Learning Model" Iconic Research And Engineering Journals, vol. 8, no. 9, Mar. 2025
Olumide Olayode AJAYI, Umar ILIYASU (2025). Prediction of Electrical Grid Collapse Using Hybrid Deep Learning Model. Iconic Research And Engineering Journals, 8(9).
Olumide Olayode AJAYI, Umar ILIYASU "Prediction of Electrical Grid Collapse Using Hybrid Deep Learning Model" Iconic Research And Engineering Journals, vol. 8, no. 9, Mar. 2025.
@article{1707685,
      author = {Olumide Olayode AJAYI, Umar ILIYASU},
      title = {Prediction of Electrical Grid Collapse Using Hybrid Deep Learning Model},
      journal = {Iconic Research And Engineering Journals},
      year = {2025},
      volume = {8},
      number = {9},
      pages = {1396-1404},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1707685.pdf},
      abstract = {Electrical power supply is crucial to the growth and development of every economy. Power system components including generators, transformers, transmission lines, buses and loads are interconnected to form an electrical grid. However, on large grids such as the Wide Area Synchronous Grid (WASG), a problem that occurs in one part of the network would result in disturbance across the entire grid leading to possible collapse of the grid and eventual blackout. Thus, an accurate and timely prediction of the power network instability is of utmost importance. This paper proposes a grid-collapse prediction model developed by hybridizing the One-Dimensional Convolutional Neural Network (1D-CNN) and Long Short-Term Memory Networks (LSTM) deep learning techniques. Historical real-time system frequency data of the national grid in the Nigeria Electricity System Industry (NESI) was used in this study. The system frequency data was divided into training and test datasets. The 1D-CNN was used to extract the features for training the LSTM network. Experimental results showed that the 1D-CNN-LSTM model gave RMSE, MAE and MAPE values of 0.0109 Hz, 0.0008 Hz, and 0.0015%, respectively, whereas using the LSTM model alone gave RMSE, MAE and MAPE values of 0.0194 Hz, 0.0099 Hz, and 0.0197%, respectively. The proposed 1D-CNN-LSTM model gives significant improvement over the LSTM model in prediction accuracy.},
      keywords = {Electrical Grid Collapse, System Frequency, 1D-CNN, LSTM, Nigeria Electricity System Industry (NESI).},
      month = {March},
  }