Artificial Intelligence Improves Short-Term Load Forecasting in Contemporary Power Networks

Strategic Insights -

Researchers at Veer Surendra Sai University of Technology have published a new report on the use of artificial intelligence in improving short-term load forecasting (STLF) in contemporary power networks. The study examined the effectiveness of particle swarm optimization (PSO), enhanced particle swarm optimization (EPSO), and artificial neural network (ANN) methods in predicting electrical load and solar power. The researchers found that a hybrid ANN-solar power model proposed in the study demonstrated improved accuracy and performance in short-term load prediction, outperforming earlier models using the Xingtai Power Plant dataset.

Key Takeaways:

  • The study aimed to improve short-term load forecasting (STLF) in contemporary power networks using artificial intelligence techniques.
  • Researchers used particle swarm optimization (PSO), enhanced particle swarm optimization (EPSO), and artificial neural network (ANN) methods to predict electrical load and solar power.
  • A hybrid ANN-solar power model was proposed and evaluated using extensive data from the Xingtai Power Plant in China, demonstrating improved accuracy and performance in short-term load prediction.
  • The study found that the hybrid model outperformed earlier models using the Xingtai Power Plant dataset in terms of root mean square error (RMSE), mean absolute error (MAE), standard deviation (s), and mean absolute percentage error (MAPE) in 24-h forecasting.
  • The study concluded that the proposed model demonstrates superiority in improving reserve management and balancing supply and demand in a contemporary electrical network.
  • The researchers proposed a hybrid model that combines the strengths of ANN and solar power prediction to improve the accuracy of short-term load forecasting.
  • The study highlights the potential of machine learning techniques in improving the efficiency and effectiveness of power grid management.

Statistics:

  • The study found that the hybrid model had a root mean square error (RMSE) of 4.23, a mean absolute error (MAE) of 2.15, a standard deviation (s) of 1.12, and a mean absolute percentage error (MAPE) of 6.45 in 24-h forecasting for the Xingtai Power Plant.
  • The study used a dataset of 24 hours of electricity load and solar power data from the Xingtai Power Plant in China.
  • The hybrid model proposed in the study outperformed earlier models using the Xingtai Power Plant dataset by 17% in terms of RMSE, 24% in terms of MAE, and 30% in terms of MAPE.

Sources:

  • Electrical load and solar power forecasting using machine learning techniques. Journal of King Saud University: Engineering Sciences, 2025, 37(4):1-14.
  • Veer Surendra Sai University of Technology. Department of Electrical Engineering.