IET Generation Transmission & Distribution · 2025 · 12 citations · 29 references
ComputingComputer ScienceDemand ForecastingMachine LearningRenewable Energy IntegrationDeep LearningEnergy Demand ForecastingLoad ForecastingRenewable EnergySmart GridsConvolutional Neural NetworksDeep Learning For Time Series ForecastingDemand PredictionEnergy ConsumptionGated Recurrent Unit NetworksHybrid Prediction ModelsLoad PredictionNeural Time Series Forecasting
ABSTRACT Integrating renewable energy sources into smart grids increases supply and demand management because renewable energy sources are intermittent and variable. To overcome this type of challenge, short‐term load forecasting (STLF) is essential for managing energy, demand‐side flexibility, and the stability of smart grids with renewable energy integration. This paper presents a new model called BiGRU‐CNN to improve the operation of STLF in smart grids. The BiGRU‐CNN model integrates bidirectional gated recurrent units (BiGRUs) to model temporal dependencies and convolutional neural networks (CNNs) to extract spatial patterns from energy consumption data. The newly developed BiGRU captures past and future contexts through bidirectional processing, and the CNN component extracts high‐level features to enhance the accuracy of load demand prediction. The BiGRU‐CNN model is compared with two other hybrid models, CNN‐LSTM and CNN‐GRU, on real‐world data from American electric power (AEP) and ISONE datasets. Simulation results show that the proposed BiGRU‐CNN outperforms other models, with single‐step forecasting yielding root mean square error (RMSE) of 121.43 (AEP) and 123.57 (ISONE), mean absolute error (MAE) of 90.95 (AEP) and 62.97 (ISONE), and mean absolute percentage error (MAPE) of 0.61% (AEP) and 0.41% (ISONE). For multi‐step forecasting, the model yields RMSE of 680.02 (AEP) and 581.12 (ISONE), MAE of 481.12 (AEP) and 411.20 (ISONE), and MAPE of 3.27% (AEP) and 2.91% (ISONE). The results show that the BiGRU‐CNN model can generate accurate and reliable STLF, which is useful in managing massive renewable energy‐integrated smart grids.
29
A Novel CNN-GRU-Based Hybrid Approach for Short-Term Residential Load Forecasting
Muhammad Sajjad, Zulfiqar Ahmad Khan, Amin Ullah et al. · IEEE Access · 2020 · 511 citations · Full text
A Short-Term Load Forecasting Method Using Integrated CNN and LSTM Network
Shafiul Hasan Rafi, Nahid‐Al Masood, Shohana Rahman Deeba et al. · IEEE Access · 2021 · 495 citations · Full text