Fourier-Enhanced Recurrent Neural Networks for Electrical Load Time Series Downscaling
This work addresses electrical load forecasting for grid management, but it is incremental as it builds on existing RNN and attention methods with added Fourier features.
The paper tackles the problem of downscaling electrical load time series by proposing a Fourier-enhanced RNN that combines recurrent backbones, Fourier seasonal embeddings, and self-attention, achieving lower and flatter RMSE across four PJM territories compared to Prophet baselines and RNN ablations.
We present a Fourier-enhanced recurrent neural network (RNN) for downscaling electrical loads. The model combines (i) a recurrent backbone driven by low-resolution inputs, (ii) explicit Fourier seasonal embeddings fused in latent space, and (iii) a self-attention layer that captures dependencies among high-resolution components within each period. Across four PJM territories, the approach yields RMSE lower and flatter horizon-wise than classical Prophet baselines (with and without seasonality/LAA) and than RNN ablations without attention or Fourier features.