Multi-Step Wind Power Forecasting with Stacked Temporal Convolutional Network (S-TCN)

Nowadays, wind power generation has become vital thanks to its advantages in cost, ecological friendliness, enormousness, and sustainability. However, the erratic intermittent nature of this energy poses significant operational management difficulties for systems. Currently, methods forecasting (WPF) are various numerous. An accurate method WPF can help system dispatchers plan unit commitment reduce risk unreliability electricity supply. In order improve accuracy short-term prediction address multi-step ahead forecasting, research presents a Stacked Temporal Convolutional Network (S-TCN) model. By using dilated causal convolutions residual connections, suggested solution addresses issue long-term dependencies performance degradation deep convolutional models sequence prediction. The simulation outcomes demonstrate that S-TCN model’s training procedure is extremely stable powerful capacity generalization. Besides, proposed model shows higher compared other existing neural networks like Vanilla Long Short-Term Memory or Bidirectional