A novel offshore wind power prediction model based on TCN-DANet-sparse transformer with dual-channel feature extraction and multi-scale temporal fusion

This paper proposes a novel hybrid deep learning architecture for offshore wind power prediction. The model integrates:

• Temporal Convolutional Network (TCN) for local temporal pattern extraction
• Dual Attention Network (DANet) for adaptive feature weighting
• Sparse Transformer for capturing long-range dependencies efficiently

The dual-channel design enables simultaneous extraction of meteorological and operational features, while multi-scale temporal fusion improves prediction accuracy across different time horizons.

Recommended citation: Chen, J., et al. (2024). "A novel offshore wind power prediction model based on TCN-DANet-sparse transformer." Energy.