Improved TCN-Vine Copula Method for Wind Speed Prediction Using 5G+ Technology Data Transmission

• Published in: 2024 6th International Conference on Industrial Artificial Intelligence (IAI) pp. 1 - 6
• Main Authors: Zhan, Jin, He, Shaowei, Wang, Chunyu, Zheng, Yafeng
• Format: Conference Proceeding
• Language: English
• Published: IEEE 21.08.2024
• Subjects: Attention mechanisms; Complexity theory; Data communication; Feature extraction; improved temporal convolutional network; kernel principal component analysis; meteorological factors; Predictive models; Sensors; vine copula; Wind energy; Wind farms; Wind speed; wind speed prediction; Wind turbines
• DOI: 10.1109/IAI63275.2024.10730724

Accurate wind speed prediction can facilitate the efficient utilization of wind energy. Therefore, a prediction method combining improved temporal convolutional network (TCN) with vine copula using 5G+ technology data transmission is proposed. Initially, the residual module in TCN is improved by using the attention mechanism and soft thresholding concept from deep residual shrinkage network (DRSN) for preliminary prediction. Subsequently, given the influence of meteorological factors on wind speed, the meteorological data are downscaled by using the kernel principal component analysis (KPCA), which reduces the complexity of the data under the premise of guaranteeing the complete features. Finally, a correction model based on vine copula is constructed, and the preliminary prediction values is corrected using the downscaled meteorological data to obtain the final prediction results. This paper takes wind turbines in a wind farm as the experimental subjects, and utilizes sensors and 5G+ technology to collect wind speed and related meteorological data. Experiments using measured data verifies the effectiveness of the proposed method.