Deep learning-based monitoring of offshore wind turbines in Shandong Sea of China and their location analysis
Longxing Liu, Mengquan Wu, Jie Zhao, Lei Bing, Longxiao Zheng, Shaopeng Luan, Yunfei Mao, Mingyue Xue, Jiayan Liu, Bowen Liu
Abstract
Offshore wind power (OWP) is one of the underpinnings of the ambitious “dual carbon” strategy in China. So, it is of great significance to monitor the spatial information and location characteristics of offshore wind turbines (OWTs) along the coast. In this study, the OWTs in Shandong sea of China were monitored using SSD (Single Shot MultiBox Detector) with ResNet34 (ResNet34-SSD) and Sentinel-2 remote sensing images. Their spatial information and location characteristics were revealed. The results show that (1) the ResNet34-SSD can identify OWTs with high accuracy and efficiency. The precision, recall and F1 are 96.58%, 91.59% and 94.02%, respectively; (2) the number of OWTs in Shandong sea continuously increases between 2021 and 2023. They are located in the sea of Dongying (149 turbines), the southern part of Laizhou Bay (88 turbines) and the sea of Haiyang (187 turbines); (3) the installed OWTs are located in areas with 10∼38 m depth and 10–37 km offshore. There are currently no OWTs located in deep and distant ocean; (4) the mean wind speed at the locations of OWTs is 6.0–7.0 m/s. The wind power density (WPD) is 250∼400 W/m 2 , with a coefficient of variation (Cv) of 1.40–1.55. The effective wind speed hours (EWSHs) are 7,100∼7,850 h, and the effective storage of wind energy (ESOWE) is about 1,800∼3,000 kWh/m 2 . This study also discussed the location advantages of OWTs from the perspective of wind energy resources.