Session: 01-11-03: Wind Energy

Paper Number: 110796

110796 - Machine Learning-Based Dynamic Wake Modelling of Large-Scale Offshore Wind Farm 

Climate change is one of the most significant challenges in the 21st century, where carbon dioxide from fossil energy sources is a major contributor. However, the current energy supply in the world still has a sharp contradiction between energy supply and demand, unreasonable structure, and low energy utilization efficiency. In this situation, there is a growing interest worldwide in increasing the production of renewable energy to reduce the emissions of climate gases, especially wind energy, a clean, accessible, and widely available renewable energy source.

Wind energy conversion into electricity using wind turbines is the main form of harnessing wind energy. With the increasing demand for electric power, the size of wind farms is becoming much larger than ever before. Power and load prediction are one of the most essential topics in wind farm layout optimization. Traditional wake modeling, such as analytic models and CFD simulation, become difficult to handle such large-scale problems accurately and efficiently. In this study, a novel dynamic wake modeling of wind turbines using machine learning is proposed to capture the spatial-temporal features of the unsteady wake field of multiple wind turbines in a wind farm. The machine learning architecture is constituted of fully connected neural networks, convolutional neural networks, and long short-term memory neural networks. Due to considering the trade-off between accuracy and efficiency, the dynamic wake meandering (DWM) model is employed to generate big datasets of wake field for training, testing, and validation of the hybrid machine learning model. This research is expected to accelerate the prediction process and improve accuracy, and it can be further applied to wind turbine design and wind farm layout optimization.

Presenting Author: Qiulei Wang The University of Hong Kong

Presenting Author Biography: Mr. Qiulei Wang obtained his MSc from Harbin Institute of Technology in 2021 and BSc from Hefei University of Technology in 2018. Currently, he is a PhD candidate at the Department of Civil Engineering, The University of Hong Kong. He was awarded the China National Scholarship from the Chinese Ministry of Education in 2015 and was selected as an Excellent Graduate in 2018. Mr. Wang’s research mainly focuses on the wake modelling of large-scale offshore wind farms using machine learning technology.

Authors:

Qiulei Wang The University of Hong Kong
Xiaowei Deng The University of Hong Kong