一种快速精确表征风电场尾流的低阶场模型开发研究

风力发电机的部署随着全球对清洁能源的需求飞速增长。同时，风力发电机的单机尺寸和安装密度越来越大，使得下游风力机的尾流损失和阻塞效应显著增加。准确评估这些因素对于优化风电场选址并最大化发电量至关重要。挑战在于，风电场的尾流特性是一个多元问题，需兼顾稳态尾流损失、动态尾迹蜿蜒和对流不稳定等问题。高精度数值模拟作为有效的验证工具在实际应用中太昂贵。而低阶模型虽可准确捕捉单个风力机的尾流，但多基于运动学并且依赖非物理的尾流合并方法。在本项研究中，我们发展一种无需采用尾流合并方法的低阶场模型表征风电场中的动态尾流。首先，通过准稳态简化形式的Navier-Stokes来模拟稳态尾流损失和较慢尺度的被动尾流蜿蜒；其次，基于得到的准稳态尾流，使用Ginzburg-Landau模型计算较快尺度的不稳定性；最后，将这些相互依赖的部分迭代组合。完成后，该方法将成为准确描述风电场中尾流信息的快速建模工具。

The deployment of wind turbines is growing at a remarkable pace owing to increasing global demand for clean energy. This is particularly true in China, which accounted for nearly half of the new global wind energy installations in 2018. Wind farms, in response, have been growing and accommodating larger and more tightly packed turbines. However, this also increases wake losses and fatigue loads for the downstream turbines and blockage effects for the entire farm. An accurate estimation of these effects is essential for optimally siting a wind farm, maximizing its power production, connecting it reliably to the power grid and minimizing its maintenance costs. The challenge is that wake characterization in wind farms is a multi-faceted problem with considerations such as the steady-state wake-deficit profile, dynamic wake meandering and convective flow instabilities. These phenomena depend on the turbine operating conditions and the fluctuations in the incoming flow. Moreover, individual turbine wakes merge together and interact with geostrophic winds, further complicating the flow. Existing solutions, such as high-fidelity numerical simulations and field experiments, are useful validation tools, but they remain too expensive for commercial applications. Low-order models can accurately capture single turbine wakes in isolation, but existing implementations are mostly kinematic and rely on unphysical wake-merging methods. We propose to develop a low-order field model capable of simulating multiple dynamically evolving wakes in a wind farm without the need for wake-merging methods. This will involve three-steps. First, we will simulate the steady-state wake-deficit and slow-scale passive wake meandering using a quasi-steady simplified form of the Navier—Stokes equations. Second, using the obtained quasi-steady wake, we will compute the fast-scale instabilities using a Ginzburg—Landau model. Third, we will iteratively combine these inter-dependent parts to characterize the full wake, and then validate the results against high-fidelity numerical simulations. On completion, this project will lead to a fast and accurate modelling tool for wake characterization in wind farms.