Abstract: As the global demand for renewable energy increases, wind power, as an important part of clean renewable energy, the accurate prediction of its power is crucial for the stable operation of the power system and the efficient use of energy. In recent years, deep learning methods have demonstrated significant advantages in the field of wind power prediction, and by constructing complex nonlinear models, deep learning models can effectively capture the intrinsic laws and changing trends of wind power data. This paper outlines the research objectives and objectives of wind power prediction from the classification of wind power prediction, the general idea of implementation, and the evaluation method. The application of deep learning technology in wind power prediction is reviewed, and on the basis of making a careful division of deep learning technology, it focuses on analyzing the problems and performance overcome by spatial structure-based deep learning models and time-based deep learning models and their related variants and summarises the limitations of the proposed modeling methods and the corresponding solutions. In addition, research progress in deep learning-based wind power prediction is outlined in data processing, parameter optimization algorithms, and optimization methods for wind power prediction models. Finally, an outlook on the future development direction of wind power prediction is given. This paper aims to provide useful references and inspiration for researchers and practitioners in related fields, promote the development of wind power prediction technology, and contribute to the realization of sustainable energy supply and green growth.

Key words: wind power prediction, neural networks, deep learning

摘要： 随着全球对可再生能源需求的增加，风电作为清洁可再生能源的重要组成部分，其功率的准确预测对于电力系统的稳定运行和能源的高效利用至关重要。近年来，深度学习方法在风功率预测领域展现出显著的优势，通过构建复杂的非线性模型，深度学习模型能够有效地捕捉风功率数据的内在规律和变化趋势。本文从风电功率预测的分类、实现一般思路和评估方法概述了风电功率预测的研究对象和目标。综述了深度学习技术在风功率预测中的应用，在对深度学习技术做出细致的划分的基础上，重点分析了基于空间结构的深度学习模型和基于时间的深度学习模型及其相关变体模型所克服的问题和性能表现，并对所提模型方法存在的局限性及对应解决方法进行总结。此外从数据处理、参数优化算法和风电功率预测模型优化方法三个方面概述了基于深度学习风电功率预测的研究进展。最后，对未来风电功率预测的发展方向进行了展望，本文旨在为相关领域的研究人员和实践者提供有益的参考和启示，推动风电功率预测技术的发展，为实现可持续能源供应和绿色发展贡献力量。

关键词: 风电功率预测, 神经网络, 深度学习