| 引用本文: | 牛哲文,余泽远,李波,唐文虎.基于深度门控循环单元神经网络的短期风功率预测模型[J].电力自动化设备,2018,(5): |
| NIU Zhewen,YU Zeyuan,LI Bo,TANG Wenhu.Short-term wind power forecasting model based on deep gated recurrent unit neural network[J].Electric Power Automation Equipment,2018,(5): |
|
| 摘要: |
| 随着新能源的不断发展，大量大容量风电机组并入电网运行，给电网的安全可靠运行以及风力发电的可持续发展都提出了新的挑战。提出一种风功率预测模型，该模型以风电场风功率历史数据以及风速、风向等数值天气预报数据作为输入对风功率进行预测。考虑到风功率预测中输入数据的波动性和不确定性，在传统门控循环单元(GRU)神经网络的基础上融合卷积神经网络(CNN)，以提高模型对原始数据的特征提取和降维能力，并引入dropout技术减少模型中的过拟合现象。工程实例分析表明，所提模型在预测准确度和运算速度方面均优于长短记忆神经网络模型。 |
| 关键词: 风功率预测 深度神经网络 门控循环单元 卷积神经网络 |
| DOI：10.16081/j.issn.1006-6047.2018.05.005 |
| 分类号:TM614 |
| 基金项目:国家自然科学基金资助项目(51477054)；国家高技术研究发展计划(863计划)项目(2015AA050201) |
|
| Short-term wind power forecasting model based on deep gated recurrent unit neural network |
|
NIU Zhewen, YU Zeyuan, LI Bo, TANG Wenhu
|
|
School of Electric Power, South China University of Technology, Guangzhou 510641, China
|
| Abstract: |
| With the development of renewable energy, a large number of wind turbines with high capacity have been connected to power grids, bringing new challenges to the safe and reliable operation of power grids and the sustai-nable development of wind power generation. A wind power forecasting model is proposed, which combines the historical data of wind power and the data of numerical weather prediction such as wind speed and wind direction as the inputs. Because of the fluctuation and uncertainty of the input data, CNN(Convolutional Neural Network)is combined with the traditional GRU(Gated Recurrent Unit) neural network to improve the ability of the model for feature extraction and dimensionality reduction of the original data. The dropout technology is introduced to reduce the over-fitting phenomenon. The analysis of engineering example shows that, the proposed model is superior to the long short-term memory neural network model in the aspects of forecasting accuracy and computational speed. |
| Key words: wind power forecasting deep neural network gated recurrent unit convolutional neural network |
