PDF(6369 KB)
Typical Wind-PV-Load Scenario Generation Based on Residual Convolutional Auto-encoders
PENG Yuzheng, LI Xiaolu, LI Congli, DING Yi
Electric Power Construction ›› 2021, Vol. 42 ›› Issue (8) : 10-17.
PDF(6369 KB)
PDF(6369 KB)
Typical Wind-PV-Load Scenario Generation Based on Residual Convolutional Auto-encoders
With the development of new energy sources, higher requirements are put forward for the regulation and operation of grids with PV and wind power. The typical scenario is one of the main ways to deal with this problem. The traditional method for generating typical scenarios is prone to data information loss and feature extraction inaccuracy. This paper proposes an uncertain wind-PV-load typical scenario generation method based on residual convolution auto-encoders. First, the residual convolution auto-encoders network is used to extract the characteristics of wind-PV-load data to reduce the data dimension while reducing the loss of data information and taking into account the coupling of wind and solar power. Then, reducing the influence of noise data on the experimental results, k-medoids is used for clustering to generate typical scenarios. The actual data collected from a power grid in northwest China is taken as the research object. Comparison with traditional clustering methods such as DBI (Davies-Bouldin Index), CHI (Calinski-Harabasz Index) and other indicators, the feasibility of the proposed method is verified.
residual neural network / multi-channel convolutional auto-encoder / wind-PV-load feature extraction / scenario generation
| [1] |
|
| [2] |
王洪涛, 刘旭, 陈之栩, 等. 低碳背景下基于改进通用生成函数法的随机生产模拟[J]. 电网技术, 2013, 37(3): 597-603.
|
| [3] |
罗庆, 晁勤, 王一波, 等. 基于场景划分方法的风光出力耦合特性机理[J]. 电力自动化设备, 2014, 34(8): 42-46.
|
| [4] |
|
| [5] |
林顺富, 田二伟, 符杨, 等. 基于信息熵分段聚合近似和谱聚类的负荷分类方法[J]. 中国电机工程学报, 2017, 37(8): 2242-2253.
|
| [6] |
|
| [7] |
|
| [8] |
石亮缘, 周任军, 张武军, 等. 采用深度学习和多维模糊C均值聚类的负荷分类方法[J]. 电力系统及其自动化学报, 2019, 31(7): 43-50.
|
| [9] |
杨晶显, 刘俊勇, 韩晓言, 等. 基于深度嵌入聚类的水光荷不确定性源场景生成方法[J]. 中国电机工程学报, 2020, 40(22): 7296-7306.
|
| [10] |
申昌, 冀俊忠. 基于双通道卷积神经网络的文本情感分类算法[J]. 模式识别与人工智能, 2018, 31(2): 158-166.
针对现有深度学习方法在文本情感分类任务中特征提取能力方面的不足,提出基于扩展特征和动态池化的双通道卷积神经网络的文本情感分类算法.首先,结合情感词、词性、程度副词、否定词和标点符号等多种影响文本情感倾向的词语特征,形成一个扩展文本特征.然后,把词向量特征与扩展文本特征分别作为卷积神经网络的两个输入通道,采用动态k-max池化策略,提升模型提取特征的能力.在多个标准英文数据集上的文本情感分类实验表明,文中算法的分类性能不仅高于单通道卷积神经网络算法,而且相比一些代表性算法也具有一定的优势.
The existing deep learning method is insufficient to extract features in the text sentiment classification task. To solve the drawback, a text sentiment classification algorithm based on the double channel convolutional neural network with extended features and a dynamic pooling is presented. Firstly, various word features influencing the sentiment orientation of text, such as emotional word, part of speech, adverb of degree, negative word and punctuation, are combined to obtain an extended text feature. Then, the word vector feature and the extended text feature are used as two individual channels of the convolutional neural network, and a new dynamic <i>k</i>-max pooling strategy is adopted to improve the efficiency of feature extraction. The experimental results on standard English datasets demonstrate that the proposed algorithm achieves better classification efficiency than traditional convolutional neural network algorithm with single channel, and it is more advantageous compared with some elitist text sentiment classification algorithms.
|
| [11] |
栗然, 孙帆, 丁星, 等. 考虑多能时空耦合的用户级综合能源系统超短期负荷预测方法[J]. 电网技术, 2020, 44(11): 4121-4134.
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
贾瑞明, 李阳, 李彤, 等. 多层级特征融合结构的单目图像深度估计网络[J]. 计算机工程, 2020, 46(12): 207-214.
|
| [16] |
张斌, 庄池杰, 胡军, 等. 结合降维技术的电力负荷曲线集成聚类算法[J]. 中国电机工程学报, 2015, 35(15): 3741-3749.
|
| [17] |
丁明, 解蛟龙, 刘新宇, 等. 面向风电接纳能力评价的风资源/负荷典型场景集生成方法与应用[J]. 中国电机工程学报, 2016, 36(15): 4064-4072.
|
| [18] |
赵兵, 王增平, 纪维佳, 等. 基于注意力机制的CNN-GRU短期电力负荷预测方法[J]. 电网技术, 2019, 43(12): 4370-4376.
|
| [19] |
|
| [20] |
潘尔生, 宋毅, 原凯, 等. 考虑可再生能源接入的综合能源系统规划评述与展望[J]. 电力建设, 2020, 41(12): 1-13.
|
/
| 〈 |
|
〉 |
