采用正则化极限学习机的短期风速预测

摘要： 高效、准确的风速预测是风电场功率预测的基础,对风力发电控制和风电场并网运行等具有重要意义。针对风速时间序列具有强烈的非线性和波动性,且难以精准预测的特点,提出一种基于正则化极限学习机（regularized extreme learning machine,RELM）的风电场短期风速预测新方法。首先,采用自相关函数（ACF）对风速时间序列的相关性进行分析,得到预测模型输入属性集合;其次,确定预测网络的输入、输出等参数,并建立RELM模型;再次,利用训练集在训练过程中确定网络参数,构建RELM预测模型;最后,以RELM预测模型开展短期风速预测,得出预测结果。采用美国风能技术中心的实测风电场风速数据开展实验证明,相对于标准的ELM和BP神经网络,新方法具有更好的预测精度。     

基于融合长短时记忆网络的风电场超短期风速预测研究

准确的风速预测是风电功率预测的重要基础,对于电力系统的安全、稳定和经济运行有着十分重要的意义。文章通过考虑临近风电场之间的风速时空相关性,提出了一种融合长短时记忆网络的多风电场超短期风速预测模型。首先,通过堆叠的长短时记忆网络提取单个风电场的时间序列特征。之后,通过张量拼接层以及全连接层融合多个风电场的时间序列特征。最后,使用线性全连接层输出所有风电场的未来风速预测值。采用江苏省3个临近风电场两年的数据来验证文章提出的模型。与4种常用方法的对比结果表明:融合长短时记忆网络在四个季节内的超短期风速预测结果均能达到最优;通过序列特征融合的方式可以考虑多个风电场之间的时空相关性。文章提出的时间序列特征提取和空间特征融合方案直观、有效,多个风电场的风速预测精度得到明显提升。     

基于混沌分析和神经网络的风速直接多步预测

针对风速序列的混沌特性,提出了一种将混沌分析和神经网络相结合的短期风速直接多步预测新方法,以提高其预测精度。首先,对风速序列进行混沌特性分析和相空间重构;然后,根据重构相空间的特征参数,结合预测需求,确定Elman网络结构;最后,利用空间欧式距离选取的样本对Elman网络进行训练,建立风速直接多步预测模型。以华北地区某风电场实测风速为例进行仿真测试,结果表明与单步迭代法和直接神经网络法相比,该文方法在进行风速直接多步预测时具有更好的整体误差指标。     

基于小波变换与Elman神经网络的短期风速组合预测

风速的准确预测对风电场发电系统的经济和安全运行有着重要的作用。为了克服风速随机性强的缺点,提高短期风速预测的精度,提出了一种将小波变换与Elman神经网络相结合的短期风速组合预测模型。该模型由小波预处理模块和神经网络预测模块组成。首先利用小波预处理模块将风速序列作多尺度分解,重构得到不同频段的子序列,然后利用Elman神经网络模块分别对其训练和预测。实际风速预测结果表明,与单一的Elman和ARMA法相比,该组合预测模型的预测精度有较大的改善,可以用于风电场短期风速的预测。     

基于特征选择的RELM风速短期预测

准确的风速预测是风电场功率预测的基础,对大规模风电并网具有重要的价值。文章提出一种基于信息增益(IG)的正则化极限学习机(RELM)短期风速预测方法。首先采用信息增益对32维风速属性序列进行特征选择,并对其进行加权;然后将正则化系数引入极限学习机(ELM)网络,构建RELM风速预测模型;最后结合美国风能技术中心的实测数据进行仿真,与传统ELM网络、BP神经网络相比,该方法具有较高的准确性和预测精度。     

基于优化近邻传播聚类的CMN风速预测

提出一种基于快速相关性约简和近邻传播聚类的卷积记忆网络短期风速预测模型。计算各风速序列及其属性序列的相关程度信息熵,运用快速相关性滤波算法进行属性约简,以降低属性维度及删除冗余属性;针对风速属性矩阵样本,采用压缩-激励模块（squeeze-and-excitation networks,SENet）构建属性表征序列,以该序列间距为样本相似度,利用近邻传播聚类实现样本集优选重构;构建卷积记忆网络,利用其挖掘深层特征及短期预测。通过对实际风场风速进行预测,对比实测数据,结果表明,该方法在风速属性数据的优选方面具有较大优势,通过保留关联紧密的属性信息,提高了预测精度。     

基于LMD-IMVO-LSSVM的短期风速预测

风速信号具有的随机性和波动性的特点给风速预测的准确性带来了巨大挑战。现有的风速预测方法较多,但大都难以满足风电场需求的预测效果。文章提出了一种基于LMD-IMVO-LSSVM的短期风速预测方法。首先采用局部均值分解(LMD)方法将原始风速序列分解为若干个平稳的风速子序列,结合改进多元宇宙优化算法(IMVO)寻优最小二乘支持向量机(LSSVM)的可调参数预测方法,建立了LMD-IMVO-LSSVM的风速预测组合模型;然后对分解得到的每个平稳子序列进行单独的预测,叠加各子序列预测结果,即得到最终的风速预测值。通过实验仿真分析得出,文章提出的组合预测模型可大大提高风速预测的准确性。     

采用时间序列预测风电场出力

基于时间序列的方法,采用自回归滑动平均(ARMA)模型进行短期风速预测;考虑风力发电机组排列布置时尾流效应的影响、风电场址地面粗糙程度、空气密度、风向变化以及不同型号风机功率特性的差异等因素,采用Jasen尾流模型建立了大型风电场的综合模型。结果表明,合理的风电场布置方案有利于减小尾流效应的影响,从而提高风电场出力。     

基于PSO优化LS-SVM的短期风速预测

提出了一种基于粒子群(PSO)算法优化最小二乘支持向量机(LS-SVM)的风电场风速预测方法。以相关性较高的历史风速序列作为输入,建立预测模型,并用粒子群算法优化模型参数。在对未来1 h风速进行预测时,文章所提出的模型比最小二乘支持向量机模型及BP神经网络模型具有较高的预测精度和运算速度。算例结果表明,经粒子群优化的最小二乘支持向量机算法是进行短期风速预测的有效方法。     

基于小波变换和LS-SVM的短期风速预测方法

针对风速序列的周期性和非平稳性,提出了基于小波变换和LS-SVM相结合的风电场风速预测模型,利用小波变换的多分辩分析法对风速序列进行分解,将风速序列投影到不同尺度上.结合LS-SVM的小样本学习能力强和计算简单等特点,将小波分解后的各风速子序列分别采用LS-SVM进行训练和预测,最后将各预测结果进行叠加得到最终的风速预测值.与LS-SVM风速预测方法进行比较,采用该文提出的方法可明显提高短期风速预测的精度,并具有较强的适应性,具有一定的工程应用前景.最后通过具体实例验证了该模型的有效性.     

Very short-term wind speed forecasting with Bayesian structural break model

This paper examines a new time series method for very short-term wind speed forecasting. The time series forecasting model is based on Bayesian theory and structural break modeling, which could incorporate domain knowledge about wind speed as a prior. Besides this Bayesian structural break model predicts wind speed as a set of possible values, which is different from classical time series model's single-value prediction This set of predicted values could be used for various applications, such as wind turbine predictive control, wind power scheduling. The proposed model is tested with actual wind speed data collected from utility-scale wind turbines.     

Short-term wind power prediction based on stacked denoised auto-encoder deep learning and multi-level transfer learning

Wind power prediction (WPP) has an important impact on the security and reliability operations of the power grid. The major difficulty in power prediction of new, expanded, or reconstructed wind farms is the lack of operational data, which leads to insufficient training of the model and makes the prediction error of wind power become enormous. A short-term WPP model based on stacked denoised auto-encoder (SDAE) deep learning and multilevel transfer learning is proposed in this paper. First, the correlation coefficient between the samples of source wind farms and the target wind farm is calculated by using a network trained with the samples from the target wind farm. Second, the samples with high correlation coefficients in source wind farms are graded and migrated to the target wind farm to assist multilevel transfer learning. Finally, the samples from different grades are each used to train a layer of SDAE, and their weights and thresholds are migrated to the final network. The case study shows that the 24-h-day-ahead normalized root-mean-square error (NRMSE) and 96-h-short-term NRMSE obtained by the proposed method are 4.48% and 5.11% lower, respectively, compared with the model without transfer learning, which proves the effectiveness of the proposed model.     

Influence of neural network training parameters on short-term wind forecasting

This paper investigates factors which can affect the accuracy of short-term wind speed prediction when done over long periods spanning different seasons. Two types of neural networks (NNs) are used to forecast power generated via specific horizontal axis wind turbines. Meteorological data used are for a specific Western Australian location. Results reveal that seasonal variations affect the prediction accuracy of the wind resource, but the magnitude of this influence strongly depends on the details of the NN deployed. Factors investigated include the span of the time series needed to initially train the networks, the temporal resolution of these data, the length of training pattern within the overall span which is used to implement the predictions and whether the inclusion of solar irradiance data can appreciably affect wind speed prediction accuracy. There appears to be a relatively complex relationship between these factors and the accuracy of wind speed prediction via NNs. Predicting wind speed based on NNs trained using wind speed and solar irradiance data also increases the prediction accuracy of wind power generated, as can the type of network selected.     

运用广义回归神经网络预测风电场功率

运用广义回归神经网络对风电场出力提前了24h预测。对引入数值气象预报信息与不引人数值气象预报信息两种情况的预测结果进行了比较分析。首先,对前15d的风功率数据进行训练,通过交叉验证,建立模型,预测了未来一天的风电场出力。然后加入历史风速数据,对历史风速和风功率进行训练,利用数值气象预报信息,预测未来1d的风功率。通过算例表明,使用广义回归神经网络模型预测未来1d的风电场出力,预测结果能够跟踪实际风功率,同时加入数值气象预报信息的预测结果较不加入数值气象预报信息的神经网络预测,精度有所提高。     

基于熵权时序模型的超短期风电功率预测

针对风电具有较强的随机性和波动性,传统的单一预测方法难以准确描述其规律且预测精度较低的问题,提出风速熵和功率熵的概念,在时间序列法的基础上分别采用基于风速和基于功率的预测方法,并根据风速熵和功率熵的计算结果动态设置预测点的权值,建立风电功率的熵权时序模型。算例分析结果表明,所提方法能有效提取风速及功率历史数据中的有用信息,提高超短期风电功率预测精度,预测结果的准确率和合格率均优于神经网络法、时间序列法和基于风速法。     

基于Elman神经网络的短期风电功率预测

为提高风电场输出功率预测精度,提出一种动态基于神经网络的功率预测方法。根据实际运行的风电场相关风速、相关风向和风电功率的历史数据,建立了基于Elman神经元网络的短期风电功率预测模型。运用多层Elman神经网络模型对西北某风电场实际1h和24h的风电输出功率预测,与BP神经网络模型对比,经仿真分析证明前者具有预测精度高的特点,三隐含层Elman神经网络模型预测效果最佳。这表明利用Elman回归神经网络建模对风电功率进行预测是可行的,能有效提高功率预测精度。     

Transfer learning for short-term wind speed prediction with deep neural networks

As a type of clean and renewable energy source, wind power is widely used. However, owing to the uncertainty of wind speed, it is essential to build an accurate forecasting model for large-scale wind power penetration. Numerical weather prediction (NWP) and data-driven modeling are two typical paradigms. NWP is usually unavailable or spatially insufficient. Data-driven modeling is an effective candidate. As to some newly-built wind farms, sufficient historical data is not available for training an accurate model, while some older wind farms may have long-term wind speed records. A question arises regarding whether the prediction model trained by data coming from older farms is also effective for a newly-built farm. In this paper, we propose an interesting trial of transferring the information obtained from data-rich farms to a newly-built farm. It is well known that deep learning can extract a high-level representation of raw data. We introduce deep neural networks, trained by data from data-rich farms, to extract wind speed patterns, and then finely tune the mapping with data coming from newly-built farms. In this way, the trained network transfers information from one farm to another. The experimental results show that prediction errors are significantly reduced using the proposed technique.     

基于深度置信网络的短期风电功率预测

为解决海量数据用作预测模型训练样本导致信息冗杂的问题,提出一种基于深度置信网络的短期风电功率预测方法.该方法首先使用历史数据作为训练样本,通过深度置信网络无监督学习提取出其相应特征,随后采用K均值算法对提取出的特征进行聚类分析,将历史数据分作几类,并通过判别分析确定待测日所属类别,以该类别所属的历史数据对设置了误差反馈...     

考虑历史相似性加权的超短期风电功率组合预测

提出一种基于历史相似性加权的超短期风电功率组合预测方法。首先,采用数值天气预报数据、风电历史数据分别作为极限学习机、长短期记忆网络的输入特征并产生预测数据;然后,通过加权灰色关联算法提取与待预测点特征近似的历史数据,评估并校正两类预测模型的预测结果。采用美国科罗拉多州某风电场实际运行数据进行训练与验证,并使用不同加权方法进行对比。结果表明,基于历史相似性优化方法可改善预测效果,缩小预测误差分布范围,验证了该文方法的有效性。