基于组合预测方法的风电场短期风速预测

基于持续法、人工神经网络法(ANN)和支持向量机(SVM)3种不同预测模型对内蒙古某风电场短期风速进行了预测研究,比较了不同单一预测模型的预测精度,并进行了4种不同预测模型的组合预测。计算结果表明,单一预测模型中支持向量机方法精度最高,而组合预测中3种方法组合的预测精度最高,并且组合预测精度均高于单一预测方法的精度。同时发现,当单一模型预测误差之间存在较强的负相关关系时,组合预测精度提高明显;而当单一模型预测误差之间存在较强的正相关关系时,则组合预测精度改进有限。     

基于灰色GMDH网络组合模型的风速预测

风电场风速具有较大的间歇性和波动性,其预测精度有待提高。针对这一问题,文章基于灰色系统理论与GMDH数据分组处理技术,建立了灰色GMDH网络组合风速预测模型;采用某风电场实测风速数据进行预测,并与灰色Verhulst模型预测方法和传统GMDH网络模型预测方法的预测结果进行了对比。算例结果表明,灰色GMDH网络组合风速预测模型能够更精确地预测风速,显示了其可行性与高效性。     

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。     

考虑气象因子的海上短期风速预测

提出一种基于气象因子的双径向基神经网络组合风速预测模型。以气温、水温、湿度等对海上风速影响较大的气象因子构成输入矢量,并通过带加权系数的模式识别方法对历史样本进行筛选,分别从时间连续规律与周期规律方向采用径向基神经网络(RBF)进行风速预测,最终通过径向基神经网络(RBF)组合预测模型得到待测日风速的预测值。对美国新泽西东北部某风电场的风速预测结果表明,该模型具有较高的预测精度与较好的预测平稳度,具有一定的实际应用价值。     

基于GRA-BPNN时变权重的光伏短期出力组合预测

针对并网情况下光伏出力预测精度低和稳定性差问题,提出了一种基于灰色关联分析(GRA)结合BP神经网络(GRA-BPNN)的变权重系数组合预测模型。首先,利用3种单一预测模型对光伏出力分别进行预测,然后,利用GRA-BPNN模型对3个单一模型不同时刻的权重系数进行预测,最后,根据权重系数计算出预测结果。文章利用武汉某并网光伏电站的实测数据对GRA-BPNN变权重组合预测模型预测结果的准确性进行检验。分析结果表明:GRA-BPNN变权重组合预测模型的相对均方根误差和相对平均误差均低于单一模型和等权重组合模型;根据各预测模型的残差直方图可知,GRA-BPNN变权重组合预测模型预测结果中出现较大残差的概率很小,有效地解决了单一模型预测结果不稳定的问题。     

基于支持向量机的风速预测模型研究

由于风速的随机性很大,风速大小的影响因素较多,风速预测的准确度不高.针对这种现象,该文基于支持向量机(SVM)理论,结合风速资料,建立支持向量机(SVM)预测模型来进行短期的风速预测,由支持向量机预测模型得到的预测风速与实际风速基本一致,预测效果较理想,预测的平均绝对百分比误差为10.07%,验证了支持向量机预测模型在风速短期预测中的可行性.     

基于优化聚类的组合风速短期预测

准确的风功率预测对电力系统安全、稳定运行具有重要意义,而风速预测是风功率预测的关键。文章提出一种基于优化模糊C均值(Optimal Fuzzy C means,OFCM)聚类的组合风速短期预测方法。首先,采用模拟退火遗传算法优化模糊C均值聚类算法的初始聚类中心;其次,基于优化模糊C均值聚类算法将初始风速属性样本数据进行分组;再根据不同风速样本组,运用极限学习机(Extremely Learning Machine,ELM)构建组合风速预测模型;最后,通过风速实测值与预测值的对比,验证了该方法的可行性。     

基于ARMA-ARCH模型和BP模型的短期风速组合预测方法研究

随着风力发电的大规模并网,由风速的波动引起的网侧不稳定现象越来越显著。为了提高风电场风速预测的精度,首先建立了ARMA模型,利用拉格朗日乘数法检验ARMA模型残差的条件异方差效应,从而建立ARMA-ARCH模型;其次建立BP神经网络预测模型;最后分别以固定权和时变权方差-协方差（MV）法将ARMA-ARCH模型和BP模型进行优选组合预测。为验证模型的适应性,分别以西班牙某风电场2016年8月和2017年1月的风速数据进行建模仿真。仿真结果表明：组合预测模型的预测结果更优,且时变权组合预测精度更高;对于单一模型来说ARMA-ARCH模型的预测精度要高于BP模型,而ARMA模型的预测精度最低。     

基于风速误差校正和ALO-LSSVM的风电功率预测

风资源的随机波动性引起的相位滞后性问题,导致风电功率预测精度不高,尤其是风速变化较快时,滞后性引起的预测误差较大。考虑到风速波动与风功率变化密切相关,提出一种非参数核密度估计和数值天气预报(NWP)相结合的方法,并对预测风速误差进行校正,改善了预测风速的相位滞后性;然后将校正后的风速和风功率作为输入数据进行风电功率预测;采用蚁狮算法(ALO)优化最小二乘支持向量机(LSSVM)参数,从而建立基于风速误差校正和ALO-LSSVM组合的风电功率预测模型。算例结果表明,所提方法风功率预测精度更高。     

基于小波分解和极限学习机的短期风速组合预测研究

提出一种基于小波分解(Wavelet Decomposition,WD)和极限学习机(Extreme Learning Machine,ELM)的新型短期风速组合预测模型。首先,采用小波分解将风速序列分解成不同频段的分量,以降低序列的非平稳性;其次,为避免极限学习机输入维数选取的随意性等问题,先对各分量进行重构相空间,再使用改进的极限学习机对各分量分别建模预测;最后,将各分量预测结果叠加得到最终预测结果。实验结果表明,文章所提的组合预测模型具有较高的预测精度。     

基于相似性样本的LSSVM短期风速和风功率预测研究

风速具有较大的随机波动性,影响电网的稳定性,良好的风速预测是解决风电并网问题的关键。为了提高风速预测的精确性,首先对风速数据进行相似性样本的提取,采用分段线性化的搜索方法,求出各小段风速的斜率与长度所占的比重,继而找出相似性距离最小的曲线簇。并以此作为训练样本,采用最小二乘支持向量机（LSSVM）模型对风速进行预测。预测结果表明,采用风速的相似曲线簇进行LSSVM模型训练所得的风速和风电功率预测结果更优。     

基于卷积和共享权系数长短时记忆网络时融合机制在风速预测上的应用研究

针对对于风能规划和应用都具有重大影响的风速存在强随机性问题,该文提出结合卷积神经网络（CNN）和共享权重长短期记忆网络（SWLSTM）的空时融合模型（CSWLSTM）,充分提取风速序列中蕴含的空域和时域信息,以提升预测精度。此外,为了获得可靠的风速概率预测结果,提出一种新的结合CNN、SWLSTM和高斯过程回归（GPR）的混合模型,称为 CSWLSTM-GPR。将CSWLSTM-GPR应用于中国内蒙古风速预测案例,从点预测精度、区间预测适用性和概率预测综合性能3个方面与相同结构的CNN和SWLSTM模型的风速预测方法进行比较。CSWLSTM-GPR的可靠性测试保证了预测结果的可靠性和说服力。实验结果表明,CSWLSTM-GPR在风速预测问题上能获得高精度的点预测、合适的预测区间和可靠的概率预测结果,也充分展现了该研究所提出CSWLSTM在风速预测方面具有较好的应用潜力。     

Development of a comprehensive MPPT for grid‐connected wind turbine driven PMSG

This paper proposes a comprehensive MPPT method by which extraction of maximum power from wind turbine and its subsequent transfer through various power stages and final delivery to the connected grid are realized. In the proposed system, the operation of the wind turbine at its maximum efficiency point is maintained by control of grid‐tied inverter such that the shaft speed of the generator is set to result the desired optimum tip speed ratio of the turbine. The proposed comprehensive MPPT estimates the required DC link voltage for each wind speed using a unified system model, uses a loss factor to account for the system losses, and then controls the inverter to push the WT extracted maximum power into the grid. The comprehensive MPPT is developed and is validated in MATLAB/Simulink platform in a wide range of operating wind speed. The results ascertain that the wind turbine is made to operate at its maximum efficiency point for all wind speeds below the rated one.     

基于数据分层预处理的短期风功率预测研究

良好的风速和风功率预测是解决风电并网问题的关键。针对样本数据中的无效点影响风功率建模问题,采用分层统计法对风功率进行统计分析后获得了风速—功率关系带,对功率进行修正,根据修正后的数据应用灰色—马尔可夫链模型进行预测,并与比恩法和经验公式法进行对比分析。结果表明,风功率分层统计法可有效地消除坏点数据,预测精度高。     

Characterizing future large,rapid changes in aggregated wind power using Numerical Weather Prediction spatial fields

A critical limiting factor to the successful deployment of a large proportion of wind power in power systems is its predictability. Power system operators play a vital role in maintaining system security, and this task is greatly aided by useful characterizations of future system operations. A wind farm power forecast generally relies on the forecast output from a Numerical Weather Prediction (NWP) model, typically at a single grid point in the model to represent the wind farm's physical location. A key limitation of this approach is the spatial misplacement of weather features often found in NWP forecasts. This paper presents a methodology to display wind forecast information from multiple grid points at hub height around the wind farm location. If the raw forecast wind speeds at hub height at multiple grid points were to be displayed directly, they would be misleading as the NWP outputs take account of the estimated local surface roughness and terrain at each grid point. Hence, the methodology includes a transformation of the wind speed at each grid point to an equivalent value that represents the surface roughness and terrain at the chosen single grid point for the wind farm site. The chosen‐grid‐point‐equivalent wind speeds for the wind farm can then be transformed to available wind farm power. The result is a visually‐based decision support tool which can help the forecast user to assess the possibilities of large, rapid changes in available wind power from wind farms. A number of methods for displaying the field for multiple wind farms are discussed. The chosen‐grid‐point‐equivalent transformation also has other potential applications in wind power forecasting such as assessing deterministic forecast uncertainty and improving downscaling results. Copyright © 2008 John Wiley & Sons, Ltd.     

基于奇异值分解与卡尔曼滤波修正多位置NWP的短期风电功率预测

考虑到数值天气预报网格点位置和系统误差对短期风电功率预测精度的影响,提出一种基于奇异值分解与卡尔曼滤波修正多位置数值天气预报的短期风电功率预测模型。首先通过奇异值分解对多位置数值天气预报数据进行特征提取与降维处理;然后使用卡尔曼滤波方法修正数值天气预报风速数据,降低数值天气预报的系统误差;最后基于极端随机森林算法,利用修正的数值天气预报数据搭建短期风电功率预测模型。通过对某风电场进行仿真,并与单位置、未降维、未修正模型比较,结果表明降维修正模型的预测效果最好,平均误差和均方根误差分别为7.94%和9.96%。     

ENDOW (efficient development of offshore wind farms): modelling wake and boundary layer interactions

While experience gained through the offshore wind energy projects currently operating is valuable, a major uncertainty in estimating power production lies in the prediction of the dynamic links between the atmosphere and wind turbines in offshore regimes. The objective of the ENDOW project was to evaluate, enhance and interface wake and boundary layer models for utilization offshore. The project resulted in a significant advance in the state of the art in both wake and marine boundary layer models, leading to improved prediction of wind speed and turbulence profiles within large offshore wind farms. Use of new databases from existing offshore wind farms and detailed wake profiles collected using sodar provided a unique opportunity to undertake the first comprehensive evaluation of wake models in the offshore environment. The results of wake model performance in different wind speed, stability and roughness conditions relative to observations provided criteria for their improvement. Mesoscale model simulations were used to evaluate the impact of thermal flows, roughness and topography on offshore wind speeds. The model hierarchy developed under ENDOW forms the basis of design tools for use by wind energy developers and turbine manufacturers to optimize power output from offshore wind farms through minimized wake effects and optimal grid connections. The design tools are being built onto existing regional‐scale models and wind farm design software which was developed with EU funding and is in use currently by wind energy developers. Copyright © 2004 John Wiley & Sons, Ltd.     

提前三小时风电机组出力组合预报研究

基于持续法和神经网络法对提前3h的风电场风速和风电机组出力进行了组合预报研究,提出误差衡量指标的选取将影响组合预报的结果;由于功率曲线的形状,风速组合预报误差最小时各种方法的权系数与功率组合预报误差最小时各种方法的权系数并不相同。计算结果表明,选取合适的误差衡量指标,以功率的最小预报误差为目标的组合预报可以提高风电机组出力的预报精度。     

Wind speed prediction in the mountainous region of India using an artificial neural network model

Measured wind speed data are not available for most sites in the mountainous regions of India. The objective of present study is to predict wind speeds for 11 locations in the Western Himalayan Indian state of Himachal Pradesh to identify possible wind energy applications. An artificial neural network (ANN) model is used to predict wind speeds using measured wind data of Hamirpur location for training and testing. Temperature, air pressure, solar radiation and altitude are taken as inputs for the ANN model to predict daily mean wind speeds. Mean absolute percentage error (MAPE) and correlation coefficient between the predicted and measured wind speeds are found to be 4.55% and 0.98 respectively. Predicted wind speeds are found to range from 1.27 to 3.78 m/s for Bilaspur, Chamba, Kangra, Kinnaur, Kullu, Keylong, Mandi, Shimla, Sirmaur, Solan and Una locations. A micro-wind turbine is used to assess the wind power generated at these locations which is found to vary from 773.61 W to 5329.76 W which is suitable for small lighting applications. Model is validated by predicting wind speeds for Gurgaon city for which measured data are available with MAPE 6.489% and correlation coefficient 0.99 showing high prediction accuracy of the developed ANN Model.     

风电场风速及风电功率预测方法研究综述

由于风能的随机性以及电力系统的非线性等原因,预测风电功率时需要考虑众多的不确定因素影响。现有预测方法主要包括物理预测方法、统计预测方法以及学习预测方法、综合预测法等。基于数字天气预报(NWP-numerical weather prediction)的物理预测方法模型复杂、计算量大,较少用于短期预测;统计预测方法模型简单,数据需求量少,较适合于数据获取有一定困难的情况;人工智能预测方法不依赖于对象的精确模型,适合于随机非线性系统;综合预测方法可一定程度地扬长避短。本文主要就风电场风速及风电功率预测方法研究进行了综合阐述,并在总结前人研究的基础上提出了一些可进一步研究的问题。