基于BiLSTM和Bootstrap方法的风电功率区间预测

风能具有较强的波动性和随机性,造成了现有风电功率预测方法的误差较大,严重影响了电力系统的安全稳定性。针对上述问题,文章提出一种基于双向长短期记忆网络(BiLSTM)和自助法(Bootstrap)的风电功率区间预测方法。该方法强调了风电功率数据间关联程度的重要性,将原始风电功率数据和测风塔提供的风速历史数据构成多变量时间序列,同时构建双向长短期记忆网络模型提高预测结果的精确度。引入Bootstrap方法增加样本的多样性,再利用人工蜂群算法(ABC)的强搜索能力对模型的超参数进行优化,最终得到区间预测的结果。以某风电场历史运行数据为例,通过与长短期记忆网络(LSTM)等现有方法在给定置信水平下的预测结果进行对比,验证了所提方法的有效性。     

西北电网区域（酒泉）调度测风网络规划研究

针对甘肃酒泉千万千瓦级风电基地的风电场规划．在大量实地查勘的基础上,根据各个风电场的面积大小、地形状况、开发规模,并充分考虑风电功率预测的需求以及测风塔实地建设的可行性,采用统计与理论分析、流体力学模型模拟典型风速分布等方法,提出合理的测风网络布局方案。为实现高精度、大区域范围的风能资源评估和风电功率预测提供技术支持,服务于风电大规模开发和上网输送。     

风电场缺失测风数据插补方法的分析

在对风电场进行风资源评估时,常需要对缺失测风数据进行插补,然后据此订正成代表年风速。文章以广东地区某风电场风资源分析为例,采用以下3种不同的线性相关插补方法 :方法一,相关关系构建基于主测风塔和参照测风塔同期所有的测风数据(除去缺测数据);方法二,相关关系构建基于不同季节的测风数据;方法三,相关关系构建基于不同风向扇区(一般16个)的测风数据。误差分析结果表明,方法三误差最小,方法二次之,方法一误差最大。     

风电场代表年数据订正方法的不确定性分析

选取陕西某复杂地形风电场测风塔2009~2011年3年逐时测风数据,利用16方位相关法计算得到该风电场测风塔代表年数据,利用该方法进行测风塔代表年数据订正过程中的不确定性,并对产生不确定性的原因进行分析。结果表明:若测风塔观测时段不同,得到的代表年数据也存在差异,说明测风塔和气象站在不同年份各风向风速差异上的不一致必然会对代表年数据计算造成一定的影响。另外,测风塔和气象站风速在月尺度和日尺度上的差异、风向变化上的差异也会导致测风塔代表年数据存在误差。利用MM5/CALMET模式提取测风塔所在位置2009~2011年3年逐时风速风向数据,将数值模拟结果、16方位相关法计算的代表年结果分别与测风塔实测数据进行对比,发现数值模拟结果误差明显小于用16方位相关法得到的代表年数据。因此认为在气象站数据与风电场测风塔数据风速风向变化一致性较差的情况下,可选择数值模拟技术对测风塔长期风况进行模拟,得到测风塔长期风况,较为准确客观地评价风电场的风能资源。     

基于FCM-等值风速模型的风功率日前预测

由于风速的随机性、间歇性,以及风电场内各机组风速、功率的分散性,给风功率预测带来了较大难度。在计算风速线性相关的权值基础上,提出了改进模糊C均值聚类算法（fuzzy c-means,FCM）的风速模型,建立了风电场等值风速与改进FCM风速的关系函数。以某风电场实测数据进行验证,结果表明：所提风电功率预测方法算法简单;该方法预测精度提高了71.35%。在该风电场不同日周期下,验证了所提预测方法的有效性和普适性。     

酒泉地区风电场风电功率预报研究

利用NOAA天气预报模式Weather Research andForecasting Model(WRF)结合统计订正方法对酒泉地区短期风电功率预报进行了预报实验。与实际出力比较24 h短期风电功率预报精度较高。并在此基础上利用风电场附近测风塔观测数据通过时间序列发进行了0~4 h超短期预报实验,预报结果显示0~2 h预报结果有利于运行调度。     

弃风限电条件下复杂地形风电场短期风功率预测对比分析

为提高弃风限电条件下风电场短期风功率预测的准确率,选取甘肃及湖北位于不同条件下5个风电场2015年5月份和9月份的数据进行对比分析。利用BJ-RUC耦合CALMET模式模拟风电场风速资料,通过线性滚动订正方法对模式资料进行订正预处理,分别采用物理法、动力统计法、集合预测法和自适应偏最小二乘回归法探讨不同条件下风电场的预报效果。结果表明:1)线性滚动订正方法能够有效降低数值预报风速误差,其中以时间步长为1 d订正效果最佳;2)动力统计法在风电场弃风率低于50%时优势明显;3)物理法3在未限电地区表现较优,可见采用实测数据建立的风功率预报模型能较好地预测未限电地区的风电功率;4)集合预测法和自适应偏最小二乘回归法可有效提高弃风限电地区的短期预测准确率,准确率最高可提升约30%;5)针对不同条件的风电场,从预报效果和普适性的角度来看,集合预测法在所有方法中稳定性最佳。     

基于风电场等效平均风速的风电功率日前预测研究

提出风电机组风速模型及风电场等效平均风速模型,并由此得到基于风电场等效平均风速的风电功率预测方法,与以往基于平均风速的预测方法进行工程应用效果对比,该方法较基于平均风速的预测方法在预测精度方面有明显提高,且模型简易可行,具有通用性。     

基于Attention-GRU风速修正和Stacking的短期风电功率预测

针对使用数值天气预报（NWP）数据进行风电功率预测时,NWP风速与实际风速存在偏差导致预测精度欠佳,提出一种基于注意力机制（Attenion）门控逻辑单元（GRU）数值天气预报风速修正和Stacking多算法融合的短期风电功率预测模型。首先,分析NWP预报风速和实际风速的皮尔逊相关系数,建立Attention-GRU风速修正模型,提高预报风速精度。其次,考虑风向、温度、湿度、气压、空气密度等气象因素,基于Stacking框架,提出融合XGBoost、LSTM、SVR、LASSO的多算法风电功率预测模型,同时采用网格搜索与交叉验证优化模型参数。最后,选取西北和东北两个典型风电场数据进行验证,算例结果表明,所提出模型能改善NWP风速精度并提升风电功率预测效果。     

基于小波分析和PSO优化神经网络的短期风电功率预测

针对风电场风速和风电功率序列起伏波动大、无明显变化规律等特点以及传统神经网络收敛速度慢、易陷入局部极小值等缺陷,提出了基于小波分析和改进粒子群算法优化神经网络的短期风电功率预测方法。首先,通过小波方法将用于神经网络训练的历史风速和风电功率序列进行分解,再针对风速和风电功率的各个分量分别建立相应的神经网络模型,采用分期变异粒子群算法对各个分量的神经网络学习算法进行优化,最后将各个分量的预测值进行小波重构得到风电功率预测结果。江苏如东某风电场风电机组的实验结果证明预测精度较传统神经网络方法有较大提高,验证了所提出方法的有效性。     

风电场风电机组的接地设计

较系统地介绍了风电场风电机组对接地电阻的要求、接地设计思路及方法,并提供实际工程中接地网布置图实例作为参考。     

风 风能 风力发电——21世纪新型清洁能源

一风的一般属性１风的形成风是人们非常熟悉的一种自然现象，人人都能感觉到它的存在。春风和煦，给万物带来生机；夏风吹拂，使人心旷神怡；秋风送爽，带来丰收的喜悦；冬风呼啸，迎来漫天飞雪。那么风是怎样形成的呢？众所周知，人类生活的地球表面被大气所包围，来自太阳的辐射不断传送到地球表面，因太阳辐射受热情况不同，地球表面各处的气温不同。在影响气压高低的因素中，气温起着最重要的作用。温度高的地区空气受热上升，气压减小；温度低的地方，空气下降，气压增大，于是产生了气压差。和水往低处流一样，空气也从气压高处向气压…     

Evaluation of wind farm efficiency and wind turbine wakes at the Nysted offshore wind farm

Here, we quantify relationships between wind farm efficiency and wind speed, direction, turbulence and atmospheric stability using power output from the large offshore wind farm at Nysted in Denmark. Wake losses are, as expected, most strongly related to wind speed variations through the turbine thrust coefficient; with direction, atmospheric stability and turbulence as important second order effects. While the wind farm efficiency is highly dependent on the distribution of wind speeds and wind direction, it is shown that the impact of turbine spacing on wake losses and turbine efficiency can be quantified, albeit with relatively large uncertainty due to stochastic effects in the data. There is evidence of the ‘deep array effect’ in that wake losses in the centre of the wind farm are under‐estimated by the wind farm model WAsP, although overall efficiency of the wind farm is well predicted due to compensating edge effects. Copyright © 2010 John Wiley & Sons, Ltd.     

Dynamic wind estimation based control for small wind turbines

We introduce a novel scheme for small wind turbines that gives dynamic estimation of wind speed from rotor angular velocity measurements. The estimation proceeds in two different dynamic observers, one giving a valid estimate for higher Tip Speed Ratios (TSRs) and which we call the Upper Wind Estimator (UWE) and the other called the Lower Wind Estimator (LWE) valid for lower TSRs. The meaning of “higher” and “lower”, and the precise regions of validity, are quantified. We further propose a coordinated control scheme using the UWE. Simulations are presented showing closed-loop performance of the turbine and the estimators both in the optimal TSR regulation condition, and the dynamic power-shedding condition caused by a wind gust. An analytic analysis of closed-loop stability and of the convergence and bias properties of the estimator is provided. Empirical data showing performance on a real turbine is also presented.     

分散式风力发电机组的电压控制

风电的分散式开发不同于大规模开发和分布式开发,由于分散风电靠近负荷中心,直接接入配电网,且不加装无功补偿调节装置SVC,配网中较大的电压波动给分散风电的并网运行带来影响。文章讨论了配网对分散风电的电压控制特点和要求,结合风电机组无功控制能力,并推导出满足配网电压调节要求的风电机组无功控制范围和对机组设备的要求。     

适用于特殊风区的风电机组塔架的再设计

由于IECⅠ级机型性能不能满足超Ⅰ类风区的要求,因此以IECⅠ级机型塔架的结构为基础,对其进行了再设计,并利用有限元方法,对再设计后的塔架的静强度、模态、稳定性进行了分析。分析结果表明:再设计后的塔架的强度、固有频率和刚度均满足要求;以满足强度、频率特性和刚度为约束条件,以减轻重量、降低成本为目标的塔架的再设计是可靠的。     

某风电场风力发电机组故障诊断

为了准确判断风电机组的运行状态及故障,提出了基于常规分析—振动幅值分析—波形频谱分析的故障诊断流程,阐述了针对风电机组的幅值分析方法和波形频谱分析方法,并通过对某机组异响的根源探究实例,准确地诊断出机组异响来源于齿轮箱太阳轮,可为风电机组故障诊断技术提供依据。     

Simulation of hourly wind speed and array wind power

Statistical summaries of wind speed are sufficient to compute many characteristics of turbine-generated power, such as the mean, variance and reliability of various power levels. However, a wind speed time series is necessary to produce a sequence of power values as used for investigating load matching and storage requirements. Since a long historical record of wind speed may not be available at a wind turbine candidate site, it is desirable to be able to generate a simulated numerical sequence of hourly wind speed values. Two such approximate procedures are developed in this paper. One procedure generates sequential wind speed values at a site based on the Weibull parameters of hourly wind speed and the lag-one autocorrelation of hourly wind speed values. Comparison with historical data at a site is made. The second procedure generates sequential hourly wind power values for a regional array of wind turbines. It utilizes the typical site wind characteristics, the spatial and lag-one cross correlation and autocorrelation of hourly wind speed values and an equivalent linearized relationship between array average wind speed and array power. Comparison with results for six different wind turbines in three different regional arrays indicates good agreement for wind power histograms, autocorrelation function and mean persistence.     

Characterization of wind speed data according to wind direction

Knowledge of the wind speed distribution and the most frequent wind directions is important when choosing wind turbines and when locating them. For this reason wind evaluation and characterization are important when forecasting output power. The data used here were collected from eleven meteorological stations distributed in Navarre, Spain. We obtained data for the period extending from 1992 to 1995, with each datum encompassing 10 minutes of time. Wind speed data of each station were gathered in eight directional sectors, each one extended over 45 degrees according to the direction from which the wind blows. The stations were grouped in two blocks: those under the influence of the Ebro valley and those in mountainous areas. For each group the Weibull parameters were estimated, (according to the Weibull probability paper because the Weibull distribution gives the best fit in this region). Kurtosis and skewness coefficients were estimated as well. The Weibull parameters, especially the scale parameter c, depend strongly on the direction considered, and both Weibull parameters show an increasing trend as the direction considered moves to the more dominant direction, while both kurtosis and skewness show a corresponding decreasing trend.     

A new formulation for rotor equivalent wind speed for wind resource assessment and wind power forecasting

The spurt of growth in the wind energy industry has led to the development of many new technologies to study this energy resource and improve the efficiency of wind turbines. One of the key factors in wind farm characterization is the prediction of power output of the wind farm that is a strong function of the turbulence in the wind speed and direction. A new formulation for calculating the expected power from a wind turbine in the presence of wind shear, turbulence, directional shear and direction fluctuations is presented. It is observed that wind shear, directional shear and direction fluctuations reduce the power producing capability, while turbulent intensity increases it. However, there is a complicated superposition of these effects that alters the characteristics of the power estimate that indicates the need for the new formulation. Data from two field experiments is used to estimate the wind power using the new formulation, and results are compared to previous formulations. Comparison of the estimates of available power from the new formulation is not compared to actual power outputs and will be a subject of future work. © 2015 The Authors. Wind Energy published by John Wiley & Sons, Ltd.