基于时空注意力-Seq2Seq模型的多风电机组多步风速预测算法

针对在多风电机组风速预测任务中，卷积运算不适用于提取排布不规则的多风电机组空间相关性的问题，提出一种基于时空注意力-Seq2Seq模型的多风电机组多步风速预测算法。首先使用空间注意力机制强化风速序列的空间相关性，并对常规空间注意力机制进行改进；之后使用Seq2Seq模型中的编码器进行编码；最后使用结合时间注意力机制的解码器计算多风电机组的多步预测结果。以河北市某风电场的实际数据为算例进行实验，结果表明相比其他对比算法，所提算法的平均绝对误差下降约4.3%～15.0%，精度有较大提高。     

大型风电场无功特性对其低电压穿越能力的影响

建立了两种典型风电机组的机电暂态模型,通过与实测数据比对校验模型的准确性。基于校验的风机模型,建立100 MW并网风电场详细仿真模型,通过仿真分析比较风电机组无功特性对风电场低电压穿越能力的影响,研究无功补偿装置的控制特性与风电场过电压的关系,为大型风电场的仿真分析及稳定控制提供了技术支撑。     

风速波动时风电场动态特性分析

为了分析风速波动时风电场的电压特性,建立了考虑风机组尾流效应的风机输入风速模型及考虑风机输入风速不同的风电场动态模型,并对风电场采用不同模型时风电场的输出进行仿真比较,结果表明在分析风速波动风电场动态特性时,采用考虑风机间尾流效应和风电场内机组排列布置的风电场模型能够相对准确地反映实际风电场的动态特性,而采用不考虑风机尾流效应的风电场单机模型,夸大了风电场对系统的影响.同时对风电场分别采用恒速风电机组和双馈变速风电机组时的输出特性进行仿真,分析了以上两种风电机组的无功电压特性.     

基于风电机组无功响应速度的风电场闪变研究

针对风电机组无功响应速度对风电场并网点电压闪变的影响,从风力发电机组特性出发,基于DIg SILENT建立了含风电机组的电网仿真模型,研究风速波动时风机无功响应速度对风电场并网点电压波动与闪变的影响。算例结果表明:风电机组无功响应速度对并网点电压闪变的影响程度与系统短路容量密切相关,加快风电机组无功响应速度能有效抑制风电场并网点的电压闪变。     

基于混合Copula优化算法的风速预测方法研究

针对风电场中各风电机组风速之间存在的复杂时空相关性问题,提出一种基于混合Copula优化算法的风电场风速预测方法。该方法首先分析单一Copula函数拟合优度检验,选取合适Copula函数进行组合;其次,构建混合Copula函数模型对风电场内多风电机组风速相关性进行分析;最后应用最大期望（EM）算法求解模型相关系数并完成风速预测。结合优化算法,改进Copula函数能很好地解决风速相关性问题,为获取准确风速预测值奠定基础。以中国某地区风电场风电机组实测风速数据为例对所提方法进行验证,实验结果表明该模型可在准确分析风速相关性的基础上提高风速预测准确性。     

基于数字孪生技术的风电机组建模研究

针对以往风电机组数字孪生建模受不同研究目的或单一软件的功能限制，难以建立风电机组整机模型的问题，提出一种新的风电机组孪生建模方法。该方法首先依托FAST风速性能模块，建立稳态风模型、随机湍流风模型以及风电场实时风速模型；接着采用空气动力学模块和结构动力学模块分别搭建风电机组叶片、塔架等关键部件的几何与动力学模型；最后在Simulink中搭建风电机组电气系统模型及控制策略，由此构建完整的风电机组孪生模型。将该孪生建模方法分别用于WindPACT 1.5 MW双馈风电机组与某风电场Fuhrl?nder 2.5 MW双馈风电机组并进行验证。结果表明：孪生模型在不同风速模型下，各重要生产参数相比设计标准及实际运行数据均具有较高的准确性。此外，通过对风电机组数字孪生系统实时仿真和现场不可测数据的孪生模拟，也进一步表明孪生模型具有可行性和有效性。     

风电场邻近风电机组缺测风速集成填充方法

针对风电场中邻近多台风电机组集中出现缺损测量风速的工况,提出基于粒子群优化广义回归神经网络的风电机组缺损测量风速集成填充方法。以"成员等同性"原则引入动态时间规整算法、空间邻点法和Pearson相关系数法,分别搜寻与缺损测量风速风电机组风速演化最为相似的若干台风电机组及对应的测量风速时序,建立基于广义回归神经网络的填充子模型,采用粒子群算法对广义回归神经网络的模型参数和训练集的构成进行全局优化,之后选取较好的子模型构造自适应的熵权集成填充模型。实验结果表明:依据相似性风速序列进行缺损风速的填充能有效提高填充精度;粒子群算法优化广义回归神经网络,不仅提高了子模型的填充效果,更使得模型参数的调节有据可依,能适应不同风电场风速数据的特点;基于熵权的集成填充策略理论依据充分,集成填充的精度和稳定性优于单个子模型。     

风电机组功率曲线建模方法对比研究

风电机组功率曲线是风电机组重要的性能指标,表征了机组的实际运行状态。准确的实测风电功率曲线可以为风电机组性能评估、风电功率曲线监测、风电功率预测、风电场数值建模等工作提供重要的参考依据。但采用耗时的功率曲线建模方法会花费大量的建模时间,从而影响建模效率。文章从功率曲线建模的数据筛选和功率曲线拟合入手,选取耗时较短的二维核密度估计模型筛选风速功率散点集中区域内的正常运行数据,并选用5种功率曲线拟合方法对正常风速功率数据进行拟合。5种模型的建模精度和建模效率对比分析表明,多项式拟合方法原理简单,拟合速度最快,且拟合精度较高,比较适用于实际功率曲线的建模工作。     

基于灰色理论和变权模糊综合评判的风电机组性能评估

为了评估风电机组性能及优化其维修策略,保证风电机组安全稳定运行,提出基于灰色理论和变权模糊综合评判的风电机组性能评估方法。首先,分析风电机组控制系统所监测的物理量,采用层次分析法构建具有反映风电机组运行性能重要特征的项目层和子项目层框架,并针对子项目层中多项评判指标同时发生严重偏离的情况,引入劣化度指标。为避免层次分析确定指标常权值所引起计算误差过大的问题,利用灰色系统理论得到各评估指标的综合权重,并结合变权理论,对某些严重偏离正常值的指标进行权重修正,建立基于灰色理论和变权模糊综合评判的风电机组性能评估模型。然后,基于某1.5 MW风电场数据采集与监控系统(Supervisory Control andData Acquisition,SCADA)的运行数据进行计算,并与常规评估方法结果及实际运行分析进行比较和验证,结果表明所提出的风电机组性能评估方法和模型正确有效。     

计及能量掺混的风电机组尾流叠加模型研究

风电机组可能受上游多台机组尾流共同影响,工程中一般应用叠加模型来模拟这种尾流叠加效应。尾流区与周围大气的能量掺混导致了尾流恢复,目前常用的尾流叠加模型无法体现这个效应。应用一维动量理论计算风电机组尾流区从周围大气吸收的能量,在能量守恒叠加模型的基础上,通过补充这部分掺混能量对其进行修正,从而提高了尾流场模拟精度。在Lillgrund海上风电场应用能量掺混叠加模型,流场模拟结果与实测数据对比表明,该模型可以准确模拟风电场内机组功率变化趋势,且相较于传统模型计算精度更高,对风电场发电量计算具有一定的参考价值。     

Dynamic models of wind farms with fixed speed wind turbines

The increasing wind power penetration on power systems requires the development of adequate wind farms models for representing the dynamic behaviour of wind farms on power systems. The behaviour of a wind farm can be represented by a detailed model including the modelling of all wind turbines and the wind farm electrical network. But this detailed model presents a high order model if a wind farm with high number of wind turbines is modelled and therefore the simulation time is long. The development of equivalent wind farm models enables the model order and the computation time to be reduced when the impact of wind farms on power systems is studied. In this paper, equivalent models of wind farms with fixed speed wind turbines are proposed by aggregating wind turbines into an equivalent wind turbine that operates on an equivalent wind farm electrical network. Two equivalent wind turbines have been developed: one for aggregated wind turbines with similar winds, and another for aggregated wind turbines under any incoming wind, even with different incoming winds.The proposed equivalent models provide high accuracy for representing the dynamic response of wind farm on power system simulations with an important reduction of model order and simulation time compare to that of the complete wind farm modelled by the detailed model.     

Aggregated dynamic model for wind farms with doubly fed induction generator wind turbines

As a result of increasing wind farms penetration in power systems, the wind farms begin to influence power system, and thus the modelling of wind farms has become an interesting research topic. Nowadays, doubly fed induction generator based on wind turbine is the most widely used technology for wind farms due to its main advantages such as high-energy efficiency and controllability, and improved power quality. When the impact of a wind farm on power systems is studied, the behavior of the wind farm at the point common coupling to grid can be represented by an equivalent model derived from the aggregation of wind turbines into an equivalent wind turbine, instead of the complete model including the modelling of all the wind turbines. In this paper, a new equivalent model of wind farms with doubly fed induction generator wind turbines is proposed to represent the collective response of the wind farm by one single equivalent wind turbine, even although the aggregated wind turbines operate receiving different incoming winds. The effectiveness of the equivalent model to represent the collective response of the wind farm is demonstrated by comparing the simulation results of equivalent and complete models both during normal operation and grid disturbances.     

Probabilistic performance assessment of wind energy conversionsystems

This paper describes the development of a general probabilistic model of an autonomous wind energy conversion system (WECS) composed of several wind turbines (wind farm) connected to a load and a battery storage. The proposed technique allows the simulation of wind farms containing identical or different wind turbines types and considers a bidirectional flow of power in and out of the battery. The model is based upon a simple procedure to estimate the joint probability distribution function of the total available wind power and that of the turbines operating modes due to hardware failure. A methodology is also developed to use the proposed model to determine an upper limit on the size of the battery storage required for a given number of turbines to satisfy the load with a certain expected energy not supplied (EENS). The model can also be used to evaluate the energy purchased from or injected to the grid in the case of grid-connected systems     

Spatial dependence in wind and optimal wind power allocation: A copula-based analysis

The investment decision on the placement of wind turbines is, neglecting legal formalities, mainly driven by the aim to maximize the expected annual energy production of single turbines. The result is a concentration of wind farms at locations with high average wind speed. While this strategy may be optimal for single investors maximizing their own return on investment, the resulting overall allocation of wind turbines may be unfavorable for energy suppliers and the economy because of large fluctuations in the overall wind power output. This paper investigates to what extent optimal allocation of wind farms in Germany can reduce these fluctuations. We analyze stochastic dependencies of wind speed for a large data set of German on- and offshore weather stations and find that these dependencies turn out to be highly nonlinear but constant over time. Using copula theory we determine the value at risk of energy production for given allocation sets of wind farms and derive optimal allocation plans. We find that the optimized allocation of wind farms may substantially stabilize the overall wind energy supply on daily as well as hourly frequency.     

Wind farm voltage dip measurements

This paper is based on continuous measurements of voltages and currents from three wind farms for a period of 1 year, and the focus is on voltage dips. The purpose is to get an overview of the characteristics and rate of voltage dips, which occur in the wind farms and to study the wind turbine responses to voltage dips. In each of the wind farms there is one measurement point at a single wind turbine and one for measuring the contribution from the whole wind farm. Different wind turbine technologies are used in the three wind farms; fixed speed turbines with directly connected induction generators in wind farm 1 and variable speed turbines with power electronics converters and synchronous generators in wind farms 2 and 3. Voltage dips are evaluated according to the standard EN 50160, by considering the durations and residual voltages of the positive sequence component voltage dips. Some examples of voltage dip events with corresponding responses in active and reactive power are shown and discussed with a view to the different technologies. Copyright © 2013 John Wiley & Sons, Ltd.     

基于RTDS的风电场等值建模与验证分析

针对风电场运行条件复杂、运行工况动态变化导致风电场输出功率的分散性问题,采用数理统计方法对风电场外特性进行稳态等值,考虑到大型风电场所处地形复杂、机群分布不规则带来的风速差异性问题,以风电场内长时间尺度实测风速数据作为特征变量,采用改进动态聚类算法进行机群划分,进而基于风电场参数对等值机模型的参数进行聚合辨识。基于RTDS实时数字建模及仿真试验分析结果表明,建立的风电场等值模型能够准确地反映风电场在不同风速及电网侧短路故障下的动态特性,可用于含双馈风电机组风电场接入电力系统稳定性分析。     

基于加装叶尖小翼风电机组发电量提升试验研究

针对老旧风电场由于风电机组风轮和风资源未能很好匹配等原因导致的发电量不理想问题,提出一种基于加装叶尖小翼风电机组发电量提升方法,该方法在增加叶片扫掠面积的同时有效改善叶片的气动性能。通过加装叶尖小翼增加叶尖部分切向力,增加主轴扭矩从而提高风轮吸收功率,达到提升发电量的目的。基于风电场实际试验数据对提功增效效果进行分析,结果表明所提方法在满足安全性的前提下可有效提高机组输出功率,提升发电量。同时该方法成本低,展现了良好的性价比。     

风电场中多台风力机的数值模拟

将NREL 5 MW风力机作为基本机型,使用致动线模型和大涡模拟相结合的数值方法,在中性大气边界层中模拟含有多台风力机的风电场。为了模拟风电场的复杂入流条件,首先模拟体积为3000 m(长)×3000 m(宽)×1000m(高)的大气边界层,并对模拟结果进行验证,结果表明:在覆盖逆温层以下,不同高度处的位温不变,平均风速满足剪切特性,脉动风速满足湍流谱特性;然后,分析了致动线模型中风轮直径上的网格节点数量(N)和高斯分布因子(ε)的取值规律,发现ε以网格尺度(η)为自变量取值时,N越大,η的系数越大,当N取63时,η的系数可取2或3,但N取25时,η只能取1.2;最后,使用致动线模型在大气边界层中布置8台风力机,模拟风电场,并对风力机间的相互干扰进行分析,发现第一排风力机功率明显大于其他风力机功率输出,占风场总功率输出的40.3%。     

Modelling the effects of environmental conditions on wind turbine failures

Operation and maintenance is one of the main cost drivers of modern wind farms and has become an emerging field of research over the past years. Understanding the failure behaviour of wind turbines (WTs) can significantly enhance operation and maintenance processes and is essential for developing reliability and strategic maintenance models. Previous research has shown that especially the environmental conditions, to which the turbines are exposed to, affect their reliability drastically. This paper compares several advanced modelling techniques and proposes a novel approach to model WT system and component failures based on the site‐specific weather conditions. Furthermore, to avoid common problems in failure modelling, procedures for variable selection and complexity reduction are discussed and incorporated. This is applied to a big failure database comprised of 11 wind farms and 383 turbines. The results show that the model performs very well in several situations such as modelling general WT failures as well as failures of specific components. The latter is exemplified using gearbox failures.