A Study on Influence of Ramp Event of Aggregated Power Output of Photovoltaic Power Generation on Electric Power System Frequency

This paper examines the influence of aggregated power output fluctuation of photovoltaic power generation system (PVS) on the power system frequency focusing on ramp events of aggregated PVS power output. A numerical simulation model of economic load dispatching control (EDC) and load frequency control (LFC) is used together with a PVS power output forecasting model and a unit commitment (UC) scheduling model developed in our preceding study. As a result, in the case of ramp event with long duration and high ramp rate, the frequency violation occurs when the power output of controllable generators with high load‐following capability reaches to upper/lower limit even if the power output of low load‐following capability generators is still available. If the load dispatching scheme is tentatively changed from the conventional EDC using an equal incremental fuel cost rule to, for example, a dispatching policy based on the capacity without the consideration of fuel cost, the aggregated load‐following capability can be kept, avoiding the frequency violation.     

风力发电系统功率波动传递特性的研究

风电具有间歇性、随机性和不确定性的特点,对电功率波动的评估是合理选择风电场储能互补设施的基础,但对于规划建设中的风电场,往往有丰富的风速数据而没有电功率数据。本文以长时间大量风速和电功率数据为样本,采用小波包将电功率和风功率分解为不同频段下的波动,通过数据曲线拟合得到风功率和电功率波动随频率的变化关系,进而得到从风功率到电功率的传递函数。风电功率传递函数可将电功率波动的估计问题转化为风功率波动的估计,为拟建风电场的功率波动研究提供一种新途径。     

The use of statistical methods to analyze electric-energy consumption by budget institutions

The possibilities of applying statistical methods to analyze the consumption of electric energy by budget organizations and control energy efficiency are studied. Statistical methods are used to consider the influence of the season and climatic conditions on consumption of electric energy, which allowed a simulation model to be created of electric-energy consumption for illumination and other needs according to data of devices considering only the total consumption of electric energy by all groups of consumers. The change in monthly consumption of electric energy within a year is a stochastic value depending on the duration of light in a day and other factors. The statistical characteristics of the monthly consumption of electric energy by individual buildings of Chuvash State University are determined. Dispersion analysis shows that distribution of the values of the monthly electric-energy consumptions by most buildings corresponds to the normal law. The correlation between the monthly electric-energy consumption and duration of light in a day permitted one to carry out a regression analysis and determine the actual average power of the lighting system and actual average power of the rest of the energy consumption of each facility. The results of the statistical analysis make it possible to compile the electric-energy consumption balance, predict electric-energy consumption by the different group of consumers, estimate the efficiency of energy-saving measures, and obtain the data necessary to fill in energy declarations and energy-performance certificates.     

计及群控电热泵的微网联络线功率平滑策略

为有效平滑电-热联合型微网的联络线功率波动,提出一种基于电热泵和蓄电池的协同控制策略。考虑电热联合微网的运行特征及能量交互方式,基于电热泵的运行特性建立电-热能量转换模型。通过分析电热泵群优先度序列,计算电热泵群可运行数量与状态切换序号指针。考虑机组启停的调控限制,采用集群控制算法计算电热泵群开关序列。基于储能荷电状态以及电热泵实际出力状态,分别建立蓄电池、电热泵功率分配权函数,将联络线的波动功率在蓄电池与电热泵群间进行分配。算例结果表明,该策略能够通过电-热能量协同控制实现微网联络线功率平滑,同时提升电-热联合型微网的互补经济性与能量综合利用效率。     

两区域互联系统交流联络线随机功率波动机制及幅值估计

正常运行时两大区互联系统交流联络线上存在随机功率随机波动的现象,限制了联络线的输电能力。研究了随机功率波动的机制与波动幅值估计。基于系统静态频率特性,推导了两区域交流联络线功率波动的公式,阐明了联络线上的低频功率波动是一种稳态过程,是负荷随机功率波动在全系统中再分配后在联络线上的反映。提出了一种联络线功率波动幅值的概率模糊估计算法。首先通过概率统计的方法来估计系统频率波动幅值的概率分布,再通过模糊数来描述贡献因子的不确定性,最后通过模糊抽样和解模糊来得到联络线功率波动的均值和标准差。通过华北–华中互联系统特高压联络线功率的实测数据验证了所提机制和算法的正确性和有效性。研究有助于掌握两大区互联电力系统的随机功率波动特性,为制定抑制联络线功率波动的措施提供依据。     

地铁供电系统无功补偿方案

在分析地铁供电系统基本特征的基础上,建立了系统平均功率因数的数学模型,并以广州地铁2号线的运营数据为例验证了模型的有效性。对于地铁供电系统而言,高压交流电缆(33/110 kV)充电功率的影响不容忽视。结合广州地铁5号线供电系统初步设计,分析了用电负荷的波动和动力照明负荷功率因数的变化对系统110kV侧功率因数的影响,估计了动力照明负荷的临界功率因数,经过技术经济比较提出了优化的无功补偿方案。     

Influence of forecast accuracy of photovoltaic power output on capacity optimization of microgrid composition under 30‐minute power balancing control

A microgrid (MG) is one of the measures for enhancing the high penetration of renewable energy (RE)‐based distributed generators (DGs). If a number of MGs are controlled to maintain the predetermined electricity demand including RE‐based DGs as negative demand, they would contribute to supply–demand balancing of the whole electric power system. For constructing an MG economically, capacity optimization of controllable DGs against RE‐based DGs is essential. Using a numerical simulation model developed on the basis of a demonstrative study on an MG using PAFC and NaS battery as controllable DGs and photovoltaic power generation system (PVS) as an RE‐based DG, this study discusses the influence of the forecast accuracy of PVS output on capacity optimization. Three forecast cases with different accuracy are compared. The main results are as follows. Even with no forecast error during every 30‐minute period, the ideal forecast method, the required capacity of NaS batteries reaches about 40% of the PVS capacity for mitigating the instantaneous forecast error within 30 minutes. The required capacity to compensate for forecast error is doubled under the actual forecast method. The influence of forecast error can be reduced by adjusting the scheduled power output of controllable DGs according to weather forecasts. Besides, the required capacity can be reduced significantly if the error of balancing control in an MG is acceptable for a few percent of periods, because the total periods of large forecast error are not frequent. © 2012 Wiley Periodicals, Inc. Electr Eng Jpn, 182(2): 20–29, 2013; Published online in Wiley Online Library ( wileyonlinelibrary.com ). DOI 10.1002/eej.22328     

基于SC-DNN和多源数据融合的新能源电力系统状态估计方法

大规模新能源并网重塑了电力系统的控制运行特性,现有的电力系统状态估计方法面临新能源波动数据识别困难、估计精度低、估计速度慢等问题。为改善现有方法的不足,提出了一种基于残差连接(skip connection, SC)-深度神经网络(deep neural network, DNN)和多源数据融合的新能源电力系统状态估计方法。首先采用基于双向长短期神经网络(bidirectional long short-term memory, BILSTM)预测的改进插值法进行多源数据融合。然后利用联合时空交叉机制和BILSTM网络的数据辨识技术替代传统的量测量突变检测法,以便更好地处理新能源波动数据。最后根据原始量测数据集建立基于SC-DNN的状态估计模型,把残差模块的拟合优势和神经网络的速度优势结合起来,从而实现状态估计精度和速度的提高。基于IEEE39节点系统和新疆某地区实网的算例分析表明,相比于传统方法,所提方法能在更准确地分辨新源波动数据与不良数据的同时提高状态估计的精度和速度。     

考虑电动汽车停泊特性的充电站-电网互动策略研究

研究充电站与电网之间的互动,对改善电动汽车的无序充电行为具有十分重要的意义。提出了一种考虑电动汽车停泊特性的充电站-电网互动策略。首先,基于历史数据,配电网运营商优化分时电价,以减少负荷波动。其次,考虑到电动汽车的停泊特性,充电站根据配电网运营商发布的充放电价格控制电动汽车的充放电功率。最后,配电网运营商通过动态重构策略改变配电网的潮流分布,以提高电压质量并降低线路损耗。基于IEEE-33节点配电网系统进行了仿真,仿真结果表明所提的互动策略可以提高充电站的利益,并能降低配电网的网损。     

基于求根多重信号分类和遗传算法的谐波间谐波频谱估计

为了改善电能质量,需要准确估计电网电压或电流的谐波、间谐波频谱.将空域处理技术的空间谱估计方法应用到电网信号的谐波、间谐波频谱估计.使用求根多重信号分类方法估计电压或电流的不同频率成分数及各成分的频率,各成分的幅值和相角估计由遗传算法完成.首先介绍频谱估计使用空间谱估计的理论依据,然后详尽地说明了求根多重信号分类方法和遗传算法的原理及算法流程.仿真结果和与其他方法的比较证明此方法在数据长度有限和噪声干扰时可以准确估计谐波、间谐波的频谱.     

Estimation method of frequency bias using normal operating data

A technique of calculating the frequency bias of an electric power system from frequency fluctuation data and total output power is described. An evaluation method for testing the calculations by the new method is developed. The results are in good agreement with estimates based on generator speed regulation but the method gives an excessively high figure for the load frequency bias.     

基于地区电网的配电网规划及其评估研究

针对地区电网依次从电网现状、负荷预测、规划目标及设计原则、电力电量平衡、电网规划方案和网架结构论证、电气计算和配电网建设项目及投资估算进行了配电网规划研究。后面接着从供电可靠性评估、线损率评估两方面进行规划评估研究。     

Accuracy of Irradiance Forecasts Using JMA Numerical Weather Prediction Data at Numerous Points Corresponding to a High Degree of PV Penetration

To assure power system stability when variable photovoltaic (PV) power generation is deployed on a large scale, accurate irradiance forecasting will be important in the near future. A substantial amount of PVs will be generating power at numerous locations in the power service area, and it will be necessary to predict the total PV output for supply/demand balancing control. Therefore, irradiance forecasting accuracy should be verified using the total irradiance at numbers of sites equivalent to areas with large‐scale PV deployment. This paper reports an investigation of the accuracy of an approach to area total irradiance prediction based on regional numerical weather prediction data from the Japan Meteorological Agency. The forecast area total irradiance is compared to the assumed total area irradiance for numbers of sites equivalent to large‐scale PV deployment.     

正交频分多路复用技术在电力光纤通信系统中的应用

提出将正交频分多路复用(orthogonal frequency division multiplexing,OFDM)技术应用到电力光纤通信系统中,以提高系统的可靠性和稳定性。针对光纤信道特性在频域上变化缓慢且相邻的信道特性高度相关的特点,在基于训练序列的信道估计中采用频域滑动平均方法来提高信道估计的准确性。仿真结果表明,相对于原始的信道估计方法,采用频域滑动平均信道估计方法后OFDM信号接收灵敏度提高1.5 dB,且相对背靠背时基本不会引入功率代价。     

基于大数据分析的县公司综合评价策略

提出一种基于大数据分析的县公司综合评价策略。以1780家县公司数据为基础,建立考虑统计分析、决策树分析、数据包络分析等大数据技术的指标挖掘策略;通过信息挖掘构建一套适用于县公司综合评价的指标体系。以经济因素和企业因素为规则,建立基于县公司聚类的差异化权重计算方法,实现对县公司差异化发展内生和外生因素的联合判别,克服综合评价结果解释的单一化。通过浙江省的实证化应用,验证了所提策略的有效性和实用性。     

Development of a three-phase unbalanced voltage fluctuation compensating system using a self-commutated static var compensator

This paper describes a self-commutated static var compensator for suppressing voltage fluctuations caused by an ac electric railway (Shinkansen, or new trunk line). The ac electric railway is a single-phase load with large changes of reactive power and it causes large voltage fluctuations and power imbalances among the three phases in the power system. The authors have developed a new voltage fluctuation compensating system using a self-commutated static var compensator. This system has the functions of reactive power compensation and negative-phase-sequence current absorption. It can suppress rapidly changing voltage fluctuations and reduce the three-phase power imbalance caused by the ac electric railway effectively. This compensating system, with a capacity of 40 MVA, was put to practical use for the first time in the world in 1993. The test results at the site show that the system has excellent performance characteristics. A theoretical analysis of the characteristics of voltage fluctuations with or without the compensating system and an evaluation of test results, including a comparison between the theoretical and measured values, are also reported. © 1998 Scripta Technica, Inc. Electr Eng Jpn, 121(3): 67–78, 1997     

考虑风电波动与电动汽车集群储能平抑控制策略

为促进风电在电网中的消纳,减轻配电网负荷压力,提出考虑风电出力波动和电动汽车集群储能系统平抑控制策略.首先,对单体电动汽车入网后行为特性进行储能建模,依据不同荷电状态(SOC)电动汽车有功响应能力,构建电动汽车集群储能模型,基于集群储能能力的差异性,利用多个电动汽车集群协调平抑联络线功率波动.其次,由集群储能系统依据联络线功率平抑波动值进行逐层自适应功率分配,确定各电动汽车蓄电池—超级电容的任务功率,充分利用车网连续调节能力.所提平抑策略可减轻大规模电动汽车连网后配电网中负荷的波动,实现储能系统内部功率相互流动,有效减少常规储能容量配置.     

基于改进马尔科夫链的风电功率时间序列模型

模拟风电功率时间序列在风电并网系统的规划和评估研究中具有重要意义,针对原始马尔科夫链在风电功率建模上无法保留其自相关性的不足,构建了一种基于改进马尔科夫链的风电功率时间序列模型。首先分析了风电功率的季节特性、日特性和波动特性;然后将风电功率数据按照不同月份及时段进行了细致划分,生成相应的状态转移概率矩阵;最后,对风电功率波动量的概率分布进行拟合,并叠加波动量,建立了基于改进马尔科夫链的风电功率时间序列模型。实例分析表明,本文所建新模型生成的风电功率序列能够保留历史序列自相关性,同时在一般统计参数、概率密度分布和自相关性三方面的准确性也优于已有模型。     

基于聚类算法的风电波动过程研究*

受电源结构和电网结构等因素制约,风电出力的随机性和波动性已给我国电网的安全运行和清洁能源的高效消纳带来了很大挑战。文中研究了基于自组织映射神经网络算法的风电波动过程划分方法,进而提出了基于聚类分析的多时空尺度风电波动特性研究框架。并在此基础上,以"三北"地区2017年实际运行数据为依据,从"风电场-省级电网-区域电网"三个层级研究了风电的分钟-小时级短期波动幅度特性和长期统计特性。分析结果表明,自组织映射聚类算法可对风电波动类别进行有效辨识,风电出力波动的时间-空间特性指标可对风电富集地区的调度运行提供量化决策依据。     

电力市场环境下云南电网AGC辅助服务问题

在电力市场环境下,为了既满足系统对调节容量和速率需求,又能使电网AGC辅助服务费用最低,基于"南方电网‘1+5’电厂辅助服务考核技术支持系统"的实际数据,统计分析了云南电网AGC辅助服务情况,并根据云南电网电源特性对云南电网AGC辅助服务提出了两个优化建议,即细化考核标准和优化考核针对性,以提高AGC调节效能、合理利用AGC容量和改善AGC运行环境。