能源大数据背景下微网风险元传递模型与优化研究

发布时间：2018-01-04 21:20

本文关键词：能源大数据背景下微网风险元传递模型与优化研究　出处：《华北电力大学(北京)》2017年博士论文　论文类型：学位论文

【摘要】：微网是在新能源发电背景下,在分布式发电基础上新兴的前沿技术。微网作为智能电网的有机组成部分,对于企业节能降耗、发电、供电质量需求具有重要意义。微网运营相关技术研究已受到了世界各国普遍关注和重视。微网集成了多种能源输入(太阳能、风能、常规化石燃料、生物质能等)、多种特性负荷、多种能源转换单元(燃料电池、微型燃气轮机、内燃机,储能系统等),是化学、热力学、电动力学等行为相互藕合的非线性复杂系统。微网中许多类型的分布式发电电源受制于自然条件,运行不确定性强,具有间歇性、复杂性、多样性、不稳定性的特点,其电能质量特征与传统电力系统有很大差异。同时,微网中的负荷受区域环境、经济和政策等社会因素的影响,也给微网带来更多的不确定性。电的流动性导致了电网中的各单元的紧密联系特性,当某个单元发生风险时也会影响到其他单元,甚至导致其他单元也发生风险,电网中这种紧密联系性也给风险可以在其间发生传递提供了可能。随着能源互联网建设的不断深入和推进,电网运行和设备检/监测产生的数据量呈现指数级增长,逐渐构成了当今信息学界所关注的能源大数据。能源大数据蕴含着能源互联网的大量知识信息,能源大数据的有效使用将为微网高效运行提供有力保障。本文站在能源大数据背景下,基于风险元传递理论,先从发电量和用电量两个方面研究如何识别微网运营中的风险元;再分别研究微网间的双向型和网络型风险元传递路径,由简入繁揭示微网风险元传递机理;最后给出减少或消弥微网风险元传递的优化方法,为管理者提供较为系统的微网风险管理决策依据。本文主要研究内容包括以下几个方面:(1)能源大数据创新微网风险管理研究。首先分析风力发电大数据、光伏发电大数据和供用电大数据等,指出微网运营过程中,微网间存在的风险动态变化现象。然后提出在能源大数据背景下,可以创新微网风险管理研究,采用风险元传递理论能够更好地分析并解释微网间的风险相互影响情况。(2)微网发电量风险元识别研究。受自然环境条件的限制,微网发电量不确定性是微网发生风险的主要原因之一,能源大数据为微网发电量风险元挖掘提供了新的途径,在风力发电和光伏发电功率预测的基础上,先对微网潮流进行确定性预测,再将马尔可夫链和拉丁超立方抽样相结合,分别对微网潮流的条件联合概率分布和非条件联合概率分布进行预测。根据风速和光照的演变特性,计算微网潮流的概率分布及置信区间可以对预测结果的不确定性进行风险元识别。(3)微网用电量风险元识别研究。在微网运行中用电量的不确定也是造成微网风险的主要因素之一。受飞蛾扑火自然现象的启发,为提高预测的精度,将其用于优化最小二乘支持向量机算法,提出一个基于最小二乘支持向量机和飞蛾扑火优化算法的混合的负荷预测模型,以达到更加精确地预测用电负荷、准确识别微网中用电量风险元的效果。(4)微网风险元双向型传递模型研究。通过对大数据分析发现,风险元不只是在微网间单向传递,还存在着互相传递的情况。以并入大电网的风光储微网的风险管理为研究对象,提出双向风险元传递模型,给出缺电风险值和影响因子计算方法。此传递路径的提出丰富了原风险元传递理论研究的内容,揭示了风险元传递的又一种客观现象。(5)微网群网络风险元传递模型研究。与大电网互联的多个微网,它们之间通过大电网相互联系,符合网络的特征,每个微网可以看作是网络中的节点,大电网互通关系是它们的边。基于复杂网络理论,构建微网群网络模型分析其复杂网络特性,并在此基础上研究微网群网络风险元传递情况。(6)微网结构智能仿生优化模型研究。研究风险的主要目的是为了规避风险和控制风险,微网是一个局部区域供能系统,规避和控制微网风险的主要措施是优化微网的结构配置。在考虑能源资源、分布式能源、储能和负载的复杂匹配关系和分析分布式电源功率外特性的基础上,建立了供电可靠性、经济成本和环境效益的微网容量优化配置的目标函数,并采用混沌优化多目标遗传算法进行求解。(7)微网备用容量优化模型研究。根据光伏、风力发电为代表的可再生能源具有间歇性、随机性及不确定性等特点,提出了用于平滑可再生能源发电系统功率输出及微网联络线功率波动的储能系统容量优化方法。利用离散傅里叶变换对可再生能源输出功率、微网平滑联络线功率所需可控功率输出进行频谱分析,优化选取满足约束的储能系统所需最优容量。本文在能源大数据背景下,将风险元传递理论引入到了微网的风险管理中。丰富了风险管理的内容,能够为微网运营风险管理提供更多的依据,具有较强的理论价值和现实意义。
[Abstract]:The microgrid is in new energy power generation under the background of new technology in distributed generation based on Microgrid. As an organic part of the smart grid, for energy saving, power generation, has important significance to the quality of power supply requirements. The micro research on related technologies of network operation has been widespread concern and attention all over the world. The micro network integration a variety of input energy (solar, wind, biomass and other conventional fossil fuels), a variety of loads, various units of energy conversion (fuel cell, micro gas turbines, internal combustion engines, storage systems, etc.) is a chemical, thermodynamic, nonlinear electrodynamics behavior of coupled complex systems. Many types of micro network. Distributed power supply is subject to natural conditions, with strong uncertainty in operation, intermittent, complexity, diversity, the characteristics of stability, the power quality characteristics are very different from traditional power system At the same time, in the microgrid load by the regional environment, social factors and economic policies, but also to the micro network brings more uncertainty. Liquidity led to a close contact electrical characteristics of each unit in the network, when a unit of risk will affect the other unit, and even lead to other the unit also has the risk, in this connection the grid closely to risk in during possible transfer. With the deepening of the construction of Internet energy and propulsion, power grid operation and equipment inspection / monitoring the amount of data generated exponentially, and gradually formed a large energy data concern in today's information science. Energy big data contains a large amount of information and knowledge of the energy of the Internet, effective use of energy data will provide a strong guarantee for efficient operation of microgrid. The energy station in the context of large data, based on the risk element transmission From the first generation theory, and Research on two aspects: how to identify the power micro grid operation risk; and then studied the microgrid between bidirectional and network type risk element transmission path, from simple to complex reveals the micro network risk element transmission mechanism; finally, to reduce or eliminate the risk element transmission of microgrid optimization method, providing network risk management decision-making system for more micro management. The main contents of this paper include the following aspects: (1) energy data innovation micro network risk management research. First analysis of wind power generation photovoltaic power big data, big data and data for radio and TV University, pointed out that the microgrid operation process, Risk Dynamic Changes of microgrid between and then energy in the background of big data, can innovation micro network risk management research, the risk element theory can better analyze and explain the micro risk network interaction . (2) microgrid power generation risk element recognition research. By the natural environment conditions, micro grid power generation uncertainty is one of the main reasons for the risks of microgrid, energy data for the micro grid power generation risk provides a new way for yuan mining, based on wind power and Guang Fufa power prediction first, deterministic prediction of microgrid trend, then the Markov chain and Latin hypercube sampling combined, respectively, the joint probability distribution of the micro grid current conditions and non conditional joint probability distribution is predicted. According to the evolution characteristics of wind speed and illumination, calculation of micro probability distribution network power flow and the confidence interval can be the risk element recognition result of forecasting uncertainty. (3) research on micro grid electricity risk recognition. One of the main factors of risk uncertainty in microgrid microgrid operation with electricity is caused by natural fire moths. Inspired by the phenomenon, in order to improve the prediction accuracy for the optimization of least squares support vector machine algorithm, put forward a prediction model of least squares support vector machine and Puhuo optimization algorithm mixed load based on, in order to achieve a more accurate prediction of power load, the accurate identification of micro grid power consumption risk element (4). A transmission model of microgrid based on two-way risk element type. Data analysis found that the risk element is not only in the microgrid between one-way transmission, there is mutual transfer. The scenery storage in grid connected microgrid risk management as the research object, put forward the two-way risk element transmission model, given the shortage of risk value the calculation method and influence factors. The proposed transfer path has enriched the original risk element transmission theory, and reveals a kind of objective phenomenon of risk transfer. (5) micro network network risk element transmission Study on the model. And a plurality of micro grid interconnection network, connected by large power grid, accord with the characteristics of the network, each micro network can be viewed as nodes in the network, network communication relationship is their side. Based on complex network theory, constructing micro network network model analysis of the characteristics of the complex network, and then based on the research of micro network network risk element transmission. (6) model of bionic optimization of micro intelligent network structure. The main objective of the study was to avoid the risks and control risks, microgrid is a local energy supply system, avoid and control risk is the main measures of micro grid optimization of microgrid. In consideration of energy resources, distributed energy, energy storage and load matching relation and complex analysis of distributed power characteristics on the basis of the establishment of the power supply reliability, capacity of microgrid economic cost and environmental benefits The objective function of optimization configuration, which is solved by using chaos optimization multi-objective genetic algorithm. (7) micro network reserve capacity optimization model. According to the research on photovoltaic, wind power as the representative of the renewable energy is intermittent, random and uncertain characteristics, for smooth power renewable energy power generation system and Microgrid power output the fluctuation of energy storage capacity optimization method is proposed. The system output power of renewable energy by using discrete Fourier transform, the microgrid smooth tie line power required for control of power output spectrum analysis, optimal selection of meet the constraints of energy storage system for optimal capacity in energy. In this paper, the background of big data, the risk element transfer theory to the risk management of microgrid. Enrich the content of risk management, it can provide more evidences for microgrid operation risk management theory, has strong value and price Real meaning.

【学位授予单位】：华北电力大学(北京)
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TM727

【参考文献】