基于新息图理论的含分布式电源配电网三相状态估计研究

发布时间：2018-06-07 07:05

本文选题：配电网状态估计 + 新息图　；参考：《哈尔滨工业大学》2014年硕士论文

【摘要】：配电网状态估计是获得配电系统准确实时状态信息的重要手段，是配电网运行控制的基础。本文针对波动性分布式电源并网后的配电网状态估计问题展开研究，以新息图理论为基础，建立含分布式电源配电网的新息图模型，对一系列不正常事件进行辨识。针对接入波动性分布式电源后，节点注入不正常事件发生频繁，辐射状配电网近根部支路因受新息误差累加的影响，而导致不正常事件误判或者湮没的问题，通过分析误差形成的原因，提出了计及网损和分布式电源输出功率的计算新息矢量，用当前时刻的节点注入测量与节点注入预报分别回推计算各支路潮流，将得到的两潮流作差，形成计算新息矢量，并代替原有的连支推算新息矢量，同时对分布式电源进行简化处理，建立含分布式电源配电网的新息图模型。通过蒙特卡洛模拟的方法对两种新息图模型的误差进行了比较，结果表明计算新息矢量在新息图计算中能够极大降低新息误差，提高了新息图在配电网中辨识不正常事件的精确度。与输电网的新息图建模不同，配电网新息图计算中必须用到节点注入测量，会出现节点注入测量误差较大的坏数据和预报误差较大的负荷突变发生在同一位置的情况，配电网中接入间歇式、波动性分布式电源后，受两种误差交叠影响的情况进一步加重。针对节点注入测量坏数据与负荷突变交叠、多相关测量坏数据与负荷突变交叠这两类问题，基于含分布式电源配电网新息图模型，提出在新息差中将测量坏数据与负荷突变分离，先辨识测量坏数据后辨识负荷突变的分类辨识方法。33母线测试系统算例表明，所提出的方法能准确辨识接入分布式电源后节点注入测量坏数据与负荷突变交叠、多相关测量坏数据与负荷突变交叠的情况。配电网正由传统的严格辐射状向弱环状发展，弱环联络开关状态的改变引起拓扑结构的变化，针对含分布式电源配电系统中弱环合环但未报告的拓扑错误，在解环模型下通过新息图理论将弱环的辨识转化为弱环合环潮流等效节点注入位置的辨识；单电源辐射状配电网结构最简单但不利于提高供电可靠性，广泛存在的是由常分联络开关连接的多电源环网，在这类结构的配电网中因线路故障或者检修经常进行负荷转移，改变了拓扑结构，通过新息图理论辨识转移负荷的方向以及负荷的大小，来辨识负荷转移引起的拓扑错误。通过33母线测试系统验证了所提出辨识拓扑错误方法的有效性。本文研究工作依托于国家高技术研究发展计划（863计划）重大项目《高渗透率间歇性能源的区域电网关键技术研究和示范》（2011AA05A105）以及国家自然科学基金项目《电网参数分检式估计方法研究》（50977017）。
[Abstract]:Distribution network state estimation is an important means to obtain accurate real-time state information of distribution system and is the basis of distribution network operation control. In this paper, based on the theory of innovation graph, the innovation graph model of distributed power distribution network is established, and a series of abnormal events are identified. After accessing the fluctuating distributed power supply, the abnormal events of node injection occur frequently, and the near root branch of the radial distribution network is affected by the accumulation of innovation error, which leads to the misjudgment or annihilation of the abnormal events. By analyzing the causes of the error formation, the computational innovation vector considering the network loss and the output power of the distributed power source is proposed. The node injection measurement and node injection prediction at the current time are used to calculate the branch power flow, respectively. The calculated innovation vector is formed by the difference of the two currents, and the new information vector is calculated instead of the original connected support. At the same time, the distributed power supply is simplified and the innovation graph model of the distribution network with distributed power is established. The errors of the two models are compared by Monte Carlo simulation. The results show that the computational innovation vector can greatly reduce the error of innovation in the calculation of innovation graph. The accuracy of the innovation graph in identifying abnormal events in distribution network is improved. Different from the modeling of innovation graph of transmission network, nodal injection measurement must be used in the calculation of innovation graph of distribution network, and the bad data with big error of node injection measurement and the sudden change of load with large prediction error will occur in the same position. After the intermittent and fluctuating distributed generation is connected in the distribution network, the influence of the overlap of the two kinds of errors is further aggravated. Aiming at the two kinds of problems such as node injection measurement bad data and load mutation overlap, multi-correlation measurement bad data and load sudden change overlap, this paper based on the innovation graph model of distribution network with distributed generation. In this paper, a new method of classification and identification of bad data and load mutation is proposed, which separates the bad data from the load mutation in the new interest rate difference. The example of the busbar test system .33 shows that the method can be used to detect the bad data and then identify the load mutation. The proposed method can accurately identify the overlap of node injection measurement bad data and load mutation after access to distributed power supply, and the overlap of multi-correlation measurement bad data and load abrupt change. The distribution network is developing from the traditional strict radiation to the weak ring. The change of the state of the weak ring connection leads to the change of the topology structure. The topology error of the weak ring closed ring in the distributed power distribution system is not reported. Based on the theory of innovation graph, the weak ring identification is transformed into the identification of the equivalent node injection position of the weak loop, the single source radial distribution network has the simplest structure but is not conducive to improving the reliability of power supply. What exists widely is the multi-power loop network connected by the constant branch contact switch. In the distribution network of this kind of structure, the load transfer is often carried out because of the line fault or maintenance, which changes the topological structure. The topological errors caused by load transfer are identified by using innovation graph theory to identify the direction and magnitude of load transfer. The validity of the proposed method is verified by 33 busbar test system. The research work in this paper is based on the national high technology research and development plan "the national high-tech research and development plan") the major project "regional power grid key technology research and demonstration of high permeability intermittent energy" (2011AA05A105) and the national natural science foundation project < grid parameters A study on the estimation method of the Partition of Inspection.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM73

【相似文献】