配电网台区用户智能识别系统研究

发布时间：2018-06-10 13:33

本文选题：蚁群算法 + 载波通信技术　；参考：《上海电力学院》2017年硕士论文

【摘要】：近年来,全国各地逐步建立了远程自动抄表系统用于远程抄表,然而,在一些老旧小区以及沿街门面,配电网台区用户混乱、不准确甚至缺失,严重影响了远程自动抄表系统准确管理台区用户及精细化统计分析台区线损等正常功能。针对这一问题本文提出了基于载波通信技术实现配电网台区用户智能识别系统,用于智能识别台区内配电变压器与采集器及用户的从属关系,并在远程监控中心显示识别结果。针对现有识别算法准确率不高的问题,本文还提出了结合蚁群、遗传和BP神经网络的等智能算法,该算法能够很好解决复杂线路情况下用户信息的准确识别,有效提高工作效率和识别准确性。经现场测试,识别仪在线识别率为97.7%,手持仪辅助现场识别可以实现用户信息全部识别,证明了方案的可行性和系统的可靠性。本课题主要开展了以下研究工作:1、搭建了配电网台区用户智能识别系统的架构。配电网台区用户智能识别系统由上位机、识别仪和手持仪三部分构成。识别仪通过广播通信主动上报方法识别台区下所有采集器信息,手持仪通过点对点通信被动查询方法识别具体某台采集器或某个用户信息,识别结果通过GPRS远程通信方式传输到远程监控中心的上位机。2、研究了配电网台区用户智能识别系统硬件设计方法。为方便系统安装和调试,系统硬件采用单元化设计方案,可方便更改电路设计和更换电路单元。识别仪和手持仪共同硬件单元主要有主控板单元、GPRS远程通信单元、电源单元和底板单元,主要区别在于识别仪采用三相载波通信单元,手持仪采用单相载波通信单元。3、研究了基于蚁群的配电网载波通信混合路由算法。为了进一步提高电力线载波通信可靠性,该算法将跳数、时延和丢包率等因素引入路径开销代价,以两节点之间的距离和通信信号品质为约束条件,求解在此条件下开销代价值最小的路由。此外,进一步改进了蚁群算法转移规则及信息素更新规则,并使用遗传算法初始化种群,采用模拟退火算法解决遗传算法内部闭环问题,可有效避免出现早熟现象。4、研究了基于BP神经网络的配电网台区智能识别方法。识别仪与手持仪的辅助应用可通过载波通信技术可初步识别出已识别用户、未识别用户和跨台区不确定用户,针对跨台区不确定用户,手持仪基于电力载波信号品质和前向BP神经网络进行台区用户智能识别,可提高台区用户识别准确度,实现台区用户全识别。5、对配电网台区用户智能识别系统进行了系统测试,测试结果显示通过识别仪和手持仪现场识别后,识别仪识别率高达97.7%,系统现场功能和性能良好,满足电力公司普查台区用户和相位信息要求。
[Abstract]:In recent years, remote automatic meter reading system has been gradually established in various parts of the country for remote meter reading. However, in some old residential areas and along the street, users in distribution network stations are confused, inaccurate or even missing. It seriously affects the normal functions of the remote automatic meter reading system, such as accurate management of users in the station area and fine statistical analysis of line loss of the station area. In order to solve this problem, this paper presents an intelligent identification system for distribution station users based on carrier communication technology, which can be used to identify the relationship between distribution transformers, collectors and users in the distribution network, and display the identification results in the remote monitoring center. In order to solve the problem that the accuracy of existing recognition algorithms is not high, this paper also puts forward some intelligent algorithms, such as ant colony, genetic and BP neural networks, which can solve the problem of accurate identification of user information in complex circuits. Effectively improve work efficiency and recognition accuracy. Field test shows that the on-line recognition rate of the recognizer is 97.7. the hand-held instrument can realize the identification of all the user information, which proves the feasibility of the scheme and the reliability of the system. In this paper, the following research work is carried out, and the structure of intelligent identification system for distribution station area is built. The intelligent identification system of distribution station area consists of upper computer, recognizer and handheld instrument. The recognizer identifies all the information of the collector in the station area by the method of active report of broadcast communication, and the handheld instrument recognizes the information of a specific collector or a user by means of passive inquiry method of point-to-point communication. The identification results are transmitted to the upper computer of the remote monitoring center by GPRS remote communication mode. The hardware design method of the intelligent identification system for distribution network stations is studied. In order to facilitate the installation and debugging of the system, the hardware of the system adopts a unit design scheme, which can easily change the circuit design and replace the circuit unit. The common hardware unit of the recognizer and the handheld instrument mainly consists of the main control board unit, the GPRS remote communication unit, the power supply unit and the bottom board unit. The main difference is that the recognition instrument adopts a three-phase carrier communication unit. A hybrid routing algorithm for distribution network carrier communication based on ant colony is studied by using single phase carrier communication unit. 3. In order to further improve the reliability of power line carrier communication, the algorithm introduces the number of hops, delay and packet loss rate into the cost of path overhead, which is constrained by the distance between two nodes and the quality of the communication signal. Under this condition, the routing with minimal overhead is solved. In addition, the ant colony algorithm transfer rule and pheromone update rule are further improved, and the genetic algorithm is used to initialize the population, and the simulated annealing algorithm is used to solve the internal closed-loop problem of genetic algorithm. The precocity phenomenon. 4 can be avoided effectively. The intelligent recognition method of distribution network based on BP neural network is studied. The auxiliary applications of the recognizer and the handheld instrument can be used to identify the identified user, the unrecognized user and the cross-station uncertain user through the carrier communication technology. Based on the power carrier signal quality and the forward BP neural network, the handheld instrument can improve the accuracy of station area user identification, realize the full identification of station area user .5. the system test of distribution network station area user intelligence identification system is carried out. The test results show that the recognition rate of the recognizer is as high as 97.7after the field identification of the instrument and the handheld instrument. The system has good field function and performance, which can meet the requirements of the users and phase information in the survey station area of the power company.
【学位授予单位】：上海电力学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM76

【参考文献】