源网荷全景同步测量系统分布式主站的设计与实现

源网荷全景同步测量系统（Full-View Synchronized Measurement System ,SYMS）实现了对高比例电力电子化的新型电力系统的全方位实时检测。SYMS 主站作为测量数据的接收与分析平台，其数据处理能力对保障SYMS稳定运行与应用效果具有重要意义。然而，由于数据类型多、装置数量激增、数据分析过程复杂，集中式架构的主站难以保证同步相量数据处理的实时性与可靠性。因此，本文提出一种适用于SYMS系统的分布式主站设计方法，并进行了实现。首先，该方法分析了SYMS不同测量装置的通信方式，提出并建立了基于开源相量数据集中器（Open Source Phasor Data Concentrator, OpenPDC）的多源异类数据适配器，并设计实现了基于HAProxy的主站前置服务器的负载均衡集群；其次，针对主站数据在线分析延时高的问题，设计并开发了基于流处理框架的Storm适配多语言、多时间窗算法的分布式计算方式，进一步搭建了包含前置数据平台与在线应用平台的SYMS分布式主站；实际系统测试结果表明，与集中式架构相比，所提架构在实际运行中可有效均衡负载，利用分布式方式提高运算速度，针对多类型实时测量数据具有更强的并发处理能力与更短的处理延时，并有效监测系统的异常状态。可为新型电力系统特性分析、建模、闭环控制等应用提供数据基础与应用平台。

Design and Realization of the Distributed Data Center of Full-view Synchronized Measurement System

The full-view synchronized measurement system (SYMS) has realized the all-directional real-time detection of power systems with high proportion of power electronics. As a platform for receiving and analyzing measurement data, the data processing ability of the SYMS master station is of great significance to ensure the stable operation and application effect of the SYMS. However, due to the reasons as the large number of data types, the surge in the number of devices, and the complexity of the data analysis process, it is difficult for SYMS master station with a centralized architecture to guarantee the real-time performance and reliability of the data processing. Therefore, a distributed master station design method for the SYMS system was proposed and implemented in this paper. Firstly, the proposed method analyzed the communication modes of different measurement devices of SYMS, established a multi-source heterogeneous data adapter based on the open source phasor data concentrator. This paper also designed and implemented a load balancing cluster of the main station front-end server based on the HAProxy. Secondly, aiming at the problem of high delay of the data online analysis of master stations, the Storm based on the stream processing framework was designed and developed to adapt the distributed computing method with multi-language and multi-time window algorithm. The SYMS distributed master station including the pre-data processing platform and the online application platform was further built. The actual system test results show that, compared with the centralized architecture, the proposed architecture can effectively balance the load in actual operation and improve the computing speed through the distributed method. The proposed architecture also have stronger concurrent processing capabilities and shorter delay for multi-type real-time measurement data. Furthermore, it can effectively monitor the abnormal state of the system. Therefore, it can provide the data foundation and application platform for characteristic analysis, modeling, closed-loop control and other applications of the new power system.