中文摘要：

数据是智能电网建设的战略资源乃至主要驱动力。如何处理智能电网中呈现海量、多样、实时、真实等4个特征的4 Vs数据,并从中提取信息,是电力系统大数据建设所面临的核心问题。描述了大数据的特征和引入了随机矩阵理论作为基础,以及提出电力系统大数据的应用思路和架构。具体电力应用方面,介绍了所开发的早期事件发现、事件诊断和定位、相关性分析、可视化3D Power Map辅助展示等一系列功能。在此基础上,建立起以随机矩阵为理论基础,以数据为主要驱动力的电力系统认知体系框架,并探讨其与传统经典认知方案的区别。进一步设计案例考查了其对坏数据的鲁棒能力,其结果表明,随机矩阵理论这种工具可以有效地处理电网中的复杂数据,具有很好的学术研究意义和工程应用价值。另通过仿真算例验证了随机矩阵方案对数据异步的鲁棒性。

英文摘要：

Data become a strategic resource and even prime driving force for smart grid construction. Essentially, it is potential contained in data that attracts attention rather than massive data themselves. Mining value from 4V data（data with features of volume, variety, velocity and veracity） within tolerant resources（time, hardware, human, etc.） is a key challenge. Definition of big data and random matrix theory, and related architecture and methods are introduced as foundations. A series of functions related to situation awareness, including early event detection, fault diagnosis and location, correlation analysis and auxiliary 3D power map are developed as concrete applications. A big data application framework, mainly based on random matrix theory and driven by data, is constructed to conduct situation awareness. Advantages of the proposed approach are elaborated, and, especially, robustness against bad data is highlighted. A case study is designed to validate this approach in dealing with unsynchronized data.