【作者】 王琦；

【导师】 陈为；

【作者基本信息】 浙江大学 ， 计算机科学， 2018， 硕士

【摘要】 因为电网系统不稳定而引发的严重停电问题可以导致巨大的经济损失,电力系统的稳定性分析一直是电力行业的核心问题。传统上,电力行业工作者采用计算机模拟的方式对电网稳定性进行定量判断,这种方法缺乏深入数据的分析工具,难以进行如事态感知、不稳定性分析推理及模式识别等的分析任务。作者所在团队与中国电力科学研究院合作,面向电网系统的暂态稳定性开展可视分析研究,开发构建了一款针对电网仿真数据的可视分析系统。本文所描述的工作具有三个重要贡献:(1)设计并实现了一款支持用户对电网的异常、不稳定及故障进行识别与推理的可视分析系统;(2)将视觉编码与统计质量控制方法相结合,提供了一种能够考察多变量对系统稳定性影响的研究方法;(3)一套用于多变量时变数据、能使分析人员随时检测电网全局变化趋势、局部扰动和多变量间相关性的可视化设计。目前,本系统已在中国电力科学研究院部署使用,以丰富的特性与高效的性能获得了电力行业工作者的广泛认可。本文主要从电力仿真数据可视化方法和系统技术选型与实现两方面介绍作者在本文系统开发构建过程中所做的贡献。更多还原

【Abstract】 Studying the stability of power system has long been a significant problem in the electricity industry because severe blackouts caused by system instability can lead to huge economic loss.Conventional approaches employ computational simulation to make quantitative judgement of the system stability.Such methods,however,lack in-depth data analysis tools to support analytical tasks such as situation-aware analysis,instability reasoning and pattern recognition.To facilitate visual exploration and reasoning on the power grid simulation data,we propose a visual analytics system in cooperation with domain experts from China Electronic Power Research Institute.In summary,our major contributions are as follows:(1)We design and implement a visual analytics system that supports the exploration,identification and inference of abnormality,instability and fault of a power grid;(2)We combine visual encoding with statistical quality control methods to support interactive visual diagnostics of overall system stability affected by multiple variables;(3)We propose a suite of visual designs for numerical time-series that empowers the analyst with capabilities of visual inspection of global trends,local variations and correlations of multiple variables over time.This system has now been deployed in China Electronic Power Research Institute and has been appreciated there because of its rich features and high performance.This paper will introduce the author’s contributions to this system from the following two perspectives:visual analytics methods for power grid simulation data and system implementation.更多还原