Visual Analysis Based On Simulation Data Of Electric Power System

Along with the expanding scale and rising voltage of modern power gird,simulation analysis plays an important role in power grid research.However,traditional analytical methods lack effective tools and means in face of big data generated from power grid simulation,therefore it is hard for analysts to quickly master the operation information,and gain insights of system rules.This paper focuses on high-dimensional and time-series data generated from simulation analysis of power system,and proposes a correlation analysis method which joins together spatial and time distribution based on visual analysis.Spatial distribution views are designed for visual analysis of high-dimensional simulation data based on geographical topology,using layers to show data of different dimensions;while time-varying graphs are used to reveal patterns of large scale time series data.Through correlation of the ranges of time and variables,analysts are able to observe synergistic change patterns of the two views with the help of interactive tools.This method combines existing professional knowledge with the simulation analysis process,making it beneficial supplement to the existing visualization tools.Main contributions of the paper include the following three aspects:1)Layered visual analysis method is proposed to show the distribution of data.Data of different variables are placed in different layers in order to display multidimensional data in the same view at the same time.2)Apply kernel density estimation to the visualization of high-dimensional time-series data.Analysts are able to observe the pattern and distribution of thousands of curves at the same time,by firstly clustering the data based on their characteristics,then mapping it to two-dimensional coordinate system based on their density distribution.3)Propose a correlation analysis method between the previous two views,allowing analysts to study the relation and rules between temporal and spatial characteristics of data.The practicality and effectiveness of the proposed method are verified by several cases.