EAGER: Real-Time: Visual Analytics for Enhanced Decision-Making and Situational Awareness in Modern Distribution Systems, with a Focus on Outage Prediction and Management

EAGER：实时：视觉分析可增强现代配电系统中的决策和态势感知，重点关注停电预测和管理

• 批准号: 1839812
• 负责人: Valentina Cecchi
• 金额: $ 29.92万
• 依托单位: University of North Carolina at Charlotte
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-10-01 至 2022-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1839812&HistoricalAwards=false
• 关键词: EAGER; Real; Time; Visual; Analytics

This project addresses the need to develop a fully visible, controllable, and resilient electric power distribution system. Although the distribution system accounts for over 75% of the outages in the power grid, power system operators currently have limited situational awareness at the distribution level, limited visibility beyond the substations, limited information on the network and its connectivity. With the rapid increase in amount and type of data accumulated in the distribution system, i.e. from Advanced Metering Infrastructure (AMI) and remotely monitored and controlled devices, the need to glean actionable intelligence from it is of paramount importance. Integrating data from these heterogeneous datasets (Geographical Information System (GIS), Supervisory Control and Data Acquisition (SCADA), AMI, Outage and Distribution Management Systems (OMS/DMS)) is a first step towards achieving this. The goal of this project is to develop a visual analytics platform to leverage the integrated dataset, enabling distribution system operators to visualize and analyze the state of the distribution system over time, empowering them to identify categorical patterns of events in space and time via highly coordinated visualizations. The project comprises three main components: 1. A data-driven approach to uncover useful information from streaming and historical data, strengthened by 2. Situationally-aware modeling and simulation of the electric power distribution system, and 3. A visual analytics system, leading to prescriptive analytics, actionable knowledge for the indispensable human in-the-loop. The developed infrastructure would enable and enhance real-time fast and confident decision-making, thus supporting overall efficiency and reliability improvements in the electric power distribution system, reducing current outage times and improving reliability indices. Benefits would accrue in terms of savings as well as in terms of customer satisfaction. The principal investigator of this project is the founder and faculty advisor of the local student chapter of IEEE Women in Engineering (WIE) and she plans to leverage that connection to broaden participation in this project. For complete situational awareness leading to decision-making, there needs to be powerful, automated analytics, enhanced by a deep understanding of the physical system, but it is also essential that the human be placed in the loop at the right place and time. For this reason, three components are integrated: 1. Data-driven real-time probabilistic outage prediction, coupled with 2. Situationally-aware modeling and simulation of the distribution system, and with 3. An interactive visual analytics system to provide visual and exploratory analytics. Utilizing real-time data streams coming from the distribution system (SCADA, AMI, Digital Fault Recorders) and from weather stations, together with historical data, advanced real-time data analytics algorithms are applied to strategically process the data, obtaining, for example, accurate loading conditions and developing probabilistic outage predictions. This sensed and processed information is then utilized, as needed, by the now situationally-aware distribution system electrical model to perform simulations. The simulations in turn provide input feedback to probabilistic scenarios and define system constraints for the analytics modules. Since human-in-the-loop is important, these analytics are closely coupled to the representations and modes of interaction in the interactive visual analytics system so that the right information is presented to the user at the right time. Thus, the developed framework results in a hybrid approach that leverages data-driven methodologies, the physical system, and visual analytics, to provide much improved decision-making capabilities. This project investigates the science of real-time learning and decision-making, while also looking closely at the technology for data-driven distribution system analysis coupled with deep understanding of the physical system. The project will provide theoretical underpinnings and novel methods to significantly move distribution modernization efforts forward, including improvements in system reliability and resiliency, new modes for management, new paradigms and paths for customer-utility cooperation, and a new approach to handling big data in the electric grid.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

该项目解决了开发完全可见，可控和弹性配电系统的需求。尽管配电系统占电网停电的75%以上，但电力系统运营商目前在配电层面的态势感知有限，变电站以外的可见性有限，有关网络及其连接的信息有限。 随着配电系统中积累的数据（即来自高级计量基础设施（AMI）和远程监控和控制设备的数据）的数量和类型的快速增加，从中收集可操作的情报的需求至关重要。整合这些异构数据集（地理信息系统（GIS），监控和数据采集（SCADA），AMI，停电和配电管理系统（OMS/DMS））的数据是实现这一目标的第一步。 该项目的目标是开发一个可视化分析平台，以利用集成数据集，使配电系统运营商能够可视化和分析配电系统随时间的状态，使他们能够通过高度协调的可视化来识别空间和时间上的事件分类模式。该项目包括三个主要部分：1.一种数据驱动的方法，从流数据和历史数据中发现有用的信息，通过2.配电系统的情景感知建模与仿真; 3.可视化分析系统，为不可或缺的人工参与者提供规范性分析和可操作的知识。 开发的基础设施将实现并加强实时快速和自信的决策，从而支持配电系统的整体效率和可靠性的提高，减少当前的停电时间并提高可靠性指标。在节省费用和提高客户满意度方面都会产生好处。该项目的主要研究者是IEEE Women in Engineering（WIE）当地学生分会的创始人和教师顾问，她计划利用这种联系来扩大对该项目的参与。 对于导致决策的完整态势感知，需要强大的自动化分析，通过对物理系统的深入理解来增强，但在正确的地点和时间将人置于循环中也是至关重要的。为此，三个组成部分被整合：1。数据驱动的实时概率停电预测，加上2。配电系统的情景感知建模与仿真，并与3。交互式可视化分析系统，提供可视化和探索性分析。 利用来自配电系统（SCADA、AMI、数字故障记录仪）和气象站的实时数据流，以及历史数据，应用先进的实时数据分析算法对数据进行战略性处理，例如获得准确的负载条件和开发概率停电预测。该感测和处理的信息然后根据需要由现在的情况感知的配电系统电气模型利用以执行仿真。模拟反过来为概率场景提供输入反馈，并为分析模块定义系统约束。由于人在回路中很重要，这些分析与交互式视觉分析系统中的表示和交互模式紧密耦合，以便在正确的时间向用户呈现正确的信息。因此，所开发的框架产生了一种混合方法，该方法利用数据驱动的方法、物理系统和可视化分析来提供更好的决策能力。 该项目研究了实时学习和决策的科学，同时还密切关注数据驱动的配电系统分析技术，以及对物理系统的深入理解。该项目将提供理论基础和新方法，以显着推动配电现代化的努力，包括提高系统的可靠性和弹性，新的管理模式，新的范例和客户公用事业合作的路径，以及处理电网大数据的新方法。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准。

项目成果

Loss Estimation and Visualization in Distribution Systems using AMI and Recloser Data

使用 AMI 和重合器数据进行配电系统的损耗估计和可视化

• DOI: 10.1109/td39804.2020.9299891
• 发表时间: 2020
• 期刊: 2020 IEEE/PES Transmission and Distribution Conference and Exposition (T&D
• 作者: Schulz, Dominik;Lawanson, Tumininu;Ravikumar, Kiran;Cecchi, Valentina
• 通讯作者: Cecchi, Valentina

Big data analytics for future electricity grids

• DOI: 10.1016/j.epsr.2020.106788
• 发表时间: 2020-12
• 期刊: Electric Power Systems Research
• 影响因子: 3.9
• 作者: M. Kezunovic;P. Pinson;Z. Obradovic;S. Grijalva;Tao Hong;R. Bessa

Analysis of Outage Frequency and Duration in Distribution Systems using Machine Learning

使用机器学习分析配电系统的停电频率和持续时间

• DOI: 10.1109/naps50074.2021.9449708
• 发表时间: 2021
• 期刊: 2020 52nd North American Power Symposium
• 作者: Lawanson, Tumininu;Sharma, Vinayak;Cecchi, Valentina;Hong, Tao
• 通讯作者: Hong, Tao

Improving Power Distribution System Situational Awareness Using Visual Analytics

使用可视化分析提高配电系统态势感知

• DOI: 10.1109/secon.2018.8479204
• 发表时间: 2018
• 期刊: IEEE SoutheastCon 2018
• 作者: Lawanson, Tumininu;Karandeh, Roozbeh;Cecchi, Valentina;Wartell, Zachary;Cho, Isaac
• 通讯作者: Cho, Isaac