EAGER: SSDIM: Simulated and Synthetic Data Generation for Interdependent Natural Gas and Electrical Power Systems Based on Graph Theory and Machine Learning

天然气和电力系统已成为美国能源基础设施的支柱。这一早期概念探索性研究补助金(AGER)项目将研究基于数据的实用方法，以产生模拟和合成数据集，从机械和人类两个方面忠实地表示两个关键基础设施系统之间的相互依赖。该项目通过为相互依存的关键能源基础设施的未来研究奠定数据基础，为能源系统现代化的重大国家挑战做出了贡献。研究成果将为研究生和本科生带来出版以及整合数据分析和能源工程的多学科培训机会。通过与中西部的公用事业公司和国家实验室建立战略联盟，将向广泛的工程师和研究人员提供关于天然气和电力网络数据分析的网络研讨会，介绍这项工作的结果。该项目将开创基于数据的方法，以了解天然气和电力网络之间的相互依存关系并建立模型。该交互数据生成方法提供具有不同时空粒度和操作条件的高保真数据集。特别是，利用基于图表和基于学习的方法，从实际数据中确定了燃气-电力系统固有的机械原理和人类影响。这些生成的数据集将使用实际数据进行验证，并通过项目网站在线提供，以及利用这些数据和现有建模方法的用例列表，以促进对天然气/电力网络在强/弱耦合效应、经济运行、级联停电等方面的相互依赖的理解。该项目促进了数据分析、图论、复杂网络以及电力和天然气工程等科学和技术领域的跨学科努力，为关键的相互依赖的基础设施提供有关合成数据生成的基础知识。

Natural gas and electric power systems have become the backbone of the U.S. energy infrastructure. This EArly-concept Grant for Exploratory Research (EAGER) project will investigate practical data-based approaches for producing simulated and synthetic datasets that faithfully represent the interdependence between the two critical infrastructure systems from mechanistic and human aspects. The project contributes to the grand national challenge of modernizing energy systems by laying the data foundation for future research in interdependent critical energy infrastructures. The research results will lead to publications as well as multi-disciplinary training opportunities that integrate data analysis and energy engineering for graduate and undergraduate students. By forging strategic alliances with the utilities in the Midwest and national laboratories, webinars on natural gas and power network data analysis will be given to a broad array of engineers and researchers on the results of this work. The project will pioneer data-based approaches to understand and model the interdependence between natural gas and power networks. The interactive data generation method provides high-fidelity datasets with different spatial-temporal granularities and operation conditions. In particular, mechanistic principles and human impacts inherent in gas-electric systems are identified from practical data using graph-based and learning-based approaches. These generated datasets will be validated using practical data, and be available online through a project website, together with a list of use cases that leverage the data and existing modeling approaches to advance the understanding of gas/power network interdependence in terms of strong/weak coupling effects, economic operations, cascading outages, etc. The project promotes an interdisciplinary effort in science and technology from data analysis, graph theories, complex networks, as well as power and natural gas engineering to provide fundamental knowledge about synthetic data generation for critical interdependent infrastructures.