CAREER: Towards Data-Driven and Field-Validated Microgrid Modeling and Analysis Techniques

这个NSF职业项目旨在通过使用微电网来提高电网的可靠性、安全性和弹性。微电网是当地的电力系统，可以与电网一起运行，并在紧急情况下与电网分开。微电网可以提高电网的弹性和可持续性，加快灾难恢复。该项目将通过解决理论研究和实际应用之间的差距，为微电网的设计和运行方式带来革命性的变化。为了实现这一目标，将采用最先进的数据驱动和机器学习算法。该项目的智力优势包括开发一种新的方法来准确建模真实世界的条件，使用机器学习来降低模型的复杂性，以及创建和现场验证微电网稳定性预测工具。该项目的更广泛影响包括改进了设计和运行微电网的方法，这将降低实施成本。通过降低成本，微电网可以在发展中国家和发达国家更快地部署。这将加快历史上被边缘化的社区的电气化进程，并提高电网的弹性、稳健性和可持续性。新创造的知识将通过电力工程方面的实践课程和讲习班进行传播。微电网的稳定性预测需要对物理系统进行精确的数学建模，以捕捉重要的动态和微妙之处。目前的建模实践没有考虑到两个关键的真实世界现象，即控制器饱和和保护动作，这两个现象对系统稳定性都有很大的影响。该项目的第一个技术贡献将通过开发一种同时模拟这两种现象的方法来解决这一差距。此外，微电网稳定性的研究也是通过线性或非线性技术来进行的。由于模型阶数和非线性的数量，稳定性技术可能会变得过于复杂。第二项技术贡献将利用科学机器学习(SciML)的进步，通过创建代理来降低系统的模型阶数。第三个技术贡献将是使用SciML的微电网稳定预测工具，该工具将预测不同运行条件和设计因素下的暂态稳定。来自工业级微电网和真实设备的数据将用于调整和确认这些替代品和工具的准确性。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This NSF CAREER project aims to increase the reliability, security, and resiliency of the electric power grid via the use of microgrids. Microgrids are local electric energy systems that can operate with the grid and separate from the grid during emergencies. Microgrids can improve grid resiliency and sustainability, and accelerate disaster recovery. The project will bring transformative change to how microgrids are designed and operated by addressing the gap between theoretical studies and real-world applications. To achieve this goal, state-of-the-art data-driven and machine learning algorithms will be employed. The intellectual merits of the project include developing a new approach to accurately model real-world conditions, using machine learning to reduce model complexity, and creating and field-validating a microgrid stability prediction tool. The broader impacts of the project include an improved method to design and operate microgrids which would reduce implementation costs. By reducing costs, microgrids can be deployed faster in both developing and developed nations. This would quicken the electrification of historically marginalized communities and improve grid resiliency, robustness, and sustainability. The newly created knowledge would be disseminated through hands-on courses and workshops on power engineering. Stability prediction for microgrids require accurate mathematical modeling of the physical system to capture important dynamics and subtleties. Current modeling practices do not account for two critical real-world phenomena, namely, controller saturation and protection action, both of which have drastic effects on system stability. The first technical contribution of this project will address this gap by developing an approach to concurrently model those two phenomena. Additionally, microgrid stability studies are approached through linear or nonlinear techniques. Stability techniques can become too complex due to model order and number of nonlinearities. The second technical contribution will leverage advances in Scientific Machine Learning (SciML) to reduce a system’s model order by creating surrogates. The third technical contribution will be a microgrid stability prediction tool using SciML that will predict transient stability under different operating conditions and design factors. Data from an industry-grade microgrid and real-world equipment will be used to tune and confirm the accuracy of those surrogates and tools.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.