CAREER: Holistic Distributed Resource Management and Discovery via Augmented Learning and Robust Optimization

NSF CAREER项目旨在开发能源工程解决方案，以体现不同住宅电力消费者的偏好和需求。这项研究将使公用事业拥有和客户拥有的资产共同管理成为可能，并为分布式能源（如屋顶太阳能）和配电系统的运行方式带来革命性的变化。这一目标将通过利用人工智能算法来分析能源系统中的人在环组件，同时考虑到偏好和能源需求的多样性来实现。该项目的智力优势包括描述最终用户的消费行为，并使行为感知智能电网解决方案的设计成为可能，而无需对行为本身进行建模。该项目提出了细致的方法来推进网格边缘资源管理，同时考虑到关键因素。该项目的更广泛影响包括提高所有人特别是低收入社区的能源恢复能力。将设计体验式学习模块，向公众传授智能电网技术，以及公用事业和消费者拥有资产的先进共同管理的好处。它还将开始设计和开发一个跨学科的研究和教育项目，重点关注具有社会意识和公平的清洁能源工程解决方案，重点是为代表性不足的少数民族提供机会。该项目将开发自下而上的方法，以克服电网边缘资源和配电系统主动管理的多重障碍。将开发创新的方法，从智能电表数据中识别设备利用率，以实现非侵入式负载发现。新的人工智能算法（例如，因果条件隐半马尔可夫模型）将被开发用于行为感知的非侵入性负荷预测。提出了一种基于深浅联合神经网络的分布式能源混合匹配高效聚合方法。提出了一种新的不平衡交流最优潮流，以促进基于逆变器的分布式能源调度，同时确定了协调的最优逆变器控制模式及其设置。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This NSF CAREER project aims to develop energy engineering solutions that embody the preferences and needs of diverse residential electricity consumers. The research will enable co-management of utility-owned and customer-owned assets, and bring transformative changes to how distributed energy resources (e.g., rooftop solar) and power distribution systems are operated. This goal will be achieved by leveraging artificial intelligence algorithms to analyze the human-in-the-loop component of energy systems along with consideration of diversity in preferences and energy needs. The intellectual merits of the project include characterizing end users’ consumption behavior and enabling design of behavior-aware smart grid solutions without modeling the behavior itself. The project proposes meticulous methodologies to advance grid-edge resource management while accounting for critical factors. The broader impacts of the project include enhancing energy resilience for all specifically for low-income communities. Experiential learning modules will be designed to educate the public on smart grid technologies and benefits of advanced co-management of utility and consumer -owned assets. It will also initiate the design and development of an interdisciplinary research and educational program focused on socially-aware and equitable clean energy engineering solutions, with emphasis on opportunities for under-represented minorities.The project will develop ground-up approaches to overcome multiple hurdles for active management of grid-edge resources and power distribution systems. Innovative methodologies will be developed to identify appliance utilization from smart meter data to enable non-intrusive load discovery. Novel artificial intelligence algorithms (for example, causal conditional hidden semi-Markov model) will be developed for behavior-aware non-intrusive load forecasting. An approach based on combined deep-shallow neural networks will be developed for efficient aggregation of mix and match of distributed energy resources with complementary control capabilities. A novel unbalanced AC optimal power flow will be enhanced to facilitate inverter-based distributed energy resources scheduling while identifying coordinated optimal inverter control modes and their settings.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.