| Description | Modeling and Control of Multi-Energy Dynamical Systems: Hidden Paths to Decarbonizaton In this talk we first illustrate several examples of microgrids, distribution systems and large-scale bulk power systems comprising diverse distributed energy resources (DERs). We describe fundamental problems with today’s operating and planning practice when attempting to integrate solar, wind intermittent resources and storage. We illustrate examples from Azores Islands and from continental power grids showing typical delivery and power balancing problems with these new resources. We show how frequency stabilization and regulation can be done by using distributed storage, such as flywheels (for managing large sudden wind gusts); STATCOMS (for managing short sudden wind gusts), and distributed roof-top PVs, HVACs and EVs. We put forward a Dynamic Monitoring and Decision Systems (DyMonDS) architecture in support of overcoming these problems by means of cyber-physical platform. Next, we propose that the main challenges in implementing interactive distributed integration of new technologies come from: 1) not having provable performance of system modules (components, subsystems), and 2) lack of physics-based protocols for operating the interconnected system by having confidence that operation would be feasible, stable and robust. To overcome these problems, we first revisit modeling and control used in today’s power systems and identify open R&D problems which must be resolved on the way to decarbonization through digitalization and more interactive distributed approaches. Second, we present our unified energy modeling for control which is multi-layered. The models of components are technology specific, yet they have the structure which utilizes the concept of interaction variables as a means of both characterizing components and supporting protocols for provable performance. Starting from first principles we show existence of such variables and their interpretation in terms of {stored energy; power; rate of change of power}. These aggregate energy models are relevant for assessing and controlling interaction dynamics and can be verified using only interface measurements, and not requiring technology specific internal knowledge of components design and control. Notably, this approach can be further generalized to characterize interaction variable for their marginal cost and emissions. This approach sets the basis for multi-energy multi-disciplinary innovations with clear understanding of potential to do useful work (exergy) and fundamentally waster energy (anergy). We discuss open questions and future work needed to technology-agnostic exploration of candidate architectures and their performance. |
|---|