DYMONDS computer platform for smart grids

Abstract

In this paper we propose an actor-oriented computer platform, which is a natural outgrowth of our Dynamic Monitoring and Decision Systems (DYMONDS) modeling, communications and control paradigm for smart grids. We summarize this approach which is physics-based and explain how its actororiented structure can support a possible multi-level computer platform. By combining principles of model-based Ptolemy II computer platform with these models, we sketch out the basis for implementing a DYMONDS-based computer platform for smart grids. Both the dynamic internal models of polymorphic actors, as well as the system model of computing and communication protocols relating their external interfaces are described. Perhaps, most important is the fact that the proposed computer platform can evolve dynamically as operating conditions and status of actors vary, and, longer-term, as different components are added to the physical system and different smarts are being embedded, rather than having to re-program every time as the physical system changes; this is because new semantics are introduced at the component interaction level rather than at the programming language level.

Computer Architecture and Multi Time-Scale Implementations for Smart Grid in a Room Simulator

“…Instead, we propose to utilize our recently introduced physics-based modeling approach which supports interfaces mainly in terms of power exchange specifications, and, at the same time captures fast dynamic response of the system by simulating dynamics of equipment interacting only with local neighboring components [1,2]. The basic novel idea is to transform the existing state-space dynamic model of electric power components characterized by the natural response (time constants) of memory components (state variables) subject to Kirchhoff's laws.…”

“…Instead, we propose a state space transformation which retains all but two state variables of each dynamic component in the interconnected system and replaces the two eliminated state variables with the interaction variables in each module: the stored energy in the component and the rate of change of stored energy (power) exchanged between the component and the rest of the system. This approach is described in more detail in [1,2].…”

“…For this reason, a novel computer platform, called the Smart Grid in a Room Simulator (SGRS), is developed. The computer platform utilizes Carnegie Mellon University's physics-based Dynamic Monitoring and Decision Systems (DyMonDS) framework for modeling, simulating and designing cyber components needed to enhance performance of the electric power grids [1]. The SGRS runs different power system objects (components) distributed over multiple processors and computers, which makes this approach scalable for large systems.…”