(92a) Integrated Grid Modeling System: A Distributed Solar Power Case Study

The future electricity system is a distributed control problem on an enormous scale: millions of generation and load devices requiring coordinated but decentralized decision making.  The interactions between a residential rooftop PV system and the household appliances will feed up into the local distribution system.  The interactions among the different residential and commercial systems feed into a distribution level supply or demand that will be further fed up into the transmission system.  At the transmission level coordination is necessary to be able to provide ancillary services or to link sub-systems where  supply and demand are mismatched with those that currently have the opposite power balance due to fluctuating conditions. The cyber (information flow) layer architecture formalizing the interactions between these different systems needs to be developed while respecting that the physical electricity interactions are dictated by the physics of the electricity and the communication systems.

Historically in the transmission and distribution networks have been modeled independently of one another due to computational limitations. However, this does not allow for an understanding of the impact of distributed devices on bulk system operations, nor of the impact of transmission level policies on distributed generation. In this work we have developed a simulation platform that ranges from the appliance-level to the Independent System Operator (ISO) level which allows for the investigation of these impacts. The open source software GridLab-D is utilized for distribution level system modeling, and has been linked with the FESTIV transmission level model to represent both ends of the power system operations spectrum.  As most transmission level substations have multiple distribution feeders, and the typical timescale for feeders is much lower than for bulk system operations, thousands of GridLab-D instances must be run to supply the information for each FESTIV time step. These simulations are only possible to be completed in realistic timescales due to the use of NREL’s Peregrine high performance computing system.

The testbed has been utilized to study the impacts of high penetration distributed solar power on transmission level system operations. Distributed solar power installations are currently growing very quickly in a number of areas of the United States, and are beginning to impact transmission system operations. For example, while system operators generally do not have any visibility into distributed solar installations or their power output, and only have visibility into their output based on deviations from their load forecasts. The case study examined includes the impact of high penetration distributed solar power on the unit commitment and economic dispatch process, bus voltage levels, and distributed solar power forecasting, among other options.