Opportunistic Hybrid Communications Systems for Distributed PV Coordination
Increasing amounts of distributed solar PV power present fundamental challenges to the way the power system has traditionally been operated. With distributed solar power providing an increasing proportion of total generation, it must also take on greater responsibilities to ensure continued reliable and cost effective grid operations. New communications systems are needed to allow for bidirectional information exchange between distributed PV generators and various levels of the power system. Therefore the aim of the proposed research is to develop a novel opportunistic hybrid communications system that is better suited to meet the needs of monitoring and control of millions of distributed PV generators, while exploiting existing communications infrastructure to greatly reduce the costs necessary to provide these services. The development of a novel communications system architecture that can enable the coordination and control of millions of distributed PV generators will be one of the main outcomes of the proposed research. The design of novel publish-subscribe pattern based communications networks utilizing the existing wireless connections to PV inverters will facilitate the secure and private communications necessary to inform existing distribution and transmission level communications networks about the current PV system state. However, non-dedicated local wireless networks alone are not reliable enough to ensure the latency and availability required for reliable grid operations. For this reason a novel robust algorithmic layer is also proposed that will allow for decentralized state estimation and prediction schemes. These algorithms will allow for the accurate estimation of critical power system variables using minimal information exchange with distributed PV systems. Additionally the algorithms will be designed to be resilient to measurement outliers and missing measurements, using the large number of distributed PV systems to overcome the data availability and quality issues associated with any particular PV system. Big data analytics, machine learning technologies, and distributed computing will play an important role in the ability of the system to efficiently access, transfer, and store the large amounts of PV and network data associated with the immense number of total PV systems contributing information. To this end a sparse system representation and distributed algorithms will help to facilitate real-time execution of the critical information exchange. To be adopted in industry any new technology must be thoroughly vetted in an intensive verification and validation process. Therefore we are not just proposing a new communications paradigm for power systems operations, but also an extensive combined laboratory and computational testing procedure that can provide more than a simple proof of concept. Scale testing will be conducted at the level of up to 5M nodes through the integration of a computational simulation of the architecture and algorithms with a combined distribution-transmission power system simulation tool, the Integrated Grid Modeling System (IGMS) already developed at NREL. This combined validation approach will ensure that the proposed framework can provide the communications necessary to operate both distribution and transmission level power systems with extremely high penetrations of distributed PV power while fulfilling all of the SI communications system metrics.
last modified Nov 02, 2015 08:27 AM