Cybersecurity and Power System Stability
The Cybersecurity and Power Systems Stability team studies dynamics in the electric power grid that arise from the complex interaction of Distributed Energy Resources (DER), such as solar photovoltaic and battery storage systems and the legacy system.
Cybersecurity and Power System Stability
As more individuals and companies utilize solar photovoltaics, battery storage devices and other distributed energy resources, and with so many of these resources connecting to the power grid, there is a growing concern among utility companies about potential cyberattacks. When solar PV and battery storage devices are connected to the power grid, utility companies have less control over the operation of these distributed resources, and this puts public infrastructure at risk.
GIG's Cybersecurity and Power System Stability team applies tools from dynamic systems and control, optimization and machine learning to identify and mitigate grid-destabilizing cybersecurity threats. Our specific projects in this area focus on grids with high penetrations of solar photovoltaic generation and battery storage systems. Our work is sponsored by the Cybersecurity for Energy Delivery Systems (CEDS) program within the Cybersecurity, Energy Security and Emergency Response (CESER) office.
Project Cybersecurity via Inverter Grid Automatic Reconfiguration (CIGAR) focuses on mitigating the effects of cyberattacks on DER smart inverters whose autonomous control settings have been maliciously adjusted to create voltage oscillations in distribution grids. In project CIGAR, our team is leveraging deep reinforcement learning to train a neural network to manipulate the settings of DER that have not been compromised via cyberattack to mitigate the destabilizing effects of hacked units. Our partners on this project are: the National Rural Electric Cooperative Association (NRECA), Siemens Corporate Technologies, and Arizona State University.
Project Supervisor Parameter Adjustment via Distribution Energy Storage (SPADES) extends the capabilities of CIGAR to include energy storage (e.g. batteries) as possible threat vectors and assets that can be used to mitigate the effect of cyber attacks on electric grids. In SPADES, our team is using deep reinforcement learning and optimal control to develop strategies to mitigate attacks on the internal power electronic control systems within storage devices, as well as the interaction of storage systems with the electric grid. Our partners on this project are: the National Rural Electric Cooperative Association (NRECA), Siemens Corporate Technologies, and Arizona State University.
This project is developing a new class of measuring instrument called the GridSweep probe to reveal subtle dynamics that can threaten electric grid stability. Although widely recognized, these dynamics have never before been directly observable. This type of probe should advance bulk grid reliability, security, resilience, and the capability to operate grids securely with high contributions from inverter-based generation.
The electric grid has incipient instabilities that are known or suspected, but not easily measured. These include reduced system inertia, vulnerability to forced oscillations, and adverse interactions from inverter controls. Grids are also vulnerable to time-synchronized dispersed load changes, whether intentional or unintentional. This project will develop a new type of probe to provide visibility of grid frequency response behavior, with the goal of mitigating the risk of blackouts and supporting intelligent decision making for critical infrastructure.
- Create a completely new class of active grid instrumentation for situational awareness of bulk grid and connected resources.
- Apply data techniques from seismology for ambient noise analysis and small-signal extraction.
- Characterize bulk grid inertia, generator control loop parameters, frequency-specific grid response and location-specific load dynamics live in situ.
- Demonstrate a synchronized GridSweeper network with geographic correlation.
This project builds on prior work funded by DARPA, the Grid Thumper, which injected a precisely timed 1-MW impulse and observed it, tens of miles away, using high-precision microPMUs with a measurement resolution of 10 PPM. The graphic illustrates the observation in San Rafael, CA, of a regular "thump" injected in Alameda, CA. GridSweep probes will be much smaller and portable than the GridThumper, made possible by new techniques for injecting a recognizable signal, along with new hardware and algorithms for measuring that signal miles away at 100 PPM resolution. One approach for making such a tiny signal observable will be an ambient noise correlation algorithm borrowed from seismology. After testing the prototype and algorithms in the lab setting, the project will deploy GridSweep probes and characterize the frequency-specific response of the grid in different geographies. In the future, this will inform strategies to protect power systems from certain types of threats, and to design inverter control systems that help stabilize the grid.
This project is funded by the 2019 Grid Modernization Lab Call, Topic Area 3: Advanced Sensors and Data Analytics, Subtopic 3.4.2: Monitoring for Critical Infrastructure Interdependencies.
Principal Investigator: Alexandra von Meier, Berkeley Lab
Co-PI: Philip Top, LLNL
Senior Personnel: Dan Arnold (Berkeley Lab), Kristina Hamachi LaCommare (Berkeley Lab), Alex McEachern (Lead at McEachern Laboratories), Eric Matzel (LLNL)
Lawrence Livermore National Laboratory
Hawaiian Electric Co.