Develop add-on control layer between existing centralized coordinator and local primary controls to promote and enforce real-time safety and resilience both locally and across the system.

PI: Sai Pushpak Nandanoori, saipushpak.n@pnnl.gov

This project will develop an add-on control layer between the existing centralized coordinator and local primary controls to promote and enforce real-time safety and resilience both locally and across the system. System-level resilience will be achieved by developing solutions at two stages—autonomous decentralized controls and a complementary semi-cooperative coordination of the local autonomous actions in a communication-sparse environment.

• Goal I: Design resilient local autonomous controls and baseline cooperative controls.

• Goal II: Preliminary validation of the autonomous controls on the 123-node microgrid use-case system.

The proactive control will be designed based on historical data and expected adversarial events to assure no setpoint that violates transient safety constraints reaches the local controller for execution. The reactive control agent, on the other hand, will provide online adaptive actions based on real-time local detection of adversarial events. The reactive agent will come into play when an unforeseen event pushes the system states outside the safe operating regime. The cross-layered cooperative control strategy will complement the actions of the local control agents by realigning the control set points across the system. The autonomous controls will be lightweight solutions derived from constrained control methods. The semi-cooperative coordination will combine information theory and network control to identify and design a minimal set of control agents for assured resilience under communication-sparse environments.

View all RD2C projects