Cybersecurity/AI, & its Use in Grid Protection/Control

The Company

The Department of Electrical and Computer Engineering at Iowa State University is a leading institution in power systems research. Under the guidance of renowned cybersecurity expert Dr. Manimaran Govindarasu, PhD candidate Vivek-Kumar Singh conducts cutting-edge R&D at the intersection of power systems, artificial intelligence, and cyber-physical security to protect critical grid infrastructure from sophisticated cyber threats.

The Challenges

• Lack of Open-Source Cybersecurity Datasets: Due to security risks and corporate confidentiality, utilities cannot share operational attack data, leaving researchers without realistic datasets to train AI/ML models.
• Complex Threat Detection: Distinguishing between benign grid control commands and malicious cyber intrusions across Wide-Area Monitoring, Protection & Control (WAMPAC) systems requires high-fidelity dynamic modeling.
• Hardware & Dynamic Interaction: Relays and Synchrophasors (PMUs) cannot natively distinguish between physical faults and cyberattacks, necessitating realistic dynamic data injection.
• Multi-Site Telecom Latency: Simulating geographically distributed control centers and substations requires accounting for real-world communication delays and network latency without compromising model accuracy.

The OPAL-RT Solution

Iowa State University leveraged OPAL-RT’s real-time hardware, software, and communication protocols to create a comprehensive cyber-physical testbed:

• Attack Staging & Dataset Generation: Used OPAL-RT real-time simulators and ePHASORSIM to inject dynamic faults and stage cyberattacks, generating proprietary-safe “good and bad” training datasets for machine learning.
• Hardware-in-the-Loop (HIL) Integration: Connected physical relays and PMUs directly to OPAL-RT hardware, enabling realistic playback and real-time dynamic response observation.
• Multi-Protocol Communication Support: Utilized native OPAL-RT drivers for IEC 61850 GOOSE, DNP3 (via OPC drivers), and C37.118 PMU protocols to emulate real-world SCADA and synchrophasor network architecture.
• Remote Co-Simulation Federation: Connected two OPAL-RT simulators remotely—one in Iowa acting as the control center and one at the US Army Research Laboratory in Maryland modeling IEEE 39-bus substations—to measure 26.7 ms synchrophasor data latency under attack conditions.

The Results

• AI Anomaly Detection: Successfully developed and validated machine learning algorithms capable of real-time anomaly detection in WAMPAC environments despite real-world network latency.
• Utility Cyber Training: Provided practical SCADA cybersecurity training to major utilities including Cedar Falls Utilities (CFU), CIPCO, MidAmerican Energy, Idaho Power Company, and Corn Belt Power Cooperative (CBPC).
• Defense & Federal Collaboration: Established a multi-substation cyber-federated testbed in partnership with the US Army Research Laboratory for defense-grade grid protection research.
• Dataset Sharing for Scientific Progress: Plan to release non-proprietary PMU attack signature datasets to accelerate global smart grid cybersecurity R&D.

Read the full story