AI-driven digital twin speeds up cyber-defense for control systems

January 31, 2019 // By Rich Pell
GE Research (Niskayuna, NY), the research and development division of General Electric, has announced a new real-time, active cyber-defense system for industrial control systems.

Called Digital Ghost, the system, says the company, offers a first-of-its-kind capability to both rapidly detect and neutralize cyber threats against industrial control systems (ICS). Residing beyond the traditional information and operational technology (IT/OT) firewalls, Digital Ghost lives inside the industrial control system itself.

It uses a network of sensors, controls, and key insights from AI-driven digital twins of key assets to continuously monitor and – when necessary – enable real-time actions to thwart cyber threats. In simulation testing and field trials, says the company, Digital Ghost has been shown to rapidly – within seconds – detect and isolate threats with 99% accuracy.

"Digital Ghost will give operators a new defense mechanism to safeguard critical infrastructure against the increasing number of cyberattacks made on their industrial control systems," says Justin John, Technology Director, Controls and Optimization at GE Research. "It not only will enable operators to more rapidly and accurately detect incoming cyber threats, it will help them neutralize these threats to keep operations running safely and reliably even while a cyberattack is occurring."

According to the company, Digital Ghost will help address two big challenges facing industrial control systems: The rapid rise in the number of cyberattacks on industrial control systems over the past five years, and the long lag times for detecting an attack. Recent ICS viruses have been shown to go undetected for months, even years.

A key enabler of Digital Ghost is GE's Digital Twin technology, in which GE researchers have created more than 1.2 million digital models of critical parts, assets, systems, and processes across the company's industrial product portfolio. GE's Digital Twins, says the company, are "living, learning models" that generate specific insights from operational and sensor data, human engineering expertise, fleet data, and simulation testing to deliver a desired business outcome.


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.