Invention Reference Number
Digital twins (DTs) have emerged as essential tools for monitoring, predicting, and optimizing physical systems by using real-time data. Traditional DTs use machine learning and simulations, but these models often face challenges in real-time analysis of imminent scenarios due to limited memory and computational resources. There is no parallel in using simulation clones, especially for large-scale systems such as power grids, transportation, and communication networks.
Description
This technology introduces a simulation cloning framework that continuously creates a speculative "tree" of simulation clones. Each branch in the tree represents a different "what-if" scenario based on probabilistic events that could impact the system in real-time. These simulation clones are evaluated and pruned based on resource availability, ensuring efficient usage of memory and computational power. This framework enables a digital twin to predict future states and guide real-world systems, such as power grids, in response to events like transformer failures due to geomagnetic disturbances.
Benefits
- Scalability: Allows the management of numerous "what-if" simulations without overwhelming memory resources.
- Real-time prediction: Continuously evaluates scenarios, providing actionable insights for physical systems in real-time.
- Operational efficiency: Enables pre-computation of potential outcomes, allowing quicker responses to physical system changes.
- Resource optimization: Uses dynamic pruning to recycle memory and computational resources, enhancing performance over long periods.
Applications and Industries
- Power grid management: Predicts and mitigates the effects of events like transformer failures during solar storms.
- Critical infrastructure Monitoring: Ensures the stability of water treatment plants, nuclear reactors, or transportation networks.
- Real-time decision support: Provides predictive maintenance and operational guidance for large, complex engineered systems.
Contact
To learn more about this technology, email partnerships@ornl.gov or call 865-574-1051.
