Bayesian Model-based Generation of Synthetic Unbalanced Distribution Networks Incorporating Reliability Indices
This new model solves a critical data privacy problem for energy infrastructure.
Researchers have developed a new AI model that generates highly realistic synthetic power distribution networks, a crucial tool as real-world grid data is often restricted. The Bayesian Hierarchical Model uniquely incorporates key reliability metrics like CAIFI and CAIDI and ensures phase consistency, which previous methods overlooked. Validated on Brazilian networks, it accurately replicates power demand, phase allocation, and reliability, producing electrically feasible networks for planning and resilience studies without using sensitive real data.
Why It Matters
It enables secure, realistic testing and planning for critical energy infrastructure without compromising private grid data.