Southern California Edison (SCE) is working with GridCARE to assess whether physics-based artificial intelligence can unlock additional capacity on constrained parts of its electricity network and accelerate connections for new customers.
An initial study of a constrained circuit in California’s Inland Empire region identified the potential for 20 MW of additional flexible capacity, which could support future customer connections while meeting reliability requirements, the companies said.
The analysis used GridCARE’s Power Acceleration technology to examine the capacity available from existing grid assets. The identified capacity remains subject to validation by utility engineers; the announcement did not specify when it might be made available for new connections.
SCE is assessing the technology at several other locations across its network. Potential applications include electric vehicle charging, manufacturing, commercial facilities and other customers whose connections may be limited by local network constraints.
“Combined with demand flexibility, it has the potential to accelerate our ability to meet new electricity demand and make the grid more affordable for customers,” said Shinjini Menon, SCE’s senior vice-president of System Planning and Engineering.
GridCARE said its approach uses physics-based AI to assess combinations of demand, weather conditions, equipment outages and flexibility resources. Utility engineers then test potential solutions using their own planning tools to determine whether operating limits and reliability requirements can be maintained.
The work comes as utilities face growing electricity demand and pressure to connect new loads without relying solely on major infrastructure upgrades. The companies did not disclose the cost of the study or set out a timetable for decisions on deploying the technology more widely.
GridCARE also cited Duke University research estimating that existing US power systems could accommodate nearly 100 GW of additional load at an annual curtailment rate of 0.5%. That estimate is a broader system-level finding and does not establish the capacity available on any individual circuit.