Watch how the same company's claims tighten over three years. In 2020, GE's grid-emulation patent fenced off a capability (an inverter that acts like the grid). In 2023, US11680558B2, granted June 20, 2023, fences off a technique: grid-forming control using virtual impedance. That is a real evolution in claim strategy.
Virtual impedance is the load-bearing limitation. The idea is to make the inverter's control loop behave as though a physical impedance — a resistance and inductance — sits at its terminals, even though no such component is there. This synthetic impedance improves how multiple grid-forming inverters share current and damps instability. Claim 1 is about implementing that synthetic element.
Why narrower is stronger here. A claim on "inverter establishes voltage and frequency" (the 2020 emulation idea) drowns in prior art and is easy to design around. A claim on "achieve grid-forming via a specific virtual-impedance method" is harder to invalidate and harder to replicate without using the same trick. The fence is smaller but the posts are deeper.
The grid stake is the central problem of inverter-dominated grids: with too few spinning machines, the inverters themselves must form the grid, and multiple grid-forming inverters must coexist without fighting each other. Virtual impedance is one of the established tools for making them play nicely — and GE is fencing its implementation of it.
The discipline: claim 1 owns a control method, not grid-forming and not the inverter. Competitors using droop control, virtual synchronous machine models, or other grid-forming approaches are outside this fence. The patent's value is exactly proportional to how essential GE's virtual-impedance formulation turns out to be — and the active citation traffic around grid-forming suggests the field is still filling with alternatives.