The distinction between a grid-following and a grid-forming inverter is the cleanest way to understand why grid-forming technology matters. A conventional inverter is grid-following: it measures the voltage and frequency already present on the grid and injects current in phase with that reference. It is a follower by design, and it needs something else on the network — historically, large spinning synchronous generators — to establish the voltage and frequency it locks onto. A grid-forming inverter inverts that relationship. It behaves as a voltage source, setting a voltage magnitude and frequency itself, so that it can help hold the grid up rather than merely feed into it. As conventional generators retire and inverter-based solar, wind, and storage make up more of the grid, that capability becomes the thing that keeps the network stable.

The patent record describes this not in marketing terms but as concrete control methods, and reading a granted claim is the most precise way to see what the technology actually consists of. Consider the granted U.S. patent US11680558B2, "Grid-forming control of inverter-based resource using virtual impedance," issued June 20, 2023 to General Electric Company. Its claim 1 sets out the method in full.

"A method for providing grid-forming control of an inverter-based resource connected to an electrical grid, the method comprising: providing the inverter-based resource, the inverter-based resource comprising: a generator or absorber of electrical power, a grid-forming terminal, and grid-forming converter controls; providing, via a processor, a virtual impedance value, XP, at a node between the generator or absorber of electrical power and the grid-forming terminal and a virtual impedance value, XG, at a node between the grid-forming terminal and a remote virtual voltage; determining a voltage drop across virtual impedance values, XP and XG, using at least one current feedback signal…"— Grid-forming control of inverter-based resource using virtual impedance, claim 1, source

Read what the claim is and is not. It claims a method of control, anchored on a specific device — an "inverter-based resource" that includes a generator or absorber of power, a grid-forming terminal, and grid-forming converter controls. The substance of the method is the virtual impedance: the processor is given two impedance values, one between the power source and the grid-forming terminal and one between that terminal and a remote virtual voltage, and it computes a voltage drop across them from current feedback. In plain terms, the inverter is made to present the network with a chosen, software-defined impedance and voltage angle rather than its raw hardware impedance — which is how it mimics the stabilizing behavior of a synchronous machine. That is the engineering content of "grid-forming" in this record: not a new box, but a control law that makes a converter act like a voltage source the grid can lean on.

Where the technology sits in the classification

The claim also makes clear how broad the disclosed application is, and a claims analyst has to note exactly where the limits sit. A dependent claim in the same patent recites that "the inverter-based resource comprises at least one of a wind turbine power system, a solar inverter, an energy storage system, a STATCOM, or a hydro-power system" — the disclosure is written to cover grid-forming control across the whole family of inverter-based resources, not solar alone. That breadth is a feature of the drafting; what protection ultimately attaches is fixed by the claim language and its construction, not by the title. The classification reflects the dual nature of the technology: grid-forming patents typically carry symbols in CPC H02J (the grid function — feeding and stabilizing the network) and in H02M (the converter doing the conversion). A related published application in the same field, US20250330022A1 from the Korea Institute of Energy Research, classifies in H02J 3/1835 and H02J 3/38 and describes "a power stage to convert electric power according to on/off controls of switching devices" together with "a control circuit to calculate a frequency command value" — the same pattern of a converter plus a control law that sets the grid's frequency.

The phrase "inverter-based resource" that recurs in these claims is itself worth pausing on, because it is the term that makes grid-forming a grid problem rather than a device problem. An inverter-based resource is any generator or storage device that reaches the grid through a power-electronic converter rather than through a directly-coupled spinning machine — a solar array, a battery, a modern wind turbine. The defining property of such resources is that they have no inherent inertia: a spinning generator resists a sudden change in grid frequency because its rotating mass physically opposes the change, but an inverter has no rotating mass and will do only what its control software tells it to. Grid-forming control is the software answer to that missing inertia. The virtual-impedance and virtual-voltage constructs in the GE claim are ways of programming an inverter to respond to grid disturbances the way a synchronous machine would, supplying the stabilizing reaction that used to come for free from spinning steel. That is why the patents read as control methods: in an inverter-based grid, stability is something you compute, not something the hardware provides on its own.

Reading the IP without overreading the technology

Two cautions keep the read honest. First, a granted claim describes what is fenced off, not what a company has installed on a live grid. The GE record establishes that a virtual-impedance method of grid-forming control was claimed and allowed; it does not, by itself, tell you which products implement it or how the field's many competing grid-forming approaches compare. Grid-forming is an active area with multiple distinct control philosophies — virtual impedance, virtual synchronous machine, droop-based, and decoupled-voltage designs among them — and a single patent covers one path through that space. Second, the technology's readiness is a deployment question the claims do not answer. A claim can recite an energy-storage or solar application without establishing that grid-forming storage is operating at scale anywhere. For a reader, the value of the patent record is that it pins down, in exact and citable language, what grid-forming control is at the level of the converter and its software — a method of making an inverter set the grid's voltage and frequency instead of chasing them — and leaves the deployment story to the filings and projects that report it.