Maximum power point tracking is the quiet workhorse of every solar installation. The panel's most-efficient operating voltage drifts all day with sun and temperature, and the MPPT algorithm constantly nudges the inverter to sit on that sweet spot. U.S. Patent No. 12,658,713, issued today to Sungrow Renewables Development Co., Ltd., proposes a twist on that workhorse: instead of optimizing only against what the panel can deliver, track the maximum power point while accounting for the losses that happen inside the inverter itself.
According to the patent's abstract, two loss terms drive the control. The first is the direct-current loss — "power loss occurring at a direct-current side of the photovoltaic inverter." The second is the conversion-efficiency loss — "power loss in the inverter circuit" as it turns DC from the panels into grid-synchronous AC. The claimed method runs MPPT "based on the direct-current loss and/or the conversion efficiency loss," and the stated result is more useful electricity at the end of the chain.
"Therefore, the efficiency of the photovoltaic inverter can be improved based on the direct-current loss and/or the conversion efficiency loss, thereby increasing energy yield of the photovoltaic system."— U.S. Patent No. 12658713, source
The reason this is more than housekeeping is that the maximum power point of the panel and the most-efficient operating point of the inverter are not the same place. A classic MPPT routine finds the voltage where the array produces the most power and parks there. But if the inverter happens to be inefficient at that operating point — wasting energy as DC-side loss or in the conversion stage — then the array's peak does not translate into the grid's gain. Sungrow's method, as described, folds the inverter's own loss profile into the optimization so the system is steered toward the point that maximizes delivered energy, not just generated energy. That is a subtle reframing of what "maximum power" should mean for a grid-connected system.
Why it lands in a grid-control class, not a solar-cell class
The classification is telling. The patent's primary CPC code is H02J 3/46, which covers controlling the sharing and supply of power to an AC network from multiple sources — squarely a grid-integration class — supported by H02J 3/0075 and H02J 2101/25. It is not classified in the H01L or H10F photovoltaic-cell families. That placement matches the invention: the contribution is not a better solar cell, it is a smarter control law for how a PV inverter behaves as a generating node on the network. For a company whose product is power-conversion equipment, that is exactly where the IP should sit.
It also connects to the deployment story the rest of this network covers. The binding constraint on solar is increasingly the interconnection point and what happens at the inverter, not the silicon. Every fraction of a percent of conversion efficiency that an inverter recovers is energy that reaches the grid without a single new panel, new permit, or new interconnection slot. A control method that squeezes the inverter's internal losses is, in effect, capacity recovered from hardware already installed — which is the cheapest megawatt-hour there is.
What the claim protects, and what it does not
The grant lists Yanhu Zhang, Pengpeng Chen, Jiankai Chen, and Shaokun Zou as inventors, and the abstract describes the method, a device, and a storage medium — the familiar method-plus-apparatus packaging that lets a control technique be asserted both as software and as the equipment running it. The "and/or" in "based on the direct-current loss and/or the conversion efficiency loss" is worth flagging: it means the method can read on a controller that uses either loss term alone or both together, which broadens the fence somewhat. Still, the claim turns on actually computing and acting on those internal loss quantities. A conventional MPPT controller that ignores inverter losses and tracks only array power would fall outside it.
As always, a granted method is not a shipping firmware build. The patent establishes that Sungrow has fenced off loss-aware MPPT control; it does not, by itself, tell us which of the company's inverters run this logic today or how much yield it recovers in the field. Those are deployment and disclosure questions, not claim questions. What the grant does is stake a position on a real and underappreciated inefficiency — the gap between the panel's peak and the inverter's best operating point — and give Sungrow something to point at if a competitor's inverter starts optimizing the same way. For a grid increasingly bottlenecked at the interconnection rather than the panel, that is IP aimed at the right place.