Grid synchronization control method, device and equipment of photovoltaic inverter and medium
By predicting the phase deviation of photovoltaic inverters through recurrent neural networks and combining it with the weighted fusion of PI controllers, the synchronization problem of traditional phase-locked loops in complex power grid environments is solved, achieving high-precision and stable grid-connected synchronization of photovoltaic inverters.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHANGSHA UNIVERSITY OF SCIENCE AND TECHNOLOGY
- Filing Date
- 2026-05-27
- Publication Date
- 2026-06-23
AI Technical Summary
In the existing grid-connected synchronous control of photovoltaic inverters, traditional phase-locked loops (PLLs) have a contradiction between response speed and stability in terms of dynamic performance. In particular, they are difficult to achieve real-time synchronous tracking of the grid phase in complex grid environments, leading to safety hazards and power quality problems.
A recurrent neural network is used for phase deviation prediction. By constructing a short-time phase deviation time series and a lightweight recurrent neural network model, and combining it with the angular frequency output by the PI controller for weighted fusion, dynamic prediction and compensation of phase changes within future control cycles can be achieved, thereby improving synchronization accuracy and stability.
It significantly improves the dynamic adaptability of photovoltaic inverters under complex grid conditions, shortens the synchronization establishment time, avoids the risk of phase overshoot or loss of lock-in, and enhances the system's response sensitivity and steady-state control capability.
Smart Images

Figure CN122267879A_ABST