Adaptive VSG-Based Coordinated Frequency Support Strategy for Grid-Following/Grid-Forming Hybrid Wind Farms

The modern power grid is a delicate balance of supply and demand, a system that has relied for over a century on massive, spinning turbines in coal and gas plants. These heavy machines act as a natural shock absorber; their sheer weight and momentum keep the system's frequency steady even when the load changes suddenly. As the world shifts toward cleaner energy, wind farms are replacing these traditional plants. However, wind turbines do not spin with the same heavy inertia. They connect to the grid through electronic switches, which means they lack that natural physical buffer. When a large amount of wind power is added without enough of this "rotational weight," the grid becomes fragile. If a sudden disturbance occurs, such as a large factory shutting down or a generator failing, the system frequency can drop too quickly, potentially triggering blackouts.

To solve this, engineers have developed two different ways to make wind turbines behave more like traditional generators. One method, called grid-following, has the turbine listen to the grid and adjust its output accordingly, much like a passenger following a driver. The other, grid-forming, allows the turbine to act as the driver, setting its own voltage and frequency to support the grid. While grid-forming turbines are powerful, they can be unstable in certain conditions, and grid-following turbines are less effective at providing immediate support. The challenge lies in making these two different types of turbines work together seamlessly, especially when the wind speed changes and the grid faces a sudden shock.

A team of researchers at Shanghai Maritime University and the Shanghai Power Equipment Research Institute has proposed a new strategy to coordinate these hybrid wind farms. They recognized that the standard way of controlling these turbines uses fixed settings that do not change, regardless of how severe a disturbance is or how much support the other turbines are providing. This rigidity can lead to a situation where the grid-following turbines wait too long to help, or help too much when it is not needed, wasting energy and wearing out equipment. The researchers developed a system where the turbines can adapt their behavior in real-time, adjusting their internal "stiffness" and "damping" based on exactly what the grid needs at that moment.

The core of their solution involves a control method known as a virtual synchronous generator. This software-based approach tricks the wind turbine into acting as if it has the heavy, spinning mass of a traditional generator, allowing it to resist changes in frequency. In the researchers' new design, the strength of this virtual mass and the amount of damping it provides are not fixed numbers. Instead, they change dynamically depending on how fast the frequency is dropping and how far it has fallen. If the frequency starts to plummet rapidly, the system instantly increases its virtual inertia to slow the drop. As the system begins to recover, the settings shift to help the frequency return to normal without overshooting or oscillating wildly. This adaptive approach ensures that the grid-forming turbines provide the right amount of support at the exact right time.

The researchers also addressed a specific coordination problem between the two types of turbines. Grid-following turbines usually have a "deadband," a small range of frequency fluctuation where they do nothing to avoid reacting to every tiny wobble. In a hybrid farm, if the grid-forming turbines are already providing strong support, the grid-following turbines should not jump in unnecessarily. However, if the grid-forming turbines are struggling, the grid-following ones need to step in immediately. The team created a dynamic deadband that changes its width based on the real-time performance of the grid-forming units. If the grid-forming turbines are doing a good job, the deadband widens, telling the grid-following turbines to stay quiet and save their energy reserves. If the grid-forming support weakens, the deadband narrows, signaling the grid-following turbines to activate their support immediately. This creates a fluid, cooperative relationship where the two types of turbines act as a single, intelligent unit rather than two separate groups.

To test this idea, the researchers built a detailed computer simulation of a power system containing two traditional generators and a hybrid wind farm with forty grid-following turbines and twenty grid-forming turbines. They subjected this virtual system to sudden load disturbances under three different wind speeds: low, medium, and high. In every scenario, they compared their new adaptive strategy against two older methods: one where the turbines used fixed settings, and another where only the grid-forming turbines adapted while the grid-following ones remained static.

The results showed a clear advantage for the adaptive approach. When a large disturbance occurred, the system using the new strategy maintained a higher minimum frequency, meaning the grid stayed much closer to its safe operating limit. In the low wind speed scenario, the frequency dropped to 49.8641 Hz with the old fixed method, but the new strategy kept it at 49.8772 Hz. In the high wind speed scenario, the improvement was even more pronounced, with the frequency staying at 49.7326 Hz compared to 49.7117 Hz with the fixed method. These higher minimum frequencies are critical because they reduce the risk of the system triggering emergency shutdowns.

Beyond just keeping the frequency higher, the new strategy also improved how the turbines shared the workload. The simulations showed that the grid-forming turbines released more of their stored kinetic energy when needed, and the grid-following turbines activated their support more effectively. Crucially, the dynamic deadband prevented the grid-following turbines from activating too often during minor fluctuations. The researchers found that this reduced the number of times the turbines had to switch their power output, which is vital for reducing wear and tear on the equipment. By coordinating the two types of turbines so that they react only when necessary and in proportion to the actual need, the system achieved a better balance between keeping the grid stable and preserving the lifespan of the wind farm.

The study concludes that this adaptive coordination strategy offers a practical way to integrate high levels of wind power without sacrificing grid stability. By allowing the virtual inertia and the activation thresholds to change in real-time, the hybrid wind farm can respond to disturbances with the same flexibility and resilience as a traditional power plant. The simulations suggest that this method not only prevents frequency from dropping too low but also ensures that the different components of the wind farm work together efficiently, providing a robust solution for the future of renewable energy grids.