A 100% PV powered microgrid operates independently and stably

The proposed adaptive GFM control enables the 100% PV park to form a reliable isolated microgrid with resilience to the large disturbances. The GFM PVs can rapidly maintain the synchronised state after the large load disturbances, unlike the conventional generations that are subject to the swing dynamics. In Linyi, Shandong, China, researchers used 100% grid-forming (GFM) photovoltaics (PVs) to power the microgrid without the conventional synchronous machine. The microgrid operates independently and stably supporting the seamless transition between islanded and grid-connected modes. Compared to the grid-following (GFL) method, GFM enables renewable sources to behave like voltage sources with the transient frequency and voltage support. GFM is the key technology for the renewable generations to replace the conventional generations to maintain the synchronous stability of the system. However, most GFM methods are developed for the battery energy storage system (BESS) and focus on the single-machine level to emulate the conventional synchronous machine, such as virtual synchronous machine (VSM)1 and droop control.2 These methods tend to lose the synchronous stability in response to the large disturbances under the multi-machine conditions. Researchers developed the adaptive GFM method with global synchronisation stability for renewable generations, which is suitable for the stable interconnection of multiple GFM sources. The applied method provides the fast active response within the millisecond time scale, ensuring the resilience against the large disturbances, including the DC voltage fluctuations, the grid faults and the load steps etc. The adaptive GFM method is also compatible with a wide range of short-circuit ratio (SCR) and supports the upper-level maximum power point tracking (MPPT) function to completely replace the GFL operation. In grid-connected mode, the GFM PVs synchronise with the utility grid and provide the transient frequency and voltage support to improve the local power quality. In islanded mode, the GFM PVs synchronise with each other to provide the reliable power supply to the local load. Instead of the traditional leader-follower structure, the GFM PVs use fully decentralised control to realise the autonomous power sharing to maintain the transient stability. In addition, the system-level coordination is utilised to optimise the steady-state points of GFM PVs for the optimal operation of the local system. The developed method is implemented in PV generation and BESS. The potential application field of the adaptive GFM method includes but is not limited to the distribution network with very high penetration of power electronic devices and the large-scale renewable stations, with variant devices such as wind generation, Static Var Generator (SVG) and voltage source converter based high voltage direct current (VSC-HVDC) system. The GFM controlled renewable generations are activated as true sources instead of negative loads to support the system stability and safety. The field test is carried out in July-August, 2024 at the industrial park in Linyi, Shandong, China. The park is integrated with seven distributed PVs for a total of 0.7 MW. To create a representative load model, a wide range of load variations were incorporated into the testing scenario. A boiler circulating pump system with one 5 kW motor pump and one adjustable boiler with 100 kW maximum heating power is equipped to emulate a load model including both the static impedance and the dynamic motor. Three 100 kW controllable loads (CL) of power electronic type and 10 kW residential building loads (RBL) are also included. The system is connected to the utility grid with two 1000 kVA 0.4/10 kV step-up transformers. Figure 1 shows the state of the system prior to the disconnection from the utility grid. The low voltage side between T1 and T2 is disconnected and the circuit breaker B1 is used to trigger the islanding and the reconnection events. The secondary control to coordinate the setpoints of each unit is implemented via the communication link to a central cluster controller. The objective of the secondary control is to regulate the voltage and frequency at the point of common coupling (PCC). Figure 2 shows the experimental setup with the cluster controller. Schematic diagram of the industrial park with distributed PVs integration and multiple load types (power electronic, static and dynamic). Experimental setup of the industrial park with cluster controller. Figure 3 shows the recorded system dynamics during the islanding operation with the secondary control enabled. As shown in Figure 2, the PV park is disconnected from the utility grid at 16:10:55 to form the islanded microgrid. The disconnection immediately changes the SCR and a voltage step is observed at the PCC. The 20% load step disturbances are introduced from 16:11:45 to 16:12:45. The maximum voltage deviation is 2% under 20% load disturbances, determined by instantaneous response of GFM PVs. The secondary control adjusts the set-points of GFM-PVs to regulate the steady-state voltage and frequency errors to improve the power quality. Figure 4 shows the recorded islanded microgrid dynamics with the secondary control disabled. It can be seen that the steady-state errors remain after the transient stability is achieved by GFM control under large load disturbances. Autonomous power sharing is observed when single GFM PV is under power dispatching. The reconnection is executed at 16:27:10, which causes a large phase jump to the PCC voltage and the microgrid seamlessly resynchronises with the utility grid. Recorded frequency and voltage at PCC and active power of each PV in the islanded operation of the park with secondary control. Recorded frequency and voltage at PCC and active power of each PV in the islanded operation of the park without secondary control. In summary, the adaptive GFM control enables the 100% PV park to form a reliable microgrid with resilience to the large disturbances including the islanding and the reconnection. The GFM PVs rapidly maintain the synchronised state after the large disturbances, unlike the conventional generations that are subject to the swing dynamics and the risk of the transient instability, which is inherited by conventional emulation-based GFM methods. The superior synchronisation stability lays a solid foundation for the upper-level secondary controller to gradually optimise the microgrid operation. Wenchuan Wu: conceptualization; funding acquisition; writing—review and editing; supervision; project administration; investigation. Zizhen Guo: methodology; software; investigation; writing–original draft. Haotian Liu: methodology; software; writing–original draft. Guannan Wu: project administration; validation. Siyun Li: methodology; software; investigation. Yuwei Cheng: methodology; software; investigation. Natural Key Research and Development Plan of China, Grant Number: 2022YFB2402900. The authors declare no potential conflict of interests. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.