Individual operation detection device, individual operation detection method, individual operation detection program, and system interconnection system comprising individual operation detection device

Abstract

Provided is an individual operation detection device which, in a power distribution system that comprises one or more system formation-type power conversion devices and that is connected to a main system via a switch capable of connection / separation, can accurately and quickly detect an independent operation on a cell grid side by means of a simple configuration. An individual operation detection device according to one embodiment of the present invention is an individual operation detection device of a cell grid system that comprises one or more system formation-type power conversion devices (grid forming (GFM) inverters) and that is connected to a main system via a connectable / separable switch. The individual operation detection device is characterized by comprising an individual operation detector that monitors an operation state of the cell grid to which the system formation-type power conversion device(s) belongs, and that detects that the cell grid is operating individually, the individual operation detector being an active detector and / or a passive detector.

Description

Islanding operation detection device, islanding operation detection method, islanding operation detection program, and grid-connected system equipped with islanding operation detection device

[0001] The present invention relates to an islanding operation detection device, an islanding operation detection method, an islanding operation detection program, and a grid-connected system equipped with the islanding operation detection device.

[0002] In recent years, there has been an increasing demand for local production and consumption of electricity generated by power generation facilities using renewable energy within the region. To address this, a system known as a mini-grid or micro-grid, which divides and manages a power system including one or more power generation facilities, is known. The AC system within this mini-grid is synchronized by an inverter installed on the output side of the power generation facility.

[0003] Because mini-grids can supply power from power generation facilities such as solar cells, in the event of a power outage in the main grid, the mini-grid can be disconnected from the main grid and operated independently, maintaining the power supply and demand within the mini-grid. However, if a fault occurs in the main grid and the mini-grid continues to be connected to the main grid and continues to control the grid connection, the reverse power flow from the mini-grid will become excessive, making it impossible to maintain the power supply and demand. For this reason, mini-grids must be equipped with an islanding operation detection function.

[0004] Patent Document 1 discloses a method for detecting grid islanding of a grid-supported voltage source converter, in which a modulation term (e.g., amplitude modulation) is added to the voltage reference continuously or for a certain period to modulate the voltage reference, thereby modulating the output voltage, thereby detecting grid islanding.

[0005] Cited Document 2 discloses an islanding operation detection system in which an active signal injection unit is provided in a high-voltage side power supply system, and this active signal injection unit detects islanding operation by injecting an active signal of an interharmonic or integer harmonic, and exemplifies that an injection current of an injection order that is a non-integer multiple (e.g., 2.25th to 2.75th) or an integer multiple (e.g., 2nd) of the fundamental wave of the power distribution system is injected as an active signal into the power distribution line.

[0006] Cited Document 3 discloses an active islanding prevention device that detects islanding by monitoring the voltage (V) and frequency (F) of the power grid in a protective relay, as well as the power traveling in either direction between the power grid and a microgrid.

[0007] Cited Document 4 discloses that islanding detection methods are generally classified into passive methods and active methods, and further cites examples of passive methods such as voltage phase jump detection, third harmonic voltage distortion sudden increase detection, and frequency change rate detection, while citing examples of active methods such as frequency shift method, slip mode frequency shift method, active power variation method, reactive power variation method, load variation method, and frequency feedback method with step injection that increases the detection speed of the islanding detection function.

[0008] Cited Document 5 discloses a method for detecting islanding operation in which the frequency of the inverter output voltage is periodically changed, the frequency of the AC current circuit is detected on the primary side of the storage device, and if there is a frequency fluctuation as a result, islanding operation is detected based on that.

[0009] JP-T-2023-517420 A JP-A-2023-119348 A JP-A-2022-136989 A JP-A-2022-047353 A JP-A-2014-122269 A

[0010] In the prior art such as Patent Documents 1 to 5, the grid-forming power conversion device (grid-forming (GFM) inverter) stabilizes its own voltage and frequency, making it difficult to detect an islanded state in which it is isolated from the power grid. This makes the islanding detection method complicated, and it is susceptible to noise, resulting in problems with detection accuracy and detection delay. Specifically, Patent Documents 1 to 5 have the following problems:

[0011] The method of detecting grid islanding of a grid support voltage source converter disclosed in Patent Document 1 utilizes an uninterruptible power supply (UPS) system based on impedance isolated single conversion (ZISC) in grid-connected mode, and detects grid islanding by continuously adding an amplitude modulation term to the voltage reference for a certain period of time to modulate the output voltage. However, the technology disclosed in Patent Document 1 detects only unplanned upstream islanding operation that is a high impedance islanding operation, which limits the systems to which it can be applied, and the need to apply a large modulation to the voltage reference is undesirable for main grid interconnected operation, which is problematic.

[0012] In the islanding detection system of Patent Document 2, a harmonic current is injected into the distribution line as an active signal, but since the harmonic current is injected into the main system during interconnected operation with the main system, there is a limit to the magnitude of the harmonic current.However, there is a problem that an active signal at an allowable level for the main system may not be able to detect islanding due to its relationship with noise.

[0013] In the active islanding prevention device of Patent Document 3, the protective relay needs to monitor not only the voltage (V) and frequency (F) of the power system, but also the power flowing in either direction between the power system and the microgrid, which poses a problem of complexity in the device for detecting islanding.

[0014] The above-mentioned Patent Document 4 discloses that islanding detection methods are generally classified into passive and active methods, but the devices for detecting islanding using conventional methods are complex and have problems with detection accuracy and detection speed due to their relationship with noise. Patent Document 4 also cites a frequency feedback method with step injection as an example of a conventional method, in which reactive power is steeply injected to further accelerate the frequency change of the power system based on the frequency change rate of the power system, and the frequency change is determined to detect islanding. However, this method requires power information from the main system, and the problem of the complexity of the devices for detecting islanding using conventional methods, including this method, remains unresolved.

[0015] The islanding operation detection method of Patent Document 5 mentioned above periodically changes the frequency of the inverter output voltage, detects the frequency of the AC circuit on the primary side of the power storage device, and if there is a frequency fluctuation as a result, detects islanding operation based on that. However, it is difficult to monitor the periodic change in frequency due to the influence of noise, etc., and there are problems with detection accuracy and detection delay.

[0016] Therefore, the object of the present invention is to provide an islanding detection device that can accurately and quickly detect islanding on the cell grid side with a simple configuration in a distribution system (hereinafter referred to as a ``cell grid'') that has one or more grid-forming power converters (grid-forming (GFM) inverters) and is connected to a main system via a connectable or disconnectable switch.

[0017] The above object of the present invention can be achieved by the following configuration: That is, an islanding operation detection device according to a first aspect of this embodiment is an islanding operation detection device for a cell grid system that includes one or more grid-forming power converters (grid-forming (GFM) inverters) and is connected to a main system via a connectable or disconnectable switch, and includes an islanding operation detector that monitors the operating state of the cell grid to which the grid-forming power converter belongs and detects that the cell grid is in islanding operation, and the islanding operation detector is at least one of an active detector and a passive detector.

[0018] The second aspect of the present embodiment is characterized in that, in the islanding operation detection device of the first aspect, the passive detector is a detector that detects islanding operation by detecting a change in the voltage phase of the grid-forming power conversion device.

[0019] The islanding operation detection device of the third aspect of this embodiment is characterized in that, in the islanding operation detection device of the first aspect, the active detector comprises at least one of the following detectors: (1) a detector that detects islanding operation by actively amplifying and detecting instability in the voltage phase of the grid-forming power conversion device; (2) a detector that detects islanding operation by adding fluctuations to the active power reference of the grid-forming power conversion device; and (3) a detector that detects islanding operation by adding fluctuations to the reactive power reference of the grid-forming power conversion device.

[0020] The fourth aspect of the islanding detection device of this embodiment is characterized in that, in the islanding detection device of any of the first to third aspects, the detection of islanding is performed using at least one of the following methods: mode-dependent frequency destabilization / stabilization method, phase shift detection method, impedance detection method, and direct fault detection method.

[0021] The fifth aspect of the present embodiment of the islanding operation detection device is characterized in that, in the islanding operation detection device of any of the first to third aspects, the grid interconnection controller and the grid-forming power conversion device further have a function of correcting the internal clock based on time synchronization information from a timestamp server that acquires time synchronization information based on the grid phase.

[0022] The sixth aspect of the present embodiment of the islanding operation detection device is characterized in that, in the islanding operation detection device of any of the first to third aspects, the grid-forming power converter performs phase synchronization control for grid connection of the cell grid based on voltage phase information of the main grid.

[0023] The seventh aspect of the present embodiment of the islanding operation detection method is an islanding operation detection device for a cell grid system that has one or more system-forming power conversion devices and is connected to a main system via a connectable or disconnectable switch, and is characterized in that it has an islanding operation detector that monitors the operating state of the cell grid to which the system-forming power conversion device belongs and detects that the cell grid is in islanding operation, and the islanding operation detector is at least one of an active detector and a passive detector.

[0024] An islanding operation detection program according to an eighth aspect of the present embodiment is characterized in that the islanding operation detection method according to the seventh aspect is executed by a computer device.

[0025] The grid interconnection system of the ninth aspect of this embodiment is characterized by comprising at least an islanding detection device of any one of the first to third aspects, and a grid interconnection controller that detects at least voltage phase information of the main grid.

[0026] A grid interconnection system according to a tenth aspect of the present embodiment is characterized by including at least the islanding operation detection device according to the sixth aspect, and a grid interconnection controller that detects at least voltage phase information of the main grid.

[0027] According to the first aspect of the islanding detection device of the present invention, in a distribution system (hereinafter referred to as a "cell grid") that has one or more grid-forming power converters (grid-forming (GFM) inverters) and is connected to a main system via a connectable or disconnectable switch, it is possible to accurately and quickly detect islanding operation of a cell grid that is operated synchronously by the grid-forming power converter on the cell grid side using a simple configuration.

[0028] For example, according to the islanding detection devices of the second and third aspects of the present invention, the cell grid system can detect the islanding operation of the cell grid when the switch is closed, so that it is possible to satisfy the provisions of the grid interconnection regulations (grid code) that standardize passive and active detection methods. In this case, if the islanding detection device can quickly detect an abnormality in the main system and the breaker with synchronism verification function can be disconnected before the main system voltage drops, there is a possibility that the cell grid will not experience a blackout and will be able to continue operating.

[0029] 4A is a block diagram of a power distribution system including a grid-forming power conversion device according to embodiment 1 of the present invention. FIG. 4B is a circuit diagram of one grid-forming power conversion device. FIG. 4C is a detailed circuit diagram of FIG. 2A. FIG. 4B is a circuit diagram of a power distribution system including a grid-forming power conversion device according to embodiment 1 of the present invention. FIG. 4C is an explanatory diagram of the rotational phase angle. FIG. 4C is an explanatory diagram of the rotational phase angle after adjustment. FIG. 4A is a block diagram of a voltage phase synchronization method according to embodiment 1 of the present invention. FIG. 4A is a block diagram of a cell grid islanding operation according to embodiment 1 of the present invention. FIG. 4B is a block diagram of a passive islanding detection device according to embodiment 1 of the present invention. FIG. 4A is a block diagram of an LPF used in FIG. 4A. FIG. 4B is a block diagram of active islanding detection according to embodiment 2 of the present invention. FIG. 5 is a block diagram of a grid-connected operation state according to embodiment 5 of the present invention. FIG. 5 is a control block diagram of another example of active islanding detection according to embodiment 2 of the present invention. FIG. 5C is a logical block diagram of active islanding perturbation according to embodiment 3 of the present invention. FIG. 5D is a block diagram of DGR control frequency dynamics according to embodiment 3 of the present invention. FIG. 5E is a block diagram of a grid-connected controller protection function for signal-based islanding detection according to embodiment 3 of the present invention. FIG. 5F is a block diagram of a grid-forming power conversion device protection function for signal-based islanding detection according to embodiment 3 of the present invention. Fig. 10 is a block diagram of a static phase angle shift protection adapted to signal-based islanding detection according to embodiment 3 of the present invention; Fig. 11 is a block diagram of a signal-based islanding detection according to embodiment 3 of the present invention; Fig. 12 is a block diagram of an islanding detection perturbation signal generation according to embodiment 3 of the present invention; Fig. 13 is a block diagram of a DGR voltage reference calculation including islanding detection signal perturbation according to embodiment 3 of the present invention; Fig. 14 is an explanatory diagram of a time stamp server of a distribution system equipped with a grid-forming power conversion device according to embodiment 5 of the present invention.

[0030] Hereinafter, an islanding operation detection device, an islanding operation detection method, an islanding operation detection program, and a grid-connected system equipped with the islanding operation detection device according to embodiments of the present invention will be described with reference to the drawings. However, the embodiments shown below are examples of an islanding operation detection device, an islanding operation detection method, an islanding operation detection program, and a grid-connected system equipped with the islanding operation detection device to embody the technical concept of the present invention, and the present invention is not limited to these, and can be equally applied to other embodiments included in the scope of the claims.

[0031] [Embodiment 1] Passive islanding detection in a grid-connected system according to embodiment 1 of the present invention will be described with reference to Figures 1 to 4. First, as an example of a grid-connected system to which this embodiment is applied, a block diagram of the grid-connected system in Figure 1 will be described.

[0032] [Overall Configuration of Grid-Connected System] Fig. 1 is a block diagram of a power distribution system including a system-forming power conversion device according to the first embodiment. Electric power generated in a power plant 10 is boosted to an extra-high voltage by an extra-high voltage substation 11, and then, via a power transmission facility 12, the power is further stepped down at a substation 13, and the power is supplied to a power distribution facility 15 (hereinafter referred to as the "main system"), which is a main system. A plurality of cell grids 20 are connected to the main system 15. The voltage distributed to the main system 15 is not limited to a high voltage (more than 600 V AC and not more than 7,000 V AC), but may be, for example, an extra-high voltage (more than 7,000 V AC) or, conversely, a low voltage (not more than 600 V AC).

[0033] In a power distribution system (hereinafter referred to as a "cell grid"), at least one location of a main power distribution system (hereinafter referred to as the "main system") is isolated by a circuit breaker 21 with a synchronism test function that can be isolated. The cell grid 20 includes the circuit breaker 21 with a synchronism test function, a cell grid system 28 connected to the main system 15 via the circuit breaker 21 with a synchronism test function, one or more system-forming power conversion devices 40 (hereinafter referred to as "Digital Grid Routers" or "DGRs") having power conversion functions connected to the cell grid system 28, and various power facilities connected to the cell grid system 28. The cell grid 20 may be, for example, the size of an apartment complex, a single city block, or the size of an industrial complex or complex. It refers to a power network that can locally produce and consume power from distributed power sources such as renewable energy power generation facilities within the area without relying on power from large-scale nuclear power plants or thermal power plants outside the area. In Japan, this term is sometimes referred to as a "microgrid," but in this specification, it is sometimes referred to as a "mini-grid" in English. These are used as synonyms for the cell grid 20. The grid interconnection controller 30 may be abbreviated to "Mini Grid Controller."

[0034] The cell grid system 28 distributes electricity to each household facility 22a serving as power demand facility 22. The power demand facility 22 is not particularly limited, but may include each household facility 22a, a power storage device 22b, an electric vehicle charging / discharging device 22c, and an electric vehicle 22d. Note that the power storage device 22b, the electric vehicle charging / discharging device 22c, and the electric vehicle 22d are capable of charging and discharging, and therefore function as power demand facility 22 when charging, and also function as power generation facility 23 when discharging. Although shown as power demand facility 22 in FIG. 1, the present embodiment is not limited to this, and they may be connected to a grid-forming power conversion device 40 as power generation facility 23.

[0035] Each power generation facility 23, such as the wind power generation facility 23a, the solar power generation facility 23b, and the internal combustion engine power generation facility 23d, and the power storage facility 23c are connected to the cell grid system 28 via a DGR 40. As described below, each DGR 40 performs synchronization control of the cell grid system 28 so that the frequency and phase of the cell grid system 28 are synchronized with the main system 15, based on a main system frequency measurement value f and a phase synchronization signal Ncyc_ref_Sync, etc., sent from a grid interconnection controller 30, for example, provided on the cloud. In addition, by adjusting the inter-system phase difference φglobal between the main system 15 and the cell grid system 28, it is possible to control the flow or reverse flow of active and reactive power. That is, when the cell grid system 28 receives power from the main system 15, i.e., when the cell grid system receives a power flow, the power flow from the main system to the cell grid system can be increased by reducing the inter-system phase difference φglobal. Conversely, when surplus power generated by a power generation facility 23 such as a solar cell power generation device in the cell grid 20 is transmitted to the main system 15, i.e., when reverse power flow is performed, the amount of power transmitted to the main system 15 can be increased by increasing the phase difference φglobal between the systems and advancing the phase angle of the cell grid system 28.

[0036] The circuit breaker with synchronism detection function (DG breaker) 21 can connect the cell grid system 28 to the main system 15 when the main system 15 and the cell grid system 28 are synchronized, and cannot connect the cell grid system 28 to the main system 15 when the main system 15 and the cell grid system 28 are not synchronized. In this way, the circuit breaker with synchronism detection function 21 is equipped with a PLL and can also detect the phase difference between the main system 15 and the cell grid system 28, and can close the circuit breaker with synchronism detection function only when both systems are synchronized. In other words, synchronism detection can also be performed by a PLL.

[0037] The circuit breaker with synchronism detection function 21 can detect the main system frequency f and the main system rotation phase angle θref. The circuit breaker with synchronism detection function 21 is also equipped with a PLL and can also detect the phase difference between the main system 15 and the cell grid system 28. The main system frequency f, the main system rotation phase angle θref, and the phase difference between the main system 15 and the cell grid system 28 detected by the circuit breaker with synchronism detection function 21 are sent to the DGR 40 and used for phase synchronization control. In this way, the circuit breaker with synchronism detection function (21) can acquire system voltage information and cell grid voltage information and perform calculations, thereby also being able to share part of the functions of the grid interconnection controller (30).

[0038] Here, an example has been described in which the main system frequency f, the main system rotation phase angle θref, and the PLL are provided in the circuit breaker with synchronism testing function 21, but the present embodiment is not limited to this, and detectors for the main system frequency f and the main system rotation phase angle θref may be provided separately from the circuit breaker with synchronism testing function 21. In addition, the PLL may also be configured to be provided separately from the circuit breaker with synchronism testing function 21.

[0039] [Equivalent Circuit of DGR 40 in Cell Grid] Figures 2A to 2E are diagrams illustrating the principle of the grid-connected system according to the first embodiment of the present invention. Figure 2A is a circuit diagram of one grid-forming power conversion device 40, showing an equivalent circuit when one DGR 40 is extracted from the cell grid system 28. As shown in Figure 2A, all DGRs 40 operate as AC voltage sources Vdgr_i having output impedances. i is a subscript representing the number of DGRs 40. Each DGR 40 is connected in parallel to the cell grid voltage Vgrid via a grid-connection resistance RG and an inductance LG, and a current Idgr_i flows out of each DGR 40. Because the cell grid voltage Vgrid is common to all DGRs 40, each DGR 40 can control Idgr_i by changing its internal voltage Vdgr_i. Under normal conditions, Idgr_i is controlled to be equal for all i, thereby equalizing the load sharing of each DGR 40. However, depending on the state of the DGR 40, the load sharing may be changed, or Idgr_i may be controlled to flow into the DGR 40 in order to charge the internal storage battery. These can be controlled depending on how the target value of Vdg_i is set.

[0040] FIG. 2B is a detailed circuit diagram of FIG. 2A, showing the half-bridge inverter, a component of the DGR 40. FIG. 2C is an explanatory diagram of the operation of FIG. 2B, showing how IL repeatedly increases and decreases as the upper and lower switches of the half-bridge are alternately turned on. Upper and lower limit bandwidths are set for the target current value Iref. When the measured current value IL exceeds the upper limit, the upper switch of the half-bridge is turned off and the lower switch is turned on. When the measured current value IL falls below the lower limit, the lower switch is turned off and the upper switch is turned on. This operation is repeatedly performed. This is called current hysteresis control.

[0041] A common issue with hysteresis control is the fluctuation of the switching period, which makes filter design difficult. To address this issue, we devised a variable bandwidth method. In this method, the lines above and below the target current value Iref represent the band; narrowing this band shortens the switching period, while widening it shortens it. Because the slope of the half-bridge reactor current IL varies depending on the difference between Vdc and Vdgr, we adjust the bandwidth accordingly to maintain a nearly constant switching frequency. This method controls the target current value Iref to be equal to Idgr and corrects the half-bridge filter voltage to be equal to the target Vdgr, achieving open-loop control. While the controlled object is current, this method ultimately enables high-speed voltage control. The following explanation assumes that Vdgr is directly controlled using this method.

[0042] FIG. 2C is a circuit diagram of a grid-connected system according to a first embodiment of the present invention, showing N DGRs 40, as shown in FIG. 2A, connected in parallel to a cell grid. The cell grid voltage Vgrid is common, and each DGR 40 is connected to the cell grid via a different output impedance ZG_i (i = 1 to N) at a different voltage Vdgr_i, drawing a current Idgr_i from the common voltage Vgrid. In this manner, synchronization has been considered important when operating multiple voltage sources in parallel because cross currents will flow between the voltage sources if the voltages are not synchronized accurately. This invention employs a standard time signal as the synchronization signal. In this embodiment, the virtual impedance technique described below is further used to virtually equalize ZG_i, thereby equalizing Vdgr_i and Idgr_i, thereby equalizing the load sharing. Furthermore, by controlling Vdgr, Idgr can also be varied, thereby changing the load sharing for each DGR 40.

[0043] Figure 2D illustrates the relationship between the current Idgr of one DGR 40, the internal voltage phase Vdgr of the DGR 40, and the cell grid voltage phase Vgrid under synchronous control. The notation "*" on the right shoulder, such as Vgrid*, indicates the target value. The figure shows a state in which the target and actual values ​​match. In the following explanation, actual values ​​will be used. Vgrid is based on the GPS time signal, so it is synchronized across all DGRs 40. Vgrid rotates at a frequency based on the global angle reference θref calculated from the GPS time base. In Figure 2E, this is represented as a fixed D axis (D axis). Idgr is a current delayed by a phase angle φ based on the load and grid impedance within the cell grid. Idgr can be decomposed into DQ axes, ID and IQ, with ID being in phase with Vgrid and IQ lagging 90°.

[0044] Voltage drops occur on the DQ axes due to virtual impedances, including Idgr and the physical impedances mentioned above. The voltage drop of the D-axis component is ID×RG + IQ×XG, and the voltage drop of the Q-axis component is ID×XG - IQ×RG. Note that in Figures 2E and 2F, the multiplication symbol "×" is represented by a period. These voltage drops of the D-axis and Q-axis components correspond to the voltage drop between Vdgr and Vgrid, so Vdgr is required to compensate for this voltage drop. The phase angle between the resulting internal voltage Vdgr and the cell grid system voltage Vgrid is δ. Because Vgrid is synchronized across all DGRs 40, Vdgr also rotates synchronously, albeit with a phase difference. Therefore, by adjusting the phase angle δ, the output share of each DGR 40 can be changed or it can function as a negative output, i.e., a load, to charge the built-in storage battery.

[0045] When the cell grid is connected to the main system, a target Vgrid* vector is created based on the frequency of the main system, and the DGR 40 internal voltage target Vdgr* and current target Idgr* are as shown in Figure 2E according to the virtual impedance. The actual cell grid voltage Vgrid, DGR 40 internal voltage Vdgr, and current Idgr are controlled to converge to their respective target values ​​Vgrid*, Vdgr*, and Idgr*, and are approximately equal. Any power surplus or deficit resulting from an imbalance between the total output of all DGRs 40 and the total demand within the cell grid is supplied from the main system or flows back to the main system.

[0046] However, when the cell grid is isolated from the main grid and operates in islanded mode, the total output of all DGRs 40 must match the total demand within the cell grid, since there is no power regulation by the main grid. Because the DGRs 40 do not have demand information, they control the actual Vdgr to match the target DGR 40 internal voltage Vdgr*. As a result, the actual current Idgr flows according to the actual demand, as shown in Figure 2E. This actual Vgrid and actual Idgr supply the active and reactive power of the demand in just the right amount. As a result, a deviation occurs between the actual current Idgr and the target current Idgr*. The actual voltage Vgrid also deviates from the target Vgrid*. The phase difference between this actual cell grid voltage Vgrid and the target cell grid voltage Vgird* is defined as Δδ. Since the Daxis* axis is synchronized with the main grid frequency, transitioning the cell grid from islanded mode to grid-connected mode requires synchronizing with the main grid by reducing Δδ to zero. This is achieved using a signal called φglobal.

[0047] Furthermore, accurate standard time must be used to synchronize all DGRs 40 during black start. Synchronization errors create voltage differences between the DGRs 40. Impedance between the DGRs 40 causes a 10% to 20% voltage drop. Even if the allowable voltage error is 1%, the current control error will be approximately ±10% to ±5%. Therefore, voltage errors should be minimized. A 1% voltage error corresponds to a phase angle of 0.0015 radians. This corresponds to a time accuracy of 5 μsec. Although some crystal oscillators have an accuracy of approximately ±10 μsec, errors between the DGRs 40 will increase if common correction is not applied. This embodiment employs voltage phase synchronization and time synchronization during black start. However, to ensure time accuracy, a GPS time signal, for example, is also used.

[0048] [Grid-Connected System Using a Grid-Forming Inverter with a Voltage Phase Synchronization Scheme] A grid-connected system using a grid-forming inverter with a voltage phase synchronization scheme according to this embodiment will be described. The voltage-based synchronization method is a standard method for grid-connected microgrids. This method has been widely implemented, tested, and deployed around the world. When connected to the main grid, fluctuations on the cell grid side are absorbed by the main grid and do not have any adverse effects.

[0049] First, the voltage-based synchronization control method has the following advantages: - The synchronization signal uses the voltage of the cell grid's distribution lines. - No special additional equipment such as a GPS signal or receiver is required. - The voltage synchronization signal is very robust. - Operation can continue as long as the power lines are not faulted or damaged. - The voltage synchronization signal automatically adjusts for the following issues: - Voltage changes and phase shifts in transformers. - Voltage drops and phase shifts due to circuit impedance. - The DGR 40 can be attached and detached without any special configuration when connected to the cell grid. - The DGR 40 is flexible for manual or automatic changes to the cell grid circuit configuration.

[0050] On the other hand, in the voltage phase synchronization method, it is necessary to note that synchronization cannot be achieved during stand-alone operation of the cell grid separated from the main grid, and that fluctuations must be prevented from becoming too large with general frequency droop control. In this embodiment, an example is described in which a synchronization signal from the grid interconnection controller 30 is used. For black start during a cell grid power outage, a synchronization signal based on, for example, GPS time may be used.

[0051] The reason why synchronization cannot be achieved during isolated operation of a cell grid separated from the main grid is explained below. When the cell grid is operating in an isolated mode, the total output of all DGRs 40 must match the total demand within the cell grid, so an actual current Idgr flows according to the actual demand, which causes a deviation from the target Idgr*. As a result, a phase difference of Δδ occurs between the actual cell grid voltage Vgrid and the target Vgrid*. If Δδ is zero, the frequency and phase are stable, but once the balance between the total output and total demand is lost and Δδ becomes positive or negative, |Δδ| increases.

[0052] For example, if the DGR40 active power output exceeds the actual load active power, the actual voltage Vgrid will lead the target voltage Vgrid* in phase, and Δδ will always be positive. The target voltage Vgrid* will try to match the actual voltage Vgrid through the PLL, so positive feedback is applied, increasing the phase of θref. As a result, the PLL frequency will continuously increase.

[0053] Conversely, if the DGR40 active power output is smaller than the actual load active power, the actual voltage Vgrid lags behind the target voltage Vgrid* in phase, and Δδ is always negative. The PLL continuously reduces the phase of θref in a positive feedback loop. As a result, the PLL frequency continuously decreases. In either case, the phase difference between the PLL and the mini-grid is not stable, and the frequency is also not stable.

[0054] Thus, voltage-based synchronization cannot be used for grid-independent cell grids because there is no robust power source to accommodate or provide the actual reactive power difference between the DGR 40 and the load. Conversely, if there is a robust power source, the actual voltage Vgrid is stable, and the target voltage Vgrid* can be controlled to approach it without falling into a positive feedback loop.

[0055] [Introduction of Inter-grid Phase Difference Signal φglobal] To avoid this positive feedback loop even in a cell grid operating in an isolated state independent of the main grid, we devised the following method. Specifically, we propose a method for stabilizing the cell grid frequency and phase by transmitting the main grid frequency signal f and the inter-grid phase difference signal φglobal to all DGRs 40 via the grid interconnection controller 30. First, as shown in Figure 3B, the phase angle signal θref generates φsync through feedback control Kθctrl 85, which reduces the error between the reference angular frequency ωref and the mini-grid angular frequency ωgrid. The phase angle signal θpll obtained by the PLL of each DGR 40 is added to this signal in adder 86 to generate the local grid phase angle signal Ncyc_ref 89. However, this control alone would cause positive feedback in the cell grid voltage phase, as described above, and would not stabilize the cell grid voltage phase.

[0056] Therefore, as shown in Figure 3B, φglobal 80 is further added in an adder 86. φglobal is a signal that represents the phase difference between the main system and the cell grid. Although the main system and the cell grid are separated, φglobal changes when the cell grid phase opens compared to the main system phase. Adding φglobal suppresses the positive feedback loop that causes the phase to open, making it possible to suppress instability in the cell grid phase and frequency. This stabilizes the cell grid frequency by using only information from the main system, even if it is not directly connected to the main system. This allows all DGRs 40 in the cell grid to be synchronized, stabilizing the frequency even during isolated operation without being connected to the main system. Furthermore, since both systems are synchronized, a circuit breaker with a synchronism test function can be closed at any time to connect the two systems.

[0057] 3A shows that φglobal is effective even when the main system and cell grid are connected. Since active power flows when the phase difference between the two systems increases, φglobal measures the power flow 17 flowing between the two systems, compares it with the target active power and reactive power, creates the necessary phase difference signal φglobal, and transmits it to all DGRs 40. This allows the output of all DGRs 40 to be increased or decreased, and the power flow between the two systems to be controlled to match the target value.

[0058] In addition, when the main system is in a power outage and the cell grid is operated independently, a reference voltage phase is created using an internal clock instead of the voltage phase of the main system, and the phase difference signal with respect to the cell grid voltage phase is transmitted as φglobal to all DGRs 40 via the system interconnection controller 30, thereby synchronizing the frequency and phase of the cell grid.

[0059] Next, the reason why a standard time signal is required when the grid interconnection controller 30 is stopped and when a black start occurs during a cell grid power outage will be explained. As described above, when the main system and the cell grid are connected, the voltage phase of the main system serves as the reference. When they are separated, voltage phase information of the main system can be obtained through the grid interconnection controller 30. Even when the main system is in a power outage, the phase based on the clock inside the grid interconnection controller 30 can be transmitted to all DGRs 40 as a common reference through the grid interconnection controller 30. However, when the grid interconnection controller 30 is stopped and the system is separated from the main system, there is no common voltage phase information for all DGRs 40. Similarly, when the cell grid is stopped, there is no common voltage phase information for all DGRs 40. In such cases, a common synchronization signal must be supplied to all DGRs 40 by another means. If the DGRs 40 acquire standard time information, they can correct their internal clocks based on the standard time information, thereby creating a voltage phase synchronized with all DGRs 40. In this way, a method is proposed for stabilizing the frequency and phase of the cell grid using standard time information during a black start after a cell grid power outage or during cell grid independent operation when the grid interconnection controller 30 is stopped.

[0060] [Detailed Description of Voltage-Based Synchronization Method] The voltage-based synchronization control method will be described with reference to Figures 3A and 3B. Figure 3A is a control block diagram of the main grid connected state of voltage synchronization according to the first embodiment of the present invention, and Figure 3B is a control block diagram of the cell grid grid independent state (synchronization control) of voltage synchronization according to the first embodiment of the present invention. The block diagrams and their control methods in each state will be described below.

[0061] [Synchronization method when connected to the main grid] In the voltage-based synchronization method, when the cell grid is connected to the main grid, the difference between the signal based on the main grid frequency and the signal based on the frequency detected by the PLL of each DGR 40 is controlled to be small, and the phase angle signal is also detected and taken into account, so that the target voltage Vgrid* and the actual voltage Vgrid approximately match. This state is shown in Figure 2E and is the same as the state when time is synchronized with GPS.

[0062] [Explanation Using a Block Diagram When Connected to the Main System] Referring to FIG. 3A , a block diagram for calculating Ncyc_ref 89 in the grid-forming power conversion device (DGR) 40A when the cell grid system is connected to the main system will be described. FIG. 3A is a block diagram of voltage-based synchronization control according to the first embodiment of the present invention, and is a control diagram for the grid-connected state. The grid-connected controller (grid-connected controller) 30A and the grid-forming power conversion device (DGR) 40A are included. In the grid-connected controller (grid-connected controller) 30A, an inter-system phase difference signal φglobal calculation unit 80 generates an inter-system phase angle φglobal from the main system active power Pgrid_main and the main system reactive power Qgrid_main measured by the grid active power Pgrid_main measurement unit 98 and the main system reactive power Qgrid_main measurement unit 99, taking into account the power flow between the main system 15 and the cell grid system 28.

[0063] First, we will explain the main system frequency measuring instrument 31. The main system voltage acquisition unit 97 acquires the main system voltage Vgrid_main. In a state connected to the main system, the main system voltage and the cell grid voltage are the same, so Vgrid_main = Vgrid_mini.

[0064] The PLL 96 calculates a main system frequency measurement value f31 from the main system voltage Vgrid_main and transmits the signal to the DGR 40 via the grid interconnection controller 30. The inter-system phase difference signal φglobal 80 calculation unit calculates an inter-system phase difference signal φglobal based on the difference between the measured main system active power Pgrid_main and main system reactive power Qgrid_main flowing between the main system and the cell grid and their respective target values ​​(not shown), and transmits the signal to the DGR 40 via the grid interconnection controller 30. If the direction of power flow toward the cell grid is considered positive, increasing the inter-system phase difference signal φglobal causes the phase angle of the cell grid system to lag relative to the main system, thereby increasing the power flow from the main system to the cell grid system. On the other hand, decreasing the inter-system phase difference signal φglobal causes the phase angle of the cell grid system to advance, thereby increasing the reverse power flow from the cell grid system to the main system. In this way, the output of the DGR 40 is changed to match the power flow between the two systems to the target value. Note that for reactive power, a voltage correction signal is sent to the DGR 40 via the grid interconnection controller 30, rather than φglobal 80, but the method is similar and detailed description is omitted.

[0065] The PLL 83 of the grid-forming power conversion device (DGR) 40A generates a cell-grid grid phase angle signal θpll and a cell-grid grid angular velocity signal ωpll from the cell grid voltage Vgrid_mini. The main grid frequency measurement value acquisition unit 41 receives the main grid frequency measurement value f measured by the main grid frequency measuring device 31 and calculates the main grid angular velocity signal ωref = 2π f. The PLL 83 calculates the cell-grid grid angular velocity signal ωpll from the cell grid voltage Vgrid_mini. Next, a comparator calculates the inter-grid angular velocity error signal ωerr = ωref - ωpll. Next, the controller kθ_ctrl 85 receives ωerr as an input, performs PI calculation, and calculates the control phase signal φsync. kθ_ctrl has proportional gain and integral gain, but the gain is constant for low-frequency components and small for high-frequency components, making it less susceptible to sudden frequency changes and the like.

[0066] An adder 86 adds the cell-grid system phase angle signal θpll and the inter-system phase angle φglobal to the control phase signal φsync, and the resulting signal is divided by 2π using a divider. Furthermore, the remainder of the division using a MOD function is calculated to extract only the decimal point, resulting in a sawtooth output Ncyc_ref 89. Furthermore, by multiplying Ncyc_ref by 2π, a cell-grid system rotation phase angle signal (hereinafter sometimes referred to as the "cell-grid system phase angle signal") θref = 2π · Ncyc_ref is obtained. While the MOD function is used in this embodiment, it can be replaced with a function that extracts the decimal point as described above.

[0067] In this way, the cell grid system phase angle signal θref = 2π·Ncyc_ref obtained by each grid-forming power conversion device (DGR) 40A is synchronized with the main system voltage (= cell grid voltage). In the case of a stable system in a steady state, the PLL frequency is constant, so ωref = ωpll, and ωerr = 0, and therefore φsync = 0. For this reason, θpll created by the PLL of each DGR 40 becomes dominant. When the inter-system phase angle φglobal is added to this, it is possible to control the power flow and reverse power flow between the cell grid system and the main system.

[0068] If the system frequency or phase changes suddenly, ωref ≠ ωpll, but a transmission time delay occurs. However, the PLL 83 monitors the cell grid voltage (= main system voltage) and creates θpll, so it follows sudden changes immediately (within about one cycle). The power flow between the main system and the cell grid changes suddenly, but a time delay occurs in the transmission of φglobal. Limits are set on the rise and fall of φglobal to prevent sudden changes. In this way, the power flow connecting the two systems and the PLL maintain the synchronization force and inertia force of the cell grid.

[0069] [Cell Grid Independence] Referring to FIG. 3B , a block diagram for calculating Ncyc_ref in the grid-forming power conversion device (DGR) 40B in the cell grid isolation state will be described. FIG. 3B is a block diagram of voltage-based synchronization control according to the first embodiment of the present invention, and is a control diagram for a cell grid isolation operation state in which the cell grid is isolated from the grid. The system includes a grid interconnection controller (grid interconnection controller) 30B and a grid-forming power conversion device (DGR) 40B. The same components as those in FIGS. 1 to 3A are assigned the same reference numerals, and their description will be omitted. The grid phase difference signal φglobal calculation unit 80 calculates the measured main grid voltage Vgrid_main and cell grid voltage Vgrid_mini using a PLL 96b, and calculates the difference in phase angle between them as the grid phase difference signal φglobal.

[0070] The input contents of φglobal in Figure 3B are different from those of φglobal in Figure 3A. In Figure 3B, both systems are isolated, so the purpose is to align the voltage phase of the DGR 40B with the main system so that the circuit breaker with synchronism detection function 21 can be closed at any time. Therefore, in Figure 3B, φglobal calculates the voltage phase difference between the two systems. When both the main system and the cell grid system are stable in a steady state during independent operation, sending a signal of φglobal = 0 matches the voltage phase of the two systems, making them ready for connection at any time. If the system frequency or phase suddenly changes, ωref ≠ ωpll. However, because both systems are isolated, the phase angle difference between the main system and the cell grid temporarily widens, but after a transmission time delay has elapsed, the voltage phases match again, and the two systems are ready for connection at any time.

[0071] A method for realizing cell-grid isolated operation by having all DGRs 40C continue synchronous operation when the main system is stopped will be described. When the main system is stopped, the circuit breaker with synchronism detection is turned off, and information from the system interconnection controller 30C is lost. In this case, the main system frequency measurement value f41 is maintained at its previous value or replaced with a rated value, and the inter-system phase difference signal φglobal is maintained at its previous value or replaced with a fixed value (including zero) prepared in advance, thereby making it possible to continue cell-grid isolated operation while all DGRs 40C maintain synchronization.

[0072] In the block diagrams of Figures 3A and 3B, an example of synchronously controlling each DGR 40 using a voltage synchronization method has been described. However, this embodiment is not limited to this and may also employ other synchronization control methods, such as a synchronization control method based on a GPS time signal. The GPS time synchronization method is characterized in that each DGR 40 receives two signals, a phase synchronization signal Ncyc_ref_Sync obtained by applying a main grid frequency signal f and a GPS time signal, via the grid interconnection controller 30. By combining these signals with the GPS time signal within the DGR 40, each DGR 40 can generate a phase angle signal θref synchronized to the main grid within ±5 μsec. For example, when the voltage phase synchronization method also uses a GPS time signal, if a reference voltage vector based on GPS time is stored within each DGR 40C, the DGR 40C can be operated using that reference voltage vector as Vgrid* in Figure 2E during black start or when the grid interconnection controller 30C is stopped.

[0073] [Passive cell grid islanding detection] Passive islanding detection in a grid-connected system according to this embodiment will be described. In this embodiment, a passive islanding detection method will be described as a method for effectively detecting islanding in a cell grid using a grid-forming inverter of the voltage phase synchronization method. This method is a method that can perform islanding detection of a cell grid without actively varying the power reference, as in active islanding detection described below.

[0074] Passive islanding detection measures and responds to circuit voltage or current characteristics that change significantly when an island is formed. Passive islanding detection is effective when distributed generation is grid-following, but when islanding load and generation are well matched, a no-detection zone (NDZ) usually exists. In a grid-connected cell-grid system using voltage-phase-locked grid-forming inverters, passive islanding detection is not effective because the mini-grid voltage and frequency are stable. Common passive islanding detection methods are shown below.

[0075]

[0076] [A: Voltage Phase Change Detection Method] In this embodiment, a phase shift method is exemplified as a detection method superior to the conventional passive cell-grid-only detection method. Phase unbalance protection includes both static and dynamic phase shift detection. Phase shift detection has an instantaneous trip characteristic with a single trip and reset level. This is not robust against grid phase angle jumps during phase-to-phase voltage drop events unless the trip level is very high. For this reason, time-delay static phase shift detection is added. This uses a low trip threshold with a trip time delay. The instantaneous and time-delay protection functions keep the phase angle distribution within a range acceptable for a typical polyphase load.

[0077] [Regarding Measurement of Differential Phase Angle Frequency] In this embodiment, the power control of the cell grid is of a grid-based type, and standard passive and active islanding detection methods and compatible passive and active islanding detection methods are described below.

[0078] The cell grid uses grid-forming power converters (DGR (Dynamic Grid-forming) inverters) for voltage synchronization, frequency stabilization, power control, and load sharing. The control is the same in both grid-connected and islanded states. This is a highly desirable feature, as it eliminates the need to adjust the DGR control mode to match the state of the synchronism-detecting circuit breaker. Mini-grid power control includes voltage-based synchronization and global time-based synchronization. Voltage-based synchronization is a grid tracking method that provides stable frequency and angle references in grid-connected states. Voltage-based synchronization in grid-independent states requires frequency stabilization control.

[0079] Grid-forming power converters adjust their voltage output angle or power reference in proportion to the measured mini-grid frequency error. This process is essentially the opposite of the frequency destabilization used in active islanding detection methods. Frequency stabilization control is essential for stable operation under grid-isolated conditions. Frequency stabilization also has the effect of frequency and angle damping and dynamic grid-integrity improvement.

[0080] Frequency stabilization control is desirable but not required for grid-tied mini-grids. Therefore, frequency stabilization control can be disabled in grid-tied states and replaced with a frequency destabilization active islanding detection method. This change does not affect power control and load sharing of the grid-tied mini-grid.

[0081] Frequency destabilization active islanding detection can be applied to mini-grids, provided that the DGR selects the frequency stabilization / destabilization mode according to the open / closed state of the synchronism detector circuit breaker. The DGR must periodically request the status of the synchronism detector circuit breaker from the grid connection controller server (MGC server). If the upstream synchronism detector circuit breaker is open, the state of the DGR's synchronism detector circuit breaker is open. There is a slight communication delay between the grid connection controller updating the synchronism detector circuit breaker status on the grid connection controller server and the DGR receiving the updated status from the grid connection controller server. A delay of 1 to 2 seconds is achievable on most public data networks. This delay results in a discrepancy between the true synchronism detector circuit breaker status and the synchronism detector circuit breaker status used in the DGR control.

[0082] The effects of these delays are outlined below. When the DGR synchronized circuit breaker status is open, the grid is in an isolated state with DGR frequency stabilization enabled and islanding protection disabled. This applies when intentional islanding is performed under normal steady state conditions.

[0083] When the DGR's synchronism detector circuit breaker status is closed, the synchronism detector circuit breaker is open, or the synchronism detector circuit breaker is closed and the main grid isolation device is open, a grid-isolated state with frequency destabilization islanding protection is in effect. The transient state that occurs during the transition from grid-tied to grid-independent state begins when the synchronism detector circuit breaker or main grid isolation device opens and ends when the DGR receives updated synchronism detector circuit breaker status and changes mode from frequency destabilization to frequency stabilization. The mini-grid operates as an unintentionally created islanded system with frequency instability. The frequency increases or decreases at a controlled rate until the grid-tied controller or DGR detects an abnormal condition. Detection methods include under-frequency, over-frequency, excessive ROCOF, and phase slip, and this state should be avoided if possible.

[0084] When transitioning from a grid-tied state to a planned isolated system, this state can be avoided by changing the status of the synchronism detector in the grid-tied controller server to "open" a certain time before the synchronism detector actually opens. The mini-grid's grid-forming power converter updates the status of the synchronism detector to "open" and changes the mode from frequency destabilization to frequency stabilization. Therefore, the frequency remains stable even after the synchronism detector opens.

[0085] The delay in opening the synchronous breaker must not exceed the allowable islanding detection time. This condition cannot be avoided if the main grid synchronous breaker or upstream isolation device opens as an automatic response to feeder circuit fault protection. A seamless transition from grid-tied to islanded operation requires the grid-tied controller to detect islanding, open the synchronous breaker, update the synchronous breaker status, and the mini-grid's grid-forming power converter to read the updated synchronous breaker status and apply frequency stabilization. To improve the probability of a seamless transition, the grid-tied controller and the DGR can coordinate their islanding detection, with the grid-tied controller providing primary protection and the DGR providing backup protection.

[0086] A certain trip latency is applied between the DGR's backup protection islanding trip and the time it de-energizes its mini-grid interface. If the DGR acquires a synchronous detector circuit breaker open status during the latency period, the DGR's protection function is reset and the mini-grid interface is not de-energized. This adjustment ensures that there is sufficient time for the grid connection controller to detect the fault, open the synchronous detector circuit breaker, and for the mini-grid DGR to read the updated synchronous detector circuit breaker status and apply frequency stabilization before the mini-grid is de-energized.

[0087] When the DGR synchronized detection function breaker status is closed and the synchronized detection function breaker is closed, the grid connection state is one in which frequency instability islanding protection is effective.Normal steady-state grid connection operation with active islanding detection.Frequency-based passive islanding detection methods are also effective.

[0088] When the DGR synchronism detector circuit breaker status is open and the synchronism detector circuit breaker is closed, the system is in a grid-connected state with DGR frequency stabilization enabled and islanding protection disabled. A transient state that occurs when transitioning from a grid-independent state to a grid-connected state: occurs after the synchronism detector circuit breaker is closed but before the DGR obtains the updated synchronism detector circuit breaker status. It also occurs when the synchronism detector circuit breaker status is updated to "open" before the grid connection controller performs delayed opening of the synchronism detector circuit breaker when a transition from a grid-connected state to a grid-independent state is scheduled.

[0089] This state is grid-tied operation without active islanding detection. Its duration should be limited to prevent sustained unintended islanding. Typically, the maximum duration should not exceed the allowable islanding detection time. In a mini-grid with multiple DGRs, the DGRs do not simultaneously obtain their synchronism-detecting circuit breaker status from the grid-tied controller. Therefore, each DGR transitions between frequency stabilization and frequency destabilization at slightly different times. The overall effect is a gradual change in frequency feedback gain between negative feedback (frequency stabilization) and positive feedback (frequency destabilization). The duration of this gradual mode change depends on the communication update rate from the grid-tied controller cloud server to the DGRs and the latency of each DGR.

[0090] In the voltage phase change detection method, a method is also available in which the differential phase angle frequency is measured and, when this differential phase angle frequency exceeds a threshold, passive islanding detection is triggered. In the event of islanding, the mini-grid frequency immediately becomes unstable, causing a sudden increase or decrease in frequency. This can trigger under-frequency protection, over-frequency protection, ROCOF protection, and sinusoidal phase shift protection. The protection functions are configured so that the positive phase shift protection trips first. This method can also be used with the active islanding detection method described below, and is discussed in more detail below. Active islanding detection is performed by the positive phase shift protection.

[0091] Figure 4A is a block diagram of a phase-shift passive islanding protection device, and Figure 4B is a block diagram of the LPF. When the circuit breaker with a synchronism detector (_closed) is detected, the detection on / off switcher turns on the detection and sends phase error signals Phi_err_i and AVdPhi_idt to the static phase shift detector and dynamic phase shift detector, respectively. On the other hand, when the circuit breaker with a synchronism detector (_closed) is not detected, i.e., when the circuit breaker with a synchronism detector (_synchronism detector) is open, the detection on / off switcher outputs a "zero" signal to the downstream decision block instead of the phase error signal to maintain islanding and prevent a trip signal from being issued. When either the static phase shift detector or the dynamic phase shift detector issues a trip signal, the output is the Phaxe_balance_trip signal. This detects that the cell grid is in an islanded state due to a grid fault or other reasons when the circuit breaker with a synchronism detector (_closed) is detected, i.e., when the grid is in grid-tied operation mode.

[0092] The block in FIG. 4A employs both static phase shift detection based on the phase difference signal Phi_err_i and dynamic phase shift detection based on the phase change difference signal AVdPhi_idt. This allows for high accuracy, error-free detection, and early detection of cell grid islanding. Furthermore, the phase shift detection of this embodiment is used for normal synchronous operation of the cell grid, thereby minimizing unnecessary perturbations to the main grid even during grid-connected operation. Furthermore, during islanding operation of the cell grid, the error in the phase signal is integrated before detection, allowing for rapid detection of an islanding abnormality. [Specific Configuration Example of Phase Shift Detection] A characteristic of local islanding operation, such as that of a cell grid, is that it is out of phase with the majority of the main grid system. This raises concerns about the rated voltage of open-type isolation devices and the risk of out-of-phase reclosing. The phase detection method of this embodiment provides a passive islanding detection method that directly detects out-of-phase conditions by measuring the phase shift of local islanding operation.

[0093] Islanding can be isolated with a synchronism-verifying circuit breaker rated to isolate two unsynchronized power systems and protected against out-of-phase reclosing. Knowledge of the main grid's positive-sequence voltage angle can be used to synchronize the mini-grid with the main grid in a grid-independent manner. Direct measurement of the local islanding phase shift requires a voltage sensor in the non-isolated part of the main grid system. Direct measurement of the local islanding phase shift is difficult because the local islanding may be isolated by upstream isolators that are physically separate from the grid-tied controller and all DGR sites.

[0094] Therefore, a grid-connection controller time server can be used as a method to measure the local islanding phase shift. The grid-connection controller time server provides highly accurate "phase reference" timestamps of main grid voltage zero-crossing events to grid-connection controllers installed within the main grid system. The grid-connection controllers providing the phase reference timestamps typically have a local clock calibrated to a global atomic clock reference using GNSS time signals. These grid-connection controllers may also have other functions, including control of local synchronous circuit breaker devices and / or associated mini-grids. In the case of grid-connection controllers with local synchronous circuit breaker devices, the voltage zero-crossings are obtained from the main grid side of the synchronous circuit breaker.

[0095] In a local mini-grid, the grid-tied controller and DGR are clients of a grid-tied controller time server connected to a remote point on the main grid. The grid-tied controller time server is a cloud server that provides the grid-tied controller main grid interface with main grid frequency measurements, fgrid_ref_MTS, and a phase reference timestamp of the most recent main grid voltage zero-crossing event. The local grid-tied controller / DGR synchronizes its time with the time server by indirect means using GNSS signals or main grid voltage zero-crossing events. The local grid-tied controller / DGR uses a global time synchronization clock to measure the timestamp of the local main grid voltage zero-crossing. By comparing the local and remote voltage zero-crossing times, the local voltage phase difference relative to the remote time server is calculated.

[0096] A local grid-tied controller / DGR can be a client of multiple remote grid-tied controller time servers. The grid-tied controller / DGR calculates the local voltage phase difference for each remote grid-tied controller time server. In grid-tied operation, the local voltage phase difference for each grid-tied controller time server is stable. When the local mini-grid is known to be in a grid-tied state, a calibrated phase difference is calculated by a low-pass filter (LPF) of the local voltage phase difference. The local voltage phase shift of the grid-tied controller / DGR is equal to the local voltage phase difference minus the calibrated phase difference. In steady-state grid-tied conditions, the local voltage phase difference for each remote grid-tied controller time server is close to zero.

[0097] When a local island forms, the island's phase drifts relative to the bulk main grid. This can be detected by comparing the absolute value of the local phase shift with an upper threshold. The associated reclosing transient is smaller in magnitude than a standard "black start" energization transient, so a 90° phase separation threshold can be applied. For example, an absolute phase shift limit of 60° can trip before the 90° threshold and reject local phase shifts due to the connection or disconnection of large loads.

[0098] If the absolute value of the local voltage phase shift is greater than the limit, it means that either the local mini-grid or the remote grid-tied controller time server is in local island operation. Another possibility is that part of the local or remote distribution network has been reconfigured, which has caused a large phase shift to its new steady-state value. Local island operation should be detected when a significant number of remote grid-tied controller time servers show an absolute local voltage phase shift that exceeds the limit. For example, two out of three remote grid-tied controller time servers show an absolute local voltage phase shift that exceeds the limit.

[0099] The data from the remote grid-connection controller time server is updated at a rate of 0.1 seconds to 1 second or even faster, which introduces a delay in calculating the local voltage phase shift. This communication delay can cause the true phase shift to exceed 90° before the grid-connection controller detects the local islanding and opens the circuit breaker. To solve this problem, the local grid-connection controller / DGR calculates the local voltage phase shift in real time by interpolating the zero-crossing data from the associated remote grid-connection controller time server. Interpolation can be a simple extrapolation process using fgrid_ref_MTS, or it can use a PLL to track the zero-crossing data from the remote time server. Interpolation is effective because the bulk grid frequency barely varies within the interpolation time window.

[0100] Local islanding and phase-shift detection can also require the condition to persist for a time delay. Typical values ​​are 60° with a 0.1-second delay for the grid-tied controller and 80° with a 0.2-second delay for the DGR as backup protection. This allows the grid-tied controller to detect islanding first and open the synchronized breaker, creating an intentional islanding operation operating in grid-independent mode. The DGR backup protection can be used to isolate the DGR site if the synchronized breaker fails to open. The grid-tied controller and DGR islanding phase-shift protection are disabled and reset when the upstream synchronized breaker opens.

[0101] The grid connection controller and DGR perform local islanding phase shift detection as described in the following algorithm: The grid connection controller / DGR detects a local islanding phase shift trip if mstPhaseTrip = True on a sufficient number of grid connection controller time servers. The grid connection controller opens the synchronism check circuit breaker after detecting a local islanding phase shift trip. The DGR opens the local site connection device or de-energizes the output after detecting a local islanding phase shift trip.

[0102] To accurately calculate the phase difference, the time synchronization between the grid-tied controller time server and the local grid-tied controller / DGR requires an accuracy of ±1 ms. This is lower than the time synchronization required for power sharing and control within the mini-grid (e.g., ±50 μs). [Evaluation of the phase shift method] The advantages of the phase shift method are as follows: - Supports both global time and voltage-based synchronization - Does not affect the power and frequency dynamics of the mini-grid or main grid - Provides direct detection and protection of phaseless islanding and reclosing conditions - Ideal for seamless transition between grid-tied and grid-independent operation - Same DGR control in both grid-tied and grid-independent operation - Standard communication from the grid-tied controller to the DGR is possible. In grid-independent state, the DGR backup islanding phase shift protection is disabled.

[0103] However, to utilize the phase-shift method, care must be taken to ensure that the passive islanding detection time does not exceed the 0.5 seconds specified in Japan's interconnection requirements. A time server must be installed on the main grid; for example, this time server may be another grid-connection controller device on another feeder circuit. Because it is not active, it may need to be used in conjunction with other methods to comply with standard RLC load qualification tests. Islanding with precisely matched RLC loads may require a long detection time before it becomes out of phase. Note that standard RLC load tests do not prevent out-of-phase islanding conditions.

[0104] Summary of Phase Shift Detection Evaluation: DGRs (distributed generation resources) are also clients of a timestamp server connected to a remote location on the main grid. The timestamp server sends timestamp information about recent voltage zero-crossing events to the cloud server. Grid-tied controllers (mini-grid controllers) and DGRs measure local voltage zero-crossing timestamps. The local clocks of the timestamp server and the grid-tied controllers / DGRs are synchronized to a global time reference. This time synchronization can be achieved via a GPS receiver or via voltage zero-crossings. During grid-connected operation, the voltage zero-crossing phase difference between each timestamp server and the local grid-tied controller / DGR is constant and stable. Therefore, the grid-tied controller / DGR can calibrate the expected phase shift.

[0105] When an islanding condition occurs, the island's phase shifts relative to the main grid. The grid-connection controller / DGR detects zero-crossing changes relative to the main grid's timestamp server. A phase change relative to one timestamp server indicates that timestamp server is in an islanding condition or connected to a reconfigured distribution circuit. A phase change relative to multiple timestamp servers indicates that the grid-connection controller / DGR is in an islanding condition or connected to a reconfigured distribution circuit. If the phase change exceeds a limit (perhaps 45 to 90 degrees), the grid-connection controller or DGR detects the islanding condition. The grid-connection controller should be configured to detect the islanding condition first and then open its breaker to continue operating in grid-independent mode. The DGR should be configured to detect the islanding condition as a backup and then shut down (the DGR's phase-shift islanding protection may be disabled when the breaker is open).

[0106] If communication from all timestamp servers to the grid-tied controller / DGR fails, a backup method must be used. The grid-tied controller uses phi-global to limit the phase shift from the mini-grid to the timestamp servers. If the phase shift reaches a threshold (e.g., 45 degrees), the grid-tied controller changes from grid power control to mini-grid synchronization control. This detects islanding conditions and prevents dangerous out-of-phase conditions.

[0107] The advantage of the phase-shifted approach is that it does not destabilize the frequency or voltage of the main grid, and it is well suited to seamless transitions.

[0108] However, if time synchronization between DGRs is required, a timestamp server or similar may be required. This may cause slight fluctuations (flicker) in the grid voltage. Details of the timestamp server will be described later.

[0109] [Embodiment 2] [B. Active Islanding Detection] Active islanding detection in a grid-connected system according to embodiment 2 of the present invention will be described with reference to FIGS. 5 to 8. The same reference numerals are used for components common to FIGS. 1 to 4, and their description will be omitted. FIGS. 5 and 6 are control block diagrams of a voltage-synchronized cell-grid system isolation state (synchronization control) according to embodiment 4 of the present invention. In conventional technology, a microgrid is assumed to operate stably both when connected to the grid and when operating independently, which poses a problem of delayed detection of an abnormality in the microgrid on the microgrid side. In this embodiment, active islanding detection allows the cell grid side to detect that the cell-grid system has unintentionally become isolated.

[0110] [Basic Description of Voltage-Based Synchronization Method] In this embodiment, cell grid independent detection is performed in a voltage phase-locked grid-forming inverter. Here, an example of the configuration of a voltage phase-locked grid-forming inverter will be described with reference to FIGS. 5 and 6.

[0111] The voltage-based synchronization control method has the following advantages: - The synchronization signal uses the voltage of the cell grid's distribution lines. - No special additional equipment such as a GPS signal or receiver is required. - The voltage synchronization signal is very robust. - Operation can continue as long as the power lines are not faulted or damaged. - The voltage synchronization signal automatically adjusts for issues such as voltage changes and phase shifts in transformers, and voltage drop and phase shifts due to circuit impedance. - The DGR 40 can be connected to and disconnected from the cell grid without any special configuration. - The DGR 40 is flexible for manual or automatic changes to the cell grid circuit configuration.

[0112] To avoid a positive feedback loop even when the cell grid is in an isolated operating state independent of the main grid, the following method is adopted. Specifically, we propose a method for stabilizing the cell grid frequency and phase by transmitting the main grid frequency signal f and the inter-grid phase difference signal φglobal to all DGRs 40 via the grid interconnection controller 30. First, as shown in Figure 5, the phase angle signal θref generates φsync through feedback control Kθctrl 85, which reduces the error between the reference angular frequency ωref and the mini-grid angular frequency ωgrid. The phase angle signal θpll obtained by the PLL of each DGR 40 is added to this signal in an adder 86 to generate a local grid phase angle signal Ncyc_ref 89.

[0113] Furthermore, φglobal 80 is further added in an adder 86. φglobal is a signal that represents the phase difference between the main system and the cell grid. Although the main system and the cell grid are separated, φglobal changes when the cell grid phase opens compared to the main system phase. Adding φglobal suppresses the positive feedback loop that causes the phase to open, making it possible to suppress instability in the cell grid phase and therefore frequency. This is intended to stabilize the cell grid frequency by using only information from the main system, even if it is not directly connected to the main system. This allows all DGRs 40 in the cell grid to be synchronized, and the frequency to be stabilized, even during isolated operation without being connected to the main system. Furthermore, since both systems are synchronized, a circuit breaker with a synchronism test function can be closed at any time to connect the two systems.

[0114] Even when the main system and the cell grid are connected, φglobal measures the power flow 17 flowing between both systems, compares it with the target active power and reactive power, creates the required phase difference signal φglobal, and transmits it to all DGRs 40. This makes it possible to increase or decrease the output of all DGRs 40 and control the power flow between both systems to match the target value. Furthermore, when the main system is in a power outage and the cell grid is operated independently, a reference voltage phase based on an internal clock is created instead of the voltage phase of the main system, and a phase difference signal with the cell grid voltage phase is transmitted as φglobal to all DGRs 40 via the grid interconnection controller 30, thereby synchronizing the frequency and phase of the cell grid.

[0115] In the examples of the block diagrams of Figures 5 and 6, an example has been described in which each DGR 40 is synchronously controlled by adopting a voltage synchronization method, but this embodiment is not limited to this and also includes other synchronization control methods, such as a synchronization control method based on a GPS time signal, or a combination of these methods.

[0116] [B. Active Method] Here, active islanding detection will be described. However, in addition to the active islanding detection method, the passive islanding detection method described later as embodiment 2 can also be adopted as a method for detecting cell grid islanding operation.

[0117] The predominant active islanding detection method for inverter-based generation is frequency destabilization. This method is preferred because it does not have steady-state perturbation signals that can cause voltage fluctuations or ripple in motor load torque. All frequency destabilization methods are compatible and work together without communication or control. This allows for successful islanding detection even in installations with multiple distributed generators installed in parallel. As distributed generators become more widespread, there is concern that frequency destabilization methods may reduce the frequency stability and transient margin of the main grid. Table 2 lists active cell-grid islanding detection methods.

[0118]

[0119] Frequency destabilization methods are incompatible with the cell grid, which forms the grid, because the mini-grid stabilizes frequency regardless of whether main grid generators are connected to the circuit. Voltage destabilization is also ineffective against the cell grid.

[0120] Active or reactive power perturbations are effective in detecting local islanding using either grid tracking or grid formation control. However, these methods are not favored due to concerns that perturbation signals from parallel distributed generators may interfere with and cancel each other out. As a result, local islanding may not be detected. Also, parallel distributed generators may provide a low-impedance path similar to the main grid. Both of these effects can degrade the measured frequency or voltage response, potentially resulting in islanding not being detected. To be effective, the perturbation signals must be coordinated among all DGRs in the mini-grid so that they function equivalently to a single large generator with an active perturbation. Below, we describe the active cell-grid islanding methods (B1-B3) of this embodiment, which offer superior detection methods not found in the prior art.

[0121] [B1] Positive Feedback-Negative Feedback Switching Method [B1a] Switching Method Based on Circuit Breaker Open / Close Detection Active islanding detection using the positive feedback-negative feedback switching method will be described with reference to Figures 5 and 6. Grid codes are established as grid connection regulations that must be followed by power sources newly connected to the grid. For example, in Europe, grid codes for connecting renewable energy to the grid are specified as the Grid Connection Requirements for Generators (RFG). In Japan, grid connection regulations are stipulated as the open access obligations stipulated in Article 17 of the Electricity Business Act, including the Electricity Transmission and Distribution Business Guidelines, the Guidelines for Grid Interconnection Technical Requirements for Ensuring Power Quality, the Grid Interconnection Regulations, the Grid Interconnection Technical Requirements (Supplement to the Terms and Conditions for Grid Interconnection, etc.), and the Grid Access Rules.

[0122] In the Japanese grid code, an active detection method on the cell grid 20 side is standardized as a method for detecting islanding of the cell grid 20. The active detection method on the cell grid 20 side is not particularly limited, but for example, a slip mode frequency shift method can be adopted. The slip mode frequency shift method is a method that employs a frequency feedback method with step injection, and is a method that performs high-speed islanding detection on the cell grid 20 side by steeply injecting reactive power from the cell grid 20 side to further promote frequency change based on the frequency change rate of the system.

[0123] 5 and 6 each include a selection circuit 160 and the like. The grid interconnection controller 30D includes a circuit breaker switching state detection unit 150. The circuit breaker switching state detection unit 150 detects the switching state of the circuit breaker with synchronism testing function 21 provided in the cell grid 20. The circuit breaker switching state detection unit 150 outputs an ON signal when the switching state of the circuit breaker with synchronism testing function 21 is closed. In addition, the circuit breaker switching state detection unit 150 outputs an OFF signal when the switching state of the circuit breaker with synchronism testing function 21 is open.

[0124] A selection circuit 160 is provided in the DGR 40D. The selection circuit 160 includes an input terminal 161 to which the output ωpll of the PLL 83 is input, two output terminals: a positive feedback terminal 163 and a negative feedback terminal 164, and a selection function unit 162 that selectively transmits the input from the input terminal 161 to either the positive feedback terminal 163 or the negative feedback terminal 164. An ON signal or an OFF signal that is output from the circuit breaker open / close state detection unit 150 is input to the selection circuit 160. The selection function unit 162 selects the positive feedback terminal 163 when an ON signal is input, and selects the negative feedback terminal 164 when an OFF signal is input.

[0125] 5 , when the circuit breaker with synchronism detection function 21 is closed and an ON signal is output from the circuit breaker open / close state detection unit 150, the selection function unit 162 selects the positive feedback terminal 163, and therefore the output signal ωpll of the PLL 83 is added as positive feedback in the adder / subtractor 170. As a result, the adder / subtractor 170 adds the output signal ωref of the multiplier 81 and the output signal ωpll of the PLL 83 to form a signal ωerr, which is input to the controller 85.

[0126] 6 , when the circuit breaker with synchronism detection function 21 is open and an OFF signal is output from the circuit breaker open / close state detection unit 150, the selection function unit 162 selects the positive feedback terminal 163, and therefore the adder / subtractor 170 subtracts the output signal ωpll of the PLL 83 as negative feedback. As a result, the adder / subtractor 170 subtracts the output signal ωpll of the PLL 83 from the output signal ωref of the multiplier 81 to obtain a signal ωerr, which is input to the controller 85.

[0127] The detection state of the DGR 40 of this embodiment is shown in Table 1.

[0128] When the operating state is the "main system connected state," as shown in FIG. 5 , the avoidance state of the circuit breaker with synchronism detection function 21 is "closed," the output signal of the circuit breaker open / close state detection unit 150 is an "ON signal," and the selection state of the selection function unit 162 is the "positive feedback terminal 163" selected. Therefore, the output signal ωpll of the PLL 83 is added as positive feedback. In this state, positive feedback of the frequency deviation is applied, and the DGR 40D functions as an islanding operation detection device using a slip mode frequency shift method. Even when the output signal ωpll of the PLL 83 is added as positive feedback, the power stabilization control of the main system is robust, so the frequency of the cell grid 20 connected to the main system does not fluctuate beyond the reference value. Therefore, during main system connected operation, the DGR detection state is determined to be "normal," i.e., normal main system connected operation, because the frequency fluctuation is within the reference value.

[0129] However, when the operating state is "main grid connected state," if a sub-grid isolated operation state is formed with the power supply and demand of the sub-grid system balanced, and if a fault or abnormality occurs in the cell grid in this state, it is necessary to trip the sub-grid system in the isolated operation state as quickly as possible to cause a power outage. Therefore, when the avoidance state of the circuit breaker with synchronism detection function 21 is "closed," the islanding operation of the sub-grid system can be detected in the DGR 40D by adopting a frequency feedback method with step injection using a slip mode frequency shift method on the sub-grid system side, i.e., in the DGR 40D.

[0130] In other words, when the operating state is "main system connected state," if the cell grid enters independent operation, the output signal ωpll of PLL 83 is added as positive feedback, causing the frequency of cell grid 20 to fluctuate and exceed the reference value. As a result, DGR 40D detects a frequency abnormality or an abnormality in the frequency change rate and enters an "abnormality detected" state, so DGR 40D performs trip control, causing the cell grid 20 to enter a power outage state.

[0131] On the other hand, when the operating state is the "cell-grid isolated operation state," as shown in Figure 6, the avoidance state of the circuit breaker with synchronism test function 21 is "open," the output signal of the circuit breaker open / close state detection unit 150 is an "OFF signal," and the selection state of the selection function unit 162 is a state in which the "negative feedback terminal 164" is selected, so the output signal ωpll of the PLL 83 is subtracted as negative feedback. In this state, negative feedback of the frequency deviation is applied, and therefore the DGR 40E can stabilize and control the frequency of the cell-grid system through synchronization control, so that the state is determined to be "normal" during cell-grid isolated operation, and the isolated operation of the cell-grid system can continue.

[0132] As described above, in the grid interconnection system of this embodiment, the circuit breaker with synchronism testing function 21 detects the open / closed state of the circuit breaker with synchronism testing function 21, and when the circuit breaker with synchronism testing function 21 is in the closed state, the selection circuit 160 is switched to the positive feedback terminal 163 to apply positive feedback of the frequency deviation, so that the DGR 40D has the function as an islanding operation detection device using the slip mode frequency shift method. On the other hand, when the circuit breaker with synchronism testing function 21 is in the open state, the selection circuit 160 is switched to the negative feedback terminal 164 to apply positive feedback of the frequency deviation, so that the DGR 40E can stabilize and control the frequency of the cell grid system through voltage phase synchronization control.

[0133] In the grid interconnection system of this embodiment, by switching between positive and negative feedback of the frequency deviation depending on the open / close state of the circuit breaker with synchronism test function 21, the DGR 40D functions as an islanding detection device using a slip mode frequency shift method when connected to the main grid, making it possible to comply with grid codes that standardize active detection methods. Moreover, in the case of a cell grid islanding operation state, the DGR 40E can continue the islanding operation of the cell grid system by stabilizing control of the frequency, as in embodiment 1. Note that while this embodiment is expressed as a simple selection circuit 160, in embodiments it is configured as a control circuit that combines a detection circuit with time delays, gains, etc.

[0134] [B1b Positive Phase Shift Method] For active cell grid islanding protection, this is called the "positive column number phase shift protection" subsystem to distinguish it from the previous B1a method. The positive phase shift method (B1b) is common to the previous switching method (B1a) based on circuit breaker open / close detection in that it is a detection element of the active islanding protection function.

[0135] The active islanding detection method is triggered by measuring the differential phase angle frequency, as described in the passive islanding detection method. The active islanding detection method measures the differential phase angle frequency. When the dynamic component of the differential phase angle frequency exceeds a threshold, an active islanding test is triggered and runs for a set time, e.g., 1.0 s. During the active islanding test, the DGR mini-grid frequency stabilization is suspended, and positive feedback from the mini-grid positive phase frequency to the DGR output voltage reference angle is applied. When the active islanding test is triggered, the output voltage reference angle is also increased in steps to accelerate islanding detection. In the case of local generation islanding, the mini-grid frequency becomes unstable immediately, causing a sudden increase or decrease in frequency. This may trigger underfrequency protection, overfrequency protection, ROCOF protection, and sinusoidal phase shift protection. The protection settings are configured so that the positive phase shift protection trips first. Therefore, active islanding detection is performed by the positive phase shift protection. Note: The proposed active islanding detection is equivalent to the frequency change triggered step reactive power method used for active islanding detection in grid-following inverters.

[0136] The protection function block diagram of the grid-tied controller and DGR is shown below. As shown in Figure 7, the damping of the slip angle due to Klpf_θ slip stops when the absolute ROCOF exceeds the ROCOF trip threshold. This improves the active islanding detection time and enhances the tolerance to frequency step and ramp events.

[0137] Figure 8 shows the active test evaluation logic, which shows that the DGR protection function includes an active islanding perturbation subsystem. A frequency destabilization positive feedback mechanism is created in the following active islanding perturbation subsystem. The mean phase angle sinusoidal frequency AVdPhi_posSeq / dt is received from the PLL filter subsystem block. The positive feedback gain Kactive_island is selected as Kactive_island_testOn_miniGrid when the active test state is true, and as Kactive_island_testOff_miniGrid when the active test state is false. Typical values ​​are Kactive_island_testOn_miniGrid = 0.1125 and Kactive_island_testOff_miniGrid = 0.

[0138] The base value of the DGR output destabilization angle, Phi_pos_shift_base, is calculated as Phi_pos_shift_base = Kactive_island * AVdPhi_posSeq / dt. The DGR output destabilization angle, Phi_pos_shift, is calculated by a range- and rate-limited LPF acting on Phi_pos_shift_base. The cutoff frequency of the LPF, flpf_Phi_posShift_miniGrid, is typically 1 Hz. The LPF is range-limited to ±ThetaActiveShiftMax_miniGrid and rate-limited to ±ThetaActiveShiftRateMax_miniGrid. Typical values ​​are ThetaActiveShiftMax_miniGrid = π / 4 and ThetaActiveShiftRateMax_miniGrid = 5.π / 4. When the active test state is triggered, Phi_pos_shift is set to PhiBoost_activeTest_miniGrid (typically π / 40) to accelerate active islanding detection. Phi_pos_shift is added to the DGR output voltage angle reference for each phase.

[0139] Here, the average differential phase angle frequency AVdPhi_i / dt is generated by the PLL filter subsystem. This is similar to the above-mentioned A: Shift Phase Method. The protection function subsystem modifies AVdPhi_i / dt to remove the residual common mode component by subtracting the phase average value from each phase value. Therefore, the complete differential value of AVdPhi_i / dt is received by the phase imbalance protection subsystem. Both subsystems apply LPFs to AVdPhi_i / dt.

[0140] If the synchronizer circuit breaker status is reported to the DGR as closed, both subsystems evaluate the absolute value of LPF(AVdPhi_i / dt). If the synchronizer circuit breaker status is reported to the DGR as open, the system is in intentional islanding, and both subsystems evaluate a zero value. This ensures that both passive islanding detection and active islanding testing are disabled during the intentional islanding condition. The calculation of AbsAVdPhi_i / dt = abs(LPF(AVdPhi_i / dt)) [synchronized_circuit_closed = True] or 0 [synchronized_circuit_closed = False] is performed by the phase imbalance protection subsystem and the active test evaluation subsystem. Conceptually, the calculation of AbsAVdPhi_i / dt can be shared between these subsystems. However, it is preferable to use separate logic to completely separate the passive and active islanding detection systems.

[0141] [Increasing Active Decision Speed ​​with DGR Dynamics] Adjusting the DGR dynamics can improve the active decision speed, enabling more stable, continuous operation of the cell grid. Figure 9 shows a block diagram of the DGR control frequency dynamics. Figure 9 incorporates the phase angle difference destabilization gain KphaseBal_shift and associated range and rate limits. This is used as an active open phase destabilization parameter. Phase angle difference destabilization is disabled when the DGR is notified that a breaker with synchronizer detection is open. This stabilizes the phase angle of the DGR output even during intentional islanding.

[0142] The frequency dynamics subsystem in Figure 9 also includes a frequency stabilization enable signal, Fstable_enable. When the active-test state is False, or when the synchronizer-enabled breaker is notified to the DGR as open, frequency stabilization is enabled (Fstable_enable = True).

[0143] In addition, by temporarily suspending frequency stabilization when Fstable_enable = False, active islanding detection can be made to operate properly.

[0144] It is also possible to improve the angle convergence rate when Fstable_enable = False. This is achieved by selecting the angle response LPF gain coefficient when Fstable_enable = False as Kg_Nresp_active_test. Typically, Kg_Nresp_active_test = 4 is set, which quadruples the cutoff frequency of the angle response LPF. This improves the response time of active islanding detection.

[0145] Furthermore, for the DGR output attenuation logic, in power and frequency damping or frequency damping, frequency damping can be suspended when Fstable_enable = False, which improves the response time of active islanding detection.

[0146] The DGR real and reactive power limiting logic can be updated to include rate limits for real and reactive power reference adjustments. Real and reactive power limit adjustments are rate limited by adding limit blocks to the input of the power limit integrator. The rate limits are asymmetric, so limits are applied relatively slowly (reducing full power with a typical duration, TmPlimitAdjRateSet = 10s) and recovered more quickly (restoring full power with a typical duration, TmPlimitAdjRateReset = 1s). This prevents limits from being applied during grid voltage sags and surge transients that require ride-through and fast recovery.

[0147] [About the DGR protection function backup mode] There are two modes for the DGR protection function. You can select the DGR protection function standard mode (UseMiniGridProtectionBackup = 0) or the DGR protection function backup mode (UseMiniGridProtectionBackup~ = 0). In standard mode, if any of the enabled protection functions trips, the DGR mini-grid interface is disabled. On the other hand, in backup mode, the DGR protection function is delayed and coordinated with the grid connection controller protection function.

[0148] In backup mode, the grid-tie controller protection function trips first, opening the circuit breaker. If coordination is successful, the DGR receives the open state of the circuit breaker from the grid-tie controller and resets the DGR protection function before tripping and disabling the DGR output. The mini-grid loads seamlessly transition from grid-tie to grid-independent operation.

[0149] The DGR backup mode protection function is implemented by adding delayed shutdown logic to the frequency destabilizing active islanding detection. The trip times of the DGR's passive and active islanding detection are not increased, so the DGR trips at the same time as the grid-connection controller. When the protection function detects islanding, the DGR enters a shutdown standby state. The DGR waits for Tm_delay_activeTrip or Tm_delay_passiveTrip (usually 1.0 s to 2.0 s for both) before disabling the mini-grid interface. During the waiting time, the active and islanding functions are disabled, allowing the frequency to stabilize. If the DGR receives a breaker with synchronism detection function open status from the grid-connection controller during the waiting time, the DGR resets the protection function and continues operating in grid-independent mode.

[0150] In the backup mode protection function, a trip wait signal (tripWait) suspends detection associated with active islanding perturbations for a wait time. The trip wait signal can be generated by active or passive islanding detection.

[0151] [Grid Tie Controller to DGR Communication] The status of the synchronous breaker is monitored and controlled by the grid tie controller. The synchronous breaker_closed signal is set to "0" for "open" and "1" for "closed." Communication from the grid tie controller to the DGR is typically either a broadcast from the grid tie controller to all DGRs or a grid tie controller server / DGR client structure. The grid tie controller's broadcast method can be configured to provide immediate broadcast updates when the synchronous breaker status changes. The grid tie controller server method relies on the DGR's rapid polling of the synchronous breaker status. The response time of the grid tie controller server method can be improved if the DGR immediately polls the synchronous breaker status (with optional fast repeat) when either the active test or trip standby state becomes active. This allows synchronous breaker status updates to be available to the DGR within 1.0 seconds on most public communication networks.

[0152] For schemes using synchronous breaker mode dependent protection: If the DGR loses communication with the grid connection controller server, the DGR assumes that the synchronous breaker status is closed. When communication from the grid connection controller to the DGR is restored, the DGR can return to normal synchronous breaker mode dependent operation.

[0153] [Effect of active cell grid isolated detection by stabilization / destabilization of frequency depending on the state of the circuit breaker with synchronism test function] The effect of active cell grid isolated detection by stabilization / destabilization of frequency depending on the state of the circuit breaker with synchronism test function of this embodiment will be described.

[0154] The advantages of this embodiment include the following: - Suitable for voltage-based synchronization - Well-known and accepted frequency destabilization islanding detection method - The proposed voltage phase balance destabilization method passes standard RLC load qualification tests - The frequency destabilization method is compatible with the islanding detection method used by most grid-tracking inverters, which is beneficial for cell grids that have a mix of DGR and standard grid-tracking inverters - Ideal for seamless transition between grid-tied and grid-independent operation - Same DGR control possible in grid-tied and grid-isolated operation - Standard communication from the grid-tied controller to the DGR When the cell grid is isolated from the main grid, the DGR active islanding perturbation and backup detection are disabled.

[0155] However, when applying this embodiment, the following points must be taken into consideration: This method is not suitable for global time-based synchronization (e.g., GPS time synchronization). When applying this method, attention must be paid to the stability of the voltage phase balance. For example, if the penetration rate of DGR is low to medium, there is no problem with stability. However, especially when the penetration rate of DGR is high, attention must be paid to the voltage phase balance of the main grid. Furthermore, in order to stabilize the voltage phase balance, attention must be paid to the sensitivity of the islanding detection protection for some polyphase loads such as transformers and motors.

[0156] [Embodiment 3] In embodiment 3, active cell grid islanding detection by detecting local impedance changes will be described. The same components as those in Figures 1 to 9 will be described using the same reference numerals.

[0157] DG sub-brids are augmented with an effective active islanding detection method that does not negate the frequency and voltage stabilization provided by grid-forming control. This requires a signal perturbation method rather than a destabilization method. This method is not commonly used in grid-tracking inverters due to concerns that perturbation signals from the units may interfere and partially cancel each other. This can lead to failure to detect local islanding. Furthermore, large perturbation signals can cause motor voltage fluctuations and torque oscillations.

[0158] In this embodiment, we propose a method to solve the problems of the conventional system. In this method, signal perturbation and detection are synchronized across the mini-grid by utilizing the time-synchronized local clock of the DGR. The local clock of the DGR is time-synchronized by an indirect method using the zero-crossing events of the GNSS signal or the mini-grid voltage. The DGR applies active periodic signal perturbation to the output real power and / or reactive power reference. The perturbation AC fraction Ncyc_pert is calculated from the local clock GPST value tref_local, the perturbation frequency fpert, and the perturbation AC fraction offset Ncyc_pert_offset. N_(cyc_pert) = frac((f_pert.t_(ref_local)) + N_(cyc_per_offset))

[0159] All DGRs operate at the same fpert, typically in the range of 5 Hz to 30 Hz (e.g., 10 Hz). This frequency range is selected to provide multiple cycles of perturbation within the required islanding detection time. All DGRs in a mini-grid, and potentially all mini-grids on the main grid feeder, have the same perturbation offset, Ncyc_pert_offset. This ensures that the active signal perturbations of the DGRs have constructive interference and do not cancel each other out. Thus, all DGRs in a mini-grid act as one large unit for driving the active perturbation signal.

[0160] Over a wider area, the perturbation signal can cause undesirable fluctuations in the voltage and frequency of the power grid. To eliminate this problem, adjacent mini-grids, or possibly adjacent feeder circuits, are configured with opposite active perturbation offsets: for example, Ncyc_pert_offset = 0 on 50% of the sites, and Ncyc_pert_offset = 0.5 on the remaining sites. This creates an overall destructive interference pattern that prevents the active perturbation signal from propagating throughout the main grid system.

[0161] Active islanding detection schemes measure mini-grid voltage or current characteristics and correlate them with a reference or measured perturbation signal. The selected characteristics should have a large difference in value between the grid-connected and local islanding states. The perturbation and measurement signals must be small to avoid unwanted disturbances in the mini-grid's voltage and frequency. In particular, the voltage of the perturbation signal must be below the recognition limit for voltage fluctuations (flicker) during normal grid-connected operation. During unintentional islanding, voltage fluctuations may exceed the flicker limit before islanding is detected. If the mini-grid continues intentional local islanding, voltage fluctuations must return to a level below the flicker limit.

[0162] The low signal levels of the perturbation and measured response characteristics are within the typical background noise levels of power converters such as DGRs. The current and voltage waveforms at the fundamental frequency fgrid are amplitude modulated by the perturbation signal. This results in the formation of sidebands at frequencies fgrid±fpert. Measurements of these current and voltage sidebands can be used to estimate the perturbation and response characteristics. Bandpass filtering of the sideband frequencies is a suitable method. Alternatively, a multi-frequency phase-locked loop (PLL)-type structure can be used to estimate both the magnitude and phase of the fundamental and sideband signals. A similar method for estimating both the magnitude of the fundamental and harmonic signals can be achieved by combining current and voltage sideband measurements to create magnitude and phase measurements of the perturbation and response characteristics. This detection method is highly selective, allowing the perturbation and response signals to be extracted from the general signal noise present in DGRs.

[0163] Some time-varying loads in mini-grids may generate perturbations at or near sideband frequencies. It is important to consider whether such time-varying loads could cause false negative (non-detection) or false positive (nuisance tripping) operation of active islanding protection. It is possible, but unlikely, that time-varying loads may have persistent periodicity with frequencies close to fpert. Periodic power loads may cause periodic time fluctuations in active power.

[0164] Because reactive power tends to be related to equipment rated power rather than specific operating conditions, periodic fluctuations in reactive power are unlikely. The impact of time-varying loads should be constrained by the limits imposed on voltage fluctuations (flicker) in standards IEC 61000-3-03 and IEC 61000-3-11. For example, at fpert = 10 Hz, the peak-to-peak fluctuation per unit of the mini-grid's fundamental rms voltage must be less than 0.4% (short-term) or 0.26% (long-term) (IEC 61000-3-3:2013). Voltage perturbations up to 50% of the long-term flicker limit have negligible flicker effects, but are significant compared to time-varying loads. Assuming the load fluctuation component is no more than five times the magnitude of the perturbation signal, islanding can be detected by the five-fold change in characteristics between grid-tied and islanded operation.

[0165] Active perturbations can be achieved by periodically varying the real or reactive power reference. DGR mini-grid control is a phase diagram-based open-loop method that calculates the magnitude and phase angle of the output voltage based on the DGR real power, reactive power, and voltage references. The active perturbation of the power reference is actually implemented as an active perturbation of the DGR output voltage calculated from the phase diagram. Therefore, the magnitude of the perturbation of the real or reactive power reference directly limits the resulting voltage fluctuations.

[0166] Phase diagram calculations are performed assuming that the total series impedance is primarily inductive and typically less than 20% per unit. Real power references primarily produce voltage perturbations in quadrature (90° out of phase) with the mini-grid voltage. Furthermore, the total quadrature voltage is small compared to the total DC voltage. Because the quadrature voltage perturbations are small, their impact on the overall DGR output voltage magnitude is negligible, resulting in negligible voltage fluctuations (flicker). Reactive power references primarily produce voltage perturbations in phase with the mini-grid voltage. This directly impacts the overall DGR output voltage magnitude and produces a corresponding level of voltage fluctuations (flicker).

[0167] In a grid-tied state, the mini-grid voltage varies with real and reactive power due to the grid impedance. This typically ranges from 0.1% to 5% (per unit) resistive impedance in series with 0.2% to 20% inductive impedance. DGRs add a virtual impedance, typically 1% resistive and 10% inductive impedance. The total effective grid impedance is therefore 1% to 6% resistive in series with 10% to 30% inductive impedance. In a grid-independent state, the mini-grid voltage varies with real and reactive power due to the load impedance in series with the grid impedance. The load impedance is typically 100% or more per unit resistive in parallel with 100% or more inductive or capacitive impedance. In an RLC islanding test, the individual inductive and capacitive impedances can be low, but are balanced so that their effects cancel each other out and only the resistive component is significant.

[0168] [Effects of Motor Load] Local islanding where the motor load is dominant may exhibit a reduced impedance to the perturbation signal at fgrid±fpert. The motor load may be an induction machine or a synchronous machine. However, motors connected with inverters do not have a significantly lower impedance to the perturbation signal. Practical synchronous machines have damping windings that act similarly to the rotor of an induction machine. The impedance of a synchronous machine at the perturbation signal frequency is dominated by the damping windings. Therefore, a synchronous machine can be considered a special case of an induction machine when evaluating the effect on perturbation signals.

[0169] Doubly-fed induction machines are sometimes used, such as in wind turbines. A doubly-fed induction machine has a wound rotor driven by an inverter. The rotor current at the perturbation signal frequency is usually passed through or somewhat attenuated by the inverter. The effective impedance of the rotor circuit and inverter to the perturbation signal is greater than that of a simple induction motor. Therefore, a doubly-fed induction machine can be considered a special case of an induction machine when evaluating its effect on perturbation signals.

[0170] The perturbation signal at fgrid±fpert indicates a slip frequency of fslip±fpert, where fslip is the fundamental slip frequency of the motor. A typical induction motor's full-load fslip ranges from 1 Hz to 2.5 Hz. For Fpert = 10 Hz and fgrid = 50 Hz, typical perturbation slip frequencies are 8 Hz and 12 Hz, with an average effect of 10 Hz. The magnitude of the "perturbation current" must be taken into account when setting trip thresholds and detecting local islanding.

[0171] On the other hand, at the expected "perturbation slip frequency," the resistive component of the motor impedance is greater than or equal to the inductive component. This is advantageous over perturbation current detection. The resistive impedance of the main grid is generally smaller than the inductive impedance, except when the total impedance is low, such as less than 5%. Therefore, the impedance characteristics of the induction motor at the perturbation frequency can be distinguished from the main grid. The simplest approach is to set the lower trip threshold on the quadrature amplitude of the measured perturbation current. This is the component of the measured perturbation current in quadrature with the measured perturbation voltage. To improve sensitivity, the lower trip threshold can be inversely proportional to the perturbation current absolute phase angle (the phase angle relative to the perturbation voltage). For example, the trip threshold increases linearly by 100% as the absolute phase angle decreases from 90° to 45°, and then remains constant from 45° to 0°. Alternatively, the measured perturbation current can be adjusted proportionally to a gain factor, KgIpertMagQ, which is proportional to the measured perturbation current absolute phase angle. From 0° to 45°, a reduced gain value of KgIpertMagQ = 0.5 is applied. From 45° to 90°, the gain increases linearly to 1.0.

[0172] A reference perturbation in real or reactive power induces quadrature and direct (in-phase) DGR output voltage perturbations, respectively. Under grid-tied conditions, the sum of the DGR and grid impedances is primarily inductive. Therefore, a reference perturbation in real or reactive power induces primarily real and reactive current (and power) perturbations, the magnitude of which is comparable to the reference value. In local islanding, the circuit impedance is dominated by the load characteristics. The magnitude of the total impedance is at least five times the magnitude of the grid impedance and typically has a significant real (resistive) component. The magnitude of the current (and power) perturbation is typically reduced by at least five times compared to the reference value, and the phase may be significantly different. Therefore, islanding can be detected when the measured current perturbation signal falls below the trip threshold. The simplest method involves comparing the sum of the current perturbation measurements to a single trip threshold. More complex methods can establish separate trip thresholds for the quadrature and quadrature components of the current perturbation measurement. Other methods may respond to the phase or phase change of the current perturbation measurement.

[0173] Next, an active power / reactive power reference periodic change method will be described as an example of the active cell grid islanding detection method based on the local impedance change detection of this embodiment.

[0174] [B2. Active Power / Reactive Power Reference Periodic Varying Method] Two specific examples of the active power / reactive power reference periodic varying method are shown below: B2a active power reference periodic varying method and B2b reactive power reference periodic varying method. However, this embodiment is not limited to these two specific examples, and any active islanding operation detection method can be adopted as long as it can realize active islanding operation detection.

[0175] As specific perturbation methods, B2a: active power reference period change method and B2b: reactive power reference period change method will be described below.

[0176] [B2a: Active power reference periodic change method] (Active power reference perturbation method) ・Very little effect on mini-grid voltage fluctuations (flicker) ・Most susceptible to time-varying loads ・Possibility of battery power ripple and DC bus voltage ripple

[0177] [B2b: Reactive power reference periodic variation method] (Reactive power reference perturbation method) - Direct impact on mini-grid voltage fluctuations (flicker) - Peak-to-peak fluctuations per unit base voltage should be limited to ≦0.14%. - Least susceptible to time-varying loads. - Mini-grid loads are not expected to have large time-varying reactive power at ff. - The magnitude of the time-varying reactive power load is limited by the system voltage variation limit. - No battery power ripple or DC bus voltage ripple.

[0178] A reactive power reference perturbation is preferred because it is considered the most robust against time-varying loads. Its magnitude should be below the standard limits for voltage fluctuations (flicker) (e.g., 0.14% per unit peak-to-peak). The resulting voltage fluctuations are naturally limited in both grid-tied and grid-independent modes. If the magnitude of the perturbation signal needs to be increased for reliable islanding detection, a scheme that adds a real power reference perturbation can be considered. In principle, active islanding detection schemes can adapt the magnitude of the power reference perturbation in response to the magnitude of the measured current or voltage perturbation.

[0179] Adaptation has the potential to reduce the magnitude of voltage and current perturbations under grid-tied conditions, especially when grid impedance is low. The magnitude adaptation must be coordinated by the grid-tied controller and communicated to the DGR so that the magnitude of the power reference perturbation signal is consistent within the mini-grid. Adaptation by the grid-tied controller is very complex and is not recommended unless excessive voltage and current perturbations under grid-tied conditions are an issue.

[0180] The active islanding perturbation signal and detector are typically disabled when the mini-grid is operating in intentional islanding mode and the synchronism detector circuit breaker is open, i.e., in a grid-independent state. The grid-tied controller communicates the open / closed state of the synchronism detector circuit breaker to the DGR. The desired response to unintentional cell-grid islanding (hereinafter sometimes referred to as "islanding") is for the grid-tied controller to detect the unintentional islanding, open the synchronism detector circuit breaker, and transition the entire mini-grid to intentional islanding operation (with islanding detection disabled). Therefore, the primary active islanding protection is operated by the grid-tied controller.

[0181] The backup active islanding protection is operated by the DGR. If the grid-tie controller does not open the circuit breaker with synchronism detection and transition to intentional islanding, the DGR detects unintentional islanding and either opens the DGR site isolation device or de-energizes the output. For effective coordination, the grid-tie controller's active islanding trip level should be more sensitive than the DGR's active islanding trip level. Also, the grid-tie controller's active islanding trip delay should be shorter than the DGR's active islanding trip delay.

[0182] [B2b: Specific Example of Reactive Power Reference Periodic Variation Method] Active islanding detection is achieved by overcurrent tripping based on the quadrature magnitude of the perturbation current via DGR (direct) voltage reference modulation at 0.14% peak-to-peak per unit (wrt grid rms voltage) at fpert = 10 Hz. The grid-tied controller islanding detection trip limit is less than 0.28% per unit of the total DGR effective rating of the mini-grid. The trip delay, including the perturbation current measurement delay, is 0.75 seconds. When the absolute phase angle decreases from 90° to 45°, the trip threshold increases by 100%. It then remains constant between 45° and 0°. If the peak value per unit voltage is less than 0.035%, tripping is disabled. This condition indicates that the mini-grid is connected to a strong main grid or the DGR has tripped.

[0183] DGR islanding detection trip limit < 0.19% peak to peak per unit wrt DGR effective rating, with a 1.5 second trip delay including perturbation current measurement delay. The trip threshold increases by 100% when the absolute phase angle decreases from 90° to 45°, and then remains constant between 45° and 0°. Tripping is disabled if the peak value per unit voltage is less than 0.035%. This condition indicates that the mini-grid is connected to a strong main grid or the DGR has tripped.

[0184] The grid analysis and protection functions of the grid interconnection controller and DGR of this embodiment will be described with reference to Fig. 10 and Fig. 11. Fig. 10 is a block diagram of the grid interconnection controller main grid protection function for signal-based islanding detection, and Fig. 11 is a block diagram of the DGR cell grid protection function for signal-based islanding detection.

[0185] The grid-tied controller has perturbation signal detection and trip logic. The DGR also includes perturbation signal generation. The grid protection functions of the grid-tied controller and DGR can continue to include phase shift and ROCOF trip logic, which were introduced for the frequency destabilization islanding detection method. ROCOF is a passive islanding detection method that is effective for all types of distributed generation. Phase angle shift tripping is effective in ensuring that the voltage balance of the mini-grid is maintained. Without active frequency destabilization, the dynamic element of phase angle shifting is not necessary.

[0186] For this reason, the block in Figure 12 includes islanding phase shift based on slip angle as a passive islanding detection method by ROCOF. The active islanding signal detection measures both the perturbation voltage magnitude, Vpert_meas, and the perturbation current regulation quadrature magnitude, Ipert_meas. The active signal trip logic applies overcurrent limit logic to Ipert_meas with a trip level of min(IpertTripLevel,(Vpert_meas.YgridTripLevel)) and a reset level of min(IpertResetLevel,(Vpert_meas.YgridResetLevel)). YgridTripLevel = 1 / ZgridTripLevel, where ZgridTripLevel is the grid impedance threshold for islanding detection. IpertTripLevel is the maximum level for the overcurrent trip limit. YgridResetLevel = 1 / ZgridResetLevel, where ZgridResetLevel is the grid impedance threshold for resetting islanding detection. IpertResetLevel is the maximum level of the undercurrent reset limit.

[0187] A trip (or non-reset) condition also requires Vpert_meas to be greater than or equal to the minimum limit (VpertMinLevel). Trip and reset timers are used to provide independent trip and reset delays, TmIp ertTripTime and TmIp ertResetTime, respectively.

[0188] Figure 13 shows the active signal trip logic block diagram. The trip and reset delay timers are implemented as a single integrator with inputs equal to the inverse of ±TmIpertTripTime (untripped) or ±TmIpertResetTime (tripped). In practice, the active signal trip logic can be implemented by a modified version of the standard low limit trip shown in the paper [4]. If the CD_closed status is FALSE, active islanding detection is disabled. The active islanding detection trip logic applies to each mini-grid phase individually. A trip on any phase indicates an islanding or open phase condition.

[0189] The active islanding signal detection measures both the perturbation voltage magnitude Vpert_meas and the perturbation current regulation quadrature magnitude Ipert_meas.

[0190] Figure 14 is an islanding detection perturbation signal generation block diagram, and Figure 15 is a DGR voltage reference calculation block diagram including the islanding detection signal perturbation. The DGR generates the voltage perturbation signal by adjusting the DGR voltage reference as shown in Figure 14. The perturbation AC fraction Ncyc_pert in Figures 14 and 15 is calculated from the local clock GPST value tref_local, the operation frequency fpert, and the perturbation AC fraction offset Ncyc_pert_offset. The value of Ncyc_pert_offset is either 0 or 0.5 and is the same for all DGRs in the mini-grid.

[0191] [Advantages] The advantages of the method of this embodiment are as follows: Supports both global time and voltage-based synchronization No impact on power and frequency dynamics of mini-grids and main grids Passes standard RLC load qualification tests Suitable for seamless transition between grid-tied and grid-independent operation Same DGR control in grid-tied and grid-independent operation Standard grid-tied controller to DGR communication DGR backup active islanding perturbation and detection are disabled in grid-independent state

[0192] However, when using this format, care must be taken to ensure that the active islanding detection time does not exceed the 1.0 second specified in Japan's interconnection requirements. Time synchronization between the cell grid DGR and the grid connection controller is also required. For example, time synchronization via GNSS signals or a grid connection controller time server is required. It is also important to note that although this is within the long-term flicker limit, grid voltage fluctuations may tend to increase.

[0193] [Embodiment 4] A cell-grid independent detection by direct fault detection in an interconnection system according to embodiment 4 of the present invention will be described. The same reference numerals are used for the same components as those in Figures 1 to 8, and their description will be omitted.

[0194] Unintended islanding frequently occurs when main grid distribution circuit isolation devices open in response to a fault in a downstream section of the distribution circuit or in the connected mini-grid and site load. The main grid isolates the fault within the unintended islanding, including the mini-grid. The fault may be within the mini-grid or within an isolated section of the main grid feeder circuit. Phase-to-phase faults may not cause significant disturbances to the phase voltages and currents within the islanding. Therefore, the mini-grid's DGR can continue to maintain local islanding with stable voltage and frequency.

[0195] A fast and effective method of detecting local islanding is for each DGR site transformer to include primary-side protection that can detect and isolate ground faults on the feeder circuit. This is also necessary to ensure that the fault is quickly de-energized to prevent potential electrical hazards and damage. If the fault is within the mini-grid, operation of the DGR site transformer's ground-fault protection will separate the mini-grid into separate load sites, each isolated from the faulted distribution circuit.

[0196] The faulted mini-grid distribution circuit section is isolated and de-energized by the fault protection. A temporary fault can be cleared once the arc is extinguished. A persistent fault is repaired by a maintenance team. Once the fault is cleared, the primary insulation of the DGR transformer can be closed and the mini-grid reconnected. The isolation device is closed only when the mini-grid feeder circuit is de-energized or energized and synchronized with the DGR site.

[0197] DGR transformer primary-side ground fault protection can be based on feeder zero-sequence voltage, zero-sequence current, or a ratio of these. For example, an ungrounded three-wire distribution circuit can use an earth voltage transformer (EVT) and a zero current transformer (ZCT) to detect ground faults. Ground fault protection opens an isolation device between the mini-grid feeder circuit and the DGR transformer primary.

[0198] [Effects] The advantages of this method include: - Rapid and direct detection of faults in local distribution circuits - Isolation of earth faults, interruption of abnormal earth fault currents and / or voltages - Provision of feeder circuit earth fault protection against sustained intentional islanding - Ideal for seamless transition between grid-tied and grid-independent operation

[0199] However, when adopting this method, additional DGR transformer primary side protection is required, which requires attention from the perspective of equipment and installation costs. It is also important to consider that islanding caused by faults other than those in the cell grid's distribution circuits or feeders may not be detected. There is also a risk that islanding caused by open-circuit type faults or isolation devices that have become open circuits due to operational faults may not be detected.

[0200] [Embodiment 5] In a grid-connected system according to embodiment 5 of the present invention, four methods of cell-grid independent detection that can be adopted will be collectively described. Method A: Mode dependent frequency destabilization / stabilization method Method B: Phase-shift detection method Method C: Impedance detection method Method D: Direct fault detection method

[0201] [Method A: Mode-Dependent Frequency Destabilization / Stabilization] The DGR (Distributed Generation Resource) is a client of the grid connection controller server and receives information including the status of the islanding control device (islanding control device). When the islanding control device (islanding device) is open, the mini-grid is in grid-isolated operation. Because the grid is small and frequency is prone to fluctuations, islanding protection based on frequency destabilization is disabled (Kactive_island_miniGrid = 0). Meanwhile, frequency stabilization is enabled (Kstable_island_miniGrid = standard value, e.g., 0.08). When the islanding control device (islanding device) is closed, the mini-grid operates in a connected state with the main grid. In this case, islanding protection based on frequency destabilization is enabled to detect islanding (Kactive_island_miniGrid = standard value, e.g., 0.0159). Meanwhile, frequency stabilization is disabled (Kstable_island_miniGrid = 0). This method is similar to the method described in Figures 7A and 7B.

[0202] The frequency destabilization and frequency stabilization methods have almost the same structure, with the positive and negative gains reversed. Frequency destabilization (active islanding detection) applies positive feedback. If the frequency increases, the output angle of the DGR increases. If the frequency decreases, the output angle of the DGR decreases. Frequency stabilization applies negative feedback. If the frequency exceeds the reference value, the output angle of the DGR decreases. If the frequency is below the reference value, the output angle of the DGR increases. If communication from the grid-connected controller cloud to the DGR fails, the DGR will enter a grid-connected protection state, enabling active islanding detection while disabling frequency stabilization.

[0203] The advantage of Method A is that this islanding detection method is a well-known standard type of slip mode frequency shift method, and is compatible with the islanding detection methods of most grid-following inverters.

[0204] On the other hand, the disadvantage of Method A is that it reduces the frequency stability of the main grid. When the DGR penetration rate is low, the impact is negligible. When the DGR penetration rate is high, this method may not be suitable. There is a delay in communicating the state of the synchronism checker circuit breaker. When the synchronism checker circuit breaker is opened, the mini-grid frequency is destabilized by this function. Therefore, the DGR and / or load may trip before the DGR switches to grid-independent mode. This method is not suitable for seamless transition from a grid-connected state to a grid-independent state. Therefore, fast transmission of the switch state is required when transitioning from a grid-connected state to a grid-independent state.

[0205] [Method C: Impedance Detection] DGRs (Distributed Generation Resources) apply small variations to their output real power and measure the corresponding change in voltage. When the impedance increases, typically above 0.5 per unit, islanding is detected. This is an active islanding protection method. Although this method is well known, it is not widely used because there is no coordination between inverters, which can cause active variations to cancel out, rendering the method ineffective. This method can be adapted by using time synchronization between DGRs to coordinate the active power variations. Therefore, all DGRs apply active power variations in phase.

[0206] The advantage of the C method is that it does not destabilize the frequency or voltage of the main grid, and is suitable for a seamless transition.

[0207] The drawback of Method C is that it requires time synchronization between DGRs, which requires a GPS or timestamp server, and may cause slight fluctuations (flicker) in the grid voltage.

[0208] Method D: Direct Fault Detection ZCTs (Zero Sequence Current Transformers) and EVTs (Earth Voltage Transformers) are used to detect faults in the distribution circuits of the mini-grid.

[0209] The advantages of the D method are that it allows for fast and direct detection of local faults, it allows for isolation of abnormal earth currents and voltages, and it is suitable for seamless transitions.

[0210] The disadvantages of the D method are that the detection circuitry is expensive, it cannot detect islanding caused by faults outside the distribution circuit, and it cannot detect open-circuit type faults and islanding.

[0211] [Verification results of active local islanding detection methods] The results of the active local islanding detection methods are explained using a simulation model. The following two methods were verified as active local islanding detection methods, and both methods were verified to be effective for cell grid islanding detection. - Verification method 1: Verification of balanced RLC islanding detection using frequency stabilization / destabilization method. Results showed that islanding detection is possible even with frequency stabilization enabled. - Verification method 2: Reactive power reference perturbation method was verified for balanced RLC islanding detection. Voltage fluctuation levels were kept below the long-term flicker limit value of IEC 61000-3-3. When applying this to an actual system, verification from the perspective of responsiveness is important.

[0212] [Internal Clock Synchronization Control] An example of internal clock synchronization in the grid interconnection system according to this embodiment will be described with reference to Fig. 16. The same reference numerals are used for components common to Figs. 1 to 15, and descriptions thereof will be omitted. Fig. 16 is an explanatory diagram of the timestamp server in the grid interconnection system according to this embodiment.

[0213] The grid interconnection controllers, μGCs, and DGRs of a grid interconnection system can each be equipped with a GPS antenna to determine accurate GPS time. However, this requires the installation of a GPS antenna for each facility, which imposes constraints on the layout, structure, cost, etc. of each facility. Therefore, the grid interconnection system of this embodiment provides a single timestamp server, which transmits accurate GPS time determined by the single GPS antenna to other grid interconnection controllers, μGCs, and DGRs via a cloud server, thereby correcting the internal clocks of each grid interconnection controller, μGC, and DGR. In this case, a timestamp server using an atomic clock can be used instead of GPS time. Although atomic clocks are expensive, they only need to be installed in one location, so the economic burden is relatively small. Atomic clocks are an effective option because they eliminate problems such as GPS communication outages and fluctuations.

[0214] In the example of Figure 16, an example will be described in which one GPS antenna is provided only in the grid interconnection controller. The grid interconnection controller equipped with a GPS antenna is used as a timestamp server. The timestamp server acquires GPS times corresponding to the zero-cross times t0, t1, and t2 of the grid voltage and transmits them to a cloud server. The other μGCs and DGRs acquire information on the zero-cross times t0, t1, and t2 of the grid voltage from the cloud server. Since the zero-cross times t0, t1, and t2 of the grid voltage are the same everywhere in the grid, the other μGCs and DGRs can accurately acquire GPS time using the zero-cross times t0, t1, and t2 of the grid voltage and the GPS time acquired from the cloud server, and can correct their internal clocks. In Figure 14, it is described that the grid interconnection controller also serves as a timestamp server and transfers timestamp information to the cloud server, but this embodiment is not limited to this. For example, as long as the function of transmitting timestamp information to each grid interconnection controller, μGC, and DGR via the server can be realized, the placement of the server, use of the cloud, etc. are arbitrary.

[0215] The above-described embodiments do not limit the present invention, and the present invention can be equally applied to other embodiments within the scope of the claims. Furthermore, the embodiments can be appropriately modified or combined. For example, in the embodiments, specific DGRs have been used as examples of the grid-forming power converter 40, but the present embodiments are not limited to these DGRs and include any grid-forming power converter as long as it functions as a grid-forming (GFM) inverter.

[0216] 10 Power plant 11 Extra-high voltage substation 12 Power transmission equipment 13 Substation 15 Main system 17 Power flow 20 Cell grid 21 Circuit breaker with synchronization test function (DG breaker) 22 Power demand equipment 22a Household equipment 22b Energy storage device 22c Electric vehicle charging / discharging device 22d Electric vehicle 23 Power generation equipment 23a Wind power generation equipment 23b Photovoltaic power generation equipment 23c Energy storage equipment 23d Internal combustion engine power generation equipment 28 Cell grid system 30, 30A, 30B, 30C, 30D Grid interconnection controller (MGC) 31 Main system frequency measuring device 40 System-forming power conversion device 41 Main system frequency measurement value acquisition unit 80 Inter-system phase difference signal φglobal calculation unit 81 Multiplier 85 Controller 86 Adder 97 Main system voltage acquisition unit 98 Main measurement unit 99 Main measurement unit 150 Circuit breaker open / close state detection unit 160 Selection circuit 161 Input terminal 162 Selection function unit 163 Positive feedback terminal 164 Negative feedback terminal 170 Adder / subtractor

Claims

1. A single - operation detection device for a cell grid system that includes one or more system - forming power conversion devices and is connected to a main system via a connectable or separable switch, comprising a single - operation detector that monitors the operating state of the cell grid to which the system - forming power conversion device belongs and detects that the cell grid is in a single - operation state, wherein the single - operation detector is at least one of an active detector and a passive detector.

2. The single - operation detection device according to claim 1, further comprising, as the passive detector, a detector that detects a single - operation state by detecting a change in the voltage phase of the system - forming power conversion device.

3. As the active detector, the single - operation detection device according to claim 1, comprising at least one of the following detectors: (1) a detector that actively expands and detects the instability of the voltage phase of the system - forming power conversion device to detect a single - operation state; (2) a detector that detects a single - operation state by adding a variation to the active - power reference of the system - forming power conversion device; and (3) a detector that detects a single - operation state by adding a variation to the reactive - power reference of the system - forming power conversion device.

4. The single - operation detection device according to any one of claims 1 to 3, wherein the detection of a single - operation state is performed by at least one of a mode - dependent frequency destabilization / stabilization method, a phase - shift detection method, an impedance detection method, and a direct - fault detection method.

5. The single - operation detection device according to any one of claims 1 to 3, further comprising a function of correcting an internal clock by the time - synchronization information from a timestamp server that acquires time - synchronization information based on a system phase for the system - forming power conversion device and the system - connection controller.

6. The single - operation detection device according to any one of claims 1 to 3, wherein the system - forming power conversion device performs phase - synchronization control for system connection of the cell grid based on the voltage - phase information of the main system.

7. A single-operation detection device for a cell grid system that includes one or more system-forming power conversion devices and is connected to a main system via a connectable or separable switch, the device comprising a single-operation detector that monitors an operating state of the cell grid to which the system-forming power conversion device belongs and detects that the cell grid is operating alone, wherein the single-operation detector is at least one of an active detector and a passive detector.

8. A single-operation detection program, characterized in that the single-operation detection method according to claim 7 is executed by a computer device.

9. A system connection system, characterized in that it comprises at least the single-operation detection device according to any one of claims 1 to 3 and a system connection controller that detects at least voltage phase information of the main system.

10. A system connection system, characterized in that it comprises at least the single-operation detection device according to claim 6 and a system connection controller that detects at least voltage phase information of the main system.

Images