Two-level islanding detection method for DC microgrids

Abstract

The present invention discloses a two-stage islanding detection method for DC microgrids, specifically: in the first stage islanding detection, the magnitude of the output voltage change rate is calculated. If it exceeds the set minimum threshold, it is judged as suspected islanding 1. Then, a five-layer wavelet analysis is performed on the PCC voltage. When the fourth-layer high-frequency coefficient d4 exceeds the set threshold, it is judged as a non-islanding event. When it is lower than the set threshold, it is judged as suspected islanding 2. In the second stage, the output voltage is sent to U PV Reference value U pvref A sinusoidal half-wave disturbance is injected into the circuit. After adjustment by the DC / DC converter, the PCC voltage in the islanding state undergoes periodic sinusoidal half-wave fluctuations. The occurrence of islanding is determined by detecting the PCC voltage fluctuation state and frequency. Combining passive and active methods to complete the entire islanding detection process leverages their strengths to overcome their weaknesses, optimizes detection performance, shortens detection time, and minimizes the non-detection zone to a negligible size.

Description

Technical Field

[0001] The present invention belongs to the technical field of direct current microgrid detection, and in particular relates to a two-stage islanding detection method applied to a direct current microgrid. Background Art

[0002] The extensive exploitation and use of primary energy sources has not only severely polluted the global environment, but also led to the increasing depletion of conventional energy sources such as coal, oil, and natural gas. Consequently, a variety of renewable clean energy sources, such as solar and wind power, have garnered widespread attention as alternatives. Distributed generation technology has rapidly developed, providing crucial power support for large power grids. However, due to the intermittent and fluctuating nature of renewable energy, direct grid integration of distributed power sources presents significant challenges. Microgrids are small-scale power generation, distribution, and consumption systems that integrate distributed power sources, energy storage systems, energy conversion devices, monitoring and protection devices, and loads. They can operate either in parallel with the main grid or in an isolated manner, helping to improve the distribution system's ability to accommodate distributed power sources and enhance the utilization of renewable energy. Depending on how a DC microgrid is connected to the main AC grid, its operating modes can be categorized as either grid-connected or islanded.

[0003] Islanding occurs when the main grid is disconnected from the microgrid due to a fault or scheduling issue, leaving the distributed generation (DG) to independently supply power to local loads, creating an uncontrollable, self-powered system. Islanding can be divided into planned and unplanned. Planned islanding occurs when the grid operator or DG operator intentionally disconnects the main grid from the microgrid due to operational scheduling requirements. Unplanned islanding occurs when the main grid fails or other reasons disconnect the microgrid from the main grid. Microgrids experiencing unintended, unplanned islanding are uncontrollable and require prompt detection and either switching to islanding mode or shutting down power. Failure to do so poses a significant threat to human and equipment safety. Therefore, islanding detection in microgrids is essential and necessary for both safety and stability.

[0004] In DC distribution systems, islanding detection is more difficult than in AC systems because electrical quantities such as frequency and phase are unavailable. The only method for detecting islanding is the voltage amplitude at the point of common coupling (PCC), making it more challenging. Current DC islanding detection methods can be categorized into three types: remote, passive, and active. Remote detection technology relies on communication within the power system and offers advantages such as fast detection speed and high reliability, but its high cost limits its practical application. Passive methods, based on threshold settings, offer simple calculations and fast response, but they can result in a large non-detection zone (NDZ) when the generated power of the distributed generation (DG) is very close to the load power. Active methods, based on small disturbance injection, offer the advantage of a small detection blind zone, but the injected disturbance can affect power quality.

[0005] Passive and active methods have lower costs, but their disadvantages are also more prominent. Consider whether it is possible to use a combination of passive and active methods for islanding detection, so as to achieve the effect of taking advantage of each other's strengths and weaknesses to ensure power supply quality while effectively reducing or eliminating detection blind spots. Summary of the Invention

[0006] The purpose of the present invention is to provide a two-stage islanding detection method for DC microgrids, which completes the entire islanding detection process by combining passive detection method with active detection method, thereby improving the accuracy of islanding detection.

[0007] The technical solution adopted by the present invention is a two-stage islanding detection method applied to a DC microgrid, which is specifically carried out in the following steps:

[0008] Step 1: Analyze the PCC voltage change after the DC microgrid islanding occurs, and set the suspected islanding threshold based on the voltage change rate under different power mismatch conditions;

[0009] Step 2: Use wavelet analysis to distinguish the islanding detection interference factors of motor startup, inter-pole high resistance fault and heavy load switching;

[0010] Step 3: Trigger the active detection part and inject disturbance into the DC microgrid through the DC / DC converter.

[0011] The present invention is also characterized in that:

[0012] In step 1, specifically:

[0013] Step 1.1: In the DC microgrid grid-connected operation mode, establish the load power consumption relationship, as shown in formula (1);

[0014]

[0015] In formula (1), P L is the load power consumption, P G is the grid output power, P DG is the output power of distributed power supply, U PCC is the PCC point voltage, R L is the load equivalent resistance;

[0016] Step 1.2: In the off-grid operation mode of the DC microgrid, the power required by the local load will be provided entirely by the distributed power supply. At this time, the PCC voltage changes due to the rebalancing of the power, and the output power relationship of the distributed power supply can be obtained as shown in formula (2);

[0017]

[0018] In formula (2), U′PCC is the voltage at the PCC point after islanding occurs, and ΔU is the voltage change;

[0019] Step 1.3: Combine equations (1) and (2) to eliminate P DG , the relationship between the voltage change ΔU at the PCC point before and after the DC microgrid islanding occurs is obtained, as shown in formula (3);

[0020]

[0021] Step 1.4: According to the analysis of formula (3), if the PCC voltage U P ' CC Trigger over / under voltage threshold, i.e.: U' PCC <0.8U N or U' PCC >1.05U N When , it is directly judged that an island occurs; U N is the rated voltage of the DC bus;

[0022] Step 1.5: Detect the voltage at PCC point and calculate the voltage change rate V CR , as shown in formula (4):

[0023]

[0024] Substituting equations (2) and (3) into equation (4) yields equation (5);

[0025]

[0026] Step 1.6: Select Power Mismatch According to the condition of (5), the corresponding voltage change rate V CR , based on which the minimum threshold is determined to be 9.9%. CR When the number exceeds the set minimum threshold, it is determined that suspected island 1 has occurred.

[0027] In step 2, specifically:

[0028] Step 2.1, the basis function of wavelet transform is ψ(x), as shown in formula (6);

[0029]

[0030] In formula (6), a is the scale expansion value, and b is the time translation value;

[0031] Step 2.2, wavelet transform coefficient is WT f (a, b), as shown in formula (7);

[0032]

[0033] Step 2.3: Use Haar as the wavelet function to perform a 5-layer decomposition of the PCC voltage. When the fourth-layer high-frequency coefficient d4 exceeds the set threshold, it is judged as a non-islanding event. When it is lower than the set threshold, it is judged as a suspected island 2, triggering the active detection part.

[0034] In step 3, specifically:

[0035] Step 3.1: After suspected island 2 occurs, use the MPPT control of photovoltaic power generation to send power to U PV Reference value U pvref A half-sine wave disturbance U is injected into dis , after the regulation of DC / DC converter, U PV The output power of the photovoltaic system fluctuates in a periodic sinusoidal half-wave, which in turn causes the PCC voltage to also fluctuate in a periodic sinusoidal half-wave.

[0036] Step 3.2: Determine whether islanding occurs based on the PCC voltage fluctuation. If the PCC voltage shows a period of T d If the PCC voltage has no obvious fluctuation or the fluctuation period is not T d , it is determined that the system is in grid-connected operation mode.

[0037] The beneficial effects of the present invention are as follows: the method of the present invention combines passive and active methods to complete the entire islanding detection process, namely, first using the passive method to determine suspected islanding, and then, when a suspected islanding occurs, triggering the active method to inject a disturbance to determine the occurrence of an islanding. This method can achieve the effect of combining strengths and weaknesses to optimize detection performance. On the other hand, the detection time is short and the accuracy of islanding detection is improved. Under normal circumstances, there is no adverse effect on the power supply quality of the microgrid or the power utilization rate of the distributed power source. The non-detection area is very small and can be almost ignored. It can also be used when dealing with multiple distributed power sources, and has a simple structure and low cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a structural diagram of the DC island detection model in the method of the present invention;

[0039] Figure 2 is an equivalent circuit diagram of a DC grid connection mode in the method of the present invention;

[0040] Figure 3 is an equivalent circuit diagram of the DC grid island mode in the method of the present invention;

[0041] Figure 4 is a flow chart of a two-stage islanding detection method applied to a DC microgrid according to the present invention;

[0042] Figure 5This is a simplified DC microgrid simulation system architecture diagram;

[0043] Figure 6 is a graph showing the relationship between the output power and voltage of the photovoltaic system in the method of the present invention;

[0044] Figure 7 is a sinusoidal half-wave disturbance graph injected in the method of the present invention;

[0045] Figure 8 This is a comparison diagram of PCC voltage after islanding under different load conditions in the method of the present invention;

[0046] Figure 9 This is the result of wavelet analysis d4 of PCC voltage under islanding event and non-islanding event;

[0047] Figure 10 This is the PCC voltage simulation result diagram when islanding occurs under the condition of large microgrid power mismatch;

[0048] Figure 11 This is the simulation result diagram of sinusoidal disturbance injection when islanding occurs in this method. DETAILED DESCRIPTION

[0049] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] The present invention is applied to a two-stage island detection method for a DC microgrid. The overall structure is as follows: Figure 1 As shown in Figure 1, it includes the DC microgrid grid connection part and the detection part. Among them, the DC microgrid grid connection equivalent model is as follows: Figure 2 As shown, it is mainly composed of three parts: power management unit, voltage control unit and load. Distributed power sources in DC microgrids, such as photovoltaic systems, often work in MPPT mode in order to obtain electric energy to the maximum extent, and can be equivalent to power management units. Since the capacity of the main power grid is much larger than that of the distributed power generation system, the grid-connected interface converter often acts as a voltage control unit to maintain the stability of the system bus voltage, thereby achieving a balance between power generation and power consumption in the DC system. When the circuit breaker CB is closed, the system operates in grid-connected mode, and the grid and distributed power sources jointly supply power to the load, and the bus voltage is controlled by the grid. When the circuit breaker CB is disconnected, as shown Figure 3 As shown in Figure 1, the system operates in island mode, and the power consumed by the load is only provided by distributed power sources. At this time, the bus voltage is in an uncontrollable state.

[0051] From the analysis of formula (3), it can be obtained that when islanding occurs, if there is power mismatch in the DC microgrid, the DC bus voltage will be offset. The greater the mismatch, the greater the bus voltage offset. GB / T 35727-2017 stipulates that the allowable voltage deviation range for low-voltage DC distribution networks below 1500V is -20%UN ~+5%U N , U N is the rated voltage of the DC bus. Therefore, when the system detects U PCC <0.8U N or U PCC >1.05U N When , the passive method can directly determine that islanding has occurred, that is, the over / under voltage detection method (UVP / OVP) principle.

[0052] Because motor startup, high-resistance faults between poles, and heavy load switching can also cause bus voltage deviations, increasing the rate of change of the output voltage, this passive method cannot accurately distinguish between islanding conditions and interference factors. This means it cannot detect islanding on its own and must be used in conjunction with an active method. The disturbances injected by the active method can affect the power quality of the microgrid. Therefore, to reduce disturbance injection caused by interference factors, this method incorporates wavelet analysis between the passive and active methods to distinguish between islanding conditions and some non-islanding conditions.

[0053] The process of the two-level island detection method for DC microgrid is as follows: Figure 4 As shown in the figure, in the first stage of the proposed islanding detection method, the PCC voltage is measured to determine whether UVP / OVP is triggered. If so, an islanding event is directly determined; otherwise, the next step is continued. The output voltage change rate (VCR) is calculated. When this variable exceeds the set minimum threshold, it is determined that islanding 1 has occurred, which triggers wavelet analysis of the PCC voltage. When the fourth-layer high-frequency coefficient d4 falls below the set threshold, it is determined that islanding 2 has occurred, triggering the second-stage active detection phase.

[0054] In the second stage, after the suspected island 2 occurs, the MPPT control of photovoltaic power generation is used to send power to U PV Reference value U pvref A half-sine wave disturbance is injected into the DC / DC converter, and the DC / DC converter adjusts the U PV The output power of the PV system undergoes periodic sinusoidal half-wave fluctuations. When the microgrid is connected to the grid, the PCC voltage is largely unaffected by the clamping effect of the main grid on the DC bus voltage. However, after an islanding event, the load voltage is determined by the PV output power, and the PCC voltage also undergoes periodic sinusoidal half-wave fluctuations. Islanding detection is achieved by exploiting multiple islanding criteria contained in the periodic sinusoidal half-wave fluctuations of the PCC voltage.

[0055] The two-stage island detection method applied to a DC microgrid of the present invention is specifically performed according to the following steps:

[0056] Step 1: Analyze the PCC voltage change after the DC microgrid islanding occurs, and set the suspected islanding threshold based on the voltage change rate under different power mismatch conditions. Specifically:

[0057] Step 1.1: In the DC microgrid grid-connected operation mode, establish the load power consumption relationship, as shown in formula (1);

[0058]

[0059] In formula (1), P L is the load power consumption, P G is the grid output power, P DG is the output power of distributed power supply, U PCC is the PCC point voltage, R L is the load equivalent resistance;

[0060] Step 1.2: In the off-grid operation mode of the DC microgrid, the power required by the local load will be provided entirely by the distributed power supply. At this time, the PCC voltage changes due to the rebalancing of the power, and the output power relationship of the distributed power supply can be obtained as shown in formula (2);

[0061]

[0062] In formula (2), U P ' CC is the voltage at the PCC point after islanding occurs, and ΔU is the voltage change;

[0063] Step 1.3: Combine equations (1) and (2) to eliminate P DG , the relationship between the voltage change ΔU at the PCC point before and after the DC microgrid islanding occurs is obtained, as shown in formula (3);

[0064]

[0065] Step 1.4: According to formula (3), the grid output power P G The larger the voltage change ΔU is, the larger the voltage change ΔU is. If the PCC voltage U P ' CC Trigger over / under voltage threshold (ie: U' PCC <0.8U N or U' PCC >1.05U N , U N is the rated voltage of the DC bus), it is directly judged that islanding occurs;

[0066] Step 1.5: Detect the voltage at PCC point and calculate the voltage change rate V CR , as shown in formula (4):

[0067]

[0068] Substituting equations (2) and (3) into equation (4) yields equation (5);

[0069]

[0070] Step 1.6: To minimize the detection blind area, select Power Mismatch. According to the condition of (5), the corresponding voltage change rate V CR , based on which the minimum threshold is determined to be 9.9%. CR When the set minimum threshold is exceeded, it is determined that suspected island 1 has occurred;

[0071] Step 2: After suspected islanding 1 occurs, wavelet analysis is used to distinguish the islanding detection interference factors of motor startup, inter-pole high resistance fault, and heavy load switching. Specifically:

[0072] Step 2.1, the basis function of wavelet transform is ψ(x), as shown in formula (6);

[0073]

[0074] In formula (6), a is the scale expansion value, and b is the time translation value;

[0075] Step 2.2, wavelet transform coefficient is WT f (a, b), as shown in formula (7);

[0076]

[0077] Step 2.3: Use Haar as the wavelet function to perform a 5-layer decomposition of the PCC voltage. When the high-frequency coefficient d4 of the fourth layer exceeds the set threshold, it is judged as a non-islanding event. When it is lower than the set threshold, it is judged as a suspected island 2, triggering the active detection part.

[0078] Step 3: Trigger the active detection part and inject disturbance into the DC microgrid through the DC / DC converter. Specifically:

[0079] Step 3.1: After suspected island 2 occurs, use the MPPT control of photovoltaic power generation to send power to U PV Reference value U pvref A half-sine wave disturbance U is injected into dis , after the regulation of DC / DC converter, U PV The output power of the photovoltaic system fluctuates in a periodic sinusoidal half-wave. Due to the lack of the grid-side clamping effect on the bus voltage, the PCC voltage also fluctuates in a periodic sinusoidal half-wave.

[0080] According to the relationship between photovoltaic output power and voltage, when the voltage U pvrefWhen it increases, it will cause the photovoltaic output power to decrease. Therefore, when the photovoltaic MPPT is working normally, the disturbance U dis , which will cause the photovoltaic MPPT to shift, causing the output power to first decrease and then increase, and thus causing the PCC voltage to exhibit periodic sinusoidal half-wave fluctuations;

[0081] Step 3.2: Determine whether islanding occurs based on the PCC voltage fluctuation. If the PCC voltage shows a period of T d If the PCC voltage has no obvious fluctuation or the fluctuation period is not T d , it is determined that the system is in grid-connected operation mode.

[0082] Aiming at a simplified DC microgrid on-grid and off-grid model, the implementation method, parameter selection and simulation verification of the detection method of the present invention are described in detail.

[0083] Figure 5 The following is a simplified DC microgrid simulation system architecture diagram. Among them, the bidirectional grid-connected converter adopts improved droop control, and the photovoltaic module adopts maximum power point tracking (MPPT) control. When the photovoltaic output power P DG Greater than the load power P L When P DG Less than P L When , the grid-connected converter works in the rectification state.

[0084] Figure 6 The figure shows the relationship between photovoltaic output power and voltage under standard conditions. pvref It corresponds to the maximum power point in the current situation. After analysis, it can be obtained that when the voltage U pvref When it increases, the photovoltaic output power will decrease. Figure 7 The figure shows the injected half-sine wave disturbance. It can be analyzed that when the photovoltaic MPPT is operating normally, the injected disturbance U dis , which will cause the photovoltaic MPPT to shift, causing the photovoltaic system output power to first decrease and then increase.

[0085] According to the architecture diagram, a simulation circuit is built in MATLAB to simulate the initial photovoltaic output power P PV is 5000W, load power P L The initial value is 5000W, the DC bus voltage is 400V, and the grid-connected state is converted to the island state at 2s. First, determine the threshold setting of suspected island 1 and set the load power to 5050W, 5150W, and 5250W respectively. At this time, the DC bus voltage will shift after the island occurs, as shown in the following example. Figure 8 As shown. According to the power mismatch (i.e., the load power is 5050W) and the voltage deviation is calculated to determine the voltage change rate. The threshold of suspected island 1 is V CR >9.9%. Secondly, the threshold setting of suspected island 2 is determined. Wavelet analysis is performed on the PCC voltage when the DC microgrid has motor start-up, inter-pole high resistance fault, large load switching and islanding events. The fourth layer high frequency coefficient d4 is obtained as follows: Figure 9 Therefore, in order to better distinguish the occurrence of some non-islanding events (such as motor startup and inter-pole high resistance fault), the threshold of suspected islanding 2 is determined to be d4 < 0.25.

[0086] Using the above thresholds, and performing simulations with a load power of 4000W, after islanding occurs, the bus voltage rises significantly due to the excessive output power of the PV system, triggering OVP and directly determining that islanding has occurred. Figure 10 When the load power is 5050W, the DC bus voltage is as follows: Figure 11 As shown, after the DC island occurs, the two-stage detection is triggered and the DC bus voltage generation period is T d The sinusoidal half-wave fluctuation indicates that the system is in islanding mode. The simulation results show that this method can detect the occurrence of islanding within 150ms.

Claims

1. A two-stage islanding detection method applied to a DC microgrid, characterized in that: Follow these steps: Step 1: Analyze the PCC voltage change after the DC microgrid islanding occurs, and set the suspected islanding threshold based on the voltage change rate under different power mismatch conditions; specifically: Step 1.1: In the DC microgrid grid-connected operation mode, establish the load power consumption relationship, as shown in formula (1); （1）； In formula (1), is the power consumed by the load, is the grid output power, is the output power of the distributed power supply, is the PCC point voltage, is the load equivalent resistance; Step 1.2: In the off-grid operation mode of the DC microgrid, the power required by the local load will be provided entirely by the distributed power supply. At this time, the PCC voltage changes due to the rebalancing of the power, and the output power relationship of the distributed power supply can be obtained as shown in formula (2); （2）； In formula (2), is the voltage at the PCC point after islanding occurs, is the voltage change; Step 1.3: Eliminate the equations (1) and (2) , the voltage change of PCC point before and after the DC microgrid islanding occurs is obtained The relationship is shown in formula (3); （3）； Step 1.4: According to the analysis of formula (3), if the PCC voltage after the change is Trigger over / under voltage threshold, namely: or When , it is directly judged that an island occurs; is the rated voltage of the DC bus; Step 1.5: Detect the voltage at the PCC point and calculate the voltage change rate , as shown in formula (4): （4）； Substituting equations (2) and (3) into equation (4) yields equation (5); （5）； Step 1.6: Select Power Mismatch =1%, the corresponding voltage change rate is obtained according to formula (5) , based on which the minimum threshold is determined to be 9.9%. When the set minimum threshold is exceeded, it is determined that suspected island 1 has occurred; Step 2: After suspected islanding 1 occurs, wavelet analysis is used to distinguish islanding detection interference factors such as motor startup, inter-pole high resistance fault, and heavy load switching; specifically: Step 2.1, the basis function of wavelet transform is , as shown in formula (6); （6）； In formula (6), is the scale expansion value, is the time shift value; Step 2.2, the wavelet transform coefficient is , as shown in formula (7); ， （7）； Step 2.3: Use Haar as the wavelet function to perform a 5-layer decomposition of the PCC voltage. When the high-frequency coefficient d4 of the fourth layer exceeds the set threshold, it is judged as a non-islanding event. When it is lower than the set threshold, it is judged as a suspected island 2, triggering the active detection part. Step 3: Trigger the active detection part and inject disturbance into the DC microgrid through the DC / DC converter; specifically: Step 3.1: After suspected island 2 occurs, use the MPPT control of photovoltaic power generation to Reference value Inject half-sine wave disturbance into , after the regulation of DC / DC converter, The output power of the photovoltaic system fluctuates in a periodic sinusoidal half-wave, which in turn causes the PCC voltage to also fluctuate in a periodic sinusoidal half-wave. Step 3.2: Determine whether islanding occurs based on the PCC voltage fluctuation. If the PCC voltage shows a period of If the PCC voltage has no obvious fluctuation or the fluctuation period is not , it is determined that the system is in grid-connected operation mode.

Images