Fault detection method applied to meshed dc microgrid

Abstract

The application discloses a fault detection method applied to a meshed direct-current microgrid, and specifically relates to the following: according to an equivalent circuit model when a fault occurs in the meshed direct-current microgrid, a fault current is analyzed, and a calculation formula of the fault current and first and second derivatives of the fault current are deduced; after the fault occurs in the direct-current microgrid, parameters are determined according to changes of the fault current, and a threshold value is designed to realize fault determination; the fault current is offset by using FCLO; different FCLO values are applied to different lines, event classification is performed by analyzing changes of first derivatives of fault currents of the different lines; and finally, fault isolation is performed according to event types of each unit. The method can complete fault clearing without any communication channel or system shutdown, greatly reduces time delay of fault recovery, and improves stability and response speed of the system.

Description

Technical Field

[0001] The present invention belongs to the technical field of DC microgrid fault detection, and in particular relates to a fault detection method applied to a meshed DC microgrid. Background Art

[0002] With the advancement of renewable clean energy and smart grid construction, flexible direct current (DC) transmission technology is maturing. As a key component of smart grids, DC microgrids are becoming critical infrastructure for the future energy internet. DC microgrids typically consist of distributed micro-sources, energy storage systems, protection and monitoring devices, and AC / DC loads. They offer significant advantages such as high efficiency, simple control, flexible operation, and ease of scalability. However, the presence of numerous power electronic devices in these systems complicates operating conditions, and the high short-circuit current surges and lack of zero-crossing points make fault detection and removal in DC microgrids extremely difficult. Therefore, rapid and reliable fault detection and removal have become a key technology driving the widespread adoption of DC microgrids.

[0003] In multi-busbar meshed DC microgrids, line short-circuit faults are the most common type of failure. Causes can include physical damage, environmental stress, electrical stress, and cable aging. Currently, differential protection technology is often used to address this issue. However, traditional differential protection technology relies on communication channels between device terminals for real-time data transmission, which can lead to communication failures or data loss, resulting in protection failure. Consequently, differential protection systems often require additional backup protection solutions, which not only increases system cost and scale but also limits their application in microgrids. Furthermore, this approach's over-reliance on communication can introduce latency, which in turn increases protection response time and reduces the operational reliability of the DC microgrid. Summary of the Invention

[0004] The present invention aims to provide a fault detection method for a meshed DC microgrid, which can complete fault clearing without any communication channels or system shutdown, greatly reducing the time delay of fault recovery.

[0005] The technical solution adopted by the present invention is a fault detection method for a meshed DC microgrid, which is specifically implemented according to the following steps:

[0006] Step 1: Analyze the fault current based on the equivalent circuit model of the meshed DC microgrid when a fault occurs, and derive the calculation formulas for the fault current and its first-order derivative and second-order derivative;

[0007] Step 2: After a DC microgrid fault occurs, parameters are determined based on the change in fault current, and a low-resistance first-order threshold, a high-resistance first-order threshold, and a high-resistance second-order threshold are designed to achieve fault determination;

[0008] Step 3: Use FCLO to offset the fault current. To determine the fault line, apply different FCLO values ​​to different lines and classify events by analyzing the changes in the first-order derivative of the fault current on different lines. Finally, isolate the fault based on the event type of each unit.

[0009] The present invention is also characterized in that:

[0010] In step 1, specifically:

[0011] Step 1.1: After a fault occurs, calculate the frequency domain fault current i(s) based on the equivalent circuit, as shown in formula (1):

[0012]

[0013] In formula (1), i L (0) and V C (0) are the current through the inductor and the voltage through the capacitor before the fault occurs, L and r are the equivalent series inductance and resistance of the cable to the fault point, R F is the fault resistance, C is the equivalent capacitance; s is the Laplace operator;

[0014] In the time domain, the fault current i(t) can be expressed as formula (2);

[0015]

[0016] Where t is time in milliseconds; p1 and p2 are the extreme points of i(s), as shown in equation (3);

[0017]

[0018] Step 1.2: Calculate the first-order derivative of the fault current; the time derivative of equation (2) is shown in equation (4);

[0019]

[0020] After the fault occurs, at t=0 + When , the magnitude of the current derivative is as shown in formula (5);

[0021]

[0022] Substituting the values ​​of p1 and p2 in formula (3) into formula (5), we get formula (6):

[0023]

[0024] Step 1.3: Calculate the second-order derivative of the fault current. Derivative equation (5) yields the second-order derivative of the line current, as shown in equation (7).

[0025]

[0026] After a fault occurs, at t=0 + When , the magnitude of the second-order derivative of current can be given by formula (8):

[0027]

[0028] Substituting the values ​​of p1 and p2 in formula (3) into formula (8), we can obtain formula (9):

[0029]

[0030] In step 2, specifically:

[0031] Step 2.1: Substitute the sampling time into the first-order derivative and calculate di / dt using the finite difference approximation, which can be expressed as formula (10):

[0032]

[0033] Δi can be written as formula (11):

[0034] Δi=i k -i k-1 (11);

[0035] Where k is the sampling moment, i k and i k-1 are the current and previous sampled line currents respectively;

[0036] Step 2.2, calculation of second-order derivative: d 2 i / dt 2 At the time interval dt, each line is calculated using the current value of di / dt and the previous calculated value, as shown in formula (12);

[0037]

[0038] Among them, Δ 2 i can be written as formula (13) or formula (14):

[0039] Δ 2 i=Δi k -Δi k-1 (13);

[0040] Δ 2 i=i k -2i k-1 +i k-2 (14);

[0041] When the fault occurs at k = 0, k increases by 1 at every sampling moment; when k = 1, since i0 = i -1 =iL (0), so Δi=Δ 2 i; Therefore, when k = 2, Δ is calculated by formula (13) or formula (14) 2 The actual value of i;

[0042] Step 2.3, Δi threshold setting; Δi threshold setting is divided into low resistance first-order threshold and low resistance first-order threshold; low resistance first-order threshold Δi max According to the given DC microgrid parameters, the high resistance first-order threshold Δi min Depends on the operation of the DC microgrid;

[0043] Step 2.4: Set the high-resistance second-order threshold Δ 2 i min ;

[0044] Step 2.5, fault determination; compare the calculated Δi with Δi max Compare and judge whether a fault occurs. If it is greater than, it is directly judged that a fault occurs. Otherwise, continue to the next step for judgment. When Δi is less than Δi min It is directly judged that no fault has occurred, otherwise continue to the next step to calculate the second-order derivative of the real-time fault current Δ 2 i, when it exceeds Δi 2 min If the fault occurs, it is judged that the fault occurs; otherwise, the fault does not occur.

[0045] In step 2.3, when a low resistance fault occurs, i L The amplitude of the first-order derivative of the (0) component is very small and approximately 0. According to formula (6), the low-resistance first-order threshold Δi can be obtained. max The calculation formula is shown in formula (15);

[0046]

[0047] Among them, Leq depends on the line inductance of each unit from the busbar and the line inductance of its parallel branch. For example, when the cable line 1-2 is close to the busbar 1, a polar fault (such as Figure 3a ), the equivalent inductance L eq , as shown in formula (16);

[0048]

[0049] If it is an inter-pole fault (such as Figure 3b ) Its equivalent inductance L eq , as shown in formula (17);

[0050]

[0051] In step 2.3, the high-resistance first-order threshold Δi minThe calculation formula is (18);

[0052]

[0053] In step 2.4, Δ 2 i min The calculation process is shown in formula (19);

[0054]

[0055] In step 3, after receiving the fault signal from step 2, an FCLO is applied to each line to offset the fault current. Because the FCLOs used for different lines have different inductances, the first-order derivative of the fault current varies. By analyzing the changes in the first-order derivatives of the currents of the PV cells, energy storage units, AC grid, and load units, the faults are classified into three event types: 1) bus fault; 2) interconnected feeder fault; and 3) adjacent feeder or busbar fault. Fault isolation is then performed based on the event type.

[0056] The beneficial effects of the present invention are that the method of the present invention determines faults by analyzing the fault current and its first-order and second-order derivatives, and achieves fault isolation through event classification. The process is simple, fast, efficient, and reliable. Furthermore, during the fault detection process, the system relies solely on locally measured parameters and can complete fault clearing without any communication channels or system downtime, greatly reducing the time delay for fault recovery. Furthermore, since the fault isolation portion does not require the transmission of real-time data such as voltage and current in traditional protection schemes, and only needs to transmit and receive event types, it significantly reduces dependence on communications and improves the stability and response speed of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a model structure diagram of a meshed DC microgrid in the method of the present invention;

[0058] Figure 2a is the short-circuit fault equivalent circuit diagram (1) in the method of the present invention;

[0059] Figure 2b This is the short-circuit fault equivalent circuit diagram (II) in the method of the present invention;

[0060] Figure 2c This is the short-circuit fault equivalent circuit diagram (III) in the method of the present invention;

[0061] Figure 3a is a polar fault and its RLC equivalent circuit diagram in the method of the present invention;

[0062] Figure 3b is an inter-pole fault and its RLC equivalent circuit diagram in the method of the present invention;

[0063] Figure 4 is a flow chart of fault judgment in the method of the present application;

[0064] Figure 5a is a fault current chart when fault F23 occurs at a distance of 35 m from the busbar in the method of the present application;

[0065] Figure 5b is a fault current chart when fault F34 occurs at a distance of 35 m from the busbar in the method of the present application;

[0066] Figure 6 is a flow chart of event classification and isolation decision in the method of the present application;

[0067] Figure 7 is a chart of the influence of different FCLOs on current deviation in the method of the present application;

[0068] Figure 8a is a chart of fault current and its first derivative when FCLO of 0.6 mH is applied to feeder 2-3 in the method of the present application;

[0069] Figure 8b is a chart of fault current and its first derivative when FCLO of 1 mH is applied to feeder 3-4 in the method of the present application;

[0070] Figure 9a is a chart of detection result when fault occurs at busbar 3;

[0071] Figure 9b is a chart of current simulation result of busbar 3;

[0072] Figure 9c is a chart of voltage simulation result of busbar 3;

[0073] Figure 10a is a chart of detection result when fault occurs at feeder 2-3;

[0074] Figure 10b is a chart of current simulation result of feeder 2-3;

[0075] Figure 10c is a chart of voltage simulation result of feeder 2-3;

[0076] Figure 11a is a chart of detection result when fault occurs at feeder 3-4;

[0077] Figure 11b is a chart of current simulation result of feeder 3-4;

[0078] Figure 11c is a chart of voltage simulation result of feeder 3-4. DETAILED DESCRIPTION

[0079] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0080] Example 1

[0081] The present invention is applied to a fault detection method of a meshed DC microgrid, wherein the structure of the meshed DC microgrid is as follows: Figure 1 As shown, it consists of four parts: photovoltaic unit, energy storage unit, AC grid unit and load unit. Figure 1 The possible bus fault and feeder fault locations are also marked, where F1 to F4 represent bus faults, and F12, F14, F23, F24, and F34 represent feeder faults. DC short circuit fault is the most serious fault type in DC microgrids. No matter where the short circuit occurs, it can be Figure 2a A short-circuit fault in a DC circuit can result in high fault currents due to the low impedance offered by the charged capacitors and cables.

[0082] In order to comprehensively analyze the response of the nonlinear circuit, the present invention performs two different stages of fault analysis.

[0083] Phase 1: Capacitor discharge. When a short circuit fault occurs, the voltage stabilizing capacitor will quickly discharge along the fault circuit, causing the voltage to drop to zero quickly and the current to rise rapidly. At this time, the equivalent circuit consists of the equivalent resistance (r), the fault resistance (R F ), equivalent capacitance (C) and equivalent inductance (L), forming a second-order discharge circuit, as shown in the following example: Figure 2b As shown;

[0084] Phase 2: Diode freewheeling phase. When the DC bus voltage drops to zero and becomes negative, the system enters the second phase, as shown in the following example. Figure 2c As shown in Figure 1, the voltage reversal affects the voltage at the converter terminal and causes the diode to conduct. The figure shows an equivalent model of two diodes conducting in any converter branch. The diodes provide an alternate path for current independent of the IGBT conduction state, thus changing the current response. During this stage, the diodes experience a current nearly ten times the nominal current, resulting in a sharp increase in current that could damage the diodes.

[0085] Therefore, the present invention aims to timely identify DC network faults through fault detection before the system enters the second stage, thereby preventing the power electronic converter and its components from being damaged by high fault currents and ensuring the stability and safety of the system.

[0086] Example 2

[0087] The present invention is applied to the fault detection method of the meshed DC microgrid, which is specifically implemented according to the following steps:

[0088] Step 1, according to the equivalent circuit model of the meshed DC microgrid when the fault occurs, the fault current is analyzed, and the calculation formula of the fault current and the first and second derivatives thereof are derived;

[0089] Step 1.1, after the fault occurs, according to the equivalent circuit as shown in Figure 2b , the frequency domain fault current i(s) is calculated, as shown in equation (1);

[0090]

[0091] In equation (1), i L (0) and V C (0) are the current through the inductance and the voltage through the capacitance before the fault occurs, L and r are the equivalent series inductance and resistance of the cable to the fault point, R F is the fault resistance, C is the equivalent capacitance; s is the Laplace operator;

[0092] In the time domain, the fault current i(t) can be expressed as equation (2);

[0093]

[0094] In the equation, t is time, in milliseconds;

[0095] p1 and p2 are the poles of i(s), as shown in equation (3);

[0096]

[0097] Step 1.2, the first derivative of the fault current is obtained; in equation (2), the fault current i(t) depends on V C (0), for a given line current i L before the fault of the DC microgrid, the capacitance voltage V C (0) is fixed, so the time derivative of equation (2) is as shown in equation (4);

[0098]

[0099] At t = 0 + after the fault occurs, the size of the current derivative is as shown in equation (5);

[0100]

[0101] Substitute the values of p1 and p2 in equation (3) to obtain equation (6):

[0102]

[0103] When R F= 0, the equivalent cable resistance before the fault point is r. For short-circuit fault current, relative to V C (0), i L The contribution of the (0) component to the first-order derivative amplitude is very small and can be ignored. Therefore, when determining the main protection threshold, only V C (0)Quantity.

[0104] Step 1.3: Calculate the second-order derivative of the fault current. By taking the time derivative of equation (5), the second-order derivative of the line current can be obtained, as shown in equation (7).

[0105]

[0106] At t = 0 + After a fault occurs, the magnitude of the second-order derivative can be given by formula (8):

[0107]

[0108] Substituting the values ​​of p1 and p2 into the circuit parameters, we can obtain equation (9):

[0109]

[0110] From formula (9), we can see that the size of the second-order derivative depends on the equivalent resistance R = R F +r. For R=R c , the initial line current contribution to the second derivative will be zero, where R c =L / C. So when 0 <R F <R c This is a low impedance fault.

[0111] Step 2: After a DC microgrid fault occurs, determine the parameters based on the change of the fault current and design appropriate setting values ​​to achieve fault determination;

[0112] Step 2.1: Substitute the sampling time into the first-order derivative and calculate di / dt using the finite difference approximation, which can be expressed as formula (10):

[0113]

[0114] Δi can be written as formula (11):

[0115] Δi=i k -i k-1 (11);

[0116] Where k is the sampling moment, i k and i k-1are the line currents of the current and the previous sampling, respectively. Since the sampling time of all measurements is the same, only the current difference is measured at consecutive sampling moments to calculate the time derivative.

[0117] Step 2.2, calculation of second-order derivative: d 2 i / dt 2 At the time interval dt, each line is calculated using the current value of di / dt and the previous calculated value, as shown in formula (12);

[0118]

[0119] Among them, Δ 2 i can be written as formula (13) or formula (14):

[0120] Δ 2 i=Δi k -Δi k-1 (13);

[0121] Δ 2 i=i k -2i k-1 +i k-2 (14);

[0122] The fault occurs when k = 0, and k increases by 1 every sampling moment. When k = 1, since i0 = i -1 =i L (0), so Δi=Δ 2 Therefore, when k=2, Δ is calculated by equation (13) or equation (14). 2 The actual value of i is used for fault detection. 2 i is compared with the respective threshold settings. Once the calculated value exceeds the threshold, a fault signal is issued.

[0123] Step 2.3, Δi threshold setting; Δi threshold setting is divided into low resistance first order threshold and low resistance first order threshold. Low resistance first order threshold (Δi max ) is calculated based on the given DC microgrid parameters, and the high-resistance first-order threshold (Δi min ) depends on the operation of the DC microgrid.

[0124] 1) Low resistance first-order threshold: The magnitude of Δi depends on the fault location and its impedance. For a direct short-circuit fault occurring at the end of the line, the threshold of Δi is calculated. The equivalent circuit considered in this case is 0≤R F ≤R c When a fault occurs, the time interval Δ t The expression of Δi can be derived from formula (7), Δi max The calculation process of is shown in formula (15);

[0125]

[0126] Where, if it is a polar fault, the equivalent inductance L eq , as shown in formula (16);

[0127]

[0128] If it is an inter-pole fault, the equivalent inductance L eq , as shown in formula (17);

[0129]

[0130] 2) High impedance first-order threshold: For high impedance faults, line load also plays an important role in determining the magnitude of Δi. In order to calculate the fault threshold Δi min , considering the case where the fault impedance is 25Ω, use formula (7) to calculate the Δi value of different line loads. F >>r,Δi min The calculation formula is (18);

[0131]

[0132] Substitute the equivalent inductance calculated from equation (16) to the fault point into the above equation, Δi min The value depends on the line current i at the time of the fault L (0), the threshold setting of Δi changes with the operation of the DC microgrid. Therefore, Δi min The setting is self-adjusted according to the different operation modes of the microgrid;

[0133] Step 2.4, high-resistance second-order threshold Δ 2 i min Setting Δ 2 i min Similar to the calculation of Δi, Δ 2 i is also derived analytically. 2 i min The value is based on the worst case R F The goal is to detect faults in the grid before starting the second phase. Substituting R = 25Ω into equation (9), we can calculate Δ 2 i min , as shown in formula (19);

[0134]

[0135] Since R F >>R c, the initial line current effect cannot be ignored. In this mode, the current is over-damped because (R / L eq ) 2 > 1 / L eq C, therefore, i L (0) will also contribute to Δ 2 i min . Since the line current depends on the way the DC microgrid works, the calculation of Δ 2 i min should be self-adjusted.

[0136] Step 2.5, fault determination. Compare the calculated Δi with Δi max to determine whether a fault has occurred. If it is greater, it is directly determined that a fault has occurred, otherwise continue to the next step to determine whether Δi is less than Δi min , then directly determine that no fault has occurred, otherwise continue to the next step to calculate the second derivative of the real-time fault current Δ 2 i, when the variable exceeds Δi 2 min , it is determined that a fault has occurred, otherwise no fault has occurred.

[0137] If it is determined that a fault has occurred, the event classification module begins to determine the specific line where the fault occurs and outputs the fault isolation scheme to isolate the fault.

[0138] Step 3, offset the fault current using FCLO; different FCLO values are applied to different lines, and event classification is performed by analyzing the changes in the first derivative of the fault current of different lines. Subsequently, the event type of each unit is transmitted to the central processing unit through data communication, and finally fault isolation is performed according to the event type; specifically:

[0139] Step 3.1, receive the fault signal through step 2, but cannot accurately determine the line where the fault occurs and isolate it. Therefore, FCLO is applied to each line to offset the fault current, thereby improving the accuracy of fault location.

[0140] Step 3.2, since the inductance taken by FCLO of different lines is different, the current offset amount is different, so the first derivative of the fault current will change differently. By analyzing the first derivative of different currents, the fault is divided into three event types: 1) bus fault; 2) interconnected feeder fault; 3) adjacent feeder or adjacent bus fault;

[0141] Step 3.3, the event type of each unit (including photovoltaic unit, energy storage unit, AC grid and load unit) is transmitted to the central processing unit through data communication, and finally fault isolation is performed according to the event type; as shown in Table 1;

[0142] Table 1 Event types and their isolation decisions

[0143]

[0144] Example 3

[0145] From formula (2), we can see that the fault current increases rapidly after the fault occurs, so it is necessary to quickly detect the fault. According to formula (7), the low impedance pole-to-ground short circuit fault and the pole-to-pole short circuit fault are + The first-order derivative of the fault current at the moment is consistent, so the pole-to-pole fault and low-impedance pole-to-ground fault can be judged by the first-order derivative of the fault current. However, when a high-impedance pole-to-ground short-circuit fault occurs, it will be missed because its first-order derivative is small. Therefore, it can be obtained from the analysis of formula (9) that the second-order derivative of the fault current is not only affected by V C Influence, line equivalent impedance R (R = r + R F ) also has a great influence on it, so the high impedance polar short circuit fault can be detected by the second-order derivative of the fault current, and the polar short circuit equivalent inductance L can be calculated according to formula (16): eq ,like Figure 3a As shown, the inter-pole fault equivalent inductance L is calculated according to formula (17): eq ,like Figure 3b As shown. Figure 4 As shown, first calculate Δi according to formula (15), formula (18), and formula (19) respectively max , Δi min and Δi 2 min , calculate the real-time first-order derivative Δi through the local real-time current data, and compare the calculated Δi with Δi max Compare and judge whether a fault occurs. If it is greater than, it is directly judged that a fault occurs. Otherwise, proceed to the next step. When Δi is less than Δi min It is directly determined that no fault has occurred, otherwise continue to the next step and calculate the second-order derivative of the real-time fault current Δ 2 i, when the variable exceeds Δi 2 min If the fault is less than , the fault has not occurred. If the fault is determined to have occurred, the event classification module begins to determine the specific line where the fault occurred and outputs a fault isolation plan to isolate the fault.

[0146] In order to optimize and refine the fault isolation process of the microgrid, it is first necessary to solve the problem of overlapping of the fault current and its first-order derivative when the fault occurs near the busbar. Figure 5a and Figure 5bIt shows that faults F23 and F34 occurred 35 meters away from the busbar, which made the fault current identification vague and the fault type unable to be accurately distinguished. This would lead to an excessively large isolation range, causing unnecessary power outages and affecting the stability and reliability of the microgrid.

[0147] In order to solve this problem, the following optimization method is proposed:

[0148] 1. Fault current offset: By introducing different FCLOs into the line, the fault current can be effectively separated from the current near the busbar, avoiding current overlap. This offset allows the fault current to be more clearly distinguished in the first-order derivative, thereby improving the accuracy of fault detection.

[0149] 2. Fault event classification: Based on the offset fault current and its first-order derivative, the event classification module can autonomously classify each fault event into one of three types:

[0150] Busbar fault (event 1)

[0151] Interconnection feeder failure (event 2)

[0152] Adjacent feeder or adjacent busbar fault (event 3)

[0153] 3. Fault isolation decision-making: By enabling each unit to autonomously classify the fault type and combining it with high-level data from other interconnected units, the system can make more accurate fault isolation decisions. Figure 6 The fault isolation decision-making process is demonstrated: after identifying the specific fault type, the area to be isolated is determined based on the different fault types and location information to reduce the scope of fault propagation and power outage.

[0154] By introducing FCLO for fault current offset, combined with precise event classification and high-level data communication, the system can more effectively identify fault types and make appropriate isolation decisions. This approach not only improves fault diagnosis accuracy but also optimizes isolation strategies, avoiding over-isolation, thereby ensuring stable microgrid operation and reducing unnecessary power outages.

[0155] Example 4

[0156] Figure 7 The short-circuit current (i) of fault F34 occurring at a distance of 35 meters from the busbar is shown when FCLO is 0.3mH, 1mH, and 3mH respectively. ACG ). By analyzing Figure 7The following conclusions can be drawn: When the FCLO is set to 0.3mH, the slope of the fault current decreases, but the maximum current value does not change much. However, when the FCLO is increased to 1mH and 3mH, the maximum fault current decreases as the slope decreases. The results show that the introduction of the FCLO and the use of current derivatives can significantly improve the accuracy of fault classification.

[0157] It should be noted that the increase of FCLO will lead to an increase in power loss. In this study, when the FCLO is selected as 1mH, the efficiency of the system under full load operation only decreases by 0.25%. Therefore, in order to take into account the effect of fault classification and the operating efficiency of the system, it is recommended that the selection range of FCLO be between 0.3mH and 1mH. Specifically, the FCLO of feeder 2-3 is set to 0.6mH, and the FCLO of feeder 3-4 is set to 1mH. When faults F23 and F34 occur at a distance of 35 meters from the busbar, the short-circuit current (i ACG ) and its first-order derivatives such as Figure 8a and Figure 8b As shown;

[0158] Example 5

[0159] In this system, when a fault occurs, the fault's presence is first determined. However, when the fault occurs close to the busbar, it may not be possible to accurately determine the isolation solution. Therefore, different FCLOs are applied to limit and offset the fault current, and then events are classified based on the first-order derivatives of the different fault currents. Each unit communicates and receives event types to collaboratively make fault isolation decisions. Through this optimized solution, the system can quickly respond to and isolate faults in various fault scenarios, as demonstrated below:

[0160] When the fault occurs at bus 3;

[0161] like Figure 9a As shown, detection and isolation time: when a fault occurs on bus 3, the present invention can complete fault detection and isolation within 0.2ms.

[0162] Current and voltage recovery: After fault isolation, the current i of bus 3 ACG With voltage V ACG like Figure 9b and Figure 9c As shown, return to normal.

[0163] Isolation solution: Because the fault occurred on bus 3, the system isolated the entire AC grid unit to ensure that the fault was quickly resolved.

[0164] The fault occurred on feeder 2-3, about 5 m away from busbar 3;

[0165] Detection and isolation time: When a fault occurs on feeder 2-3 and is about 5m away from busbar 3, the present invention can complete fault detection and isolation within 0.7ms.

[0166] Example 6

[0167] like Figure 10a As shown, the detection and isolation time is: when the fault occurs in feeder 2-3 and is about 5m away from busbar 3, the present invention can complete fault detection and isolation within 0.7ms.

[0168] Current and voltage recovery: After fault isolation, the current i of feeder 2-3 ACG and voltage V ACG Return to normal. The specific recovery process is as follows: Figure 10b and Figure 10c shown.

[0169] Isolation solution: Since the fault occurs on feeder 2-3, the system will isolate only this feeder.

[0170] The fault occurred on feeder 3-4, about 5 m away from busbar 3;

[0171] like Figure 11a As shown, the detection and isolation time is: when the fault occurs in feeder 3-4 and is about 5m away from busbar 3, the present invention can complete fault detection and isolation within 0.85ms.

[0172] Current and voltage recovery: After fault isolation, the current i of feeder 3-4 ACG and voltage V ACG Back to normal, the recovery process is as follows Figure 11b and Figure 11c shown.

[0173] Isolation solution: Similarly, the system will only isolate feeder 3-4 where the fault is located to ensure minimal impact on the power grid.

[0174] The present invention first uses the fault current and its first- and second-order derivatives of each unit to identify faults. It then outputs the fault type based on the differential impact of the FCLO on the first-order derivative of the fault current. This information is combined with the outputs of other interconnected units to make a fault isolation decision. To validate the proposed protection scheme, the present invention established a precise microgrid model in the MATLAB / Simulink simulation environment and conducted simulation verification.

Claims

1. A fault detection method applied to a meshed DC microgrid, characterized in that: Please follow the steps below to implement it: Step 1: Analyze the fault current based on the equivalent circuit model of the meshed DC microgrid when a fault occurs, and derive the calculation formulas for the fault current and its first-order derivative and second-order derivative; specifically: Step 1.1: After a fault occurs, calculate the frequency domain fault current i(s), as shown in formula (1); In formula (1), i L (0) and V C (0) are the current through the inductor and the voltage through the capacitor before the fault occurs, L and r are the equivalent series inductance and resistance of the cable to the fault point, R F is the fault resistance, C is the equivalent capacitance; s is the Laplace operator; In the time domain, the fault current i(t) can be expressed as formula (2); Where t is time in milliseconds; p1 and p2 are the extreme points of i(s), as shown in Equation (3); Step 1.2: Calculate the first-order derivative of the fault current; the time derivative of equation (2) is shown in equation (4); After the fault occurs, at t=0 + When , the magnitude of the current derivative is as shown in formula (5); Substituting the values ​​of p1 and p2 in formula (3) into formula (5), we get formula (6): Step 1.3: Calculate the second-order derivative of the fault current. By taking the time derivative of equation (5), the second-order derivative of the line current can be obtained, as shown in equation (7). After a fault occurs, at t=0 + When , the magnitude of the second-order derivative of current can be given by formula (8): Substituting the values ​​of p1 and p2 in formula (3) into formula (8), we can obtain formula (9): Step 2: After a DC microgrid fault occurs, parameters are determined based on the change in fault current, and a low-resistance first-order threshold, a high-resistance first-order threshold, and a high-resistance second-order threshold are designed to achieve fault determination; specifically: Step 2.1: Substitute the sampling time into the first-order derivative and calculate di / dt using the finite difference approximation, which can be expressed as formula (10): Δi can be written as formula (11): Δi=i k -i k-1 (11); Where k is the sampling moment, i k and i k-1 are the current and previous sampled line currents respectively; Step 2.2, calculation of second-order derivative: d 2 i / dt 2 At the time interval dt, each line is calculated using the current value of di / dt and the previous calculated value, as shown in formula (12); Among them, Δ 2 i can be written as formula (13) or formula (14): Δ 2 i=Δi k -Δi k-1 (13); Δ 2 and=i k -2i k-1 +and k-2 (14); When the fault occurs at k = 0, k increases by 1 at every sampling moment; when k = 1, since i0 = i -1 =i L (0), so Δi=Δ 2 i; Therefore, when k = 2, Δ is calculated by formula (13) or formula (14) 2 The actual value of i; Step 2.3, Δi threshold setting; Δi threshold setting is divided into low resistance first-order threshold and low resistance first-order threshold; low resistance first-order threshold Δi max According to the given DC microgrid parameters, the high resistance first-order threshold Δi min Depends on the operation of the DC microgrid; Step 2.4: Set the high-resistance second-order threshold Δ 2 i min ; Step 2.5, fault determination; compare the calculated Δi with Δi max Compare and judge whether a fault occurs. If it is greater than, it is directly judged that a fault occurs. Otherwise, continue to the next step for judgment. When Δi is less than Δi min It is directly judged that no fault has occurred, otherwise continue to the next step to calculate the second-order derivative of the real-time fault current Δ 2 i, when it exceeds Δi 2 min If the fault occurs, it is judged that the fault occurs, otherwise the fault does not occur; Step 3: Use FCLO to offset the fault current; apply different FCLO values ​​to different lines and classify events by analyzing the changes in the first-order derivatives of the fault currents of different lines; finally, isolate the fault based on the event type of each unit; specifically: After receiving the fault signal in step 2, an FCLO is applied to each line to offset the fault current. Since the inductance of the FCLO varies between lines, the current offset varies, resulting in different changes in the first-order derivative of the fault current. By analyzing the different first-order derivatives of the current, the fault is divided into three event types: 1) bus fault; 2) interconnected feeder fault; and 3) adjacent feeder or adjacent busbar fault. The event type is determined by the photovoltaic unit, energy storage unit, AC grid, and load unit, and ultimately the fault is isolated based on the event type.

2. The fault detection method for a meshed DC microgrid according to claim 1, wherein: In step 2.3, the low-resistance first-order threshold Δi max The calculation process of L is shown in formula (15); eq is the equivalent inductance; 3. The fault detection method for a meshed DC microgrid according to claim 2, wherein: If it is a polar fault, the equivalent inductance L eq , as shown in formula (16); If it is an inter-pole fault, the equivalent inductance L eq , as shown in formula (17);

Images