Direct current power distribution network fault location method based on cosine similarity of detection signals

Abstract

The application discloses a DC power distribution network fault positioning method based on a detection signal and cosine similarity, and the method steps are as follows: according to a double-end modular multilevel converter (MMC) type medium-voltage flexible DC power distribution network topological structure and equipment parameters, the injection strategy of the detection signal is determined; positive and negative voltage signals at the outlet of the MMC are collected, fault identification is carried out, and the fault pole is determined; the zero-mode current of each line head is calculated, and a band-pass filter (BPF) is used to extract the detection signal; in combination with a fault line selection criterion, fault line selection is completed based on cosine similarity. The application has higher accuracy under the condition that the power distribution network structure is complex and the transition resistance is large, and has obvious advantages in positioning time, method stability and universality.

Description

Technical Field

[0001] This invention belongs to the field of DC distribution network fault location technology, and in particular, it is a DC distribution network fault location method based on the similarity between the detection signal and the cosine similarity. Background Technology

[0002] With the development of new power systems, flexible DC distribution networks based on MMC (Multi-Level Capacitor) have received increasing attention due to their high controllability and compatibility with DC loads. However, the unique topology and operating principle of MMC significantly increase the complexity of faults in flexible DC distribution networks, potentially causing serious damage to the power system. Therefore, rapid and accurate fault location is a crucial aspect of ensuring the safe operation of flexible DC distribution networks and is a hot topic and focus of research in DC distribution network protection.

[0003] Given the complex fault mechanisms and difficulty in fault location in flexible DC distribution networks, there is an urgent need to study rapid and accurate fault location strategies for flexible DC distribution networks to ensure the safe and reliable operation of the power system and avoid the harm caused by difficult-to-identify faults. Summary of the Invention

[0004] The purpose of this invention is to provide a fault location method for DC distribution networks based on the similarity between detection signals and cosine similarity, thereby avoiding the harm caused to the system by the difficulty in identifying faults and ensuring the safe and reliable operation of the power system.

[0005] The technical solution to achieve the purpose of this invention is: a DC distribution network fault location method based on the similarity between the detection signal and cosine, comprising the following steps:

[0006] Step 1: Determine the injection strategy for the detection signal based on the topology and equipment parameters of the dual-ended MMC type flexible DC distribution network;

[0007] Step 2: Collect the positive and negative voltage signals at the MMC outlet to identify the faulty pole;

[0008] Step 3: Calculate the zero-mode current at the beginning of each line and extract the characteristic signal;

[0009] Step 4: Combine the fault route selection criteria and extract fault features based on cosine similarity to complete the fault route selection.

[0010] Furthermore, the converter station control strategy in step 1 is as follows:

[0011] The detection signal injection system is a dual-terminal MMC type medium-voltage flexible DC distribution network. The MMC1 converter station adopts constant DC voltage control and constant AC voltage control, i.e., U dc / U acControl; The MMC2 converter station adopts constant active power control and constant reactive power control, i.e., P / Q control; the modulation method is nearest-level approximation modulation;

[0012] The DC distribution network topology in step 1 is as follows:

[0013] The converter station outlet is a DC bus for connecting to the power supply feeder. The two buses are connected by a tie line, which is a cable line. The load types include DC loads, AC loads and distributed photovoltaics.

[0014] Furthermore, the detection signal injection strategy in step 1 is as follows:

[0015] For MMC converters, sinusoidal signals introduce frequency and phase characteristics to fault analysis, therefore the type of detection signal is a sinusoidal signal;

[0016] In terms of control strategy, the sinusoidal detection signal is superimposed with the MMC modulation wave, requiring no additional signal attachment device. The control strategy formula is as follows:

[0017]

[0018] In the formula, u comd_abc The output voltage of the circulating current suppression controller, u ctrl To detect the signal, u abc_ref For AC voltage reference value, u abc_P_ref u abc_N_ref These are the voltage modulation waves of the positive and negative bridge arms of the MMC, U dc This is the DC side voltage;

[0019] The frequency of the probe signal is much smaller than the switching frequency of the MMC submodule. The expression for the open-loop time constant T of the MMC is:

[0020]

[0021] In the formula, ω AC L is the fundamental frequency of the AC system. arm For the MMC bridge arm reactance, C, n, R L These are the submodule capacitor, number of submodules, and equivalent loss resistance, respectively.

[0022] The modulus |Z| of the unit line-to-ground impedance is:

[0023]

[0024] In the formula, R, L, and C are the unit resistance, inductance, and capacitance of the DC system line, respectively;

[0025] At this time, the line impedance to ground is at its minimum, and the fault response characteristics are most obvious; therefore, the detection signal frequency ω is... ctrlfor:

[0026]

[0027] Furthermore, the specific method for fault identification in step 2 is as follows:

[0028] The startup criteria are set so that when a single-pole fault is detected in the system, the MMC additional control strategy is activated to inject a probe signal into the system.

[0029] After a single-pole ground fault, the DC-side fault pole voltage rapidly drops to 0, while the non-fault pole voltage rises to twice its rated value under the control of the MMC constant DC voltage. For a dual-terminal MMC DC system, the absolute values ​​of the zero-mode voltages on both MMC DC sides are used as the start-up criterion.

[0030] |U P1 +U N1 |>0.2U dcB ∪|U P2 +U N2 |>0.2U dcB

[0031] In the formula, U P1 and U N1 These are the positive and negative voltages on the DC side of MMC1, respectively; U P2 and U N2 These are the positive and negative voltages on the DC side of MMC2, respectively; U dcB This is the rated voltage of the DC system;

[0032] To avoid transient interference from the environment or internal system, the criterion is activated when all three consecutive sampling points are true. Then, the fault pole is identified using a fault pole selection criterion, which is as follows:

[0033]

[0034] Furthermore, the specific method for feature signal extraction in step 3 is as follows:

[0035] Taking the injected probe signal as the research object, the proposed line selection criterion is applicable to the probe signal. After collecting the positive and negative currents at the beginning of the line, the zero-mode current is calculated first. The frequency of the probe signal is used as the center frequency of the filter. The cutoff frequency of the BPF is calculated based on the center frequency, and the filter bandwidth is reduced for signal extraction.

[0036] Furthermore, the fault selection criterion in step 4 is:

[0037] a) When a single-pole grounding fault occurs in a feeder, the characteristic signal extracted from the faulty feeder is out of phase with the characteristic signal of the non-faulty line. When there is a bus fault or a tie line fault, the characteristic signals of each feeder are in the same phase.

[0038] b) When a busbar fault occurs, the characteristic signal of the feeder connected to the faulty busbar is in phase with the characteristic signal of the tie line on the busbar side, while the characteristic signal of the feeder connected to the non-faulty busbar is out of phase with the characteristic signal of the tie line on the busbar side.

[0039] c) When a fault occurs in the tie line, the characteristic signal of any feeder connected to any busbar is out of phase with the characteristic signal of the tie line on that busbar side.

[0040] Furthermore, the method for extracting fault features based on cosine similarity in step 4 is as follows:

[0041] First, faulty feeders are identified by calculating the sum of cosine similarities between the characteristic signals of each feeder and other feeders. If the value is negative, it indicates that a single-pole grounding fault has occurred on the line. If there is no negative value, the fault location is not on the feeder.

[0042] When the fault location is determined not to be on the feeder, bus fault determination is performed. The sum of the cosine similarity of the detection signals at both ends of the tie line and the feeder on the same side is calculated. If the value is positive, it indicates that the bus on that side has a fault. If there is no positive value, it is determined to be a tie line fault.

[0043] A fault location system for DC distribution networks based on the similarity between detection signals and cosine similarity is disclosed. The system achieves fault location of DC distribution networks based on the method described above.

[0044] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it performs fault location of a DC distribution network based on the method described above, using the similarity between a detection signal and a cosine similarity.

[0045] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, enables fault location in a DC distribution network based on the similarity between a detection signal and a cosine similarity, according to the method described thereon.

[0046] Compared with the prior art, the significant advantages of this invention are:

[0047] (1) Existing detection signal methods directly use zero-mode current for phase discrimination. The current components in the fault zero-mode current are complex and the frequency distribution is irregular. This invention separates the detection signal from the original signal through BPF, enhances the phase difference between characteristic signals, and improves the line selection accuracy.

[0048] (2) Existing detection signal methods can only identify feeder faults in single-ended MMC type flexible DC distribution networks. This invention performs transient analysis on the fault network of the most widely used double-ended MMC type flexible DC distribution network. The proposed criteria can identify faults on any feeder, bus, and tie line in the system.

[0049] (3) The existing cosine similarity method can only reflect the phase correlation between two lines. The present invention calculates the sum of the cosine similarities between a single line and other lines, which can more intuitively show the difference between a certain line and all other lines.

[0050] (4) Compared with other fault location methods, the method of the present invention is not affected by the transition resistance, has good stability and universality, and has obvious advantages in the case of complex distribution network structure and large transition resistance. Attached Figure Description

[0051] Figure 1 This is a structural diagram of a dual-end power supply type flexible DC distribution network system.

[0052] Figure 2 This is a block diagram of the additional control strategy for injecting probe signals.

[0053] Figure 3 This is a fault selection flowchart. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0055] This invention relates to a fault location method for flexible DC distribution networks based on the detection signal and cosine similarity. The method includes the following steps: determining the injection strategy of the detection signal according to the topology and equipment parameters of a medium-voltage flexible DC distribution network using a Modular Multilevel Converter (MMC) type; acquiring the positive and negative voltage signals at the MMC outlet for fault identification and determining the faulty pole; calculating the zero-mode current at the beginning of each line and extracting the detection signal using a band-pass filter (BPF); and combining the fault selection criteria with cosine similarity to complete the fault selection. Compared with other fault location methods, the method proposed in this invention has higher accuracy in complex distribution network structures and with large transition resistance, and also has significant advantages in location time, method stability, and universality.

[0056] This invention relates to a DC distribution network fault location method based on the similarity between detection signals and cosine similarity, comprising the following steps:

[0057] Step 1: Based on the topology and equipment parameters of the dual-ended MMC type flexible DC distribution network, determine the injection strategy for the detection signal:

[0058] The detection signal injection system is a dual-terminal MMC type medium-voltage flexible DC distribution network, such as Figure 1 As shown, the detection signal injection system is a dual-terminal MMC type medium-voltage flexible DC distribution network, and the MMC1 converter station adopts U... dc / U ac The MMC2 converter station uses P / Q control with NLM (Nearest Level Modulation) modulation. The converter station output is a DC bus for connecting to the power supply feeder. The two buses are connected by interconnecting cable lines, which handle DC loads, AC loads, and distributed photovoltaic loads.

[0059] For MMC converters, sinusoidal signals introduce frequency, phase and other characteristics into fault analysis, effectively improving the problem of unclear characteristics of single-pole grounding faults. Therefore, the type of detection signal is a sinusoidal signal.

[0060] The control block diagram for injecting the probe signal is as follows: Figure 2 As shown, a modulated wave is obtained through decoupling control and circulating current suppression control. A detection signal is then superimposed on the modulated wave. The control strategy formula is as follows:

[0061]

[0062] In the formula, u comd_abc The output voltage of the circulating current suppression controller, u ctrl To detect the signal, u abc_ref For AC voltage reference value, u abc_p_ref u abc_n_ref These are the voltage modulation waves of the positive and negative bridge arms of the MMC, U dc This is the DC side voltage.

[0063] The frequency of the probe signal is much lower than the switching frequency of the MMC submodule, while ensuring that the system fault response characteristics are most pronounced. The open-loop time constant expression for the MMC is:

[0064]

[0065] In the formula ω AC L is the fundamental frequency of the AC system. arm For the MMC bridge arm reactance, C, n, R L These are the submodule capacitor, number of submodules, and equivalent loss resistance, respectively.

[0066] The magnitude of the impedance of a unit line to ground is:

[0067]

[0068] In the formula, R, L, and C represent the unit resistance, inductance, and capacitance of the DC system circuit, respectively.

[0069] At this time, the line impedance to ground is at its minimum, and the fault response characteristics are most obvious; therefore, the detection signal frequency is:

[0070]

[0071] Step 2: Collect the positive and negative voltage signals at the MMC outlet to identify the faulty pole;

[0072] A startup criterion is set up so that when a unipolar fault is detected in the system, the MMC additional control strategy is activated, injecting a detection signal into the system. After a unipolar ground fault, the DC-side fault pole voltage drops rapidly to 0, while the non-fault pole voltage rises to twice its rated value under the control of the MMC's constant DC voltage. Therefore, the absolute value of the zero-mode voltage on the DC side of the dual-terminal MMC is used as the startup criterion.

[0073] |U P1 +U N1 |>0.2U dcB ∪|U P2 +U N2 |>0.2U dcB (5)

[0074] In the formula U P1 and U N1 These are the positive and negative voltages on the DC side of MMC1, respectively; U P2 and U N2 These are the positive and negative voltages on the DC side of MMC2, respectively; U dcB This is the rated voltage for the DC system.

[0075] To avoid transient interference from the environment or internal system, this criterion is activated when all three consecutive sampling points are met. The fault pole is then identified using a fault pole selection criterion, which is as follows:

[0076]

[0077] Step 3: Calculate the zero-mode current at the beginning of each line and extract the characteristic signal;

[0078] This method takes the injected detection signal as the research object. Therefore, the proposed line selection criterion is applicable to the detection signal. After collecting the positive and negative currents at the beginning of the line, the zero-mode current is calculated first. Since the frequency of the injected signal is significantly different from the fault current, the BPF is effective in extracting the detection signal. The frequency of the detection signal is used as the center frequency of the filter. The cutoff frequency of the BPF is calculated based on the center frequency. At the same time, the filter bandwidth should be minimized to ensure the accuracy of signal extraction.

[0079] Step 4: Combine the fault route selection criteria and extract fault features based on cosine similarity to complete the fault route selection;

[0080] The specific idea behind the line selection criterion proposed in this method is as follows:

[0081] a) When a single-pole grounding fault occurs in a feeder, the characteristic signal extracted from the faulty feeder is out of phase with the characteristic signal of the non-faulty line. When there is a bus fault or a tie line fault, the characteristic signals of each feeder are in the same phase.

[0082] b) When a busbar fault occurs, the characteristic signal of the feeder connected to the faulty busbar is in phase with the characteristic signal of the tie line on the busbar side, while the characteristic signal of the feeder connected to the non-faulty busbar is out of phase with the characteristic signal of the tie line on the busbar side.

[0083] c) When a fault occurs in the tie line, the characteristic signal of any feeder connected to any busbar is out of phase with the characteristic signal of the tie line on that busbar side.

[0084] The cosine similarity is used to describe the zero-mode current phase correlation between line i and line j, and its calculation formula is as follows:

[0085]

[0086] In the formula, vector x = {x1, x2, ..., x...} N}, y = {y1, y2, ..., y N} represents the zero-mode current sampling value measured by the current transformers of the two lines; N is the number of sampling points.

[0087] Cosine similarity only reflects the phase correlation between two lines. To more intuitively illustrate the difference between a line and all other lines, this method calculates the sum of the cosine similarities between a line i and the characteristic signals of all other feeders, denoted as C. i That is, C1 = cos(I1,I2) + cos(I1,I3) + ..., C2 = cos(I2,I1) + cos(I2,I3) + ..., and so on. When there is an element in {C1,C2,C3,...} that is negative, it indicates that a single-pole grounding fault has occurred on the feeder. The smallest element C iThis corresponds to the faulty line i. If there are no negative numbers in the set, it indicates that the fault location is not on the feeder line.

[0088] When the fault location is determined not to be on the feeder, bus fault determination is performed. Characteristic signals from both sides of the tie line are extracted, and the sum of their cosine similarities with the characteristic signals of the feeder on the same side is calculated, i.e., C. B1 With C B2 If C B1 If >0, it indicates a fault in bus BUS1. If C B2 If >0, it indicates a fault in bus BUS2. If C B1 With C B2 If all values ​​are less than 0, it is determined to be a fault in the L0 connection line. Figure 3 The implementation flowchart for steps 2, 3, and 4 can be divided into the control signal injection stage, the fault identification stage, and the fault selection stage.

[0089] The present invention also proposes a DC distribution network fault location system based on the similarity between the detection signal and the cosine similarity. The system realizes DC distribution network fault location based on the method described above.

[0090] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it performs fault location of a DC distribution network based on the method described above, using the similarity between a detection signal and a cosine similarity.

[0091] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, enables fault location in a DC distribution network based on the similarity between a detection signal and a cosine similarity, according to the method described thereon.

[0092] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0093] Example

[0094] Based on the appendix Figure 1 The effectiveness of the fault location method was verified using the flexible DC distribution network system shown in Table 1 (referencing the Shenzhen Flexible DC Distribution Demonstration Project). The system parameters are shown in Table 1.

[0095] Table 1 Parameters of Dual-Ended MMC Flexible DC Distribution Network System

[0096]

[0097] According to Table 1 and Equations (2), (3) and (4), the frequency of the detection signal is 240Hz. Therefore, the center frequency of BPF is 240Hz. The low-end cutoff frequency of BPF is 235Hz and the high-end cutoff frequency is 245Hz.

[0098] Taking the L3 terminal positive grounding fault, BUS2 negative grounding fault, and L0 midpoint positive grounding fault as examples, transition resistors of 10Ω, 300Ω, and 1000Ω were set at the fault points respectively. The fault identification results are shown in Table 2.

[0099] Table 2. Route selection results under different fault conditions.

[0100]

[0101] Regardless of which line experiences a single-pole grounding fault, the fault characteristics are unrelated to the presence or magnitude of transition resistance, and the phase relationship between zero-mode currents is not affected. Therefore, as the transition resistance increases, the feeder fault identification effect remains essentially unchanged, exhibiting good accuracy and resistance to transition resistance.

[0102] In summary, existing detection signal methods directly use zero-mode current for phase discrimination. However, the current components in the fault zero-mode current are complex and the frequency distribution is irregular. This invention separates the detection signal from the original signal through BPF, enhancing the phase difference between characteristic signals and improving the accuracy of fault location. Existing detection signal methods can only identify feeder faults in single-ended MMC flexible DC distribution networks. This method performs transient analysis on the fault network of the most widely used double-ended MMC flexible DC distribution network, and the proposed criteria can identify faults on any feeder, bus, and tie line within the system. Existing cosine similarity methods can only reflect the phase correlation between two lines. This invention calculates the sum of the cosine similarities between a single line and all other lines, which can more intuitively show the difference between a certain line and all other lines. Compared with other fault location methods, this method is not affected by transition resistance, has good stability and universality, and has significant advantages in cases of complex distribution network structures and large transition resistance.

[0103] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0104] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for fault location in DC distribution networks based on the similarity between detection signals and cosine similarity, characterized in that, Includes the following steps: Step 1: Determine the injection strategy for the detection signal based on the topology and equipment parameters of the dual-ended MMC type flexible DC distribution network; Step 2: Collect the positive and negative voltage signals at the MMC outlet to identify the faulty pole; Step 3: Calculate the zero-mode current at the beginning of each line and extract the characteristic signal; Step 4: Combine the fault route selection criteria and extract fault features based on cosine similarity to complete the fault route selection; The specific method for fault identification in step 2 is as follows: The startup criteria are set so that when a single-pole fault is detected in the system, the MMC additional control strategy is activated to inject a probe signal into the system. After a single-pole ground fault, the DC-side fault pole voltage rapidly drops to 0, while the non-fault pole voltage rises to twice its rated value under the control of the MMC constant DC voltage. For a dual-terminal MMC DC system, the absolute values ​​of the zero-mode voltages on both MMC DC sides are used as the start-up criterion. In the formula, U P1 and U N1 These are the positive and negative voltages on the DC side of MMC1, respectively; U P2 and U N2 These are the positive and negative voltages on the DC side of MMC2, respectively; U dcB This is the rated voltage of the DC system; To avoid transient interference from the environment or internal system, the criterion is activated when all three consecutive sampling points are true. Then, the fault pole is identified using a fault pole selection criterion, which is as follows: The specific method for feature signal extraction in step 3 is as follows: Taking the injected probe signal as the research object, the proposed line selection criterion is applicable to the probe signal. After collecting the positive and negative currents at the beginning of the line, the zero-mode current is calculated first. The frequency of the probe signal is used as the center frequency of the filter. The cutoff frequency of the BPF is calculated based on the center frequency. At the same time, the filter bandwidth is reduced for signal extraction. The method for extracting fault features based on cosine similarity in step 4 is as follows: First, faulty feeders are identified by calculating the sum of cosine similarities between the characteristic signals of each feeder and other feeders. If the value is negative, it indicates that a single-pole grounding fault has occurred on the line. If there is no negative value, the fault location is not on the feeder. When the fault location is determined not to be on the feeder, bus fault determination is performed. The sum of the cosine similarity of the detection signals at both ends of the tie line and the feeder on the same side is calculated. If the value is positive, it indicates that the bus on that side has a fault. If there is no positive value, it is determined to be a tie line fault.

2. The DC distribution network fault location method based on the similarity between the detection signal and cosine as described in claim 1, characterized in that, The converter station control strategy in step 1 is as follows: The detection signal injection system is a dual-terminal MMC type medium-voltage flexible DC distribution network. The MMC1 converter station adopts constant DC voltage control and constant AC voltage control, i.e., U dc / U ac Control; The MMC2 converter station adopts constant active power control and constant reactive power control, i.e., P / Q control; the modulation method is nearest-level approximation modulation; The DC distribution network topology in step 1 is as follows: The converter station outlet is a DC bus for connecting to the power supply feeder. The two buses are connected by a tie line, which is a cable line. The load types include DC loads, AC loads and distributed photovoltaics.

3. The DC distribution network fault location method based on the similarity between the detection signal and cosine as described in claim 1, characterized in that, The detection signal injection strategy in step 1 is as follows: For MMC converters, sinusoidal signals introduce frequency and phase characteristics to fault analysis, therefore the type of detection signal is a sinusoidal signal; In terms of control strategy, the sinusoidal detection signal is superimposed with the MMC modulation wave, requiring no additional signal attachment device. The control strategy formula is as follows: In the formula, u comd_abc For the output voltage of the circulating current suppression controller, u ctrl To detect the signal, u abc_ref For AC voltage reference value, u abc_P_ref u abc_N_ref These are the voltage modulation waves of the positive and negative bridge arms of the MMC, respectively, U dc This is the DC side voltage; The frequency of the probe signal is much smaller than the switching frequency of the MMC submodule. The expression for the open-loop time constant T of the MMC is: In the formula, L is the fundamental frequency of the AC system. arm For the MMC bridge arm reactance, C, n, R L These are the submodule capacitor, number of submodules, and equivalent loss resistance, respectively. Modulus of unit line-to-ground impedance for: In the formula, R, L, and C are the unit resistance, inductance, and capacitance of the DC system line, respectively; At this time, the line impedance to ground is at its minimum, and the fault response characteristics are most obvious; therefore, the detection signal frequency is... for: 。 4. The DC distribution network fault location method based on the similarity between the detection signal and cosine as described in claim 1, characterized in that, The fault selection criterion in step 4 is: a) When a single-pole grounding fault occurs in a feeder, the characteristic signal extracted from the faulty feeder is out of phase with the characteristic signal of the non-faulty line. When there is a bus fault or a tie line fault, the characteristic signals of each feeder are in the same phase. b) When a busbar fault occurs, the characteristic signal of the feeder connected to the faulty busbar is in phase with the characteristic signal of the tie line on the busbar side, while the characteristic signal of the feeder connected to the non-faulty busbar is out of phase with the characteristic signal of the tie line on the busbar side. c) When a fault occurs in the tie line, the characteristic signal of any feeder connected to any busbar is out of phase with the characteristic signal of the tie line on that busbar side.

5. A DC distribution network fault location system based on the similarity between a detection signal and a cosine similarity, characterized in that, The system is based on the method described in any one of claims 1-4 to realize DC distribution network fault location based on the similarity between the detection signal and the cosine.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements DC distribution network fault location based on the similarity between the detection signal and the cosine similarity according to any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it realizes DC distribution network fault location based on the similarity between the detection signal and the cosine of the method according to any one of claims 1-4.

Images