Flexible DC power grid line protection method based on modal voltage adaptive threshold

Abstract

This invention discloses a flexible DC power grid line protection method based on a modal voltage adaptive threshold. Specifically, it involves: integrating the 1-mode and 0-mode fault components of the DC line voltage, and determining the upper limit of the integration time using a mathematical morphological gradient; constructing an adaptive threshold, where the product of the absolute value of the difference between the two modal voltage integrals and the adaptive setting coefficient is used as an adjustment term to change the threshold; constructing a parameter setting method for the adaptive threshold criterion, where the preset threshold is the 1-mode voltage integral of a metallic ground fault on the far-end DC bus, and the adaptive coefficient is the preset threshold divided by the absolute value of the difference between the two modal voltage integrals when there is a high-resistance fault at the near end; and constructing a fault identification criterion, where if the 1-mode voltage integral is greater than the adaptive threshold and is a positive fault, the line is identified as a faulty line. This invention can accurately identify lines experiencing polar faults, improves the sensitivity of the main protection to high-resistance faults, can be applied under different sampling frequencies and weak boundary conditions, and has a certain degree of noise resistance.

Description

Technical Field

[0001] This invention belongs to the field of DC power grid relay protection, specifically relating to a flexible DC power grid line protection method based on mode voltage adaptive threshold. Background Technology

[0002] Large-scale development of new energy sources is one way to achieve decarbonization goals. High-voltage direct current (HVDC) transmission based on modular multilevel converters can achieve flexible grid connection and long-distance transmission of new energy sources. True bipolar HVDC transmission systems using overhead lines can transmit energy at high capacity and have flexible operating modes after a fault, thus attracting widespread attention. To ensure the continuity and reliability of power supply, DC grids require the main protection methods of DC line relay protection systems to identify faults and issue trip commands within 3ms. Existing main protection methods for true bipolar flexible DC grid lines are susceptible to excessive resistance; when a high-resistance fault occurs, the protection method cannot correctly detect the fault. When smaller current-limiting reactors are configured on both sides of the DC line, the sensitivity of these main protection methods to grounding faults via transition resistance will be further reduced, making it impossible to accurately detect the faulty line. Summary of the Invention

[0003] To address the shortcomings of existing DC power grid main protection methods, this invention provides a flexible DC power grid line protection method based on modal voltage adaptive threshold.

[0004] The present invention provides a flexible DC power grid line protection method based on modal voltage adaptive threshold, comprising the following steps:

[0005] Step 1: Obtain n sampled values ​​of the positive and negative voltages on the line side of the line reactor. Calculate the 1-mode voltage fault component and the 0-mode voltage fault component based on the sampled values. Use the voltage gradient algorithm as the starting criterion for the protection method. Select the mode voltage fault components 1 second before and 1 second after the activation time of the starting criterion using the data window. Calculate the mathematical morphological gradients of the two modes of voltage respectively. The first minimum point of the mathematical morphological gradient after the starting time is the upper limit of the time for mode voltage integration.

[0006] Step 2: Construct an adaptive threshold: Determine the upper limit of integration time using the constructed mathematical morphological gradient. Use the absolute value of the difference between the 1-mode voltage integral and the 0-mode voltage integral and the product of the adaptive coefficient as the adjustment term. The adaptive threshold is the sum of the adjustment term and the preset threshold.

[0007] Step 3: Construct a parameter tuning method for the adaptive threshold criterion:

[0008] The preset threshold is set to the integral of the first-mode voltage for a metallic ground fault on the far-end DC bus, and the adaptive coefficient is set to the absolute value of the difference between the integral of the first-mode voltage and the integral of the zero-mode voltage when the near-end high-resistance fault is divided by the preset threshold.

[0009] Step 4: Construct fault line criteria for a flexible DC power grid using adaptive thresholds:

[0010] After the protection algorithm is started, if the integral of the 1-mode voltage is greater than the adaptive threshold and is identified as a positive fault by the direction criterion, then the line is a faulty line and a trip command is sent to the corresponding faulty pole.

[0011] Furthermore, step 1 specifically involves:

[0012] and These represent the 1-mode and 0-mode voltage fault components at the line protection measurement point, respectively. The protection method uses a voltage gradient algorithm as the activation criterion.

[0013]

[0014] Where k represents the current sampling time; To trigger the threshold for the criterion; Voltage gradient The absolute value of.

[0015] voltage gradient for:

[0016]

[0017] Where kj is the sampling time; when the start-up criterion is met, the modal voltage fault components are extracted 1 second before and 1 second after the start-up time. and Calculate mathematical morphological gradient

[0018] Calculate according to the following formula:

[0019]

[0020] in, and These are the dilation operator and the erosion operator, respectively; b is the structuring element; f 1(0) (k) and g 1(0) (k) is The two input signals, with superscripts 1 and 0, represent the 1-mode and 0-mode components, respectively.

[0021] Operator and The operation rules are as follows:

[0022]

[0023]

[0024] Where max and min are the maximum and minimum values, respectively; h is the sampled signal, i.e., the 1-mode or 0-mode voltage fault component; and y is the element number in b.

[0025] f 1(0) (k) and g 1(0) The construction method of (k) is as follows:

[0026]

[0027]

[0028] Among them, the operator and · are the opening and closing operators, respectively.

[0029] Operator The operation method for s and · is as follows:

[0030]

[0031]

[0032] After the initiation criterion is established, The time corresponding to the first minimum point is taken as the upper limit of the time for the integration of the corresponding mode voltage.

[0033] Furthermore, step 2 specifically involves:

[0034] Adaptive threshold u set The construction method is as follows:

[0035]

[0036] Among them, u set1 η and η are the preset threshold and the adaptive coefficient, respectively; and The integral of the absolute values ​​of the mode 1 and mode 0 voltages is calculated by the following formula:

[0037]

[0038]

[0039] in, and The upper time limits for the 1-mode voltage integral and the 0-mode voltage integral are respectively determined using the mathematical morphological gradient method; t is the time variable, and dt represents the time derivative.

[0040] Furthermore, step 3 specifically involves:

[0041] Preset threshold u set1 Tuning based on the maximum integral value of faults outside the positive zone:

[0042]

[0043] in, The fault component of the mode 1 voltage during a positive external fault; when a metallic ground fault occurs on the remote DC bus, the mode 1 voltage reaches its maximum value within a defined interval.

[0044]

[0045] in, This is the mode 1 voltage fault component when a metallic grounding fault occurs on the remote DC bus.

[0046] The adaptive coefficient η is based on the preset threshold u set1 The absolute value of the difference between the integral of the 1-mode voltage and the integral of the 0-mode voltage is used for tuning; the tuning rule for η is: when a high-impedance fault occurs in the region, the adaptive threshold u... set It should automatically adjust to 0, that is, the minimum absolute value of the voltage integral difference should be used in the setting;

[0047]

[0048] Among them, the minimum value of voltage integral difference The setting is based on a near-end high-resistance fault.

[0049] Furthermore, step 4 specifically involves:

[0050] After the protection algorithm is activated, if the integral of the 1-mode voltage is greater than the adaptive threshold and the fault is identified as a forward fault based on the direction criterion, then the line is a faulty line, and a trip command is sent to the corresponding faulty pole; the fault direction criterion is:

[0051]

[0052] Among them, S d The voltage integral of the boundary reactor is calculated by the following formula:

[0053]

[0054] in, The first-order voltage fault component at the busbar; T d T is the upper limit of the integration time for the direction criterion. d =0.5ms.

[0055] The adaptive threshold criterion for modal voltage is:

[0056]

[0057] Among them, K rel K is the reliability coefficient. rel=1.2; When the adaptive threshold criterion is met, a ground fault is determined to have occurred on the line, and a trip command is sent to the corresponding faulty pole according to the following pole selection criterion:

[0058]

[0059] Where Δ is the pole selection criterion threshold, which is the fault pole when a high-resistance fault is correctly detected, Δ=10kV·ms; S p The integral of the zero-mode voltage fault component is calculated using the following formula:

[0060]

[0061] The beneficial technical effects of this invention are as follows:

[0062] 1) This invention utilizes the modal voltage fault components at one end of the line to design a main protection method for DC lines. The integrated value of the modal voltage fault component is compared with an adaptive threshold to detect the fault. This invention eliminates the need for laying additional communication optical cables, and the fault detection speed is unaffected by line length, meeting the speed requirements of DC power grids for main protection methods.

[0063] 2) Within the integration interval defined by this invention, the integral of the 1-mode voltage for faults within the zone is less than the integral of the 0-mode voltage, while the integral of the 1-mode voltage for faults outside the zone is greater than the integral of the 0-mode voltage. This characteristic is not affected by excessive resistance. This invention incorporates the difference between the integral of the 1-mode voltage and the integral of the 0-mode voltage into the threshold, dynamically changing the threshold and improving the sensitivity of the main protection method to high-resistance faults.

[0064] 3) The line protection method of the present invention is not affected by fault location, AC side fault, or power flow transfer, and can accurately identify polar faults even when the transition resistance is 600 ohms.

[0065] 4) The line protection method of the present invention can operate under noise interference and can still identify high-resistance faults under weak boundary conditions with small-specification current-limiting reactors. Attached Figure Description

[0066] Figure 1 This is a flowchart of the flexible DC power grid line protection method based on modal voltage adaptive threshold according to the present invention.

[0067] Figure 2 This is a four-terminal DC power grid test system.

[0068] Figure 3 It is a composite modulus network for faults within the region.

[0069] Figure 4 This is a graph showing the integral variation of the modal voltage fault component.

[0070] Figure 5 This is a diagram of DC line towers.

[0071] Figure 6 The simulation results are shown for a positive ground fault occurring at l1 380km.

[0072] Figure 7 The simulation results are shown in the figure.

[0073] Figure 8 The simulation results are shown in the figure.

[0074] Figure 9 The simulation results of the current reversal are shown in the figure.

[0075] Figure 10 The simulation results are shown for 200mL and 100mL current-limiting reactors.

[0076] Figure 11 The simulation results are shown for 80mL and 60mL current-limiting reactors.

[0077] Figure 12 The simulation results are shown in the figure with 40dB noise interference.

[0078] Figure 13 The simulation results are shown in the figure for different sampling frequencies. Detailed Implementation

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

[0080] The flowchart of a flexible DC power grid line protection method based on modal voltage adaptive threshold according to the present invention is as follows: Figure 1 As shown, firstly, n sampled values ​​of the positive and negative voltages on the line side of the line reactor are obtained. Based on these sampled values, the 1-mode voltage fault component and the 0-mode voltage fault component are calculated. A voltage gradient algorithm is used as the activation criterion for the protection method. Using a data window, the mode voltage fault components are selected one second before and one second after the activation time of the activation criterion. The mathematical morphological gradients of the two mode voltages are calculated respectively. The first minimum point of the mathematical morphological gradient after the activation time is taken as the upper limit of the mode voltage integration time. The integration interval of the 1-mode voltage and the 0-mode voltage is the upper limit of the integration time from the activation time to the corresponding mode voltage. The absolute value of the difference between the 1-mode voltage integral and the 0-mode voltage integral is multiplied by the adaptive setting coefficient as the adjustment term. The adaptive threshold is the sum of the adjustment term and the preset threshold. The parameters in the adaptive threshold are obtained from the proposed setting scheme. If the 1-mode voltage integral is greater than the adaptive threshold and is identified as a positive fault by the direction criterion, then the line is a faulty line, and a trip command is sent to the corresponding faulty pole. The specific method is as follows:

[0081] 1) For example Figure 2The diagram shows a four-terminal DC grid, where MMC1 to MMC4 represent four converter stations; l1, l2, l3, and l3 represent four DC lines; B1, B2, B3, and B4 represent four DC buses, and L... dc The current-limiting reactors installed on both sides of the DC line are generally larger than 100 mL, and 60 mL-100 mL under weak boundary conditions. In the example of this invention, the reactor is 150 mH. Furthermore, 60 mL and 80 mL reactors are used to verify the high-resistance fault identification capability of the proposed method under weak boundary conditions; P ab For the protection device and DC circuit breaker, the subscripts a and b are the local converter station and the counterpart converter station numbers; M, N and L are the measurement points; F1 to F6 are the fault points set in the example of this invention.

[0082] With line l1 P 12 The circuit protection method of the present invention will be illustrated using an example. Figure 3 The figure shows a composite modulus network when a polar fault occurs on the line, where all electrical quantities are voltage fault components; M 1 and M 0 These represent the corresponding points of measurement point M in the composite modulus network. and These represent the corresponding points of DC bus B1 in the composite modulus network; L eq For the converter's equivalent reactance, and These represent the impedances of mode 1 and mode 0, respectively. and ε(t) represents the initial voltage traveling wave amplitudes of mode 1 and mode 0 at the fault point, respectively, and ε(t) represents the step function. R is the modulo-1 current at the fault point. f For the transition resistance, U dc This represents the rated voltage of the DC power grid.

[0083] The measurement points collect n sampled values ​​of the positive and negative voltages on the line side of the line reactor, and calculate the modulus voltage fault component based on the sampled values. and 0-mode voltage fault component The voltage gradient algorithm is used as the activation criterion for the protection method:

[0084]

[0085] Where k represents the current sampling time; To trigger the threshold for the criterion; Voltage gradient The absolute value of.

[0086] voltage gradient for:

[0087]

[0088] Where kj represents the sampling time. When the start-up criterion is met, the modal voltage fault components are extracted 1 second before and 1 second after the start-up time. and Calculate mathematical morphological gradient

[0089] It can be calculated using the following formula:

[0090]

[0091] in, and These are the dilation operator and the erosion operator, respectively; b is the structuring element; f 1(0) (k) and g 1(0) (k) is The two input signals, with superscripts 1 and 0, represent the 1-mode and 0-mode components, respectively.

[0092] Operator and The operation rules are as follows:

[0093]

[0094]

[0095] Where max and min are the maximum and minimum values, respectively; h is the sampled signal, i.e., the 1-mode or 0-mode voltage fault component; and y is the element number in b.

[0096] f 1(0) (k) and g 1(0) The construction method of (k) is as follows:

[0097]

[0098]

[0099] in, and · are the opening and closing operators, respectively.

[0100] Operator The operation method for s and · is as follows:

[0101]

[0102]

[0103] After the initiation criterion is established, The time corresponding to the first minimum point is taken as the upper limit of the time for the integration of the corresponding mode voltage.

[0104] 2) Construct an adaptive threshold, use the constructed mathematical morphological gradient to determine the upper limit of integration time, and use the product of the absolute value of the difference between the 1-mode voltage integral and the 0-mode voltage integral and the set adaptive coefficient as the adjustment term. The adaptive threshold is the sum of the adjustment term and the preset threshold.

[0105] Adaptive threshold u set The construction method is as follows:

[0106]

[0107] Among them, u set1 η and η are the preset threshold and the adaptive coefficient, respectively; and The integral of the absolute values ​​of the mode 1 and mode 0 voltages can be calculated by the following formula:

[0108]

[0109]

[0110] in, and The upper time limits for the 1-mode voltage integral and the 0-mode voltage integral are respectively determined by the mathematical morphological gradient method according to claim 2; t is a time variable, and dt represents the time derivative.

[0111] 3) Parameter tuning method for adaptive threshold criterion: The preset threshold is tuned to the 1-mode voltage integral of the metallic grounding fault of the far-end DC bus, and the adaptive coefficient is tuned to the absolute value of the difference between the 1-mode voltage integral and the 0-mode voltage integral when the near-end high-resistance fault is divided by the preset threshold.

[0112] Preset threshold u set1 The tuning is based on the maximum integral value of faults outside the positive zone.

[0113]

[0114] in, This represents the mode 1 voltage fault component during a positive out-of-range fault. When a metallic ground fault occurs on the remote DC bus, the mode 1 voltage reaches its maximum value within a defined interval.

[0115]

[0116] in, This is the mode 1 voltage fault component when a metallic grounding fault occurs on the remote DC bus.

[0117] The adaptive coefficient η is based on the preset threshold u set1The tuning is based on the difference between the integral of the 1-mode voltage and the integral of the 0-mode voltage. The tuning rule for η is: when a high-impedance fault occurs within the region, the adaptive threshold u... set It should be automatically adjusted to 0, meaning that the minimum absolute value of the voltage integral difference should be used in the setting.

[0118]

[0119] During an in-zone fault, the 1-mode voltage and 0 mode voltage The integral analytical expression is as follows:

[0120]

[0121]

[0122] in, The upper limit of the integral of the modal voltage, i.e., in step 2 and and τ is the corresponding modal voltage integral function; c The dispersion coefficient of a voltage traveling wave decreases as the fault point approaches the measurement point; K c The attenuation coefficient of the voltage traveling wave, with superscripts 1 and 0 indicating the 1-mode component and the 0-mode component, respectively. and It can be represented as:

[0123]

[0124]

[0125]

[0126] Known Modal voltage integral function f(τ) under different parameters c Z c The image is attached. Figure 4 As shown. For remote faults, and Approximately 10 -4 Order of magnitude. (From the appendix) Figure 4 It can be seen that, within the corresponding order of magnitude of the parameters, Significantly smaller than For near-end faults and The order of magnitude is less than 10 -4 , and It can be simplified to:

[0127]

[0128]

[0129] As can be seen from monotonicity, At the same time, by the appendix Figure 4 It can be seen that at τ c When it is small, f(τ) c Z c The change is minimal. Therefore, during a near-end fault... and The difference is smaller for remote faults, while the minimum voltage integral difference occurs for near-end faults. The setting is based on a near-end high-resistance fault.

[0130] 4) Fault line criterion for flexible DC power grid based on adaptive threshold:

[0131] After the protection algorithm is activated, if the integral of the 1-mode voltage is greater than the adaptive threshold and the fault is identified as a forward fault based on the direction criterion, then the line is a faulty line, and a trip command is sent to the corresponding faulty pole. The fault direction criterion is:

[0132]

[0133] Among them, S d The voltage integral of the boundary reactor can be calculated by the following formula:

[0134]

[0135] in, The first-order voltage fault component at the busbar; T d As the upper limit of the integration time for the direction criterion, T in this invention d =0.5ms. The adaptive threshold criterion for modal voltage is:

[0136]

[0137] Among them, K rel For the reliability coefficient, K of the present invention rel =1.2. When the adaptive threshold criterion is met, a ground fault is determined to have occurred on the line, and a trip command is sent to the corresponding faulty pole according to the following pole selection criterion:

[0138]

[0139] Where Δ is the pole selection criterion threshold, which is the fault pole when a high-resistance fault is correctly detected. In this invention, Δ = 10 kV·ms; S p The integral of the zero-mode voltage fault component can be calculated using the following formula:

[0140]

[0141] Three factors affect the detection speed of the proposed solution: data window length, calculation time of the mathematical morphological gradient, and system signal transmission delay. The data window length is 2ms, requiring data acquisition 1ms after the fault occurs. At a 50kHz sampling rate, the data window contains 100 sampling points. When the structuring element length is 3, the operator... The requirement is 100 × 3 = 300 addition operations, using the operator... It also requires 300 addition operations, therefore the mathematical morphological gradient of claim 2 requires 3100 addition operations. Existing digital signal processing chips can achieve a processing speed of 1GHz, therefore the calculation delay of the mathematical morphological gradient is less than 0.01ms. The system signal transmission delay is generally less than 0.3ms. In summary, the detection speed of the proposed solution is approximately 1 + 0.01 + 0.3 = 1.31ms. This invention can identify faulty lines within 1.4ms, meeting the requirements of DC power grids for the fast operation of main line protection.

[0142] Example:

[0143] Power grid topology diagram as follows Figure 2 As shown, the electromagnetic transient software PSCAD / EMTDC was used to build... Figure 2 The experimental data collected from the four-terminal DC power grid is shown.

[0144] The MMC-HVDC power grid adopts a true bipolar connection structure, and the rated voltage of the power grid U dc The voltage is ±500kV. Each converter station adopts a half-bridge submodule, and the DC line adopts a frequency-dependent model. Its parameters and tower structure are shown in Table 2 and Appendix. Figure 5 As shown in Table 1, converter station MMC1 uses constant DC grid voltage and constant reactive power control, while converter stations MMC2-MMC4 use constant active power and constant reactive power control. The main parameters of the test system are shown in Table 1. The system sampling frequency f... N =50kHz, time window length T n =2ms, extracting the modal voltage fault components 1 second before and 1 second after the start-up time. and All failure times are set to 2.5 seconds; all examples in this embodiment are attached. Figure 2 Line L1 protection device P 12 To verify the effectiveness of this invention, M represents P. 12 The measurement points. Based on the tuning method constructed in step three and considering a transition resistance of 600Ω, the preset threshold in the criterion can be obtained: u set1 =146.39kV·ms; Where the reliability coefficient K η =0.95; η =41.83.

[0145] Table 1 Parameters of a Four-Terminal DC Grid

[0146]

[0147] Table 2 DC line parameters

[0148]

[0149] Calculation example 1:

[0150] To verify the selectivity of the line protection method of this invention, positive metallic ground faults (PGFs) were respectively set at L1 (fault locations at distances of 0km, 60km, 120km, 250km, 380km, 440km, and 500km from B1), a remote DC bus fault (F3), and an adjacent line beginning fault (10km from B2) to verify the selectivity of the method for faults within the area. A three-phase fault (AC2) was set on the AC side of converter station MMC2 to test the impact of AC side faults on this invention. The identification results of this invention are shown in Table 3.

[0151] Table 3 Fault Identification Results

[0152]

[0153] As shown in Table 3, for positive faults at different locations on line l1, if the integral value of the first-mode voltage is greater than the corresponding adaptive threshold, the method will send a trip command to the protection device. For polar faults on the remote DC bus, adjacent lines, and three-phase AC sides, if the integral value of the first-mode voltage is less than the corresponding adaptive threshold, the method will not send a trip command to the protection device. The method of this invention can identify polar faults at various locations on the line and has good selectivity.

[0154] Appendix Figure 6 and attached Figure 7 The waveforms for the mode voltage, mathematical morphological gradient, mode voltage integral, and adaptive threshold of the PGF generated at l1 380km and the PGF generated at the remote DC bus are shown in the appendix. Figure 6 and attached Figure 7 It can be seen that the mathematical morphological gradient method constructed in this invention can select the upper limit of integration based on the waveform; within the integration interval constructed in this invention, the 1-mode voltage integral of the fault within the region is less than the 0-mode voltage integral, while the 1-mode voltage integral of the fault outside the region is greater than the 0-mode voltage integral.

[0155] Calculation example 2:

[0156] The effectiveness and sensitivity of this invention for high-resistance faults were verified by setting up negative faults (NGF) grounded through a 300Ω transition resistor and positive faults grounded through a 600Ω transition resistor. The identification results are shown in Tables 4 and 5.

[0157] Table 4. Fault identification results via 300Ω grounded negative electrode

[0158]

[0159] Table 5. Fault identification results via a 600Ω grounded positive electrode.

[0160]

[0161] As shown in Tables 4 and 5, for high-resistivity faults within the fault zone, the integral value of the mode voltage is greater than the corresponding adaptive threshold. This invention can accurately and sensitively identify high-resistivity faults within the fault zone. The waveforms of the mode voltage, mathematical morphological gradient, mode voltage integral, and adaptive threshold are shown in the attached figure when the near-end positive terminal within the fault zone is grounded through a 600Ω transition resistor. Figure 8 As shown. (From the appendix) Figure 8 It can be seen that for high-resistance faults near the region, the adaptive threshold remains negative. The threshold tuning method constructed by this invention is correct and effective, and this invention can accurately identify high-resistance faults.

[0162] Calculation example 3:

[0163] The identification results of the protection method should not be affected by the power flow direction. At 2.5s, the active power reference value of MMC2 changes from -1pu to 1pu to test the operating characteristics of this invention during power flow reversal; after the power flow reversal, a positive ground fault is set at l1 250km at 3s to test the identification capability of this invention under different power flow directions. The waveforms of modal voltage, mathematical morphological gradient, modal voltage integral, and adaptive threshold are attached. Figure 9 As shown in the figure, the proposed solution will not malfunction when the current reverses, and can still correctly identify faults in the area in different current directions.

[0164] Calculation example 4:

[0165] To verify the performance of the proposed method under different current-limiting reactors and under weak boundary conditions, the current-limiting reactors were set to 60mL, 80mL, 100mL, and 200mL, respectively. The 80mL and 60mL current-limiting reactors were reserved for testing the performance of the proposed method under weak boundary conditions. In-zone faults were defined as positive terminals grounded via a 300Ω transition resistor, and out-of-zone faults were defined as metallic positive terminal grounding faults on the distant DC bus. The modulo-1 voltage integral and the corresponding adaptive threshold are attached. Figure 10 and attached Figure 11 As shown in the figure, the integral of the modulo-1 voltage for faults within the fault zone is greater than the adaptive threshold, while the integral of the modulo-1 voltage for faults outside the fault zone is less than the adaptive threshold. The performance of the proposed method for detecting high-resistance faults is not affected by boundary conditions; even under weak boundary conditions, the proposed method can still correctly identify high-resistance faults.

[0166] Calculation example 5:

[0167] To verify the performance of the proposed method under noise interference, 40dB of white noise was superimposed on the voltage data. The fault was set to ground via a 300Ω transition resistor, and the fault location was 250km away. The voltage waveform, voltage integral, and adaptive threshold after noise superposition are shown in the attached figure. Figure 12 As shown. The integration method reduces the impact of white noise on the proposed solution, and the proposed method can operate under 40dB noise interference. Table 6 shows the identification results of different points of the negative line grounded through a 300Ω transition resistor under 40dB noise interference.

[0168] Table 6. Identification results of 40dB noise interference

[0169]

[0170] As shown in Table 6, the method proposed in this invention can correctly identify faulty lines under 40dB noise interference.

[0171] Calculation example 6:

[0172] To verify the effectiveness of the proposed method at different sampling frequencies, the sampling frequencies were changed to 20kHz and 100kHz, while other parameters remained unchanged. The fault was set to ground via a 300Ω transition resistor. The experimental results are attached. Figure 13 As shown in the figure, lower sampling frequencies reduce the difference between the line-mode voltage integral and the threshold. Although the sampling frequency affects both the adaptive threshold and the 1-mode voltage integral, the difference is significant enough to accurately detect faults. The proposed method can correctly identify faults within the region at different sampling frequencies.

Claims

1. A method of flexible HVDC grid line protection based on modal voltage adaptive threshold, characterized in that, The method comprises the following steps: Step 1: obtaining n sampling values of positive and negative electrode voltages of a line reactor line side, calculating 1-mode voltage fault component and 0-mode voltage fault component according to the sampling values, and adopting a voltage gradient algorithm as a starting criterion of the protection method; The modal voltage fault components of the 1 second before and after the starting criterion activation time point are selected by using a data window, the mathematical shape gradients of the two modal voltages are calculated respectively, and the first minimum point of the mathematical shape gradient after the starting time point is the upper limit of time of the modal voltage integration; Step 2: constructing an adaptive threshold: determining the integration upper limit time by using the constructed mathematical shape gradient, taking the product of the absolute value of the difference between the 1-mode voltage integration and the 0-mode voltage integration and the adaptive coefficient as an adjustment term, and taking the sum of the adjustment term and a preset threshold as the adaptive threshold; Step 3: parameter setting method of the adaptive threshold criterion: The preset threshold is set as the 1-mode voltage integration of the remote DC bus metallic ground fault, and the adaptive coefficient is set as the preset threshold divided by the absolute value of the difference between the 1-mode voltage integration and the 0-mode voltage integration in the near-end high-resistance fault; Step 4: fault line criterion of the flexible DC power grid using the adaptive threshold: After the protection algorithm is started, if the 1-mode voltage integration is greater than the adaptive threshold and is identified as a positive fault by a direction criterion, the line is a fault line, and a trip command is sent to the corresponding fault pole.

2. A method for flexible HVDC grid line protection based on modal voltage adaptive threshold as claimed in claim 1, wherein, The step 1 is specifically: and are the 1-mode and 0-mode voltage fault components at the line protection measurement point, respectively, and the protection method uses a voltage gradient algorithm as the start criterion: wherein k denotes the current sampling instant; is a start criterion threshold value; is the absolute value of the voltage gradient is the absolute value of the voltage gradient is: Where k-j is the sampling time; when the starting criterion is established, the modal voltage fault component of 1 second before and after the starting time is intercepted and Computing the mathematical morphology gradient is computed according to the following formula: where, and are dilation and erosion operators, respectively; b is a structuring element; f 1(0) (k) and g 1(0) (k) are two input signals, superscripts 1 and 0 represent 1-mode and 0-mode components, respectively; operators and operate according to the following rules: where max and min are the maximum and minimum functions, respectively; h is the sampled signal, i.e., the 1-mode or 0-mode voltage fault component; y is the element number in b; f 1(0) (k) and g 1(0) The construction method of (k) is shown as follows: where the operators and • are open and closed operators, respectively; the operators and • are calculated as follows: After the start criterion is established, the corresponding time of the first minimum point is taken as the upper limit of the time of the corresponding modal voltage integral. The corresponding time of the first minimum point is taken as the upper limit of the time of the corresponding modal voltage integral.

3. The method of claim 2, wherein the method further comprises: The step 2 is specifically: Adaptive threshold u set The construction method is: wherein u set1 and η are a preset threshold value and an adaptive coefficient, respectively; and is the integral of the absolute value of the 1-mode and 0-mode voltages, calculated by the following equation: wherein and are the upper time limits of the 1-mode voltage integral and 0-mode voltage integral, respectively, determined according to the mathematical morphological gradient method; t is the time variable and dt represents the time differential.

4. The method of claim 3, wherein the method further comprises: The step 3 is specifically: Pre-set threshold value u set1 Setting according to the maximum integral value of forward external faults: wherein, is the 1-mode voltage fault component for a forward zone external fault; when a metallic ground fault occurs on the remote DC bus, the 1-mode voltage reaches a maximum value when integrated over the determination interval; wherein is the 1-mode voltage fault component for a metallic ground fault of the remote DC bus; The adaptive coefficient η is according to a preset threshold value u set1 And the absolute value of the difference between the 1-mode voltage integral and the 0-mode voltage integral is set; the setting rule of η is that when a high-resistance fault occurs in the area, the adaptive threshold value u set Should be automatically adjusted to 0, that is, the minimum value of the absolute value of the voltage integral difference should be used in the setting; Wherein the minimum value of the voltage integral difference Set by the near-end high impedance fault.

5. A method of flexible HVDC grid line protection based on modal voltage adaptive threshold as claimed in claim 4, wherein, The step 4 is specifically: After the protection algorithm is started, if the 1-mode voltage integration is greater than the adaptive threshold and is identified as a positive fault by a direction criterion, the line is a fault line, and a trip command is sent to the corresponding fault pole; and the fault direction criterion is: where S d is the voltage integral of the boundary reactor, calculated by the following equation: wherein, is the 1-mode voltage fault component at the bus; T d is the directional criterion integration time upper limit, T d = 0.5 ms; the adaptive threshold criterion for the modal voltage is: Wherein, K rel is the reliability coefficient, K rel = 1.2; when the adaptive threshold criterion is established, it is determined that a ground fault occurs on the line, and a trip command is sent to the corresponding fault pole according to the following pole selection criterion: Wherein, Δ is the selected pole criterion threshold, is the fault pole when the high resistance fault is detected correctly, Δ = 10 kV·ms; S p is the 0-mode voltage fault component integral, which is calculated by the following formula:

Images