A flexible dc power grid protection method based on inner product of fault vectors

Abstract

The application discloses a kind of flexible DC power grid protection methods based on fault vector inner product, belong to the field of power system relay protection, this method includes: the line mode fault voltage, current vector of fixed data window length is collected;The transient inner product of both is calculated, and the amplitude thereof is used to constitute single-ended quantity main protection criterion to identify low resistance fault;The inner angle thereof is used to constitute direction criterion, and is matched with the information of opposite end to constitute longitudinal back-up protection to identify high resistance fault and no boundary element scene.The application forms "multi-section" protection by the comprehensive utilization of voltage and current transient inner product, expands the fault characteristic difference, does not depend on singular point information, has the advantages of threshold clear, strong anti-interference ability, high reliability, and is suitable for extremely high flexible DC power grid.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of new power system relay protection, and particularly relates to a flexible DC power grid protection method based on fault vector inner product. BACKGROUND

[0002] In the construction of a new type of power system mainly composed of wind power, photovoltaic and other new energy sources, voltage source converter HVDC (VSC-HVDC), also known as flexible DC (flexible DC) power transmission in China, is an important technical means to solve the large-scale grid connection of new energy with random and fluctuating characteristics. Compared with the traditional point-to-point line-commuted converter HVDC (LCC-HVDC), the flexible DC power transmission does not have commutation failure, has fewer harmonic components, has independent control of active and reactive power, and has DC networking capability, and can suppress random fluctuations through time and space complementation in a wide range. After a fault occurs in the DC side of the flexible DC system, the energy storage capacity of the sub-modules in the VSC is rapidly discharged through the low-impedance fault loop, and within a few milliseconds, the fault current may exceed the rated capacity of the turn-off semiconductor device and the DC circuit breaker (DCCB). Therefore, it is generally believed that the flexible DC system has extremely high speed requirements for fault protection (2-3 ms), and it is a challenge to accurately and reliably obtain fault characteristics within a limited data length and design protection based on single-ended transient quantities.

[0003] In the traditional LCC-HVDC, the DC side filter capacitor and the line protection boundary composed of reactance are used as the protection boundary. In the flexible DC power grid, the current limiting reactor (CLR) connected in series at both ends of the line can be used as the protection boundary. The existing methods mainly use the transient characteristics of the fault traveling wave after passing through the CLR to design the protection, which can be divided into two categories: 1) using the characteristics that the change of the fault traveling wave is more gentle after passing through the CLR, such as the first document “Multi-terminal MMC-HVDC fault fast detection scheme based on new type of directional criterion” (Shuyong Yu et al., Proceedings of the CSEE, Vol. 41, No. 24, 2021) which uses the voltage traveling wave change rate to distinguish the fault area, and the second document “Research on DC arc fault characteristics and detection method of photovoltaic system” (Xinzhou Dong et al., Power System Technology, Vol. 42, No. 6, 2018) which uses the wavelet transform modulus maximum value to design the protection scheme; 2) using the characteristics that the CLR blocks high-frequency components and passes low-frequency components, such as the third document “Research on DC arc fault characteristics and detection method of photovoltaic system” (Bin Li et al., Proceedings of the CSEE, Vol. 36, No. 21, 2016) which uses the transient energy identified by the wavelet transform detail coefficient to distinguish the fault area, and in addition, the fault area can also be distinguished by the characteristic frequency band, energy entropy and transient energy of the variational mode decomposition. However, after the fault occurs at the DC side, the fault location and transition resistance vary greatly, resulting in significant differences in transient characteristics. The methods that simply use the characteristics of the CLR element have problems such as difficult coordination of the protection “four properties”, complex threshold setting, and dependence on the line protection boundary.

[0004] Therefore, the existing methods have the following limitations: 1) difficult coordination of the protection “four properties”: the fault location and transition resistance vary greatly, resulting in significant differences in transient characteristics, and it is difficult to coordinate the selectivity, speed, sensitivity and reliability of the protection; 2) complex threshold setting: the threshold of the methods that depend on the characteristics of the protection boundary is greatly affected by the line parameters and network structure, and the setting process is complex; 3) dependence on the boundary element: the protection performance is highly dependent on the CLR composed of the boundary, and in the case of shared installation of the reactor, the protection may fail; 4) weak anti-interference ability: some methods depend on the singular point information in the fault signal, which is easily disturbed by noise. Therefore, there is an urgent need for a new flexible DC power grid line protection principle that does not depend on a single boundary characteristic, has clear threshold, strong anti-interference ability and can adapt to complex fault conditions. SUMMARY

[0005] The purpose of the present application is to provide a flexible DC power grid protection method based on fault vector inner product, which does not depend on a single boundary characteristic, has clear threshold, strong anti-interference ability and can adapt to complex fault conditions.

[0006] Technical scheme: The flexible DC power grid protection method based on fault vector inner product of the present application comprises:

[0007] Data preprocessing: measuring the positive line voltage, negative line voltage, positive line current and negative line current of the line where the signal measurement point is located in the flexible HVDC power grid in real time; calculating the variation of the positive line voltage, negative line voltage, positive line current and negative line current by using the steady-state value; decoupling the variation of the positive line voltage and negative line voltage by using the line mode decoupling matrix to obtain the line mode fault voltage component and the zero mode fault voltage component; decoupling the variation of the positive line current and negative line current by using the line mode decoupling matrix to obtain the line mode fault current component and the zero mode fault current component;

[0008] Starting criterion: judging whether the line mode fault voltage component meets the preset judgment condition; if the line mode fault voltage component meets the preset judgment condition, recording the sampling time Corresponding time , and continuously sampling to , wherein, is the sampling data length, the line mode fault voltage vector and the line mode fault current vector are obtained according to the line mode fault voltage component and the line mode fault current component of a sampling length, and the fault pole criterion is entered;

[0009] Fault pole criterion: judging the fault type by using the cumulative amount of the zero mode fault voltage component, and the fault type includes positive ground fault, negative ground fault and pole-to-pole short circuit fault; if the fault type is single-pole ground fault or double-pole ground fault, setting different threshold values of the main protection criterion for the single-pole ground fault and the double-pole ground fault, and starting the ultra-fast main protection;

[0010] Ultra-fast main protection criterion: calculating the transient inner product of the line mode fault voltage vector and the line mode fault current vector ; comparing with the threshold value of the main protection criterion; if is less than the threshold value of the main protection criterion, an intra-zone fault occurs, the main protection device acts, otherwise, the longitudinal directional backup protection is started;

[0011] Longitudinal directional backup protection criterion: substituting the angle between the line mode fault voltage vector and the line mode fault current vector into the preset fault direction criterion of the two ends of the longitudinal line, so as to judge the fault direction; if the positive direction fault occurs at the two ends of the longitudinal line, judging whether the intra-zone high resistance fault occurs by the fault label information and time difference of the two ends of the longitudinal line; if the intra-zone high resistance fault occurs, the backup protection device acts.

[0012] Further, the judgment line mode fault voltage component whether satisfies preset judgment condition, wherein, the expression of judgment condition is as follows:

[0013] ;

[0014] Wherein, is the sampling time; is the real-time measurement at the time and after the line mode decoupling matrix decoupling line mode fault voltage component; is the threshold value of starting criterion.

[0015] Further, the fault type is judged by using the cumulative amount of zero mode fault voltage component, and the calculation formula is as follows:

[0016] ;

[0017] Wherein, is the cumulative amount of zero mode fault voltage component; is the real-time measurement at the time and after the line mode decoupling matrix decoupling zero mode fault voltage component; is the threshold value of fault type criterion; N-PTG is negative ground fault; P-PTG is positive ground fault; PTP is pole-to-pole short circuit fault.

[0018] Further, the transient inner product of line mode fault voltage vector and line mode fault current vector is calculated , and the expression is as follows:

[0019] ;

[0020] Wherein, is the sampling data length; is the real-time measurement at the time and after the line mode decoupling matrix decoupling line mode fault voltage component; is the real-time measurement at the time and after the line mode decoupling matrix decoupling line mode fault current component.

[0021] Further, the product of line mode fault voltage vector and line mode fault current vector transfer function has the characteristics of passing low frequency and resisting high frequency, and the transient characteristic difference of low resistance fault in and out of fault area is expanded by .

[0022] Further, the step of comparing the threshold value of the main protection criterion with the main protection criterion is specifically as follows:

[0023] ;

[0024] wherein, is the threshold value of the main protection criterion; if the fault type is single-pole ground fault, if the fault type is double-pole ground fault, , and .

[0025] Further, the step of substituting the angle between the line mode fault voltage vector and the line mode fault current vector into the preset fault direction criterion at both ends of the pilot line to determine the fault direction, wherein the expression of the fault direction criterion is as follows:

[0026] ;

[0027] wherein, is the angle between the line mode fault voltage vector and the line mode fault current vector ; is the first reliability coefficient.

[0028] Further, the angle between the line mode fault voltage vector and the line mode fault current vector is , and the expression is as follows:

[0029] ;

[0030] wherein, is the transient inner product of the line mode fault voltage vector and the line mode fault current vector .

[0031] Further, if the positive direction fault occurs at both ends of the pilot line, the protection device determines whether the in-zone high resistance fault occurs through the fault label information and the time difference at both ends of the pilot line, comprising:

[0032] If the positive direction fault occurs at both ends of the pilot line, the protection device sends the corresponding fault label information and the time at the local end of the line to the opposite end of the line through the communication system, and the opposite end of the line receives the fault label information and the time; the protection device determines whether the in-zone high resistance fault occurs through the fault label information and the time difference at both ends of the pilot line, and the expression is as follows:

[0033] ;

[0034] wherein, is the fault label information received by the line pair end; is the time when the information is received by the line pair end; is the time difference threshold between the two ends of the line after considering various signal delays.

[0035] Further, the time difference threshold between the two ends of the line after considering various signal delays , and the calculation formula is as follows:

[0036] ;

[0037] wherein, is the length of the corresponding protection line; is the second reliability coefficient; is the line mode fault traveling wave speed.

[0038] Beneficial effects: Compared with the prior art, the significant technical effects of the present application are as follows: (1) expanding the characteristic difference: in the super-fast main protection step, the transient inner product is calculated, which is the product characteristic of the voltage and current transfer function, because the transient inner product amplitude can significantly expand the transient characteristic difference between the low-resistance fault in the area and the low-resistance fault outside the area (especially the remote line), thereby improving the sensitivity and reliability of the main protection; (2) not dependent on singular points: continuous sampling to , and is the length of the sampling data, by collecting data in a data window, for subsequent calculation of inner product or inner angle, rather than singular values of a single sampling point, the anti-interference ability and robustness are stronger; (3) natural direction discrimination ability: the polarity of the transient inner product is natural, which contains fault direction information, without the need for additional complex direction discrimination elements; (4) strong anti-noise ability: the fault direction criterion uses the inner angle characteristic, and its mathematical characteristics make the criterion result stably concentrate in the no-fault state in a random noise environment, effectively preventing misoperation; (5) strong adaptability: the proposed fault direction criterion only depends on the topological constraints of the line port, and does not depend on the protection boundary formed by the current limiting reactor, so it is still effective in the scenes of reactor sharing installation; (6) forming a "multi-section" protection: through the cooperation of the inner product amplitude main protection and the inner angle direction backup protection, a coordinated "multi-section" protection strategy is formed, which can reliably cope with various fault conditions from low resistance to high resistance. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is the flowchart of the present application;

[0040] Figure 2A flexible direct current power grid topology structure for the present application;

[0041] Figure 3 A direct current global equivalent circuit diagram after line fault in the present application;

[0042] Figure 4 A direction protection principle based on an internal angle in the present application;

[0043] Figure 5 A protection result diagram of R12 and R21 main protection criterion in the positive pole grounding fault after area of the embodiment of the present application;

[0044] Figure 6 A protection result diagram of R14 and R41 main protection criterion in the positive pole grounding fault after area of the embodiment of the present application;

[0045] Figure 7 A direction protection simulation result diagram of the embodiment of the present application;

[0046] Figure 8 A fault waveform diagram of line mode fault voltage in different positions under the condition of current limiting reactor sharing installation;

[0047] Figure 9 A fault waveform diagram of line mode fault current in different positions under the condition of current limiting reactor sharing installation. DETAILED DESCRIPTION

[0048] The technical solutions of the present application will be described in detail below in combination with specific embodiments and the drawings of the specification.

[0049] A flexible direct current power grid protection method based on fault vector inner product of the present application is suitable for flexible direct current power grids with complex types (ring, double-end, etc.) network structure, different voltage levels (±500kV, ±320kV, ±160kV, etc.), such as Figure 1 The flexible direct current power grid protection method includes the following steps:

[0050] S1, data preprocessing: measuring the positive pole line voltage, negative pole line voltage, positive pole line current and negative pole line current of the line on which the signal measurement point in the flexible direct current power grid is located in real time; calculating the change amount of the positive pole line voltage, negative pole line voltage, positive pole line current and negative pole line current by using the steady-state value; decoupling the change amount of the positive pole line voltage and negative pole line voltage by using the line mode decoupling matrix to obtain the line mode fault voltage component and zero mode fault voltage component; decoupling the change amount of the positive pole line current and negative pole line current by using the line mode decoupling matrix to obtain the line mode fault current component and zero mode fault current component.

[0051] The specific implementation process of step S1 is as follows:

[0052] S1.1, measuring the positive line voltage, the negative line voltage, the positive line current and the negative line current of the line where the signal measurement point is located in the flexible DC power grid in real time; using the steady-state value to calculate the variation of the positive line voltage, the negative line voltage, the positive line current and the negative line current, that is,

[0053]

[0054] wherein, is the variation of the positive or negative line voltage; is the variation of the positive or negative line current; , is the positive or negative voltage and current measured at the kth sampling time, and is the rated value of the positive or negative voltage and current.

[0055] S1.2, decoupling the variations of the positive line voltage and the negative line voltage by using the line mode decoupling matrix to obtain the line mode fault voltage component and the zero mode fault voltage component, that is,

[0056]

[0057] wherein, and are the line mode fault voltage component and the zero mode fault voltage component at the th sampling time, respectively.

[0058] S1.3, decoupling the variations of the positive line current and the negative line current by using the line mode decoupling matrix to obtain the line mode fault current component and the zero mode fault current component, that is,

[0059]

[0060] wherein, and are the line mode fault current component and the zero mode fault current component at the th sampling time, respectively.

[0061] S2, starting criterion: judging whether the line mode fault voltage component satisfies the preset judgment condition; if the line mode fault voltage component satisfies the preset judgment condition, recording the sampling time corresponding to the time and continuously sampling to , wherein, is the sampling data length, the line mode fault voltage vector and the line mode fault current vector are obtained according to the line mode fault voltage component and the line mode fault current component of a sampling length, and the fault pole criterion is entered.

[0062] ​​​​The starting criterion should have sufficient sensitivity to distinguish between normal operation and disturbance. The present scheme uses the smoothed gradient of the sampling points to identify the normal operation and disturbance state. When a single-pole ground fault occurs, the initial voltage amplitude of the fault point is:

[0063] ;

[0064] When is 800 Ω, after smoothing with 3 data points, 20 kV is taken in the present embodiment.

[0065] The starting criterion is used to distinguish between normal operation and disturbance state. According to theoretical analysis, only the voltage of the fault region jumps, so the smoothed gradient of 3 sampling points is used to detect the disturbance on the DC side.

[0066] In the present embodiment, it is judged whether the line mode fault voltage component satisfies the preset judgment condition, wherein the expression of the judgment condition is as follows:

[0067] ;

[0068] wherein, is the sampling time; is the real-time measurement at the time and the line mode fault voltage component decoupled by the line mode decoupling matrix ; is the threshold value of the starting criterion.

[0069] It is worth noting that under a certain sampling rate, the data point length affects the selection of the criterion threshold value, the longer the data point used for calculation, but the HVDC power grid has a high requirement for the speed of protection, so the sampling rate is 100 kHz by comprehensive consideration, 100, i.e. 1 ms data length.

[0070] The typical four-terminal ring HVDC power grid structure is shown in Figure 2 , which is simplified from an actual project. The true bipolar connection is used, and the four converter stations are converter stations 1-4. Each converter station is composed of two modular multilevel converters (MMC). The neutral points between the converter stations are connected to the equipotential surface of the ground. In Figure 1 , the line is composed of overhead transmission lines, and the lengths of the lines are , , and as shown in Figure 2The two ends of each line are configured with DC circuit breakers (DCCB) and current limiting reactors (CLR), the DCCB is used for rapid isolation of the fault area, and the CLR both inhibits the fault current rising speed and provides a protection boundary. In addition, the fault signal is measured at the line side of the CLR, and the signal measurement points R12, R21 and R14, etc. are also the installation positions of each line protection.

[0071] The DC line can be equivalent to the process of the traveling wave along the line by using the Thevenin model, which can be expressed as:

[0072]

[0073] wherein the subscripts and represent the line port and the fault position, the subscript represents the line mode component “1” or the zero mode component “0”, and are the backward and forward voltage traveling waves, and are the fault voltage components of the line port and the fault position, is the fault voltage transfer function, and a first-order inertia link can fit the delay, attenuation and distortion characteristics of the transmission process, is the length of the fault distance R12.

[0074] Each bridge arm of the MMC contains half-bridge sub-modules, and the bridge arm reactance is . After the DC side fault, the MMC can be equivalent to:

[0075]

[0076] wherein: is the half-bridge sub-module capacitor.

[0077] In Figure 2 , when the positive pole ground fault (F1) occurs in the line 12, the additional circuit of the fault can be obtained according to the superposition theorem. By solving the additional circuit state equation of the fault point and using decoupling, the boundary condition can be obtained, so as to establish the line mode-zero mode composite mode global equivalent circuit diagram after the single-pole ground fault, as shown in Figure 3 . At the fault point position, there is a coupling between the line mode network and the zero mode network, and the initial voltage traveling wave of the fault position can be solved, which is a negative step function.

[0078] In Figure 3 ,​​ are the line mode fault voltage components of the line port in the area (R12), the non-fault line port of the DC bus (R14), the DC bus, the remote end of the non-fault line (R41) and the line port of the opposite end in the area (R21) respectively; are the fault current components of R12, R14, R41 and the flow direction zero mode network at the fault location respectively; are the CLR sizes of each converter station respectively; is the line rated voltage, is the transition resistance.

[0079] S3, fault pole criterion: the cumulative amount of the zero mode fault voltage component is used to determine the fault type, the fault type includes positive pole grounding fault, negative pole grounding fault and pole-to-pole short circuit fault; if the fault type is single pole grounding fault or double pole grounding fault, different threshold values of the main protection criterion are set for the single pole grounding fault and the double pole grounding fault, and the ultra-fast main protection is started.

[0080] Different from the prior art of determining the fault area first and then determining the fault type, the application considers that the fault type should be determined first, and then different protection threshold values are set according to the fault type. Since the pole-to-pole fault is a symmetrical fault, no zero mode fault component is generated, so the cumulative amount of the zero mode fault component is used to determine the fault type.

[0081] In the embodiment, the cumulative amount of the zero mode fault voltage component is used to determine the fault type, and the calculation formula is as follows:

[0082]

[0083] wherein, is the cumulative amount of the zero mode fault voltage component; is the real-time measurement at the time and the zero mode fault voltage component decoupled by the line mode decoupling matrix , is accumulated from 0 to ; is the threshold value of the fault type criterion; N-PTG is negative pole grounding fault; P-PTG is positive pole grounding fault; PTP is pole-to-pole short circuit fault. When the pole-to-pole short circuit fault occurs, the amplitude of the initial fault voltage is higher, at this time, the main protection criterion can set a higher protection threshold value.

[0084] ​​​​​​​​​​​It is worth noting that the threshold of the fault type criterion The transfer time difference and other interferences, mainly used to distinguish noise, line mode and zero mode components, are distinguished in the embodiment Take 200 kV.

[0085] S4, ultra-fast main protection criterion: calculate the transient inner product of line mode fault voltage vector And line mode fault current vector ; Compare With the threshold of the main protection criterion; if Less than the threshold of the main protection criterion, an internal fault occurs, the main protection device acts, otherwise, the longitudinal direction backup protection is started.

[0086] The ultra-fast main protection criterion of the application is set based on the transient inner product of line mode fault voltage vector and line mode fault current vector, and the ultra-fast main protection criterion uses the amplitude of the transient inner product to distinguish the fault area. Taking line mode fault component as an example, solve Figure 3 At the fault line port (R12), the product of line mode fault voltage and current transfer function is:

[0087] ;

[0088] Wherein, The product of R12 voltage and current transfer function is, and according to , we can get . The fault equivalent voltage at the fault point after multiple refraction and reflection of the traveling wave on the fault line; The line mode fault voltage component of the internal line port (R12); The line mode fault current component of R12; The line mode wave impedance; The fault voltage transfer function, a first-order inertial link can fit the delay, attenuation and distortion characteristics of the transfer process, The length of the fault distance R12; The reflection coefficient of the line port line mode component, .

[0089] Similarly, the product of the transfer function of the external line near end (R14) is obtained as:

[0090] ;

[0091] Wherein, The product of R14 voltage and current transfer function is, and . The refraction coefficient of the line mode component at the fault point position; the line mode fault voltage component at the non-faulted line end (R14) of the DC bus, is the line mode fault current component at R14. Furthermore, the product of the transfer functions of the line outside the fault zone (R41) is obtained as:

[0092] ;

[0093] wherein, is the product of the voltage and current transfer functions of R41, and has . is the line mode fault voltage component at the non-faulted line end (R41) of the DC bus; is the length of line 14; is the line mode fault current component at R12.

[0094] At any time after the fault occurs, at the fault distance, for any type of fault, . Therefore, at the faulted line end (R12), the product of the post-fault line mode fault voltage and current is greater than zero, i.e. , while at the non-faulted line end (R14) of the DC bus, it is less than zero, i.e. . The specific value is related to the equivalent voltage at the fault point, and in order to expand the difference between the inside and outside of the fault zone, the product over a period of time can be superimposed, i.e.

[0095] ;

[0096] wherein, is the transient inner product of the line mode fault voltage vector and the line mode fault current vector ; is the length of the sampled data, as the length of the vector for calculating the fault component; is the real-time measurement at the time and the line mode fault voltage component after decoupling by the line mode decoupling matrix ; is the real-time measurement at the time and the line mode fault current component after decoupling by the line mode decoupling matrix .

[0097] Since the product of the line mode fault voltage vector and the line mode fault current vector has higher low-pass and high-pass characteristics, but the input function at this time is , and is affected by the fault point transition resistance , the amplitude of the inner product is more susceptible to the fault point The influence of the change. However, for low resistance faults, there is better performance (enlarging the difference between R12 and R41) than using voltage signals or current signals alone, that is, by enlarging the transient characteristic difference between low resistance faults inside and outside the fault area. Therefore, the application proposes that for faults with transition resistance less than 200-300 Ω, the amplitude criterion can be used as the super-fast main protection criterion.

[0098] The application uses the amplitude of the inner product of the line mode fault voltage vector and the line mode fault current vector to detect line faults, that is:

[0099] The is compared with the threshold value of the main protection criterion, and the specific comparison is as follows:

[0100] ;

[0101] wherein, is the threshold value of the main protection criterion; according to the polarity can naturally distinguish faults of the same DC bus and healthy lines. If the fault type is a single-pole ground fault, then , if the fault type is a double-pole ground fault, then , and When the main protection criterion is met, it is identified as an intra-zone fault, and the main protection device acts.

[0102] It should be noted that the main protection device has low sensitivity, and is generally used to identify intra-zone low resistance faults. In some cases, an extremely high resistance fault occurs in the intra-zone, and the main protection device may refuse to act, at which time the backup protection criterion with higher sensitivity is used for protection (i.e., step S5), forming a "multi-section" protection.

[0103] The threshold value of the main protection criterion is selected according to the following principles: less than the protection amplitude of the intra-zone line end high resistance fault, and greater than the intra-zone line end high resistance fault. For single-pole ground faults and pole-to-pole short circuit faults, the application takes , .

[0104] S5, longitudinal direction backup protection criterion: the angle between the line mode fault voltage vector and the line mode fault current vector is substituted into the preset fault direction criterion of the longitudinal line, so as to determine the fault direction; if a positive direction fault occurs at the two ends of the longitudinal line, then whether an intra-zone high resistance fault occurs is determined through the fault label information and time difference at the two ends of the longitudinal line; if an intra-zone high resistance fault occurs, then the backup protection device acts.

[0105] The longitudinal direction backup protection criterion of the application uses the inner angle of the line mode fault voltage vector and the line mode fault current vector to determine the fault direction. When the transition resistance is large, the interference may be generated at the remote end (R41) of the non-fault line of the DC bus. At this time, for the fault line or the remote end (R41) of the non-fault line of the DC bus, is greater than zero. For the near end (R14) of the non-fault line of the DC bus, is less than zero, and the feature difference is increased by accumulation, so the positive and negative of can be used to determine the fault direction.

[0106] Generally, when the high resistance fault is started, the starting criterion is sensitive, and the noise may cause the starting criterion to malfunction, the positive and negative have randomness. In order to further extract this feature and exclude the interference of noise disturbance, the inner angle of the line mode fault voltage vector and the current vector can be expressed as:

[0107] ;

[0108] Among them, is the included angle between the line mode fault voltage vector and the line mode fault current vector ; is the transient inner product of the line mode fault voltage vector and the line mode fault current vector .

[0109] At the fault line port (R12), the voltage and current are approximately negatively correlated, and the included angle exceeds 90°; at the non-fault line port (R21), the theoretical included angle is zero degree; for random noise, the probability density function of the included angle of two n-dimensional random vectors in space is:

[0110] ;

[0111] When , the maximum probability is , and the numerical analysis can obtain and 100, the variances are 0.02 and 0.01 respectively, which indicates that when the vector dimension n is high, the included angle of any vector is theoretically concentrated near . Therefore, threshold 1 (1) and threshold 2 can be set to determine the fault direction, as shown in Figure 4 , wherein is the first reliability coefficient, and the threshold 2 takes the middle value of .

[0112] In the CLR sharing installation case, the fault equivalent circuit is solved, and the line mode fault voltage and current transfer function at different positions are obtained as follows:

[0113]

[0114] Wherein, the reflection coefficient .

[0115] Since there is no CLR protection boundary, the voltage of the DC fault line (R12) and the non-fault line (R14) is completely the same, that is, the voltage measured at the far end (R41) of the non-fault line of the DC bus only passes through the line attenuation, so it is difficult to determine the fault area by the single-ended transient quantity, and the longitudinal method is needed to determine the fault. By observing the transfer function, only the reflection coefficient changes, and the reflection coefficient is still in the interval, and the direction criterion is not affected by the CLR position, because the direction criterion is constrained by the line port topology, and is not constrained by the line protection boundary element.

[0116] In order to exclude the interference of random noise, the fault direction criterion proposed by the application is as follows:

[0117]

[0118] Wherein, is the included angle between the line mode fault voltage vector and the line mode fault current vector . is the first reliable coefficient. In the embodiment, the value of is 1.2.

[0119] If a positive direction fault occurs at both ends of the longitudinal line, whether an intra-zone high resistance fault occurs is determined by the fault label information and time difference at both ends of the longitudinal line, including:

[0120] If a positive direction fault occurs at both ends of the longitudinal line, the protection device sends the corresponding fault label information and the time at the local end of the line to the opposite end of the line through the communication system, and the opposite end of the line receives the fault label information and the time; finally, whether an intra-zone high resistance fault occurs is determined by the fault label information and the time difference at both ends of the longitudinal line, and the expression is as follows:

[0121]

[0122] Wherein, is the corresponding fault label information at the local end of the line; is the fault label information received at the opposite end of the line; is the time when the information is received at the opposite end of the line; ​​​The time difference threshold of both ends of the line after considering various signal delays.

[0123] The time difference of both ends of the line should be theoretically less than the time of the fault traveling wave passing through the whole line, but also includes communication delay, signal processing delay, etc. After comprehensive consideration, in the embodiment, the time difference threshold of both ends of the line after considering various signal delays The calculation formula is as follows:

[0124] ;

[0125] Among them, is the length of the corresponding protection line; is the second reliability coefficient; is the line mode fault traveling wave speed, which is approximately equal to the speed of light. In the embodiment, 1.5 is taken.

[0126] In summary, the fault area is determined by the inner product amplitude when the low resistance fault occurs, the fault direction criterion of both ends of the longitudinal line is used to reliably identify when the high resistance fault occurs, and through the reasonable combination of the protection criterion, the "x multi-section" protection is formed, and the complex and variable DC line fault can be reliably identified.

[0127] The core of the application is to use the transient inner product and the vector inner angle characteristics of the line mode fault voltage vector and the current vector after the fault to construct the "multi-section" protection criterion.

[0128] The application designs a single-ended main protection based on the inner product amplitude and a longitudinal protection based on the inner angle direction criterion, and through the "multi-section" protection, the line fault can be reliably identified. A flexible HVDC grid protection method based on the transient inner product of the line mode fault voltage and current is provided, which can expand the characteristic difference between the inside and outside of the protection zone when the low resistance fault occurs, and naturally distinguish the fault line and the healthy line of the same bus. The fault direction is determined by the inner angle of the fault voltage and current vector, which is greater than when the direction is positive, close to when the direction is negative, and when the random noise occurs, which improves the misoperation of the direction criterion caused by random noise when the starting criterion is high sensitivity.

[0129] The application considers that the HVDC protection should not only improve the single-ended quantity protection high-resistance fault identification capability, but also allow longer protection time during high-resistance fault, and a "multi-section" protection can be designed to improve sensitivity. To this end, first, the equivalent circuit after the fault of the HVDC power grid is established, and the transfer functions of the line mode fault voltage, the line mode fault current and the product of the two are analyzed. Compared with the single use of the line mode fault voltage or the line mode fault current, the product of the transfer functions can expand the transient characteristic difference of the low-resistance faults in and out of the fault area, the inner product amplitude can be used to design the main protection criterion, and the positive and negative values and the inner angle of the inner product can naturally distinguish the fault line and the healthy line of the same DC bus. Finally, the single-ended quantity main protection based on the inner product amplitude and the longitudinal protection based on the new type of directional criterion are designed, and through the "multi-section" protection, the line fault can be reliably identified.

[0130] The application collects the line mode fault voltage and current vectors with a fixed data window length, calculates the transient inner product of the two, uses the amplitude to constitute a single-ended quantity main protection criterion to identify low-resistance faults, and uses the inner angle to constitute a directional criterion, which is combined with the information of the opposite end to constitute longitudinal backup protection to identify high-resistance faults and no-border element scenarios. Through the comprehensive use of the voltage and current transient inner product, the application forms a "multi-section" protection, expands the fault characteristic difference, does not depend on the singular point information, has the advantages of clear threshold, strong anti-interference ability and high reliability, and is suitable for the flexible HVDC power grid with extremely high speed.

[0131] The effect of the application is verified through a simulation experiment.

[0132] Simulation system and parameter setting:

[0133] A ±500 kV HVDC power grid shown in Figure 2 is built in PSCAD / EMTDC. In the simulation, the neutral points of the converters are grounded, the converter 1 controls the DC side voltage to be ±500 kV, and the remaining converters control the active power. The transmission line is an overhead transmission line, and a phase region frequency-dependent model is used in the simulation. Different types (single-pole ground fault and pole-to-pole short circuit fault), different transition resistances (Rf∈[0, 600] Ω) and different fault distances (d∈[2, 204] Ω) of line faults are simulated in the line 12 area, and the performance of the protection scheme is verified by taking the in-zone protection positions R12, R21 and the out-zone protection positions R14, R41 and R23 as examples. In the fault signal, Gaussian white noise with a signal to noise ratio (SNR) of 55 dB is added by default.

[0134] PTG fault main protection results:

[0135] After the positive pole ground fault of different fault distances and transition resistances occurs in the line 12, the protection results of the in-zone R12 and R21 main protections are as follows: Figure 5It can be found from the figure that the protection value is sensitive to the transition resistance, and the protection value decreases significantly with the increase of the transition resistance, but is less than zero. When the transition resistance is less than 300 Ω, the protection value is significantly less than the protection threshold, and the main protection can reliably identify the fault. When the transition resistance is greater than 400 Ω, the main protection may refuse to act, and the backup protection is needed to identify the high resistance fault.

[0136] The protection results of the R14 and R41 main protection in the area are shown in Figure 6 It can be found from the figure that for the non-fault line protection R14 of the DC bus, the protection value also decreases with the increase of the transition resistance, but is greater than zero, so the direction of the fault can be judged by the polarity information of the protection value. For the remote protection R41 of the non-fault line, it may have an impact on the main protection scheme. In the simulation results, the amplitude of R41 is significantly less than the threshold, and can reliably act.

[0137] The high resistance fault pilot protection results are shown in

[0138] For convenience of explanation, a positive ground fault with a transition resistance of 600 Ω is taken as an example, which occurs at a position 30 km away from the R12 protection on line 12. The fault outside the area is calculated to be inactive, and the starting criterion of the area protection R12 and R21 is correctly active. The fault pole criterion is calculated to be , which can correctly identify the positive ground fault. The transient inner product calculated by R12 and R21 is-950.1 MVA and-1638.4 MVA respectively, and the corresponding included angle is 141.5° and 136.0° respectively. In order to study the protection outside the area, the sensitivity of the starting criterion is improved, and the included angle of the outside R14 and R41, and the random noise is as shown in Figure 7 .

[0139] The protection results in the CLR shared installation case are shown in

[0140] In the CLR shared installation case, a positive ground fault is simulated to occur at a position 100 km away from the R12 protection. The fault waveforms are shown in Figure 8 and Figure 9 , the line mode fault voltages at different positions are shown in Figure 8 , and the line mode fault currents at different positions are shown in Figure 9 . Using the proposed method, under the condition of a sampling rate of 20 kHz and N=60 data points used for calculation, the transient inner products of R12, R21, R14 and R41 at four positions are calculated to be-2.7e4, -2.8e4, 8.2e3, -7.1e3 respectively, i.e. the single-ended quantity is difficult to identify the complex and variable fault. The proposed direction criterion is still applicable, and the inner angles calculated at this time are , , and , correctly identify the fault. Although the use of the positive and negative of the transient inner product also allows correct determination, it is easier to distinguish random noise interference after using the inner angle.

Claims

1. A flexible DC grid protection method based on fault vector inner product, characterized in that, include: Data preprocessing: Real-time measurement of the positive line voltage, negative line voltage, positive line current, and negative line current of the line where the signal measurement point is located in the flexible DC power grid; calculation of the changes in positive line voltage, negative line voltage, positive line current, and negative line current using steady-state values; decoupling the changes in positive and negative line voltages using a line-mode decoupling matrix to obtain the line-mode fault voltage component and the zero-mode fault voltage component; decoupling the changes in positive and negative line currents using a line-mode decoupling matrix to obtain the line-mode fault current component and the zero-mode fault current component. Triggering criterion: Determine whether the fault voltage component of the line mode meets the preset judgment conditions; If the fault voltage component of the line mode meets the preset judgment condition, then record the sampling time. corresponding time and continuously sampled ,in, Given the sampled data length, the line-mode fault voltage vector is obtained based on the line-mode fault voltage component and the line-mode fault current component within a sampled length. and line-mode fault current vector And enter the fault determination criteria; Fault criterion: The cumulative amount of zero-mode fault voltage component is used to determine the fault type. Fault types include positive ground fault, negative ground fault, and pole-to-pole short circuit fault. If the fault type is a single-pole ground fault or a double-pole ground fault, different threshold values ​​for the main protection criteria are set for single-pole ground fault and double-pole ground fault, and ultra-fast main protection is activated. Ultra-fast main protection: Calculating the line-mode fault voltage vector and line-mode fault current vector transient inner product ;Will Compare with the threshold of the main protection criterion; if If the value is less than the threshold of the main protection criterion, an intra-zone fault occurs and the main protection device operates; otherwise, the longitudinal backup protection is activated. Longitudinal directional backup protection: Vectoring the line-mode fault voltage and line-mode fault current vector The included angle between the two ends is substituted into the preset fault direction criterion at both ends of the longitudinal connection line to determine the fault direction; if a fault occurs in the positive direction at both ends of the longitudinal connection line, the fault tag information and time difference at both ends of the longitudinal connection line are used to determine whether a high resistance fault in the zone has occurred; if a high resistance fault in the zone occurs, the backup protection device will activate.

2. The flexible DC grid protection method based on fault vector inner product according to claim 1, characterized in that, The determination of whether the fault voltage component of the line mode meets the preset judgment condition is expressed as follows: ； in, Sampling time; For the first Real-time measurement of time and processed by the line-mode decoupling matrix Decoupled line-mode fault voltage components; The threshold for triggering the criterion.

3. The flexible DC grid protection method based on fault vector inner product according to claim 1, characterized in that, The formula for determining the fault type using the cumulative amount of the zero-mode fault voltage component is as follows: ； in, This is the cumulative amount of the zero-mode fault voltage component; For the first Real-time measurement of time and processed by the line-mode decoupling matrix The decoupled zero-mode fault voltage component; The threshold is used for fault type determination; N-PTG is for negative grounding fault; P-PTG is for positive grounding fault; PTP is for pole-to-pole short circuit fault.

4. The flexible DC grid protection method based on fault vector inner product according to claim 1, characterized in that, The calculation of the line mode fault voltage vector and line-mode fault current vector transient inner product The expression is as follows: ； in, The length of the sampled data; For the first Real-time measurement of time and processed by the line-mode decoupling matrix Decoupled line-mode fault voltage components; For the first Real-time measurement of time and processed by the line-mode decoupling matrix The fault current component after decoupling of the line mode.

5. The flexible DC grid protection method based on fault vector inner product according to claim 4, characterized in that: The line-mode fault voltage vector and line-mode fault current vector The product of transfer functions has the characteristic of passing low frequencies and blocking high frequencies. The transient characteristics of low-resistance faults inside and outside the fault zone are expanded.

6. The flexible DC grid protection method based on fault vector inner product according to claim 1, characterized in that, The The comparison with the threshold of the main protection criterion is as follows: ； in, The threshold for the main protection criterion; if the fault type is a single-pole ground fault, then If the fault type is a double-pole grounding fault, then ,and .

7. The flexible DC grid protection method based on fault vector inner product according to claim 1, characterized in that, The line mode fault voltage vector and line-mode fault current vector The included angle between the two ends is substituted into the preset fault direction criterion at both ends of the longitudinal connection line to determine the fault direction. The expression for the fault direction criterion is as follows: ； in, Linear fault voltage vector and line-mode fault current vector The angle between them; It is the first reliability coefficient.

8. The flexible DC grid protection method based on fault vector inner product according to claim 7, characterized in that, The line-mode fault voltage vector and line-mode fault current vector The angle between Its expression is as follows: ； in, Linear fault voltage vector and line-mode fault current vector The transient inner product.

9. The flexible DC grid protection method based on fault vector inner product according to claim 1, characterized in that, If a forward fault occurs at both ends of the longitudinal connecting line, the fault tag information and time difference at both ends of the longitudinal connecting line are used to determine whether a high-resistance fault has occurred within the zone, including: If a forward fault occurs at either end of the longitudinal connection line, the protection device sends the corresponding fault tag information to the local end of the line via the communication system. and time Upon reaching the other end of the line, the other end receives the fault tag information and time. The fault tag information and time difference between the two ends of the longitudinal connection line determine whether a high-resistance fault has occurred within the area, as shown in the following expression: ； in, This refers to the fault tag information corresponding to this end of the line; This refers to the fault tag information received at the other end of the line. The time at which the other end of the line receives information; To account for the time difference threshold between the two ends of the line after considering various signal delays.

10. The flexible DC grid protection method based on fault vector inner product according to claim 9, characterized in that, The time difference threshold between the two ends of the line after considering various signal delays The calculation formula is as follows: ； in, This corresponds to the length of the protection line; The second reliability coefficient; The traveling wave velocity of the line mode fault.

Images