Fault phase selection method for new energy bundling and sending line based on resistance and inductance parameter characteristics

Abstract

The new energy external transmission line fault phase selection method based on the same frequency minimum characteristic of resistance and inductance parameters comprises the following steps: establishing an equivalent RLC network on the opposite side of a single-ended measurement signal; analyzing the characteristics of the fault phase and the non-fault phase respectively; analyzing the resistance and inductance parameter characteristics of the receiving end system; constructing the same frequency minimum phase selection characteristic of the resistance and inductance parameters; calculating the resistance and inductance parameters of the RLC network by using the recursive least square method RLS; and establishing a fault phase selection criterion. The method only applies the single-ended measurement signal to construct the criterion, has small calculation amount, and has a significant phase selection characteristic, so that the method can provide phase selection guidance for the single-ended local protection such as distance protection, and assist the execution of the reclosing process.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power system fault identification, and particularly relates to a new energy bunched external transmission line fault phase selection method based on a resistance-inductance parameter homodyne minimum feature. BACKGROUND

[0002] New energy centralized development and long-distance transmission are the key to promoting efficient consumption of new energy and resolving the source / load reverse distribution problem, and thus in some areas of China, high-proportion new energy and thermal power are bunched into the power system. The external transmission line of such a system has a high voltage level, mostly at 500 kV and above; the energy structure on the power supply side is complex, containing multiple types of new energy and having a high new energy penetration rate. The fault phase selection configured for the current high-voltage line is an important basis for realizing local quantity protection phase tripping. The high-proportion and multi-type new energy access on the power supply side has coupled and complex fault characteristics, seriously interfering with the implementation of conventional fault phase selection methods, and the accuracy of fault phase selection faces great challenges.

[0003] Currently, there are mainly some improved fault phase selection methods for new energy station transmission lines, and they are divided according to the grid-connected type of new energy.

[0004] For example, for inverter-type power sources, document [1]: Liang Yingyu, Lu Zhengjie. Enhanced sequence component phase selection element adaptability photovoltaic grid-connected inverter sequence impedance angle reconstruction scheme [J]. Electric Power Automation Equipment, 2022, 42(1): 133-139. Through the control-protection coordination idea, by adding certain control to the inverter-type controller after the fault, the risk caused by the negative sequence suppression of the inverter-type power source is resolved, and the electrical conditions for the establishment of the traditional phase selection criterion are restored.

[0005] Document [2]: Zhang Junfeng, Gao Liang, Shen Yifei, et al. Fault voltage sequence component phase selection element suitable for double-fed wind farm [J]. Power System Protection and Control, 2018, 46(10): 136-143. For double-fed type power sources, a new sequence component phase selection principle is constructed by compensating the phase angle of the impedance.

[0006] Obviously, the above-mentioned improved phase selection methods for new energy access have obvious power dependence characteristics, that is, they are better adapted to a certain type of new energy access. At the same time, the control-protection coordination idea applied to fault phase selection has high controllability and flexibility, but may cover or conflict with the existing control objectives of new energy, and also depends on the power type and injection source.

[0007] In summary, the known improved phase selection methods do not fully consider the scenario of high-proportion and multi-type cluster access of new energy on the power supply side, and the existing improved methods may not be suitable for the above-mentioned scenario, and there is a gap in the research on fault phase selection methods suitable for high-proportion new energy bunched external transmission lines. SUMMARY

[0008] In order to solve the problem that the conventional fault phase selection criterion is difficult to adapt to high proportion of new energy bundled access, the application provides a new energy bundled outgoing line fault phase selection method based on impedance parameter homology minimum feature, which is based on single-end on-site quantity protection, and an equivalent operation network of the opposite side network is constructed by using single-end measurement signals; through fault network analysis, the impedance parameter homology minimum selection feature existing after the fault is determined; the single-end abrupt variable measurement signal and the recursive least square (RLS) method are used to calculate the impedance parameter of the opposite side equivalent, and finally a homology change criterion distinguishing symmetry and asymmetry is constructed, and a fault phase selection criterion is constructed based on the discrimination result of the homology change, so that the fault phase selection after high proportion of multi-type new energy access is accurately judged.

[0009] The technical scheme adopted by the application is as follows:

[0010] The new energy bundled outgoing line fault phase selection method based on impedance parameter homology minimum feature comprises the following steps:

[0011] Step 1: establishing an opposite side equivalent RLC network of single-end measurement signals;

[0012] Step 2: analyzing the features of the fault phase and the non-fault phase respectively;

[0013] Step 3: analyzing the impedance parameter characteristics of the receiving end system;

[0014] Step 4: constructing the impedance parameter homology minimum selection feature;

[0015] Step 5: calculating the resistance and inductance parameters of the RLC network by using the recursive least square (RLS) method;

[0016] Step 6: establishing a fault phase selection criterion.

[0017] In the step 1, the single-end measurement signals of the power supply side line are taken as the reference to construct the opposite side equivalent RLC network, wherein the line is equivalent to the series-parallel combination of R 、 L 、 C , the receiving end system is equivalent to a passive impedance, and the power supply behavior is described by the numerical change of the impedance parameter; in the steady state, the parameters of the equivalent RLC network satisfy the following relationship:

[0018] (1);

[0019] In the formula (1), the subscript 0 represents the parameters of the steady state operation before the fault; R’ 、 L’ 、 C’ respectively represent the resistance, inductance and capacitance parameters of the equivalent RLC network; R、 L 、 C R, L, C represent the resistance, inductance, capacitance parameters of the line, respectively; and R, L, C represent the equivalent resistance, inductance, capacitance parameters of the receiving end system before the fault.

[0020] In step 2, for the fault phase, the equivalent R, L, C parameters after the fault can be derived from the fault network as follows:

[0021] (2);

[0022] (3);

[0023] wherein, R’ 、 L’ 、 C’ R, L, C represent the resistance, inductance, capacitance parameters of the equivalent RLC network, respectively; R 1 and R 2 represent the line resistance on the left and right sides of the fault point, respectively; R e represents the equivalent resistance of the receiving end system after the fault; A represents the resistance additional coefficient on the right side of the fault point; L 1 and L 2 represent the line inductance on the left and right sides of the fault point, respectively; L e represents the equivalent inductance of the receiving end system after the fault; B represents the inductance additional coefficient on the right side of the fault point; is the transition resistance; is the rated angular frequency.

[0024] It can be seen from equation (3) that the numerator is always less than the denominator regardless of the fault location and the transition resistance, so the coefficient B is always less than 1; while A will be affected by R f , L e , and the fault location.

[0025] For the non-fault phase, since its topological structure is the same as before the fault, the equivalent parameters of resistance, inductance, and capacitance are the same as expressed in equation (1).

[0026] Since the capacitive reactance is much larger than the values of resistance and inductance, from equations (1) and (2), the differences in the equivalent RLC network parameters after the fault are reflected between R and L , C remains unchanged, so only R and Lcharacteristic changes.

[0027] In step 3, through the analysis of step 2, the equivalent R-L parameters of the fault phase are related to the equivalent impedance parameters of the receiving end infinite system; taking the steady-state power flow direction of the line as the reference positive direction, the receiving end infinite system absorbs power which can be equivalent to a positive impedance during steady-state operation; and after the fault, the receiving end infinite system supplies power to the fault point in the reverse direction, and the R-L parameters between the phases are different.

[0028] (1) Fault phase: due to the reverse power supply of the receiving end infinite system to the fault point, and the maximum fault phase current provided by the receiving end infinite system, the R e and L e significantly drops compared to before the fault; at the same time, due to the reverse power supply, the polarity is reversed, R e and L e tends to be negative after the fault;

[0029] (2) Non-fault phase: in actual operation, considering the line mutual inductance and coupling factors, the non-fault line current may have a small amplitude fluctuation, so the R e and L e changes little compared to before the fault.

[0030] In summary, from the aspects of numerical size and numerical polarity, the behavior of the R-L parameters of the receiving end infinite system can be described, and it can be known that the fault phase tends to a small negative value, while the non-fault phase is basically unchanged compared to before the fault.

[0031] In step 4, the R-L parameters of the receiving end infinite system obtained in step 3 are R e and L e characteristics are brought into step 2, and the change rule of the R-L parameters between the fault phase and the non-fault phase can be analyzed;

[0032] (1) Non-fault phase: since the equivalent RLC network structure is unchanged, and the R e and L e changes little compared to before the fault, and finally from the expression of formula (1), it can be known that the equivalent R-L parameters of the non-fault phase represented by the single-ended measurement signal of the power supply side line and only have a small change after the fault;

[0033] (2) Fault phase: for the inductance L’ parameters, since the coefficientB is always less than 1, and L e Decreases and tends to negative values. Combining formula (2) and formula (3), it can be seen that the inductance of the fault phase is L’ The parameters after the fault are much smaller than those in the non-fault phase and steady state. L’ The change pattern is significantly different from that of the non-fault phase L’ For resistors R’ Parameters, due to the non-fault phase R’ unchanged, so the resistance of the fault phase R’ The changing trend of the parameters is different from that of the non-fault phase;

[0034] In addition, for the inductance parameters L’ , due to the fault phase L’ The value after the fault is much smaller than the steady-state value before the fault, so its change is negative. The change of the resistance and inductance parameters in the equivalent RLC network is characterized by the sudden change of the three-phase voltage and current measured at a single end, that is,

[0035] (4);

[0036] In formula (4): is the change of inductance parameter; L’ is the inductance parameter after the fault ； is the inductance parameter before the fault.

[0037] In step 5, the voltage and current mutations measured at the power supply side line port are used as inputs, and the recursive least squares (RLS) method is applied to calculate the changes in the resistance and inductance parameters in real time. The voltage and current equations of the single-port network are:

[0038] (5);

[0039] in, u and i Respectively represent the voltage and current mutation amounts on the M side; t represents the time variable, R 、 L 、 C are the resistance, capacitance and inductance parameters of the equivalent network respectively; X 1. X 2. X 3. X 4 are the four parameters to be determined output by the RLS algorithm.

[0040] Rewrite Equation (5) into a system linear equation:

[0041] (6);

[0042] In formula (5),Y n ) represents the system equation dependent variable matrix, which is based on formula (5), Y n )=[ u (1) u (2) … u n ] T , u (1) u (2) … u n ) respectively represent the values of the 1st, 2nd, …, the nth voltage mutation variable; n H ) is an input matrix, n H n )=[ h (1) h (2) … h n ] T , h (1) h (2) … h n ) respectively represent the values of the 1st, 2nd, …, the nth sampling point in the input matrix; n h n d 2 u / dt 2 du / dtdi / dti ; X n ) is a to-be-estimated parameter matrix, X n X 1 X 2 X 3 X 4 T .

[0043] The recursion and parameter update formula of the RLS algorithm with a forgetting factor are as follows:

[0044] (7)

[0045] In the formula, n represents the current number of sampling points; represents the forgetting factor. represents the value of the gain vector corresponding to the (n+1)th sampling point; represents the value of the estimated error covariance matrix corresponding to the (n+1)th sampling point;​​​​​​​​​​​​​​​ Represents the value of the parameter matrix to be estimated corresponding to the n+1th sampling point; Represents the value of the n+1th sampling point in the input matrix; Represents the value of the dependent variable matrix of the system equation corresponding to the n+1th sampling point, that is, the value of the n+1th voltage mutation; The value of the estimated error covariance matrix corresponding to the nth sampling point; the value of the parameter matrix to be estimated corresponding to the nth sampling point; The transposed vector of the value at sample point (n+1) in the input matrix.

[0046] The changes in resistance and inductance parameters obtained using RLS are:

[0047] ;

[0048] in, Represents the change in resistance parameters, which is equal to the fourth value of the parameter matrix to be estimated by the RLS algorithm X 4; Indicates the change in inductance parameters, which is equal to the third value of the parameter matrix to be estimated by the RLS algorithm. X 3.

[0049] In step 6, the resistance and inductance parameters of the three-phase equivalent network calculated in step 5 are used to calculate the Euclidean distance between the resistance and inductance of the three phases to characterize the coherent change trend, and a method for identifying the coherent change trend of symmetrical and asymmetrical faults is constructed, and finally a fault phase selection criterion is established. It includes:

[0050] (1) Identification of coherent change trends:

[0051] The Euclidean distance of the phase-to-phase resistance and inductance parameters is calculated using a sliding data window as an indicator of the coherence change trend. The calculation is as follows:

[0052] (8);

[0053] In formula (8), Represents the inductive Euclidean distance between phases m and n at the nth sampling point; the values ​​of m and n are A, B, or C; D Indicates the total number of data points in the sliding data window. Considering the requirement of feature smoothness, the total number of data points in one cycle is taken; n Indicates the n sampling points; N 0 is the sampling point number after the criterion is started. Indicates the change in inductance parameter of phase m at the ith sampling point in the data window; Indicates the change in inductance parameter of phase n at the ith sampling point in the data window.

[0054] The Euclidean distance of resistance is the same as the above formula, only replace L with R , and the calculation is as follows:

[0055] ;

[0056] wherein, represents the Euclidean distance of resistance between m and n at the nth sampling point; m and n are A, B or C; D represents the total number of data points in the sliding data window, considering the requirement of feature smoothness, the total number of data points of one cycle is taken; n represents the n th sampling point; N 0 is the sampling point number after the criterion is started. represents the change amount of resistance parameter of the m-phase of the i th sampling point in the data window; represents the change amount of resistance parameter of the n-phase of the i th sampling point in the data window,

[0057] Different coherence change criteria are constructed for symmetric and asymmetric fault characteristics.

[0058] The Euclidean distance of three-phase parameters of symmetric fault is small, and theoretically tends to 0. The criterion shown in formula (9) is constructed to identify symmetric fault.

[0059] (9);

[0060] wherein, D Rmn represents the three-phase Euclidean distance of resistance parameter; subscripts m and n are phase indicators, taking values of A, B or C; N 0 is the sampling point number after the criterion is started; is the resistance Euclidean distance of the i th sampling point; M is the average value calculation window; D set is the symmetric fault discrimination threshold;

[0061] The inductance parameter calculation is the same, replace R with L . The calculation is as follows:

[0062] ;

[0063] wherein, D Lmn represents the three-phase Euclidean distance of inductance parameter; subscripts m and n are phase indicators, taking values of A, B or C; N 0 is the sampling point number after the criterion is started. is the Euclidean distance of inductance for the m-th sampling point; M is the average value calculation window; i is the Euclidean distance of inductance for the m-th sampling point; M is the average value calculation window; D set is the symmetry fault discrimination threshold;

[0064] The non-symmetry fault always has a coherent change trend in two phases, so the coherent change criterion of the non-symmetry fault is:

[0065] (11);

[0066] In formula (10), represents the resistance Euclidean distance between the m phase and the n phase; represents the resistance Euclidean distance between the x phase and the y phase; represents the inductance Euclidean distance between the m phase and the n phase; represents the inductance Euclidean distance between the x phase and the y phase. For example, mn is AB, and xy is BC or CA.

[0067] a and b respectively resistance and inductance adaptive comparison coefficient.

[0068] (2) Fault phase selection criterion:

[0069] On the basis of the coherent change phase distinguished above, the fault phase selection criterion is constructed:

[0070] (13);

[0071] wherein, represents the inductance parameter change amount of the phase; p represents the inductance parameter change amount of the phase; represents the inductance parameter change amount of the phase; represents the inductance parameter change amount of the phase; m represents the inductance parameter change amount of the phase; represents the inductance parameter change amount of the phase; n represents the inductance parameter change amount of the phase; represents the logical "and" operation.

[0072] The new energy bundling outgoing line fault phase selection method based on the coherent minimum feature of resistance and inductance parameters has the following technical effects:

[0073] 1) The present application carries out fault phase selection based on the network parameter consistency change law, and does not depend on the power type, and has good adaptability to high proportion of new energy bundling access on the power side.

[0074] 2) The present application only needs to calculate the voltage and current sudden change on one side of the line, without additional control, which meets the basic principle of mutual cooperation of fault phase selection and local protection such as distance protection.

[0075] 3) The method has effect in the whole length of the line, has good tolerance to transition resistance, and can adapt to different fault ride-through behaviors of new energy power sources.

[0076] 4) The method has good adaptability to different types and high proportion of new energy access, is not affected by fault location, and has certain tolerance to transition resistance.

[0077] 5) The method only applies single-ended measurement signals to construct the criterion, has small calculation amount, obvious phase selection characteristics, can provide phase selection guidance for single-ended local protection such as distance protection, and assists the execution of the reclosing process. BRIEF DESCRIPTION OF DRAWINGS

[0078] The application will be further described below in combination with the drawings and examples.

[0079] Figure 1 A schematic diagram of a high-proportion new energy bundled sending scene.

[0080] Figure 2 is a network diagram of opposite side equivalent operation of single-ended measurement.

[0081] Figure 3 is a fault network diagram of AG fault.

[0082] Figure 4 is a schematic diagram of resistance and inductance parameter behavior of a receiving end infinite system.

[0083] Figure 5 is a flowchart of a fault phase selection method based on resistance and inductance parameter homodyne minimum characteristics.

[0084] Figure 6(a) is the phase selection result (resistance and inductance parameters) of AG fault;

[0085] Figure 6(b) is the phase selection result (resistance and inductance parameter Euclidean distance) of AG fault.

[0086] Figure 7(a) is the phase selection result (resistance and inductance parameters) of AB fault;

[0087] Figure 7(b) is the phase selection result (resistance and inductance parameter Euclidean distance) of AB fault.

[0088] Figure 8(a) is the phase selection result (resistance and inductance parameters) of ABC fault;

[0089] Figure 8(b) is the phase selection result (resistance and inductance parameter Euclidean distance) of ABC fault.

[0090] Figure 9(a) is the phase selection result (resistance and inductance parameters) of AG fault transition resistance 300Ω;

[0091] Figure 9(b) is the phase selection result (resistance and inductance parameter Euclidean distance) of AG fault transition resistance 300Ω.

[0092] Figure 10(a) is the phase selection result of the line head AG fault (resistance inductance parameters);

[0093] Figure 10(b) is the phase selection result of the line head AG fault (resistance inductance parameter Euclidean distance). DETAILED DESCRIPTION

[0094] The high-proportion new energy bundling outgoing line fault phase selection method based on the coherent minimum characteristic of resistance inductance parameters. The method aims at the problem that the new energy cluster access to the power system causes the failure of the conventional fault phase selection method, and identifies the resistance inductance parameter characteristics of the equivalent system through single-ended measurement signals, so as to realize the judgment of the fault phase. The method starts from the change rule of the fault network parameter consistency, describes the opposite side equivalent of the equivalent RLC network by the single-ended measurement signal, and on the basis of analyzing the fault impedance behavior of the receiving end infinite power source, constructs the fault phase selection criterion based on the coherent minimization characteristic of the resistance inductance parameters. The method includes the following steps:

[0095] Step one: modeling of the opposite side equivalent RLC network of the single-ended measurement signal. The fault phase selection serves the single-ended local protection, and should only use the electrical quantities measured by the single-ended measurement to construct the criterion. The opposite side of the measurement point on the power source side is equivalent to an equivalent RLC network, which includes the line and the receiving end infinite system. The line is in series-parallel connection with R, L and C; the receiving end infinite system is equivalent to a series model of R and L, and the fault behavior of the infinite system is considered in the change rule of its impedance.

[0096] A typical high-proportion new energy bundling outgoing system is shown in Figure 1 , the outgoing line voltage level is 500 kV, and the power source side includes typical new energy bundling access such as inverter type and double-fed type. The fault phase selection cooperates with the local protection such as distance protection, so the measurement signal of the single-ended line on the power source side (M side) is taken as the reference to construct the RLC passive operational network of the opposite side. The line is equivalent to R , L , C in series-parallel combination, the receiving end system is equivalent to a passive impedance, and the numerical change of the resistance inductance parameters is used to describe the power source behavior. The basic schematic of the equivalent is shown in Figure 2 . In the steady state, the parameters of the equivalent RLC network have the following relationships:

[0097] (1);

[0098] wherein subscript 0 represents the parameters of the steady state before the fault; R’ , L’ , C’ respectively represent the resistance, inductance and capacitance parameters of the equivalent RLC network; R , L , C respectively represent the resistance, inductance and capacitance parameters of the line; and R and L represent the equivalent resistance and inductance parameters of the receiving infinite system before fault.

[0099] Step 2: Fault network analysis. Analyze the equivalent resistance and inductance parameters of different phases after fault according to the fault phase.

[0100] Fault phase, its network is shown in Figure 3 The equivalent R, L and C parameters after fault can be derived from the fault network shown in Figure 3

[0101] (2);

[0102] (3);

[0103] wherein, R’ , L’ , C’ R, L and C represent the resistance, inductance and capacitance parameters of the equivalent RLC network respectively; R 1 and R 2 represent the line resistance on the left and right sides of the fault point respectively; R e R represents the equivalent resistance of the receiving infinite system after fault; A K represents the resistance additional coefficient on the right side of the fault point; L 1 and L 2 represent the line inductance on the left and right sides of the fault point respectively; L e L represents the equivalent inductance of the receiving infinite system after fault; B K represents the inductance additional coefficient on the right side of the fault point; Rt represents the transition resistance;ω represents the rated angular frequency.

[0104] As can be seen from equation (3), no matter the fault location and the transition resistance, the numerator is always less than the denominator, so the coefficient B is always less than 1; while A will be affected by R f , L e and the fault location.

[0105] For the non-fault phase, since its topological structure is the same as that before fault, the equivalent parameters of resistance, inductance and capacitance are the same as equation (1).

[0106] Since the capacitive reactance is much larger than the resistance and inductance, from equation (1) and equation (2), the main difference of the network parameters after fault is R and L ​between, C is basically unchanged, so only the R and L characteristics are concerned.

[0107] Step three: analysis of the resistance and inductance parameter characteristics of the receiving end infinite system. Taking the steady-state power flow direction as the positive direction, the equivalent resistance and inductance parameter variation law of the infinite system after the line fault is determined, and the size and polarity of the resistance and inductance parameters are used to represent the state after the fault. As can be seen from the analysis in step two, the equivalent resistance and inductance parameters of the fault phase are related to the equivalent impedance parameters of the receiving end infinite system. Taking the steady-state power flow direction of the line as the reference positive direction, the receiving end infinite system absorbs power in the steady state, which can be equivalent to a positive impedance;

[0108] After the fault, the receiving end infinite system supplies power to the fault point in the reverse direction, and there is a large difference between the resistance and inductance parameters of different phases.

[0109] (1) Fault phase. Since the receiving end infinite system supplies power to the fault point in the reverse direction, and the fault phase current provided by the receiving end infinite system is the largest, the R e and L e significantly drops compared to before the fault; at the same time, due to the reverse power supply, the polarity is reversed, R e and L e tends to be negative after the fault.

[0110] (2) Non-fault phase. In actual operation, considering factors such as line mutual inductance and coupling, the non-fault line current may have a small amplitude fluctuation, so the R e and L e changes little compared to before the fault.

[0111] In summary, from the aspects of numerical size and numerical polarity, the behavior of the resistance and inductance parameters of the receiving end infinite system can be described, and it can be seen that the fault phase tends to a small negative value, while the non-fault phase is basically unchanged compared to before the fault.

[0112] Step four: construction of the coherent minimum selection phase feature. Considering both resistance and inductance parameters, the infinite impedance law obtained in step three is brought into the fault network analysis conclusion in step two to construct the coherent minimum selection phase feature of the resistance and inductance parameters. By bringing the R e and L e features of the receiving end infinite system obtained in step three into step two, the variation law of the resistance and inductance parameters between the fault phase and the non-fault phase can be analyzed.

[0113] (1) Non-faulty phase: since the equivalent RLC network structure is unchanged, and the non-faulty phase of the receiving end is infinite system R e and L e Compared with the change before the fault, finally, according to the expression of formula (1), the equivalent resistance and inductance parameters of the non-faulty phase represented by the single-ended measurement signal of the line on the power side and only have small changes after the fault.

[0114] (2) Fault phase: for the inductance L’ parameter, since the coefficient B is always less than 1, and L e decreases and tends to a negative value, combined with formula (2) and formula (3), the inductance L’ parameter of the fault phase tends to a negative value after the fault, and is much smaller than that of the non-faulty phase and the steady state value, and the change rule of the fault phase L’ is significantly different from that of the non-faulty phase L’ ; for the resistance R’ parameter, although it is coupled by many factors, since the resistance R’ of the non-faulty phase is basically unchanged, the change trend of the resistance R’ of the fault phase is significantly different from that of the non-faulty phase.

[0115] In summary, after the fault, the R’ and L’ parameters in the equivalent RLC network have the same change trend, that is, for asymmetric faults, there is always a same change (synchronization) of the two-phase parameters.

[0116] In addition, for the inductance parameter L’ , since the value of the fault phase L’ after the fault is much smaller than the steady state value before the fault, the change amount should be negative. The change of the resistance and inductance parameters in the equivalent RLC network is represented by the single-ended measurement of the three-phase voltage and current jump variables, that is

[0117] (4);

[0118] wherein: is the change amount of the inductance parameter; L’ is the inductance parameter after the fault ； is the inductance parameter before the fault.

[0119] In summary, based on the equivalent RLC network method, the resistance and inductance parameters have obvious same change minimization phase selection characteristics.

[0120] Step five: online driving method of equivalent RLC network. The voltage and current jump variables measured at the line port of the power supply side are taken as inputs, and the resistance and inductance parameters of the equivalent network are calculated online by recursive least squares (RLS).

[0121] The voltage and current equations of the single-port network are shown in Figure 2

[0122] (5);

[0123] wherein, u and i represent the voltage and current jump variables on the M side, respectively; t represents the time variable, R , L , C are the resistance, capacitance, and inductance parameters of the equivalent network, respectively; X 1, X 2, X 3, X 4 are the four parameters to be solved output by the RLS algorithm.

[0124] Rewrite equation (5) as a system of linear equations:

[0125] (6);

[0126] wherein, Y ( n ) represents the dependent variable matrix of the system equation, and based on equation (5), Y ( n )=[ u (1) u (2) … u ( n )] T ; H ( n ) is the input matrix, H ( n )=[ h (1) h (2) … h ( n )] T , h ( n )=[ d 2 u / dt 2 du / dtdi / dti ], wherein the first and second derivatives of the discretely sampled voltage and current are calculated by the difference method; X ( n ​) is a parameter matrix to be estimated, X ( n )=[ X 1 X 2 X 3 X 4] T .

[0127] The recursion and parameter update formula of the RLS algorithm with a forgetting factor are as follows:

[0128] (7);

[0129] wherein, K and P are a gain vector matrix and a covariance matrix respectively; n denotes the current sampling point number; h ( n ) is an independent variable matrix; y ( n ) is an input variable; denotes a forgetting factor, and the present application takes a value of 0.96.

[0130] Initial values of the RLS algorithm iteration X ( n ) and P ( n ) are calculated adaptively by taking the first four groups of data, so as to increase the adaptability of the algorithm.

[0131] Step six: fault phase selection criterion. The coherent minimization phase selection feature possessed by the resistance and inductance parameters of the equivalent RLC network is used to complete the phase selection work according to the flow of coherent trend identification and phase selection criterion. Specifically, the resistance and inductance parameters of the three-phase equivalent network calculated in step five are used to calculate the Euclidean distance of the resistance and inductance between the three phases to represent the coherent change trend, construct the coherent change identification method of symmetric and asymmetric faults, and finally establish the fault phase selection criterion.

[0132] (1) Coherent change trend identification:

[0133] The Euclidean distance of the resistance and inductance between the three phases is calculated by using a sliding data window, which is used as an index of the coherent change trend of the resistance and inductance. The calculation method of the inductance parameter is as follows:

[0134] (8);

[0135] wherein, L mn denotes the Euclidean distance of the inductance; the subscripts m and n are phase indicators, and take values of A, B or C; DN represents the total number of data points in the sliding data window, and the total number of data points in one cycle is taken considering the requirement of feature smoothness; n N represents the total number of data points in the sliding data window, and the total number of data points in one cycle is taken considering the requirement of feature smoothness; n N represents the total number of data points in the sliding data window, and the total number of data points in one cycle is taken considering the requirement of feature smoothness; N 0 is the sampling point number after the criterion is started. The Euclidean distance of resistance is the same as the above formula, only replace L with R .

[0136] Symmetrical and asymmetrical faults have different harmonic variation characteristics: the Euclidean distance of three-phase parameters of symmetrical faults is small, and theoretically approaches 0; while asymmetrical faults always have harmonic variation of two-phase parameters, and the Euclidean distance between the two phases should be a certain large value.

[0137] First, construct the criterion to identify symmetrical faults as shown in formula (7).

[0138] (9);

[0139] wherein, D Rmn N represents the total number of data points in the sliding data window, and the total number of data points in one cycle is taken considering the requirement of feature smoothness; m and n are phase indicators, taking values of A, B or C; N 0 is the sampling point number after the criterion is started. N represents the total number of data points in the sliding data window, and the total number of data points in one cycle is taken considering the requirement of feature smoothness; i M is the average value calculation window, which is 5 in this paper; D set N represents the total number of data points in the sliding data window, and the total number of data points in one cycle is taken considering the requirement of feature smoothness; R replace L .

[0140] (10);

[0141] If it meets formula (8), it is identified as a three-phase symmetrical fault, otherwise, the harmonic trend of asymmetrical fault is judged.

[0142] Asymmetrical faults always have harmonic variation trend of two-phase, so the harmonic variation criterion of asymmetrical faults is:

[0143] (11);

[0144] In the formula, subscript xy represents two-phase combination except mn , for example, mn is BC, xy is AB or CA; a and b are respectively the adaptive comparison coefficients of resistance and inductance. Coefficients a ,b The calculation formula is shown as formula (10).

[0145] (12);

[0146] In the formula, K rel is a reliability coefficient, and the present application takes a smaller value to avoid the influence of early-stage sudden increase, K rel 0.2 is taken; M The calculation length of the adaptive threshold value is 5 in the present application. i The first sampling value is represented by X1. i The second sampling value is represented by X2.

[0147] (2) Fault phase selection criterion.

[0148] Based on the coherent change distinguished by the above criterion, a fault phase selection criterion based on the minimization of the fault phase inductance parameter can be constructed:

[0149] (13);

[0150] In the formula, the subscript represents the phase distinction mark; represents the logical AND operation; represents the inductance parameter calculated by the voltage and current sudden change variables.

[0151] The flow chart of the fault phase selection method based on the coherent minimum feature of the resistance-inductance parameter is shown in Figure 5 .

[0152] Verification example:

[0153] To verify the effectiveness of the fault phase selection method proposed in the present application, a high-proportion new energy bundling transmission model is built in PSCAD as shown in Figure 1 . The rated power of various types of energy on the power supply side is: 500 MW of thermal power, 500 MW of double-fed wind power, and 250 MW of inverter photovoltaic power. The system rated voltage is 500 kV, and the line length is 300 km; the new energy is configured with low voltage ride through strategy and negative sequence suppression, crowbar protection and other strategies according to the general control method. The voltage and current sudden change variables of M side are taken for fault phase selection.

[0154] Scenario 1: The phase selection result of AG fault. The AG metallic ground fault is set at the midpoint of the line, and the phase selection result is shown in Fig. 6(a) and Fig. 6(b). As shown in Fig. 6(a), the resistance-inductance parameters have significant coherent change trends, and the Euclidean distance of the calculated resistance and inductance parameters can be seen. The non-fault phase (B, C two phases are the smallest). The adaptive threshold value calculated by formula (10) a , bare 1.4 and 6.6 respectively, which satisfies the coherence change criterion of BC shown in formula (11). According to the criterion shown in formula (13), only Δ L A If it is less than zero, it is determined to be a phase A fault.

[0155] Scenario 2: Phase selection results for an AB fault. An AB fault is set at the line midpoint, and the results are shown in Figures 7(a) and 7(b). The resistance and inductance parameters of the two faulty phases (phases A and B) exhibit a coherent change trend. Correspondingly, the Euclidean distance between the resistance and inductance parameters is the smallest for AB. Equations (11) and (12) reliably identify the phase selection as AB coherent. Based on this result, we further verify that only the inductance parameters of phases A and B are negative, thus identifying the phase selection as an AB fault.

[0156] Scenario 3: Phase selection results for an ABC fault. An ABC fault is set at the midpoint of the line, and the results are shown in Figures 8(a) and 8(b). The resistance and inductance parameters of the fault phases (A, B, and C) exhibit a coherent change trend. Correspondingly, the Euclidean distances of the resistance and inductance parameters are close to each other for phases AB, BC, and CA. Calculating the average distances of the resistance and inductance parameters reveals that the average maximum distances of the resistance and inductance are 0.22 and 0.01, respectively. According to Equation (9), these values ​​are far less than the threshold setting, and are therefore identified as a three-phase coherent change. Finally, according to the criterion shown in Equation (13), the inductance parameters for all three phases are negative, indicating an ABC fault.

[0157] Scenario 4: The results of the AG fault transition resistance of 300Ω are shown in Figures 9(a) and 9(b). As the transition resistance increases, the trend of synchronous changes in the resistance and inductance parameters does not disappear. In particular, according to Equations (2) and (3), the inductance parameters are not affected by the transition resistance. Therefore, the convergence and magnitude of the inductance parameters in Figures 6(a), 6(b) and 9(a), 9(b) are almost the same. The adaptive threshold calculated by Equation (12) a 、 b are 2.6 and 5.4 respectively, which satisfies the criterion of the two homophonic changes of BC shown in formula (11). According to the criterion shown in formula (13), it can be seen that only Δ L A If it is less than zero, it is determined to be a phase A fault.

[0158] Scenario 5: Phase selection for different penetration rates and wind-solar ratios, the results are shown in Table 1. Set the power supply side penetration rates to 0, 40, 60, 80, and 100, and set different wind-solar ratios under each penetration rate. Even in the scenario where 100% new energy is connected to the power supply side, or there is no new energy on the power supply side (conventional synchronous power supply system), the method of the present invention can select the correct fault phase based on the established synchronization minimization index. It can be seen that the fault phase selection method constructed based on the synchronization minimization of resistance-inductance parameters is not affected by the new energy penetration rate on the power supply side, does not limit the power supply type, has good flexibility, and can adapt to the scenario of high-proportion new energy bundled transmission.

[0159] Table 1 Phase selection results for different penetration rates and different wind-solar ratios

[0160]

[0161] Scenario 6: Phase selection results at different fault locations. The results of the AG fault at the head end of the line are shown in Figures 10(a) and 10(b). As the fault location changes, the trend of synchronous changes in the resistance and inductance parameters does not disappear. In particular, according to Equations (2) and (3), the inductance parameter is not affected by the fault location. Therefore, the convergence and magnitude of the inductance parameters in Figures 6(a), 6(b) and 10(a), 10(b) are almost the same. The adaptive threshold calculated by Equation (12) a 、 b are 1.7 and 4.1 respectively, which satisfies the criterion of the two homophonic changes of BC shown in formula (11). Further according to the criterion shown in formula (13), it can be seen that only Δ L A If it is less than zero, it is determined to be a phase A fault. The results at the end of the line are similar to those at the beginning, and the coherence minimization feature is significant. The adaptive threshold a 、 b They are 3.2 and 7.2 respectively, and both can be reliably identified as phase A fault.

Claims

1. A fault phase selection method for renewable energy bundled transmission lines based on the minimum coherent characteristics of resistance and inductance parameters, characterized by The following steps are involved: Step 1: Establish an equivalent RLC network on the opposite side of the single-ended measurement signal; Step 2: Analyze the characteristics of the fault phase and the non-fault phase respectively; Step 3: Analyze the resistance and inductance parameter characteristics of the receiving system; Step 4: Construct the phase selection feature of the coherent minimization of the resistance and inductance parameters; Step 5: Calculate the resistance and inductance parameters of the RLC network using the recursive least squares method RLS; Step 6: Establish fault phase selection criteria; In step 2, for the fault phase, the equivalent R, L, and C parameters after the fault can be derived from the fault network: （2）； （3）； in, R’ 、 L’ 、 C’ They represent the resistance, inductance, and capacitance parameters of the equivalent RLC network respectively; R 1 and R 2 represents the line resistance on the left and right sides of the fault point respectively; R e Indicates the equivalent resistance of the receiving system after a fault; A Indicates the additional resistance coefficient on the right side of the fault point; L 1 and L 2 represents the line inductance on the left and right sides of the fault point respectively; L e Indicates the equivalent inductance of the receiving system after the fault; B Indicates the additional inductance coefficient on the right side of the fault point; is the transition resistance; is the rated angular frequency; It can be seen that in formula (3), no matter what the fault location and transition resistance are, the numerator is always smaller than the denominator, so the coefficient B is always less than 1; and A will be affected R f 、 L e , and the impact of fault location; For the non-fault phase, since its topology is the same as before the fault, the equivalent parameters of resistance, inductance, and capacitance are the same as those expressed in formula (1); Since the capacitive reactance of the capacitor is much larger than the resistance and inductance, it can be seen from equations (1) and (2) that the difference in the equivalent RLC network parameters after the fault is reflected in R and L between, C Unchanged, so only focus on R and L characteristic changes.

2. The fault phase selection method for renewable energy bundled transmission lines based on the minimum coherent characteristic of resistance and inductance parameters according to claim 1 is characterized by: In step 1, the single-ended measurement signal of the power supply side line is used as a reference to construct an equivalent RLC network on the opposite side, wherein the line is equivalent to R 、 L 、 C The receiving end system is equivalent to a passive impedance, and the power supply behavior is described by the numerical changes of the resistance and inductance parameters. In steady state, the parameters of the equivalent RLC network have the following relationship: （1）； In formula (1): subscript 0 represents the parameters of steady-state operation before the fault; R’ 、 L’ 、 C’ They represent the resistance, inductance, and capacitance parameters of the equivalent RLC network respectively; R 、 L 、 C Respectively represent the resistance, inductance, and capacitance parameters of the circuit; and They represent the equivalent resistance and inductance parameters of the receiving system before the fault.

3. The fault phase selection method for renewable energy bundled transmission lines based on the minimum coherent characteristic of resistance and inductance parameters according to claim 1 is characterized by: In step 3, the analysis in step 2 shows that the equivalent resistance and inductance parameters of the fault phase are related to the equivalent impedance parameters of the infinite receiving system. Taking the steady-state power flow direction of the line as the reference positive direction, during steady-state operation, the power absorbed by the infinite receiving system can be equivalent to a positive impedance. After a fault occurs, the infinite receiving system supplies reverse power to the fault point, resulting in differences in resistance and inductance parameters between phases.

4. The fault phase selection method for renewable energy bundled transmission lines based on the minimum coherent characteristic of resistance and inductance parameters according to claim 3 is characterized by: In step 3: (1) Fault phase: Since the receiving-end infinite system supplies reverse power to the fault point, and the fault phase current provided by the receiving-end infinite system is the largest, R e and L e Significant drop compared to before the failure; At the same time, due to reverse power supply, the polarity is reversed. R e and L e Tends to be negative after a fault; (2) Non-fault phase: In actual operation, considering the mutual inductance and coupling factors of the line, the current of the non-fault line may fluctuate slightly. Therefore, the current of the non-fault phase after the fault R e and L e The change is small compared with that before the fault.

5. The fault phase selection method for renewable energy bundled transmission lines based on the minimum coherent characteristic of resistance and inductance parameters according to claim 1 is characterized by: In step 4, the infinite system of the receiving end obtained in step 3 is converted into R e and L e By bringing the characteristics into step 2, we can analyze the change pattern of the resistance and inductance parameters between the fault phase and the non-fault phase; (1) Non-fault phase: Since the equivalent RLC network structure remains unchanged and the receiving end is infinite, the non-fault phase of the system R e and L e Compared with the change before the fault, it is smaller. Finally, it can be seen from the expression of formula (1) that the non-fault equal value resistance-inductance parameter represented by the single-ended measurement signal of the power supply side line is and There were only minor changes after the failure; (2) Fault phase: For inductance L’ Parameters, due to the coefficient B is always less than 1, and L e Decreases and tends to negative values. Combining formula (2) and formula (3), it can be seen that the inductance of the fault phase is L’ The parameters after the fault are much smaller than those in the non-fault phase and steady state. L’ The change pattern is significantly different from that of the non-fault phase L’ For resistors R’ Parameters, due to the non-fault phase R’ unchanged, so the resistance of the fault phase R’ The changing trend of the parameters is different from that of the non-fault phase; In addition, for the inductance parameters L’ , due to the fault phase L’ The value after the fault is much smaller than the steady-state value before the fault, so its change is negative; the change of the resistance and inductance parameters in the equivalent RLC network is characterized by the sudden change of the three-phase voltage and current measured at a single end, that is, （4）； In formula (4): is the change of inductance parameter; L’ is the inductance parameter after the fault ； is the inductance parameter before the fault.

6. The fault phase selection method for renewable energy bundled transmission lines based on the minimum coherent characteristic of resistance and inductance parameters according to claim 1 is characterized by: In step 5, the voltage and current mutations measured at the power supply side line port are used as inputs, and the recursive least squares method RLS is applied to calculate the changes in the resistance and inductance parameters in real time. The voltage and current equations of the single-port network are: （5）； in, u and i Respectively represent the voltage and current mutation amounts on the M side; t represents the time variable, R 、 L 、 C are the resistance, capacitance and inductance parameters of the equivalent network respectively; X 1. X 2. X 3. X 4 are the four parameters to be determined output by the RLS algorithm; Rewrite Equation (5) into a system linear equation: （6）； In formula (6), Y ( n ) represents the dependent variable matrix of the system equation. Based on formula (5), we can know that Y ( n )=[ u (1) u (2) … u ( n )] T , u (1) u (2) … u ( n ) represent the 1st, 2nd, ... n The value of the voltage mutation; H ( n ) is the input matrix, H ( n )=[ h (1) h (2) … h ( n )] T , h (1) h (2) … h ( n ) represent the 1st, 2nd, ... n The value of the sampling point; h ( n )=[ d 2 u / dt 2 du / dtdi / dti ]; X ( n ) is the parameter matrix to be estimated, X ( n )=[ X 1 X 2 X 3 X 4] T ; The recursion and parameter update formula of the RLS algorithm with forgetting factor is: （7）； Where, n Indicates the current number of sampling points; represents the forgetting factor; Represents the value of the gain vector corresponding to the n+1th sampling point; Represents the value of the estimated error covariance matrix corresponding to the n+1th sampling point; Represents the value of the parameter matrix to be estimated corresponding to the n+1th sampling point; Represents the value of the n+1th sampling point in the input matrix; Represents the value of the dependent variable matrix of the system equation corresponding to the n+1th sampling point, that is, the value of the n+1th voltage mutation; The value of the estimated error covariance matrix corresponding to the nth sampling point; the value of the parameter matrix to be estimated corresponding to the nth sampling point; The transposed vector of the value of the n+1th sampling point in the input matrix; The changes in resistance and inductance parameters obtained using RLS are: ； in, Represents the change in resistance parameters, which is equal to the fourth value of the parameter matrix to be estimated by the RLS algorithm X 4; Indicates the change in inductance parameters, which is equal to the third value of the parameter matrix to be estimated by the RLS algorithm. X 3.

7. The fault phase selection method for renewable energy bundled transmission lines based on the minimum coherent characteristic of resistance and inductance parameters according to claim 6 is characterized by: In step 6, the resistance and inductance parameters of the three-phase equivalent network calculated in step 5 are used to calculate the Euclidean distance of the resistance and inductance between the three phases to characterize the coherent change trend, construct a coherent change trend identification method for symmetrical and asymmetrical faults, and finally establish a fault phase selection criterion; including: (1) Identification of coherent change trends: The Euclidean distance of the phase-to-phase resistance and inductance parameters is calculated using a sliding data window as an indicator of the coherence change trend. The calculation is as follows: （8）； In formula (8), Represents the inductive Euclidean distance between phases m and n at the nth sampling point; the values ​​of m and n are A, B, or C; D Indicates the total number of data points in the sliding data window. Considering the requirement of feature smoothness, the total number of data points in one cycle is taken; n Indicates the n sampling points; N 0 is the sampling point number after the criterion is started; Indicates the change in inductance parameter of phase m at the ith sampling point in the data window; Indicates the change in inductance parameter of phase n at the ith sampling point in the data window; The Euclidean distance of the resistor is the same as the above formula, only need to L Replace with R , calculated as follows: ； in, Represents the Euclidean distance between phases m and n at the nth sampling point; the values ​​of m and n are A, B, or C; D Indicates the total number of data points in the sliding data window. Considering the requirement of feature smoothness, the total number of data points in one cycle is taken; n Indicates the n sampling points; N 0 is the sampling point number after the criterion is started; Indicates the resistance parameter change of phase m at the i-th sampling point in the data window; Indicates the resistance parameter change of phase n at the ith sampling point in the data window, According to the characteristics of symmetrical and asymmetrical faults, different coherence change criteria are constructed; The Euclidean distance of the three-phase parameters of symmetrical faults is small and theoretically approaches 0. The criterion shown in formula (9) is constructed to identify symmetrical faults. （9）； in, D Rmn The three-phase Euclidean distance representing the resistance parameter; subscript m and n Phase identification mark, the value is A, B or C; N 0 is the sampling point number after the criterion is started; For the i The resistance Euclidean distance of the sampling points; M is the average value calculation window; D set is the symmetric fault discrimination threshold; The calculation of inductance parameters is similar to that of R Replace with L ; The calculation is as follows: ； in, D Lmn The three-phase Euclidean distance representing the inductance parameter; m and n Phase identification mark, the value is A, B or C; N 0 is the sampling point number after the criterion is started; For the i The inductance Euclidean distance of the sampling points; M is the average value calculation window; D set is the symmetric fault discrimination threshold; Asymmetric faults always have a trend of synchronous changes between two phases. Therefore, the synchronous change criterion for asymmetric faults is: （11）； In formula (11), represents the Euclidean distance of resistance between phases m and n; represents the resistance Euclidean distance between the x and y phases; represents the inductive Euclidean distance between phases m and n; Represents the inductive Euclidean distance between phases x and y; if mn is AB, then xy is BC or CA; a and b Adaptive comparison coefficients for resistance and inductance respectively; (2) Fault phase selection criteria: Based on the above-identified coherent phase distinction, the fault phase selection criterion is constructed: （13）； in, express p Phase inductance parameter change; express m Phase inductance parameter change; express n Phase inductance parameter change; Represents a logical "AND" operation.

Images