Fault location method for offshore wind power dc collection line based on traveling wave multi-scale gradient analysis

Abstract

The application discloses a kind of offshore wind power dc collection line fault location methods based on travelling wave multiscale gradient analysis, S1.distinguishes that fault occurs;S2.sampling the voltage and current of positive pole and negative pole of fault collection line bus side, calculate 1 mode voltage reverse travelling wave;S3.based on mathematical morphology algorithm, 1 mode voltage reverse travelling wave is filtered, and filter signal is calculated;S4.the multiscale gradient of filter signal is calculated, and the transient abrupt point of waveform is identified, and the arrival time and wave head polarity information of fault voltage travelling wave are extracted;S5.calculating fault distance.The method of the application realizes fault location using fault voltage data in a short time after fault, and can quickly determine the specific fault position of collection line, and has important significance for accelerating fault clearing and guaranteeing the safe and stable operation of offshore wind power system.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of electric power systems, and relates to the technical field of relay protection of offshore wind power DC collection lines, in particular to an offshore wind power DC collection line fault positioning method based on traveling wave multi-scale gradient analysis. BACKGROUND

[0002] In the face of increasingly severe energy crisis and climate change, all countries in the world have proposed development plans to change from fossil energy to renewable energy, and have clearly stated that renewable energy generation is the future development direction. Wind energy, as one of the most mature and widely used renewable energies, is expected to become a pillar energy industry in China. Offshore wind power has the advantages of high wind energy quality, large single machine capacity, and no occupation of land resources, and can be consumed nearby, so it has broad development prospects. Nowadays, due to the increasing tension of offshore wind farm sites, in order to adapt to the trend of deep-sea development, the full DC transmission system of offshore wind power based on DC collection and DC transmission technology has emerged as the times require, which reduces the number of system commutations and effectively reduces the system energy consumption, and is of great research significance.

[0003] The line fault of the full DC transmission system of offshore wind power is mostly permanent fault, and the fault needs to be promptly investigated and cleared to restore the normal operation and power transmission of the system. According to the fault positioning principle, the existing DC grid fault positioning methods can be divided into three categories: 1) transient characteristic method: the mathematical relationship between the fault location and the fault characteristics after the fault occurs is obtained by modeling the fault line, and the fault positioning is realized. Among them, the single-ended method has the advantages of good economy and fast distance measurement speed, but it is difficult to eliminate the influence of transition resistance on the fault distance measurement result; the double-ended method has the advantages of reliable principle and high distance measurement accuracy, but it is restricted by communication conditions and cost; 2) active injection method: an additional circuit or a power electronic converter is used to inject a detection signal into the fault circuit, and the fault distance is measured according to the characteristics of the collected signal. This method requires additional signal injection equipment, which increases the construction cost of the DC distribution network; 3) traveling wave method: when a fault occurs in the DC line, a fault traveling wave will be generated at the fault point and propagate to both sides of the line, and the fault location is realized by extracting the amplitude, polarity and frequency of the transient traveling wave. However, for short lines, the difference in wave head arrival time is small, making it more difficult to identify the wave head, and putting forward very high requirements for the wave head identification algorithm. In summary, the sending end of the full DC system of offshore wind power is a multi-terminal parallel network with distributed converters in series, the equivalent capacitance of the converter is in series with the line impedance, forming a complex and non-uniform line topology, which makes it difficult to locate the fault of the collection line. The existing positioning methods have poor adaptability, and it is urgent to propose a high-precision collection line fault positioning method to determine the specific fault location of the collection line, speed up fault clearance, and ensure the safe and stable operation of the offshore wind power system. SUMMARY

[0004] To solve the problems in the prior art, the offshore wind power DC collection line fault positioning method based on traveling wave multi-scale gradient analysis is provided, the method does not have the communication synchronization problem, the fault positioning method has high positioning accuracy, strong robustness to noise and interference, and good engineering application value.

[0005] To achieve the above object, the technical scheme adopted by the present application is as follows:

[0006] A kind of offshore wind power DC collection line fault positioning method based on traveling wave multi-scale gradient analysis, comprising the following steps:

[0007] Step one: the protection element discriminates that fault occurs;

[0008] After the offshore wind power DC collection network fault occurs, the protection element starts, discriminates the fault collection line and fault type;

[0009] Step two: sampling the voltage and current of positive pole and negative pole of fault collection line bus side, calculating 1 module voltage reverse traveling wave;

[0010] According to formula (1), 1 module voltage u1 (k) and 1 module current i1 (k) are calculated;

[0011]

[0012] In the formula, k is sampling serial number, N is the total number of sampling window, u p (k) is positive pole voltage, u n (k) is negative pole voltage, i p (k) is positive pole current, i n (k) is negative pole current;

[0013] According to formula (2), 1 module voltage reverse traveling wave u b1 (k) is calculated;

[0014] u b1 (k)=[u1(k)-i1(k)·Z C ] / 2 (2)

[0015] In the formula, Z C is wave impedance;

[0016] Step three: based on mathematical morphology algorithm, 1 module voltage reverse traveling wave is filtered, and filtered signal is calculated;

[0017] Select flat structure element g (x), and its definition domain is D (g)={0,1,…,P};

[0018] According to formula (3), 1 module voltage reverse traveling wave is dilated and corroded;

[0019]

[0020] where 1-modulus voltage counter wave u b1 (k) is defined in D(u b1 ) = {0, 1, …, N}, denotes dilation operation,! denotes erosion operation;

[0021] According to formula (4), open operation and closed operation are performed;

[0022]

[0023] where denotes open operation, · denotes closed operation;

[0024] According to formula (5), the filter signal Filter(u b1 (k)) is calculated;

[0025]

[0026] where OC and CO respectively denote open-closed operation and closed-open operation, and the results are calculated according to formula (6);

[0027]

[0028]

[0029] Step four: calculate the multi-scale gradient of the filter signal, identify the transient mutation point of the waveform, and extract the arrival time and wave head polarity information of the fault voltage traveling wave;

[0030] Based on the Laplacian pyramid algorithm, the filter signal is divided into multiple resolution scales, and Filter(u b1 (k)) is the bottom layer signal G0of the Gaussian pyramid, and the lth layer signal G l of the Gaussian pyramid decomposition is calculated according to formula (7);

[0031]

[0032] where N g is the layer number of the top layer of the Gaussian pyramid, m is the independent variable of the window function, and the value is m = -2, -1, 0, 1, 2, is the window function,

[0033] The lth layer extension sequence G

[0034]

[0035] where

[0036] The l-th layer signal LP of the Laplacian pyramid decomposition is calculated according to formula (9) l and the top layer Laplacian pyramid signal

[0037]

[0038] In the formula, is the (l+1)-th layer expansion sequence, is the top layer signal of the Gaussian pyramid;

[0039] The gradient SG of each layer of the Laplacian pyramid sequence of the 1-modulus voltage traveling wave is calculated according to formula (10) l ;

[0040] SG l (k) = (LP l * ker)(k) (10)

[0041] In the formula, the convolution operation kernel ker of the Sobel gradient operator is [-1 0 1];

[0042] The multi-scale gradient MSG(k) of the 1-modulus voltage traveling wave is calculated according to formula (11)

[0043]

[0044] In the formula, Expand represents the up-sampling operation;

[0045] According to the calculation result of the multi-scale gradient, the transient abrupt point of the voltage waveform is identified, and the sampling sequence number of the arrival time of the measured initial traveling wave is extracted, denoted as k r0 ; the sampling sequence number of the arrival time of the second traveling wave is extracted, and if the polarity of the wave head of the second traveling wave is the same as that of the initial traveling wave, it is denoted as k r1 , and if the polarity of the wave head of the second traveling wave is opposite to that of the initial traveling wave, it is denoted as k t1 ;

[0046] Step five: calculate the fault distance;

[0047] The fault distance x is calculated according to formula (12)

[0048]

[0049] Wherein, v is the traveling wave speed, L is the full length of the fault collection line, Δt same is the arrival time difference of the initial two consecutive traveling waves of the same polarity, and Δt diff is the arrival time difference of the initial two consecutive traveling waves of opposite polarity, and the fault distance x is calculated according to formula (13)

[0050]

[0051] In the formula, T s is a sampling step length;

[0052] Preferably, the total number of sampling points N of the sampling window is 500, the layer number N of the top layer of the Gaussian pyramid is 2, the sampling step length T is 1e-6, and the sampling step length T of the Gaussian pyramid is 1e-6. g is 2, and the sampling step length T s is 1e-6.

[0053] Compared with the prior art, the present application has the following advantages:

[0054] (1) Since fault positioning can be realized based on fault voltage data at the collection bus, only a traveling wave collection device needs to be installed on one side of the collection line, so that the fault positioning method of the present application has lower communication and measurement costs;

[0055] (2) Since linear filtering operation based on mathematical morphology is performed before extracting traveling wave information, Gaussian noise and impulse noise that have a greater impact on signal mutation point recognition can be effectively filtered out, so that the fault positioning method of the present application has strong anti-interference ability;

[0056] (3) Since the multi-scale gradient algorithm retains the rich details of sequence mutation points at small scales while also taking into account the strong anti-noise interference ability of large scales, the fault positioning method of the present application has strong positioning accuracy;

[0057] In summary, the present application can quickly determine the specific fault position of the collection line, can speed up the fault recovery speed, and has important significance for ensuring the safe and stable operation of the offshore wind power system. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 is a topology diagram of an offshore wind power DC collection and transmission system suitable for the method of the present application.

[0059] Figure 2 is a flowchart of the method of the present application.

[0060] Fig. 3(a) is a line positive electrode voltage signal under a collection line fault.

[0061] Fig. 3(b) is a line negative electrode voltage signal under a collection line fault.

[0062] Fig. 3(c) is a line positive electrode current signal under a collection line fault.

[0063] Fig. 3(d) is a line negative electrode current signal under a collection line fault.

[0064] Figure 4 is a line 1-mode voltage counter-traveling wave shape under a collection line fault.

[0065] Figure 5 is the filtered 1-modulus voltage backwave signal.

[0066] Figure 6 is the multi-scale gradient calculation result. DETAILED DESCRIPTION

[0067] The application will be further described in detail below in combination with the drawings and examples.

[0068] As shown in Figure 1 , at 0 ms, the positive ground fault (denoted as fault F) occurs at the positive pole of the gathering line A at a distance of 7.5 km from the gathering bus, and the fault resistance is set to 0.1Ω.

[0069] When the fault F occurs, the line voltage and current waveforms measured by the sampling element are shown in FIG. 3(a), FIG. 3(b), FIG. 3(c) and FIG. 3(d).

[0070] A fault locating method for offshore wind power DC gathering lines based on multi-scale gradient analysis of traveling wave, the flowchart is shown in Figure 2 , including the following steps:

[0071] Step one: after the fault of the offshore wind power DC gathering network occurs, the protection element is started, and it is determined that the fault gathering line is line A and the fault type is single-pole fault.

[0072] Step two: the positive pole voltage u p (k) and the negative pole voltage u n (k) of the sampling line, the positive pole current i p (k) and the negative pole current i n (k) are measured, respectively as shown in FIG. 3(a), FIG. 3(b), FIG. 3(c) and FIG. 3(d).

[0073] The 1-modulus voltage u1(k) and the 1-modulus current i1(k) are calculated according to formula (1).

[0074] The 1-modulus voltage backwave u b1 (k) is calculated according to formula (2), and the result is shown in Figure 4 .

[0075] Step three: the 1-modulus voltage backwave is filtered based on the mathematical morphology algorithm, and the filtered signal is calculated according to formula (5), and the result is shown in Figure 5 .

[0076] Step four: the multi-scale gradient of the filtered voltage signal is calculated according to formula (11), and the result is shown in FIG. 6.

[0077] According to the multi-scale gradient calculation result, the transient mutation point of the voltage waveform is identified, and the arrival time and the wave head polarity information of the fault voltage traveling wave are extracted, wherein the arrival sampling sequence number corresponding to the initial traveling wave is 63, and the arrival sampling sequence number of the second same polarity traveling wave is 185;

[0078] Step five: the arrival time difference of the initial two continuous same polarity traveling waves is calculated according to formula (13) as 122 μs;

[0079] The traveling wave velocity v is 122.89 km / ms, and the fault distance x is calculated according to formula (12) as 7.496 km;

[0080] It can be seen from the data in step five that the protection method of the application can reliably and quickly identify the fault position, and has high positioning accuracy.

Claims

1. A fault location method for offshore wind power DC collection lines based on traveling wave multi-scale gradient analysis, characterized in that: Includes the following steps: Step 1: The protection element detects a malfunction; After a fault occurs in the offshore wind power DC collection network, the protection components are activated to identify the faulty collection line and the type of fault. Step 2: Sample the voltage and current of the positive and negative terminals on the fault collection line bus side, and calculate the 1-mode voltage reverse traveling wave; Calculate the first-mode voltage u1(k) and the first-mode current i1(k) according to equation (1); In the formula, k is the sampling sequence number, N is the total number of samples in the sampling window, and u p (k) is the positive voltage, u n (k) is the negative terminal voltage, i p (k) represents the positive current, i n (k) represents the negative electrode current; According to equation (2), calculate the mode 1 voltage reverse traveling wave u. b1 (k); at b1 (k)=[u1(k)-i1(k)·Z C ] / 2 (2) In the formula, Z C Wave impedance; Step 3: Filter the 1-mode voltage inverse traveling wave based on mathematical morphology algorithm and calculate the filtered signal; Choose a flat structuring element g(x), and denote its domain as D(g) = {0, 1, ..., P}; According to equation (3), expansion and erosion calculations are performed on the reverse traveling wave of the 1-mode voltage; (u b1 ! g)(k)=min{u b1 (k+x)-g(k)|(k+x)∈D(u b1 In the formula, k∈D(g), the mode 1 voltage inverse traveling wave u b1 The domain of (k) is D(u b1 )={0,1,…,N}, `!` represents the expansion operation, and `!` represents the erosion operation. According to equation (4), perform opening and closing operations; In the formula, · represents the opening operation, and · represents the closing operation; According to equation (5), the filtered signal Filter(u) is calculated. b1 (k)); In the formula, OC and CO represent the opening-closing operation and the closing-opening operation, respectively, and the results are calculated according to formula (6); Step 4: Calculate the multi-scale gradient of the filtered signal, identify the transient change points of the waveform, and extract the arrival time and wavefront polarity information of the fault voltage traveling wave. The filtered signal is divided into multiple resolution scales based on the Laplace pyramid algorithm, denoted as Filter(u). b1 (k) represents the bottom layer signal G0 of the Gaussian pyramid. The l-th layer signal G of the Gaussian pyramid decomposition is calculated according to equation (7). l ; In the formula, N g Let m be the layer number of the top layer of the Gaussian pyramid, and m be the independent variable of the window function, taking values ​​of m = -2, -1, 0, 1, 2. For window functions, The l-th layer extended sequence is calculated according to equation (8). In the formula, The signal LP of the l-th layer of the Laplace pyramid decomposition is calculated according to equation (9). l and the top Laplace pyramid signal In the formula, For the (l+1)th layer extended sequence, This is the signal at the top of the Gaussian pyramid; The gradient SG of each layer of the Laplace pyramid sequence of the mode-1 voltage back-traveling wave is calculated according to equation (10). l ; SG l (k)=(LP l *ker)(k) (10) In the formula, the convolution kernel of the Sobel gradient operator is ker = [-1 0 1]; The multiscale gradient MSG(k) of the mode-1 voltage back-traveling wave is calculated according to equation (11); In the formula, Expand represents the upsampling operation; Based on the multi-scale gradient calculation results, transient abrupt changes in the voltage waveform are identified, and the arrival time sampling number of the measured initial traveling wave is extracted and denoted as k. r0 Extract the arrival time sampling number of the second traveling wave. If the polarity of the second traveling wave's wavefront is the same as that of the initial traveling wave, then record it as k. r1 If the polarity of the second traveling wavefront is opposite to that of the initial traveling wave, it is denoted as k. t1 ; Step 5: Calculate the distance to the fault; Calculate the fault distance x according to equation (12); Where v is the traveling wave velocity, L is the total length of the fault collection line, and Δt is the arrival time difference between two initial consecutive traveling waves of the same polarity. same The arrival time difference Δt between the initial two consecutive traveling waves of opposite polarity diff Calculated according to formula (13); In the formula, T s This is the sampling step size.

2. The method for fault location of offshore wind power DC collection lines based on traveling wave multi-scale gradient analysis according to claim 1, characterized in that: The total number of sampling points N in the sampling window is 500, and the layer number N of the top layer of the Gaussian pyramid is... g The value is 2, and the sampling step size T s The value is 1e -6 .

Images