Fault classification and pole selection method for flexible DC power transmission system

Abstract

The invention relates to a fault classification and pole selection method for a flexible direct-current power transmission system, which realizes fault classification and pole selection according to the difference of high-frequency components of bus and line fault voltages when a fault occurs, and comprises the following steps: S1, when the system has a fault, collecting bus and line fault voltage data of a single converter station, constructing a data set and dividing the data set into a training set and a test set; s2, decomposing the fault voltage by adopting discrete wavelet transform, and extracting bus and line fault voltage characteristics; s3, calculating a fine composite multi-scale discrete entropy; s4, taking an extreme random tree model as a data mining model, and taking the calculated RCMDE value as the input of an ET classifier; after data in the training set and the test set are processed through the steps S2-S3, the extreme random tree model is trained and evaluated, and the trained extreme random tree model is obtained and can be used for fault classification and pole selection. According to the method, the fault detection precision and reliability can be improved, and accurate classification and pole selection of different fault types are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection of direct current transmission lines, and in particular to a fault classification and pole selection method for a flexible direct current transmission system based on the fusion of wavelet walk entropy and extreme random tree. Background Art

[0002] HVDC transmission has the advantages of flexible power control, larger power supply capacitor, lower transmission line loss, and easier interaction with renewable energy. However, overhead transmission lines are susceptible to extreme weather events and have poor operating environment. The complex structure and flow of multi-terminal HVDC architectures make it difficult to apply traditional AC system fault classification methods. When a fault occurs on the DC side, the sharp increase in fault current threatens the stability of the DC transmission line and damages the health of the equipment. Due to the high cost of HVDC circuit breakers, a large number of circuit breakers cannot be deployed in long-distance HVDC transmission. Therefore, it is necessary to adopt a fast and accurate fault detection method to reduce the requirements for DC circuit breakers and damage to transmission lines.

[0003] At present, the research on fault detection of flexible DC transmission lines can be divided into the following two aspects: one is to detect faults based on the amplitude change and increment of fault voltage or current. This method relies on expert experience to set the threshold, resulting in low accuracy of the fault detection method; the other is to use corresponding mathematical methods to extract and analyze the amplitude, polarity and duration of transient traveling waves of the faulty part to detect faults. However, this method still requires manual determination of the threshold after signal analysis, and the accuracy is also affected by subjective factors. Summary of the invention

[0004] The object of the present invention is to provide a method for fault classification and pole selection in a flexible direct current transmission system, which can improve the accuracy and reliability of fault detection and achieve accurate classification and pole selection for different fault types.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: a method for fault classification and pole selection of a flexible DC transmission system, which realizes fault classification and pole selection according to the difference in high-frequency components of busbar and line fault voltages when a fault occurs, and includes the following steps:

[0006] S1. When a system fault occurs, the bus and line fault voltage data of a single converter station are collected, including bus voltage and line voltage, and a data set is constructed and divided into a training set and a test set;

[0007] S2. Decompose the fault voltage using discrete wavelet transform (DWT) to extract bus and line fault voltage characteristics;

[0008] S3, based on the feature quantity extracted by discrete wavelet transform, calculate the refined composite multi-scale discrete entropy RCMDE;

[0009] S4. Use the extreme random tree model as the data mining model, and use the RCMDE value calculated in step S3 as the input of the ET classifier; for the data in the training set and the test set, after being processed through steps S2-S3, input the data into the ET classifier, train and evaluate the extreme random tree model, and obtain the trained extreme random tree model, which can be used for fault classification and pole selection.

[0010] Furthermore, in step S1, bus and line fault voltage data within 3 ms after the fault begins are collected.

[0011] Furthermore, in step S2, the discrete wavelet function is expressed as:

[0012]

[0013] Where a is the proportional factor, b is the displacement coefficient, j and k are both integers, and t is time; then the DWT of the fault voltage u(t) is:

[0014]

[0015] in, express conjugation of;

[0016] The db4 wavelet is used to perform a 2-layer DWT decomposition of the fault voltage. The fault voltage signal is passed through a low-pass filter and a high-pass filter, and the output is downsampled to obtain the approximate coefficient A1 and the detail coefficient D1. The approximate coefficient is further decomposed by a 2-layer DWT to obtain the detail coefficient D of the required scale. j .

[0017] Furthermore, in step S3, the detail coefficient extracted by discrete wavelet transform is used as the feature quantity, which is a single variable signal with a length of L: x = {x1, x2, ..., x L}, multi-scale coarse-graining is performed on x, then the f-th coarse-grained sequence is:

[0018]

[0019] in, f=1,2,...,τ, τ is the scale factor;

[0020] Use the following normal distribution function to Mapping to y = {y i, |i=1,2…,F}:

[0021]

[0022] Among them, y∈(0,1), y i is the element in y, xi is the i-th sequence of the input univariate signal, σ and μ are the standard deviation and mean of the normal distribution function, respectively;

[0023] Apply a linear algorithm to map y to the range [1,2,…,c]:

[0024]

[0025] Among them, c is the number of categories, and round() means taking an integer;

[0026] Under the given embedding dimension m and time delay d, the coarse-grained sequence is reconstructed in phase space to obtain the embedding vector

[0027]

[0028] Where m represents the embedding dimension and d is the time delay;

[0029] Define walking mode if but The corresponding walking mode is Each pattern consists of c data, and each value has m possible values, so the corresponding distribution entropy patterns have c m kind;

[0030] For each scaling factor, the RCMDE obtained from the fault voltage is expressed as:

[0031]

[0032] in, Represents the embedding vector Mapping to walking mode The number of is a coarse-grained sequence The probability value of max is the maximum scale factor.

[0033] Compared with the prior art, the present invention has the following beneficial effects: the present invention provides a method for fault classification and pole selection of flexible direct current transmission system based on wavelet walk entropy and extreme random tree fusion, so as to solve the problems of low accuracy, insensitivity to high resistance faults and poor pole selection performance of the existing fault detection method in MTDC system, improve the accuracy and reliability of fault detection, realize accurate classification and pole selection of different fault types, and have strong anti-interference ability, anti-transition resistance ability and low dependence on sampling frequency. The method introduces RCMDE to extract features of detail coefficients obtained by DWT. The method can quantify the nonlinearity and uncertainty of fault signals, and is less affected by mutation signals. It considers the amplitude relationship between amplitudes, has better stability and faster calculation speed, and can more accurately reflect the differences between different fault types. In addition, the method uses ET intelligent classifier for fault classification and pole selection. ET has good performance and strong generalization ability, and only needs a small number of samples to obtain high diagnostic accuracy. By combining with DWT and RCMDE, accurate classification and pole selection of bus and line faults in MTDC system are realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of a method implementation of an embodiment of the present invention;

[0035] Figure 2 Schematic diagram of a four-terminal MMC high-voltage direct current transmission system according to an embodiment of the present invention;

[0036] Figure 3 Schematic diagram of the propagation process of traveling waves in an embodiment of the present invention;

[0037] Figure 4 is a fault transient equivalent circuit diagram when a bipolar short circuit fault occurs in the OHL24 line in an embodiment of the present invention;

[0038] Figure 5 is a frequency response diagram of the voltage between the current limiting reactors in an embodiment of the present invention;

[0039] Figure 6 Schematic diagram of the discrete wavelet decomposition feature extraction process in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0041] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0042] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0043] like Figure 1 As shown, this embodiment provides a method for fault classification and pole selection of a flexible direct current transmission system, and realizes fault classification and pole selection according to the difference in high-frequency components of busbar and line fault voltage when a fault occurs. The method adopts a discrete wavelet transform (DWT) and a refined composite multiscale dispersion entropy (RCMDE) and an extremely randomized tree (ET) fusion fault classification method for multi-terminal high-voltage direct current (MTDC) system fault classification and pole selection, including busbar fault identification. Wavelet transform is used to analyze the detail coefficients of fault voltage in the time-frequency domain. In order to extract multi-scale fault features from the fault voltage detail coefficients, RCMDE is introduced into the signal analysis of the MTDC system for the first time. Entropy is used to measure the complexity and uncertainty of the signal as the feature input of ET, which greatly reduces the dimensionality of data training. The method draws on the mature experience of fault signal analysis, and the fault features can better reflect the differences between different fault types. At the same time, combined with machine learning algorithms, the classification and pole selection of bus faults and line faults in the four-port DC transmission architecture can be considered. The method specifically includes the following steps:

[0044] S1. When a system fault occurs, the bus and line fault voltage data of a single converter station are collected, including bus voltage and line voltage, and a data set is constructed and divided into a training set and a test set.

[0045] S2. Use discrete wavelet transform (DWT) to decompose the fault voltage and extract the bus and line fault voltage characteristics.

[0046] S3. Based on the feature quantity extracted by discrete wavelet transform, the refined composite multi-scale discrete entropy RCMDE is calculated.

[0047] S4. Use the extreme random tree model as the data mining model, and use the RCMDE value calculated in step S3 as the input of the ET classifier; for the data in the training set and the test set, after being processed through steps S2-S3, input the data into the ET classifier, train and evaluate the extreme random tree model, and obtain the trained extreme random tree model, which can be used for fault classification and pole selection.

[0048] The following is a further detailed description of the relevant contents involved in this method.

[0049] A four-port MMC-HVDC system model corresponding to the MTDC system is constructed. The model includes four 300km overhead transmission lines (OHL 12 ,OHL 13 ,OHL 24 ,OHL 34 ), install current limiting reactors (L xy ) is used to limit the fault current, and a hybrid high-voltage DC circuit breaker is installed in the system to achieve fault isolation. The parameters of each converter station (MMC1-MMC4) are set as follows Figure 2 shown.

[0050] When a transmission line or bus fails, a fault wave will be generated and propagated with the sharp voltage drop at the fault point. When the wave encounters an impedance change point, refraction and reflection of the wave will occur. Figure 3 shown.

[0051] The natural frequency of the fault voltage traveling wave can be expressed as:

[0052]

[0053] in, represents the speed of the nth traveling wave, x fi Represents the distance between the measurement point and the fault point.

[0054] The line voltage at a frequency of 2ωf can be expressed as:

[0055]

[0056] Among them, F1(ω) and F2(ω) are impedance coefficients determined by the line boundary conditions, β(ω) is the line attenuation coefficient, and σ(ω) is the phase shift coefficient of the line. The propagation speed of the traveling wave is defined by the moving speed of the phase point, that is:

[0057] ωt+σ(ω)x fi =const (3)

[0058] Differentiating equation (3) yields the traveling wave velocity:

[0059]

[0060] According to formula (1), the smaller the distance between the fault point and the measurement point, the greater the natural frequency of the traveling wave. According to formula (4), traveling wave components of different frequencies have different propagation speeds. Therefore, when a fault occurs, the original fault voltage contains many components of different frequencies, and the different components are superimposed on the reflected traveling wave, so there is a high-frequency transient component at the measurement point.

[0061] The fault types of MTDC systems are mainly divided into single-pole grounding faults and double-pole short-circuit faults.

[0062] Bipolar short circuit fault is the most serious fault. Taking bipolar short circuit fault as an example, the transient equivalent circuit of the fault when bipolar short circuit fault occurs in OHL24 line is as follows: Figure 4 shown.

[0063] When F 24 When a bipolar short-circuit fault occurs in the converter, the equivalent inductance and capacitance of the converter are discharged, which is equivalent to superimposing a negative voltage source -V at the fault point. f .

[0064] MMC i L in the station eqi and C eqi It can be expressed as:

[0065]

[0066] Where i = 1, 2, 3, 4; L0 is MMC i The bridge arm reactor of the station; C0 is the submodule capacitance; N is the number of submodules in each MMC bridge arm; Figure 3 As shown, the current limiting reactor L dc24 The ratio of the DC voltage on the upper line side to the DC voltage on the bus side is:

[0067]

[0068] From formula (6) and Figure 5 It can be seen that as the frequency increases, the value of lg(H(jω)) approaches 2, and the current limiting reactor L dc24 The high-frequency fault voltage component can be effectively reduced.

[0069] According to the above analysis, fault classification and pole selection can be realized according to the difference in high-frequency components of busbar and line fault voltages when a fault occurs.

[0070] Step S1: Fault data collection and processing

[0071] When a fault occurs in the system, the bus and line fault voltage data of a single converter station are collected, including the bus voltage (V dc2, E1, E2) and line voltage (V dc24 ), the sampling frequency is 100kHz. The data within 3ms after the fault occurs are collected to form a data set and the data set is divided into a training set (70%) and a test set (30%).

[0072] Step S2: Discrete Wavelet Transform (DWT) decomposition of fault voltage

[0073] When a fault occurs, the line voltage and bus voltage will change dramatically. Different fault types cause different voltage drop amplitudes and voltage oscillation intensities. Since the fault voltage contains abundant transient high-frequency components, wavelet transform is a suitable solution to enhance the high-frequency characteristics of burst signals in the time domain and frequency domain. In addition, compared with the continuous wavelet function, the DWT method significantly reduces the amount of data. Therefore, this method uses discrete wavelet transform to decompose the fault voltage and extract the bus and line fault voltage characteristics.

[0074] The discrete wavelet function can be expressed as:

[0075]

[0076] Where a is the proportional factor, b is the displacement coefficient, j and k are both integers, and t is time. Then the DWT of the fault voltage u(t) is:

[0077]

[0078] in, express The conjugation of.

[0079] The db4 wavelet is used to perform a 2-layer DWT decomposition of the fault voltage. The fault voltage signal is passed through a low-pass filter and a high-pass filter, and the output is downsampled to obtain the approximate coefficient A1 and the detail coefficient D1. The approximate coefficient is then decomposed in the same steps, that is, a further 2-layer DWT decomposition is performed to obtain the detail coefficient D of the required scale. j (This article takes D1 and D2), such as Figure 6 shown.

[0080] Original signal x hf (kT) The high-frequency components D1 and D2 can be extracted from the fault traveling wave.

[0081] In order to obtain the time-frequency characteristics of the fault voltage in each frequency band, db4 wavelet is more suitable for power transient analysis than other wavelet decompositions. This paper uses the db4 wavelet of discrete wavelet transform to perform multi-layer decomposition of the fault voltage and conducts time-frequency analysis in different frequency ranges.

[0082] Step S3: Calculate the refined composite multiscale discrete entropy (RCMDE)

[0083] In order to analyze the characteristics of high-frequency components under MTDC fault conditions, entropy is a good solution to quantify the nonlinearity and uncertainty of fault signals, and has been widely used in feature extraction and fault diagnosis. RCMDE is an entropy calculation method that can measure the complexity and uncertainty of input signals while reducing the dimensionality of training data. The detail coefficients D1 and D2 obtained by DWT are taken as feature quantities, and they are assumed to be univariate signals of length.

[0084] First, the feature quantity is the series detail coefficients D1 and D2 obtained by DWT transformation of the fault voltage. It is a univariate signal of length L: x = {x1, x2, …, x L}, multi-scale coarse-graining is performed on x, then the f-th coarse-grained sequence is:

[0085]

[0086] in, f=1,2,...,τ, τ is the scale factor.

[0087] Using the normal distribution function as shown in formula (10), Mapping to y = {y i, |i=1,2…,F}:

[0088]

[0089] Among them, y∈(0,1), y i is the element in y, x i is the i-th sequence of the input univariate signal, σ and μ are the standard deviation and mean of the normal distribution function, respectively.

[0090] Apply a linear algorithm to map y to the range [1,2,…,c]:

[0091]

[0092] Where c is the number of categories and round() means taking an integer. This step can also be achieved by other linear and nonlinear mapping techniques. Under the given embedding dimension m and time delay d, the coarse-grained sequence is reconstructed in phase space to obtain the embedding vector

[0093]

[0094] Here, m represents the embedding dimension and d is the time delay.

[0095] Define walking mode if but The corresponding walking mode is Each pattern consists of c data, and each value has m possible values, so the corresponding distribution entropy patterns have c m kind.

[0096] For each scaling factor, the RCMDE obtained from the fault voltage can be expressed as:

[0097]

[0098] in, Represents the embedding vector Mapping to walking mode The number of is a coarse-grained sequence The probability value of max is the maximum scale factor.

[0099] Step S4: Fault classification and pole selection

[0100] There are many methods for data mining, such as support vector machine (SVM), random forest (RF), decision tree (DT), etc. Extreme random tree (ET) has good performance and strong generalization ability. In addition, by using only a small number of samples, high diagnostic accuracy can be obtained. Therefore, this method uses the extreme random tree model as the data mining model.

[0101] Data within 3ms after the start of the fault is collected to construct two data sets, one is the training set and the other is the test set. The training set accounts for 70% of the data set, and the test set accounts for 30% of the data set. For the data in the training set and the test set, the discrete wavelet transform is performed on the fault voltage waveform to obtain the detail coefficients D1 and D2 with time-frequency characteristics. The RCMDE values ​​of D1 and D2 are used as the input of the ET classifier to train and evaluate the extreme random tree model to obtain a trained extreme random tree model. Using the trained model, it is possible to make predictions based on new data points to obtain the final fault classification and pole selection results.

[0102] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0103] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0104] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0105] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0106] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

Claims

1. A method for fault classification and pole selection in a flexible direct current transmission system, characterized in that: Fault classification and pole selection are realized according to the difference of high frequency components of busbar and line fault voltage when a fault occurs, including the following steps: S1. When a system fault occurs, the bus and line fault voltage data of a single converter station are collected, including bus voltage and line voltage, and a data set is constructed and divided into a training set and a test set; S2. Decompose the fault voltage using discrete wavelet transform (DWT) to extract bus and line fault voltage characteristics; S3, based on the feature quantity extracted by discrete wavelet transform, calculate the refined composite multi-scale discrete entropy RCMDE; S4. Use the extreme random tree model as the data mining model, and use the RCMDE value calculated in step S3 as the input of the ET classifier; for the data in the training set and the test set, after being processed through steps S2-S3, input the data into the ET classifier, train and evaluate the extreme random tree model, and obtain the trained extreme random tree model, which can be used for fault classification and pole selection.

2. A method for fault classification and pole selection of a flexible direct current transmission system according to claim 1, characterized in that: In step S1, bus and line fault voltage data are collected within 3 ms after the fault begins.

3. A method for fault classification and pole selection of a flexible direct current transmission system according to claim 1, characterized in that: In step S2, the discrete wavelet function is expressed as: Where a is the proportional factor, b is the displacement coefficient, j and k are both integers, and t is time; then the DWT of the fault voltage u(t) is: in, express conjugation of; The db4 wavelet is used to perform a 2-layer DWT decomposition of the fault voltage. The fault voltage signal is passed through a low-pass filter and a high-pass filter, and the output is downsampled to obtain the approximate coefficient A1 and the detail coefficient D1. The approximate coefficient is further decomposed by a 2-layer DWT to obtain the detail coefficient D of the required scale. j .

4. A method for fault classification and pole selection of a flexible direct current transmission system according to claim 1, characterized in that: In step S3, the detail coefficient extracted by discrete wavelet transform is used as the feature quantity, which is a single variable signal with a length of L: x = {x1, x2, ..., x L }, multi-scale coarse-graining is performed on x, then the f-th coarse-grained sequence is: in, f=1,2,...,τ, τ is the scale factor; Use the following normal distribution function to Mapping to y = {y i, |i=1,2…,F}: Among them, y∈(0,1), y i is the element in y, x i is the i-th sequence of the input univariate signal, σ and μ are the standard deviation and mean of the normal distribution function, respectively; Apply a linear algorithm to map y to the range [1,2,…,c]: Among them, c is the number of categories, and round() means taking an integer; Under the given embedding dimension m and time delay d, the coarse-grained sequence is reconstructed in phase space to obtain the embedding vector Where m represents the embedding dimension and d is the time delay; Define walking mode if but The corresponding walking mode is Each pattern consists of c data, and each value has m possible values, so the corresponding distribution entropy patterns have c m kind; For each scaling factor, the RCMDE obtained from the fault voltage is expressed as: in, Represents the embedding vector Mapping to walking mode The number of is a coarse-grained sequence The probability value of τ max is the maximum scale factor.

Images