New energy grid-connected tie line permanent fault identification method based on wavelet packet decomposition

Abstract

The invention relates to the technical field of high-voltage power transmission network fault identification, and discloses a new energy grid-connected tie line permanent fault identification method based on wavelet packet decomposition, which comprises the following steps: acquiring starting protection action information of a circuit breaker of a power transmission line in real time; acquiring port voltage information of a fault phase, and sequentially dividing the port voltage information of the circuit breaker after A power frequency periods of starting protection action are adopted backwards according to a set time window and a step length to obtain a plurality of pieces of port voltage waveform data divided by the time window; performing multi-layer wavelet packet decomposition on each segment of port voltage waveform data to obtain a plurality of frequency band results; respectively calculating Euclidean distances of the reference result and the experimental result based on the same frequency band; and judging whether the fault phase is an instantaneous fault or a permanent fault according to the Euclidean distance. The safe and stable operation level of the power transmission network is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of high-voltage transmission network fault identification technology, specifically to a method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition. Background Technology

[0002] In power systems, transmission lines are the most frequently faulted components. Based on operational experience, on high-voltage overhead lines of 110kV and above systems with high grounding currents, over 70% of faults are single-phase grounding short circuits, and more than 80% of these are transient faults. Therefore, when a single-phase grounding short circuit occurs, tripping only the faulty phase and reclosing it is one of the important measures to ensure the safe and stable operation of the power system. However, this also brings some adverse effects to the power system: the impact of reclosing on a permanent fault is even greater than that of a short circuit under normal conditions. If the fault is not cleared in time, it may lead to equipment damage, electrical fires, and a decrease in grid stability. Therefore, pre-judging whether a fault is permanent or transient to determine whether to reclose (reclosing is permissible for transient faults, but not for permanent faults) can reduce the impact on the system and has significant practical implications for the power system. Currently, after the circuit breakers at both ends of a line trip, the voltage at the line terminals is weak and the transient characteristics are complex, making it difficult to accurately extract effective fault characteristics, resulting in a low accuracy rate in fault nature judgment. Summary of the Invention

[0003] This invention provides a method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition. By comparing the time-frequency domain characteristics of the phase-end voltage before and after the arc extinction of a transient fault in the transmission network, digital processing and waveform morphology methods are used to determine the nature of the fault. This method can quickly and accurately identify permanent faults in the transmission network, which is of great practical significance for maintaining the safe and stable operation of the system and solving the problem of fault nature identification during reclosing of transmission lines under the current background.

[0004] This invention is achieved through the following technical solution:

[0005] A method for identifying permanent faults in renewable energy grid-connected tie lines based on wavelet packet decomposition includes:

[0006] Real-time acquisition of activation and protection information of circuit breakers on transmission lines;

[0007] The port voltage information of the faulty phase of the circuit breaker that took the start protection action information is obtained, and the port voltage information of the circuit breaker after taking the start protection action A power frequency cycles is divided sequentially with a set time window and step size to obtain a number of port voltage waveform data divided by time window, where A is a positive integer;

[0008] Each segment of the port voltage waveform data is subjected to multi-level wavelet packet decomposition to obtain several frequency band results corresponding to each time window within a set frequency range;

[0009] Preprocessing the results of several frequency bands for the first time window yields a baseline result. Preprocessing the results of several frequency bands for all time windows other than the first time window yields the experimental results for the corresponding time windows. The Euclidean distance between the baseline result and the experimental result based on the same frequency band is calculated.

[0010] The Euclidean distance is used to determine whether the fault phase is a transient or permanent fault.

[0011] As an optimization, the port voltage information of the circuit breaker after taking the start-up protection action for A power frequency cycles is divided sequentially according to a set time window and step size, thereby obtaining several port voltage data sequences divided by time windows, specifically:

[0012] Obtain the port voltage information of the circuit breaker after three power frequency cycles of taking the start protection action;

[0013] The time window is set to the size of one power frequency cycle, and the step size is the size of one power frequency cycle. The port voltage information after three power frequency cycles of starting the protection action is divided sequentially to obtain port voltage waveform data for several time windows before reclosing.

[0014] As an optimization, the specific process of performing multi-level wavelet packet decomposition on each segment of the port voltage waveform data to obtain several frequency band results based on each time window within a set frequency range is as follows:

[0015] The sampling frequency is set to B Hz, and the port voltage waveform data is sampled to obtain a port voltage sequence composed of several discrete sampling points, where B is a positive integer;

[0016] The port voltage sequence is decomposed into 32 bandwidths using the db6 wavelet basis with a 5-level wavelet packet decomposition. The sub-signals, wherein one of the sub-signals contains There are 1 sampling point, and C is one power frequency cycle;

[0017] The sub-signal with the smallest value among the first 8 frequency bands is obtained as the frequency band result for the corresponding frequency band.

[0018] As an optimization, the specific formula for wavelet packet decomposition is as follows:

[0019]

[0020] Where i is the node number, j is the decomposition level, g(n) represents the low-frequency component signal obtained by low-pass filtering coefficient decomposition of the port voltage sequence, h(n) represents the high-frequency component signal obtained by high-pass filtering coefficient decomposition of the port voltage sequence, and g(n) and h(n) are a pair of orthogonal mirror filters. These are the wavelet basis functions constructed based on the two-scale equation. s(t) represents the amplitude at the k-th sampling point under the j-th layer decomposition and the i-th node, s(t) represents the port voltage waveform data, and n represents the index of the filter coefficient.

[0021] As an optimization, the specific formula for the wavelet basis function is as follows:

[0022]

[0023] Where i is the node number, j is the decomposition level, and when i = 0... Represents the scaling function, when i=1, Describing wavelet functions, These are wavelet basis functions constructed based on the two-scale equation.

[0024] As an optimization, the preprocessing of several frequency band results for the first time window to obtain the baseline results, and the preprocessing of several frequency band results for all time windows other than the first time window to obtain the experimental results for the corresponding time windows are as follows:

[0025] The amplitude of the sampling points in several sub-signals of the first time window is normalized based on the divided frequency bands to obtain the reference result corresponding to each frequency band;

[0026] The amplitudes of the sampling points in several sub-signals corresponding to time windows other than the first time window are normalized based on the divided frequency bands to obtain the experimental results corresponding to each frequency band.

[0027] As an optimization, determining whether the fault phase is a transient or permanent fault based on the Euclidean distance specifically involves:

[0028] The sum of the Euclidean distances between the benchmark results and each experimental result within the set frequency range and based on the same frequency band is calculated respectively. If the sum of the Euclidean distances is greater than a threshold, the fault phase is determined to be a transient fault, and the fault arc is extinguished within the time window corresponding to the sum of the Euclidean distances being greater than the threshold; otherwise, the fault phase is determined to be a permanent fault.

[0029] As an optimization, the specific formula for calculating the sum of the Euclidean distances between the benchmark result and one of the experimental results within the set frequency range and based on the same frequency band is as follows:

[0030]

[0031] Where D represents the sum of Euclidean distances, m represents the number of frequency bands within the set frequency range, and d k x represents the Euclidean distance between the baseline result and the experimental result in the k-th frequency band. i y represents the amplitude of the i-th sampling point in the benchmark result. i Let represent the amplitude of the i-th sampling point in the experimental results, and n represent the number of sampling points.

[0032] This invention also discloses a permanent fault identification system for renewable energy grid-connected tie lines based on wavelet packet decomposition, comprising:

[0033] The data acquisition module is used to collect real-time information on the activation and protection actions of circuit breakers in transmission lines.

[0034] The segmentation module is used to obtain the port voltage information of the faulty phase of the circuit breaker that has taken the start protection action information, and to segment the port voltage information of the circuit breaker after taking the start protection action A power frequency cycles in sequence with a set time window and step size, so as to obtain a number of port voltage waveform data segmented by time window, where A is a positive integer;

[0035] The decomposition module is used to perform multi-level wavelet packet decomposition on each segment of the port voltage waveform data to obtain several frequency band results based on each time window within a set frequency range.

[0036] The calculation module is used to preprocess the results of several frequency bands in the first time window to obtain the benchmark results, preprocess the results of several frequency bands in all time windows other than the first time window to obtain the experimental results of the corresponding time windows, and calculate the Euclidean distance between the benchmark results and the experimental results based on the same frequency band.

[0037] The judgment module is used to determine whether the fault phase is a transient fault or a permanent fault based on the Euclidean distance.

[0038] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the steps of the aforementioned method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition.

[0039] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0040] In this invention, by comparing the time-frequency domain characteristics of the phase-end voltage before and after the arc extinction of a transient fault in the power transmission network, digital processing and waveform morphology methods are used to determine the nature of the fault. This enables the rapid and accurate identification of the arc extinction moment and the permanent fault identification of the power transmission network. This has important practical significance for maintaining the safe and stable operation of the system and solving the problem of fault nature identification during reclosing of transmission lines under the current background.

[0041] In this invention, signal processing algorithms are used to improve the ability to identify and extract fault signal characteristics. By forming a stable time-frequency domain signal, a complete line fault nature discrimination process is formed to solve the problem of difficulty in identifying the nature of single-phase grounding faults, thereby effectively improving the safe and stable operation level of the power transmission network. Attached Figure Description

[0042] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0043] Figure 1 This is a flowchart of the identification method in an embodiment of the present invention;

[0044] Figure 2 This is a flowchart illustrating another form of the identification method in an embodiment of the present invention.

[0045] Figure 3 This is a schematic diagram of the 5-layer wavelet packet decomposition of a signal in an embodiment of the present invention;

[0046] Figure 4 This is a flowchart of wavelet packet decomposition in an embodiment of the present invention;

[0047] Figure 5 This is a flowchart for determining the nature of a fault in an embodiment of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0049] To address the challenges posed by increased fault voltage harmonic content due to new energy grid connection, which hinders fault identification methods in traditional power supply scenarios, and the complex transient characteristics resulting from circuit breaker tripping at both ends of the line, making accurate extraction of effective fault features difficult and leading to low accuracy in fault identification, this embodiment 1 provides a method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition. Figures 1-5 As shown, it includes the following steps:

[0050] S1. Real-time acquisition of the activation and protection action information of circuit breakers on power transmission lines.

[0051] The circuit breaker trips and disconnects the faulty phase after a fault occurs in a certain phase of the line, which serves as the starting criterion for fault nature identification. If the circuit breaker of a certain phase activates its protection action signal, the next step of fault nature identification is carried out; otherwise, it is not included in the fault nature identification scheme.

[0052] S2. Obtain the port voltage information of the faulty phase of the circuit breaker that has taken the start protection action information, and divide the port voltage information of the circuit breaker after taking the start protection action A power frequency cycles in sequence with a set time window and step size to obtain a number of port voltage waveform data divided by time window, where A is a positive integer.

[0053] In this embodiment, A is 3, and the set time window and step size are one power frequency cycle, that is, the specific process of S2 is as follows:

[0054] S2.1 Obtain the port voltage information of the circuit breaker after three power frequency cycles of taking the start protection action;

[0055] S2.2 Set the time window to the size of one power frequency cycle and the step size to the size of one power frequency cycle. Divide the port voltage information after three power frequency cycles of starting the protection action into several time windows before reclosing to obtain the port voltage waveform data.

[0056] One power frequency cycle is 20ms. Therefore, if the circuit breaker initiates a protection action (single-phase trip), the port voltage of the faulty phase after the circuit breaker initiates a single-phase trip is acquired. The waveform of the port voltage approximately three cycles after the circuit breaker protection action is divided into time windows of 20ms each, with a step size of 20ms, until the reclosing time is reached, which is approximately 1 second. Therefore, in this embodiment, the port voltage information is divided into 497 time windows with a step size of 20ms each, up to 1 second. In other words, the voltage waveform is captured starting three power frequency cycles after the circuit breaker protection action, and each 20ms segment of the waveform is captured sequentially to obtain the port voltage information for each sequence.

[0057] S3. Perform multi-level wavelet packet decomposition on each segment of the port voltage waveform data to obtain several frequency band results corresponding to each time window within a set frequency range.

[0058] In some embodiments, the specific process of S3 is as follows:

[0059] S3.1 Set the sampling frequency to B Hz, sample the port voltage waveform data to obtain a port voltage sequence composed of several discretized sampling points, where B is a positive integer;

[0060] S3.2. Perform 5-level wavelet packet decomposition on the port voltage sequence using the db6 wavelet basis to obtain 32 bandwidths corresponding to the port voltage sequence. The sub-signals, wherein one of the sub-signals contains There are 1 sampling point, and C is one power frequency cycle;

[0061] S3.3 Obtain the frequency band result of the sub-signal with the smallest value among the first 8 frequency bands.

[0062] Among them, utilizing Figure 3 The wavelet packet decomposition flowchart shown demonstrates a 5-level wavelet packet decomposition of the signal using the db6 wavelet basis. With a sampling frequency of 1600Hz, 32 sub-signals with a bandwidth of 25Hz are obtained. The frequency ranges of the first 8 sub-signals are 0-25Hz, 25-50Hz, 50-75Hz, 75-100Hz, 100-125Hz, 125-150Hz, 150-175Hz, and 175-200Hz, respectively.

[0063] Figure 3 The reason it's 0-800Hz is because, according to the Nyquist sampling theorem, the sampling frequency f... s The highest frequency at which the signal can be effectively analyzed is determined to be f. s / 2. Therefore, when the sampling frequency is 1600Hz, the effective frequency range is 0-800Hz. Thus, 0-800Hz is obtained by 5-layer wavelet packet decomposition. 5 = There are 32 sub-signals, each with a bandwidth of 800 / 32 = 25, and each sub-signal has 0.02 * 1600 = 32 sampling points.

[0064] To determine the number of sampling points contained in a sub-signal, it is necessary to comprehensively analyze the sampling frequency, time window length, and characteristics of wavelet packet decomposition. Let's assume the sampling frequency of the original signal is f. s (In this embodiment, f) s =1600Hz), and the time window length is T (e.g., T = 20ms in this embodiment), then: the number of original sampling points within a single time window is: N 原始 =f s *T, which represents 32 sampling points.

[0065] Wavelet packet decomposition is an orthogonal transform in the time-frequency domain, and the number of effective sampling points of the signal remains unchanged before and after decomposition (without considering boundary effects). That is, if the original time window has N sampling points, the number of sampling points of each sub-signal after decomposition is still N. Therefore, if the original time window has 32 sampling points, after 5 layers of wavelet packet decomposition, the number of sampling points of each sub-signal is still 32.

[0066] The essence of wavelet packet decomposition is frequency domain subdivision (dividing the frequency range of the original signal into multiple sub-bands), rather than "time domain downsampling". Therefore, the number of sampling points for each sub-signal is the same as that of the original signal (32 in total); the only difference is that the original signal contains information across the entire frequency band, while the sub-signal only contains information from a narrow frequency band (such as 0-25Hz, 25-50Hz, etc.).

[0067] In some embodiments, the specific formula for wavelet packet decomposition is as follows:

[0068]

[0069] Where i is the node number, j is the decomposition level, g(n) represents the low-frequency component signal obtained by low-pass filtering coefficient decomposition of the port voltage sequence, h(n) represents the high-frequency component signal obtained by high-pass filtering coefficient decomposition of the port voltage sequence, and g(n) and h(n) are a pair of orthogonal mirror filters. These are the wavelet basis functions constructed based on the two-scale equation. s(t) represents the amplitude at the k-th sampling point under the j-th layer decomposition and the i-th node, s(t) represents the port voltage waveform data, and n represents the index of the filter coefficient.

[0070] It should be noted that s(t) in the formula is the original signal (port voltage waveform data), which is a continuous voltage signal. In practice, wavelet packet decomposition is used to process discrete signals, that is, to process the port voltage sequence.

[0071] In some embodiments, the specific formula for the wavelet basis function is as follows:

[0072]

[0073] Where i is the node number, j is the decomposition level, and when i = 0... Represents the scaling function, when i=1, Describing wavelet functions, These are wavelet basis functions constructed based on the two-scale equation.

[0074] S4. Preprocess the results of several frequency bands for the first time window to obtain the baseline results, and preprocess the results of several frequency bands for all time windows other than the first time window to obtain the experimental results for the corresponding time windows. Calculate the Euclidean distance between the baseline results and the experimental results based on the same frequency band.

[0075] In some embodiments, the specific process of preprocessing the frequency band results of the first time window to obtain the baseline results, and preprocessing the frequency band results corresponding to all time windows other than the first time window to obtain the experimental results of the corresponding time windows is as follows:

[0076] S4.1. The amplitude of the sampling points in several sub-signals of the first time window is normalized based on the divided frequency bands to obtain the reference result corresponding to each frequency band.

[0077] S4.2. Normalize the amplitude of the sampling points in several sub-signals corresponding to the time windows other than the first time window based on the divided frequency bands to obtain the experimental results corresponding to each frequency band.

[0078] In other words, the first window is selected as the reference voltage, and the decomposition result of the reference voltage is compared with the decomposition result of each subsequent window. Specifically, the waveform decomposition results of the two voltage sequences in the same frequency band after wavelet packet decomposition are obtained, and the differences between the two after normalization are compared.

[0079] The reference voltage and window voltage are decomposed into corresponding sub-signals using wavelet packet decomposition. The waveforms of the sub-signals in each frequency band are normalized for comparison. Specifically:

[0080]

[0081] In the formula, A i A represents the amplitude at the i-th sampling point in a certain frequency band. max A min These represent the maximum and minimum amplitudes at the sampling points in this frequency band, respectively, A′ i This represents the normalized amplitude of the i-th sampling point.

[0082] S5. Determine whether the faulty phase is a transient or permanent fault based on the Euclidean distance.

[0083] In some embodiments, S5 specifically refers to:

[0084] The sum of the Euclidean distances between the benchmark results and each experimental result within the set frequency range and based on the same frequency band is calculated respectively. If the sum of the Euclidean distances is greater than a threshold, the fault phase is determined to be a transient fault, and the fault arc is extinguished within the time window corresponding to the sum of the Euclidean distances being greater than the threshold; otherwise, the fault phase is determined to be a permanent fault.

[0085] In some embodiments, the threshold value is 3.

[0086] In other words, in this embodiment, the Euclidean distance between the waveform formed by the amplitude of the sampling point corresponding to the non-first time window in each frequency band and the reference waveform (the waveform formed by the amplitude of the sampling point corresponding to the first time window) is calculated. If the calculated distance is greater than the threshold value of 3, it indicates that it is a transient fault and the fault arc is extinguished within the time of the window. If it is always less than 3, the fault is considered to be a permanent fault.

[0087] Using the time-frequency distribution difference between the reference voltage waveform and the subsequent window waveform as a fault characteristic, the sum of the Euclidean distances between the two waveforms within the same 200Hz frequency band is calculated. The specific formula is as follows:

[0088]

[0089]

[0090] Where D represents the sum of Euclidean distances, m represents the number of frequency bands within the set frequency range, and here m is 8, which is the sum of the Euclidean distances between the experimental results and the benchmark results across 8 frequency bands within a 200Hz frequency range, and d k x represents the Euclidean distance between the baseline result and the experimental result in the k-th frequency band. i y represents the amplitude of the i-th sampling point in the benchmark result. i This represents the amplitude of the i-th sampling point in the experimental results, that is, x i and y i The values ​​represent the decomposed amplitudes of the reference waveform and the subsequent window waveform at the same sampling point in the same frequency band. n represents the number of sampling points.

[0091] x here i y i That is, the aforementioned A i .

[0092] like Figure 2 If the calculated Euclidean distance D is greater than the threshold value of 3, it indicates a transient fault, and the fault arc is extinguished within the time of the window being drawn; if D is always less than 3, the fault is considered a permanent fault.

[0093] Specifically, the following explanation will be further illustrated by building a new energy grid-connected transmission line model in the electromagnetic transient simulation software PSCAD:

[0094] The tie line is 60 km long, with a positive sequence impedance Z1 = 0.01 + j0.41 Ω / km and a unit positive sequence capacitance to ground X'. C1 = -j3.5886×10 8 Ω·km, zero-sequence impedance per unit length Z0=0.33+j1.32Ω / km, zero-sequence capacitance per unit length to ground X' C0 = -j5.1175×10 8 The fault time is set to Ω·km, with a fault occurrence time of 1s. The circuit breaker protection trips 10ms after the fault occurs, and voltage waveform is captured starting 60ms later. When the calculated Euclidean distance begins to exceed the threshold value of 3, it indicates a transient fault, and the fault arc extinguishes within the time frame of the window. If it remains less than 3, the fault is considered permanent. As shown in Table 1, the arc extinguishing time for a transient fault is 1.22s.

[0095] Table 1

[0096]

[0097] Different fault locations and transition resistances were set for both cases with and without parallel reactors to verify the accuracy of the proposed fault nature judgment method. The judgment results are shown in Tables 2 and 3.

[0098] Table 2 Configuration of Parallel Reactors

[0099]

[0100]

[0101] Table 3. No parallel reactors are configured.

[0102]

[0103] Based on the fault nature judgment results in Tables 2 and 3, the analysis of the fault nature judgment results shows that this method can correctly judge the fault nature in the line, and is not affected by the fault location, transition resistance, and parallel reactor, thus having a very good fault nature judgment capability.

[0104] Example 2 discloses a permanent fault identification system for new energy grid-connected tie lines based on wavelet packet decomposition, used to execute the permanent fault identification method for new energy grid-connected tie lines based on wavelet packet decomposition in Example 1, including:

[0105] The data acquisition module is used to collect real-time information on the activation and protection actions of circuit breakers in transmission lines.

[0106] The segmentation module is used to obtain the port voltage information of the faulty phase of the circuit breaker that has taken the start protection action information, and to segment the port voltage information of the circuit breaker after taking the start protection action A power frequency cycles in sequence with a set time window and step size, so as to obtain a number of port voltage waveform data segmented by time window, where A is a positive integer;

[0107] The decomposition module is used to perform multi-level wavelet packet decomposition on each segment of the port voltage waveform data to obtain several frequency band results based on each time window within a set frequency range.

[0108] The calculation module is used to preprocess the results of several frequency bands in the first time window to obtain the benchmark results, preprocess the results of several frequency bands in all time windows other than the first time window to obtain the experimental results of the corresponding time windows, and calculate the Euclidean distance between the benchmark results and the experimental results based on the same frequency band.

[0109] The judgment module is used to determine whether the fault phase is a transient fault or a permanent fault based on the Euclidean distance.

[0110] Example 3 also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the permanent fault identification method for new energy grid-connected tie lines based on wavelet packet decomposition described in Example 1.

[0111] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for identifying permanent faults in renewable energy grid-connected tie lines based on wavelet packet decomposition, characterized in that, include: Real-time acquisition of activation and protection information of circuit breakers on transmission lines; The port voltage information of the faulty phase of the circuit breaker that took the start protection action information is obtained, and the port voltage information of the circuit breaker after taking the start protection action A power frequency cycles is divided sequentially with a set time window and step size to obtain a number of port voltage waveform data divided by time window, where A is a positive integer; Each segment of the port voltage waveform data is subjected to multi-level wavelet packet decomposition to obtain several frequency band results corresponding to each time window within a set frequency range; Preprocessing the results of several frequency bands for the first time window yields a baseline result. Preprocessing the results of several frequency bands for all time windows other than the first time window yields the experimental results for the corresponding time windows. The Euclidean distance between the baseline result and the experimental result based on the same frequency band is calculated. The Euclidean distance is used to determine whether the fault phase is a transient or permanent fault.

2. The method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition according to claim 1, characterized in that, The port voltage information of the circuit breaker after taking the start-up protection action for A power frequency cycles is divided sequentially according to a set time window and step size to obtain several port voltage data sequences divided by time windows, specifically: Obtain the port voltage information of the circuit breaker after three power frequency cycles of taking the start protection action; The time window is set to the size of one power frequency cycle, and the step size is the size of one power frequency cycle. The port voltage information after three power frequency cycles of starting the protection action is divided sequentially to obtain port voltage waveform data for several time windows before reclosing.

3. The method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition according to claim 1, characterized in that, The specific process of performing multi-level wavelet packet decomposition on each segment of the port voltage waveform data to obtain several frequency band results corresponding to each time window within a set frequency range is as follows: The sampling frequency is set to B Hz, and the port voltage waveform data is sampled to obtain a port voltage sequence composed of several discrete sampling points, where B is a positive integer; The port voltage sequence is decomposed into 32 bandwidths using the db6 wavelet basis with a 5-level wavelet packet decomposition. A sub-signal of Hz, wherein one of the sub-signals contains There are 1 sampling point, and C is one power frequency cycle; The sub-signal with the smallest value among the first 8 frequency bands is obtained as the frequency band result for the corresponding frequency band.

4. The method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition according to claim 3, characterized in that, The specific formula for wavelet packet decomposition is as follows: ； in, i For node number, j The number of decomposition layers, This represents the low-frequency component signal obtained by decomposing the port voltage sequence using low-pass filtering coefficients. This represents the high-frequency component signal obtained by decomposing the port voltage sequence using high-pass filtering coefficients. and It is a pair of orthogonal mirror filters. , These are the wavelet basis functions constructed based on the two-scale equation. This represents the amplitude at the k-th sampling point under the j-th layer decomposition and the i-th node. This represents the port voltage waveform data, where n represents the index of the filter coefficient.

5. The method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition according to claim 4, characterized in that, The specific formula for the wavelet basis function is as follows: ； in, i For node number, j The number of decomposition layers, hour, Represents the scaling function. hour, Describing wavelet functions, , These are wavelet basis functions constructed based on the two-scale equation.

6. The method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition according to claim 4, characterized in that, The specific process of preprocessing the frequency band results of the first time window to obtain the baseline results, and preprocessing the frequency band results of all other time windows to obtain the experimental results of the corresponding time windows is as follows: The amplitude of the sampling points in several sub-signals of the first time window is normalized based on the divided frequency bands to obtain the reference result corresponding to each frequency band; The amplitudes of the sampling points in several sub-signals corresponding to time windows other than the first time window are normalized based on the divided frequency bands to obtain the experimental results corresponding to each frequency band.

7. The method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition according to claim 1, characterized in that, The specific method for determining whether a fault is transient or permanent based on the Euclidean distance is as follows: The sum of the Euclidean distances between the benchmark results and each experimental result within the set frequency range and based on the same frequency band is calculated respectively. If the sum of the Euclidean distances is greater than a threshold, the fault phase is determined to be a transient fault, and the fault arc is extinguished within the time window corresponding to the sum of the Euclidean distances being greater than the threshold; otherwise, the fault phase is determined to be a permanent fault.

8. The method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition according to claim 7, characterized in that, The specific formula for calculating the sum of the Euclidean distances between the benchmark result and one of the experimental results within the set frequency range and based on the same frequency band is as follows: ； ； Where D represents the sum of Euclidean distances, and m represents the number of frequency bands within the set frequency range. This represents the Euclidean distance between the baseline result and the experimental result in the k-th frequency band. This represents the amplitude of the i-th sampling point in the benchmark result. Let represent the amplitude of the i-th sampling point in the experimental results, and n represent the number of sampling points.

9. A permanent fault identification system for new energy grid-connected tie lines based on wavelet packet decomposition, characterized in that, include: The data acquisition module is used to collect real-time information on the activation and protection actions of circuit breakers in transmission lines. The segmentation module is used to obtain the port voltage information of the faulty phase of the circuit breaker that has taken the start protection action information, and to segment the port voltage information of the circuit breaker after taking the start protection action A power frequency cycles in sequence with a set time window and step size, so as to obtain a number of port voltage waveform data segmented by time window, where A is a positive integer; The decomposition module is used to perform multi-level wavelet packet decomposition on each segment of the port voltage waveform data to obtain several frequency band results based on each time window within a set frequency range. The calculation module is used to preprocess the results of several frequency bands in the first time window to obtain the benchmark results, preprocess the results of several frequency bands in all time windows other than the first time window to obtain the experimental results of the corresponding time windows, and calculate the Euclidean distance between the benchmark results and the experimental results based on the same frequency band. The judgment module is used to determine whether the fault phase is a transient fault or a permanent fault based on the Euclidean distance.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for identifying permanent faults in new energy grid-connected tie lines based on wavelet packet decomposition as described in any one of claims 1 to 8.

Images