Direct current system fault positioning method based on improved frequency conversion signal injection method

Abstract

The invention provides a DC system fault positioning method based on an improved variable frequency signal injection method, and the method comprises the steps: firstly, injecting AC signals of different frequencies into a bus of a DC system, and collecting the response current of each branch after the AC signals are injected; then, screening the response current through a signal processing algorithm to obtain a response characteristic quantity, and calculating a response ratio of each branch according to the response characteristic quantity; and finally, based on the response ratio of each branch, judging the branch with the fault through a numerical comparison method. According to the method, the frequency conversion signals are injected at the two ends, branch feedback detection and corresponding signal algorithms are combined, and the feedback result is processed by utilizing ratios, so that the operation steps are simplified, the requirements of detection equipment are reduced, the influence of interference signals on the final judgment result can be reduced through algorithm combination processing, and the overall detection accuracy of the method is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical data processing, and in particular to a DC system fault locating method based on an improved variable frequency signal injection method. Background Art

[0002] The DC system in a substation provides power for control equipment, relay protection devices, and other components, ensuring the normal operation of the power system. Due to the complex topology and diverse operating conditions of DC systems, component failures, secondary circuit failures, and ground faults are common. Ground faults are the most damaging of these.

[0003] When a DC system experiences a ground fault, the system can still operate normally. However, if the faulty branch is not repaired promptly, a second ground fault could generate an excessive short-circuit current, causing the fuse to trip. This could result in a loss of power to the DC system, and render control equipment and relay protection devices inoperable, leading to further losses. Therefore, when a DC system experiences a ground fault, repair it as soon as possible to avoid two simultaneous ground faults.

[0004] Traditional methods for locating ground faults include signal injection and leakage current. The leakage current method is unaffected by the distributed capacitance of the DC system relative to the earth. However, when the insulation conditions of the positive and negative poles are comparable, the leakage current method cannot detect simultaneous grounding of the positive and negative poles. The signal injection method is easily affected by factors such as distributed capacitance, ring network harmonics, and ripple voltage, resulting in significant interference in the measured signal and large errors in the detection results.

[0005] The aforementioned issues result in low accuracy in existing ground fault location detection. Furthermore, methods like the traveling wave method require more complex monitoring equipment to accurately capture the traveling wave, the bridge method can suffer from large errors for long-distance branches, and insulation monitoring devices also have accuracy issues. Consequently, existing technologies also suffer from high detection equipment costs, low execution efficiency, and difficulty in quickly locating the branch where the fault occurs in complex circuit networks. Summary of the Invention

[0006] In response to the problems existing in the prior art, the present application provides a method for locating branch faults in a DC system of a substation based on an improved variable frequency signal injection method. By injecting variable frequency signals at both ends of the busbar and combining branch detection and signal processing algorithms, the method can eliminate the influence of interference factors such as distributed capacitance and ring network harmonics on the branch response current, thereby improving the detection accuracy of the variable frequency signal injection method.

[0007] To achieve the above objectives, the technical solutions adopted in this application are as follows: The present application provides a DC system fault location method based on an improved variable frequency signal injection method, which mainly includes the following steps: Injecting a first AC signal into the positive busbar of the DC system and collecting a first response current of each branch; After collecting the first response current of each branch, injecting a second AC signal into the negative bus of the DC system, and collecting the second response current of each branch; The frequency of the first AC signal is different from the frequency of the second AC signal; The first response current and the second response current in each branch collected are decomposed into multiple modal components respectively by using a modal decomposition algorithm, and the modal components whose center frequencies are consistent with the frequencies of the injected AC signals are screened out and recorded as response modal components; Solving the amplitude of the response modal component based on a signal amplitude algorithm to obtain a response current amplitude; Solve the ratio of the two response current amplitudes in each branch to obtain the response ratio of each branch; Based on the response ratio of each branch, the faulty branch is determined by numerical comparison method.

[0008] Optionally, the modal decomposition algorithm is a successive variational modal decomposition algorithm; and the signal amplitude algorithm is a Hilbert transform.

[0009] Optionally, the response modal component solution includes the following steps: The Fourier transform formula of the current order mode is updated in the Fourier domain to obtain the current order Fourier transform formula; Calculate and update the center frequency of the current-order mode according to the current-order Fourier transform formula; The Lagrange multipliers are updated by the dual ascent method; Repeat the update operation until the internal convergence condition is met and output the current modal component; The modal order is increased, and the update operation and convergence judgment are continued until the external convergence conditions are met to complete the decomposition of the branch response current.

[0010] Optionally, solving the amplitude of the response modal component includes the following steps: A Hilbert transform process is performed on the response modal component to construct an analytical signal, and the amplitude of the response modal component is calculated based on the analytical signal.

[0011] Optionally, solving the amplitude of the response modal component further includes the following steps: performing a Hilbert transform on the response modal component through a convolution operation to obtain a Hilbert transform result; The analytical signal is a signal function constructed by the response modal component and the Hilbert transform result; The magnitude of the response modal component is calculated based on the signal function.

[0012] Optionally, the frequency of the first AC signal is greater than the frequency of the second AC signal.

[0013] Optionally, the solution of the response ratio includes the following steps: Calculating, in each branch, a response current amplitude corresponding to the first AC signal and a response current amplitude corresponding to the second AC signal; The ratio of the two response current amplitudes in the branch is solved to obtain the response ratio of the branch.

[0014] Optionally, the response ratio is the amplitude of a subsequent response current divided by the amplitude of a previous response current in the time sequence; Determining the faulty branch includes the following steps: Sort the branches according to the numerical values ​​of the response ratios to obtain a branch sequence; A branch with a response ratio greater than a threshold or the largest value in the branch sequence is selected as a branch with a fault.

[0015] Optionally, the following steps are also included: In the branch with the fault, dividing the branch into a plurality of branch sections based on branch nodes; Inject branch AC signals of different frequencies at both ends of each branch section; The response current of each branch section after the branch AC signal is injected is collected in sequence, and the step of determining the branch with a fault is repeated to determine the branch section with a fault.

[0016] Optionally, the branch AC signal includes a third AC signal and a fourth AC signal; the response current following the branch AC signal includes a third response current and a fourth response current; injecting the third AC signal into the upstream node of the branch section, and collecting a third response current of the downstream node of the branch section; After the third response current is collected, injecting the fourth AC signal into the downstream node of the branch section, and collecting the fourth response current of the upstream node of the branch section; The frequency of the third AC signal is greater than the frequency of the fourth AC signal.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The method of the present invention injects variable frequency signals at both ends, combines branch feedback detection and corresponding signal algorithms, and utilizes ratio processing to feedback results. On the one hand, the operation steps are simplified and the demand for detection equipment is reduced. On the other hand, interference components such as harmonics and ripple voltage in the measured signal are filtered out through algorithm combination, thereby alleviating the influence of distributed capacitance. In addition, the influence of interference signals on the final judgment result is reduced through ratio processing of the variable frequency feedback signal, thereby improving the overall detection accuracy of the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 is a flow chart of a method according to a specific embodiment of the present invention; Figure 2 This is a DC system model of a substation in a specific embodiment of the present invention; Figure 3 This is a normal branch model of the DC system of a substation in a specific embodiment of the present invention; Figure 4 This is a branch model of a substation DC system with a ground fault in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0020] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0021] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0022] It is worth noting that the methods used in the present invention are all conventional methods unless otherwise specified; the raw materials and devices used are all conventional commercially available products, and their sources are not specifically limited unless otherwise specified.

[0023] This embodiment provides a DC system fault location method based on an improved variable frequency signal injection method, such as Figure 1 As shown, it mainly includes the following steps: When a fault is detected in the DC system of the substation, AC signals of different frequencies are first injected into the positive bus and negative bus of the DC system, respectively. In this embodiment, a double-ended single signal injection is performed, and thus the AC signals of different frequencies include a first AC signal and a second AC signal.

[0024] Specifically, a signal generator is used to inject a first AC signal into the positive bus of the DC system. In this embodiment, the frequency is selected as AC signal As the first AC signal, the first response current of each branch is measured using a mutual inductor .

[0025] In the above example, n Represents the sequence number, such as the superscript of the response current represents the branch sequence number, and there are n Branch road.

[0026] After collecting the first response current of each branch, the second AC signal is injected into the negative bus of the DC system, that is, Figure 2 As shown, a signal generator is used to inject a frequency of AC signal As the second AC signal, the second response current of the branch is measured using a transformer. .

[0027] The frequency of the first AC signal is greater than the frequency of the second AC signal. .

[0028] The response current is processed by a signal processing algorithm; wherein the signal processing algorithm includes a modal decomposition algorithm, which decomposes the collected response current into multiple modal components, and screens out the modal components whose center frequencies are consistent with the frequencies of the injected corresponding AC signals, which are recorded as response modal components.

[0029] Specifically, the modal decomposition algorithm uses successive variational modal decomposition (SVMD) to decompose the current signal of each branch, including the following steps: Initialize the iteration counter , represents the number of modal components that have been extracted; Repeat the following iterative process until the internal convergence condition is met and output the current modal component : (a) For the current modal component , update its Fourier transform in the Fourier domain , the specific update formula is as follows: ; Where, For the The first mode Fourier transform after the update; is the Fourier transform of the input AC signal; is the Lagrange multiplier Fourier transform of It is The first mode The updated center frequency; is the balance parameter; is the frequency variable in the Fourier domain, It is i The center frequency of the modal component.

[0030] (b) Update the current modal component Center frequency , the specific update method is as follows: ; Where, For the The first mode The updated center frequency; For the The first mode The Fourier transform after the update.

[0031] (c) Update the Lagrange multiplier by the dual ascent method , the specific update method is as follows: ; Where, It is Updated Lagrange multiplier Fourier transform of is the dual ascending step length; is the Fourier transform of the input signal; For the The first mode Fourier transform after the update; For the Fourier transform of the first mode; is the Fourier transform of the unprocessed signal.

[0032] Update iteration counter , to increase the modal order; specifically, repeat the above steps, complete the update operation and convergence judgment, extract the modal components one by one until the external convergence conditions are met, and then obtain the intrinsic mode functions (IMFs) of each order; at this point, the branch response current of each branch is decomposed to obtain the set of intrinsic mode functions of each order [IMF1, IMF2, ..., IMFn].

[0033] The amplitude of the response modal component is solved based on the signal amplitude algorithm. The signal amplitude algorithm used in this embodiment is Hilbert transform. Solving the amplitude of the response modal component includes the following steps: Perform Hilbert transform on the response modal components to construct analytical signals, and then perform calculations based on the analytical signals. The specific steps are as follows: Performing Hilbert transform on the response modal component through convolution operation to obtain Hilbert transform results; The analytical signal is a signal function constructed by the response modal component and the Hilbert transform result; Calculates the magnitude of the modal components of the response based on a signal function.

[0034] Therefore, in this embodiment, the response currents of the first and second AC signals are calculated separately.

[0035] First, the Hilbert transform is used to calculate the value of each branch under the injected signal. The response current amplitude under the action of The amplitude of the branch For example, specifically: No. The response current of each branch is In the previous steps, it has been decomposed into modal components .

[0036] Select center frequency With the injection signal frequency The same first-order modal components , perform Hilbert transform on it, the specific transformation method is shown in the formula: ; Where, For the first modal component Hilbert transform; For the The modal component, whose center frequency The frequency of the injected AC signal equal; Represents the convolution operation.

[0037] Construct the analytical signal. The specific construction method is shown in the formula: ; Where, is the analytical signal obtained by Hilbert transform; For the The modal component, whose center frequency The frequency of the injected AC signal equal; For the first-order modal components Hilbert transform; Is an imaginary unit.

[0038] Thus, the amplitude function can be obtained as follows: ; Where, For the The amplitude function of the first-order mode is also the response current of the branch The amplitude of For the The modal component, whose center frequency The frequency of the injected AC signal equal; For the Hilbert transform of the modal component.

[0039] Furthermore, the Hilbert transform is used to calculate the influence of each branch on the injected signal. The response current amplitude under the action of The amplitude of the branch For example, specifically: No. The response current of each branch is In the previous steps, it has been decomposed into modal components ; Select center frequency With the injection signal frequency The same first-order modal components , perform Hilbert transform on it, the specific transformation method is shown in the formula: ; Where, For the first-order modal components Hilbert transform; For the The modal component, whose center frequency The frequency of the injected AC signal equal; Represents the convolution operation.

[0040] Construct the analytical signal. The specific construction method is shown in the formula: ; Where, is the analytical signal obtained by Hilbert transform; For the The modal component, whose center frequency The frequency of the injected AC signal equal; For the first-order modal components Hilbert transform; Is an imaginary unit.

[0041] Thus, the amplitude function can be obtained as follows: ; Where, For the The amplitude function of the first-order mode is also the response current of the branch The amplitude of For the The modal component, whose center frequency The frequency of the injected AC signal equal; For the first-order modal components Hilbert transform.

[0042] Solve the ratio of the two response current amplitudes in each branch to obtain the response ratio of each branch. In this embodiment, the ratio discrimination method first calculates the response current amplitude corresponding to the AC signal of different frequencies in each branch, that is, the amplitude of the first and second response currents in each branch, and compares the two response feature quantities in the branch in time sequence to obtain the response ratio of the branch. Among them, the time sequence comparison method is to collect and calculate the response current amplitude later. Divide by the previously collected and calculated response current amplitude .

[0043] Specifically, the Take the branch as an example, inject signals of different frequencies, and the response ratio is .

[0044] Based on the response ratios of each branch, a numerical comparison method is used to determine the faulty branch. The method of this embodiment sorts the response ratios of each branch according to the magnitude of the response current amplitude ratio to obtain a branch sequence. The branch with a response ratio greater than a threshold or the largest value in the branch sequence is selected as the faulty branch.

[0045] Specifically, the amplitude ratios of the response currents of each branch are compared. The branch with the larger amplitude ratio is the branch where the ground fault occurs. The principle is as follows, assuming there is no interference: Only when the positive busbar is injected with a frequency of After receiving the AC signal, the response current of each branch is measured using a mutual inductor. Assuming that the response current in the normal branch is , the response current in the branch where the ground fault occurs is ; Only when the negative busbar injection frequency is After receiving the AC signal, the response current of each branch is measured using a mutual inductor. Assuming that the response current in the normal branch is , the response current in the branch where the ground fault occurs is ; like Figure 3 As mentioned above, the response current in the normal branch and The amplitude ratio is calculated as follows: ; Where, The injection frequency for the positive bus is The response current in the normal branch after the AC signal; The injection frequency for the negative bus is The response current in the normal branch after the AC signal; is the capacitance of the branch to ground; is the voltage of the AC signal; is the frequency of the AC signal injected from the positive bus; is the frequency of the AC signal injected from the negative bus.

[0046] like Figure 4 As shown, the response current in the branch where the ground fault occurs is and The amplitude ratio is calculated as follows: ; Where, The injection frequency for the positive bus is The response current in the branch where the ground fault occurs after the AC signal is received; The injection frequency for the negative bus is The response current in the branch where the ground fault occurs after the AC signal is received; is the capacitance of the branch to ground; is the voltage of the AC signal; is the grounding resistance; is the frequency of the AC signal injected from the positive bus; is the frequency of the AC signal injected from the negative bus.

[0047] because ,The relationship shown in the above formula can be obtained as follows, which indicates that the branch with a larger amplitude ratio is the branch where the ground fault occurs: ; Where, The injection frequency for the positive bus is The response current in the normal branch after the AC signal; The injection frequency for the negative bus is The response current in the normal branch after the AC signal; The injection frequency for the positive bus is The response current in the branch where the ground fault occurs after the AC signal is received; The injection frequency for the negative bus is The response current in the branch where the ground fault occurs after the AC signal is received.

[0048] From this, it can be analyzed that regardless of the branch load conditions, when a ground fault occurs, the node position presents a resistance characteristic, which causes the ratio to increase abnormally. This can eliminate a large number of interference factors and conclude that there is a fault in the branch.

[0049] Accordingly, in order to simplify the judgment process, a better way is to combine the threshold judgment method, collect the response ratio of each of the above-mentioned branches under normal working conditions multiple times, and take the maximum value of the multiple test results for each branch as the threshold of the branch, and compare the response ratio obtained above with the threshold of the corresponding branch in subsequent measurements. When the response ratio is greater than the threshold, it means that there is a ground fault in the branch.

[0050] Optionally, in order to further quickly locate the fault location in the branch, you can also refer to the above method to perform preliminary location judgment. The specific steps are as follows: Through the above steps, the branch with fault is determined, and multiple branch sections are divided based on the nodes of the branch; thus, it is equivalent to a branch.

[0051] Inject branch AC signals of different frequencies at both ends of each branch section; wherein the branch AC signals include a third AC signal and the fourth AC signal , and the frequency of the third AC signal is greater than the frequency of the fourth AC signal ( ); Accordingly, the response current after the branch AC signal includes a third response current and a fourth response current.

[0052] The injection method is performed in sections and in sequence, and in each section, the third AC signal is injected into the upstream node of the branch section and the third response current of the downstream node of the branch section is collected; After the third response current is collected, a fourth AC signal is injected into the downstream node of the branch section, and a fourth response current of the upstream node of the branch section is collected.

[0053] The third and fourth response currents of each branch section after the branch AC signal is injected are collected in sequence, and the step of determining the branch with the fault is repeated. That is, in this embodiment, the response ratio of each branch section is also calculated, and the response ratios of all sections in the branch with the fault are collected. The section ratio sequence is obtained by sorting the sections from large to small according to the ratios. The section with the largest ratio can be preliminarily determined to be the branch section with the fault, and thus on-site detection is performed. If it is found after detection that there is no grounding fault, the detection is continued in sequence according to the section ratio sequence until the fault point is found.

[0054] Furthermore, in the implementation of the above segmentation, the segmentation rules on the branch load side need to exclude the load, that is, the node before the branch load to the load DC positive connection is a segment, and the load DC negative connection to the node after the load is a segment.

[0055] In the above specific embodiments, the present application can effectively reduce the impact of interference by injecting variable frequency AC signals at both ends and performing fault judgment in the form of characteristic quantity ratios, thereby eliminating the need for complex mathematical calculation steps or a large number of high-precision detection equipment, reducing investment costs, and improving system response speed, thereby quickly troubleshooting and reducing the occurrence of multiple grounding faults; at the same time, at the algorithm level, successive variational mode decomposition (SVMD) and Hilbert transform are used to filter out interference components such as harmonics and ripple voltage in the measured signal, alleviate the influence of distributed capacitance, and further improve the detection accuracy of the variable frequency injection method; further, by utilizing the concept of equivalent branch circuits, two-level rapid fault location and judgment is achieved, which also saves costs and can reduce the workload of maintenance personnel.

[0056] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions of the technical solution of the present invention by ordinary technicians in this field do not deviate from the essence and scope of the technical solution of the present invention.

Claims

1. A DC system fault location method based on an improved variable frequency signal injection method, characterized in that: The steps include: Injecting a first AC signal into the positive busbar of the DC system and collecting a first response current of each branch; After collecting the first response current of each branch, injecting a second AC signal into the negative bus of the DC system, and collecting the second response current of each branch; The frequency of the first AC signal is different from the frequency of the second AC signal; The first response current and the second response current in each branch collected are decomposed into multiple modal components respectively by using a modal decomposition algorithm, and the modal components whose center frequencies are consistent with the frequencies of the injected AC signals are screened out and recorded as response modal components; Solving the amplitude of the response modal component based on a signal amplitude algorithm to obtain a response current amplitude; Solve the ratio of the two response current amplitudes in each branch to obtain the response ratio of each branch; Based on the response ratio of each branch, the faulty branch is determined by numerical comparison method.

2. The DC system fault location method based on the improved variable frequency signal injection method according to claim 1, characterized in that: The modal decomposition algorithm is a successive variational modal decomposition algorithm; the signal amplitude algorithm is a Hilbert transform.

3. The DC system fault location method based on the improved variable frequency signal injection method according to claim 2, characterized in that: The response modal component solution includes the following steps: The Fourier transform formula of the current order mode is updated in the Fourier domain to obtain the current order Fourier transform formula; Calculate and update the center frequency of the current-order mode according to the current-order Fourier transform formula; The Lagrange multipliers are updated by the dual ascent method; Repeat the update operation until the internal convergence condition is met and output the current modal component; The modal order is increased, and the update operation and convergence judgment are continued until the external convergence conditions are met to complete the decomposition of the branch response current.

4. The DC system fault location method based on the improved variable frequency signal injection method according to claim 2, characterized in that: Determining the amplitude of the response modal component includes the following steps: A Hilbert transform process is performed on the response modal component to construct an analytical signal, and the amplitude of the response modal component is calculated based on the analytical signal.

5. The DC system fault location method based on the improved variable frequency signal injection method according to claim 4 is characterized in that: Solving the amplitude of the response modal component also includes the following steps: performing a Hilbert transform on the response modal component through a convolution operation to obtain a Hilbert transform result; The analytical signal is a signal function constructed by the response modal component and the Hilbert transform result; The magnitude of the response modal component is calculated based on the signal function.

6. The DC system fault location method based on the improved variable frequency signal injection method according to claim 1, characterized in that: The frequency of the first AC signal is greater than the frequency of the second AC signal.

7. The DC system fault location method based on the improved variable frequency signal injection method according to claim 6, characterized in that: The solution of the response ratio comprises the following steps: Calculating, in each branch, a response current amplitude corresponding to the first AC signal and a response current amplitude corresponding to the second AC signal; The ratio of the two response current amplitudes in the branch is solved to obtain the response ratio of the branch.

8. The DC system fault location method based on the improved variable frequency signal injection method according to claim 7, characterized in that: The response ratio is the amplitude of the subsequent response current divided by the amplitude of the previous response current in the time sequence; Determining the faulty branch includes the following steps: Sort the branches according to the numerical values ​​of the response ratios to obtain a branch sequence; A branch with a response ratio greater than a threshold or the largest value in the branch sequence is selected as a branch with a fault.

9. The DC system fault location method based on the improved variable frequency signal injection method according to claim 1, characterized in that: The following steps are also included: In the branch with the fault, dividing the branch into a plurality of branch sections based on branch nodes; Inject branch AC signals of different frequencies at both ends of each branch section; The response current of each branch section after the branch AC signal is injected is collected in sequence, and the step of determining the branch with a fault is repeated to determine the branch section with a fault.

10. The DC system fault location method based on the improved variable frequency signal injection method according to claim 9, characterized in that: The branch AC signal includes a third AC signal and a fourth AC signal; the response current after the branch AC signal includes a third response current and a fourth response current; injecting the third AC signal into the upstream node of the branch section, and collecting a third response current of the downstream node of the branch section; After the third response current is collected, injecting the fourth AC signal into the downstream node of the branch section, and collecting the fourth response current of the upstream node of the branch section; The frequency of the third AC signal is greater than the frequency of the fourth AC signal.

Images