A method for diagnosing a fault of a circuit breaker contact

Abstract

This invention discloses a fault diagnosis method for circuit breaker contacts, comprising the following steps: Step S1, injecting DC current into the circuit breaker contact and simultaneously acquiring the voltage drop U(t) and injected current I(t) across the contact, and calculating the dynamic resistance measurement curve; Step S2, inputting the DRM signal acquired in Step S1 into the SVMD algorithm, directly outputting the IMF component, and constructing a multi-domain feature vector based on the main IMF components decomposed by SVMD and the multi-domain features corresponding to the IMF components; Step S3, using the AHP algorithm to model and calculate the total feature vector corresponding to the multi-domain feature vector F, given the total fault calculation formula, where X is the total fault index. The beneficial effect of this invention is that it can ultimately quantitatively evaluate the current fault status of the circuit breaker contacts to predict the probability of future faults in the circuit breaker contacts.

Description

Technical Field

[0001] This invention belongs to the field of circuit breaker technology, and more specifically, relates to a fault diagnosis method for circuit breaker contacts. Background Technology

[0002] As an indispensable control and protection device in power systems, circuit breakers function as switches to interrupt or close current during normal operation. When abnormal faults such as overload, undervoltage, or short circuits occur, they can quickly and automatically disconnect the circuit through their internal tripping mechanism, thereby protecting downstream electrical equipment and preventing fires. Circuit breakers are currently widely used in power grids at all levels, from household distribution boxes and industrial control cabinets to high-voltage substations. With the development of smart grids, circuit breakers are evolving from traditional mechanical protection to intelligent systems with sensing, communication, and remote fault diagnosis capabilities, becoming a crucial node for achieving reliable operation and refined management of power systems.

[0003] In power systems, whether circuit breaker contacts are faulty is one of the most crucial factors determining whether a circuit breaker can function properly. However, existing methods for fault diagnosis of circuit breaker contacts largely rely on collecting a specific electrical variable during circuit breaker operation and determining whether a fault has occurred using simple logic based on a fixed threshold. This threshold-based fault determination method primarily focuses on qualitative fault diagnosis, merely determining the presence or absence of a fault diagnosis. When no fault has occurred in the circuit breaker contacts, it cannot further analyze the possibility of a fault occurring. Summary of the Invention

[0004] Existing methods for fault diagnosis of circuit breaker contacts mostly rely on simple logical judgments using fixed thresholds. This makes it difficult to quantitatively analyze the future probability of a circuit breaker contact failure when no failure has occurred, which is inconvenient. To address this problem, the following invention is proposed: A fault diagnosis method for circuit breaker contacts includes the following steps: Step S1: Inject DC current into the circuit breaker outlet and simultaneously collect the voltage drop U(t) and injected current I(t) across the outlet to calculate the dynamic resistance measurement curve. Step S2: Input the DRM signal acquired in Step S1 into the SVMD algorithm, directly outputting the IMF components. Based on the main IMF components decomposed by SVMD, construct a multi-domain feature vector according to the multi-domain features corresponding to the IMF components. ; Step S3: Use the AHP algorithm to model and calculate the total eigenvector corresponding to the multi-domain eigenvector F. The given total fault calculation formula is: X is the total failure index.

[0005] Furthermore, in step S1, a high-precision constant current source is used to inject DC current into the circuit breaker outlet, with a current amplitude of at least 100A. A high-speed, high-resolution A / D converter is used to synchronously acquire the voltage drop U(t) and the injected current I(t) across the outlet, with a sampling rate of at least 10kHz recommended.

[0006] Furthermore, in step S2, the multi-domain features include the arc contact time f1, waveform kurtosis f2, centroid frequency f3, multi-scale arrangement entropy f4, and energy moment f5.

[0007] Furthermore, in step S3, the total eigenvector satisfy ,in For expert subjective weight vector, The entropy weight is the objective weight vector. This is the preference coefficient.

[0008] Furthermore, ,in Multi-domain features The corresponding entropy weight objective weight satisfies , Represents multi-domain features Information entropy. The beneficial effect of this invention is that it can ultimately provide a quantitative evaluation of the current fault status of the circuit breaker contacts in order to predict the likelihood of future faults in the circuit breaker contacts. Attached Figure Description

[0009] Figure 1 This is a flowchart of the fault diagnosis method in this invention. Detailed Implementation

[0010] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments.

[0011] like Figure 1 As shown, this is a preferred embodiment of the present invention, a fault diagnosis method for circuit breaker contacts, comprising the following steps.

[0012] Step S1: Dynamic resistance measurement and data acquisition This embodiment uses dynamic resistance measurement (DRM) as the data basis for measuring contact fault status.

[0013] A high-precision constant current source is used to inject DC current into the circuit breaker contacts. The current amplitude must meet the requirements of ensuring the signal-to-noise ratio and breaking down the oxide film on the contact surface. Typically, it is at least 100A. A high-speed, high-resolution A / D converter (sampling rate recommended to be at least 10kHz) is used, and the voltage drop U(t) and injected current I(t) across the contacts are collected simultaneously.

[0014] According to Ohm's Law The dynamic resistance measurement curve, also known as the DRM curve, can be calculated. This embodiment selects the DRM curve during the circuit breaker's opening process. During the opening process, the main contacts separate first, the current transfers to the arc contacts, and then the arc contacts separate again. Therefore, the DRM curve clearly shows the evolution of the main contact contact stage (low resistance), the arc contact only stage (medium resistance), and complete disconnection (infinite resistance) in the time domain. This process directly reflects the geometric wear of the contacts.

[0015] Step S2: Direct extraction of multi-domain features based on successive variational mode decomposition Traditional variational mode decomposition (VMD) suffers from low computational efficiency due to the need to pre-determine the number of modes K and its reliance on complex parameter optimization algorithms. Therefore, this embodiment employs successive variational mode decomposition (SVMD). This method does not require prior knowledge of the number of signal modes or iterative optimization, and can directly and quickly extract modal components containing fault information from complex signals and calculate multi-domain features.

[0016] SVMD-based direct signal decomposition: Principle: SVMD transforms the mode extraction process into a process of finding the optimal mode one by one by constructing a new variational problem. Unlike VMD, which extracts all modes simultaneously, SVMD extracts one intrinsic mode function (IMF) at a time that minimizes the signal residual energy, until the residual satisfies the stopping condition.

[0017] Specific implementation method: The non-stationary DRM signal acquired in step S1 is input into the SVMD algorithm. The algorithm automatically locks the frequency band with concentrated energy in the signal and directly outputs a series of physically meaningful IMF components.

[0018] Direct acquisition of multi-domain features: Based on the main IMF components decomposed by SVMD, multi-domain feature vectors are constructed according to the multi-domain features corresponding to the IMF components. .

[0019] In this embodiment, several representative multi-domain features are selected, including, in order: Arc contact time f1: The resistance change point is directly identified from the first main mode of the DRM signal, and the time difference between the separation of the main contact and the separation of the arc contact is calculated.

[0020] Waveform kurtosis f2: the kurtosis value of the IMF component, used to characterize the severity of impact during mechanical operation.

[0021] Center of gravity frequency f3: This is the center of gravity frequency of the IMF component, which reflects the main frequency drift of the mechanism's motion.

[0022] Multiscale arrangement entropy f4: The arrangement entropy is calculated using key IMF components to quantify the complexity and disorder during the contact process. This index is extremely sensitive to changes in micro-roughness caused by ablation of the contact surface.

[0023] Energy moment f5: Reflects the distribution of signal energy on the time axis and is used to identify jamming or hysteresis of the operating mechanism.

[0024] The larger the value of the above five multi-domain features, the more irregular the DRM curve is during the circuit breaker's tripping process, which means the tripping process is more unstable and the circuit breaker is closer to a fault state.

[0025] Step S3: AHP algorithm modeling and given total fault calculation formula This step is the core algorithm of this invention, designed to address the misjudgment problem of traditional models when facing complex faults. First, a three-layer evaluation system is constructed: Target Layer (A): Overall health index of the circuit breaker. Criterion Layer (B): Multi-domain features extracted in step S2. Indicator Layer (C): Multi-domain feature vector F extracted in step S2.

[0026] The following requires the total eigenvector corresponding to the multi-domain eigenvector F. ,in General characteristics Corresponding multi-domain features In this embodiment, the weighting is assigned using a combination of subjective and entropy weighting methods to simultaneously consider subjective factors.

[0027] After normalizing the multi-domain feature vector F, solve for: (1) the expert subjective weight vector. ,satisfy ,in Multi-domain features The corresponding expert subjective weights are constructed based on expert experience, and the expert subjective weight vector is output through the judgment matrix. It is necessary to ensure that the consistency ratio CR < 0.1; (2) Entropy weight objective weight vector ,satisfy ,in Multi-domain features The corresponding entropy weight objective weight is equivalent to calculating the information entropy of each multi-domain feature using historical monitoring data. , Represents multi-domain features The information entropy. Finally, the formula for calculating the total eigenvector is: , This is the preference coefficient.

[0028] Finally, the total failure calculation formula can be given, specifically satisfying: X is the total failure index. The smaller the output value, the lower the probability of the circuit breaker contacts failing. The larger the output value, the higher the probability of the circuit breaker contacts failing. An upper limit threshold can be set. When X exceeds the upper limit threshold, the circuit breaker contacts are directly determined to be in a fault state.

[0029] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A fault diagnosis method for circuit breaker contacts, characterized in that: Includes the following steps, Step S1: Inject DC current into the circuit breaker break and simultaneously collect the voltage drop U(t) and injected current I(t) across the break, and calculate the dynamic resistance measurement curve. Step S2: Input the DRM signal acquired in Step S1 into the SVMD algorithm, directly outputting the IMF components. Based on the main IMF components decomposed by SVMD, construct a multi-domain feature vector according to the multi-domain features corresponding to the IMF components. ; Step S3: Use the AHP algorithm to model and calculate the total eigenvector corresponding to the multi-domain eigenvector F. The given total fault calculation formula is: X is the total failure index.

2. The fault diagnosis method according to claim 1, characterized in that: In step S1, a high-precision constant current source is used to inject DC current into the circuit breaker outlet, with a current amplitude of at least 100A. A high-speed, high-resolution A / D converter is used to synchronously acquire the voltage drop U(t) and the injected current I(t) across the outlet, with a sampling rate of at least 10kHz recommended.

3. The fault diagnosis method according to claim 2, characterized in that: In step S2, the multi-domain features include arc contact time f1, waveform kurtosis f2, centroid frequency f3, multi-scale arrangement entropy f4, and energy moment f5.

4. The fault diagnosis method according to claim 3, characterized in that: In step S3, the total eigenvector satisfy ,in For expert subjective weight vector, The entropy weight is the objective weight vector. This is the preference coefficient.

5. The fault diagnosis method according to claim 4, characterized in that: ,in Multi-domain features The corresponding entropy weight objective weight satisfies , Represents multi-domain features Information entropy.

Images