Cable fault detection method and cable connecting pipe

Abstract

The invention discloses a cable fault detection method and a cable connecting pipe, and the method comprises the following steps: S1, collecting a current signal of a cable in an operation process through a high-frequency current sensor, and carrying out the preprocessing of the current signal to obtain a standardized current signal; s2, constructing a Duffing chaotic oscillator kinetic equation of the fault signal, and solving by using a fourth-order Runge-Kutta method to obtain a phase trajectory; s3, extracting a phase trajectory from three parts, namely a chaotic quantitative index, a geometric morphology feature and a phase diagram grid entropy, so as to obtain a plurality of phase trajectory features; and S4, performing normalization processing on the phase trajectory features, giving a fault index calculation formula, setting a plurality of thresholds, and performing fault detection and classification according to the threshold range of the fault index. And a high-frequency current sensor is arranged in the cable connecting pipe. The cable fault detection device has the beneficial effects that even if a fault electric signal generated by the cable is relatively small, fault detection can be carried out on the cable, and the sensing performance is high.

Description

Technical Field

[0001] This invention belongs to the field of power transmission and distribution technology, specifically, it relates to a cable fault detection method and a cable connection pipe. Background Technology

[0002] With the acceleration of urbanization, high-voltage cross-linked polyethylene cables have gradually replaced overhead lines and become the main artery of urban power transmission and distribution networks. However, during long-term operation, the insulation performance of cable systems will gradually deteriorate due to the coupling effect of multiple factors such as electricity, heat, mechanics and environment.

[0003] Current methods for cable fault detection primarily involve acquiring electrical signals generated during cable operation using high-frequency current sensors or transient voltage sensors, and then determining whether the cable is in a faulty state based on the fluctuations in these signals. However, this method typically faces the following problem in practical applications: fault signals attenuate extremely rapidly during transmission within the cable. When the signal acquisition point is far from the point of fault occurrence, the fault signal becomes extremely small, making it difficult to directly detect cable faults through signal fluctuations. In other words, the method has poor ability to detect minute fault signals. Summary of the Invention

[0004] Existing cable fault detection methods struggle to directly detect cable faults using electrical signals when the fault signal attenuates significantly. To address this issue, the following invention is proposed: A cable fault detection method includes the following steps: Step S1, collecting the current signal of the cable during operation using a high-frequency current sensor and preprocessing it to obtain a standardized current signal; Step S2, constructing the Duffing chaotic oscillator dynamics equation of the fault signal and solving it using the fourth-order Runge-Kutta method to obtain the phase trajectory; Step S3, extracting the phase trajectory from three parts: chaotic quantization index, geometric morphological features, and phase diagram grid entropy, to obtain multiple phase trajectory features; Step S4, normalizing the phase trajectory features, setting multiple thresholds based on the given fault index calculation formula, and performing fault detection and classification according to the threshold range of the fault index.

[0005] Furthermore, in step S1, the preprocessing includes signal denoising and signal normalization.

[0006] Furthermore, the signal denoising method is a moving average filtering method. For the current value of the current signal at any time, the average of the current values ​​at the current point and several non-outlier values ​​before and after it is taken to replace the original current value.

[0007] Furthermore, the signal standardization method involves standardizing the current value. ,satisfy , For the i-th time point The denoised current value obtained after signal denoising processing is as follows. Represents all maximum noise reduction current values. Representing the minimum noise reduction current, all standardized current values ​​are sorted according to time. Arrange to obtain standard current signals .

[0008] Furthermore, in step S2, the dynamic equation of the Duffing chaotic oscillator is:

[0009] in For inertia, Here, k is the damping term, and k is the damping ratio. It is a nonlinear restoring force term. As a built-in periodic driving force of the system, It is the coupling coefficient. The standard current signal collected in step S1.

[0010] Furthermore, in step S2, the method for solving the phase trajectory is as follows: given the following first-order differential equation, the following formula exists. in Define vector The first-order differential equation is abbreviated as: ; For time arrive The time difference is h. The slope of the four steps is calculated sequentially using the following formula. sum vector

[0011]

[0012] Repeat the above steps to obtain multiple vectors Y, and plot the time series on the horizontal axis. The vertical axis is On the phase plane, the phase trajectory is obtained.

[0013] Furthermore, in step S3, the chaos quantification index includes three phase trajectory features: box dimension, correlation dimension, and maximum Lyapunov exponent; the geometric morphology features include two phase trajectory features: phase diagram area and trajectory centroid; and the phase diagram grid entropy uses Shannon entropy as the phase trajectory feature.

[0014] Further, in step S4, the formula for calculating the fault index is: ,in The i-th phase trajectory feature after normalization. for The corresponding weight value, For a small given value.

[0015] The present invention also proposes a cable connector tube, wherein a connector hole is provided at both ends of the cable connector tube, and the ends of two cables are electrically connected to each other after being inserted into the connector hole. A high-frequency current sensor electrically connected to the cable is provided inside the connector tube.

[0016] The fault detection method proposed in this invention has the advantage that even if the fault electrical signal is severely attenuated, the phase trajectory can be directly obtained by solving the Duffing chaotic oscillator dynamics equation, and then the fault of the cable can be detected by the state of the phase trajectory, thus greatly improving the ability to sense fault electrical signals. Attached Figure Description

[0017] Figure 1 This is a flowchart of the cable fault detection method in this invention. Detailed Implementation

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments.

[0019] like Figure 1 As shown, this is a preferred embodiment of the present invention, which includes a cable fault detection method and a cable connection pipe.

[0020] The cable fault detection method includes the following steps: Step S1: Acquisition and preprocessing of cable current signals During the actual operation of a cable, the changes in its current signal best reflect its current operating status. Therefore, it is necessary to collect the current signals generated by the cable during operation and perform appropriate preprocessing.

[0021] In this embodiment, a high-frequency current sensor electrically connected to the cable collects the current value of the cable at various times t during operation. The frequency of current value collection is set to at least 100 kHz / s. The collected current values ​​are then arranged according to time t to obtain a discrete time series. , That is, the current signal of the cable, and the fault signal is included in the current signal.

[0022] After acquiring the current signal, it needs to be preprocessed. In this embodiment, the preprocessing includes signal denoising and signal standardization. Since the environment is usually filled with harsh electromagnetic interference or random white noise, which has a high frequency and is messy, the acquired current signal has too many spikes. Therefore, signal denoising is required. In this embodiment, the moving average filtering method is used for signal denoising. The specific steps are as follows: for the current value at any time t, take the current value at the current point and the N non-outlier current values ​​before and after it, and use the average of these current values ​​to replace the original current value.

[0023] Signal standardization primarily aims to eliminate the influence of dimensions, thereby eliminating differences in current signals between cables of different specifications and improving the applicability of this embodiment. The steps are as follows: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] The denoised current value obtained after signal denoising processing is denoted as follows: The standardized current value is ,satisfy , Represents all maximum noise reduction current values. The minimum noise reduction current will be determined by the standardized current value according to time. The discrete-time sequence obtained by arranging the signals is denoted as the standard current signal. .

[0024] Step S2: Fault signal detection based on chaotic oscillator.

[0025] When a cable fault occurs, a small, periodically varying fault signal is typically generated within the standard current signal for subsequent fault detection. In this embodiment, a Duffing chaotic oscillator system is used as the detector, and the dynamic equation of the Duffing chaotic oscillator is introduced as follows:

[0026] in For inertia, Here, k is the damping term, and k is the damping ratio. It is a nonlinear restoring force term. As a built-in periodic driving force of the system, It is the coupling coefficient. The standard current signal collected in step S1.

[0027] The driving force critical value of the chaos threshold in the Duffing chaotic oscillator equation was calculated using the Melnikov method. The amplitude F of the periodic driving force is set to .

[0028] The dynamic equations of this oscillator typically do not have analytical solutions and are usually solved numerically. The fourth-order Runge-Kutta method is used below to solve the dynamic equations of the Duffing chaotic oscillator. The basic steps are as follows: The original dynamic equations are transformed into first-order differential equations, as follows: in Let vector The above first-order differential equation can be simplified as follows: .

[0029] Based on the above, for time... arrive The time difference is h, and the slope needs to be calculated in four steps sequentially. Here, the time difference h is related to It is related to the length of time.

[0030] If the current time t satisfies Then, the vector Y in the corresponding first-order differential equation satisfies Nonlinear restoring force term The corresponding potential energy function is This is a double potential well system, and its minimum potential energy satisfies If the initial vector is , usually satisfy The following four-step slope calculation formula is given:

[0031] Furthermore, it can be done through vectors Calculate the vector This results in multiple sets of vectors Y, all of which are plotted on the x-axis. The vertical axis is On the phase plane, the phase trajectory is obtained. If the phase trajectory is always in a chaotic state, it means that the input standard current signal contains only noise. If the phase trajectory becomes a regular limit cycle at a certain moment, it means that the input standard current signal contains a weak periodic fault signal.

[0032] Step S3: Phase trajectory feature extraction After obtaining the phase trajectory that meets the above conditions, it is necessary to extract the phase trajectory features to determine whether the phase trajectory features are closer to a periodic state or a chaotic state. In this embodiment, the phase trajectory features are extracted from three parts: chaotic quantification index, geometric morphology features, and phase diagram grid entropy.

[0033] Chaos quantification metrics include box dimension, correlation dimension, and maximum Lyapunov exponent. Box dimension represents the number of grids required to cover the phase trajectory after dividing the phase plane into grids; the smaller the value, the simpler the phase trajectory structure, and the closer it is to a periodic state. Correlation dimension describes fractal dimension and is usually calculated using the GP algorithm. Like box dimension, the smaller the value, the simpler the phase trajectory structure, and the closer it is to a periodic state. Maximum Lyapunov exponent (MLE) describes the exponential separation rate between two adjacent trajectories in phase space and is usually calculated using the Wolf algorithm or Rosenstein algorithm. The closer the MLE is to 0, the more stable the periodic cycle of the phase trajectory.

[0034] Geometric features include the phase diagram area and the trajectory centroid. The phase diagram area can be calculated using either the convex hull method or the mesh method. The convex hull method calculates the area of ​​the smallest convex polygon containing all phase points, while the mesh method involves meshing the phase plane and counting the number of meshes into which trajectory points fall. The trajectory centroid is the phase trajectory. The statistical distribution center can be calculated using the formulas for centroid and average radius. The smaller the area of ​​the phase diagram and the centroid of the trajectory, the higher the degree of clustering of the phase trajectories, and the closer it is to a periodic state.

[0035] Phase diagram grid entropy is usually represented by Shannon entropy, the steps of which involve dividing the phase plane into... After mapping the grid, calculate the probability of a point falling into each grid j. Calculate Shannon entropy The lower the Shannon entropy, the closer the phase trajectory is to a periodic state. All the above phase trajectory features are combined into a phase trajectory feature vector, which is a 6-dimensional vector in this embodiment.

[0036] Step S4: Perform fault detection and classification To eliminate the influence of different element dimensions, the signal standardization method in step S1 is used to normalize each element in the phase trajectory feature vector to obtain the standard phase trajectory feature vector V. The current signal of the cable is collected n times, and after steps S1-S3 and normalization processing, n standard phase trajectory feature vectors can be obtained.

[0037] The following section classifies these standard phase trajectory feature vectors. In this embodiment, a weighted fault index is used to classify the standard phase trajectory feature vectors.

[0038] The steps are as follows: For the standard phase trajectory feature vector V, its fault index is defined as... Satisfies the calculation formula The weaker the fault signal, the smaller the fault index. Let V be the i-th element, which is also the i-th phase trajectory feature. for The corresponding weight value; the larger the weight value, the greater the... The smaller the impact on cable faults, the lower the weight value can be. This can be determined using either an objective or subjective method. The objective method pre-determines the weight based on the fluctuations of various elements during past cable faults; generally, the greater the fluctuation, the smaller the pre-determined value. The subjective method involves a person skilled in the art directly assigning weight values ​​to each element based on its importance; the more important the element, the smaller its weight value. For small given values, prevent the denominator from being 0.

[0039] Then, based on the required fault detection accuracy, multiple thresholds of different sizes are set for the fault index. A threshold range is formed between any adjacent thresholds. When the fault index is less than the minimum threshold, it means that the fault signal can be ignored and the cable is in normal condition. When the fault index falls into other threshold ranges, the cable enters the fault state within that threshold range. To determine the specific fault state, it is only necessary to refer to the fault state of the cable corresponding to the fault index that fell into that threshold range in the past.

[0040] The cable connecting pipe has two frustum-shaped ends and each has a connecting hole. The ends of the two cables are electrically connected to each other after being inserted into the connecting holes. High-frequency current sensors electrically connected to the two cables are installed in the cable connecting pipe to collect the current signals generated by the two cables during operation.

[0041] 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 method of cable fault detection, characterized by: The method comprises the following steps: Step S1, collecting the current signal of the cable during operation by a high-frequency current sensor, and pre-processing to obtain a standardized current signal; Step S2, constructing a Duffing chaotic oscillator dynamics equation of the fault signal, and using a fourth-order Runge-Kutta method to solve to obtain a phase trajectory; Step S3, extracting the phase trajectory from three parts of a chaotic quantitative index, a geometric morphological feature, and a phase diagram grid entropy to obtain multiple phase trajectory features; 2. The cable fault detection method of claim 1, wherein: Step S4, normalizing the phase trajectory features, giving a fault index calculation formula, setting multiple threshold values, and performing fault detection and classification according to the threshold value range of the fault index.

3. The method of claim 2, wherein: In step S1, the preprocessing includes signal denoising and signal standardization.

4. The method of claim 3, wherein: The signal standardization method is to standardize the current value , satisfies , is the i-th moment , the de-noised current value obtained by de-noising the signal, represents all the maximum de-noised current values, represents the minimum de-noised current, and all the standardized current values are arranged according to the moment to obtain the standard current signal .

5. The method of claim 4, wherein: The signal denoising method is a moving average filtering method, and for the current value of the current signal at any time, the mean value of the current values of the current point and the multiple current values before and after the current point which are not abnormal values is used to replace the original current value. wherein is an inertial term, is a damping term, k is a damping ratio, is a non-linear restoring force term, is a periodic excitation force built into the system, is a coupling coefficient, is a standard current signal collected in step S1.

6. The method of cable fault detection according to claim 5, characterized in that: In step S2, the dynamics equation of the Duffing chaotic oscillator is where , define the vector , the first-order differential equation is briefly written as ; for the time from to , the time difference is h, and the four-step slope and the vector Repeat the above steps to obtain a plurality of vectors Y, and plot the time sequence on the phase plane with the horizontal coordinate and the vertical coordinate to obtain the phase trajectory.

7. The method of claim 6, wherein: In step S2, the method for solving the phase trajectory is that a first-order differential equation is given, and there is a formula In step S3, the chaotic quantitative index includes three phase trajectory features of a box dimension, a correlation dimension, and a maximum Lyapunov exponent; The geometric morphological feature includes two phase trajectory features of a phase diagram area and a trajectory barycenter; 8. The method of cable fault detection according to claim 7, characterized in that: In step S4, the fault index is calculated by the formula wherein is the normalized i-th phase trajectory feature, is the corresponding weight value, is a small given value.

9. A cable connection tube characterized by: The phase diagram grid entropy uses Shannon entropy as a phase trajectory feature. The cable connection pipe is provided with a connection hole at both ends, the end portions of the two cables are electrically connected to each other after being inserted into the connection hole, and the inside is provided with the high-frequency current sensor according to any one of claims 1-8.

Images