Cable insulation fault detection method and device based on sheath ring current

Abstract

The invention discloses a cable insulation fault detection method and device based on sheath ring current, and belongs to the technical field of intelligent monitoring, and the method comprises the steps: synchronously collecting cable sheath ring current data of a to-be-detected cable and related branch load current data; smoothing the cable sheath ring current data to obtain a smooth current data sequence; summing the load current data on each branch to obtain a load current sum sequence; calculating an instantaneous ratio resistance sequence based on the smooth current data sequence and the load current sum sequence; judging whether the to-be-detected cable has a fault risk or not according to the instantaneous specific resistance sequence; if yes, utilizing a multi-physics field finite element simulation model to extract electromagnetic features and thermal features; inputting the electromagnetic features and the thermal features into a fault classification model to obtain a predicted fault type; and when the insulation fault is predicted, an insulation fault alarm is given out. According to the invention, the problem of low insulation fault detection accuracy in the prior art can be solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent monitoring, in particular to a cable insulation fault detection method and device based on sheath loop current. BACKGROUND

[0002] In the transmission and distribution network of electric power energy, cables, as indispensable infrastructure, undertake the key task of efficiently and stably transporting electric energy from the power generation end to the power consumption end. Due to the long-term exposure of cables to complex and changeable environmental conditions, they not only have to bear the heat accumulation generated by electrical load, but also have to cope with the combined effects of environmental temperature fluctuations, humidity changes, chemical corrosion and mechanical vibration, etc. These factors can cause the insulation performance of cables to decline, and further cause insulation faults.

[0003] With the rapid development of sensor technology, signal processing technology and artificial intelligence technology, cable insulation fault detection methods based on online monitoring have gradually become a research hotspot. Online monitoring technology installs various sensors on cables to collect parameters in real time during their operation, and uses advanced signal processing algorithms and artificial intelligence models to analyze and process the collected data, thereby realizing real-time monitoring and fault warning of cable operation status. However, existing online monitoring technologies mostly focus on the monitoring and analysis of a single parameter, ignoring the important characteristic parameter of cable sheath loop current. Cable sheath loop current is the current generated in the cable sheath during the operation of the cable due to electromagnetic induction and capacitive coupling, etc. Its characteristics are closely related to the insulation state of the cable. When the cable has an insulation fault, the size, phase, waveform and other characteristic parameters of the sheath loop current will change significantly, and these changes contain rich fault information. SUMMARY

[0004] The present application provides a cable insulation fault detection method and device based on sheath loop current, which can solve the problem of low accuracy of insulation fault detection in the prior art.

[0005] To solve the above technical problems, the present application provides a cable insulation fault detection method based on sheath loop current, comprising:

[0006] According to the preset data collection interval, the cable sheath loop current data and the related branch load current data of the cable to be tested in the power system are synchronously collected; wherein the related branch load current data is the load current data on the branches connected in series with the cable to be tested;

[0007] The sliding average method is used to smooth the cable sheath loop current data to obtain a smooth current data sequence;

[0008] Summing up the load current data on each branch at each time, a load current sum sequence is obtained;

[0009] Based on the smoothed current data sequence and the load current sum sequence, an instantaneous ratio resistance sequence is calculated;

[0010] By analyzing whether the instantaneous ratio resistance sequence is abnormal, it is determined whether the cable under test has a risk of failure;

[0011] When it is determined that the cable under test has a risk of failure, electromagnetic features and thermal features are extracted from the smoothed current data sequence and the instantaneous ratio resistance sequence using a multi-physical field finite element simulation model;

[0012] The electromagnetic features and the thermal features are input into a fault classification model to obtain a predicted fault type;

[0013] When the predicted fault type is an insulation fault, an insulation fault alarm is issued.

[0014] As a preferred solution, based on the smoothed current data sequence and the load current sum sequence, the instantaneous ratio resistance sequence is calculated, including:

[0015] According to the following formula, the instantaneous ratio resistance corresponding to each time in the instantaneous ratio resistance sequence is calculated:

[0016]

[0017] In the formula, R(t) is the instantaneous ratio resistance at time t; I c (t) is the cable sheath circulating current at time t; I n (t) is the load current sum sequence at time t.

[0018] As a preferred solution, by analyzing whether the instantaneous ratio resistance sequence is abnormal, it is determined whether the cable under test has a risk of failure, including:

[0019] The data standard deviation and the load change rate of the instantaneous ratio resistance sequence are calculated respectively;

[0020] According to the data standard deviation and the load change rate, it is determined whether the cable under test has a risk of failure.

[0021] As a preferred solution, according to the data standard deviation and the load change rate, it is determined whether the cable under test has a risk of failure, including:

[0022] When the data standard deviation is greater than a preset standard deviation threshold or the load change rate is greater than a preset change rate threshold, it is determined whether the cable under test has a risk of failure.

[0023] As a preferred solution, the construction process of the multi-physical field finite element simulation model is:

[0024] Obtaining topology data of the power system, and constructing a three-dimensional model according to the topology data;

[0025] Obtaining a plurality of fault data of the cable to be tested; wherein the fault data includes fault parameters and fault types;

[0026] Training the three-dimensional model using the fault data to learn fault characteristics corresponding to each fault type; wherein the fault characteristics include fault electromagnetic characteristics and fault thermal characteristics;

[0027] Determining the trained three-dimensional model as a multi-physical field finite element simulation model.

[0028] As a preferred solution, the basic equations of the multi-physical field finite element simulation model include Maxwell's equation set, heat conduction equation, heat convection equation and heat radiation equation, and electromagnetic and thermal coupling equation;

[0029] The Maxwell equation set is:

[0030]

[0031] In the formula, is a vector differential operator; E is the electric field intensity; B is the magnetic flux density; H is the magnetic field intensity; J is the current density; D is the electric field displacement; p is the charge density; t is the time variable;

[0032] The heat conduction equation is:

[0033]

[0034] In the formula, k is the thermal conductivity of the material; T is the surface temperature of the object; Q is the heat source in unit volume;

[0035] The heat convection equation and the heat radiation equation are:

[0036] q conv = h (T ∞ -T) A

[0037]

[0038] In the formula, q conv is the heat convection heat flux density; q rad is the heat radiation heat flux density; h is the convection heat transfer coefficient; ∈ is the emissivity; σ is the Stefan-Boltzmann constant; T ∞ is the ambient temperature; A is the fluid surface area normal vector;

[0039] The electromagnetic and thermal coupling equation is:

[0040]

[0041] wherein, a is the thermal diffusivity.

[0042] As a preferred solution, when it is determined that the to-be-tested cable has a risk of failure, electromagnetic features and thermal features are extracted from the smoothed current data sequence and the instantaneous specific resistance sequence by using a multi-physical field finite element simulation model, including:

[0043] The smoothed current data sequence and the instantaneous specific resistance sequence are input into the multi-physical field finite element simulation model, so that the multi-physical field finite element simulation model performs simulation to obtain resulting electromagnetic features and thermal features;

[0044] The simulation process of the multi-physical field finite element simulation model is as follows:

[0045] The high-frequency component in the smoothed current data sequence is extracted, and the repetition rate of the high-frequency component is calculated to obtain a high-frequency pulse repetition rate;

[0046] According to the instantaneous specific resistance sequence, the contact resistance change rate of the to-be-tested cable is calculated, the failure time point is identified through the contact resistance change rate, and the resistance growth rate before and after the failure is calculated based on the failure time point;

[0047] The spatial derivative of the magnetic flux density is extracted based on the smoothed current data sequence, and the spatial gradient of the magnetic flux density is calculated;

[0048] The amplitude and duration of the pulse are obtained according to the smoothed current data sequence, and the partial discharge pulse energy is calculated based on the amplitude and duration of the pulse;

[0049] The hotspot temperature and temperature gradient are extracted in the smoothed current data sequence;

[0050] The temperature curve is fitted based on the smoothed current data sequence, and the thermal time constant is extracted from the temperature curve;

[0051] The high-frequency pulse repetition rate, the resistance growth rate before and after the failure, the spatial gradient of the magnetic flux density, and the partial discharge pulse energy are determined as electromagnetic features;

[0052] The hotspot temperature, the temperature gradient, and the thermal time constant are determined as thermal features.

[0053] As a preferred solution, the training process of the failure classification model is as follows:

[0054] A plurality of the failure data are input into the multi-physical field finite element simulation model to obtain failure electromagnetic features and failure thermal features corresponding to each of the failure data;

[0055] The fault electromagnetic feature, the fault thermal feature and the fault type corresponding to each fault data are input into a support vector machine model, and the support vector machine model is trained to learn the mapping relationship between the fault electromagnetic feature and the fault thermal feature and the fault type.

[0056] The trained support vector machine model is determined as a fault classification model.

[0057] As a preferred solution, the fault classification function of the fault classification model is:

[0058]

[0059] In the formula, f(x) is a fault classification function; K(x i ,x) is a kernel function; N is the total number of samples; b is a bias term; α i is a sample weight; y i is a fault type;

[0060] The kernel function is:

[0061] K(x i ,x)=exp(-γ‖x i -x‖ 2 )

[0062] In the formula, γ is a control kernel function width; ‖·‖ is a Euclidean distance; x i is the i-th sample variable; x is input feature data.

[0063] Correspondingly, the application provides a cable insulation fault detection device based on a sheath circulating current, comprising a data acquisition module, a smoothing processing module, a current summation module, a resistance calculation module, a risk judgment module, a feature extraction module, a fault prediction module and a fault warning module.

[0064] The data acquisition module is used for synchronously acquiring cable sheath circulating current data and related branch load current data of a to-be-tested cable in a power system according to a preset data acquisition interval; wherein the related branch load current data is load current data on a plurality of branches connected in series with the to-be-tested cable.

[0065] The smoothing processing module is used for performing smoothing processing on the cable sheath circulating current data by using a sliding average method to obtain a smoothed current data sequence.

[0066] The current summation module is used for summing up load current data on a plurality of branches corresponding to each time point to obtain a load current total sequence.

[0067] The resistance calculation module is configured to calculate an instantaneous ratio resistance sequence based on the smoothed current data sequence and the load current sum sequence.

[0068] The risk judgment module is configured to judge whether the cable under test has a risk of failure by analyzing whether the instantaneous ratio resistance sequence is abnormal.

[0069] The feature extraction module is configured to extract electromagnetic features and thermal features from the smoothed current data sequence and the instantaneous ratio resistance sequence by using a multi-physical field finite element simulation model when it is determined that the cable under test has a risk of failure.

[0070] The fault prediction module is configured to input the electromagnetic features and the thermal features into a fault classification model to obtain a predicted fault type.

[0071] The fault early warning module is configured to issue an insulation fault alarm when the predicted fault type is an insulation fault.

[0072] Compared with the prior art, the embodiment of the present application has the following beneficial effects:

[0073] The present application provides a cable insulation fault detection method based on sheath circulating current, according to a preset data collection interval, the cable sheath circulating current data and the related branch load current data of the cable under test in a power system are synchronously collected; the cable sheath circulating current data is smoothed by using a moving average method to obtain a smoothed current data sequence; the load current data on a plurality of branches corresponding to each time point is summed respectively to obtain a load current sum sequence; an instantaneous ratio resistance sequence is calculated based on the smoothed current data sequence and the load current sum sequence; whether the cable under test has a risk of failure is judged by analyzing whether the instantaneous ratio resistance sequence is abnormal; when it is determined that the cable under test has a risk of failure, electromagnetic features and thermal features are extracted from the smoothed current data sequence and the instantaneous ratio resistance sequence by using a multi-physical field finite element simulation model; the electromagnetic features and the thermal features are input into a fault classification model to obtain a predicted fault type; when the predicted fault type is an insulation fault, an insulation fault alarm is issued. The cable sheath circulating current data and the related branch load current data of the cable under test are collected, whether there is a risk of failure is judged by analyzing the collected information, early warning of the fault is realized; when there is a risk of failure, rich feature information is extracted by using a multi-physical field finite element simulation model, the fault classification model is used to predict the fault type, rapid positioning of the fault is realized, the insulation fault detection accuracy is improved, and the safe and stable operation of the power system is ensured. BRIEF DESCRIPTION OF DRAWINGS

[0074] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, the drawings described below only constitute some of the embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0075] Figure 1 A flowchart of an embodiment of the cable insulation fault detection method based on the sheath circulating current provided by the present application;

[0076] Figure 2 An embodiment of the moving average method provided by the present application;

[0077] Figure 3 A structural diagram of an embodiment of the cable insulation fault detection device based on the sheath circulating current provided by the present application. DETAILED DESCRIPTION

[0078] In order to make the objects, technical solutions and advantages of the present application clearer, the following will combine the drawings in the embodiments of the present application to make a clear and complete description of the technical solutions in the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0079] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "include" and "have" and any variations thereof in the specification and claims of the present application and the above description of drawings are intended to cover non-exclusive inclusion.

[0080] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly and specifically limited.

[0081] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearance of the phrase in various places in the specification does not necessarily all refer to the same embodiment, nor is it necessarily independent or alternative embodiments to other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0082] In the description of the embodiments of the present application, the term "and / or" is only to describe an association relationship of associated objects, which means that there can be three relationships, for example, A and / or B, which can represent three cases of A alone, A and B together, and B alone. In addition, the character " / " in this paper generally represents that the front and rear associated objects are a "or" relationship.

[0083] In the description of the embodiments of the present application, the term "a plurality of" refers to two or more (including two), and similarly, "a plurality of groups" refers to two or more groups (including two groups), and "a plurality of pieces" refers to two or more pieces (including two pieces).

[0084] In the description of the embodiments of the present application, unless otherwise explicitly specified and limited, the technical terms "mounting", "connecting", "connecting", "fixing" and the like should be understood in a broad sense, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be mechanical connection, or it can be electrical connection; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meaning of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0085] Reference Figure 1 To solve the problem of low accuracy of insulation fault detection in the prior art, an embodiment of the present application provides a cable insulation fault detection method based on sheath circulating current, which comprises steps 101 to 108, and each step is as follows:

[0086] Step 101: According to the preset data acquisition interval, the cable sheath circulating current data and the related branch load current data of the cable to be tested in the power system are synchronously collected; wherein the related branch load current data is the load current data on the several branches connected in series with the cable to be tested.

[0087] In the embodiment of the present application, when detecting the fault state of the cable to be detected in the power system, it is not necessary to stop the running state thereof, and the cable sheath loop current data of a period of time is collected when the cable to be detected is in the running state according to the preset data collection interval. The device for collecting the data can adopt an electromagnetic induction type current transformer with high sensitivity, wide range and high stability, and the electromagnetic induction type current transformer is installed at a suitable position on the cable according to the physical layout, electrical parameters of the cable and the topological structure of the power system. The range of the electromagnetic induction type current transformer covers 0.1A to 10kA, so as to ensure that the sensor can adapt to the normal operation in the power system and the extreme current condition when the short-circuit fault occurs. The measurement accuracy of the sensor is ±0.1%FS, and high-precision current data is provided, so as to provide reliable basis for fault detection and system monitoring. In addition, the bandwidth of the sensor is ≥100kHz, so as to accurately capture the transient current signal and meet the fast response requirement of the power system. The current sensor is symmetrically installed with three groups of sensors on the surface of the three-phase conductor of the cable, so as to effectively avoid the magnetic field interference, ensure that the current signals measured by each group of sensors are not affected by the surrounding electromagnetic environment, and ensure the accuracy and reliability of the data.

[0088] In the embodiment of the present application, according to the topological structure data of the power system, a plurality of load branches in series with the cable to be detected are determined, and the load current data of each load branch is synchronously collected when the cable sheath loop current data of the cable to be detected is collected by installing sensors on the load branches. By collecting the related branch load current data of the cable to be detected, the current condition of each branch can be tracked in real time during the operation of the power system, and sufficient data support is provided for subsequent fault diagnosis and fault type classification.

[0089] Step 102: smoothing the cable sheath loop current data by using the moving average method to obtain a smooth current data sequence.

[0090] In the embodiment of the present application, after the cable sheath loop current data is collected, in order to effectively eliminate the random noise and interference in the current signal and improve the data quality, the original current data collected can be smoothed by using the moving average method. The moving average method calculates the arithmetic mean value of the current data in a sliding window with a fixed length, and replaces the original data value of the center point of the window with the mean value, so as to realize the smoothing of the current signal. The moving average method is as follows: for a given current data sequence, I1, I2,..., In, the moving average value of the current data sequence is as follows: n , the moving average value of which is The value at time t is:

[0091]

[0092] In the formula, is the moving average value; M is the size of the moving window, usually selected as a time step; I i is the cable sheath circulating current data at the i-th moment.

[0093] As an example of an embodiment of the present application, refer to Figure 2 is a schematic diagram of the moving average method provided by the present application. Assuming that the length of the moving window is N = 5, which corresponds to a sampling interval of 50 ms, the time covered by the moving window is 250 ms. When a new data point is collected, the value queue in the moving window is updated in real time, and the arithmetic mean of the data in the window is recalculated to obtain the smoothed current value. Using the moving average method ensures the real-time and efficiency of the smoothing process, and the processing delay after each data update is not more than 1 ms, meeting the strict requirements of the power system for real-time data processing.

[0094] Step 103: Summing the load current data on each branch corresponding to each moment to obtain a load current sum sequence.

[0095] In an embodiment of the present application, after obtaining the load current data of a plurality of related branches of the cable to be tested within a period of time, the load current data of each related branch is added for each moment to form the load current sum at that moment. The calculation formula of the load current sum at each moment is represented as:

[0096]

[0097] In the formula, I t (t) is the load current sum at time t; m is the number of branches; I i (t) is the load current data of the i-th branch at time t.

[0098] In an embodiment of the present application, after calculating the load current sum at each moment, these data are integrated to form a load current sum sequence. After forming the load current sum sequence, each load current sum in the load current sum sequence is normalized by determining the maximum load current sum in the load current sum sequence. The formula for normalization can be represented as:

[0099]

[0100] In the formula, I n (t) is the normalized load current sum at time t; max(I t (t)) is the maximum load current sum in the load current sum sequence.

[0101] Step 104: Based on the smoothed current data sequence and the load current sum sequence, an instantaneous ratio resistance sequence is calculated.

[0102] In the embodiment of the present application, after the smooth current data sequence and the load current sum sequence are obtained, the instantaneous ratio resistance between the cable sheath loop current data of the cable to be measured and the related branch load current data is calculated, the change of the instantaneous ratio resistance is tracked in real time, and the load distribution of the cable line in the power system and the electrical coupling characteristics between the cable line and other loads can be intuitively reflected.

[0103] As a preferred scheme of the embodiment, the calculation of the instantaneous ratio resistance adopts a high-precision floating-point operation algorithm to ensure the accuracy of the calculation result. The instantaneous ratio resistance at each time in the instantaneous ratio resistance sequence is calculated according to the following formula:

[0104]

[0105] In the formula, R(t) is the instantaneous ratio resistance at time t; I c (t) is the cable sheath loop current at time t; and I n (t) is the load current sum sequence at time t.

[0106] After the instantaneous ratio resistance at each time is calculated, the data is integrated to form the instantaneous ratio resistance sequence.

[0107] Step 105: judging whether the cable to be measured has a risk of failure by analyzing whether the instantaneous ratio resistance sequence is abnormal.

[0108] As a preferred scheme of the embodiment, whether the cable to be measured has a risk of failure is judged by analyzing whether the instantaneous ratio resistance sequence is abnormal, including:

[0109] calculating the data standard deviation and the load change rate of the instantaneous ratio resistance sequence respectively;

[0110] judging whether the cable to be measured has a risk of failure according to the data standard deviation and the load change rate.

[0111] In the embodiment of the present application, whether the cable to be measured has a risk of failure can be detected by judging whether the instantaneous ratio resistance sequence is abnormal. Specifically, the data standard deviation and the load change rate of the instantaneous ratio resistance sequence are calculated respectively, the calculated data standard deviation and load change rate are compared with preset threshold values respectively, and whether the cable to be measured has a risk of failure is judged.

[0112] In the formula, σ

[0113]

[0114] In the formula, σ Ris the data standard deviation of the instantaneous ratio resistance sequence; R(t) is the instantaneous ratio resistance at t moment; is the average value of the instantaneous ratio resistance sequence.

[0115] The load change rate of the instantaneous ratio resistance sequence can be calculated by the following formula:

[0116]

[0117] In the formula, is the load change rate; R(t) is the instantaneous ratio resistance at t moment; R(t-Δt) is the instantaneous ratio resistance at t-Δt moment; and Δt is the time interval.

[0118] As a preferred scheme of the embodiment, judging whether the to-be-tested cable has a failure risk according to the data standard deviation and the load change rate comprises:

[0119] When the data standard deviation is greater than a preset standard deviation threshold or the load change rate is greater than a preset change rate threshold, it is determined whether the to-be-tested cable has a failure risk.

[0120] In the embodiment, by setting the standard deviation threshold and the change rate threshold, whether the to-be-tested cable has a failure risk can be judged according to the calculated data standard deviation and load change rate. For example, when the data standard deviation is greater than a preset standard deviation threshold or the load change rate is greater than a preset change rate threshold, it is determined whether the to-be-tested cable has a failure risk.

[0121] Step 106: When it is determined that the to-be-tested cable has a failure risk, electromagnetic features and thermal features are extracted from the smoothed current data sequence and the instantaneous ratio resistance sequence by using a multi-physical field finite element simulation model.

[0122] In the embodiment, after the data acquisition and failure risk judgment of the to-be-tested cable are completed, in order to further explore the internal mechanism of cable failure and realize accurate identification of the failure type and severity grading judgment, a multi-physical field finite element simulation model can be constructed to analyze the physical characteristics of the cable failure. Based on various typical failure scenarios that the cable may encounter in the actual operation of the power system, a multi-physical field finite element simulation model that integrates the coupling effect of electromagnetic field and thermal field needs to be constructed to fully consider the complex geometric structure, material properties of the cable and the interaction mechanism between the multi-physical fields, and the numerical calculation technology of the finite element method is used to finely simulate the electromagnetic field and thermal field distribution of the cable under normal working state and different failure modes.

[0123] As a preferred scheme of the embodiment, the construction process of the multi-physical field finite element simulation model is as follows:

[0124] acquire topological data of the power system, and construct a three-dimensional model according to the topological data;

[0125] acquire several kinds of fault data of the cable to be measured; wherein the fault data comprises fault parameters and fault types;

[0126] train the three-dimensional model by using the fault data to learn fault features corresponding to each fault type; wherein the fault features comprise fault electromagnetic features and fault thermal features;

[0127] determine the trained three-dimensional model as a multi-physical field finite element simulation model.

[0128] In the embodiment of the application, firstly, based on the industrial CT scanning data (resolution 0.05mm / pixel) of the cable to be measured, a 1:1 scale three-dimensional model is constructed by using reverse engineering software. The model finely reconstructs the contact interfaces of the conductor, the insulating layer, the semiconductor shielding layer and the metal sheath, and especially the micro topography of the conductor crimping part: the indentation depth is 0.3-0.5mm, and the surface roughness Ra is less than or equal to 1.6μm. Parameterized modeling is performed to ensure the simulation accuracy of the contact resistance.

[0129] In the meshing stage, an adaptive hybrid mesh strategy is adopted:

[0130] The interface area between the conductor and the insulating layer: a hexahedron-dominated hybrid mesh is used, the contact surface is locally encrypted to 0.1mm (to ensure the skin effect capture), and the transition zone adopts a pyramid element for smooth transition;

[0131] The inside of the insulating layer: tetrahedral elements are used for unstructured discretization, and the grid size is dynamically adjusted according to the electric field strength gradient (high field strength zone ≤0.5mm, low field strength zone ≤2mm);

[0132] The outer shielding layer and the air domain: a gradual mesh is adopted, and the grid density in the near-field zone reaches 50 / λ (λ is the wavelength of electromagnetic wave), and the grid size in the far-field zone is increased by 1.5 times.

[0133] The total number of elements of the final three-dimensional model is about 12 million, and the single-step solving time is compressed to 15 minutes through parallel computing. After the three-dimensional model is constructed, a power frequency sinusoidal voltage excitation is applied at the end of the three-phase conductor, an interpolation function is used to realize continuous loading of the voltage, and numerical oscillation is avoided; a perfect matched layer (PML) is set on the outer boundary of the air domain, the thickness is set to λ / 4, the attenuation coefficient α is set to 100dB / m, the electromagnetic wave is ensured to be transmitted without reflection, and an infinite open space is simulated; the conductor adopts a temperature-dependent conductivity model; and the insulating layer introduces a nonlinear dielectric constant.

[0134] In the embodiment of the present application, after the three-dimensional model is constructed, a plurality of fault data of the cable to be tested is obtained in the database, including different fault types and corresponding fault parameters. According to the fault data, fault injection is realized by modifying the model parameters of the three-dimensional model, so that the three-dimensional model learns the fault characteristics.

[0135] As an example of the embodiment of the present application, according to the fault types and fault parameters recorded in the following table, the three-dimensional model is realized to realize fault injection, and the corresponding characteristic parameters can be extracted. For example, a 0.1-1.0mm air gap is embedded at the conductor crimping position, and the contact resistance is dynamically changed to simulate the growth process of the oxide layer, so that the hot spot temperature is extracted . A 0.5mm×0.5mm×0.5mm air gap is implanted in the insulating layer, the air pressure in the air gap is set to 0.05-0.2MPa, and a step voltage (ΔU=5kV) is applied to trigger discharge, so that the current harmonic content is extracted . The load current is increased to 1.2-2.0 times the rated value (I =1000A), and the ambient temperature is set to linearly increase from 25℃ to 80℃ to simulate the thermal aging process, so that the temperature gradient is extracted .

[0136] As a preferred scheme of the embodiment, the basic equations of the multi-physical field finite element simulation model include Maxwell's equations, heat conduction equation, heat convection equation and heat radiation equation, and electromagnetic and thermal coupling equation;

[0137] Among them, the Maxwell equation group is:

[0138]

[0139] In the formula, is a vector differential operator; E is the electric field intensity; B is the magnetic flux density; H is the magnetic field intensity; J is the current density; D is the electric field displacement; ρ is the charge density; t is the time variable;

[0140] The heat conduction equation is:

[0141]

[0142] In the formula, k is the thermal conductivity of the material; T is the surface temperature of the object; Q is the heat source in unit volume, which is usually represented by the joule heat generated by the current;

[0143] The heat convection equation and the heat radiation equation are:

[0144] q conv = h (T ∞ -T) A

[0145]

[0146] where q conv is the heat convection heat flux density; q rad is the heat radiation heat flux density; h is the convection heat transfer coefficient; ∈ is the emissivity; σ is the Stefan-Boltzmann constant; T ∞ is the ambient temperature; A is the fluid surface area normal vector;

[0147] The electromagnetic and thermal coupling equation is:

[0148]

[0149] where α is the thermal diffusivity.

[0150] In the embodiment of the present application, by the above electromagnetic field and thermal field equation, a multi-physical field coupled finite element model can be constructed, which can effectively describe the physical property evolution of the cable under normal working state and different fault modes. Under the normal working state of the cable, the electromagnetic field and the thermal field remain relatively stable, and the temperature distribution is uniform. Under the fault mode, the multi-physical field simulation model can accurately simulate the electromagnetic field distortion and the sharp change of temperature caused by the uneven distribution of current. From the simulation results, a feature parameter set capable of distinguishing the fault type is extracted, including current harmonic component, temperature gradient change rate and other key indicators.

[0151] As a preferred scheme of the present embodiment, when it is determined that the to-be-tested cable has a fault risk, electromagnetic features and thermal features are extracted from the smoothed current data sequence and the instantaneous specific resistance sequence by using the multi-physical field finite element simulation model, including:

[0152] The smoothed current data sequence and the instantaneous specific resistance sequence are input into the multi-physical field finite element simulation model, so that the multi-physical field finite element simulation model performs simulation to obtain result electromagnetic features and thermal features;

[0153] The simulation process of the multi-physical field finite element simulation model is as follows:

[0154] The high-frequency component in the smoothed current data sequence is extracted, the repetition rate of the high-frequency component is calculated, and the high-frequency pulse repetition rate is obtained;

[0155] According to the instantaneous specific resistance sequence, the contact resistance change rate of the to-be-tested cable is calculated, the fault time point is identified through the contact resistance change rate, and based on the fault time point, the resistance growth rate before and after the fault is calculated;

[0156] Based on the smoothed current data sequence, the spatial derivative of the magnetic flux density is extracted, and the spatial gradient of the magnetic flux density is calculated;

[0157] derive the amplitude and duration of the pulse from the smoothed current data sequence, and calculate a partial discharge pulse energy based on the amplitude and duration of the pulse;

[0158] extract a hot spot temperature and a temperature gradient in the smoothed current data sequence;

[0159] fit a temperature curve based on the smoothed current data sequence, and extract a thermal time constant from the temperature curve;

[0160] determine the high-frequency pulse repetition rate, the resistance growth rate before and after the fault, the spatial gradient of the magnetic flux density, and the partial discharge pulse energy as electromagnetic features;

[0161] determine the hot spot temperature, the temperature gradient, and the thermal time constant as thermal features.

[0162] In the embodiments of the present application, in the simulation field, the multi-physical field finite element simulation model can accurately calculate the key electromagnetic parameters such as the electric field intensity, the magnetic field intensity, and the magnetic flux density of the to-be-tested cable body and the surrounding space, deeply analyzes the influence of uneven current distribution on electromagnetic field distribution, and the feedback effect of electromagnetic field change on the electrical performance of the cable joint. Considering various heat transfer modes such as Joule heat generated by the current, material heat conduction, heat convection, and heat radiation, the temperature distribution of the cable joint under different load conditions is accurately simulated, and especially the sharp change process of the local temperature when the fault occurs.

[0163] The feature parameters extracted in the simulation results include electromagnetic features and thermal features. The electromagnetic features include a high-frequency pulse repetition rate N PD , a resistance growth rate β before and after the fault, a spatial gradient of the magnetic flux density , and a partial discharge pulse energy E PD . The calculation formula of the partial discharge pulse energy E PD is as follows:

[0164]

[0165] In the formula, E PD is the partial discharge pulse energy; P(t) is the pulse power; t1 is the starting time of the pulse; and t2 is the ending time of the pulse.

[0166] The thermal features include a hot spot temperature T max (℃), a temperature gradient ΔT / Δx (℃ / cm), and a thermal time constant τ th (s). The thermal time constant is obtained by fitting a temperature curve with a double exponential.

[0167] Step 107: input the electromagnetic features and the thermal features into a fault classification model to obtain a predicted fault type.

[0168] As a preferred scheme of the embodiment, the training process of the fault classification model is as follows:

[0169] The plurality of fault data is input into the multi-physical field finite element simulation model to obtain the fault electromagnetic features and fault thermal features corresponding to each fault data;

[0170] The fault electromagnetic features, the fault thermal features and the fault types corresponding to each fault data are input into the support vector machine model to train the support vector machine model to learn the mapping relationship between the fault electromagnetic features and the fault thermal features and the fault types;

[0171] The trained support vector machine model is determined as the fault classification model.

[0172] In the embodiment, after the electromagnetic features and the thermal features of the to-be-tested cable are extracted, the machine learning algorithm is introduced to realize intelligent classification and grading of faults and improve detection accuracy. The training process of the fault classification model is as follows: the plurality of fault data of the to-be-tested cable is input into the support vector machine model for training, and the kernel function parameters, the penalty factor and other hyperparameters of the model are adjusted through an optimization algorithm, so that the model can accurately divide the decision boundary of different fault types in the feature vector space. The support vector machine (SVM) model is as follows:

[0173]

[0174] In the formula, w is a hyperplane normal vector in the SVM, which is a weight vector used to determine the direction and position of the decision boundary; b is a bias vector in the SVM, also called an intercept, which is a constant term in the hyperplane equation; C is a penalty factor in the support vector machine, which is used to control the tolerance of the classification model to the wrong classification of the training set; and ξ i is a relaxation variable used to represent the error degree of the i th sample, and N is the total number of samples.

[0175] The constraint condition of the support vector machine (SVM) model is as follows:

[0176] y i (w·x i +b)≥1-ξ i ,ξ i ≥0,i=1,2,...,N

[0177] The support vector machine (SVM) model can adopt a Gaussian radial basis function (RBF) kernel, and the kernel function form is as follows:

[0178] K(x i ,x j )=exp(-γ‖x i -x j ‖ 2 )

[0179] where γ is the control of kernel function width; ‖·‖ is the Euclidean distance; x i is the i-th sample variable; x j is the j-th sample variable; the grid search algorithm is used to search the preset parameter space, such as {C ∈ [1, 1000], γ ∈ [0.001, 1]} for 5-fold cross-validation:

[0180]

[0181] where (C * ,γ * ) is the final optimal hyperparameter pair; K is the number of cross-validation, if 5-fold cross-validation is performed, K is 5; Accuracy k (C,γ) is the classification accuracy obtained by training the model with hyperparameters C and γ in the k-th cross-validation.

[0182] Therefore, the final optimal parameter combination is C * and γ * , C * is used to balance the classification boundary and the penalty for misclassification, and γ * is used to control the complexity of feature space mapping.

[0183] Finally, model validation and performance evaluation are performed, and stratified 10-fold cross-validation is used to ensure that the proportion of each type of fault sample in each fold is consistent with the original data set, avoiding data distribution deviation. The performance indicators can be used to judge the performance of the model, including classification accuracy, Kappa coefficient, and confusion matrix.

[0184] When the support vector machine (SVM) model is trained, a fault classification model is formed, and the fault classification function of the fault classification model is:

[0185]

[0186] where f(x) is the fault classification function; K(x i ,x) is the kernel function; N is the total number of samples; b is the bias term; α i is the sample weight; y i is the fault type;

[0187] The kernel function is:

[0188] K(x i ,x)=exp(-γ‖x i -x‖ 2 )

[0189] where γ is the control of kernel function width; ‖·‖ is the Euclidean distance; x i is the i-th sample variable; x is the input feature data.

[0190] Step 108: when the predicted fault type is insulation fault, an insulation fault alarm is issued.

[0191] In the embodiment of the present application, when the predicted fault type is obtained through the fault classification model, if the predicted fault type is insulation fault, an insulation fault alarm is issued, and the cable insulation fault detection is realized. The insulation fault includes electrical faults such as insulation breakdown and partial discharge; thermal performance faults of insulation overheating and aging; mechanical damage faults such as external force damage and excessive bending; and chemical corrosion and moisture faults.

[0192] The above embodiment has the following effects:

[0193] The present application provides a cable insulation fault detection method based on sheath circulating current, according to the preset data acquisition interval, the cable sheath circulating current data and the related branch load current data of the cable to be tested in the power system are synchronously collected; the sliding average method is used to smooth the cable sheath circulating current data, and the smoothed current data sequence is obtained; the load current data on a plurality of branches corresponding to each time is summed respectively, and the load current sum sequence is obtained; based on the smoothed current data sequence and the load current sum sequence, the instantaneous ratio resistance sequence is calculated; by analyzing whether the instantaneous ratio resistance sequence is abnormal, it is judged whether the cable to be tested has a fault risk; when it is determined that the cable to be tested has a fault risk, the electromagnetic characteristics and the thermal characteristics are extracted from the smoothed current data sequence and the instantaneous ratio resistance sequence by using the multi-physical field finite element simulation model; the electromagnetic characteristics and the thermal characteristics are input into the fault classification model, and the predicted fault type is obtained; when the predicted fault type is insulation fault, an insulation fault alarm is issued. The present application collects the cable sheath circulating current data and the related branch load current data of the cable to be tested, analyzes the collected information to determine whether there is a fault risk, realizes early warning of the fault, when there is a fault risk, rich feature information is extracted by using the multi-physical field finite element simulation model, and the fault classification model is used to predict the fault type, realizes rapid positioning of the fault, thereby improves the insulation fault detection accuracy, and ensures the safe and stable operation of the power system.

[0194] As shown in the above method item embodiment, corresponding device item embodiments are provided; Figure 3

[0195] An embodiment of the present application provides a cable insulation fault detection device based on sheath circulating current, which comprises a data acquisition module, a smoothing processing module, a current summation module, a resistance calculation module, a risk judgment module, a feature extraction module, a fault prediction module and a fault warning module.

[0196] ​The data collection module is configured to synchronously collect cable sheath loop current data of a to-be-tested cable and related branch load current data in a power system according to a preset data collection interval.

[0197] The smoothing processing module is configured to perform smoothing processing on the cable sheath loop current data by using a sliding average method to obtain a smoothed current data sequence.

[0198] The current summation module is configured to sum up load current data on a plurality of branches corresponding to each time point to obtain a load current summation sequence.

[0199] The resistance calculation module is configured to calculate an instantaneous ratio resistance sequence based on the smoothed current data sequence and the load current summation sequence.

[0200] The risk judgment module is configured to judge whether the to-be-tested cable has a fault risk by analyzing whether the instantaneous ratio resistance sequence is abnormal.

[0201] The feature extraction module is configured to extract electromagnetic features and thermal features from the smoothed current data sequence and the instantaneous ratio resistance sequence by using a multi-physical field finite element simulation model when it is determined that the to-be-tested cable has a fault risk.

[0202] The fault prediction module is configured to input the electromagnetic features and the thermal features into a fault classification model to obtain a predicted fault type.

[0203] The fault early warning module is configured to issue an insulation fault alarm when the predicted fault type is an insulation fault.

[0204] It can be understood that the above-mentioned device item embodiments correspond to the method item embodiments of the present application, and can realize the cable insulation fault detection method based on the sheath loop current provided by any one of the above-mentioned method item embodiments.

[0205] It should be noted that the device embodiments described above are only schematic, and part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. In addition, in the device embodiment provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0206] The above-described specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above-described specific embodiments are merely examples of the present application and are not intended to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for detecting cable insulation faults based on sheath circulating current, characterized in that, include: According to the preset data acquisition interval, the cable sheath circulating current data and related branch load current data of the cable under test in the power system are collected synchronously; wherein, the related branch load current data are the load current data of several branches connected in series with the cable under test. The circulating current data of the cable sheath was smoothed using the moving average method to obtain a smoothed current data sequence. The load current data of several branches corresponding to each time moment are summed to obtain the total load current sequence; Based on the smoothed current data sequence and the sum of the load currents, the instantaneous ratio resistance sequence is calculated; By analyzing whether the instantaneous ratio resistance sequence is abnormal, it can be determined whether the cable under test has a fault risk; When it is determined that the cable under test has a fault risk, electromagnetic and thermal features are extracted from the smooth current data sequence and the instantaneous ratio resistance sequence using a multiphysics finite element simulation model. The electromagnetic and thermal features are input into the fault classification model to obtain the predicted fault type. When the predicted fault type is an insulation fault, an insulation fault alarm is issued.

2. The cable insulation fault detection method based on sheath circulating current according to claim 1, characterized in that, The instantaneous ratio resistance sequence calculated based on the smoothed current data sequence and the sum of the load currents includes: Calculate the instantaneous ratio resistance at each moment in the instantaneous ratio resistance sequence using the following formula: In the formula, R(t) is the instantaneous ratio resistance at time t; I c (t) represents the circulating current in the cable sheath at time t; I n (t) is the sequence of the total load current at time t.

3. The cable insulation fault detection method based on sheath circulating current according to claim 1, characterized in that, The step of determining whether the cable under test has a fault risk by analyzing whether the instantaneous ratio resistance sequence is abnormal includes: Calculate the standard deviation of the instantaneous ratio resistance sequence and the load change rate, respectively; Based on the data standard deviation and the load change rate, determine whether the cable under test has a risk of failure.

4. The cable insulation fault detection method based on sheath circulating current according to claim 3, characterized in that, The step of determining whether the cable under test has a fault risk based on the data standard deviation and the load change rate includes: When the standard deviation of the data is greater than a preset standard deviation threshold or the load change rate is greater than a preset change rate threshold, it is determined whether the cable under test has a fault risk.

5. The cable insulation fault detection method based on sheath circulating current according to claim 1, characterized in that, The construction process of the multiphysics finite element simulation model is as follows: Obtain the topology data of the power system and construct a three-dimensional model based on the topology data; Acquire several types of fault data for the cable under test; wherein, the fault data includes fault parameters and fault types; The fault data is used to train the three-dimensional model to learn the fault characteristics corresponding to each fault type; wherein, the fault characteristics include fault electromagnetic characteristics and fault thermal characteristics; The trained 3D model is designated as a multiphysics finite element simulation model.

6. The cable insulation fault detection method based on sheath circulating current according to claim 5, characterized in that, The basic equations of the multiphysics finite element simulation model include Maxwell's equations, the heat conduction equation, the heat convection equation, the heat radiation equation, and the electromagnetic and thermal coupling equation. The Maxwell equations are as follows: In the formula, Here, is the vector differential operator; E is the electric field strength; B is the magnetic flux density; H is the magnetic field strength; J is the current density; D is the electric field displacement; ρ is the charge density; and t is the time variable. The heat conduction equation is: In the formula, k is the thermal conductivity of the material; T is the surface temperature of the object; and Q is the heat source per unit volume. The heat convection equation and the heat radiation equation are: q conv =h(T ∞ -T)A In the formula, q conv q represents the heat flux density during heat convection. rad ρ is the radiative heat flux density; h is the convective heat transfer coefficient; ∈ is the emissivity; σ is the Stefan-Boltzmann constant; T ∞ A represents the ambient temperature; A represents the normal vector of the fluid surface area. The electromagnetic-thermal coupling equation is as follows: In the formula, α is the thermal diffusivity.

7. The cable insulation fault detection method based on sheath circulating current according to claim 6, characterized in that, When it is determined that the cable under test has a fault risk, electromagnetic and thermal characteristics are extracted from the smoothed current data sequence and the instantaneous ratio resistance sequence using a multiphysics finite element simulation model, including: The smoothed current data sequence and the instantaneous ratio resistance sequence are input into the multiphysics finite element simulation model so that the multiphysics finite element simulation model can perform simulation and obtain the electromagnetic and thermal characteristics of the results. The simulation process of the multiphysics finite element simulation model is as follows: Extract the high-frequency components from the smoothed current data sequence, calculate the repetition rate of the high-frequency components, and obtain the high-frequency pulse repetition rate; Based on the instantaneous ratio resistance sequence, the contact resistance change rate of the cable under test is calculated, the fault time point is identified by the contact resistance change rate, and the resistance growth rate before and after the fault is calculated based on the fault time point. The spatial derivative of the magnetic flux density is extracted based on the smoothed current data sequence, and the spatial gradient of the magnetic flux density is calculated. The amplitude and duration of the pulse are obtained from the smoothed current data sequence, and the partial discharge pulse energy is calculated based on the amplitude and duration of the pulse. Extract hotspot temperatures and temperature gradients from the smoothed current data sequence; A temperature curve is fitted based on the smoothed current data sequence, and a thermal time constant is extracted from the temperature curve; The high-frequency pulse repetition rate, the resistance growth rate before and after the fault, the spatial gradient of magnetic flux density, and the partial discharge pulse energy are defined as electromagnetic characteristics. The hot spot temperature, the temperature gradient, and the thermal time constant are defined as thermal characteristics.

8. The cable insulation fault detection method based on sheath circulating current according to claim 7, characterized in that, The training process of the fault classification model is as follows: Several types of fault data are input into a multiphysics finite element simulation model to obtain the fault electromagnetic characteristics and fault thermal characteristics corresponding to each type of fault data. The electromagnetic features, thermal features, and fault types corresponding to each fault data are input into the support vector machine model, and the support vector machine model is trained to learn the mapping relationship between the electromagnetic features and thermal features of the fault and the fault type. The trained support vector machine model was selected as the fault classification model.

9. The cable insulation fault detection method based on sheath circulating current according to claim 8, characterized in that, The fault classification function of the fault classification model is: In the formula, f(x) is the fault classification function; K(x) i (x) is the kernel function; N is the total number of samples; b is the bias term; α i For sample weights; y i Fault type; The kernel function is: K(x i ,x)=exp(-γ‖x i -x‖ 2 ) In the formula, γ is the width of the control kernel function; ||·|| is the Euclidean distance; x i Let be the i-th sample variable; x is the input feature data.

10. A cable insulation fault detection device based on sheath circulating current, characterized in that, include: The system includes a data acquisition module, a smoothing module, a current summation module, a resistance calculation module, a risk assessment module, a feature extraction module, a fault prediction module, and a fault early warning module. The data acquisition module is used to synchronously acquire cable sheath circulating current data and related branch load current data of the cable under test in the power system according to a preset data acquisition interval; wherein, the related branch load current data are the load current data of several branches connected in series with the cable under test. The smoothing module is used to smooth the circulating current data of the cable sheath using the moving average method to obtain a smoothed current data sequence. The current summation module is used to sum the load current data of several branches corresponding to each time moment to obtain the total load current sequence. The resistance calculation module is used to calculate the instantaneous ratio resistance sequence based on the smoothed current data sequence and the sum of the load current sequence; The risk assessment module is used to determine whether the cable under test has a fault risk by analyzing whether the instantaneous ratio resistance sequence is abnormal. The feature extraction module is used to extract electromagnetic and thermal features from the smooth current data sequence and the instantaneous ratio resistance sequence using a multiphysics finite element simulation model when it is determined that the cable under test has a fault risk. The fault prediction module is used to input the electromagnetic features and the thermal features into the fault classification model to obtain the predicted fault type. The fault warning module is used to issue an insulation fault alarm when the predicted fault type is an insulation fault.

Images