A diagnostic method for insulation faults of power cables

Abstract

The present invention provides a method for diagnosing insulation faults of power cables, including collecting cable insulation parameters, preprocessing the cable insulation parameters to determine the true insulation state of the power cable; calculating the prior probability of nodes in the fault layer of a preset cable insulation network model according to the true insulation state of the power cable, and correcting the cable insulation network model according to the prior probability; diagnosing the current fault type of the cable through the corrected cable insulation network model to obtain a final diagnosis result. In the present invention, according to the actual working conditions of the cable insulation state, a variety of power cable monitoring methods are used as the fault characterization layer, and by resetting the prior probability of the fault characterization nodes, the probability of the parent nodes in the model fault layer can be corrected, truly reflecting the reasons for cable insulation faults, providing guidance for cable maintenance and repair work, and providing guidance for the maintenance cycle and key inspection parts.

Description

Technical Field

[0001] The present invention relates to the technical field of diagnosing insulation faults of power cables, and particularly to a method for diagnosing insulation faults of power cables. Background Art

[0002] Power cables are one of the most widely used power transmission carriers in modern power systems. Due to their wide coverage area, complex line structures, and the mixing of cable batches and processes during operation, existing cable insulation performance monitoring means cannot well cover various types of insulation faults. The insulation performance of power cables is affected by multiple factors, such as line overload, air gap ionization in the cable insulation layer, heat recession problems caused by poor connection processes; mechanical damage problems caused by construction; insulation dampness problems caused by humid environments and sealing processes; insulation degradation problems caused by thermal aging, electrical aging, and chemical corrosion; and cable joint surface flashovers caused by induced lightning overvoltage and switching overvoltage, etc., all of which may cause irreversible damage to the cable insulation.

[0003] There are many existing cable insulation performance monitoring means. The DC superposition method can judge the cable insulation damage caused by through-water treeing, but it does not have a good detection effect on cable insulation damage that is not through or relatively severely aged as a whole; while the monitoring method using the tangent value of the dielectric loss angle can effectively reflect the overall insulation aging degree of the cable and can reflect the cable insulation damage characteristics caused by thermal aging and dampness, but this method cannot reflect local insulation deterioration of the cable and insulation performance degradation such as surface flashover at the cable joints; the ground wire current method can effectively verify the relationship between the capacitive current and the cable insulation state and thus judge the cable insulation performance, but the implementation of this method depends too much on the historical data collection of the same cable and requires engineering implementation in combination with probability statistics; the partial discharge method can detect local defects in the cable insulation when applied to cable lines and relatively accurately locate the defect positions, but in actual work, its accuracy is greatly lost due to noise or electromagnetic signal interference. To make up for this defect, it is mostly combined with the DC superposition method or the tangent value detection method of the dielectric loss angle for application.

[0004] Among the numerous methods for cable insulation condition assessment, there are a series of defects caused by the method itself, which affect the universality of cable insulation condition assessment. Moreover, among the construction of many existing cable assessment models, they mostly target single measurement methods. For example, based on the Canny algorithm or the K-means clustering algorithm, background noise filtering and precise extraction of required target quantities are carried out based on the infrared partial discharge method. However, such methods are limited by the complex branch structure of the cable and the distribution mode of the front-end equipment in the pipe well, and have great limitations in application. Developing a method that can, according to the actual state of cable insulation, use the fault characterization layer constructed by multiple online monitoring methods as the posterior event probability and conduct real-time analysis and diagnosis of the insulation fault occurrence mechanism has practical engineering significance for improving cable insulation safety situation awareness and maintenance. Summary of the Invention

[0005] The object of the present invention is to propose a diagnostic method for power cable insulation faults, and solve the technical problem of how to conduct real-time analysis and diagnosis of the insulation fault occurrence mechanism and improve cable insulation safety situation awareness.

[0006] On the one hand, a diagnostic method for power cable insulation faults is provided, including:

[0007] Collect cable insulation parameters, preprocess the cable insulation parameters, and determine the true insulation state of the power cable;

[0008] Calculate the prior probability of the nodes of the fault layer in the preset cable insulation network model according to the true insulation state of the power cable, and correct the cable insulation network model according to the prior probability;

[0009] Diagnose the current cable fault type through the corrected cable insulation network model to obtain the final diagnostic result.

[0010] Preferably, the preset cable insulation network model includes:

[0011] A fault layer for classifying cable insulation fault types, and the nodes of the fault layer at least include cable insulation moisture absorption, cable insulation aging, cable insulation breakdown, flashover along the surface of the cable joint, and cable fire;

[0012] A characterization layer for characterizing cable insulation fault parameter information corresponding to the cable insulation fault type, and the nodes of the characterization layer at least include the direct current component of the cable insulation layer, the tangent value of the dielectric loss angle of the cable insulation, the capacitive current value of the cable grounding wire, the local pulse discharge current value of the cable insulation, the temperature value in the cable pipe well, and the water immersion angle in the cable pipe well.

[0013] Preferably, the preprocessing of the cable insulation parameters includes:

[0014] Normalize the cable insulation parameters through a preset semi-trapezoidal model and conduct real-time evaluation of the status information.

[0015] Preferably, the cable insulation parameters are processed through the following semi-trapezoidal model:

[0016]

[0017] where x is the real-time status quantity of the power cable insulation performance among the cable insulation parameters, V(x) is the generalization evaluation result of the semi-trapezoidal model, and ω na , ω nb are respectively the safety thresholds characterizing the power cable insulation performance parameters.

[0018] Preferably, calculating the prior probability of the nodes in the fault layer of the preset cable insulation network model according to the real insulation state of the power cable includes:

[0019] Determine the ternary coding values of the nodes in the cable insulation fault characterization layer according to the real insulation state of the power cable;

[0020] Determine the corresponding external symptom types according to the ternary coding values, and reset the prior probability of the nodes in the fault layer of the cable insulation network model with the external symptom types.

[0021] Preferably, determine the ternary coding values of the nodes in the cable insulation fault characterization layer according to the following formula:

[0022]

[0023] where Ω(V n ) is the ternary coding value, x is the real-time status quantity of the power cable insulation performance among the cable insulation parameters, and V n is the corresponding external symptom type.

[0024] Preferably, the external symptom types include:

[0025] DC current component of the cable insulation layer, and its corresponding external symptom type is V1;

[0026] Tangent value of the dielectric loss angle of the cable insulation, and its corresponding external symptom type is V2;

[0027] Capacitive current value of the cable grounding wire, and its corresponding external symptom type is V3;

[0028] Local impulse discharge current value of the cable insulation, and its corresponding external symptom type is V4;

[0029] Temperature value in the cable pipe well, and its corresponding external symptom type is V5;

[0030] The water immersion angle in the cable pipe shaft, and its corresponding external symptom type is V6.

[0031] Preferably, reset the prior probability of the fault layer nodes in the cable insulation network model according to the following formula:

[0032]

[0033] Wherein, pa(V n ) is the fault layer node, and P(pa(V n )) is the prior probability of the fault layer node.

[0034] Preferably, it further includes

[0035] Taking the actual working state information of the cable insulation state as the inference evidence, updating the fault layer and diagnosing the current fault type of the cable by calculating the marginal conditional probability of the fault layer nodes;

[0036] Wherein, the fault layer nodes at least include cable insulation moisture absorption S1, cable insulation aging S2, cable insulation breakdown S3, flashover along the surface of the cable joint S4, and cable fire S5.

[0037] In summary, implementing the embodiments of the present invention has the following beneficial effects:

[0038] The diagnostic method for power cable insulation faults provided by the present invention, according to the actual working conditions of the cable insulation state, uses multiple power cable monitoring methods as the fault characterization layer, and by resetting the prior probability of the fault characterization nodes, can correct the probability of the parent nodes of the model fault layer, truly reflect the causes leading to cable insulation faults, provide guidance for cable maintenance and repair work, and provide guidance for the maintenance cycle and key inspection parts. Integrating the advantages of multiple insulation monitoring methods and complementing each other's defects, a complete insulation state evaluation method for power cables is given. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, obtaining other drawings without creative efforts still belongs to the scope of the present invention.

[0040] Figure 1 It is a main flow schematic diagram of a diagnostic method for power cable insulation faults in an embodiment of the present invention.

[0041] Figure 2 It is a schematic diagram of a cable insulation network model in an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0042] To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0043] As Figure 1 shown, it is a schematic diagram of an embodiment of a method for diagnosing insulation faults of power cables provided by the present invention. In this embodiment, the method includes the following steps:

[0044] Step S1, collect cable insulation parameters, and preprocess the cable insulation parameters to determine the true insulation state of the power cable; that is, collect cable insulation parameters through a cable insulation data monitoring system, perform normalization processing on the original data in the data sample and constrain it between 0 and 1, and use a semi-ladder model to perform real-time evaluation of the status information of the power cable insulation monitoring parameters.

[0045] In a specific embodiment, a cable insulation network model needs to be established before preprocessing the cable insulation parameters. This model is used to evaluate power cable insulation faults; among them, as Figure 2 shown, the preset cable insulation network model includes:

[0046] A fault layer, used to classify cable insulation fault types. The nodes of the fault layer at least include cable insulation moisture absorption, cable insulation aging, cable insulation breakdown, surface flashover of cable joints, and cable fire;

[0047] A characterization layer, used to characterize cable insulation fault parameter information corresponding to the cable insulation fault types. The nodes of the characterization layer at least include the DC current component of the cable insulation layer, the tangent value of the dielectric loss angle of the cable insulation, the capacitive current value of the cable grounding wire, the local impulse discharge current value of the cable insulation, the temperature value in the cable pipe well, and the water immersion angle in the cable pipe well.

[0048] It can be understood that a Bayesian network evaluation model for insulation faults of 35 kV and below cross-linked polyethylene power cables is established. The cable insulation fault types are classified as: cable insulation moisture absorption, cable insulation aging, cable insulation breakdown, surface flashover of cable joints, and cable fire and used as the fault layer; the DC current component of the cable insulation layer, the tangent value of the dielectric loss angle of the cable insulation, the capacitive current value of the cable grounding wire, the local impulse discharge current value of the cable insulation, the temperature value in the cable pipe well, and the water immersion angle in the cable pipe well that characterize the cable insulation faults are used as the characterization layer of the cable insulation faults.

[0049] Specifically, the preprocessing of the cable insulation parameters includes: normalizing the cable insulation parameters through a preset semi-trapezoidal model and performing real-time evaluation of the status information. That is: the insulation monitoring parameters of the power cable are subjected to real-time evaluation of the status information using the semi-trapezoidal model, where x is the real-time status quantity of the insulation performance of the power cable, and V(x) is the generalized evaluation result of the semi-trapezoidal model, representing the corresponding external symptom description that meets the performance parameters, ω na and ω nb are respectively the safety thresholds characterizing the insulation performance parameters of the power cable. When it is lower than ω na , it is in good condition. When it is between , it is in a state that needs attention. When it is higher than ω nb , it is in a bad state. Among them, the cable insulation parameters are processed through the following semi-trapezoidal model:

[0050]

[0051] where x is the real-time status quantity of the insulation performance of the power cable in the cable insulation parameters, V(x) is the generalized evaluation result of the semi-trapezoidal model, and ω na and ω nb are respectively the safety thresholds characterizing the insulation performance parameters of the power cable. Then, the insulation performance parameters of the power cable are corrected and synthesized to judge the true insulation state e of the power cable.

[0052] Step S2: Calculate the prior probability of the nodes in the fault layer of the preset cable insulation network model according to the true insulation state of the power cable, and correct the cable insulation network model according to this prior probability; that is, through the actual working state obtained by the cable insulation data monitoring system, the nodes in the cable insulation fault characterization layer are ternary coded and valued using formula (2), and the prior probability P(pa(V n )) of the nodes in the fault layer of the cable insulation Bayesian network model is reset. n )

[0053] In a specific embodiment, calculating the prior probability of the nodes in the fault layer of the preset cable insulation network model according to the true insulation state of the power cable includes: determining the ternary coding and value of the nodes in the cable insulation fault characterization layer according to the true insulation state of the power cable; determining the corresponding external symptom type according to the ternary coding and value, and resetting the prior probability of the nodes in the fault layer of the cable insulation network model according to this external symptom type. Among them, the ternary coding and value of the nodes in the cable insulation fault characterization layer are determined according to the following formula:

[0054]

[0055] where Ω(V n) is a ternary coding value, x is the real-time state variable of the power cable insulation performance among the cable insulation parameters, and V n is the corresponding external symptom type.

[0056] Specifically, the external symptom types include:

[0057] The DC current component of the cable insulation layer, and its corresponding external symptom type is V1;

[0058] The tangent value of the dielectric loss angle of the cable insulation, and its corresponding external symptom type is V2;

[0059] The capacitive current value of the cable grounding wire, and its corresponding external symptom type is V3;

[0060] The partial pulse discharge current value of the cable insulation, and its corresponding external symptom type is V4;

[0061] The temperature value in the cable pipe well, and its corresponding external symptom type is V5;

[0062] The water immersion angle in the cable pipe well, and its corresponding external symptom type is V6.

[0063] It can be understood that through the actual working state obtained by the cable insulation data monitoring system, the nodes of the cable insulation fault characterization layer are ternary coded using the above formula, and 0, 1, and 2 respectively represent the judgments on the cable insulation fault that occur in the fault symptom layer: good, attention required, and bad. The expectation maximization algorithm (EM) is used to determine the maximum likelihood estimate value of the performance parameters that cause insulation faults and obtain the conditional probabilities between nodes: According to the data obtained from the data set and the current parameter values, calculate the logarithmic likelihood expectation of the sample data:

[0064]

[0065] Then calculate the probability of the fault type node occurring under the characterization layer node:

[0066] θ = arg θ maxQ(θ, θ(t))

[0067] Through repeated iteration, a stable solution is finally obtained, and the conditional probability of the characterization layer occurring in the true fault layer is obtained.

[0068] And, reset the prior probability of the fault layer node in the cable insulation network model according to the following formula:

[0069]

[0070] Among them, pa(V n ) is the fault layer node, and P(pa(V n )) is the prior probability of the fault layer node.

[0071] Step S3: Diagnose the current cable fault type through the corrected cable insulation network model to obtain the final diagnosis result. That is, after determining the prior probability, the probability of the parent node of the model fault layer in the original cable insulation network model is corrected to truly reflect the cause of the cable insulation fault.

[0072] In a specific embodiment, it further includes using the actual working state information of the cable insulation state as inference evidence, updating the fault layer, and diagnosing the current cable fault type by calculating the marginal conditional probability of the fault layer nodes. Among them, the fault layer nodes at least include cable insulation moisture absorption S1, cable insulation aging S2, cable insulation breakdown S3, flashover along the surface of the cable joint S4, and cable fire S5. It can be understood that a Bayesian network model of the cable insulation state is established using Netica joint tree. The actual working state information e of the cable insulation state is used as the inference evidence, and cable insulation moisture absorption S1, cable insulation aging S2, cable insulation breakdown S3, flashover along the surface of the cable joint S4, and cable fire S5 are used as the fault layer. The current cable fault type is diagnosed by calculating the marginal conditional probability p(Sn|e) of the fault layer node Sn.

[0073] In summary, implementing the embodiments of the present invention has the following beneficial effects:

[0074] The diagnostic method for power cable insulation faults provided by the present invention, according to the actual working conditions of the cable insulation state, uses a variety of power cable monitoring methods as the fault characterization layer, and by resetting the prior probability of the fault characterization nodes, the probability of the parent node of the model fault layer can be corrected to truly reflect the cause of the cable insulation fault, providing guidance for cable maintenance and repair work, and providing guidance for the maintenance cycle and key inspection parts. By integrating the advantages of various insulation monitoring methods and complementing each other's defects, a complete insulation state evaluation method for power cables is given.

[0075] The above-disclosed are only the preferred embodiments of the present invention. Of course, the scope of the rights of the present invention cannot be limited by this. Therefore, equivalent changes made according to the claims of the present invention still fall within the scope covered by the present invention.

Claims

1. A diagnostic method for power cable insulation faults, characterized in that, Including: Collecting cable insulation parameters, preprocessing the cable insulation parameters, and determining the true insulation state of the power cable; Calculating the prior probability of the nodes of the fault layer in the preset cable insulation network model according to the true insulation state of the power cable, and correcting the cable insulation network model according to the prior probability; Diagnosing the current cable fault type through the corrected cable insulation network model to obtain the final diagnosis result; Among them, the calculating the prior probability of the nodes of the fault layer in the preset cable insulation network model according to the true insulation state of the power cable includes: Determining the ternary coding values of the nodes of the cable insulation fault characterization layer according to the true insulation state of the power cable; Determining the corresponding external symptom type according to the ternary coding values, and resetting the prior probability of the nodes of the fault layer in the cable insulation network model according to the external symptom type; Resetting the prior probability of the nodes of the fault layer in the cable insulation network model according to the following formula: Among them, pa(V n ) is the node of the fault layer, and P(pa(V n )) is the prior probability of the node of the fault layer, where V n is the corresponding external symptom type.

2. The method according to claim 1, wherein The preset cable insulation network model includes: A fault layer for classifying cable insulation fault types, and the nodes of the fault layer at least include cable insulation moisture absorption, cable insulation aging, cable insulation breakdown, flashover along the surface of the cable joint, and cable fire; A characterization layer for characterizing cable insulation fault parameter information corresponding to the cable insulation fault type, and the nodes of the characterization layer at least include the DC current component of the cable insulation layer, the tangent value of the dielectric loss angle of the cable insulation, the capacitive current value of the cable grounding wire, the local impulse discharge current value of the cable insulation, the temperature value in the cable shaft, and the water immersion angle in the cable shaft.

3. The method according to claim 2, wherein The preprocessing of the cable insulation parameters includes: Normalizing the cable insulation parameters through a preset half-ladder model and performing real-time evaluation of the state information.

4. The method according to claim 3, wherein Processing the cable insulation parameters through the following half-ladder model: Among them, x is the real-time state quantity of the power cable insulation performance in the cable insulation parameters, V(x) is the generalization evaluation result of the semi-trapezoidal model, and ω na , ω nb are the safety thresholds representing the power cable insulation performance parameters respectively.

5. The method according to claim 4, wherein Determining the ternary coding values of the nodes of the cable insulation fault characterization layer according to the following formula: Among them, Ω(V n ) is the ternary coding value, and x is the real-time state quantity of the power cable insulation performance among the cable insulation parameters.

6. The method according to claim 5, wherein The external symptom types include: The DC current component of the cable insulation layer, and its corresponding external symptom type is V1; The tangent value of the dielectric loss angle of the cable insulation, and its corresponding external symptom type is V2; The capacitive current value of the cable grounding wire, and its corresponding external symptom type is V3; The local impulse discharge current value of the cable insulation, and its corresponding external symptom type is V4; The temperature value in the cable shaft, and its corresponding external symptom type is V5; The water immersion angle in the cable shaft, and its corresponding external symptom type is V6.

7. The method according to claim 1, wherein It also includes Taking the actual working state information of the cable insulation state as the inference evidence, updating the fault layer, and diagnosing the current cable fault type by calculating the marginal conditional probability of the nodes of the fault layer; Among them, the nodes of the fault layer at least include cable insulation moisture absorption S1, cable insulation aging S2, cable insulation breakdown S3, flashover along the surface of the cable joint S4, and cable fire S5.

Images