GIS fault diagnosis and identification method based on multi-source information fusion

Abstract

The embodiment of the invention discloses a GIS fault diagnosis and identification method based on multi-source information fusion. The method comprises the steps that GIS signals are collected through a plurality of sensors installed on GIS equipment; extracting a plurality of pieces of feature information based on the GIS signal; obtaining a fuzzy rule base, wherein the fuzzy rule base comprises a matching relationship between the feature information corresponding to the historical GIS signal and the fault type; fuzzifying the plurality of pieces of feature information to obtain fuzzified feature information; determining total fuzzy output of the fuzzification feature information based on the fuzzy rule base, wherein the total fuzzy output comprises multi-dimensional feature information; and performing defuzzification on the total fuzzy output to determine the GIS equipment fault type. According to the invention, a plurality of sensors installed on GIS equipment are used for acquiring signals, and the sensors acquire equipment operation data from different angles and layers, so that multi-source data fusion is realized. The multi-source data can reflect the operation state of the equipment more comprehensively, and the risk of fault misjudgment is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power failure identification, and particularly to a method for GIS fault diagnosis and identification based on multi-source information fusion. Background Art

[0002] With the continuous development of the power system, GIS has been widely used in the power system due to its advantages such as small floor area, high reliability, and convenient maintenance. However, various faults may occur during the operation of GIS, such as partial discharge, insulation breakdown, mechanical faults, etc. These faults will not only affect the normal operation of the power system, but may also cause serious safety accidents, resulting in large-scale power outages, bringing huge losses to social production and life.

[0003] Traditional detection means rely only on partial discharge detection or single-sensor monitoring, and it is difficult to comprehensively and accurately diagnose the fault type in this way. Summary of the Invention

[0004] Based on this, it is necessary to propose a method for GIS fault diagnosis and identification based on multi-source information fusion for the above problems.

[0005] A method for GIS fault diagnosis and identification based on multi-source information fusion, the method includes the following steps:

[0006] Collect GIS signals through a plurality of sensors installed on the GIS device;

[0007] Extract a plurality of characteristic information based on the GIS signals;

[0008] Obtain a fuzzy rule base, where the fuzzy rule base includes the matching relationship between the characteristic information corresponding to the historical GIS signals and the fault types;

[0009] Fuzzify the plurality of characteristic information to obtain fuzzified characteristic information;

[0010] Determine the total fuzzy output based on the fuzzy rule base for the fuzzified characteristic information, and the total fuzzy output includes multi-dimensional characteristic information;

[0011] Defuzzify the total fuzzy output to determine the fault type of the GIS device.

[0012] In the above solution, the collecting GIS signals through a plurality of sensors installed on the GIS device specifically includes:

[0013] The plurality of sensors include partial discharge sensors, vibration sensors, temperature sensors, and gas sensors;

[0014] The partial discharge sensor, vibration sensor, temperature sensor, and gas sensor synchronously collect GIS signals.

[0015] In the above solution, the partial discharge sensor, vibration sensor, temperature sensor, and gas sensor synchronously collect GIS signals, specifically including:

[0016] The partial discharge sensor collects the partial discharge signal PD(t) of the GIS device according to the first sampling frequency f1;

[0017] The vibration sensor collects the vibration signal V(t) of the GIS device according to the second sampling frequency f2;

[0018] The temperature sensor collects the temperature signal T(t) of the GIS device according to the third sampling frequency f3;

[0019] The gas sensor collects the gas signal G(t) of the GIS device according to the fourth sampling frequency f4.

[0020] In the above solution, before extracting several characteristic information based on the GIS signal, the method further includes:

[0021] Denoise the partial discharge signal PD(t) and perform N-layer decomposition on the denoised signal;

[0022] Reconstruct the decomposed signal based on a preset threshold to obtain the optimal partial discharge signal PD d (t);

[0023] According to the passband frequency range [f l , f h filter the vibration signal V(t) to obtain the optimal vibration signal V f (t);

[0024] Perform data cleaning on the temperature signal T(t) and the gas signal G(t) to obtain the corrected temperature signal T1(t) and the corrected gas signal G1(t).

[0025] In the above solution, extracting several characteristic information based on the GIS signal specifically includes:

[0026] Convert the optimal partial discharge signal PD d (t) to obtain the discharge amount Q, discharge phase discharge repetition rate f r ;

[0027] Perform time-frequency analysis on the optimal vibration signal Vf ( t) to obtain the vibration energy E, frequency component F, and vibration amplitude A;

[0028] Calculate the temperature change rate α of the temperature signal based on the corrected temperature signal T1(t);

[0029] Determine the change rate β of the decomposition product content according to the key components of the corrected gas signal G1(t).

[0030] In the above solution, the process of fuzzifying the several pieces of characteristic information to obtain the fuzzified characteristic information specifically includes:

[0031] According to the discharge quantity Q, discharge phase discharge repetition rate f r and the actual change ranges of vibration energy E, frequency component E, vibration amplitude A, temperature change rate α, and decomposition product content change rate β, obtain several corresponding fuzzy sets;

[0032] Respectively determine the membership functions corresponding to the discharge quantity Q, discharge phase discharge repetition rate f r vibration energy E, frequency component F, vibration amplitude A, temperature change rate α, and decomposition product content change rate β;

[0033] Based on the corresponding membership functions, calculate the membership values of the discharge quantity Q, discharge phase discharge repetition rate f r vibration energy E, frequency component F, vibration amplitude A, temperature change rate α, and decomposition product content change rate β belonging to each fuzzy set;

[0034] Determine the fuzzified characteristic information according to the membership values of each fuzzy set.

[0035] In the above solution, the process of determining the total fuzzy output based on the fuzzy rule base for the fuzzified characteristic information specifically includes:

[0036] Based on the membership values of each characteristic information on the corresponding fuzzy sets, determine the credibility of each rule according to the fuzzy logic operation rules;

[0037] According to the credibility of each rule and the output fuzzy set of the rule, calculate the output membership degree of each rule;

[0038] Aggregate the output membership degrees of all rules to obtain the total fuzzy output.

[0039] In the above solution, the process of defuzzifying the total fuzzy output to determine the GIS equipment fault type specifically includes:

[0040] Defuzzify the total fuzzy output, calculate the centroid position under the membership function curve of the fuzzy output, and obtain the fault type judgment score. The abscissa of the membership function curve of the fuzzy output represents the fault type, the ordinate represents the membership value, and the coordinate value represents the judgment score of the corresponding fault type;

[0041] Compare the judgment scores of different fault types, and take the fault type with the highest judgment score as the final fault type diagnosis result.

[0042] In the above solution, after defuzzifying the total fuzzy output to determine the fault type of the GIS device, the method further includes:

[0043] Collect the time difference of the partial discharge signal reaching different sensors;

[0044] Determine the propagation speed of the partial discharge signal in the GIS device;

[0045] Calculate the distance from the current fault point to each sensor;

[0046] According to the calculated distances from the fault point to each sensor and the known positions of the sensors, use the geometric positioning principle to determine the position of the fault point.

[0047] The present application also proposes a system for GIS fault diagnosis and identification based on multi-source information fusion, including:

[0048] A sensor module; configured in the GIS device for collecting partial discharge, vibration, temperature and gas signals;

[0049] A data preprocessing module; used for preprocessing the signals collected by the sensor module, including denoising, filtering and cleaning;

[0050] A feature extraction module; used for extracting the discharge amount, discharge phase, discharge repetition rate, vibration energy, frequency component, vibration amplitude, temperature change rate and gas decomposition product content change rate from the preprocessed signals;

[0051] A fuzzy inference module; including a fuzzy rule base and an inference engine, used to obtain the fuzzy rule base, the fuzzy rule base includes the matching relationship between the characteristic information corresponding to the historical GIS signal and the fault type, and fuzzify the extracted characteristic information, and determine the total fuzzy output based on the fuzzy rule base;

[0052] A defuzzification module; used for defuzzifying the total fuzzy output output by the fuzzy inference module to determine the fault type of the GIS device.

[0053] The present application also proposes a readable storage medium storing a computer program, which when executed by a processor causes the processor to perform the following steps:

[0054] Collect GIS signals through a number of sensors installed on the GIS device;

[0055] Extract a number of characteristic information based on the GIS signals;

[0056] Obtain a fuzzy rule base, which includes the matching relationship between the characteristic information corresponding to the historical GIS signals and the fault types;

[0057] Fuzzify the number of characteristic information to obtain fuzzified characteristic information;

[0058] Determine the total fuzzy output based on the fuzzy rule base for the fuzzified characteristic information, and the total fuzzy output includes multi-dimensional characteristic information;

[0059] Defuzzify the total fuzzy output to determine the fault type of the GIS device.

[0060] This application also proposes a computer device, including a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor as follows:

[0061] Collect GIS signals through a number of sensors installed on the GIS device;

[0062] Extract a number of characteristic information based on the GIS signals;

[0063] Obtain a fuzzy rule base, which includes the matching relationship between the characteristic information corresponding to the historical GIS signals and the fault types;

[0064] Fuzzify the number of characteristic information to obtain fuzzified characteristic information;

[0065] Determine the total fuzzy output based on the fuzzy rule base for the fuzzified characteristic information, and the total fuzzy output includes multi-dimensional characteristic information;

[0066] Defuzzify the total fuzzy output to determine the fault type of the GIS device.

[0067] Adopting the embodiments of the present invention has the following beneficial effects: By collecting GIS signals from different dimensions through multiple sensors installed on GIS devices, the operating information of the devices can be comprehensively obtained, providing a rich data basis for subsequent analysis; Based on multiple feature information extracted from these signals, the device status can be more accurately reflected; The obtained fuzzy rule base stores the matching relationship between historical GIS signal feature information and fault types, providing an empirical basis for fault identification; Fuzzifying the feature information effectively addresses the uncertainty and ambiguity of feature information in actual operation, making the processing more in line with the actual situation; Finally, based on the fuzzy rule base, reasoning is performed on the fuzzified feature information to obtain a total fuzzy output containing multi-dimensional feature information. This process comprehensively considers the mutual relationship between each feature and its impact on faults, realizing the comprehensive judgment of multi-dimensional features. Finally, the total fuzzy output is defuzzified to determine the fault type, which not only utilizes the advantage of fuzzy logic in dealing with uncertainty but also obtains a clear fault judgment result through defuzzification. The whole process realizes the effective fusion and utilization of multi-source data, fully exerts the ability of fuzzy logic to handle complex problems, thereby significantly improving the accuracy and efficiency of a GIS fault diagnosis and identification method based on multi-source information fusion, and reducing the misjudgment rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] 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 the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0069] Among them:

[0070] Figure 1 It is a schematic flow chart of a GIS fault diagnosis and identification method based on multi-source information fusion in an embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0071] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all of them; Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0072] In the following description, numerous specific details are given to provide a more thorough understanding of the present invention; however, it will be apparent to those skilled in the art that the present invention may be practiced without one or more of these details; in other instances, to avoid obscuring the present invention, some technical features well known in the art are not described. It should be understood that the present invention can be implemented in different forms and should not be construed as limited to the embodiments presented herein; on the contrary, providing these embodiments will make the disclosure thorough and complete and will fully convey the scope of the present invention to those skilled in the art.

[0073] The purpose of the terms used herein is only to describe specific embodiments and is not a limitation of the present invention. As used herein, the singular forms "a", "an", and "the" are also intended to include the plural forms unless the context clearly dictates otherwise. It should also be understood that the terms "comprising" and / or "including", when used in this specification, determine the presence of the stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups. As used herein, the term "and / or" includes any and all combinations of the related listed items.

[0074] To thoroughly understand the present invention, detailed structures will be presented in the following description to illustrate the technical solutions proposed by the present invention; the optional embodiments of the present invention are described in detail below. However, in addition to these detailed descriptions, the present invention may also have other embodiments.

[0075] As Figure 1 shown, in one embodiment, a method for GIS fault diagnosis and identification based on multi-source information fusion is provided. The method for GIS fault diagnosis and identification based on multi-source information fusion includes steps S101 to S106, which are described in detail as follows:

[0076] S101. Collect GIS signals through a plurality of sensors installed on the GIS device;

[0077] Multiple sensors can monitor the GIS device from different angles and dimensions. For example, a temperature sensor can sense the temperature change of the device in real time, a pressure sensor can monitor the gas or liquid pressure inside the device, and a vibration sensor can detect the mechanical vibration condition of the device, etc. These different types of signals are combined to comprehensively reflect the operating state of the GIS device, providing a rich data basis for subsequent fault identification.

[0078] Through the collaborative work of multiple sensors, these abnormal signals can be detected more sensitively, thereby timely discovering potential faults in the device and avoiding the further expansion of faults.

[0079] In some embodiments, GIS signals are collected by a number of sensors installed on GIS devices, specifically including:

[0080] The number of sensors includes partial discharge sensors, vibration sensors, temperature sensors, and gas sensors;

[0081] The partial discharge sensors, vibration sensors, temperature sensors, and gas sensors collect GIS signals synchronously.

[0082] Among them, the acquisition time interval is Δt.

[0083] In some embodiments, the partial discharge sensors, vibration sensors, temperature sensors, and gas sensors collect GIS signals synchronously, specifically including:

[0084] The partial discharge sensor collects the partial discharge signal PD(t) of the GIS device according to the first sampling frequency f1;

[0085] The vibration sensor collects the vibration signal V(t) of the GIS device according to the second sampling frequency f2;

[0086] The temperature sensor collects the temperature signal T(t) of the GIS device according to the third sampling frequency f3;

[0087] The gas sensor collects the gas signal G(t) of the GIS device according to the fourth sampling frequency f4.

[0088] Preferably, the sampling frequency (the first sampling frequency) f1 of the partial discharge sensor = 50 MHz, the sampling frequency (the second sampling frequency) f2 of the vibration sensor = 10 kHz, the sampling frequency (the third sampling frequency) f3 of the temperature sensor = 1 Hz, the sampling frequency (the fourth sampling frequency) f4 of the gas sensor = 0.1 Hz, and the acquisition time interval Δt = 1 s.

[0089] S102. Extract a number of characteristic information based on the GIS signals;

[0090] Extracting characteristic information from the collected GIS signals can remove irrelevant data interference and focus on the key characteristics closely related to equipment failures. For example, by analyzing characteristics such as the frequency, amplitude, and change trend of the signals, characteristic parameters reflecting abnormal equipment operation states can be extracted, and these characteristic parameters can more accurately reflect the fault conditions of the equipment.

[0091] In addition, the amount of original GIS signal data is usually large, and directly processing it will face problems such as a large amount of calculation and a long processing time. Extracting characteristic information can convert a large amount of original data into a small number of key characteristics, reduce the data dimension, thereby reducing the complexity of data processing and improving the efficiency of fault identification.

[0092] In some embodiments, before extracting a plurality of characteristic information based on the GIS signal, the method further includes:

[0093] Denoise the partial discharge signal PD(t) and decompose the denoised signal into N layers;

[0094] Reconstruct the decomposed signal based on a preset threshold to obtain the optimal partial discharge signal PD d (t);

[0095] Filter the vibration signal V(t) according to the passband frequency range [f l , f h to obtain the optimal vibration signal V f (t);

[0096] Clean the data of the temperature signal T(t) and the gas signal G(t) to obtain the corrected temperature signal T1(t) and the corrected gas signal G1(t).

[0097] Preferably, for the temperature signal, if the deviation between the temperature value at a certain moment and the temperature value at the adjacent moment exceeds the set threshold Δt = 10°C, it is determined as an abnormal value and corrected; for the gas signal, if the content value of a certain gas component is missing or significantly abnormal, interpolation is performed according to historical data or data of adjacent devices.

[0098] Preferably, the wavelet threshold denoising method is used to denoise the partial discharge signal PD(t). After selecting the wavelet basis function and performing N-layer decomposition to determine the threshold, PD d (t) is reconstructed; the Butterworth band-pass filter is used to filter the V(t), and the passband frequency range is set as [f l , f h to obtain V f (t); the data of T(t) and G(t) are cleaned to remove abnormal values and missing values, and the corrected temperature signal T1(t) and the corrected gas signal G1(t) are obtained.

[0099] Among them, in the wavelet threshold denoising method, the Daubechies wavelet basis 'db4' is selected, the decomposition layer number N = 5, and the passband frequency range of the Butterworth band-pass filter is [100lHz, 1kHz].

[0100] In some embodiments, extracting a plurality of characteristic information based on the GIS signal specifically includes:

[0101] Convert the optimal partial discharge signal PD d (t) to obtain the discharge amount Q, the discharge phase the discharge repetition rate f r ;

[0102] Perform time-frequency analysis on the optimal vibration signal V f (t) to obtain the vibration energy E, frequency components F, and vibration amplitude A;

[0103] Calculate the temperature change rate α of the temperature signal according to the corrected temperature signal T1(t);

[0104] Determine the change rate β of the decomposition product content according to the key components of the corrected gas signal G1(t).

[0105] Preferably, calculate the discharge quantity Q by integrating PD d (t) within a discharge cycle, and the discharge phase is determined according to the phase relationship between the occurrence time of partial discharge and the power supply voltage, and the discharge repetition rate f r is calculated by counting the number of discharge pulses per unit time.

[0106] Specifically, perform short-time Fourier transform on the optimal vibration signal V f (t) to obtain a time-frequency diagram, and extract the vibration energy E, frequency components F, and vibration amplitude A from it. The vibration energy E is calculated by integrating the energy of each frequency component in the time-frequency diagram, the frequency components F are determined by analyzing the frequency spectrum distribution of the time-frequency diagram, and the vibration amplitude A is obtained by finding the maximum amplitude of the vibration signal in the time-frequency diagram.

[0107] Calculate the temperature change rate α of the temperature signal through the following formula:

[0108]

[0109] where Δt is the acquisition time interval.

[0110] S103. Obtain a fuzzy rule base, which includes the matching relationship between the characteristic information corresponding to historical GIS signals and the fault types;

[0111] The fuzzy rule base stores the matching relationship between the characteristic information corresponding to historical GIS signals and the fault types. These historical data have been verified through a large number of practices and contain rich fault identification experiences. By obtaining the fuzzy rule base, the system can use these historical experiences to judge the current equipment status, improving the accuracy and reliability of fault identification.

[0112] The fault modes of GIS equipment are diverse and are affected by various factors such as the operating environment and equipment aging. The fuzzy rule base can establish corresponding rules according to different historical situations and can flexibly handle various complex fault scenarios. When encountering a new fault situation, as long as its characteristic information matches some rules in the rule base, the fault type can be accurately identified.

[0113] The fuzzy rule base is established based on a large amount of historical and real-time experimental data as well as expert experience. It contains the fuzzy relationships between different input feature combinations of GIS signals and fault types, providing an intelligent reasoning mechanism for fault identification. The system can automatically match and reason in the fuzzy rule base according to the input feature information to obtain the corresponding fault type.

[0114] S104. Fuzzify several pieces of feature information to obtain fuzzified feature information.

[0115] In actual operation, the feature information of GIS equipment often has a certain degree of uncertainty and fuzziness. For example, the change range of some features may not be clear, or the data is inaccurate due to measurement errors. Fuzzifying the feature information can transform this uncertain information into a fuzzy set, which is more in line with the actual situation, can better handle the uncertainty in the feature information, and improve the accuracy of fault identification.

[0116] The fuzzification process can make the system insensitive to small changes in the feature information and enhance the robustness of the system. Even if there are certain fluctuations or errors in the feature information, the fuzzified feature information can still remain relatively stable, thus ensuring the reliability of the fault identification result.

[0117] The fuzzified feature information conforms to the operation rules of fuzzy logic, facilitating subsequent fuzzy reasoning based on the fuzzy rule base. Through fuzzy reasoning, the mutual relationships between different feature information and their impacts on faults can be comprehensively considered to obtain a more accurate fault judgment result.

[0118] In some embodiments, fuzzifying several pieces of feature information to obtain fuzzified feature information specifically includes:

[0119] According to the actual change ranges of the discharge quantity Q, discharge phase discharge repetition rate f r , vibration energy E, frequency component F, vibration amplitude A, temperature change rate α, and decomposition product content change rate β, obtain corresponding several fuzzy sets;

[0120] Respectively determine the membership functions corresponding to the discharge quantity Q, discharge phase discharge repetition rate f r , vibration energy E, frequency component F, vibration amplitude A, temperature change rate α, and decomposition product content change rate β;

[0121] Based on the corresponding membership functions, respectively calculate the membership values of the discharge quantity Q, discharge phase discharge repetition rate f r , vibration energy E, frequency component F, vibration amplitude A, temperature change rate α, and decomposition product content change rate β belonging to each fuzzy set.

[0122] Determine the fuzzified feature information according to the membership degree values of each fuzzy set.

[0123] In some embodiments, for the discharge amount Q, according to its minimum value Q min and maximum value Q max , three fuzzy sets of "small", "medium", and "large" are defined, and a triangular membership function is used to describe the degree to which Q belongs to each fuzzy set. Fuzzify each feature value to obtain its membership degree value on the corresponding fuzzy set. For example, if the current value of the discharge amount Q is Q0, the membership degrees of Q0 belonging to the "small", "medium", and "large" fuzzy sets are calculated by the membership function as μ 小 (Q0), μ 中 (Q0), μ 大 (Q0). The dynamic fuzzy rule base contains a large number of rules such as "IF (the discharge amount is large) AND (the vibration energy is normal) AND (the temperature change rate is normal) AND (the change rate of the gas decomposition product content is normal) THEN (the fault type is partial discharge fault)". According to the fuzzy membership degree values of the input features, through fuzzy implication (such as Mamdani implication or Lukasiewicz implication) and fuzzy composition (such as max-min composition or max-product composition) operations, the output fuzzy set of each rule is obtained.

[0124] S105. Determine the total fuzzy output based on the fuzzified feature information according to the fuzzy rule base, and the total fuzzy output includes multi-dimensional feature information;

[0125] The total fuzzy output includes multi-dimensional feature information, which means that when performing fault identification, the system comprehensively considers the influence of multiple feature information on the fault. Different feature information may reflect the operating state of the device from different aspects. By analyzing them comprehensively, it is possible to more comprehensively and accurately judge the fault situation of the device and avoid the one-sidedness that may be brought by single feature judgment.

[0126] During the fuzzy reasoning process, the system can explore the potential relationships between different feature information. For example, some feature information may not have an obvious fault indication effect when analyzed alone, but when combined with other feature information, it may reveal the fault hidden danger of the device. By determining the total fuzzy output, the system can make full use of these potential relationships to improve the accuracy of fault identification.

[0127] In some embodiments, determining the total fuzzy output based on the fuzzified feature information according to the fuzzy rule base specifically includes:

[0128] Based on the membership degree values of each feature information on the corresponding fuzzy set, determine the credibility of each rule according to the fuzzy logic operation rules;

[0129] Calculate the output membership degree of each rule according to the credibility of each rule and the output fuzzy set of the rule.

[0130] Aggregate the output membership degrees of all rules to obtain the total fuzzy output.

[0131] In some embodiments, perform an aggregation operation (such as union operation) on the output fuzzy sets of all rules to obtain the total fuzzy output.

[0132] S106. Defuzzify the total fuzzy output to determine the GIS device fault type.

[0133] In some embodiments, use the centroid method to convert the total fuzzy output into a definite fault type diagnosis result, including: calculating the centroid coordinates of the fuzzy output, and determining the final diagnosis result according to the fault type area where the centroid coordinates are located. For example, if the centroid coordinates fall in the "partial discharge fault" area, the diagnosis result is partial discharge fault; if they fall in the "mechanical fault" area, the diagnosis is mechanical fault, etc.

[0134] Defuzzification is the process of converting the fuzzy output into a definite fault type. Through defuzzification, the system can obtain the specific fault type from the result of fuzzy reasoning, making the fault identification result more intuitive and definite, which is convenient for the staff to carry out fault handling and maintenance decision-making. This makes the entire fault identification process achieve automation and intelligence, improves the efficiency of fault identification, and reduces the subjectivity and error of manual judgment.

[0135] In some embodiments, defuzzify the total fuzzy output to determine the GIS device fault type, which specifically includes:

[0136] Defuzzify the total fuzzy output, calculate the centroid position under the membership degree function curve of the fuzzy output to obtain the fault type judgment score. The abscissa of the membership degree function curve of the fuzzy output represents the fault type, the ordinate represents the membership degree value, and the coordinate value represents the judgment score of the corresponding fault type.

[0137] Compare the judgment scores of different fault types, and take the fault type with the highest judgment score as the final fault type diagnosis result.

[0138] In some embodiments, after defuzzifying the total fuzzy output to determine the GIS device fault type, the method further includes:

[0139] Collect the time difference of the partial discharge signal arriving at different sensors.

[0140] Determine the propagation speed of the partial discharge signal in the GIS device.

[0141] Calculate the distance from the current fault point to each sensor.

[0142] Based on the calculated distances from the fault point to each sensor and the known positions of the sensors, the position of the fault point is determined using the principle of geometric positioning.

[0143] Preferably, in fault location, the times t1, etc. when the partial discharge signals reach each sensor (such as S1, S2, etc.) are recorded. The propagation speed v of the partial discharge signal in the GIS is known (which can be determined in advance through experiments, assuming v = 2×10 8 m / s). Given two partial discharge sensors S1 and S2, and the times when the partial discharge signals reach S1 and S2 are t1 and t2 respectively, the distance from the fault point to S1 is The distance from the fault point to S2 is

[0144] Furthermore, by using three or more partial discharge sensors, multiple sets of distance data can be obtained to form a system of equations. Optimization algorithms such as the least squares method are used to solve this system of equations to determine the position coordinates of the fault point. The calculated position of the fault point is marked on the schematic diagram of the GIS device to visually display the estimated position of the fault point.

[0145] Analyze the information of the vibration sensor and the temperature sensor to assist in positioning. Check the amplitude distribution of the vibration signals collected by the vibration sensor to determine the area where the vibration amplitude increases significantly, which may be the area near the occurrence of the fault. Analyze the temperature data of the temperature sensor to find the area where the temperature rises significantly, which can also be used as a reference for fault location. Integrate the calculation results of the time difference positioning method and the vibration and temperature assisted positioning information to further optimize the judgment of the fault point position and improve the accuracy of fault location. For example, if the position of the fault point calculated by the time difference positioning method is close to the area where the vibration amplitude increases and the area where the temperature rises, then the confidence in the fault location result is stronger; if there is a certain deviation, then further analyze the reasons, and it may be necessary to adjust the sensor layout or optimize the parameters of the positioning algorithm.

[0146] In some embodiments, in a two-dimensional plane, with each sensor as the center and the distance from the fault point to the sensor as the radius, draw circles. The intersection points of these circles are the possible positions of the fault point. In three-dimensional space, a similar method of spherical intersection needs to be used to determine the position of the fault point.

[0147] Preferably, compare the fault diagnosis result and the fault location result with the actually simulated fault situation. If the diagnosis result is consistent with the actual fault type, and the error between the fault location and the actual fault point is within an acceptable range (e.g., the error is less than 10 cm), the result is considered accurate and reliable. Record the detailed information of this fault diagnosis and location, including the collected data, feature extraction values, diagnostic reasoning process, location calculation process, etc., for subsequent analysis and experience summary. If the diagnosis result or the location result is inaccurate, analyze the possible reasons. For example, check whether the sensors are working properly, whether there are sensor failures or improper installations resulting in inaccurate data collection; review whether the fuzzy rule base is perfect and whether it needs to be supplemented or adjusted according to new fault situations; evaluate whether the feature extraction method is appropriate and whether the feature extraction algorithm needs to be improved to enhance the representativeness of features, etc. Whether the result of this time is accurate or not, after completing the repair (if there is a fault) or confirming the normal operation of the equipment, return to the multi-source information collection step and continue to monitor the GIS equipment in real time. Continuously repeat the above processes of data collection, preprocessing, feature extraction, information fusion and fault diagnosis, fault location, etc., to achieve cyclic monitoring of the GIS equipment, timely discover and handle possible faults, and ensure the safe and stable operation of the GIS equipment. During the cyclic monitoring process, statistically analyze the monitoring data and diagnosis results regularly (e.g., monthly or quarterly), evaluate the trend of the equipment operation status, and early warn of potential fault risks.

[0148] In some embodiments, an operation time period containing various fault situations is selected from an actually operating GIS substation for experimental monitoring:

[0149] Partial discharge sensors (sampling frequency f1 = 50 MHz), vibration sensors (sampling frequency f2 = 10 kHz), temperature sensors (sampling frequency f3 = 1 Hz), and gas sensors (sampling frequency f4 = 0.1 Hz) are installed on the GIS equipment according to standard specifications. Each sensor synchronously collects data, and the data collection time interval Δt = 1 s. During this period, approximately 10,000 groups of data samples are collected in total. Among them, there are approximately 6,000 groups of data samples in the normal operation state, approximately 2,000 groups of simulated partial discharge fault data samples, approximately 1,000 groups of mechanical fault data samples, and approximately 1,000 groups of insulation overheat fault data samples. These data samples cover information under different fault degrees and different operation conditions, providing a rich data basis for subsequent analysis.

[0150] The collected partial discharge signal PD(t) is processed using the wavelet threshold denoising method (Daubechies wavelet basis 'db4', decomposition level N = 5). Taking a set of typical partial discharge fault data as an example, the signal-to-noise ratio of the original signal is about 10 dB. After denoising, the signal-to-noise ratio is increased to about 25 dB, the signal waveform becomes clearer, and the noise interference is effectively suppressed. The discharge quantity Q, discharge phase and discharge repetition rate f r are extracted from the denoised signal. In the partial discharge fault samples, the average value of the discharge quantity Q is about 1000 pC, and the discharge phase is mainly concentrated in the 30° - 60° interval. The discharge repetition rate f r is about 50 times per second on average. These characteristic values are consistent with the known partial discharge fault characteristics, indicating the accuracy of feature extraction.

[0151] After the vibration signal V(t) is filtered by a Butterworth band-pass filter (passband frequency range: [100 Hz, 1 kHz]), in the mechanical fault samples, the low-frequency equipment foundation vibration interference and high-frequency sensor noise are successfully removed. For example, there are obvious noise peaks in the vibration signal spectrum below 50 Hz and above 2 kHz before filtering. These noise peaks disappear after filtering, and obvious frequency components related to mechanical faults appear in the 200 Hz - 500 Hz interval. The vibration energy E, frequency component F, and vibration amplitude A are extracted from the filtered signal. The vibration energy E in the mechanical fault samples is about 50% higher on average than in the normal operating state. The proportion of the frequency around 300 Hz in the frequency component F increases from about 10% in the normal state to about 30%. The vibration amplitude A increases by about 3 times on average. These characteristic changes can effectively reflect the mechanical fault situation.

[0152] In the insulation overheating fault samples, the temperature change rate α of the temperature signal T(t) increases significantly, reaching an average of about 5 °C / s, while the temperature change rate during normal operation is generally below 0.5 °C / s. For gas signals, taking the decomposition product SO2 of SF6 gas as an example, in the insulation overheating fault and partial discharge fault samples, the change rate β of its decomposition product content both increases. In the insulation overheating fault, it reaches an average of 0.05 ppm / s, and in the partial discharge fault, it is about 0.03 ppm / s on average, while it is almost 0 during normal operation, indicating that the temperature and gas feature extraction can effectively capture the fault information.

[0153] The extracted eigenvalue is input into the dynamic fuzzy inference system for fault diagnosis. Taking 1000 groups of partial discharge fault samples as an example, after fuzzy inference and defuzzification, the number of samples correctly diagnosed as partial discharge faults is 920 groups, and the diagnostic accuracy rate reaches 92%. For mechanical fault samples, the diagnostic accuracy rate is about 88%, and for insulation overheating fault samples, the diagnostic accuracy rate is about 90%. Generally speaking, among all 10,000 groups of data samples, the average diagnostic accuracy rate of fault diagnosis reaches about 89%, which indicates that the dynamic fuzzy inference system can effectively fuse multi-source information and accurately diagnose different types of GIS faults.

[0154] Among the fault samples, 50 groups were selected for fault location verification. Using the time difference location method of partial discharge signals combined with vibration and temperature sensors for auxiliary location, in these 50 groups of samples, the average error between the calculated fault point location and the actual fault location is about 8 cm. For example, in a partial discharge fault, the actual fault point is located at a certain connection part of the GIS bus. The deviation of the fault point calculated by the time difference location method from this actual position is within 5 cm. Combining with the large vibration amplitude detected by the vibration sensor in this area and the temperature rise monitored by the temperature sensor, the fault location is further determined, proving the high accuracy of the fault location method.

[0155] In some complex fault situations, such as when there are both partial discharge and mechanical faults and the fault locations are close, by comprehensively analyzing multi-source information and optimizing the location algorithm, the fault point location can still be determined relatively accurately. For example, in a group of simulated composite fault samples, although there is a certain deviation in the position calculated by the initial time difference location method, after combining vibration and temperature information, the final fault location error is controlled within 10 cm, indicating that the method also has good effectiveness in complex fault scenarios.

[0156] In summary, through multi-source information fusion, this method synthesizes various information such as partial discharge, vibration, temperature, and gas, overcomes the limitations of single detection means, can diagnose GIS fault types more comprehensively and accurately, improves the accuracy and reliability of diagnosis, and at the same time uses a dynamic fuzzy inference system for fault diagnosis, can better adapt to complex situations in GIS fault diagnosis, realize the dynamic optimization of the rule base, and finally, through the time difference location method combined with other sensor information for auxiliary location, can more accurately determine the fault location, reduce the maintenance cost and workload.

[0157] This application also proposes a system for GIS fault diagnosis and identification based on multi-source information fusion, including:

[0158] A sensor module; configured in the GIS device for collecting partial discharge, vibration, temperature, and gas signals;

[0159] Data preprocessing module; used to preprocess the signals collected by the sensor module, including denoising, filtering, and cleaning;

[0160] Feature extraction module; used to extract the discharge amount, discharge phase, discharge repetition rate, vibration energy, frequency components, vibration amplitude, temperature change rate, and gas decomposition product content change rate from the preprocessed signals;

[0161] Fuzzy inference module; including a fuzzy rule base and an inference engine, used to obtain the fuzzy rule base, the fuzzy rule base includes the matching relationship between the feature information corresponding to the historical GIS signals and the fault types, and fuzzify the extracted feature information, and determine the total fuzzy output based on the fuzzy rule base;

[0162] Defuzzification module; used to defuzzify the total fuzzy output output by the fuzzy inference module to determine the GIS device fault type.

[0163] This application also proposes a readable storage medium storing a computer program, which when executed by a processor causes the processor to perform the following steps:

[0164] Collect GIS signals through a plurality of sensors installed on the GIS device;

[0165] Extract a plurality of feature information based on the GIS signals;

[0166] Obtain the fuzzy rule base, which includes the matching relationship between the feature information corresponding to the historical GIS signals and the fault types;

[0167] Fuzzify the plurality of feature information to obtain fuzzified feature information;

[0168] Determine the total fuzzy output based on the fuzzy rule base for the fuzzified feature information, and the total fuzzy output includes multi-dimensional feature information;

[0169] Defuzzify the total fuzzy output to determine the GIS device fault type.

[0170] This application also proposes a computer device, including a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor as follows:

[0171] Collect GIS signals through a plurality of sensors installed on the GIS device;

[0172] Extract a plurality of feature information based on the GIS signals;

[0173] Obtain the fuzzy rule base, which includes the matching relationship between the feature information corresponding to the historical GIS signals and the fault types;

[0174] Fuzzify a number of characteristic information to obtain fuzzified characteristic information;

[0175] Based on the fuzzy rule base, determine the total fuzzy output for the fuzzified characteristic information, where the total fuzzy output includes multi-dimensional characteristic information;

[0176] Defuzzify the total fuzzy output to determine the GIS device fault type.

[0177] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0178] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0179] The above-described embodiments merely represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the patent of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. The above-disclosed is only the preferred embodiment of the present invention, and of course, it cannot be used to limit the scope of the rights of the present invention. Therefore, the equivalent changes made according to the claims of the present invention still fall within the scope covered by the present invention.

Claims

1. A method for GIS fault diagnosis and recognition based on multi-source information fusion, characterized in that, The method includes: Collecting GIS signals through a plurality of sensors installed on the GIS device; Extracting a plurality of characteristic information based on the GIS signals; Obtaining a fuzzy rule base, where the fuzzy rule base includes the matching relationship between the characteristic information corresponding to the historical GIS signals and the fault types; Fuzzifying the plurality of characteristic information to obtain fuzzified characteristic information; Determining the total fuzzy output based on the fuzzy rule base for the fuzzified characteristic information, where the total fuzzy output includes multi-dimensional characteristic information; Defuzzifying the total fuzzy output to determine the fault type of the GIS device.

2. The method for GIS fault diagnosis and recognition based on multi-source information fusion according to claim 1, wherein, The collecting GIS signals through a plurality of sensors installed on the GIS device specifically includes: The plurality of sensors include a partial discharge sensor, a vibration sensor, a temperature sensor, and a gas sensor; The partial discharge sensor, the vibration sensor, the temperature sensor, and the gas sensor collect GIS signals synchronously.

3. A method for GIS fault diagnosis and recognition based on multi-source information fusion according to claim 2, characterized in that The partial discharge sensor, the vibration sensor, the temperature sensor, and the gas sensor collecting GIS signals synchronously specifically includes: The partial discharge sensor collects the partial discharge signal PD(t) of the GIS device according to the first sampling frequency f1; The vibration sensor collects the vibration signal V(t) of the GIS device according to the second sampling frequency f2; The temperature sensor collects the temperature signal T(t) of the GIS device according to the third sampling frequency f3; The gas sensor collects the gas signal G(t) of the GIS device according to the fourth sampling frequency f4.

4. A method for GIS fault diagnosis and recognition based on multi-source information fusion according to claim 1, characterized in that, Before the extracting a plurality of characteristic information based on the GIS signals, the method further includes: Denosing the partial discharge signal PD(t) and performing N-layer decomposition on the denoised signal; Reconstruct the decomposed signal based on a preset threshold to obtain the optimal partial discharge signal PD d (t); Filter the vibration signal V(t) according to the passband frequency range [f l , f h to obtain the optimal vibration signal V f (t); Performing data cleaning on the temperature signal T(t) and the gas signal G(t) to obtain the corrected temperature signal T1(t) and the corrected gas signal G1(t).

5. The method for GIS fault diagnosis and recognition based on multi-source information fusion according to claim 4, characterized in that, The extracting a plurality of characteristic information based on the GIS signals specifically includes: Convert the optimal partial discharge signal PD d (t) to obtain the discharge quantity Q, discharge phase and discharge repetition rate f r ; Perform time-frequency analysis on the optimal vibration signal V f (t) to obtain vibration energy E, frequency component F, and vibration amplitude A; Calculating the temperature change rate α of the temperature signal according to the corrected temperature signal T1(t); Determining the change rate β of the decomposition product content according to the key components of the corrected gas signal G1(t).

6. The method for GIS fault diagnosis and recognition based on multi-source information fusion according to claim 5, wherein The fuzzifying the plurality of characteristic information to obtain fuzzified characteristic information specifically includes: According to the actual change ranges of the discharge quantity Q, discharge phase discharge repetition rate f r , vibration energy E, frequency component F, vibration amplitude A, temperature change rate α, and decomposition product content change rate β, a plurality of corresponding fuzzy sets are obtained; Determine the membership functions corresponding to the discharge amount Q, discharge phase discharge repetition rate f r , vibration energy E, frequency component F, vibration amplitude A, temperature change rate α, and decomposition product content change rate β respectively; Calculate the discharge quantity Q and the discharge phase respectively based on the corresponding membership functions discharge repetition rate f r , the membership values of the vibration energy E, the frequency component F, the vibration amplitude A, the temperature change rate α, and the change rate β of the decomposition product content belonging to each fuzzy set; Determining the fuzzified characteristic information according to the membership degree values of each fuzzy set.

7. A method for GIS fault diagnosis and identification based on multi-source information fusion according to claim 6, characterized in that, The determining the total fuzzy output based on the fuzzy rule base for the fuzzified characteristic information specifically includes: Based on the membership degree values of each characteristic information on the corresponding fuzzy set, determining the credibility of each rule according to the fuzzy logic operation rules; Calculating the output membership degree of each rule according to the credibility of each rule and the output fuzzy set of the rule; Aggregating the output membership degrees of all rules to obtain the total fuzzy output.

8. A method for GIS fault diagnosis and recognition based on multi-source information fusion according to claim 1, characterized in that, The defuzzifying the total fuzzy output to determine the fault type of the GIS device specifically includes: Defuzzify the total fuzzy output, calculate the centroid position under the membership function curve of the fuzzy output, and obtain the fault type judgment score. The abscissa of the membership function curve of the fuzzy output represents the fault type, the ordinate represents the membership value, and the coordinate value represents the judgment score of the corresponding fault type; Compare the judgment scores of different fault types, and take the fault type with the highest judgment score as the final fault type diagnosis result.

9. A method for GIS fault diagnosis and recognition based on multi-source information fusion according to claim 1, characterized in that After defuzzifying the total fuzzy output to determine the fault type of the GIS device, the method further includes: Collect the time difference of the partial discharge signal reaching different sensors; Determine the propagation speed of the partial discharge signal in the GIS device; Calculate the distance from the current fault point to each sensor; Based on the calculated distances from the fault point to each sensor and the known positions of the sensors, use the geometric positioning principle to determine the position of the fault point.

10. A GIS fault diagnosis and recognition system based on multi-source information fusion, characterized in that, It includes: Sensor module; Configured in the GIS device for collecting partial discharge, vibration, temperature, and gas signals; Data preprocessing module; Used to preprocess the signals collected by the sensor module, including denoising, filtering, and cleaning; Feature extraction module; Used to extract the discharge quantity, discharge phase, discharge repetition rate, vibration energy, frequency component, vibration amplitude, temperature change rate, and gas decomposition product content change rate from the preprocessed signals; Fuzzy inference module; including a fuzzy rule base and an inference engine, used to obtain the fuzzy rule base. The fuzzy rule base includes the matching relationship between the characteristic information corresponding to the historical GIS signals and the fault type, and fuzzify the extracted characteristic information, and determine the total fuzzy output based on the fuzzy rule base; Defuzzification module; Used to defuzzify the total fuzzy output output by the fuzzy inference module to determine the fault type of the GIS device.

11. A readable storage medium stores a computer program. When the computer program is executed by a processor, the processor is caused to execute the steps of the method according to any one of claims 1 to 9.

12. A computer device includes a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor is caused to execute the steps of the method according to any one of claims 1 to 9.

Images