Tunnel cable fault diagnosis method and device based on gas component analysis

Through the multi-trajectory chaotic fuzzy spatial clustering algorithm based on gas component analysis, the problems of low sensitivity and poor reliability of cable fault detection in the prior art are solved, and accurate detection of high-voltage cable insulation defects and overheating faults are achieved.

CN119619718APending Publication Date: 2025-03-14STATE GRID HUBEI ELECTRIC POWER CO LTD WUHAN POWER SUPPLY CO
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411743651.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-30
Publication Date
2025-03-14

AI Technical Summary

Technical Problem

The existing cable fault detection methods have problems such as low sensitivity, poor reliability and low detection accuracy, especially in the detection of high-voltage cable insulation defects and overheating faults.

Method used

A multi-trajectory chaotic fuzzy spatial clustering algorithm based on gas component analysis is used to collect and analyze gas data near the cable to form a cable gas component data set, and the clustering results are determined using the multi-trajectory chaotic fuzzy spatial clustering algorithm to obtain the fault diagnosis clustering reference model, and finally the fault type of the cable is judged based on the gas component data.

Benefits of technology

It realizes high sensitivity and reliability detection of cable insulation defects and overheating faults, can accurately judge the fault type of cable, and has the ability to detect multiple faults.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119619718A_ABST
    Figure CN119619718A_ABST
Patent Text Reader

Abstract

The invention discloses a tunnel cable fault diagnosis method and device based on gas component analysis. The method comprises the following steps: collecting gas data near a normal type tunnel cable and a known fault type tunnel cable; introducing gas near the tunnel cable into a gas detector for analysis and calculation to obtain normal and fault gas component data; forming a cable gas component data set G based on the gas component data; carrying out initialization processing on the data set G, and setting a parallel computing track number c, a chaotic sequence y, iteration precision a and a maximum iteration number M; determining a clustering result of the data set G based on a multi-track chaos fuzzy space clustering algorithm, and obtaining a cable fault diagnosis clustering reference model based on the gas component data; and acquiring gas component data G'of the to-be-diagnosed cable, calculating the membership degree of the G 'and typical data in G, and judging the fault type of the to-be-diagnosed cable according to the membership degree. According to the invention, the cable fault problem can be diagnosed accurately by using the analysis result of the gas components near the cable.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of cable fault diagnosis, and in particular to a method and device for diagnosing tunnel cable faults based on gas component analysis. Background Art

[0002] Cross-linked polyethylene (XLPE) power cable has become the mainstream equipment for power transmission due to its superior performance such as simple structure, large current carrying capacity, easy laying and small footprint, under the condition of increasingly improved urban power grids. In order to reduce accidents and ensure power supply reliability, cable insulation defects and insulation aging have gradually become the focus of power researchers.

[0003] At present, the detection methods for high-voltage cable insulation defects include temperature detection method, infrared detection method and partial discharge detection method. However, the installation cost and maintenance cost of the temperature detection method are relatively high; the infrared detection method has high work intensity and poor targeting; the partial discharge detection method is easily affected by surrounding noise and other high-frequency signals, and the detection accuracy is not high.

[0004] In addition, the main cause of high-voltage cable insulation aging is cable overheating failure. Severe insulation aging and overheating failures pose the risk of cable fire accidents.

[0005] When there are insulation defects in high-voltage cables, it is usually accompanied by partial discharge, which accelerates the destruction of cable insulation. During the partial discharge of the cable, the energy generated by the discharge will cause the air around the cable to react and produce special gases. Similarly, when the cable has an overheating fault, special gases will also be produced. Therefore, by detecting different gas components near the cable, it is possible to detect both cable insulation defects and cable overheating defects.

[0006] However, there are not enough existing studies on the gas composition around cables, and the types of gases produced near cables due to insulation defects and overheating failures are relatively complex, making it difficult to intuitively determine the specific fault type of the cable from the complex gas composition data.

[0007] The use of clustering algorithms can, to a certain extent, solve cable fault diagnosis problems based on gas component analysis. Compared with machine learning algorithms, clustering algorithms have faster solving capabilities and meet the requirements of rapid cable fault detection. However, general clustering algorithms have problems such as sensitivity to initial values ​​and insufficient convergence.

[0008] Therefore, how to use a reasonable method to implement a cable fault diagnosis method and device based on gas component analysis is an urgent problem that technicians in this field need to solve. Summary of the invention

[0009] In view of this, the present invention provides a tunnel cable fault diagnosis method based on gas component analysis to solve the problem of how to accurately diagnose cable faults using the analysis results of gas components near the cable.

[0010] The present invention provides a tunnel cable fault diagnosis method based on gas component analysis, comprising:

[0011] Collect a certain amount of gas data near the tunnel cable of normal type and known fault type. The fault types include overheating fault, bubble discharge inside the cable, insulation surface discharge and space corona discharge. Each type of data should be no less than 20 items.

[0012] The gas near the tunnel cable is passed into the gas detector to analyze and calculate the normal and fault gas component data. The gas categories include hydrogen H 2 、Methane CH 4 , ethane C 2 H 6 , Ethylene C 2 H 4 , acetylene C 2 H 2 , carbon monoxide CO, carbon dioxide CO 2 , sulfur dioxide SO 2 、Ammonia NH 3 and hydrogen cyanide HCN;

[0013] Based on the normal and fault gas composition data, a cable gas composition data set G = {g 1 ,g 2 ,…,g n}, initialize the data set G, set the number of parallel calculation trajectories c, chaotic sequence y, iteration accuracy a and maximum number of iterations M;

[0014] The clustering result of the data set G is determined based on the multi-trajectory chaotic fuzzy spatial clustering algorithm, and a cable fault diagnosis clustering reference model based on gas component data is obtained;

[0015] The gas component data G' of the cable to be diagnosed is obtained, and the membership between the gas component data G' of the cable to be diagnosed and the typical data in the data set G is calculated based on the cable fault diagnosis clustering reference model, and the fault type of the cable to be diagnosed is determined according to the calculated membership.

[0016] Furthermore, the method of collecting a certain amount of gas data near tunnel cables of normal types and known fault types specifically includes: using a gas collection device with exhaust and gas outlet capabilities, fixed within 0.1m near the middle joint of the cable, selecting cables that have been diagnosed to have four single faults of overheating, bubble discharge inside the cable, insulation surface discharge and space corona discharge for gas data collection, and collecting gas near the cable under normal operating conditions, the volume of gas collected each time is fixed, and the number of times each type of gas data is collected exceeds 20 times.

[0017] Furthermore, the gas detector includes: a gas sensing circuit, an amplifying circuit, a control unit, a data storage unit, and a power supply. The gas sensing circuit is connected to the amplifying circuit, and the amplifying circuit, the data storage unit and the power supply are respectively connected to the control unit. The gas sensing circuit is used to detect the gas component concentration of the characteristic gas in the air around the tunnel cable; the amplifying circuit is used to amplify the gas component concentration signal measured by the gas sensing circuit to improve the concentration resolution; the control unit is used to store the gas component concentration information obtained after detection and amplification as time series data and save it in the data storage unit; the power supply is used to supply power to the control unit.

[0018] Furthermore, the gas composition data set G = {g 1 ,g 2 ,…,g n} includes: gas components obtained by analyzing the gas near normal cables, gas components obtained by analyzing the gas near overheating fault cables, gas components obtained by analyzing the gas near internal bubble discharge fault cables, gas components obtained by analyzing the gas near insulation surface discharge fault cables, and gas components obtained by analyzing the gas near space corona discharge fault cables, where g 1 ,g 2 ,…,g n are gas component vectors, each of which contains hydrogen H 2 、Methane CH 4 , ethane C 2 H 6 , Ethylene C 2 H 4 , acetylene C 2 H 2 , carbon monoxide CO, carbon dioxide CO 2 , sulfur dioxide SO 2 、Ammonia NH 3 and hydrogen cyanide HCN, the ordered real number vector of the concentrations of the 10 gas components, such as g i =[g i1 ,g i2 ,g i3 ,gi4 ,g i5 ,g i6 ,g i7 ,g i8 ,g i9 ,g i,10 ].

[0019] Furthermore, the initialization processing of the data set G includes data normalization processing.

[0020] Furthermore, the process of establishing a cable fault diagnosis clustering reference model using a multi-trajectory chaotic fuzzy spatial clustering algorithm includes:

[0021] A. Select c cluster starting points as the initial cluster centers of l trajectories V(1) = {v 1 ,v 2 ,…,v c}, the cluster center at the kth iteration is V(k), and the membership degree D(k) of the cluster center V(k) at the kth iteration is calculated according to the following formula;

[0022]

[0023] Where D = (d im ) n×c , d im is the i-th gas component sample g i For cluster center v m membership, γ is the shape coefficient, which is used to adjust the sensitivity of the similarity function to the sample distance, V = (v mj ) c×10 , v mj is the cluster center value of the jth gas component of the mth fault;

[0024] B. Substitute the k-th iteration membership D(k) into the following formula to obtain the cluster center V'(k) = [v' 1 (k),v' 2 (k),…,v' c (k)];

[0025]

[0026] C. Calculate the objective function value of the kth iteration according to the following formula F(k) = [F 1 (k),F 2 (k),…,F c (k)];

[0027]

[0028] Among them, s im is the i-th gas component gi With the mth cluster center v m Grayscale distance, s im =||g i -v m ||, α is the weight parameter of the control domain item, N n For sample g i Neighborhood sample set, N is N n The number of samples in the middle neighborhood;

[0029] D. Find the historical minimum value F of each trajectory min =[F 1-min ,F 1-min ,…,F c-min ], where the historical minimum value F of the i-th trajectory i-min =min(F i (t)), t=[1,2,…,k], and F i-min The corresponding cluster center v i-min ; Determine the historical minimum value F of all trajectories up to the current step glob =min(F i-min ), and F glob The corresponding cluster center v glob ;

[0030] E. Calculate the adjustment amount ΔV(k) of the cluster center according to the following formula;

[0031]

[0032] F. Calculate the cluster center V(k+1)=V(k)+ΔV(k) of the k+1th iteration;

[0033] G. Update y chaotic sequences N(k)=[x 1 (k),x 2 (k),..,x y (k)];

[0034] x(k)=βx(k-1)(1-x(k-1))

[0035] Among them, β is the control parameter, 1<β≤4;

[0036] H. Map the above y chaotic sequences to the corresponding adjustment value interval to obtain the tentative clustering center, namely N(k)ΔV(k)=[x 1 (k)ΔV(k),x 2 (k)ΔV(k),..,x y (k)ΔV(k)];

[0037] I. Calculate the objective function value according to step C in If the xth j The objective function value of the tentative cluster centers Better than the existing objective function value F of the corresponding i-th trajectory in V(k+1) i (k+1), the cluster center of the trajectory is Δv determined by step E i Update to x j (k)Δv i , the cluster center of the k+1th iteration of the corresponding i-th trajectory is updated as follows:

[0038] v i (k+1)=v i (k)+x j (k)Δv i

[0039] J. Repeat steps AJ until the clustering process converges to the preset accuracy a or reaches the maximum number of iterations M, completing the multi-trajectory chaotic fuzzy space clustering process and obtaining the clustering result of the data set G.

[0040] Furthermore, the process of calculating the membership between G' and typical data in G includes: calculating the membership according to the formula in step A, and the corresponding type with the largest membership is the specific fault type of the cable G' to be diagnosed.

[0041] A tunnel cable fault diagnosis device based on gas component analysis includes a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the tunnel cable fault diagnosis method based on gas component analysis is implemented.

[0042] Compared with the prior art, the beneficial effects of the present invention include:

[0043] 1. The gas component analysis method provided has the advantages of high sensitivity and strong reliability. It can not only detect insulation defect faults of cables, such as bubble discharge inside the cable, discharge on the surface of the insulation layer and space corona discharge, but also detect overheating faults of cables and has the ability to detect various cable faults.

[0044] 2. The multi-trajectory chaotic fuzzy spatial clustering algorithm provided uses multi-trajectory parallel analysis to optimize the algorithm's global historical information extraction ability, solving the problem of general clustering algorithms being sensitive to initial values; it uses chaotic sequence analysis to optimize the algorithm's global optimization ability, reducing the probability of the algorithm falling into a local optimal solution; the chaotic fuzzy spatial clustering algorithm integrates spatial information and improves the problem of insufficient noise resistance of the fuzzy clustering algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A schematic diagram of a flow chart of a tunnel cable fault diagnosis method based on gas component analysis provided by the present invention;

[0046] Figure 2 A schematic diagram of gas collection near a cable provided by the present invention;

[0047] Figure 3 A schematic diagram of the composition of the gas detector provided by the present invention;

[0048] Figure 4 A schematic flow chart of step S4 is provided for the present invention;

[0049] Figure 5 This is a diagram of the tunnel cable environment of an embodiment provided by the present invention. DETAILED DESCRIPTION

[0050] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0051] Example 1

[0052] The embodiment of the present invention provides a tunnel cable fault diagnosis method based on gas component analysis, Figure 1 Come and see, Figure 1 The present invention provides a schematic flow chart of a tunnel cable fault diagnosis method based on gas component analysis. The tunnel cable fault diagnosis method based on gas component analysis comprises steps S1 to S5, wherein:

[0053] In step S1, a certain amount of gas data of normal type and known fault type near the tunnel cable is collected. The fault types include overheating fault, bubble discharge inside the cable, insulation layer surface discharge and space corona discharge. The data of each type is not less than 20 pieces.

[0054] In step S2, the gases are passed into a gas detector to analyze and calculate various gas components, including hydrogen H 2 、Methane CH 4 , ethane C 2 H 6 , Ethylene C 2 H 4 , acetylene C 2 H 2 , carbon monoxide CO, carbon dioxide CO 2 , sulfur dioxide SO 2 、Ammonia NH 3 and hydrogen cyanide HCN and other gases;

[0055] In step S3, a cable gas composition data set G is formed based on the normal and fault gas composition data. 1 ,g 2 ,…,g n Initialize the data set G, set the number of parallel calculation trajectories c, chaotic sequence y, iteration accuracy a and maximum number of iterations M;

[0056] In step S4, the clustering result of the data set G is determined based on the multi-trajectory chaotic fuzzy spatial clustering algorithm, and a cable fault diagnosis clustering reference model based on gas component data is obtained;

[0057] In step S5, the gas component data G' of the cable to be diagnosed is obtained, the membership degree between G' and the typical data in G is calculated, and the fault type of the cable to be diagnosed is determined according to the calculated membership degree.

[0058] Preferably, the process of acquiring the gas data of a certain amount of normal type and known fault type near the tunnel cable includes: using a gas collection device with the ability to extract and discharge gas, fixed within 0.1m near the middle joint of the cable, selecting the cable that has been diagnosed and confirmed to have four single faults of overheating fault, cable internal bubble discharge, insulation layer surface discharge and space corona discharge for gas data collection, and collecting the gas near the cable under normal operation. As a specific embodiment, the volume of gas collected each time in the embodiment of the present invention is fixed, and the number of times each type of gas data is collected is 30 times.

[0059] It should be noted that, combined with Figure 2 Come and see, Figure 2 The schematic diagram of gas collection near the cable provided by the present invention is a specific embodiment. The gas collection device used in the embodiment of the present invention is composed of a gas collection box, an air intake valve, an air extraction valve, an air guide tube and an air storage bag. Generally, the gas collection box is first clamped at the position of the middle joint of the cable to be tested, the entire gas collection device is connected, the air intake valve and the air extraction valve are started, and the air storage bag is connected to the small hole on the air guide tube. After continuous gas storage for 30 seconds, the air intake valve and the air extraction valve are closed, and the air storage bag is sealed to obtain a gas data. Repeat the above process to obtain 30 times of gas near the cable with four single faults including overheating fault, bubble discharge inside the cable, insulation layer surface discharge and space corona discharge, as well as gas near the normal cable, and a total of 150 gas data are obtained.

[0060] Preferably, combined Figure 3 Come and see, Figure 3 The gas detector provided by the present invention is a schematic diagram of the composition, including: a gas sensor circuit, a data storage circuit, an amplification circuit, a control unit and a power supply module. As a specific embodiment, the gas detector used in the embodiment of the present invention includes hydrogen H 2 、Methane CH4 , ethane C 2 H 6 , Ethylene C 2 H 4 , acetylene C 2 H 2 , carbon monoxide CO, carbon dioxide CO 2 , sulfur dioxide SO 2 、Ammonia NH 3 The sensing circuits for 10 gases such as hydrogen cyanide (HCN) and the corresponding amplifying circuits are finally uniformly input into the control unit for processing and the obtained concentration information of the 10 gas components is stored in the data storage unit. The power module is used to power the entire gas detector.

[0061] Preferably, the gas composition data set G = {g 1 ,g 2 ,…,g 150} includes: gas components obtained by analyzing the gas near normal cables, gas components obtained by analyzing the gas near overheating fault cables, gas components obtained by analyzing the gas near internal bubble discharge fault cables, gas components obtained by analyzing the gas near insulation surface discharge fault cables, and gas components obtained by analyzing the gas near space corona discharge fault cables, where g 1 ,g 2 ,…,g 150 are all gas component vectors, each of which is an ordered real number vector containing the concentrations of 10 gas components, such as g i =[g i1 ,g i2 ,g i3 ,g i4 ,g i5 ,g i6 ,g i7 ,g i8 ,g i9 ,g i,10 ].

[0062] Preferably, the initialization processing of the data set G includes data normalization processing.

[0063] Preferably, combined Figure 4 Come and see, Figure 4 The flowchart of step S4 provided by the present invention includes steps S41 to S410, wherein:

[0064] In step S41, c cluster starting points are selected as the initial cluster centers V of the trajectory, and the membership degree D corresponding to the cluster centers is calculated;

[0065] In step S42, a new cluster center V' is calculated according to the degree of membership;

[0066] In step S43, the objective function value F is calculated;

[0067] In step S44, the historical minimum value F of each trajectory is calculated min , and its corresponding cluster center V min , and the historical minimum value F of all trajectories glob and V glob ;

[0068] In step S45, the adjustment amount ΔV(k) of the cluster center is calculated;

[0069] In step S46, the cluster center ΔV(k+1) of the k+1th iteration is calculated;

[0070] In step S47, update y chaotic sequences N(k);

[0071] In step S48, y chaotic sequences are mapped to corresponding adjustment value intervals to obtain a tentative clustering center N(k)ΔV(k);

[0072] In step S49, the objective function value F corresponding to the chaotic sequence is calculated. N(k) And update the cluster center corresponding to the corresponding trajectory;

[0073] In step S410, the above steps are repeated until the clustering process converges or reaches the maximum number of iterations, and the clustering is completed to obtain the clustering result;

[0074] As a specific embodiment, the embodiment of the present invention determines the clustering result of the data set G by using a multi-trajectory chaotic fuzzy spatial clustering algorithm to obtain a cable fault diagnosis clustering reference model based on gas component data. The specific process is as follows:

[0075] In the first step, five clustering starting points are selected as the initial clustering centers of the five trajectories V(1) = {v 1 ,v 2 ,…,v 5}, calculate the membership degree D(k) of the cluster center V(k) of the kth iteration according to the following formula;

[0076]

[0077] The second step is to substitute the membership degree D(k) of the kth iteration into the following formula to obtain the cluster center V'(k) = [v' 1 (k),v' 2 (k),…,v' 5 (k)];

[0078]

[0079] The third step is to calculate the objective function value of the kth iteration according to the following formula: F(k) = [F 1 (k),F 2 (k),…,F 5 (k)];

[0080]

[0081] The fourth step is to find the historical minimum value F of each trajectory min =[F 1-min ,F 1-min ,…,F 5-min ], where the historical minimum value F of the i-th trajectory i-min =min(F i (t)), t=[1,2,…,k], and F i-min The corresponding cluster center v i-min ; Determine the historical minimum value F of all trajectories up to the current step glob =min(F i-min ), and F glob The corresponding cluster center v glob ;

[0082] Step 5: Calculate the adjustment amount ΔV(k) of the cluster center according to the following formula;

[0083]

[0084] Step 6: Calculate the cluster center V(k+1)=V(k)+ΔV(k) of the k+1th iteration;

[0085] Step 7: Update the five chaotic sequences N(k)=[x 1 (k),x 2 (k),..,x 5 (k)];

[0086] x(k)=4x(k-1)(1-x(k-1))

[0087] The eighth step is to map the above five chaotic sequences to the corresponding adjustment value intervals to obtain the tentative clustering center, namely N(k)ΔV(k)=[x 1 (k)ΔV(k),x 2 (k)ΔV(k),..,x 5 (k)ΔV(k)];

[0088] Step 9: Calculate the objective function value F N(k) =[F x1 ,F x2 ,…,F x5 ], where F xj=[F 1,xj ,F 2,xj ,…,F 5,xj ], if the xth j The objective function value F of the tentative cluster centers i,xj Better than the existing objective function value F of the corresponding i-th trajectory in V(k+1) i (k+1), the cluster center of the trajectory is Δv determined by step E i Update to x j (k)Δv i , the corresponding cluster center of the k+1th iteration of the ith trajectory is updated as follows:

[0089] v i (k+1)=v i (k)+x j (k)Δv i

[0090] Step 10: Repeat the above steps until the clustering process converges to the preset accuracy of 0.001, completing the multi-trajectory chaotic fuzzy space clustering process and obtaining the final clustering result of the data set G, as shown in Table 1:

[0091] Table 1

[0092]

[0093]

[0094] Preferably, combined Figure 5 Come and see, Figure 5 The tunnel cable environment diagram of the embodiment provided by the present invention, in a specific embodiment of the present invention, the embodiment of the present invention is Figure 5 Gas is collected from the tunnel cable shown in the figure, and the gas components are analyzed to obtain the gas data set G' near the on-site tunnel cable;

[0095] Further, the membership degree of G' and the typical data in G is calculated according to the above step S41, and the calculation result is shown in Table 2

[0096] Table 2

[0097]

[0098] Example 2

[0099] An embodiment of the present invention provides a tunnel cable fault diagnosis device based on gas component analysis, including a processor and a memory, wherein a computer program is stored in the memory. When the computer program is executed by the processor, the calculation method for tunnel cable fault diagnosis based on gas component analysis as described above is implemented.

[0100] The present invention discloses a tunnel cable fault diagnosis method and device based on gas component analysis. Firstly, a certain amount of gas data of normal type and known fault type near the tunnel cable is collected. The fault types include overheating fault, cable internal bubble discharge, insulation layer surface discharge and space corona discharge. The data of each type is not less than 20 pieces. Based on passing these gases into a gas detector to analyze and calculate multiple gas components, the specific gas categories include hydrogen H 2 、Methane CH 4 , ethane C 2 H 6 , Ethylene C 2 H 4 , acetylene C 2 H 2 , carbon monoxide CO, carbon dioxide CO 2 , sulfur dioxide SO 2 、Ammonia NH 3 and hydrogen cyanide HCN and other gases; based on the normal and fault gas component data, a cable gas component data set G = {g 1 ,g 2 ,…,g n Initialize the data set G, set the number of parallel calculation trajectories c, chaotic sequence y, iteration accuracy a and maximum number of iterations M; determine the clustering result of the data set G based on the multi-trajectory chaotic fuzzy space clustering algorithm, and obtain the cable fault diagnosis clustering reference model based on gas component data; obtain the gas component data G' of the cable to be diagnosed, calculate the membership degree between G' and the typical data in G, and judge the fault type of the cable to be diagnosed according to the calculated membership degree.

[0101] The technical solution of the present invention utilizes a multi-trajectory chaotic fuzzy space clustering algorithm to perform cluster analysis on the distribution of different gas components under various fault types, forming a cable fault diagnosis clustering reference model, fully considering the global optimization ability and generalization ability of the clustering algorithm, and realizing fault diagnosis of tunnel cables.

[0102] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A tunnel cable fault diagnosis method based on gas component analysis, characterized in that: include: Collect a certain amount of gas data near the tunnel cable of normal type and known fault type. The fault types include overheating fault, bubble discharge inside the cable, insulation surface discharge and space corona discharge. Each type of data should be no less than 20 items. The gas near the tunnel cable is passed into the gas detector to analyze and calculate the normal and fault gas component data. The gas categories include hydrogen H2, methane CH4, ethane C2H6, ethylene C2H4, acetylene C2H2, carbon monoxide CO, carbon dioxide CO2, sulfur dioxide SO2, ammonia NH3 and hydrogen cyanide HCN; Based on the normal and fault gas composition data, a cable gas composition dataset G = {g1, g2, …, g n }, initialize the data set G, set the number of parallel calculation trajectories c, chaotic sequence y, iteration accuracy a and maximum number of iterations M; The clustering result of the data set G is determined based on the multi-trajectory chaotic fuzzy spatial clustering algorithm, and a cable fault diagnosis clustering reference model based on gas component data is obtained; The gas component data G' of the cable to be diagnosed is obtained, and the membership between the gas component data G' of the cable to be diagnosed and the typical data in the data set G is calculated based on the cable fault diagnosis clustering reference model, and the fault type of the cable to be diagnosed is determined according to the calculated membership.

2. The method for diagnosing tunnel cable faults based on gas component analysis according to claim 1 is characterized in that: The method of collecting a certain amount of gas data near tunnel cables of normal types and known fault types specifically includes: using a gas collection device with exhaust and gas outlet capabilities, fixed within 0.1m near the middle joint of the cable, selecting cables that have been diagnosed and confirmed to have four single faults of overheating, bubble discharge inside the cable, insulation layer surface discharge and space corona discharge for gas data collection, and collecting gas near the cable under normal operating conditions, the volume of gas collected each time is fixed, and the number of times each type of gas data is collected exceeds 20 times.

3. The method for diagnosing tunnel cable faults based on gas component analysis according to claim 1 is characterized in that: The gas detector includes: a gas sensing circuit, an amplifying circuit, a control unit, a data storage unit, and a power supply. The gas sensing circuit is connected to the amplifying circuit, and the amplifying circuit, the data storage unit and the power supply are respectively connected to the control unit. The gas sensing circuit is used to detect the gas component concentration of the characteristic gas in the air around the tunnel cable; the amplifying circuit is used to amplify the gas component concentration signal measured by the gas sensing circuit to improve the concentration resolution; the control unit is used to store the gas component concentration information obtained after detection and amplification as time series data and save it in the data storage unit; the power supply is used to supply power to the control unit.

4. The method for diagnosing tunnel cable faults based on gas component analysis according to claim 1, characterized in that: The gas composition data set G = {g1, g2, ..., g n } includes: gas components obtained by analyzing the gas near normal cables, gas components obtained by analyzing the gas near overheating fault cables, gas components obtained by analyzing the gas near internal bubble discharge fault cables, gas components obtained by analyzing the gas near insulation surface discharge fault cables, and gas components obtained by analyzing the gas near space corona discharge fault cables, where g1, g2, …, g n They are all gas component vectors, and each vector is an ordered real number vector containing the concentrations of 10 gas components: hydrogen H2, methane CH4, ethane C2H6, ethylene C2H4, acetylene C2H2, carbon monoxide CO, carbon dioxide CO2, sulfur dioxide SO2, ammonia NH3 and hydrogen cyanide HCN.

5. The method for diagnosing tunnel cable faults based on gas component analysis according to claim 1, characterized in that: The initialization process for the data set G includes data normalization process.

6. The method for diagnosing tunnel cable faults based on gas component analysis according to claim 1, characterized in that: The process of establishing a cable fault diagnosis clustering reference model using a multi-trajectory chaotic fuzzy spatial clustering algorithm includes: A. Select c cluster starting points as the initial cluster centers of l trajectories V(1) = {v1, v2, …, v c }, the cluster center at the kth iteration is V(k), and the membership degree D(k) of the cluster center V(k) at the kth iteration is calculated according to the following formula; Where D = (d im ) n×c , d im is the i-th gas component sample g i For cluster center v m membership, γ is the shape coefficient, which is used to adjust the sensitivity of the similarity function to the sample distance, V = (v mj ) c×10 , v mj is the cluster center value of the jth gas component of the mth fault; B. Substitute the membership degree D(k) of the kth iteration into the following formula to obtain the cluster center V'(k) = [v'1(k), v'2(k), ..., v' c (k)]; C. Calculate the objective function value of the kth iteration according to the following formula: F(k) = [F1(k), F2(k), …, F c (k)]; Among them, s im is the i-th gas component g i With the mth cluster center v m Grayscale distance, s im =||g i -v m ||, α is the weight parameter of the control domain item, N n For sample g i Neighborhood sample set, N is N n The number of samples in the middle neighborhood; D. Find the historical minimum value F of each trajectory min =[F 1-min ,F 1-min ,…,F c-min ], where the historical minimum value F of the i-th trajectory i-min =min(F i (t)), t=[1,2,…,k], and F i-min The corresponding cluster center v i-min ; Determine the historical minimum value F of all trajectories up to the current step glob =min(F i-min ), and F glob The corresponding cluster center v glob ; E. Calculate the adjustment amount ΔV(k) of the cluster center according to the following formula; F. Calculate the cluster center V(k+1)=V(k)+ΔV(k) of the k+1th iteration; G. Update y chaotic sequences N(k)=[x1(k),x2(k),..,x y (k)]; x(k)=βx(k-1)(1-x(k-1)) Among them, β is the control parameter, 1<β≤4; H. Map the above y chaotic sequences to the corresponding adjustment value interval to obtain the tentative cluster center, that is, N(k)ΔV(k)=[x1(k)ΔV(k),x2(k)ΔV(k),..,x y (k)ΔV(k)]; I. Calculate the objective function value according to step C in If the xth j The objective function value of the tentative cluster centers Better than the existing objective function value F of the corresponding i-th trajectory in V(k+1) i (k+1), the cluster center of the trajectory is Δv determined by step E i Update to x j (k)Δv i , the cluster center of the k+1th iteration of the corresponding i-th trajectory is updated as follows: v i (k+1)=v i (k)+x j (k)Δv i J. Repeat steps AJ until the clustering process converges to the preset accuracy a or reaches the maximum number of iterations M, completing the multi-trajectory chaotic fuzzy space clustering process and obtaining the clustering result of the data set G.

7. The method for diagnosing tunnel cable faults based on gas component analysis according to claim 6 is characterized in that: The process of calculating the membership between G' and the typical data in G includes: calculating the membership according to the formula in step A, and the corresponding type with the largest membership is the specific fault type of the cable G' to be diagnosed.

8. A tunnel cable fault diagnosis device based on gas component analysis, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the method for diagnosing tunnel cable faults based on gas component analysis according to any one of claims 1 to 7 is implemented.