CN114994138B

CN114994138B - Defect detection method, device and equipment for cable buffer layer

Info

Publication number: CN114994138B
Application number: CN202210935884.4A
Authority: CN
Inventors: 房晟辰; 宋鹏先; 李隆基; 张弛; 贺春; 孙成; 李维博; 路菲
Original assignee: State Grid Corp of China SGCC; State Grid Tianjin Electric Power Co Ltd; Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd
Current assignee: State Grid Corp of China SGCC; State Grid Tianjin Electric Power Co Ltd; Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd
Priority date: 2022-08-05
Filing date: 2022-08-05
Publication date: 2022-11-08
Anticipated expiration: 2042-08-05
Also published as: CN114994138A

Abstract

The invention discloses a method, a device and equipment for detecting defects of a cable buffer layer, aiming at coiled cables, two Gaussian regression models which are obtained by training by adopting the same input data and different output data as sample data are used as volume resistivity calculation models, the data output by the two models can improve the accuracy of data detection, in the actual measurement, cable basic parameters, cable coil basic parameters and resistance measurement parameters are respectively input into the two models, after first volume resistivity and second volume resistivity which are respectively output by the two models are obtained through calculation, the total volume resistivity is calculated according to the first volume resistivity and the second volume resistivity, and finally the total volume resistivity is compared with preset evaluation parameters to obtain the detection result of the defects of the buffer layer of the cable to be detected. By adopting the embodiment of the invention, the volume resistivity of the cable buffer layer can be accurately calculated, and then whether the buffer layer of the coiled cable has ablation defects or not can be accurately judged according to the volume resistivity.

Description

Defect detection method, device and equipment for cable buffer layer

Technical Field

The invention relates to the technical field of cables, in particular to a method, a device and equipment for detecting defects of a cable buffer layer.

Background

With the high-speed development of economy in China, the power demand is continuously improved, the requirements on power transmission quality and reliability are higher and higher, and cable power transmission lines are basically adopted in urban power transmission networks and have the characteristics of stable power transmission performance, high safety and high adaptability. In the cable transmission line, the buffer layer of the cable not only plays a role of mechanical force buffering, but also more importantly realizes the electrical connection between the insulation shield and the grounding metal sheath. In recent years, the number of faults caused by ablation of the buffer layer of the high-voltage power cable is gradually increased, and the ablation hidden danger of the buffer layer becomes one of important hidden dangers threatening the safety of a power grid. The ablation hidden trouble of the buffer layer mainly comprises the situations of partial discharge ablation, current-induced thermal ablation, electrochemical ablation and the like, the hidden trouble situations are all caused by the fact that the volume resistivity of the buffer layer is seriously increased after the buffer layer is wetted, and the volume resistivity is increased to cause that the insulation shield and the metal sheath cannot form good electrical connection, so that the fault is caused. Therefore, whether the cable has ablation defects or not can be directly reflected by calculating the volume resistivity of the buffer layer.

Patent 202111148373.x provides that the volume resistivity of the internal buffer layer of the cable is calculated by measuring related data in a laid-flat laying state, but in actual work, once the cable is laid, if an unqualified item is detected, even if a supplier receives measures such as goods return, the economic cost required for re-reeling the laid cable and the influence on the engineering progress are not small, so that the practical work is still inconvenient for application. Therefore, before the cable is laid, the defect detection is carried out on the cable when the cable is positioned on the disc within the most reasonable time, however, when the cable is positioned on the disc, under the action of the gravity of the insulated wire core, the contact between the corrugated metal sleeve and the buffer layer is greatly different from that in the flat laying state, the contact between the corrugated metal sleeve and the buffer layer in the cable near the upper and lower parts of the axis of the cable disc is likely to be tight, the contact between the corrugated metal sleeve and the buffer layer is likely to be loose at the left and right sides of the axis, the volume resistivity error calculated by the method of patent 202111148373.X for the complete moving of the cable into the disc is very large, and whether the cable into the disc has ablation defect or not can not be accurately detected.

Disclosure of Invention

The embodiment of the invention aims to provide a method, a device and equipment for detecting defects of a cable buffer layer, which can accurately calculate the volume resistivity of the cable buffer layer and further can accurately judge whether the cable buffer layer of a coiled cable has ablation defects according to the volume resistivity.

In order to achieve the above object, an embodiment of the present invention provides a method for detecting defects of a cable buffer layer, including:

acquiring cable basic parameters, resistance measurement parameters and cable reel basic parameters of a cable to be measured; the cable to be measured is arranged on the cable drum, and the resistance measurement parameters comprise cable inner port resistance and cable outer port resistance between a corrugated metal sleeve and an insulation shielding layer of the cable to be measured, which are respectively positioned on the inner side and the outer side of the cable drum;

taking the cable basic parameters, the cable reel basic parameters and the resistance measurement parameters as input data, inputting the input data and a preset first super-parameter vector into a preset first volume resistivity calculation model, and inputting the input data and a preset second super-parameter vector into a preset second volume resistivity calculation model; the first volume resistivity calculation model and the second volume resistivity calculation model are Gaussian regression models obtained by training by using the same input data and different output data as sample data in a training process, and the output data are obtained by calculating in different calculation modes respectively;

acquiring a first volume resistivity output by the first volume resistivity calculation model and a second volume resistivity output by the second volume resistivity calculation model;

calculating a total volume resistivity from the first volume resistivity and the second volume resistivity;

and comparing the total volume resistivity with preset evaluation parameters to obtain a buffer layer defect detection result of the cable to be detected.

As an improvement of the above scheme, the acquiring of the resistance measurement parameter of the cable to be measured includes:

a hole through which the end of the cable to be tested is arranged on the inner side of the cable reel is positioned on the same horizontal plane with the axis of the cable reel;

after standing for a preset time period, stripping the waterproof plug, the short-section outer sheath, the corrugated metal sleeve and the buffer layer of the cable to be tested positioned on the inner side and the outer side of the cable reel;

respectively connecting a conductor mesh belt to the corrugated metal sleeve and the insulating shielding layer of the cable to be tested positioned on the inner side of the cable drum, respectively connecting two conductor mesh belts to two ends of a direct current resistance tester for resistance measurement to obtain a plurality of first measurement values, and taking the average value of the first measurement values as the port resistance of the inner side of the cable;

and respectively connecting the corrugated metal sleeve and the insulating shielding layer of the cable to be tested positioned outside the cable reel with conductor mesh belts, respectively connecting the two conductor mesh belts to two ends of a direct current resistance tester for resistance measurement so as to obtain a plurality of second measurement values, and taking the mean value of the second measurement values as the resistance of the outer port of the cable.

As an improvement of the above, the training method of the first volume resistivity calculation model and the second volume resistivity calculation model includes:

acquiring basic cable parameters, resistance measurement parameters and corresponding basic cable reel parameters of a plurality of groups of sample cables as input data, and taking the input data and output data which is obtained by pre-calculation and corresponds to the sample cables as training data;

constructing a target kernel function according to the cable basic parameters, the cable reel basic parameters and the resistance measurement parameters;

constructing a Gaussian regression model using the target kernel function, the input data and the output data;

and optimizing the Gaussian regression model to obtain a volume resistivity calculation model, and outputting the optimized hyper-parameters of the Gaussian regression model.

As an improvement of the foregoing solution, if the output data includes first output data calculated by using a first calculation strategy and second output data calculated by using a second calculation strategy, the volume resistivity calculation model includes a first volume resistivity calculation model and a second volume resistivity calculation model, and the optimized hyper-parameter includes a first hyper-parameter vector corresponding to the first volume resistivity calculation model and a second hyper-parameter vector corresponding to the second volume resistivity calculation model.

As an improvement of the above scheme, the basic parameters of the cable drum include the inner width of the cable drum and the diameter of the drum body of the cable drum, and the basic parameters of the cable include the length of the cable, the bending radius of the outermost layer of the cable on the drum, the average outer diameter of the cable, the average thickness of the outer sheath, the relevant parameters of the corrugated metal sleeve, the average wrapping thickness of the buffer layer, the relevant parameters of the insulation shielding layer, the relevant parameters of the conductor and the resistivity of the insulation shielding layer, so that the target kernel function is constructed according to the basic parameters of the cable, the basic parameters of the cable drum and the measured parameters of the resistance, and the method includes the following steps:

substituting the cable length, the bending radius of the outermost layer of the cable on the drum, the average value of the outer diameter of the cable and the basic parameters of the cable drum into a combined function of a Gaussian kernel function and a periodic kernel function to obtain a first kernel function;

substituting the cable length, the average thickness of the outer sheath, the related parameters of the corrugated metal sleeve, the average wrapping thickness of the buffer layer, the related parameters of the insulating shielding layer and the related parameters of the conductor into a combined function of Gaussian kernel functions to obtain a second kernel function;

substituting the resistance measurement parameter, the cable length, the resistivity of the insulation shielding layer, the related parameter of the insulation shielding layer and the related parameter of the conductor into a combined function of Gaussian kernel functions to obtain a third kernel function;

and integrating the first kernel function, the second kernel function and the third kernel function to obtain the target kernel function.

As an improvement of the above solution, the first kernel function satisfies the following formula:

；

wherein,k ₁ in order to be said first kernel function,k _s is a function of a gaussian kernel, and is,k _p in the form of a periodic kernel function,d _pd the diameter of the drum body of the cable drum,d _pw the width of the cable drum is the inner width,d _pc the outermost bend radius of the cable on the tray,d _d is the average value of the outer diameter of the cable,d _cable is the cable length.

As an improvement of the above solution, the parameters related to the corrugated metal sleeve include an average value of outside diameter of the corrugated metal sleeve, an average thickness of the corrugated metal sleeve, an average value of pitch of the corrugations and an average value of depth of the corrugations, the parameters related to the insulation shielding layer include an average thickness of the insulation shielding layer and an average thickness of the insulation layer, the parameters related to the conductor include an average thickness of the conductor shielding layer and an average value of outside diameter of the conductor core, and the second kernel function satisfies the following formula:

；

wherein,k ₂ for the purpose of the second kernel function,k _s in the form of a gaussian kernel function,d _t is the average thickness of the outer jacket,d _al is the average value of the outside diameter of the corrugated metal sleeve,d _len is the average value of the pitches of the wrinkles,d _dep is the average value of the depth of the wrinkles,t _al is the average thickness of the corrugated metal sleeve,t _hc for the average thickness of the said buffer layer wrapped around the package,t _op is the average thickness of the insulating-shielding layer,t _ins is the average thickness of the insulating layer,t _ip for the thickness of the conductor shield layer,d _cu is the average value of the outer diameter of the conductor wire core,d _cable is the cable length.

As an improvement of the above scheme, the insulation shielding layer related parameters include an average insulation shielding layer thickness and an average insulation layer thickness, and the conductor related parameters include a conductor shielding layer thickness and an average conductor core outer diameter, then the third kernel function satisfies the following formula:

；

wherein,k ₃ in order to be said third kernel function,k _s in the form of a gaussian kernel function,r _i is the port resistance at the inner side of the cable,r _o is the resistance of the outer port of the cable,u _op is the resistivity of the insulating shield layer and,t _op is the average thickness of the insulating-shielding layer,t _ins is the average thickness of the insulating layer,t _ip for the thickness of the conductor shield layer,d _cu is the average value of the outer diameter of the conductor wire core,d _cable is the cable length.

In order to achieve the above object, an embodiment of the present invention further provides a defect detection apparatus for a cable buffer layer, including:

the parameter acquisition module is used for acquiring the cable basic parameters, the resistance measurement parameters and the cable reel basic parameters of the cable to be measured; the cable to be measured is arranged on the cable drum, and the resistance measurement parameters comprise cable inner port resistance and cable outer port resistance between a corrugated metal sleeve and an insulation shielding layer of the cable to be measured, which are respectively positioned on the inner side and the outer side of the cable drum;

the data input module is used for taking the cable basic parameters, the cable reel basic parameters and the resistance measurement parameters as input data, inputting the input data and a preset first super-parameter vector into a preset first volume resistivity calculation model, and inputting the input data and a preset second super-parameter vector into a preset second volume resistivity calculation model; the first volume resistivity calculation model and the second volume resistivity calculation model are Gaussian regression models obtained by training by using the same input data and different output data as sample data in a training process, and the output data are obtained by calculating in different calculation modes respectively;

the volume resistivity calculation module is used for acquiring a first volume resistivity output by the first volume resistivity calculation model and a second volume resistivity output by the second volume resistivity calculation model, and calculating the total volume resistivity according to the first volume resistivity and the second volume resistivity;

and the defect detection module is used for comparing the total volume resistivity with preset evaluation parameters to obtain a buffer layer defect detection result of the cable to be detected.

In order to achieve the above object, an embodiment of the present invention further provides a defect detection apparatus for a cable buffer layer, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, where the processor implements the defect detection method for the cable buffer layer according to any one of the above embodiments when executing the computer program.

Compared with the prior art, the method, the device and the equipment for detecting the defects of the cable buffer layer disclosed by the invention have the advantages that aiming at coiled cables, two Gaussian regression models which are obtained by training by adopting the same input data and different output data as sample data are used as volume resistivity calculation models, the data output by the two models can improve the accuracy of data detection, in the actual measurement, the cable basic parameters, the cable reel basic parameters and the resistance measurement parameters are respectively input into the two models, after the first volume resistivity and the second volume resistivity which are respectively output by the two models are obtained through calculation, the total volume resistivity is calculated according to the first volume resistivity and the second volume resistivity, and finally the total volume resistivity is compared with the preset evaluation parameters to obtain the defect detection result of the buffer layer of the cable to be detected. By adopting the embodiment of the invention, the volume resistivity of the cable buffer layer can be accurately calculated, and then whether the buffer layer of the coiled cable has ablation defects or not can be accurately judged according to the volume resistivity.

Drawings

Fig. 1 is a flowchart of a method for detecting defects of a cable buffer layer according to an embodiment of the present invention;

FIG. 2 is a schematic structural diagram of a cable provided by an embodiment of the present invention;

fig. 3 is a block diagram of a defect detection apparatus for a cable buffer layer according to an embodiment of the present invention;

fig. 4 is a block diagram of a defect detecting apparatus for a cable buffer layer according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Referring to fig. 1, fig. 1 is a flowchart of a defect detection method for a cable buffer layer according to an embodiment of the present invention, where the defect detection method for the cable buffer layer includes:

s1, acquiring cable basic parameters, resistance measurement parameters and cable reel basic parameters of a cable to be measured;

s2, taking the cable basic parameters, the cable reel basic parameters and the resistance measurement parameters as input data, inputting the input data and a preset first hyper-parameter vector into a preset first volume resistivity calculation model, and inputting the input data and a preset second hyper-parameter vector into a preset second volume resistivity calculation model;

s3, acquiring a first volume resistivity output by the first volume resistivity calculation model and a second volume resistivity output by the second volume resistivity calculation model;

s4, calculating the total volume resistivity according to the first volume resistivity and the second volume resistivity;

and S5, comparing the total volume resistivity with preset evaluation parameters to obtain a buffer layer defect detection result of the cable to be detected.

Specifically, in step S1, referring to fig. 2, the cable according to the embodiment of the present invention includes a battery core (conductor) 10, a conductor shielding layer 20, an insulating layer 30, an insulating shielding layer 40, a buffer layer 50, and a corrugated sheath 60, where the cable to be tested according to the embodiment of the present invention is a cable placed on a cable drum after the cable leaves a factory, and the cable drum is a drum providing a function of winding wires and cables for industrial and mining enterprises.

Specifically, the basic parameter of the cable reel comprises the inner width of the cable reeld _pw And diameter of cable drumd _pd The cable basic parameter comprises the cable lengthd _cable Outermost layer bending radius of cable on discd _pc Average value of outer diameter of cabled _d Average thickness of outer sheathd _t Parameters related to the corrugated metal sleeve and average thickness of the buffer layer wrappingt _hc Insulation shield layer-related parameter, conductor-related parameter, and insulation shield layer resistivityu _op . The parameters related to the corrugated metal sleeve comprise the average value of the outside diameter of the corrugated metal sleeved _al Average thickness of corrugated metal sheatht _al Average value of wrinkle pitchd _len And average value of wrinkle depthd _dep The insulation shielding layer related parameters comprise the average thickness of the insulation shielding layert _op And average thickness of insulating layert _ins The conductor-related parameter includes a conductor shielding layer thicknesst _ip And the average value of the outer diameter of the conductor wire cored _cu 。

Illustratively, cable drum diameterd _pd Inner width of cable reeld _pw Length of cabled _cable May be obtained from the specifications provided by the cable supplier and sent with the cable drum. Outermost layer bend radius of cable on discd _pc Average value of outside diameter of corrugated metal sleeved _al Generally, the measurement result is obtained through field actual measurement. Average value of wrinkle pitchd _len Average value of wrinkle depthd _dep Resistivity of the insulating shield layeru _op Generally, the method can be obtained in a cable production process control file, and can also be obtained through field actual measurement. Mean value of cable outer diameterd _d Average thickness of outer sheathd _t Average thickness of corrugated metal sheatht _al Average thickness of buffer layer wrappingt _hc Average thickness of insulating shielding layert _op Average thickness of insulating layert _ins Thickness of the conductor shield layert _ip Average value of outer diameters of conductor coresd _cu Typically as provided in cable test reports.

Illustratively, the resistance measurement parameter needs to be obtained by actual measurement, and the resistance measurement parameter includes the cable inside port resistance between the corrugated metal sleeve and the insulation shielding layer of the cable to be measured, which are respectively located inside and outside the cable drumr _i And cable outer port resistorr _o . Further, the method for obtaining the resistance measurement parameters of the cable to be measured comprises the following steps of S11-S14:

s11, enabling a hole through which the end of the cable to be tested, which is positioned on the inner side of the cable reel, penetrates to be positioned on the same horizontal plane with the axis of the cable reel;

s12, after standing for a preset time period, stripping the waterproof plug, the short-section outer sheath, the corrugated metal sleeve and the buffer layer of the cable to be tested, which are positioned on the inner side and the outer side of the cable drum;

s13, respectively connecting a conductor mesh belt to the corrugated metal sleeve and the insulation shielding layer of the cable to be tested positioned on the inner side of the cable drum, respectively connecting the two conductor mesh belts to two ends of a direct current resistance tester to perform resistance measurement so as to obtain a plurality of first measurement values, and taking the average value of the first measurement values as the port resistance of the inner side of the cable;

s14, respectively connecting the corrugated metal sleeve and the insulating shielding layer of the cable to be tested positioned outside the cable reel with conductor mesh belts, respectively connecting the two conductor mesh belts to two ends of a direct current resistance tester for resistance measurement, so as to obtain a plurality of second measurement values, and taking the mean value of the second measurement values as the resistance of the outer port of the cable.

For example, when a cable is wound on the cable drum, the end of the cable is located at the innermost side of the cable drum, in order to unify the winding starting positions of the cables on different cable drums, a hole through which the end of the cable penetrates needs to be located on the same horizontal plane with the axis of the cable drum, and then the position of the cable drum is fixed. And keeping the cable reel position for more than 4 hours (preset time period), so that the buffer layer and the corrugated metal sleeve in the cable enter a new contact state under the action of gravity again after the position is adjusted, the waterproof plugs on the inner side and the outer side of the cable are stripped, and the short-section outer sheath, the corrugated metal sleeve and the buffer layer are stripped on the two sides respectively. Wrapping conductor mesh belts outside the corrugated metal sleeve and the insulating shielding layer on the inner side of the cable, keeping the outer end of the cable unchanged, connecting a direct current resistance tester between the two conductor mesh belts, measuring the direct current resistance for multiple times, and averaging to obtain the average valuer _i (ii) a Removing two conductor mesh belts on the inner side of the cable, wrapping the conductor mesh belts outside the corrugated metal sleeve and the insulating shielding layer of the cable, keeping the inner side end of the cable unchanged, connecting a direct current resistance tester between the two conductor mesh belts, measuring the direct current resistance for multiple times, and averaging to obtain the direct current resistance testerr _o And removing the conductor mesh belts at two positions outside the cable, additionally installing waterproof plugs at the inner side and the outer side of the cable, and finishing the measurement of the resistance of the ports at the inner side and the outer side of the cable.

Specifically, in step S2, two gaussian regression models, namely a first volume resistivity calculation model and a second volume resistivity calculation model, are constructed in advance, and the two models can respectively obtain output data according to corresponding input data, where the output data is the volume resistivity calculated by each model. The first volume resistivity calculation model and the second volume resistivity calculation model are Gaussian regression models obtained by training by using the same input data and different output data as sample data in a training process, and the output data are obtained by calculating in different calculation modes respectively.

Further, the training method of the first volume resistivity calculation model and the second volume resistivity calculation model includes steps S21 to S24:

s21, collecting basic cable parameters, resistance measurement parameters and corresponding basic cable reel parameters of a plurality of groups of sample cables as input data, and taking the input data and output data which is obtained by pre-calculation and corresponds to the sample cables as training data;

s22, constructing a target kernel function according to the cable basic parameters, the cable reel basic parameters and the resistance measurement parameters;

s23, constructing a Gaussian regression model by using the target kernel function, the input data and the output data;

and S24, optimizing the Gaussian regression model to obtain a volume resistivity calculation model, and outputting the optimized hyper-parameters of the Gaussian regression model.

It is worth to be noted that, in the embodiment of the present invention, a gaussian process regression model is used as a nonlinear regression model, and according to specific applications, a combined kernel function mode is adopted to improve the generalization capability of the model. As a statistical learning model, the Gaussian Process (GP) is determined by a mean function and a covariance function. For a real process

Its mean function and covariance function (also called asKernel functions, hereinafter not distinguished from each other) are defined as:

；

wherein

Representing two different sample inputs in the training sample data set and the test sample data set. The gaussian process can be expressed as:

；

in general, the mean function may be set to a zero function. Assuming that the output term contains an independent and identically distributed Gaussian noise

The mean is 0 and the variance is

I.e. labels

Then the covariance matrix for the labels of the training samples is:

；

wherein,

is a matrix of the unit, and is,

represents a covariance matrix ofiGo to the firstjThe column element isiAn input of training samplesx _i And a firstjAn input of training samplesx _j The value of the covariance function between, i.e.

. And

in a similar manner to that described above,

to middleiGo to the firstjThe column element isiAn input of training samplesx _i And a firstjA test sample inputx* _j The covariance function value of (a) to (b). By the same token can obtain

A covariance matrix. Given zero mean function, kernel functionkAnd on the premise of hyper-parameters, the prior distribution of the output values of the GP model to the test samples is a Gaussian distribution:

wherein,

in order to test the output value of the sample,Mfor the number of samples tested, 0 is the zero vector. The above formula can be abbreviated as:

；

similar to the above equation, the joint distribution of the output values of the label and the test sample of the training sample is a gaussian distribution:

wherein,Nis the number of training samples. Also, it can be abbreviated as:

；

then, a certain matrix operation is performed on the above formula to derive a conditional distribution:

focusing on the mean and covariance parts among them:

；

wherein

；

For a statistical learning regression model, given training set input and label data, the mean of the condition distribution given test set input

Is used as a point predictor for the GP model, and

the diagonal elements of (a) are the variances of the predicted distributions for all points in the test set。

The training process of the gaussian process regression model is equivalent to maximizing the log-edge likelihood:

；

wherein,

determining a kernel function for the hyperparametric vectorkForm of (2) and noise

In the context of machine learning, the hyper-parameter vector is a parameter set before the learning process is started, under normal conditions, the hyper-parameter needs to be optimized, the quality of the hyper-parameter determines the accuracy rate of a model, and a group of optimal hyper-parameter vectors are selected for machine learning so as to improve the learning performance and effect. The edge likelihood includes fitting terms to the training data

While also including a regularization term

. Therefore, in the GP model training, when the maximum edge likelihood is selected, a trade-off between the data fitting and the model complexity can be obtained. The training method is characterized in that a verification set is not required to be set, and corresponding verification error calculation is not required. The training process can use a conjugate gradient method or a BFGS quasi-Newton method and a modified form thereof to carry out optimization calculation.

Specifically, the output data includes first output data calculated by using a first calculation strategy and second output data calculated by using a second calculation strategy, and then the volume resistivity calculation model includes a first volume resistivity calculation model and a second volume resistivity calculation model, and the optimized hyper-parameter includes a first hyper-parameter vector corresponding to the first volume resistivity calculation model and a second hyper-parameter vector corresponding to the second volume resistivity calculation model. With reference to the above steps S21 to S24, the training processes of the first volume resistivity calculation model and the second volume resistivity calculation model are respectively illustrated.

Training the first volume resistivity calculation model:

1) Collecting model input characteristic information including cable basic parameters, resistance measurement parameters and corresponding cable drum basic parameters of a plurality of groups of sample cables, and recording the input characteristic information of each drum as a vector x _total . Then all training sample input feature vectors may be merged into the input feature matrix X _total Wherein each row corresponds to a disc cable input feature vector x _total 。

2) The first calculation strategy is as follows: for newly produced high-voltage power cables with different suppliers and different structure sizes, calculation parameters required by calculation of a patent 202111148373.X are obtained by collecting factory test reports, production process records and carrying out on-site actual measurement, the volume resistivity of buffer layers of different cables is calculated according to the method described in the patent 202111148373.X, and the volume resistivity of the buffer layer of each cable is recorded asy _total . Then all training sample outputs may be combined into an output vector y _total Wherein each row corresponds to a disc cable output datay _total 。

3) Constructing a target kernel function according to the cable basic parameters, the cable reel basic parameters and the resistance measurement parametersk(x _i , x _j )。

4) Utilizing the target kernel functionk(x _i , x _j ) The input data matrix X _total And the output data vector y _total Constructing a covariance matrix of labels of a Gaussian regression model training sample:

wherein,

is a matrix of the units,

represents a covariance matrix ofiGo to the firstjColumn elements being input data matrix numberiThe row vector and the secondjKernel function values between row vectors; hyper-parameters in kernel functions and

together forming a first hyperparametric vector

。

5) And optimizing the Gaussian regression model to obtain a first volume resistivity calculation model, and outputting a first hyperparametric vector optimized by the Gaussian regression model. Solving the following optimization problems by adopting methods such as a conjugate gradient method or a BFGS quasi-Newton method and an improved form thereof:

；

the optimization calculation is completed, and the first hyper-parameter vector of the model is determined

And finishing the storage and packaging of the first hyper-parameter vector after the training is finished.

Training process of the second volume resistivity calculation model:

1) Collecting model input characteristic information including cable basic parameters, resistance measurement parameters and corresponding cable drum basic parameters of a plurality of groups of sample cables, and recording the input characteristic information of each drum as a vector x _sample . Then all training sample input feature vectors may be merged into the input feature matrix X _sample Wherein each row corresponds to a disc cable input featureVector x _sample 。

2) And a second calculation strategy is as follows: for newly produced high-voltage power cables with different suppliers and different structure sizes, parameters required for calculation in a method for obtaining JB/T10259-2014 cable and optical cable water-blocking tapes or other standards by collecting factory test reports, production process records and carrying out on-site actual measurement are adopted, the volume resistivities of the buffer layers of different coils of cables are respectively measured and calculated according to the methods in the JB/T10259-2014 cable and optical cable water-blocking tapes or other standards, and the volume resistivities of the buffer layers of each coil of cables are recorded as the volume resistivities of the buffer layers of the coils of the cablesy _sample . Then all training sample outputs may be combined into an output vector y _sample Wherein each row corresponds to a disc cable output datay _sample 。

4) Utilizing the target kernel functionk(x _i , x _j ) The input data matrix X _sample And the output data vector y _sample Constructing a covariance matrix of labels of a Gaussian regression model training sample:

wherein,

is a matrix of the units,

together forming a second hyperparametric vector

。

5) And optimizing the Gaussian regression model to obtain a second volume resistivity calculation model, and outputting a second hyperparametric vector optimized by the Gaussian regression model. Solving the following optimization problems by adopting methods such as a conjugate gradient method or a BFGS quasi-Newton method and an improved form thereof:

；

the optimization calculation is completed, and a second hyperparametric vector of the model is determined

And finishing the storage and packaging of the second hyperparameter vector after the training is finished.

Further, according to a gaussian regression model, the first volume resistivity calculation model satisfies the following formula:

；

the second volume resistivity calculation model satisfies the following formula:

；

wherein X _* Is input data in the process of actually applying the model.

After the first volume resistivity calculation model and the second volume resistivity calculation model are obtained, in the practical application process, the input data (cable basic parameters, cable reel basic parameters and resistance measurement parameters) X of the cable to be measured are obtained _* Training sample input data X _total Training sample output data y _total Kernel functionk(Kernel function)kFor calculating all elements in each covariance matrix K) and the first hyperparametric vector

Inputting into the first volume resistivity calculation model, and inputting the input data (cable basic parameter, the cable drum basic parameter and the resistance measurement parameter) X _* Training sample input data X _sample Training sample output data y _sample Kernel functionkAnd the second hyperparametric vector

Inputting the data into the second volume resistivity calculation model, and calculating to obtain the first volume resistivity according to the first volume resistivity calculation model and the second volume resistivity calculation model respectively

And a second volume resistivity

. And the first hyper-parameter vector and the second hyper-parameter vector are used as the optimal hyper-parameters of the model obtained by pre-training optimization, and the optimal hyper-parameters are respectively substituted into the corresponding models so as to improve the calculation accuracy of the first volume resistivity calculation model and the second volume resistivity calculation model.

Further, in step S22, the constructing a target kernel function according to the cable basic parameters, the cable drum basic parameters and the resistance measurement parameters includes steps S211 to S214:

s211, substituting the length of the cable, the bending radius of the outermost layer of the cable on the drum, the average outer diameter of the cable and the basic parameters of the cable drum into a combined function of a Gaussian kernel function and a periodic kernel function to obtain a first kernel function;

s212, substituting the cable length, the average thickness of the outer sheath, the related parameters of the corrugated metal sleeve, the average wrapping thickness of the buffer layer, the related parameters of the insulating shielding layer and the related parameters of the conductor into a combined function of a Gaussian kernel function to obtain a second kernel function;

s213, substituting the resistance measurement parameter, the cable length, the resistivity of the insulating shielding layer, the relevant parameter of the insulating shielding layer and the relevant parameter of the conductor into a combined function of Gaussian kernel functions to obtain a third kernel function;

s214, integrating the first kernel function, the second kernel function and the third kernel function to obtain the target kernel function.

In an exemplary practical application, the mean function can be kept unchanged, and the performance of the GP model can be adjusted by setting different covariance functions. Any one of the semi-positive kernel functions can be used as the covariance function of the GP. The selection of different kernel functions will directly affect the type of function learned by the GP model and affect its generalization ability. The kernel functions are applied to different dimensions of the input vector respectively, and multiplication is carried out on the kernel functions, so that interaction among different dimensions can be introduced. The resulting product, combined with the kernel function, will represent a more complex coupling relationship between the different dimensions. The resistance of two ports of the coiled cable is a series-parallel resistance network between the corrugated metal sleeve and the insulating shielding layer, and the state quantity of the network is related to the following three aspects of information: 1. the distribution of cable winding positions on the cable drum; 2. the mutual position condition between the buffer layer and the corrugated metal sleeve inside the cable; 3. and measuring the resistance at two ends of the cable and the resistance of the insulating shielding layer. The three parts have a coupling relation to the volume resistivity common influence. Therefore, after the corresponding kernel functions are respectively established for the three parts, the kernel functions of the three parts are multiplied to obtain the final kernel function. To simplify the expression, the kernel function is described below

It is briefly described as

。

In particular, the distribution of the position of the cable windings is related on the one hand to the parameters of the cable drum and to the cable physical dimensions, and on the other hand to the periodicity of the cable length winding on the drum. Then, the first kernel function satisfies the following formula:

；

wherein,k ₁ in order to be said first kernel function,k _s is a function of a gaussian kernel, and is,k _p is a function of the periodic kernel and is,d _pd the diameter of the drum body of the cable drum,d _pw in order to make the inside of the cable drum wide,d _pc the outermost bend radius of the cable on the tray,d _d is the average value of the outer diameter of the cable,d _cable is the cable length.

In particular, the mutual position between the buffer layer and the corrugated metal sheath inside the cable is related to the dimensions of the buffer layer and the corrugated metal sheath on the one hand, and to the weight of the insulated wire core, which is related to the mass of the parts of the insulated wire core. Then, the second kernel function satisfies the following formula:

；

wherein,k ₂ for the purpose of the second kernel function,k _s in the form of a gaussian kernel function,d _t is the average thickness of the outer jacket,d _al is the average value of the outside diameter of the corrugated metal sleeve,d _len is the average value of the pitches of the wrinkles,d _dep is the average value of the depth of the wrinkles,t _al is the average thickness of the corrugated metal sleeve,t _hc for the average thickness of the said buffer layer wrapped around the package,t _op is the average thickness of the insulating-shielding layer,t _ins is the insulating layerThe average thickness of the film is measured,t _ip for the thickness of the conductor shield layer,d _cu is the average value of the outer diameter of the conductor wire core,d _cable is the cable length.

Specifically, the resistance measured at the two ends of the cable and the resistance of the insulating shielding layer directly form a kernel function. Then, the third kernel function satisfies the following formula:

；

wherein,k ₃ in order to be said third kernel function,k _s in the form of a gaussian kernel function,r _i is the port resistance at the inner side of the cable,r _o is the resistance of the outer port of the cable,u _op is the resistivity of the insulating-shielding layer,t _op is the average thickness of the dielectric shield layer,t _ins is the average thickness of the insulating layer,t _ip for the thickness of the conductor shield layer,d _cu is the average value of the outer diameter of the conductor wire core,d _cable is the cable length.

Exemplaryly,k _s the isotropic square exponential kernel, or gaussian kernel, radial basis function kernel,k _s has the following form:

；

wherein,

and

and respectively representing two hyperparameters of the signal variance and the scaling length of the square exponential kernel function.

k _p The following periodic kernel functions may be selected:

；

hyper-parameterpThe period length of the function to be approximated is expressed.

It should be noted that the form of the combined kernel function is addition or multiplication, and multiplication is selected in the embodiment of the present invention. Then, the target kernel function satisfies:

(ii) a Removing the repeated redundant items to obtain the form of the target kernel function as follows:

when the kernel function is adopted, the corresponding first hyperparametric vector

And a second hyperparametric vector

The forms are respectively:

wherein the first hyper-parameter vector

In

For being brought into the first volume resistivity calculation model to participate in the calculation, and dividing

The other hyper-parameters are used to carry in kernel functionskCalculating all elements in each covariance matrix K; second hyperparametric vector

In (1)

For taking into account in the second volume resistivity calculation model, and dividing

The other hyper-parameters are used to carry in kernel functionskAll elements in each covariance matrix K are calculated.

Specifically, in steps S3 to S4, after the first volume resistivity and the second volume resistivity are obtained, the total volume resistivity is calculated according to the first volume resistivity and the second volume resistivity, and the following formula is satisfied:

；

wherein,wis a weight value which is preset in the weight value,

(ii) a And calculating the total volume resistivity by setting a weight coefficient. Such as: when in usewWhen the value is 0, the total volume resistivity is equal to the second volume resistivity; when in usewWhen the value is 1, the total volume resistivity is equal to the first volume resistivity; when in usewWhen the value is 0.5, the total volume resistivity is equal to the sum of half the first volume resistivity and half the second volume resistivity.

Specifically, in step S5, the total volume resistivity is compared with a preset evaluation parameter to obtain a buffer layer defect detection result of the cable to be detected. Currently, JB/T10259-2014 requires the volume resistivity of the buffer layer to be less than or equal to 1000 Ω · m (evaluation parameter), so when the volume resistivity is less than or equal to 1000 Ω · m, the buffer layer is judged to be free of defects; and when the total volume resistivity is more than 1000 omega-m, judging that the buffer layer has defects.

Compared with the prior art, the defect detection method of the cable buffer layer, disclosed by the invention, aims at coiled cables, two Gaussian regression models which are obtained by training by adopting the same input data and different output data as sample data are used as volume resistivity calculation models, the data output by the two models can improve the accuracy of data detection, in the actual measurement, cable basic parameters, cable reel basic parameters and resistance measurement parameters are respectively input into the two models, after a first volume resistivity and a second volume resistivity which are respectively output by the two models are obtained through calculation, the total volume resistivity is calculated according to the first volume resistivity and the second volume resistivity, and finally the total volume resistivity is compared with a preset evaluation parameter to obtain the defect detection result of the buffer layer of the cable to be detected. By adopting the embodiment of the invention, the volume resistivity of the cable buffer layer can be accurately calculated, and then whether the buffer layer of the coiled cable has ablation defects or not can be accurately judged according to the volume resistivity.

Referring to fig. 3, fig. 3 is a block diagram of a defect detecting apparatus 100 for a cable buffer layer according to an embodiment of the present invention, where the defect detecting apparatus 100 for a cable buffer layer includes:

the parameter acquisition module 11 is used for acquiring cable basic parameters, resistance measurement parameters and cable reel basic parameters of a cable to be measured; the cable to be measured is arranged on the cable drum, and the resistance measurement parameters comprise cable inner port resistance and cable outer port resistance between a corrugated metal sleeve and an insulation shielding layer of the cable to be measured, which are respectively positioned on the inner side and the outer side of the cable drum;

the data input module 12 is configured to use the cable basic parameters, the cable drum basic parameters, and the resistance measurement parameters as input data, input the input data and a preset first hyper-parameter vector into a preset first volume resistivity calculation model, and input the input data and a preset second hyper-parameter vector into a preset second volume resistivity calculation model; the first volume resistivity calculation model and the second volume resistivity calculation model are Gaussian regression models obtained by training by using the same input data and different output data as sample data in a training process, and the output data are obtained by calculating in different calculation modes respectively;

a volume resistivity calculation module 13, configured to obtain a first volume resistivity output by the first volume resistivity calculation model and a second volume resistivity output by the second volume resistivity calculation model, and calculate a total volume resistivity according to the first volume resistivity and the second volume resistivity;

and the defect detection module 14 is configured to compare the total volume resistivity with a preset evaluation parameter to obtain a buffer layer defect detection result of the cable to be detected.

Specifically, the parameter obtaining module 11 is specifically configured to:

and respectively connecting the corrugated metal sleeve and the insulating shielding layer of the cable to be tested positioned outside the cable reel with conductor mesh belts, respectively connecting the two conductor mesh belts to two ends of a direct current resistance tester to carry out resistance measurement so as to obtain a plurality of second measurement values, and taking the average value of the second measurement values as the resistance of the outer port of the cable.

Specifically, the training method of the first volume resistivity calculation model and the second volume resistivity calculation model includes:

constructing a Gaussian regression model using the target kernel function, the input data, and the output data;

Specifically, the output data includes first output data calculated by using a first calculation strategy and second output data calculated by using a second calculation strategy, and then the volume resistivity calculation model includes a first volume resistivity calculation model and a second volume resistivity calculation model, and the optimized hyper-parameter includes a first hyper-parameter vector corresponding to the first volume resistivity calculation model and a second hyper-parameter vector corresponding to the second volume resistivity calculation model.

Specifically, the basic parameters of the cable drum include the inner width of the cable drum and the diameter of the drum body of the cable drum, and the basic parameters of the cable include the length of the cable, the bending radius of the outermost layer of the cable on the drum, the average outer diameter of the cable, the average outer sheath thickness, the relevant parameters of a corrugated metal sleeve, the average wrapping thickness of a buffer layer, the relevant parameters of an insulating shielding layer, the relevant parameters of a conductor and the resistivity of the insulating shielding layer, so that a target kernel function is constructed according to the basic parameters of the cable, the basic parameters of the cable drum and the measured resistance parameters, and the method includes:

substituting the resistance measurement parameter, the cable length, the resistivity of the insulating shielding layer, the related parameter of the insulating shielding layer and the related parameter of the conductor into a combined function of Gaussian kernel functions to obtain a third kernel function;

Specifically, the first kernel function satisfies the following formula:

；

wherein,k ₁ in order to be said first kernel function,k _s in the form of a gaussian kernel function,k _p in the form of a periodic kernel function,d _pd the diameter of the drum body of the cable drum,d _pw in order to make the inside of the cable drum wide,d _pc the outermost bend radius of the cable on the tray,d _d is the average value of the outer diameter of the cable,d _cable is the cable length.

Specifically, the parameters related to the corrugated metal sleeve include an average value of outside diameter of the corrugated metal sleeve, an average thickness of the corrugated metal sleeve, an average value of pitch of the corrugations and an average value of depth of the corrugations, the parameters related to the insulating shielding layer include an average thickness of the insulating shielding layer and an average thickness of the insulating layer, the parameters related to the conductor include an average thickness of the conductor shielding layer and an average value of outer diameter of the conductor wire core, and then the second kernel function satisfies the following formula:

；

wherein,k ₂ for the purpose of the second kernel function,k _s in the form of a gaussian kernel function,d _t is the average thickness of the outer jacket,d _al is the average value of the outside diameter of the corrugated metal sleeve,d _len is the average value of the pitches of the wrinkles,d _dep is the average value of the depth of the wrinkles,t _al is the average thickness of the corrugated metal sleeve,t _hc the average thickness of the buffer layer around the package is,t _op is the average thickness of the dielectric shield layer,t _ins is the average thickness of the insulating layer,t _ip for the thickness of the conductor shield layer,d _cu is the average value of the outer diameter of the conductor wire core,d _cable is the cable length.

Specifically, the insulation shielding layer related parameters include an average insulation shielding layer thickness and an average insulation layer thickness, and the conductor related parameters include a conductor shielding layer thickness and an average conductor core outer diameter, then the third kernel function satisfies the following formula:

；

wherein,k ₃ in order to be said third kernel function,k _s in the form of a gaussian kernel function,r _i is the port resistance at the inner side of the cable,r _o is the resistance of the outer port of the cable,u _op is the resistivity of the insulating shield layer and,t _op is the average thickness of the dielectric shield layer,t _ins is the average thickness of the insulating layer,t _ip for the thickness of the conductor shield layer,d _cu is the average value of the outer diameter of the conductor wire core,d _cable is the cable length.

Specifically, the total volume resistivity is calculated according to the first volume resistivity and the second volume resistivity, and the following formula is satisfied:

；

wherein,wis a preset weight value, and is used as a weight value,

(ii) a And calculating the total volume resistivity by setting a weight coefficient. Such as: when in usewWhen the value is 0, the total volume resistivity is equal to the second volume resistivity; when the temperature is higher than the set temperaturewWhen the value is 1, the total volume resistivity is equal to the first volume resistivity; when in usewWhen the value is 0.5, the total volume resistivity is equal to the sum of half the first volume resistivity and half the second volume resistivity.

Specifically, the total volume resistivity is compared with preset evaluation parameters to obtain a buffer layer defect detection result of the cable to be detected. Currently, JB/T10259-2014 requires the volume resistivity of the buffer layer to be less than or equal to 1000 Ω & m (evaluation parameter), so when the volume resistivity is less than or equal to 1000 Ω & m, the buffer layer is judged to be free of defects; and when the total volume resistivity is more than 1000 omega-m, judging that the buffer layer has defects.

It should be noted that, in the working process of each module in the defect detection apparatus for a cable buffer layer according to the embodiment of the present invention, reference may be made to the working process of the defect detection method for a cable buffer layer according to the above embodiment, which is not described herein again.

Compared with the prior art, the defect detection device 100 for the cable buffer layer, disclosed by the invention, aims at coiled cables, two Gaussian regression models which are obtained by training by adopting the same input data and different output data as sample data are used as volume resistivity calculation models, the data output by the two models can improve the accuracy of data detection, in the actual measurement, the cable basic parameters, the cable reel basic parameters and the resistance measurement parameters are respectively input into the two models, after the first volume resistivity and the second volume resistivity which are respectively output by the two models are obtained through calculation, the total volume resistivity is calculated according to the first volume resistivity and the second volume resistivity, and finally the total volume resistivity is compared with the preset evaluation parameters to obtain the defect detection result of the buffer layer of the cable to be detected. By adopting the embodiment of the invention, the volume resistivity of the cable buffer layer can be accurately calculated, and then whether the buffer layer of the coiled cable has ablation defects or not can be accurately judged according to the volume resistivity.

Referring to fig. 4, fig. 4 is a block diagram illustrating a defect detecting apparatus 200 for a cable buffer layer according to an embodiment of the present invention, where the defect detecting apparatus 200 for a cable buffer layer includes: a processor 21, a memory 22 and a computer program stored in the memory 22 and operable on the processor 21, wherein the processor 21 executes the computer program to implement the steps of the above-mentioned embodiments of the method for detecting defects of a cable buffer layer, such as the steps S1 to S5 shown in fig. 1.

Illustratively, the computer program may be divided into one or more modules/units, which are stored in the memory 22 and executed by the processor 21 to accomplish the present invention. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution process of the computer program in the defect detecting apparatus 200 of the cable buffer layer. For example, the computer program may be divided into a parameter acquisition module 11, a data input module 12, a volume resistivity calculation module 13 and a defect detection module 14.

The defect detecting device 200 of the cable buffer layer may be a desktop computer, a notebook, a palm computer, a cloud server, or other computing devices. The defect detecting apparatus 200 for a cable buffer layer may include, but is not limited to, a processor 21 and a memory 22. It will be understood by those skilled in the art that the schematic diagram is merely an example of the defect detecting apparatus 200 for a cable buffer layer, and does not constitute a limitation of the defect detecting apparatus 200 for a cable buffer layer, and may include more or less components than those shown, or combine some components, or different components, for example, the defect detecting apparatus 200 for a cable buffer layer may further include an input-output device, a network access device, a bus, etc.

The Processor 21 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. The general purpose processor may be a microprocessor or the processor 21 may be any conventional processor or the like, the processor 21 being the control center of the defect detecting apparatus 200 for the cable buffer layer, and various interfaces and lines connecting the various parts of the defect detecting apparatus 200 for the entire cable buffer layer.

The memory 22 may be used to store the computer program and/or module, and the processor 21 may implement various functions of the defect detecting apparatus 200 for a cable buffer layer by running or executing the computer program and/or module stored in the memory 22 and calling data stored in the memory 22. The memory 22 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data, a phonebook, etc.) created according to the use of the cellular phone, and the like. In addition, the memory 22 may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.

Wherein, the modules/units integrated with the defect detecting apparatus 200 for a cable buffer layer can be stored in a computer readable storage medium if they are implemented in the form of software functional units and sold or used as independent products. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer readable storage medium, and when the computer program is executed by the processor 21, the steps of the method embodiments described above may be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, read-Only Memory (ROM), random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like.

It should be noted that the above-described device embodiments are merely illustrative, where the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. In addition, in the drawings of the embodiment of the apparatus provided by the present invention, the connection relationship between the modules indicates that there is a communication connection therebetween, and may be specifically implemented as one or more communication buses or signal lines. One of ordinary skill in the art can understand and implement it without inventive effort.

While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention.

Claims

1. A method for detecting defects of a cable buffer layer is characterized by comprising the following steps:

2. The method for detecting defects of a cable buffer layer according to claim 1, wherein the obtaining of the measured resistance parameter of the cable to be measured includes:

3. The method for detecting defects in a cable buffer layer according to claim 1, wherein the method for training the first and second volume resistivity calculation models comprises:

4. The method of claim 3, wherein the output data includes first output data calculated using a first calculation strategy and second output data calculated using a second calculation strategy, and the volume resistivity calculation model includes a first volume resistivity calculation model and a second volume resistivity calculation model, and the optimized hyper-parameter includes a first hyper-parameter vector corresponding to the first volume resistivity calculation model and a second hyper-parameter vector corresponding to the second volume resistivity calculation model.

5. The method for detecting defects of the cable buffer layer according to claim 3, wherein the cable drum basic parameters include a cable drum inner width and a cable drum body diameter, the cable basic parameters include a cable length, a cable outermost layer bending radius on a drum, a cable outer diameter average value, an outer sheath average thickness, a wrinkle metal sleeve related parameter, a buffer layer wrapping average thickness, an insulation shielding layer related parameter, a conductor related parameter and an insulation shielding layer resistivity, and the constructing the target kernel function according to the cable basic parameters, the cable drum basic parameters and the resistance measurement parameters includes:

substituting the cable length, the average thickness of the outer sheath, the related parameters of the corrugated metal sleeve, the average wrapping thickness of the buffer layer, the related parameters of the insulating shielding layer and the related parameters of the conductor into a combined function of a Gaussian kernel function to obtain a second kernel function;

6. The method of defect detection of a cable buffer layer according to claim 5, wherein the first kernel function satisfies the following formula:

；

7. The method of claim 5, wherein the parameters associated with the corrugated metal jacket include an average value of an outside diameter of the corrugated metal jacket, an average thickness of the corrugated metal jacket, an average value of a corrugation pitch, and an average value of a corrugation depth, the parameters associated with the insulating shield include an average thickness of the insulating shield and an average thickness of the insulating layer, the parameters associated with the conductor include an average thickness of the conductor shield and an average value of an outside diameter of the conductor core, and the second kernel function satisfies the following equation:

；

wherein,k ₂ for the purpose of the second kernel function,k _s is a function of a gaussian kernel, and is,d _t is the average thickness of the outer jacket,d _al is the average value of the outside diameter of the corrugated metal sleeve,d _len is the average value of the pitches of the wrinkles,d _dep is the average value of the depth of the wrinkles,t _al is the average thickness of the corrugated metal sleeve,t _hc for the average thickness of the said buffer layer wrapped around the package,t _op is the average thickness of the dielectric shield layer,t _ins is the average thickness of the insulating layer,t _ip for the thickness of the conductor shield layer,d _cu is the average value of the outer diameter of the conductor wire core,d _cable is the cable length.

8. The method for detecting defects of a cable buffer layer according to claim 5, wherein the insulation shielding layer related parameters include an average thickness of the insulation shielding layer and an average thickness of the insulation layer, the conductor related parameters include an average thickness of the conductor shielding layer and an average value of an outer diameter of the conductor core, and the third kernel function satisfies the following formula:

；

wherein,k ₃ in order to be said third kernel function,k _s in the form of a gaussian kernel function,r _i is the port resistance at the inner side of the cable,r _o is the resistance of the port at the outer side of the cable,u _op is the resistivity of the insulating shield layer and,t _op is the average thickness of the dielectric shield layer,t _ins is the average thickness of the insulating layer,t _ip for the thickness of the conductor shield layer,d _cu is the average value of the outer diameters of the conductor wire cores,d _cable is the cable length.

9. A defect detection apparatus for a cable buffer layer, comprising:

the data input module is used for taking the cable basic parameters, the cable reel basic parameters and the resistance measurement parameters as input data, inputting the input data and a preset first super-parameter vector into a preset first volume resistivity calculation model, and inputting the input data and a preset second super-parameter vector into a preset second volume resistivity calculation model; the first volume resistivity calculation model and the second volume resistivity calculation model are Gaussian regression models obtained by training by adopting the same input data and different output data as sample data in the training process, and the output data are obtained by calculating in different calculation modes respectively;

10. A defect detection device of a cable buffer layer, comprising a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, the processor implementing the defect detection method of a cable buffer layer according to any one of claims 1 to 8 when executing the computer program.