Distribution cable operation state prediction method based on multi-source data, medium and equipment

Abstract

The invention relates to the technical field of manufacturing of power distribution switch control equipment such as intelligent power distribution systems and facilities or power cables and cable accessories, in particular to a multi-source data-based power distribution cable running state prediction method applied to energy type power distribution system cables, a medium and equipment. The method comprises the following steps: S1, collecting feature information, defect types and defect criteria of a distribution cable, collecting real-time signals of multi-source parameters, and preprocessing to obtain a distribution cable running state multi-source database; s2, reading the multi-source data of the operation state of the distribution cable, constructing a data model of the distribution cable and a channel, and obtaining a prediction model of the operation state of the distribution cable; and S3, predicting to obtain a partial discharge spectrogram after the current moment, and judging a partial discharge state. According to the invention, the capability of discovering and disposing internal hidden dangers of the distribution cable is improved, and the capabilities of risk prediction, fault early warning and auxiliary decision making of the distribution cable in a sudden abnormal state are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent power distribution system, facilities and other power distribution switch control equipment manufacturing or power cable and cable accessories technology, and specifically to a distribution cable operating status prediction method, medium and equipment based on multi-source data applied to energy-type power distribution system cables. Background Art

[0002] Compared to overhead lines, distribution cables face a more uncertain operating environment due to various factors, including the external environment, operating conditions, and external forces. Cable health gradually deteriorates during operation, potentially leading to defects. Severe defects can lead to systemic grid failures, potentially threatening the safe and stable operation of the grid.

[0003] Defect identification for high-voltage cables primarily focuses on the pre-fault phase. Once a cable breakdown occurs, the energy of the fault increases exponentially, causing irreparable damage. Given the increasing importance of distribution cables in urban power distribution networks, research into cable defect prediction technology is essential to facilitate timely detection and repair of cable defects.

[0004] In the existing technology, there are mainly the following methods for detecting the operating status of distribution cables: partial discharge pattern recognition method, pulse current method, high-frequency current method, ultrasonic detection method, optical detection method, chemical detection method, infrared thermal imaging method and radio frequency detection method. The above detection methods have the following corresponding defects:

[0005] (1) It is too simple and cannot accurately identify cable defects, which has great limitations. In addition, it is also impossible to accurately identify the type of cable defects through a single judgment;

[0006] (2) The sensitivity, accuracy, resolution, and dynamic range of the measurement are limited by hardware parameters;

[0007] (3) It is only applicable to the case where the power equipment has external shielding grounding. Otherwise, it is difficult to measure the partial discharge signal. In addition, since the detection principle is electromagnetic coupling, the electromagnetic interference is relatively serious.

[0008] (4) It is only applicable to qualitatively determine the presence or absence of partial discharge signals, and to physically locate the partial discharge source by combining electrical pulse signals or directly using ultrasonic signals;

[0009] (5) The equipment is complex and expensive, has low sensitivity, and requires the substance to be detected to be transparent to light, making it difficult to apply in practice;

[0010] (6) In practical applications, a unified judgment standard is required. The detection accuracy is not high, and it is more sensitive to discovering early latent faults but cannot reflect sudden faults.

[0011] (7) There are still great difficulties in the field of quantitative research;

[0012] (8) It is impossible to distinguish three-phase power and the signal is easily affected by external interference. Summary of the Invention

[0013] The present invention provides a distribution cable operating status prediction method, medium and equipment based on multi-source data with higher prediction accuracy, wider application scope and simpler implementation, which can solve at least one of the above technical problems.

[0014] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0015] A method for predicting the operating status of a distribution cable based on multi-source data includes the following steps:

[0016] S1. Collect characteristic information, defect types and defect criteria of distribution cables, acquire real-time signals of multi-source parameters, pre-process and obtain a multi-source database of distribution cable operation status;

[0017] S2. Based on the multi-source database of distribution cable operation status, read the multi-source data of distribution cable operation status, form a data vector, build a data model of distribution cables and channels, and obtain a distribution cable operation status prediction model;

[0018] S3. Calculate the current discharge amount of the distribution cable and input it into the distribution cable operation status prediction model to predict the partial discharge spectrum after the current moment and determine the partial discharge status.

[0019] Furthermore, the S1 further includes:

[0020] S11. Collect and organize characteristic information such as the different laying methods, types, and structural characteristics of distribution cables, as well as the installation methods, binding relationships, and numbers of sensing devices. Also collect and organize defect types of distribution cables. Furthermore, collect and organize existing criteria and defect criteria such as typical artificial defects. Based on the characteristic information, defect types, and defect criteria, construct a basic database of distribution cable operating status.

[0021] S12. Using sensing devices installed at communication nodes, collect real-time data on distribution cable status and environment, including surface temperature, abnormal vibration, partial discharge, ambient temperature and humidity, and water immersion of connectors and terminal cables;

[0022] S13. Based on the basic database of distribution cable operation status, analyze the changing patterns of multi-source data such as thermal, electrical, vibration, and environmental status of distribution cables and channels under typical defects or fault conditions, screen out differences, and integrate data from different sources of distribution cables to form a sample data set of distribution cable operation status that is uniformly stored, updated, and managed, and build a multi-source database of distribution cable operation status.

[0023] Furthermore, the S2 further includes:

[0024] S21. The temperatures of the distribution cable joints and body are related to the ambient temperature, ambient humidity, core current, and sheath current, respectively. Therefore, the ambient temperature, ambient humidity, sheath / core current ratio, and historical temperatures of the distribution cable joints are used as input samples. Based on multi-source data such as load, temperature, humidity, vibration, and partial discharge, a data model for the distribution cable and channel is constructed.

[0025] S22. Based on the multi-source data information in the data model of distribution cables and channels, a particle swarm-least squares support vector machine algorithm is used to construct a distribution cable operation status prediction model.

[0026] Furthermore, the construction process of the distribution cable operation status prediction model includes:

[0027] S22.1. Establish the Lagrangian solution equation:

[0028]

[0029] In the formula, w is the weight vector, b is the bias, ξ is the relaxation factor, C is the penalty factor, and a i is a Lagrange multiplier, i=1,2,...k,φ(x i ) is a nonlinear mapping that maps the training sample data from the original space to a high-dimensional feature space, y i Output value for training sample data vector;

[0030] S22.2. Calculate the optimal parameters a and b:

[0031] The following expression is obtained by Lagrange multipliers and the Callus-Kuhn-Tucker condition:

[0032]

[0033] Eliminating w and ξ, the equation becomes:

[0034]

[0035] In the formula, y=[y1, y2,…y k ] T, a=[a1,a2,...a k ] T , Θ=[1,1,...1] T , Ω is a square matrix, the element in the i-th row and j-th column of Ω is

[0036] S22.3. Solving kernel function:

[0037] The inner product calculation of high-dimensional space is equivalent to a kernel function of the original input space. When dealing with nonlinear problems, no nonlinear transformation is performed and the kernel function is directly used instead of the inner product calculation. The expression is:

[0038]

[0039] S22.4. Establish a least squares support vector machine regression model:

[0040] After solving for parameters a and b, the expression of the least squares support vector machine regression model is obtained:

[0041]

[0042] S22.5. Optimize the parameters of the least squares support vector machine regression model using the particle swarm optimization algorithm:

[0043] The particle swarm optimization algorithm is initialized first to generate a group of random particles, which are random solutions, and then finds the optimal solution through iteration;

[0044] In each iteration, particles update themselves by tracking two extreme values. One extreme value is the optimal solution achieved by each particle in the previous searches, which is called the individual extreme value p. ibest , is the historical optimal position, and the other extreme value is the optimal solution achieved by all particles in the entire particle swarm in all generations of search, which is called the global extreme value g best , is the global optimal position;

[0045] Among them, the position of the i-th particle in the group in the n-dimensional space is represented by x i =(x i1 , x i2 ,...x in ), velocity is represented by v i＝ (v i1 , v i2 ,...v in ), the individual extreme value of the i-th particle is expressed as p ibest =(p i1 , p i2 ,...p in ), the global extreme value of the entire particle swarm is expressed as g best =(g1, g2, ... g n).

[0046] Furthermore, the S22.5 further includes:

[0047] S22.51. Setting the value range of the penalty factor C and the kernel function parameters to be optimized, and setting other initialization parameters including at least the local search capability and the maximum number of populations, and randomly initializing a group of particles;

[0048] S22.52, initialize the position and velocity information of each particle, and set the optimal solution achieved by each particle in the previous searches as p ibest , which is the historical optimal position, and the optimal solution reached by all particles in the entire particle swarm in the previous searches is set as g best , which is the global optimal position;

[0049] S22.53. Use the regression error of the sample data as the objective function to calculate the fitness of each particle;

[0050] S22.54, for each particle, its fitness value and the historical optimal position p ibest For comparison, if the fitness is better than the p ibest , then the fitness of the current position replaces the p ibest , becoming the new historical optimal position p ibest ;

[0051] S22.55, for each particle, its fitness value and the global optimal position g best For comparison, if the fitness is better than that g best , then the fitness of the current position replaces the g best , becomes the new global optimal position g best ;

[0052] S22.56. Update the particle's velocity and position using the following formula:

[0053] V i (k+1)=wv i (k)+c1rand1(p ibest -x i (k))+c2rand2(g best -x i (k))

[0054] x i (k+1)=x i (k)+v i (k+1)

[0055] Where c1 and c2 are learning factors, usually between (0, 2), rand1 and rand2 are random numbers between (0, 1), and w is the momentum coefficient;

[0056] S22.57. Determine whether the iteration termination condition is met. If so, stop the operation. If not, return to S22.53 and continue iterating until the condition is met.

[0057] The iteration termination condition is selected as the maximum number of iterations G k Or the optimal position searched by the particle swarm meets the predetermined minimum adaptation threshold.

[0058] Furthermore, the S3 further includes:

[0059] S31, measuring partial discharge signals in distribution cables;

[0060] S32. Calculate the current density in the current discharge air gap based on the measured partial discharge signal data:

[0061]

[0062] Where: q max is the maximum charge measured, r is the air gap radius, N e is the electron density, σ cavmax is the current density, which is also represented by J. is the correlation coefficient between electron energy distribution and mean free path, e is the charge of the electron, λ e is the mean free path of electrons, m e is the mass of the electron, c e is the thermal velocity of electrons;

[0063] S33. Determine the evolution of the severity of partial discharge in distribution cables:

[0064] Current density J is a physical quantity that describes the distribution of current per unit area, and its unit is ampere per square meter A / m 2 , that is, J = I / A, where J represents the current density, I represents the current intensity passing through the cross section, and A represents the area of ​​the cross section;

[0065] The current density J is integrated over the surface of the ground electrode to give the current I(t):

[0066] I(t)=∫Jda

[0067] The discharge amount q is obtained by integrating the current. In a discharge time interval t, the discharge amount q is obtained by integrating the current I(t) flowing through the ground electrode:

[0068]

[0069] The calculated discharge quantity q data is input into the distribution cable operation status prediction model to predict the partial discharge spectrum after the current moment, which can be used to determine the evolution process of the partial discharge severity.

[0070] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the processor executes the steps of the above-mentioned distribution cable operating status prediction method based on multi-source data.

[0071] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the above-mentioned distribution cable operating status prediction method based on multi-source data.

[0072] The beneficial effects of the present invention are embodied in:

[0073] 1. The method proposed in the present invention can better predict the temperature of distribution cable joints with high prediction accuracy, providing a reliable judgment basis for cable temperature monitoring and early warning systems.

[0074] 2. Compared with other existing methods, the method proposed in this invention has better relative error, and the obtained optimization parameters are completely effective.

[0075] 3. The present invention realizes the early warning of associated faults by identifying the discharge amount, trend and pattern of each sensor node.

[0076] 4. The present invention is based on the particle swarm-least squares support vector machine algorithm, which can quickly predict the temperature rise and hot spots of transformer windings.

[0077] 5. The method proposed in the present invention is directed to the pulse current method and utilizes the high-speed computing performance of computers to make detection more accurate and simpler.

[0078] 6. The present invention realizes intelligent analysis of monitoring data and intelligent diagnosis of cable status through intelligent diagnostic algorithms and multi-dimensional comparative analysis, so as to grasp the operating status in real time, discover defects in time, provide data support for operation and maintenance personnel to formulate maintenance strategies, and effectively avoid sudden power outages, fires, and personal and equipment injuries.

[0079] 7. The method proposed in the present invention can realize the automated operation and maintenance of distribution cables, make up for the current problem of insufficient number of manual inspection personnel, provide assistance for the rational planning of operation and maintenance personnel, and enable personnel to be applied to the lines that need maintenance. At the same time, it can avoid unexpected power outages caused by distribution line faults and improve user satisfaction of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] The drawings described herein are used to provide further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0081] Figure 1 It is a schematic diagram of the development process of high-voltage cable status.

[0082] Figure 2 It is a schematic diagram of the overall flow of the method according to an embodiment of the present invention.

[0083] Figure 3 1 is a schematic diagram of data vectors of multi-source data on the operating status of a distribution cable at different times according to an embodiment of the present invention.

[0084] Figure 4 It is a schematic diagram of a two-dimensional axisymmetric geometric model of internal discharge in a cable according to an embodiment of the present invention.

[0085] Figure 5 It is a schematic diagram of partial discharge spectrum prediction according to an embodiment of the present invention.

[0086] Figure 6 It is a structural block diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0087] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. In the absence of conflict, the embodiments in this application and the features in the embodiments can be combined with each other. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0088] It should be noted that the meaning of "and / or" appearing throughout the text includes three parallel solutions. Taking "A and / or B" as an example, it includes solution A, solution B, or solutions in which both A and B are satisfied. In addition, "multiple" refers to more than two. In addition, the technical solutions between the various embodiments can be combined with each other, but this must be based on the fact that ordinary technicians in this field can implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0089] The health status of distribution cables gradually degrades during operation. As the health status continues to degrade, defects also appear, as shown in Table 1 below.

[0090] Table 1 Causes and forms of distribution cable degradation

[0091]

[0092] The evolution of defects is as follows Figure 1 As shown, when defects become severe, they can evolve into systemic grid failures, posing a potential threat to the safe and stable operation of the grid. Defect identification for distribution cables primarily focuses on the pre-fault phase, as once a cable breakdown occurs, the fault energy increases exponentially, causing irreparable damage.

[0093] See also Figure 2 The embodiment of the present invention provides a method for predicting the operating status of a distribution cable based on multi-source data, comprising the following steps:

[0094] S1. Collect characteristic information, defect types and defect criteria of distribution cables, acquire real-time signals of multi-source parameters, pre-process and obtain a multi-source database of distribution cable operation status;

[0095] S2. Based on the multi-source database of distribution cable operation status, read the multi-source data of distribution cable operation status, form a data vector, build a data model of distribution cables and channels, and obtain a distribution cable operation status prediction model;

[0096] S3. Calculate the current discharge amount of the distribution cable and input it into the distribution cable operation status prediction model to predict the partial discharge spectrum after the current moment and determine the partial discharge status.

[0097] In this embodiment, the S1 further includes:

[0098] S11. Collect and organize characteristic information such as the different laying methods, types, and structural characteristics of distribution cables, as well as the installation methods, binding relationships, and numbers of sensing devices. Also collect and organize defect types of distribution cables. Furthermore, collect and organize existing criteria and defect criteria such as typical artificial defects. Based on the characteristic information, defect types, and defect criteria, construct a basic database of distribution cable operating status.

[0099] S12. Using sensing devices installed at communication nodes, collect real-time data on distribution cable status and environment, including surface temperature, abnormal vibration, partial discharge, ambient temperature and humidity, and water immersion of connectors and terminal cables;

[0100] S13. Based on the basic database of distribution cable operation status, analyze the changing patterns of multi-source data such as thermal, electrical, vibration, and environmental status of distribution cables and channels under typical defects or fault conditions, screen out differences, and integrate data from different sources of distribution cables to form a sample data set of distribution cable operation status that is uniformly stored, updated, and managed, and build a multi-source database of distribution cable operation status.

[0101] Here, for the preprocessing methods of multi-source data, technologies such as attribute information integration, data cleaning, normalization, and normalization processing can be used to screen out abnormal data.

[0102] The sensing device in this application can use a multi-state micro-power integrated sensing device to collect signals or data of multiple source parameters in real time. The specific device model or type is not specifically limited here.

[0103] See also Figure 3 In this embodiment, S2 further includes:

[0104] S21. The temperatures of the distribution cable joints and body are related to the ambient temperature, ambient humidity, core current, and sheath current, respectively. Therefore, the ambient temperature, ambient humidity, sheath / core current ratio, and historical temperatures of the distribution cable joints are used as input samples. Based on multi-source data such as load, temperature, humidity, vibration, and partial discharge, a data model for the distribution cable and channel is constructed.

[0105] The sheath current is generated in the sheath loop by the sheath induced voltage and has a certain proportional relationship with the core current.

[0106] S22. Based on the multi-source data information in the data model of distribution cables and channels, a particle swarm-least squares support vector machine algorithm is used to construct a distribution cable operation status prediction model;

[0107] Support vector machines (SVM) are a machine learning method based on statistical theory. Least squares support vector machines (LSSVM), as an improved algorithm of support vector machines, have unique advantages in dealing with small sample and nonlinear problems, and solve the problem of slow computing speed during large sample training. The performance of LSSVM depends on the model parameters, especially the kernel function, which has a significant impact on the model's prediction accuracy.

[0108] Particle swarm optimization (PSO) is an evolutionary computing technology that starts from a random solution, searches for the optimal solution through iteration, and evaluates the quality of the solution through fitness. However, its rules are simpler than those of genetic algorithms. It does not have the "crossover" and "mutation" operations of genetic algorithms. It searches for the global optimum by following the currently searched optimal value. The algorithm is easy to implement, has high accuracy, and converges quickly.

[0109] The particle swarm optimization support vector machines (PSO-SVM) model can improve the prediction accuracy and stability of the model. However, in practical applications, the PSO-SVM model has some limitations in parameter selection, and the training results are prone to gradually converge to the local optimal solution.

[0110] The present invention utilizes the particle swarm optimization algorithm to optimize and improve the parameters of the LSSVM model, which can effectively improve the prediction accuracy of the particle swarm optimization-least squares support vector machines (PSO-LSSVM) model.

[0111] In this embodiment, the process of constructing the distribution cable operating status prediction model includes:

[0112] S22.1. Establish the Lagrangian solution equation:

[0113]

[0114] Where w is the weight vector, b is the bias, ξ is the relaxation factor, C is the penalty factor (a constant that can find a compromise between training error and model complexity to make the function have better generalization ability), a i is a Lagrange multiplier, i=1,2,...k,φ(x i ) is a nonlinear mapping that maps the training sample data from the original space to a high-dimensional feature space, y i Output value for training sample data vector;

[0115] S22.2. Calculate the optimal parameters a and b:

[0116] The following expression is obtained by Lagrange multipliers and the Callus-Kuhn-Tucker condition:

[0117]

[0118] Eliminating w and ξ, the equation becomes:

[0119]

[0120] In the formula, y=[y1, y2,…y k ] T , a=[a1,a2,...a k ] T , Θ=[1,1,...1] T, Ω is a square matrix, the element in the i-th row and j-th column of Ω is

[0121] S22.3. Solving kernel function:

[0122] The inner product calculation of high-dimensional space is equivalent to a kernel function of the original input space. When dealing with nonlinear problems, no nonlinear transformation is performed and the kernel function is directly used instead of the inner product calculation. The expression is:

[0123]

[0124] The least squares support vector machine has good nonlinear approximation ability and can better fit the data;

[0125] S22.4. Establish a least squares support vector machine regression model:

[0126] After solving for parameters a and b, the expression of the least squares support vector machine regression model is obtained:

[0127]

[0128] S22.5. Optimize the parameters of the least squares support vector machine regression model using the particle swarm optimization algorithm:

[0129] The particle swarm optimization algorithm is initialized first to generate a group of random particles, which are random solutions, and then finds the optimal solution through iteration;

[0130] In each iteration, particles update themselves by tracking two extreme values. One extreme value is the optimal solution achieved by each particle in the previous searches, which is called the individual extreme value p. ibest , is the historical optimal position, and the other extreme value is the optimal solution achieved by all particles in the entire particle swarm in all generations of search, which is called the global extreme value g best , is the global optimal position;

[0131] Among them, the position of the i-th particle in the group in the n-dimensional space is represented by x i =(x i1 , x i2 ,...x in ), velocity is represented by v i＝ (v i1 , v i2 ,...v in ), the individual extreme value of the i-th particle is expressed as p ibest =(p i1 , p i2 ,...p in ), the global extreme value of the entire particle swarm is expressed as g best =(g1, g2, ... g n ).

[0132] In this embodiment, the S22.5 further includes:

[0133] S22.51. Setting the value range of the penalty factor C and the kernel function parameters to be optimized, and setting other initialization parameters including at least the local search capability and the maximum number of populations, and randomly initializing a group of particles;

[0134] S22.52, initialize the position and velocity information of each particle, and set the optimal solution achieved by each particle in the previous searches as p ibest , which is the historical optimal position, and the optimal solution reached by all particles in the entire particle swarm in the previous searches is set as g best , which is the global optimal position;

[0135] S22.53. Use the regression error of the sample data as the objective function to calculate the fitness of each particle;

[0136] S22.54, for each particle, its fitness value and the historical optimal position p ibest For comparison, if the fitness is better than the p ibest , then the fitness of the current position replaces the p ibest , becoming the new historical optimal position p ibest ;

[0137] S22.55, for each particle, its fitness value and the global optimal position g best For comparison, if the fitness is better than that g best , then the fitness of the current position replaces the g best , becomes the new global optimal position g best ;

[0138] S22.56. Update the particle's velocity and position using the following formula:

[0139] V i (k+1)=wv i (k)+c1rand1(p ibest -x i (k))+c2rand2(g best -x i (k))

[0140] x i (k+1)=x i (k)+v i (k+1)

[0141] Where c1 and c2 are learning factors, usually between (0, 2), rand1 and rand2 are random numbers between (0, 1), and w is the momentum coefficient;

[0142] S22.57. Determine whether the iteration termination condition is met. If so, stop the operation. If not, return to S22.53 and continue iterating until the condition is met.

[0143] The iteration termination condition is selected as the maximum number of iterations G k Or the optimal position searched by the particle swarm meets the predetermined minimum adaptation threshold.

[0144] See also Figure 4-Figure 5 In this embodiment, S3 further includes:

[0145] S31, measuring partial discharge signals in distribution cables;

[0146] S32. Calculate the current density in the current discharge air gap based on the measured partial discharge signal data:

[0147]

[0148] Where: q max is the maximum charge measured, r is the air gap radius, N e is the electron density, σ cavmax is the current density, which is also represented by J. is the correlation coefficient between electron energy distribution and mean free path, e is the charge of the electron, λ e is the mean free path of electrons, m e is the mass of the electron, c e is the thermal velocity of electrons;

[0149] S33. Determine the evolution of the severity of partial discharge in distribution cables:

[0150] Current density J is a physical quantity that describes the distribution of current per unit area, and its unit is ampere per square meter A / m 2 , that is, J = I / A, where J represents the current density, I represents the current intensity passing through the cross section, and A represents the area of ​​the cross section;

[0151] The current density J is integrated over the surface of the ground electrode to give the current I(t):

[0152] I(t)=∫Jda

[0153] The discharge amount q is obtained by integrating the current. In a discharge time interval t, the discharge amount q is obtained by integrating the current I(t) flowing through the ground electrode:

[0154]

[0155] During the discharge process, MATLAB (commercial mathematical software) and COMSOL (multi-physics simulation software) interact in real time, and MATLAB extracts parameters from COMSOL to calculate the charge;

[0156] The calculated discharge quantity q data is input into the distribution cable operation status prediction model to predict the partial discharge spectrum after the current moment, which can be used to determine the evolution process of the partial discharge severity.

[0157] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of the above-mentioned method for predicting the operating status of a distribution cable based on multi-source data.

[0158] See also Figure 6 An embodiment of the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the above-mentioned method for predicting the operating status of a distribution cable based on multi-source data.

[0159] An embodiment of the present invention further provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the steps of the above-mentioned method for predicting the operating status of a distribution cable based on multi-source data.

[0160] It is understandable that the system, device and storage medium provided in the embodiments of the present invention correspond to the method provided in the embodiments of the present invention. The explanation, examples and beneficial effects of the relevant contents can refer to the corresponding parts of the above-mentioned distribution cable operation status prediction method based on multi-source data.

[0161] It should be noted that those skilled in the art will understand that all or part of the steps implemented in the embodiments of the present invention can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using hardware, it can be implemented in whole or in part in the form of purchased standard parts or modified parts. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiments of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (eg, a solid state disk (SSD)).

[0162] In summary, current online monitoring technologies for power equipment, especially distribution cables, are primarily focused on fault diagnosis or status assessment. Fault diagnosis means that a diagnosis can only be performed when a cable fault occurs, while status assessment evaluates the current state of the cable, lacking an understanding of the cable's future state and state development trends. Therefore, the present invention predicts future operating states based on current information about the distribution cable to avoid faults. Specifically, a computer-aided testing system is combined with traditional testing methods. The measured partial discharge signal is amplified and filtered, then A / D converted. The analog quantity is converted into a digital quantity and then fed into a distribution cable operating state prediction model for data processing and analysis. Various spectra and statistics are generated, thereby predicting and analyzing partial discharge conditions.

[0163] It should be understood that the examples and implementation methods described herein are for illustrative purposes only and are not intended to limit the present invention. Those skilled in the art may make various modifications or changes based on them. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for predicting the operating status of distribution cables based on multi-source data, characterized in that: The following steps are involved: S1. Collect characteristic information, defect types and defect criteria of distribution cables, acquire real-time signals of multi-source parameters, pre-process and obtain a multi-source database of distribution cable operation status; S2. Based on the multi-source database of distribution cable operation status, read the multi-source data of distribution cable operation status, form a data vector, build a data model of distribution cables and channels, and obtain a distribution cable operation status prediction model; S3. Calculate the current discharge amount of the distribution cable and input it into the distribution cable operation status prediction model to predict the partial discharge spectrum after the current moment and determine the partial discharge status.

2. The method for predicting the operating status of a distribution cable based on multi-source data according to claim 1, characterized in that: Said S1 further comprises: S11. Collect and organize characteristic information such as the different laying methods, types, and structural characteristics of distribution cables, as well as the installation methods, binding relationships, and numbers of sensing devices. Also collect and organize defect types of distribution cables. Furthermore, collect and organize existing criteria and defect criteria such as typical artificial defects. Based on the characteristic information, defect types, and defect criteria, construct a basic database of distribution cable operating status. S12. Using sensing devices installed at communication nodes, collect real-time data on distribution cable status and environment, including surface temperature, abnormal vibration, partial discharge, ambient temperature and humidity, and water immersion of connectors and terminal cables; S13. Based on the basic database of distribution cable operation status, analyze the changing patterns of multi-source data such as thermal, electrical, vibration, and environmental status of distribution cables and channels under typical defects or fault conditions, screen out differences, and integrate data from different sources of distribution cables to form a sample data set of distribution cable operation status that is uniformly stored, updated, and managed, and build a multi-source database of distribution cable operation status.

3. The method for predicting the operating status of a distribution cable based on multi-source data according to claim 1, characterized in that: Said S2 further comprises: S21. The temperatures of the distribution cable joints and body are related to the ambient temperature, ambient humidity, core current, and sheath current, respectively. Therefore, the ambient temperature, ambient humidity, sheath / core current ratio, and historical temperatures of the distribution cable joints are used as input samples. Based on multi-source data such as load, temperature, humidity, vibration, and partial discharge, a data model for the distribution cable and channel is constructed. S22. Based on the multi-source data information in the data model of distribution cables and channels, a particle swarm-least squares support vector machine algorithm is used to construct a distribution cable operation status prediction model.

4. The method for predicting the operating status of a distribution cable based on multi-source data according to claim 3, characterized in that: The construction process of the distribution cable operation status prediction model includes: S22.

1. Establish the Lagrangian solution equation: In the formula, w is the weight vector, b is the bias, ξ is the relaxation factor, C is the penalty factor, and a i is a Lagrange multiplier, i=1,2,...k,φ(x i ) is a nonlinear mapping that maps the training sample data from the original space to a high-dimensional feature space, y i Output value for training sample data vector; S22.

2. Calculate the optimal parameters a and b: The following expression is obtained by Lagrange multipliers and the Callus-Kuhn-Tucker condition: Eliminating w and ξ, the equation becomes: In the formula, y=[y1, y2,…y k ] T , a=[a1,a2,...a k ] T , Θ=[1,1,...1] T , Ω is a square matrix, the element in the i-th row and j-th column of Ω is S22.

3. Solving kernel function: The inner product calculation of high-dimensional space is equivalent to a kernel function of the original input space. When dealing with nonlinear problems, no nonlinear transformation is performed and the kernel function is directly used instead of the inner product calculation. The expression is: S22.

4. Establish a least squares support vector machine regression model: After solving for parameters a and b, the expression of the least squares support vector machine regression model is obtained: S22.

5. Optimize the parameters of the least squares support vector machine regression model using the particle swarm optimization algorithm: The particle swarm optimization algorithm is initialized first to generate a group of random particles, which are random solutions, and then finds the optimal solution through iteration; In each iteration, particles update themselves by tracking two extreme values. One extreme value is the optimal solution achieved by each particle in the previous searches, which is called the individual extreme value p. ibest , is the historical optimal position, and the other extreme value is the optimal solution achieved by all particles in the entire particle swarm in all generations of search, which is called the global extreme value g best , is the global optimal position; Among them, the position of the i-th particle in the group in n-dimensional space is represented by x i =(x i1 , x i2 ,...x in ), velocity is represented by v i＝ (v i1 , v i2 ,...v in ), the individual extreme value of the i-th particle is expressed as p ibest =(p i1 , p i2 ,...p in ), the global extreme value of the entire particle swarm is expressed as g best =(g1, g2, ... g n ).

5. The method for predicting the operating status of a distribution cable based on multi-source data according to claim 4, characterized in that: Said S22.5 further includes: S22.

51. Setting the value range of the penalty factor C and the kernel function parameters to be optimized, and setting other initialization parameters including at least the local search capability and the maximum number of populations, and randomly initializing a group of particles; S22.52, initialize the position and velocity information of each particle, and set the optimal solution achieved by each particle in the previous searches as p ibest , which is the historical optimal position, and the optimal solution reached by all particles in the entire particle swarm in the previous searches is set as g best , which is the global optimal position; S22.

53. Use the regression error of the sample data as the objective function to calculate the fitness of each particle; S22.54, for each particle, its fitness value and the historical optimal position p ibest For comparison, if the fitness is better than the p ibest , then the fitness of the current position replaces the p ibest , becoming the new historical optimal position p ibest ; S22.55, for each particle, its fitness value and the global optimal position g best For comparison, if the fitness is better than that g best , then the fitness of the current position replaces the g best , becomes the new global optimal position g best ; S22.

56. Update the particle's velocity and position using the following formula: V i (k+1)=wv i (k)+c1rand1(p ibest -x i (k))+c2rand2(g best -x i (k)) x i (k+1)=x i (k)+v i (k+1) Where c1 and c2 are learning factors, usually between (0, 2), rand1 and rand2 are random numbers between (0, 1), and w is the momentum coefficient; S22.

57. Determine whether the iteration termination condition is met. If so, stop the operation. If not, return to S22.53 and continue iterating until the condition is met. The iteration termination condition is selected as the maximum number of iterations G k Or the optimal position searched by the particle swarm meets the predetermined minimum adaptation threshold.

6. The method for predicting the operating status of a distribution cable based on multi-source data according to claim 1, characterized in that: Said S3 further comprises: S31, measuring partial discharge signals in distribution cables; S32. Calculate the current density in the current discharge air gap based on the measured partial discharge signal data: Where: q max is the maximum charge measured, r is the air gap radius, N e is the electron density, σ cavmax is the current density, which is also represented by J. is the correlation coefficient between electron energy distribution and mean free path, e is the charge of the electron, λ e is the mean free path of electrons, m e is the mass of the electron, c e is the thermal velocity of electrons; S33. Determine the evolution of the severity of partial discharge in distribution cables: Current density J is a physical quantity that describes the distribution of current per unit area, and its unit is ampere per square meter A / m 2 , that is, J = I / A, where J represents the current density, I represents the current intensity passing through the cross section, and A represents the area of ​​the cross section; The current density J is integrated over the surface of the ground electrode to give the current I(t): I(t)=∫Jda The discharge amount q is obtained by integrating the current. In a discharge time interval t, the discharge amount q is obtained by integrating the current I(t) flowing through the ground electrode: The calculated discharge quantity q data is input into the distribution cable operation status prediction model to predict the partial discharge spectrum after the current moment, which can be used to determine the evolution process of the partial discharge severity.

7. A computer-readable storage medium, characterized in that A computer program is stored, and when the computer program is executed by a processor, the processor is caused to perform the steps of the method for predicting the operating status of a distribution cable based on multi-source data as described in any one of claims 1 to 6.

8. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for predicting the operating status of a distribution cable based on multi-source data as claimed in any one of claims 1 to 6.

Images