A method for online monitoring of cable thermal aging based on power line communication technology

Abstract

This invention discloses a method for online monitoring of cable thermal aging based on power line communication technology, implemented according to the following steps: Step 1: Establish a power line channel model to obtain the broadband channel frequency response under high-frequency conditions; Step 2: Based on the different distribution parameters of the cable under different thermal aging types and degrees, obtain the power line channel model after the cable undergoes thermal aging, and obtain the broadband channel frequency response under thermal aging conditions; Step 3: Use a random forest classification algorithm to classify cable aging and obtain the aging type of the cable; Step 4: Use a random forest regression prediction algorithm to train the cable groups obtained in Step 3 that have undergone overall thermal aging and local thermal aging, and predict the degree of thermal aging of the cable. This method can not only identify whether the cable has undergone thermal aging, but also predict the degree of cable aging.

Description

Technical Field

[0001] This invention belongs to the field of online monitoring technology for cable thermal aging, specifically a method for online monitoring of cable thermal aging based on power line communication technology. Background Technology

[0002] Power line communication (PLC) technology is a crucial technology for information transmission in smart grids. Currently, with the continuous growth of urban loads, the requirements for power supply reliability and urban aesthetics are increasing, leading to a corresponding increase in cable lines. Thermal aging is a significant cause of cable insulation degradation, making online monitoring of thermal aging essential for the safe and stable operation of the power grid. Employing PLC technology, which requires no additional investment, for online monitoring of cable thermal aging not only improves power supply reliability but also offers excellent economic benefits.

[0003] Currently, there are few mature technologies applicable to on-site cable aging diagnosis, mainly represented by sampling analysis and 0.1Hz ultra-low frequency dielectric loss detection. Sampling analysis can be used to identify water treeing and thermal aging in cables, with the latter being more effective at identifying water treeing. However, sampling analysis is destructive and only suitable for aging diagnosis of distribution cables with relatively low sampling costs and uniform aging. Although 0.1Hz ultra-low frequency dielectric loss detection has been included in cable field testing, whether it is completely equivalent to power frequency dielectric loss detection and reflects the insulation state of cables operating under power frequency conditions still requires further research, and the applied voltage may cause secondary damage to the cable.

[0004] In this context, there is an urgent need to find a new method for monitoring cable thermal aging, which would help to accelerate the solution of the large-scale cable aging problem in my country. Summary of the Invention

[0005] The purpose of this invention is to provide a method for online monitoring of cable thermal aging based on power line communication technology, which can not only identify whether the cable has undergone thermal aging, but also predict the degree of cable aging.

[0006] The technical solution adopted in this invention is a method for online monitoring of cable thermal aging based on power line communication technology, which is implemented according to the following steps:

[0007] Step 1: Obtain the characteristic impedance and propagation constant of the cable based on the distributed parameter model of the cable per unit length in high frequency; establish the power line channel model to obtain the broadband channel frequency response under high frequency conditions;

[0008] Step 2: Based on the different distributed parameters of the cable under different types and degrees of thermal aging, obtain the power line channel model of the cable after thermal aging, and obtain the broadband channel frequency response under thermal aging conditions.

[0009] Step 3: Using the random forest classification algorithm, the channel frequency response obtained in Step 2 is used as input and the cable aging type is used as output to classify the cable aging and obtain the cable aging type.

[0010] Step 4: Use the random forest regression prediction algorithm to train the cable groups that have undergone overall thermal aging and local thermal aging obtained in Step 3, and predict the degree of thermal aging of the cable.

[0011] The invention is further characterized in that,

[0012] The specific implementation method of step 1 is as follows:

[0013] Step 1.1: First, divide the complex network topology of cable distribution in reality into N independent basic units;

[0014] Then, calculate the distributed parameters of the cable: resistance R, inductance L, conductance G, and capacitance C;

[0015] The resistance R is shown in equation (1):

[0016]

[0017] In formula (1), r s As shown in equation (2):

[0018]

[0019] In equation (2), μ0 is the free permeability, and μ0 = 4π × 10 -7 V·s / (A·m), f is the frequency, σ c r is the conductivity of the conductor. c The radius of the cable core;

[0020] In equation (1), α R It is a correction factor, as shown in equation (3):

[0021]

[0022] In the formula, n e r represents the number of multi-strand wires in the outer ring of the cable. z The radius of the multi-strand wires forming the outer ring of the cable is given by δ, where:

[0023]

[0024]

[0025] In equation (4), n s The total number of strands that make up the conductor core;

[0026] The inductance L is shown in equation (6):

[0027]

[0028] In equation (6), L s For the self-inductance of a conductor, L m The mutual inductance of the conductors is shown in equation (7):

[0029]

[0030] In equation (7), d is the distance between the two conductors, which is 16 mm;

[0031] The capacitor C is shown in equation (8):

[0032] C=μ0ε0ε t L -1 (8)

[0033] In equation (8), ε0 is the dielectric constant in vacuum and ε0 = 8.8 × 10 -12 F / m, ε t The relative permittivity of the aged portion of the cable;

[0034] The conductance G is shown in equation (9):

[0035] G=2πfμ0ε0ε t L -1 (9)

[0036] In the distributed parameter model, the characteristic impedance of each line segment is:

[0037]

[0038] The propagation constant is:

[0039]

[0040] In the formula, R is the resistance per unit length in the distributed parameter model, G is the conductance per unit length in the distributed parameter model, L is the inductance per unit length in the distributed parameter model, and C is the capacitance per unit length in the distributed parameter model.

[0041] Step 1.2: Based on the voltage ratios at various frequency points between the receiver and transmitter of the nth basic unit, obtain the channel frequency response H of the nth basic unit. (n) (f), as shown in equation (12):

[0042]

[0043] In equation (12), V r (n)V is the voltage at the receiving end of the nth basic unit. t (n) V is the voltage at the emitter of the nth basic unit. p (n) The voltage at the branch point of the nth basic unit;

[0044] in, As shown in equation (13), As shown in equation (14), that is:

[0045]

[0046]

[0047] In equation (13), γ2 (n) Let l2 be the propagation constant of the receiving side line of the nth basic unit. (n) Let Γ be the line length on the receiving side of the nth basic unit. L (n) Let be the reflection coefficient of the receiver of the nth basic unit, expressed by equation (15), that is:

[0048]

[0049] In equation (15), Z L (n) Z is the equivalent load impedance of the nth basic unit. c2 (n) The characteristic impedance of the receiving line;

[0050] In equation (14), γ1 (n) Let l1 be the propagation constant of the transmitting side line of the nth basic unit. (n) Let Γ be the line length on the transmitting side of the nth basic unit. P (n) Let be the reflection coefficient of the nth basic unit branch point, expressed by equation (16), that is:

[0051]

[0052] In equation (16), Z p (n) Z represents the equivalent load impedance at the nth basic unit branch point. c1 (n) The characteristic impedance of the transmitting side line;

[0053] Step 1.3: Using the broadband channel frequency response of each basic unit in the complex network obtained in Step 1.2, the channel frequency response of the entire network topology is obtained, as shown in Equation (17):

[0054]

[0055] The specific implementation method for step 2 is as follows:

[0056] Step 2.1: Based on the characteristic that the complex dielectric constant of the cable insulation varies differently in the high-frequency range under different degrees of thermal aging, the complex dielectric constants of the insulation under the following conditions in the overall thermal aging of the cable (unaged, slightly aged, and severely aged) and the complex dielectric constants of the insulation under the following conditions in the local thermal aging (unaged, slightly aged, and severely aged) are fitted respectively in the frequency range of 1MHz-100MHz, as shown in Equation (18):

[0057]

[0058] In equation (18), ε' is the real part of the complex permittivity, ε” is the imaginary part of the complex permittivity, ε0 ​​is the permittivity in vacuum, w is the angular frequency, and M, N, and p are all fitting coefficients.

[0059] Step 2.2: Based on the relationship between the complex permittivity ε(w) and the capacitance C and conductance G in the distributed parameters, the distributed parameter model of the cable after thermal aging is obtained. Then, according to Equation (12), the power line channel model after the overall thermal aging of the cable when the entire cable is aged and the local thermal aging of the cable when the local cable is aged are obtained respectively.

[0060] Step 2.3, for overall thermal aging and local thermal aging, the channel frequency response H1 of the nth basic unit under different degrees of thermal aging. (n) (f) As shown in equation (19);

[0061]

[0062] In equation (19), V r1 (n) V represents the voltage at the receiving end of the nth basic unit after cable aging. t1 (n) V represents the voltage at the transmitter of the nth basic unit after cable aging. p1 (n) This represents the voltage at the nth basic unit branch point after the cable has aged.

[0063] Step 2.4: Set the total length of the backbone line of each basic unit and set it to either overall thermal aging or local thermal aging. The aging degree of each aging type is randomly set to no aging, slight thermal aging, and severe thermal aging, respectively. Set the line length of the branch section and the load at the end of the branch to obtain the channel frequency response under different aging degrees corresponding to different aging types.

[0064] The specific implementation method for step 3 is as follows:

[0065] Step 3.1: Use the Sparrow Algorithm to optimize the Random Forest classification algorithm, set the population size in the optimization algorithm, set the maximum number of iterations, and assign the optimal parameters obtained from the optimization algorithm to the network;

[0066] Step 3.2, take the aged channel frequency response H1 obtained in step 2. (n) (f) is used as input and the cable thermal aging type is used as output for classification and prediction.

[0067] The specific implementation method for step 4 is as follows:

[0068] Step 4.1: Use the random forest regression prediction algorithm, set the number of decision trees in the random forest, set the maximum number of features, do not limit the maximum depth of subtrees, set the minimum number of samples required for internal node subdivision, and set the minimum number of samples for leaf nodes.

[0069] Step 4.2: Using the channel frequency response of the cable groups classified as overall thermal aging and local thermal aging obtained in Step 3 as input, the machine is trained to obtain the degree of cable aging.

[0070] The beneficial effects of this invention are:

[0071] This invention is based on transmission line theory and utilizes existing carrier devices in the power system to monitor the channel frequency response in the cable in real time. No additional equipment is required, thus enabling online monitoring of whether the cable is experiencing thermal aging and the degree of thermal aging. The accuracy of predicting the type of cable thermal aging reaches 100%, and the prediction accuracy of the degree of cable thermal aging reaches 99%. This solves the problems of difficulty in measuring changes in related electrical parameters caused by cable aging and the high cost of measurement. Attached Figure Description

[0072] Figure 1 This is a typical topology network diagram in the method for online monitoring of cable thermal aging based on power line communication technology of the present invention;

[0073] Figure 2 This is the equivalent circuit diagram of the nth basic unit in the method for online monitoring of cable thermal aging based on power line communication technology of the present invention;

[0074] Figure 3 This is a simplified network topology diagram of the method for online monitoring of cable thermal aging based on power line communication technology in this invention;

[0075] Figure 4 This is a diagram illustrating the identification effect of cable thermal aging in the method for online monitoring of cable thermal aging based on power line communication technology of the present invention.

[0076] Figure 5This is a cable thermal aging identification confusion matrix diagram in the method for online monitoring of cable thermal aging based on power line communication technology of the present invention;

[0077] Figure 6 This is a diagram showing the predicted effect of the overall thermal aging degree of the cable in the method for online monitoring of cable thermal aging based on power line communication technology of the present invention.

[0078] Figure 7 This is a prediction error diagram of the overall thermal aging degree of the cable in the method for online monitoring of cable thermal aging based on power line communication technology of the present invention;

[0079] Figure 8 This is a diagram illustrating the predicted effect of the local thermal aging degree of a cable in the method for online monitoring of cable thermal aging based on power line communication technology of the present invention.

[0080] Figure 9 This is a prediction error diagram of the local thermal aging degree of the cable in the method for online monitoring of cable thermal aging based on power line communication technology of the present invention. Detailed Implementation

[0081] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0082] This invention relates to a method for online monitoring of cable thermal aging based on power line communication technology. The method utilizes power line communication technology and is implemented according to the following steps:

[0083] Step 1: Based on the distributed parameter model (resistance R, inductance L, conductance G, capacitance C) of the cable per unit length in high frequency (1MHz-100MHz), obtain the characteristic impedance and propagation constant of the cable. Use the voltage ratio of each frequency point between the receiver and transmitter to obtain the transmission characteristics of each basic unit, establish the power line channel model, and obtain the broadband channel frequency response in the high frequency case.

[0084] Step 1.1: First, divide the complex network topology of actual cable distribution into N independent basic units, such as... Figure 1 As shown. The structure of each basic unit includes a trunk and branches (branches may not exist). Branches may have only one level or multiple levels. The number of branches at each level can be single or multiple. For basic units with multiple levels of branches, the calculation is simplified by using the equivalent impedance of the branches in parallel with the trunk. For basic units without branches, the branch length is set to 0, and no simplified calculation is required.

[0085] Then, calculate the distributed parameters of the cable: resistance R, inductance L, conductance G, and capacitance C;

[0086] The resistance R is shown in equation (1):

[0087]

[0088] In formula (1), r s As shown in equation (2):

[0089]

[0090] In equation (2), μ0 is the free permeability, and μ0 = 4π × 10 -7 V·s / (A·m), f is the frequency, σ c Let σ be the conductivity of the conductor. c =5.7×10 7 S / m,r c The core radius is 4mm;

[0091] In equation (1), α R It is a correction factor, as shown in equation (3):

[0092]

[0093] In the formula, n e r represents the number of multi-strand wires in the outer ring of the cable. z The radius of the multi-strand wires forming the outer ring of the cable is given by δ, where:

[0094]

[0095]

[0096] In equation (4), n s The total number of strands that make up the conductor core;

[0097] The inductance L is shown in equation (6):

[0098]

[0099] In equation (6), L s For the self-inductance of a conductor, L m The mutual inductance of the conductors is shown in equation (7):

[0100]

[0101] In equation (7), d is the distance between the two conductors, which is 16 mm;

[0102] The capacitor C is shown in equation (8):

[0103] C=μ0ε0ε t L -1 (8)

[0104] In equation (8), ε0 is the dielectric constant in vacuum and ε0 = 8.8 × 10*12 F / m, ε t The relative permittivity of the aged portion of the cable;

[0105] The conductance G is shown in equation (9):

[0106] G=2πfμ0ε0ε t L -1 (9)

[0107] The equivalent circuit diagram of the nth basic unit is as follows: Figure 2 As shown, in the distributed parameter model, the characteristic impedance of each line segment is:

[0108]

[0109] The propagation constant is:

[0110]

[0111] In the formula, R is the resistance per unit length in the distributed parameter model, G is the conductance per unit length in the distributed parameter model, L is the inductance per unit length in the distributed parameter model, and C is the capacitance per unit length in the distributed parameter model.

[0112] Step 1.2: Based on the voltage ratios at various frequency points between the receiver and transmitter of the nth basic unit, obtain the channel frequency response H of the nth basic unit. (n) (f), as shown in equation (12):

[0113]

[0114] In equation (12), V r (n) V is the voltage at the receiving end of the nth basic unit. t (n) V is the voltage at the emitter of the nth basic unit. p (n) The voltage at the branch point of the nth basic unit;

[0115] in, As shown in equation (13), As shown in equation (14), that is:

[0116]

[0117]

[0118] In equation (13), γ2 (n) Let l2 be the propagation constant of the receiving side line of the nth basic unit. (n) Let Γ be the line length on the receiving side of the nth basic unit. L(n) Let be the reflection coefficient of the receiver of the nth basic unit, expressed by equation (15), that is:

[0119]

[0120] In equation (15), Z L (n) Z is the equivalent load impedance of the nth basic unit. c2 (n) The characteristic impedance of the receiving line;

[0121] In equation (14), γ1 (n) Let l1 be the propagation constant of the transmitting side line of the nth basic unit. (n) Let Γ be the line length on the transmitting side of the nth basic unit. P (n) Let be the reflection coefficient of the nth basic unit branch point, expressed by equation (16), that is:

[0122]

[0123] In equation (16), Z p (n) Z represents the equivalent load impedance at the nth basic unit branch point. c1 (n) The characteristic impedance of the transmitting side line;

[0124] Step 1.3: Using the broadband channel frequency response of each basic unit in the complex network obtained in Step 1.2, the channel frequency response of the entire network topology is obtained, as shown in Equation (17):

[0125]

[0126] Step 2: Based on the different distributed parameters of the cable under different types and degrees of thermal aging, obtain the power line channel model of the cable after thermal aging, and obtain the broadband channel frequency response under thermal aging conditions.

[0127] Step 2.1: Based on the characteristic that the complex dielectric constant of the cable insulation varies differently in the high-frequency range under different degrees of thermal aging, the complex dielectric constants of the insulation under the following conditions in the overall thermal aging of the cable (unaged, slightly aged, and severely aged) and the complex dielectric constants of the insulation under the following conditions in the local thermal aging (unaged, slightly aged, and severely aged) are fitted respectively in the frequency range of 1MHz-100MHz, as shown in Equation (18):

[0128]

[0129] In equation (18), ε' is the real part of the complex permittivity, ε” is the imaginary part of the complex permittivity, ε0 ​​is the permittivity in vacuum, w is the angular frequency, and M, N, and p are fitting coefficients, as shown in Table 1.

[0130] Table 1. Fitting parameter values ​​under different degrees of thermal aging.

[0131]

[0132] Step 2.2: Based on the relationship between the complex permittivity ε(w) and the capacitance C and conductance G in the distributed parameters, the distributed parameter model of the cable after thermal aging is obtained. Then, according to Equation (12), the power line channel model after the overall thermal aging of the cable when the entire cable is aged and the local thermal aging of the cable (the local cable with a length of 1-2m) is obtained respectively.

[0133] Among them, the basic unit selected is the one that simplifies the process of transmitting the equivalent impedance of the branches back to the main node, such as... Figure 3 As shown, the specific structure is as follows: a basic unit contains a trunk and a branch. When overall thermal aging occurs in this simple network or local thermal aging occurs in the cable, since the channel frequency response is obtained by iterative backtracking from the last basic unit, the total channel frequency response will change accordingly when a simple network undergoes thermal aging, and the oscillation attenuation of the channel frequency response is different under different thermal aging conditions.

[0134] Step 2.3: When the cable undergoes different degrees of thermal aging, its characteristic impedance and propagation constant will change accordingly, thus affecting the voltage at the transmitting end, receiving end, and branch point in the line. Therefore, for overall thermal aging and local thermal aging, the channel frequency response H1 of the nth basic unit under different degrees of thermal aging... (n) (f) As shown in equation (19);

[0135]

[0136] In equation (19), V r1 (n) V represents the voltage at the receiving end of the nth basic unit after cable aging. t1 (n) V represents the voltage at the transmitter of the nth basic unit after cable aging. p1 (n) This represents the voltage at the nth basic unit branch point after the cable has aged.

[0137] Step 2.4: Set the total length of the backbone line of each basic unit to 100m, and set it to either overall thermal aging or local thermal aging. The aging degree of each aging type is randomly set to no aging, slight thermal aging, and severe thermal aging, respectively. The line length of the branch section is set to 20m, and the load at the end of the branch is set to 50Ω. This results in 900 sets of channel frequency responses under different aging degrees corresponding to different aging types.

[0138] Step 3: Use the random forest classification algorithm to classify the channel frequency response H1 obtained in Step 2. (n) (f) Taking the cable aging type as input and the cable aging type as output, the cable aging is classified to obtain the cable aging type.

[0139] Step 3.1: Use the sparrow algorithm to optimize the random forest classification algorithm. Set the population size in the optimization algorithm to 5, set the maximum number of iterations to 100, and assign the optimal parameters obtained from the optimization algorithm to the network.

[0140] Step 3.2, take the aged channel frequency response H1 obtained in step 2. (n) (f) Using the cable thermal aging type as input and the cable thermal aging type as output, classification and prediction are performed, and the identification effect is as follows: Figure 4 , Figure 5 As shown, Figure 4 Category 0 indicates the cable has not aged, Category 1 indicates the cable has undergone overall thermal aging, and Category 2 indicates the cable has undergone localized thermal aging. Figure 5 This indicates that the groups classified as non-aged, overall thermally aged, and locally thermally aged were 311, 309, and 280, respectively. From... Figure 4 As can be seen, the classification accuracy reached 100%. When different types of thermal aging occur in the cable, they will affect the attenuation of the channel frequency response to varying degrees. Therefore, the random forest classification algorithm can identify whether thermal aging has occurred. If it has occurred, it can predict whether it is uniform thermal aging or local thermal aging.

[0141] Random forest is an algorithm that integrates multiple trees using the Bagging concept of ensemble learning. Its basic unit is the decision tree. It can handle input samples with high-dimensional features without dimensionality reduction, and can evaluate the importance of each feature in classification problems. During the generation process, it can obtain an unbiased estimate of the internal generation error. It is often used for classification and regression prediction problems. In the random forest algorithm, to classify an input sample, it needs to be input into each tree for classification. The classification results of several weak classifiers are voted to form a strong classifier, which ultimately achieves the classification of cable aging types.

[0142] Step 4: Use the random forest regression prediction algorithm to train the cable groups that have undergone overall thermal aging and local thermal aging obtained in Step 3, and predict the degree of thermal aging of the cable.

[0143] Step 4.1: Use the random forest regression prediction algorithm, set the number of decision trees in the random forest to 100 (range 80-100), the maximum number of features to 32 (range 9-32), do not limit the maximum depth of subtrees, set the minimum number of samples required for internal node subdivision to 2, and set the minimum number of samples for leaf nodes to 1.

[0144] Step 4.2: Using the channel frequency responses of the 309 cable groups classified as overall thermal aging and the 280 cable groups classified as local thermal aging obtained in Step 3 as input, the machine is trained to obtain the aging degree of the cable. The prediction accuracy A of the cable aging degree is shown in Equation (20), and the prediction error P... a As shown in equation (21);

[0145]

[0146] In the formula, T1 is the number of groups with correct prediction results, and N1 is the total number of groups;

[0147]

[0148] In the formula, Y predict Y represents the aging degree predicted by the random forest regression prediction algorithm. real This indicates the actual degree of aging.

[0149] The overall thermal aging degree prediction effect is as follows: Figure 6 As shown, the prediction accuracy reached 99%, and the prediction error rate was as follows: Figure 7 As shown, the maximum prediction error does not exceed 1.2%, and the prediction effect of the degree of local thermal aging is as follows. Figure 8 As shown, the prediction accuracy also reached 99%, and the prediction error rate was as follows: Figure 9 As shown, the maximum prediction error does not exceed 0.4%. In summary, whether it is overall thermal aging or local thermal aging, the prediction accuracy reaches over 99%, and the maximum error is less than 1.2%. Therefore, this method can effectively achieve online monitoring of cable thermal aging.

Claims

1. A method for online monitoring of thermal aging of a cable based on power line communication technology, characterized in that, The specific steps are as follows: Step 1: Obtain the characteristic impedance and propagation constant of the cable based on the distributed parameter model of the cable per unit length in high frequency; establish the power line channel model to obtain the broadband channel frequency response under high frequency conditions; The specific implementation method of step 1 is as follows: Step 1.1: First, divide the complex network topology of cable distribution in reality into N independent basic units; Then, the distributed parameters of the cable are calculated: resistance , inductance , conductance and capacitance ; wherein the resistance As shown in equation (1): （1） in formula (1) as shown in formula (2): （2） In formula (2), is the vacuum permeability, and = 4π x 10-7 H / m , is the frequency, is the electrical conductivity of the conductor, is the cable core radius; In equation (1), It is a correction coefficient, as shown in equation (3): （3） In the formula, This refers to the number of multi-strand wires in the outer ring of the cable. To form the radius of the multi-strand wires in the outer ring of the cable, For skin depth, among which: （4） （5） In equation (4), The total number of strands that make up the conductor core; inductance As shown in equation (6): （6） In equation (6), For the self-inductance of a conductor, The mutual inductance of the conductors is shown in equation (7): （7） In equation (7), The distance between the two conductors is 16 mm; capacitance As shown in equation (8): （8） In equation (8), The dielectric constant in vacuum is and , The relative permittivity of the aged portion of the cable; electrical conductivity As shown in equation (9): （9） In the distributed parameter model, the characteristic impedance of each line segment is: （10） The propagation constant is: （11） In the formula, The resistance per unit length in the distributed parameter model. The conductivity per unit length in the distributed parameter model. For the inductance per unit length in the distributed parameter model, The capacitance per unit length in the distributed parameter model; Step 1.2: Based on the voltage ratios at various frequency points between the receiver and transmitter of the nth basic unit, obtain the channel frequency response of the nth basic unit. As shown in equation (12): （12） In equation (12), The voltage at the receiving end of the nth basic unit. The voltage at the transmitter of the nth basic unit. The voltage at the branch point of the nth basic unit; in, As shown in equation (13), As shown in equation (14), that is: （13） （14） In equation (13), Let n be the propagation constant of the receiving side line of the nth basic unit. Let n be the line length on the receiving side of the nth basic unit. Let be the reflection coefficient of the receiver of the nth basic unit, expressed by equation (15), that is: （15） In equation (15), Let be the equivalent load impedance of the nth basic unit. The characteristic impedance of the receiving line; In equation (14), Let n be the propagation constant of the transmitting side line of the nth basic unit. Let n be the line length on the transmitting side of the nth basic unit. Let be the reflection coefficient of the nth basic unit branch point, expressed by equation (16), that is: （16） In equation (16), This represents the equivalent load impedance at the nth basic unit branch point. The characteristic impedance of the transmitting side line; Step 1.3: Using the broadband channel frequency response of each basic unit in the complex network obtained in Step 1.2, the channel frequency response of the entire network topology is obtained, as shown in Equation (17): （17）； Step 2: Based on the different distributed parameters of the cable under different types and degrees of thermal aging, obtain the power line channel model of the cable after thermal aging, and obtain the broadband channel frequency response under thermal aging conditions. The specific implementation method for step 2 is as follows: Step 2.1: Based on the characteristic that the complex dielectric constant of the cable insulation varies differently in the high-frequency range under different degrees of thermal aging, the complex dielectric constants of the insulation under the following conditions in the overall thermal aging of the cable (unaged, slightly aged, and severely aged) and the complex dielectric constants of the insulation under the following conditions in the local thermal aging (unaged, slightly aged, and severely aged) are fitted respectively in the frequency range of 1MHz-100MHz, as shown in Equation (18): （18） In equation (18), Let be the real part of the complex permittivity. This represents the imaginary part of the complex permittivity. The dielectric constant in vacuum. Angular frequency, , , All are fitting coefficients; Step 2.2, based on the complex permittivity With the capacitance in the distributed parameters Electrical conductivity The relationship between the two is used to obtain the distributed parameter model of the cable after thermal aging. Then, according to Equation (12), the power line channel model after the overall thermal aging of the cable when the whole cable is aged and the local thermal aging of the cable when the local cable is aged are obtained respectively. Step 2.3: For overall thermal aging and local thermal aging, the channel frequency response of the nth basic unit under different degrees of thermal aging. As shown in equation (19); （19） In equation (19), This represents the voltage at the receiving end of the nth basic unit after the cable has aged. This represents the voltage at the transmitter of the nth basic unit after the cable has aged. This represents the voltage at the nth basic unit branch point after the cable has aged. Step 2.4: Set the total length of the backbone line of each basic unit and set it to either overall thermal aging or local thermal aging. The aging degree of each aging type is randomly set to no aging, slight thermal aging, and severe thermal aging, respectively. Set the line length of the branch section and the load at the end of the branch to obtain the channel frequency response under different aging degrees corresponding to different aging types. Step 3: Using the random forest classification algorithm, take the channel frequency response obtained in Step 2 as input and the cable aging type as output to classify the cable aging and obtain the cable aging type. Step 4: Use the random forest regression prediction algorithm to train the cable groups that have undergone overall thermal aging and local thermal aging obtained in Step 3, and predict the degree of thermal aging of the cable.

2. The method for online monitoring of cable thermal aging based on power line communication technology according to claim 1, characterized in that, The specific implementation method for step 3 is as follows: Step 3.1: Use the Sparrow Algorithm to optimize the Random Forest classification algorithm, set the population size in the optimization algorithm, set the maximum number of iterations, and assign the optimal parameters obtained from the optimization algorithm to the network; Step 3.2, take the aged channel frequency response obtained in Step 2. The cable thermal aging type is used as input and as output for classification and prediction.

3. The method for online monitoring of cable thermal aging based on power line communication technology according to claim 2, characterized in that, The specific implementation method for step 4 is as follows: Step 4.1: Use the random forest regression prediction algorithm, set the number of decision trees in the random forest, set the maximum number of features, do not limit the maximum depth of subtrees, set the minimum number of samples required for internal node subdivision, and set the minimum number of samples for leaf nodes. Step 4.2: Using the channel frequency response of the cable groups classified as overall thermal aging and local thermal aging obtained in Step 3 as input, the machine is trained to obtain the degree of cable aging.

Images