A method and system for predicting dynamic current carrying capacity of three-core cables based on multi-dimensional correlation

Abstract

The present invention relates to the technical field of power cable current carrying capacity, and discloses a method and system for predicting the dynamic current carrying capacity of a three-core cable based on multidimensional correlation. The method obtains the cable temperature and the corresponding cable load of the three-core cable under different laying environment conditions, fits the cable temperature-load change series obtained after classification, calculates the characteristic value of the cable temperature, and optimizes the network parameters of a GRNN generalized regression neural network using an artificial bee colony algorithm. Iterative training is also performed using the characteristic value of the cable temperature as input and the cable load corresponding to the temperature data as output. The cable temperature measured at a future time is input into the trained GRNN generalized regression neural network, and a corresponding cable load prediction value at a future time is output. The dynamic current carrying capacity of the three-core cable is calculated based on the cable load prediction value at the future time and a preset maximum load limit of the three-core cable, thereby improving the prediction accuracy of the dynamic current carrying capacity of the three-core cable.

Description

Technical Field

[0001] The present invention relates to the technical field of power cable current carrying capacity, and in particular to a method and system for predicting the dynamic current carrying capacity of a three-core cable with multi-dimensional correlation. Background Art

[0002] Power cables, as crucial components of power transmission lines, are ubiquitous in urban construction. Their installation methods and environments are becoming increasingly complex as cities develop. During high-voltage cable operation, if the core temperature remains above the rated value for extended periods, cable losses will increase, thermal aging may accelerate, and even thermal breakdown may occur, shortening the cable's service life. Traditional cable current-carrying capacity calculations often differ significantly from actual values, leading to misjudgments of the cable's continuous current-carrying capacity and causing the cable to operate under long-term overload or heavy load conditions. Furthermore, the complex installation environment, diverse installation methods, and varying installation locations all contribute to a wide range of factors influencing cable current-carrying capacity. These factors, along with varying degrees of influence, contribute to the diverse temperature distribution of cables, further impacting the economic and reliability of cable operation and maintenance. Calculating and predicting the dynamic current-carrying capacity of cables has long been a challenge for the industry.

[0003] Chinese invention patent application publication number CN111707888A discloses a method for dynamically predicting cable conductor temperature, current carrying capacity, and withstand time. This method uses historical temperature and current information to fit the parameters of comprehensive factors affecting the thermal balance distribution of power cables. An iterative algorithm is then used to derive the conductor temperature, real-time current carrying capacity, and withstand time at different times. Using historical curves of ambient temperature, cable conductor temperature, and cable conductor load current, along with a set time step, the values ​​of these parameters at the previous five corresponding moments in the current state and time are obtained. Parameters affecting cable heat absorption and dissipation are substituted into the heat balance equation to solve for them, and the heat balance equation is modified based on the new parameters. Using the modified equation, current planning scheme, and iterative algorithm, the cable conductor temperature, current carrying capacity, and withstand time can be predicted. Because the modified equation can be adjusted in real time based on the current cable heat absorption and dissipation status, the prediction accuracy is high.

[0004] However, when predicting the current-carrying capacity of three-core cables / cable joints, due to the complex laying environment of three-core cables / cable joints, diverse laying methods, and variable laying locations, different laying methods and environmental conditions in different geographical locations have produced various factors that affect the current-carrying capacity of three-core cables. Moreover, due to different environmental conditions, the degree of influence of various factors on the current-carrying capacity of three-core cables is also different. Therefore, the selection of influencing factors has a large dispersion. The above method only selects some influencing factors as the overall model input parameters, which has great uncertainty.

[0005] At the same time, since the environmental conditions of the three-core cable may change, such as changes in ambient temperature and humidity, changes in surface temperature caused by long-term power supply of the three-core cable, and changes in the thermal conductivity of the surrounding laying environment, the degree of influence of various influencing factors on the current-carrying capacity of the three-core cable / cable joint will change. However, the above methods do not reflect the changes in influencing factors, resulting in low prediction accuracy of the dynamic current-carrying capacity of the three-core cable. In this case, if applied to industrial practice, it will cause serious misjudgments and omissions in power grid scheduling and fault diagnosis. Summary of the Invention

[0006] The present invention provides a method and system for predicting the dynamic current carrying capacity of a three-core cable based on multi-dimensional correlation, which solves the technical problem of low prediction accuracy of the dynamic current carrying capacity of the three-core cable.

[0007] In view of this, a first aspect of the present invention provides a method for predicting the dynamic current carrying capacity of a three-core cable based on multi-dimensional correlation, comprising the following steps:

[0008] Obtain the cable temperature and corresponding cable load of three-core cables under different laying environment conditions;

[0009] The cable temperature and its corresponding cable load are classified according to the corresponding laying environment conditions, and a set of cable temperature-load change series corresponding to multiple laying environment conditions is obtained;

[0010] Fit each set of cable temperature-load change series to obtain a temperature-load fitting change curve, and calculate the characteristic values ​​of the cable temperature in each set of temperature-load fitting change curves, including mean, mean square value, root mean square amplitude, standard deviation, peak value, peak-to-valley value, crest factor, form factor, variance, and skewness;

[0011] Constructing a GRNN generalized regression neural network, and optimizing network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain an optimized GRNN generalized regression neural network;

[0012] Based on the optimized GRNN generalized regression neural network, the characteristic value of cable temperature is used as input and the cable load corresponding to the temperature data is used as output to perform iterative training to obtain the trained GRNN generalized regression neural network.

[0013] Inputting the cable temperature measured at a future moment into the trained GRNN generalized regression neural network, and outputting the corresponding cable load prediction value at the future moment;

[0014] The dynamic current carrying capacity of the three-core cable is calculated based on the cable load forecast value at a future moment and the preset maximum load limit of the three-core cable.

[0015] Preferably, the laying environmental conditions include cable trench laying, laying exposed to the air and direct underground laying, wherein the influencing factors of the cable trench laying method include whether the cable is immersed in water and whether the cables are stacked; wherein the influencing factors of the direct underground laying method include soil thermal conductivity and boundary soil thermal conductivity; wherein the influencing factors of the laying method exposed to the air include direct sunlight time, rain time and water immersion time.

[0016] Preferably, before constructing a GRNN generalized regression neural network and optimizing the network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain the optimized GRNN generalized regression neural network, the step further includes:

[0017] The factor analysis method is used to perform dimensionality reduction processing on the eigenvalues, specifically including:

[0018] Convert all the eigenvalues ​​to The order matrix is:

[0019] Formula 1

[0020] In formula 1, Y is rank matrix, q is the number of eigenvalues, m is the number of eigenvalue samples, is the qth eigenvalue;

[0021] Will Defined as the original variable, factor analysis is used to decompose each original variable to obtain:

[0022] Formula 2

[0023] In formula 2, u q is the sample mean of the original variable, k qn is the factor loading of the original variable, n is the number of factors extracted from the original variable, n<q, is the common factor vector matrix, G n is the nth element in the common factor vector matrix, is a special factor vector matrix;

[0024] Simplify Equation 2 to:

[0025] Formula 3

[0026] Convert Equation 3 into a parameter sequence:

[0027] Formula 4

[0028] In formula 4, is the factor loading matrix, where

[0029]

[0030] SPSS statistical software was used to solve Formula 4 to obtain the common factor vector, and a scree plot of the common factor vector sequence was generated based on the common factor vector.

[0031] The slope between two adjacent common factor vectors is calculated according to the scree plot of the common factor vector sequence, and each common factor vector is traversed in sequence from left to right to determine whether the slope is greater than a preset slope threshold. If the slope is greater than the preset slope threshold, the traversed common factor vector is intercepted as the eigenvalue after dimensionality reduction.

[0032] Preferably, the steps of constructing a GRNN generalized regression neural network and optimizing the network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain the optimized GRNN generalized regression neural network specifically include:

[0033] Constructing a GRNN generalized regression neural network, wherein the GRNN generalized regression neural network includes an input layer, a pattern layer, a summation layer, and an output layer connected in sequence;

[0034] The input layer is used to receive input and pass the input to the model layer;

[0035] The transfer function of the mode layer is:

[0036] Formula 5

[0037] In formula 5, is the output of neuron i, X is the input, is the training sample of the i-th neuron, is the smoothing factor, n is the number of neurons;

[0038] The summation layer includes two types of neurons, wherein the first type of neurons sums the output values ​​of all neurons in the pattern layer, and the summation formula is:

[0039] Formula 6

[0040] In formula 6, S D is the sum value, T is the transpose symbol;

[0041] The second type of neurons assign weights to the output values ​​of all neurons in the pattern layer, thereby performing weighted summation. The weighted summation formula is:

[0042] Formula 7

[0043] In formula 7, S Nj is the weighted sum value, y ij is the output sample y i The jth element of Y i is the i-th training sample, j is the number of elements, and k is the total number of elements;

[0044] The number of neurons in the output layer is the dimension of the output vector. The output layer representation function is obtained by dividing the weighted sum of the neurons in each layer by the unweighted sum:

[0045]

[0046] In formula 8, y i is the i-th output;

[0047] The artificial bee colony algorithm is used to optimize the smoothing factor, including:

[0048] Randomly generate a population and initialize it. The initialization formula is:

[0049] Formula 9

[0050] In Equation 9, s is the sth particle in the population, w is the dimension, is a random number between 0 and 1, is the location of the w-th dimension food source of the s-th solution in the population, are the maximum value and the minimum value of the food source location respectively;

[0051] Hire bees to find food sources and update the location of the food sources to:

[0052] Formula 10

[0053] In formula 10, is the location of the food source after the location update, a is a randomly generated value, It is a random number between -1 and 1;

[0054] After the food source is mined, observe the location of the bees towards the food source The probability of selection is calculated as follows:

[0055]

[0056] In formulas 11 and 12, S is the number of particles, is the fitness function value of the food source, is the objective function of the food source, which is the error function between the calculated output of the GRNN generalized regression neural network and the training sample value under a specific smooth factor condition;

[0057] Find the position of the observer bee relative to the food source The smoothing factor corresponding to the maximum probability is selected as the optimal smoothing factor, and the optimal smoothing factor is applied to the GRNN generalized neural network.

[0058] Preferably, the step of calculating the dynamic current carrying capacity of the three-core cable according to the cable load prediction value at a future moment and the preset maximum load limit of the three-core cable specifically includes:

[0059] The difference between the predicted cable load value at a future moment and the preset maximum load limit of the three-core cable is calculated, and the difference result is the dynamic current carrying capacity of the three-core cable.

[0060] In a second aspect, the present invention provides a multi-dimensional correlation three-core cable dynamic current carrying capacity prediction system, comprising:

[0061] A data acquisition module is used to obtain the cable temperature and corresponding cable load of the three-core cable under different laying environment conditions;

[0062] A classification module is used to classify the cable temperature and its corresponding cable load according to the corresponding laying environment conditions, and obtain a set of cable temperature-load change series corresponding to multiple laying environment conditions;

[0063] The characteristic value module is used to fit each set of cable temperature-load change series to obtain a temperature-load fitting change curve, and calculate the characteristic values ​​of the cable temperature in each set of temperature-load fitting change curves. The characteristic values ​​include mean, mean square value, root mean square amplitude, standard deviation, peak value, peak-to-valley value, crest factor, shape factor, variance and skewness;

[0064] A network construction module is used to construct a GRNN generalized regression neural network, and optimize the network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain an optimized GRNN generalized regression neural network;

[0065] The training module is used to perform iterative training based on the optimized GRNN generalized regression neural network, taking the characteristic value of the cable temperature as the input and the cable load corresponding to the temperature data as the output, to obtain a trained GRNN generalized regression neural network;

[0066] A load prediction module is used to input the cable temperature measured at a future time into the trained GRNN generalized regression neural network and output a corresponding cable load prediction value at a future time;

[0067] The current carrying capacity prediction module is used to calculate the dynamic current carrying capacity of the three-core cable based on the cable load prediction value at a future moment and the preset maximum load limit of the three-core cable.

[0068] Preferably, the laying environmental conditions include cable trench laying, laying exposed to the air and direct underground laying, wherein the influencing factors of the cable trench laying method include whether the cable is immersed in water and whether the cables are stacked; wherein the influencing factors of the direct underground laying method include soil thermal conductivity and boundary soil thermal conductivity; wherein the influencing factors of the laying method exposed to the air include direct sunlight time, rain time and water immersion time.

[0069] Preferably, the system further includes:

[0070] The dimensionality reduction module is used to perform dimensionality reduction processing on the eigenvalues ​​using factor analysis, specifically including:

[0071] Convert all the eigenvalues ​​to The order matrix is:

[0072] Formula 1

[0073] In formula 1, Y is rank matrix, q is the number of eigenvalues, m is the number of eigenvalue samples, is the qth eigenvalue;

[0074] Will Defined as the original variable, factor analysis is used to decompose each original variable to obtain:

[0075] Formula 2

[0076] In formula 2, u q is the sample mean of the original variable, k qn is the factor loading of the original variable, n is the number of factors extracted from the original variable, n<q, is the common factor vector matrix, G n is the nth element in the common factor vector matrix, is a special factor vector matrix;

[0077] Simplify Equation 2 to:

[0078] Formula 3

[0079] Convert Equation 3 into a parameter sequence:

[0080] Formula 4

[0081] In formula 4, is the factor loading matrix, where

[0082]

[0083] SPSS statistical software was used to solve Formula 4 to obtain the common factor vector, and a scree plot of the common factor vector sequence was generated based on the common factor vector.

[0084] The slope between two adjacent common factor vectors is calculated according to the scree plot of the common factor vector sequence, and each common factor vector is traversed in sequence from left to right to determine whether the slope is greater than a preset slope threshold. If the slope is greater than the preset slope threshold, the traversed common factor vector is intercepted as the eigenvalue after dimensionality reduction.

[0085] Preferably, the network construction module specifically includes:

[0086] A network initialization module is used to construct a GRNN generalized regression neural network, wherein the GRNN generalized regression neural network includes an input layer, a pattern layer, a summation layer, and an output layer connected in sequence;

[0087] The input layer is used to receive input and pass the input to the model layer;

[0088] The transfer function of the mode layer is:

[0089] Formula 5

[0090] In formula 5, is the output of neuron i, X is the input, is the training sample of the i-th neuron, is the smoothing factor, n is the number of neurons;

[0091] The summation layer includes two types of neurons, wherein the first type of neurons sums the output values ​​of all neurons in the pattern layer, and the summation formula is:

[0092] Formula 6

[0093] In formula 6, S D is the sum value, T is the transpose symbol;

[0094] The second type of neurons assign weights to the output values ​​of all neurons in the pattern layer, thereby performing weighted summation. The weighted summation formula is:

[0095] Formula 7

[0096] In formula 7, S Nj is the weighted sum value, y ij is the output sample y i The jth element of Y i is the i-th training sample, j is the number of elements, and k is the total number of elements;

[0097] The number of neurons in the output layer is the dimension of the output vector. The output layer representation function is obtained by dividing the weighted sum of the neurons in each layer by the unweighted sum:

[0098] Formula 8

[0099] In formula 8, y i is the i-th output;

[0100] The network optimization module is used to optimize the smoothing factor using the artificial bee colony algorithm, specifically including:

[0101] Randomly generate a population and initialize it. The initialization formula is:

[0102] Formula 9

[0103] In Equation 9, s is the sth particle in the population, w is the dimension, is a random number between 0 and 1, is the location of the w-th dimension food source of the s-th solution in the population, are the maximum value and the minimum value of the food source location respectively;

[0104] Hire bees to find food sources and update the location of the food sources to:

[0105] Formula 10

[0106] In formula 10, is the location of the food source after the location update, a is a randomly generated value, It is a random number between -1 and 1;

[0107] After the food source is mined, observe the location of the bees towards the food source The probability of selection is calculated as follows:

[0108]

[0109] In formulas 11 and 12, S is the number of particles, is the fitness function value of the food source, is the objective function of the food source, which is the error function between the calculated output of the GRNN generalized regression neural network and the training sample value under a specific smooth factor condition;

[0110] Find the position of the observer bee relative to the food source The smoothing factor corresponding to the maximum probability is selected as the optimal smoothing factor, and the optimal smoothing factor is applied to the GRNN generalized neural network.

[0111] Preferably, the current carrying capacity prediction module is specifically used to calculate the difference between the cable load prediction value at a future moment and the preset maximum load limit of the three-core cable, and the difference result is the dynamic current carrying capacity of the three-core cable.

[0112] It can be seen from the above technical solutions that the present invention has the following advantages:

[0113] The present invention obtains the cable temperature and the corresponding cable load of the three-core cable under different laying environment conditions to consider multi-dimensional influencing factors, classifies the cable temperature and the corresponding cable load according to the corresponding laying environment conditions, and fits the cable temperature-load change series set obtained after classification to obtain a temperature-load fitting change curve, calculates the characteristic value of the cable temperature in each set of temperature-load fitting change curves, and uses an artificial bee colony algorithm to optimize the network parameters of the GRNN generalized regression neural network to obtain an optimized GRNN generalized regression neural network. It also uses the characteristic value of the cable temperature as an input and the cable load corresponding to the temperature data as an output to perform iterative training to obtain a trained GRNN generalized regression neural network, inputs the cable temperature measured at a future time into the trained GRNN generalized regression neural network, outputs the corresponding cable load prediction value at the future time, and calculates the dynamic current-carrying capacity of the three-core cable based on the cable load prediction value at the future time and the preset maximum load limit of the three-core cable, thereby improving the prediction accuracy of the dynamic current-carrying capacity of the three-core cable. BRIEF DESCRIPTION OF THE DRAWINGS

[0114] Figure 1 A flowchart of a method for predicting dynamic current carrying capacity of a three-core cable based on multi-dimensional correlation provided by an embodiment of the present invention;

[0115] Figure 2 A scree map provided by an embodiment of the present invention;

[0116] Figure 3 A schematic structural diagram of a multi-dimensional correlation three-core cable dynamic current carrying capacity prediction system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0117] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0118] For easier understanding, see Figure 1The present invention provides a method for predicting the dynamic current carrying capacity of a three-core cable based on multi-dimensional correlation, comprising the following steps:

[0119] S1. Obtain the cable temperature and corresponding cable load of the three-core cable under different laying environment conditions.

[0120] Among them, the laying environment conditions include cable trench laying method, laying method exposed to the air and direct burial underground laying method. Among them, the influencing factors of the cable trench laying method include whether the cable is immersed in water and whether the cables are stacked; among them, the influencing factors of the direct burial underground laying method include soil thermal conductivity and boundary soil thermal conductivity; among them, the influencing factors of the laying method exposed to the air include direct sunlight time, rain time and water immersion time.

[0121] It should be noted that, considering the complex laying environment of three-core cables, the diverse laying methods and the changing laying locations, different laying methods and environmental conditions in different geographical locations have produced a variety of factors that affect the current-carrying capacity of three-core cables / cable connectors. Moreover, due to different environmental conditions, the degree of influence of various factors on the current-carrying capacity of three-core cables / cable connectors is also different. Therefore, the possible influencing factors are selected as comprehensively as possible.

[0122] The laying environment conditions include cable trench laying, laying exposed in the air and laying directly underground.

[0123] For the personalized influencing factors in different laying environment conditions: the influencing factors for the cable trench laying method include: whether it is immersed in water (including full immersion and partial immersion), whether there is cable stacking (that is, multiple cables are entangled together).

[0124] The factors affecting the direct underground laying method are: for uniform soil environment, consider the soil thermal conductivity; for uneven soil environment, consider the boundary soil thermal conductivity.

[0125] For laying cables exposed to the air, consider the time of direct sunlight, short-term rain, long and short-term immersion in water, and long and short-term wind.

[0126] For each of the above situations, after confirming that the cable structure is normal, conduct a flow temperature measurement experiment. Record the temperature and load changes to form a surface temperature-load data set.

[0127] S2. Classify the cable temperature and its corresponding cable load according to the corresponding laying environment conditions to obtain a set of cable temperature-load change series corresponding to multiple laying environment conditions;

[0128] S3. Fit each set of cable temperature-load change series to obtain a temperature-load fitting change curve, and calculate the characteristic values ​​of the cable temperature in each set of temperature-load fitting change curves. The characteristic values ​​include mean, mean square value, root mean square amplitude, standard deviation, peak value, peak-to-valley value, crest factor, form factor, variance, and skewness.

[0129] The horizontal axis of the temperature-load fitting change curve can be the load value, and the vertical axis is the temperature value. The characteristic values ​​of the cable temperature in each set of temperature-load fitting change curves are calculated. The characteristic values ​​include mean, mean square value, root mean square amplitude, standard deviation, peak value, peak-to-valley value, crest factor, form factor, variance, and skewness.

[0130] The calculation formula for the mean is:

[0131]

[0132] Where, is the mean, Z is the total number of samples of the eigenvalue, z is the zth sample of the eigenvalue, and T(z) is the cable temperature of the zth sample;

[0133] The formula for calculating the mean square value is:

[0134]

[0135] Where, F rms is the mean square value;

[0136] The formula for calculating the square root amplitude is:

[0137]

[0138] Where, F sra is the square root amplitude;

[0139] The formula for calculating the standard deviation is:

[0140]

[0141] Where, F std is the standard deviation;

[0142] The peak value is calculated as:

[0143]

[0144] Where, F p is the peak value;

[0145] The calculation formula for the peak-to-valley value is:

[0146]

[0147] Where, Fptv is the peak-to-valley value, are the maximum and minimum cable temperature respectively;

[0148] The formula for calculating the crest factor is:

[0149]

[0150] Where, F pi is the peak factor;

[0151] The calculation formula for the form factor is:

[0152]

[0153] Where, F wi is the form factor;

[0154] The formula for calculating variance is:

[0155]

[0156] Where, F var is the variance;

[0157] The calculation formula for skewness is:

[0158]

[0159] Where, F sk is the skewness.

[0160] S4, constructing a GRNN generalized regression neural network, and optimizing the network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain an optimized GRNN generalized regression neural network;

[0161] S5. Based on the optimized GRNN generalized regression neural network, iterative training is performed with the characteristic value of the cable temperature as input and the cable load corresponding to the temperature data as output to obtain a trained GRNN generalized regression neural network;

[0162] S6. Input the cable temperature measured at the future moment into the trained GRNN generalized regression neural network, and output the corresponding cable load prediction value at the future moment;

[0163] S7. Calculate the dynamic current carrying capacity of the three-core cable based on the cable load prediction value at a future time and the preset maximum load limit of the three-core cable.

[0164] The present invention provides a method for predicting the dynamic current carrying capacity of a three-core cable with multi-dimensional correlation. The method obtains the cable temperature and the corresponding cable load of the three-core cable under different laying environment conditions, considers multi-dimensional influencing factors, classifies the cable temperature and the corresponding cable load according to the corresponding laying environment conditions, fits the cable temperature-load change series obtained after the classification to obtain a temperature-load fitting change curve, calculates the characteristic value of the cable temperature in each group of temperature-load fitting change curves, optimizes the network parameters of a GRNN generalized regression neural network using an artificial bee colony algorithm to obtain an optimized GRNN generalized regression neural network, and iteratively trains the eigenvalue of the cable temperature as an input and the cable load corresponding to the temperature data as an output to obtain a trained GRNN generalized regression neural network. The cable temperature measured at a future time is input into the trained GRNN generalized regression neural network, and a corresponding cable load prediction value at the future time is output. The dynamic current carrying capacity of the three-core cable is calculated based on the cable load prediction value at the future time and a preset maximum load limit of the three-core cable, thereby improving the prediction accuracy of the dynamic current carrying capacity of the three-core cable.

[0165] In a specific embodiment, before step S4, the method further includes:

[0166] S31. Use factor analysis to reduce the dimensionality of eigenvalues.

[0167] Among them, factor analysis is used to reduce the dimensionality of the existing data set. Factor analysis reduces the complex relationships between these variables to a few comprehensive factors. The number of factors is smaller than the original variables, but they have all the information of the original variables.

[0168] The process of using factor analysis to reduce the dimensionality of eigenvalues ​​specifically includes:

[0169] Convert all eigenvalues ​​to The order matrix is:

[0170] Formula 1

[0171] In formula 1, Y is rank matrix, q is the number of eigenvalues, m is the number of eigenvalue samples, is the qth eigenvalue;

[0172] Will Defined as the original variable, factor analysis is used to decompose each original variable to obtain:

[0173] Formula 2

[0174] In formula 2, u q is the sample mean of the original variable, k qnis the factor loading of the original variable, n is the number of factors extracted from the original variable, n<q, is the common factor vector matrix, G n is the nth element in the common factor vector matrix, is a special factor vector matrix;

[0175] Simplify Equation 2 to:

[0176] Formula 3

[0177] Convert Equation 3 into a parameter sequence:

[0178] Formula 4

[0179] In formula 4, is the factor loading matrix, where

[0180]

[0181] SPSS statistical software was used to solve Formula 4 to obtain the common factor vector, and a scree plot of the common factor vector sequence was generated based on the common factor vector.

[0182] Among them, using statistical SPSS software to solve formula 4, we can get all the parameter values, including the common factor vector matrix G, the special factor vector matrix , the sample mean of the original variable and the factor loading matrix K, where the common factor vector is the characteristic factor representing the eigenvalue. Figure 2 As shown, the horizontal axis is the serial number and the vertical axis is the characteristic value.

[0183] The slope between two adjacent common factor vectors is calculated according to the scree plot of the common factor vector sequence. Each common factor vector is traversed from left to right in the sequence to determine whether the slope is greater than the preset slope threshold. If the slope is greater than the preset slope threshold, the traversed common factor vector is intercepted as the eigenvalue after dimensionality reduction.

[0184] Among them, such as Figure 2 If the slopes of the first two points change greatly, only G1 and G2 are taken and the rest are discarded. If the slopes of the first three points change greatly, G1, G2, and G3 are taken and the rest of the G values ​​are discarded.

[0185] According to Equation 4, since each row can be expanded into q items, after the above calculation, only the first several of the q items are taken for each row, which reduces the amount of data and the dimension, thus achieving dimensionality reduction.

[0186] In a specific embodiment, step S4 specifically includes:

[0187] S401, constructing a GRNN generalized regression neural network, wherein the GRNN generalized regression neural network includes an input layer, a pattern layer, a summation layer, and an output layer connected in sequence;

[0188] Among them, the input layer is used to receive input and pass the input to the model layer;

[0189] The transfer function of the pattern layer is:

[0190] Formula 5

[0191] In formula 5, is the output of neuron i, X is the input, is the training sample of the i-th neuron, is the smoothing factor, n is the number of neurons;

[0192] The summation layer includes two types of neurons. The first type of neurons sums the output values ​​of all neurons in the pattern layer. The summation formula is:

[0193] Formula 6

[0194] In formula 6, S D is the sum value, T is the transpose symbol;

[0195] The second type of neurons assign weights to the output values ​​of all neurons in the pattern layer, thereby performing weighted summation. The formula for weighted summation is:

[0196] Formula 7

[0197] In formula 7, S Nj is the weighted sum value, y ij is the output sample y i The jth element of Y i is the i-th training sample, j is the number of elements, and k is the total number of elements;

[0198] The number of neurons in the output layer is the dimension of the output vector. The output layer representation function is obtained by dividing the weighted sum of the neurons in each layer by the unweighted sum:

[0199] Formula 8

[0200] In formula 8, y i is the i-th output;

[0201] It should be noted that according to the theory of generalized regression neural network, the generalized regression neural network does not need to set the form of the mapping equation in advance, but the smoothing factor It has a great influence on the prediction results. Selecting the best parameters can significantly improve the prediction accuracy.

[0202] S402: Optimizing the smoothing factor using an artificial bee colony algorithm, specifically including:

[0203] S4021. Randomly generate a population and initialize the population. The initialization formula is:

[0204] Formula 9

[0205] In Equation 9, s is the sth particle in the population, w is the dimension, is a random number between 0 and 1, is the location of the w-th dimension food source of the s-th solution in the population, are the maximum value and the minimum value of the food source location respectively;

[0206] S4022. Hire bees to find a food source and update the location of the food source to:

[0207] Formula 10

[0208] In formula 10, is the location of the food source after the location update, a is a randomly generated value, It is a random number between -1 and 1;

[0209] S4023. After the food source is mined, observe the location of the bees to the food source. The probability of selection is calculated as follows:

[0210]

[0211] In formulas 11 and 12, S is the number of particles, is the fitness function value of the food source, is the objective function of the food source, which is the error function between the calculated output of the GRNN generalized regression neural network and the training sample value under a specific smooth factor condition;

[0212] S4024. Obtain the position of the observing bee relative to the food source The smoothing factor corresponding to the maximum probability is selected as the optimal smoothing factor, and the optimal smoothing factor is applied to the GRNN generalized neural network.

[0213] Among them, based on experience Assign values ​​within a large range, for example, 1-100. Generate random matrices s and w within this range. S and w can be randomly selected. A larger dimension yields better optimization results, but also increases the computational complexity. However, s and w cannot be less than q. In this example, q is 10. When the maximum number of optimizations is reached, the iteration stops and the optimal smoothing factor is output.

[0214] In a specific embodiment, step S7 specifically includes:

[0215] The difference between the predicted cable load value at a future moment and the preset maximum load limit of the three-core cable is calculated, and the difference result is the dynamic current carrying capacity of the three-core cable.

[0216] The above is a detailed description of an embodiment of a method for predicting the dynamic current carrying capacity of a three-core cable with multi-dimensional correlation provided by the present invention. The following is a detailed description of an embodiment of a system for predicting the dynamic current carrying capacity of a three-core cable with multi-dimensional correlation provided by the present invention.

[0217] For easier understanding, see Figure 3 The present invention provides a multi-dimensional correlation three-core cable dynamic current carrying capacity prediction system, comprising:

[0218] The data acquisition module 100 is used to obtain the cable temperature and the corresponding cable load of the three-core cable under different laying environment conditions;

[0219] A classification module 200 is used to classify the cable temperature and its corresponding cable load according to the corresponding laying environment conditions, and obtain a set of cable temperature-load change series corresponding to multiple laying environment conditions;

[0220] The characteristic value module 300 is used to fit each set of cable temperature-load change series to obtain a temperature-load fitting change curve, and calculate the characteristic values ​​of the cable temperature in each set of temperature-load fitting change curves. The characteristic values ​​include mean, mean square value, root mean square amplitude, standard deviation, peak value, peak-to-valley value, crest factor, shape factor, variance and skewness;

[0221] The network construction module 400 is used to construct a GRNN generalized regression neural network and optimize the network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain an optimized GRNN generalized regression neural network.

[0222] The training module 500 is used to perform iterative training based on the optimized GRNN generalized regression neural network, using the characteristic value of the cable temperature as input and the cable load corresponding to the temperature data as output, to obtain a trained GRNN generalized regression neural network;

[0223] The load prediction module 600 is used to input the cable temperature measured at a future time into the trained GRNN generalized regression neural network and output the corresponding cable load prediction value at the future time;

[0224] The current carrying capacity prediction module 700 is used to calculate the dynamic current carrying capacity of the three-core cable based on the cable load prediction value at a future moment and the preset maximum load limit of the three-core cable.

[0225] In a specific embodiment, the laying environment conditions include cable trench laying, laying exposed to the air, and direct underground laying. Among them, the influencing factors of the cable trench laying method include whether the cable is immersed in water and whether the cables are stacked; among them, the influencing factors of the direct underground laying method include soil thermal conductivity and boundary soil thermal conductivity; among them, the influencing factors of the laying method exposed to the air include direct sunlight time, rain time, and water immersion time.

[0226] In a specific embodiment, the system further includes:

[0227] The dimensionality reduction module is used to reduce the dimensionality of eigenvalues ​​using factor analysis, specifically including:

[0228] Convert all eigenvalues ​​to The order matrix is:

[0229] Formula 1

[0230] In formula 1, Y is rank matrix, q is the number of eigenvalues, m is the number of eigenvalue samples, is the qth eigenvalue;

[0231] Will Defined as the original variable, factor analysis is used to decompose each original variable to obtain:

[0232] Formula 2

[0233] In formula 2, u q is the sample mean of the original variable, k qn is the factor loading of the original variable, n is the number of factors extracted from the original variable, n<q, is the common factor vector matrix, G n is the nth element in the common factor vector matrix, is a special factor vector matrix;

[0234] Simplify Equation 2 to:

[0235] Formula 3

[0236] Convert Equation 3 into a parameter sequence:

[0237] Formula 4

[0238] In formula 4, is the factor loading matrix, where

[0239]

[0240] SPSS statistical software was used to solve Formula 4 to obtain the common factor vector, and a scree plot of the common factor vector sequence was generated based on the common factor vector.

[0241] The slope between two adjacent common factor vectors is calculated according to the scree plot of the common factor vector sequence. Each common factor vector is traversed from left to right in the sequence to determine whether the slope is greater than the preset slope threshold. If the slope is greater than the preset slope threshold, the traversed common factor vector is intercepted as the eigenvalue after dimensionality reduction.

[0242] In a specific embodiment, the network construction module specifically includes:

[0243] The network initialization module is used to construct the GRNN generalized regression neural network, which includes an input layer, a pattern layer, a summation layer, and an output layer connected in sequence;

[0244] Among them, the input layer is used to receive input and pass the input to the model layer;

[0245] The transfer function of the pattern layer is:

[0246] Formula 5

[0247] In formula 5, is the output of neuron i, X is the input, is the training sample of the i-th neuron, is the smoothing factor, n is the number of neurons;

[0248] The summation layer includes two types of neurons. The first type of neurons sums the output values ​​of all neurons in the pattern layer. The summation formula is:

[0249] Formula 6

[0250] In formula 6, S D is the sum value, T is the transpose symbol;

[0251] The second type of neurons assign weights to the output values ​​of all neurons in the pattern layer, thereby performing weighted summation. The formula for weighted summation is:

[0252] Formula 7

[0253] In formula 7, S Nj is the weighted sum value, y ij is the output sample y i The jth element of Y i is the i-th training sample, j is the number of elements, and k is the total number of elements;

[0254] The number of neurons in the output layer is the dimension of the output vector. The output layer representation function is obtained by dividing the weighted sum of the neurons in each layer by the unweighted sum:

[0255] Formula 8

[0256] In formula 8, y i is the i-th output;

[0257] The network optimization module is used to optimize the smoothing factor using the artificial bee colony algorithm, specifically including:

[0258] Randomly generate a population and initialize it. The initialization formula is:

[0259] Formula 9

[0260] In Equation 9, s is the sth particle in the population, w is the dimension, is a random number between 0 and 1, is the location of the w-th dimension food source of the s-th solution in the population, are the maximum value and the minimum value of the food source location respectively;

[0261] Hire bees to find food sources and update the location of the food sources to:

[0262] Formula 10

[0263] In formula 10, is the location of the food source after the location update, a is a randomly generated value, It is a random number between -1 and 1;

[0264] After the food source is mined, observe the location of the bees towards the food source The probability of selection is calculated as follows:

[0265]

[0266] In formulas 11 and 12, S is the number of particles, is the fitness function value of the food source, is the objective function of the food source, which is the error function between the calculated output of the GRNN generalized regression neural network and the training sample value under a specific smooth factor condition;

[0267] Find the position of the observer bee relative to the food source The smoothing factor corresponding to the maximum probability is selected as the optimal smoothing factor, and the optimal smoothing factor is applied to the GRNN generalized neural network.

[0268] In a specific embodiment, the current carrying capacity prediction module is specifically used to calculate the difference between the cable load prediction value at a future moment and the preset maximum load limit of the three-core cable, and the difference result is the dynamic current carrying capacity of the three-core cable.

[0269] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working process of the system described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0270] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0271] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0272] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0273] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for predicting the dynamic current carrying capacity of a three-core cable based on multi-dimensional correlation, characterized in that: The following steps are involved: Obtain the cable temperature and corresponding cable load of three-core cables under different laying environment conditions; The cable temperature and its corresponding cable load are classified according to the corresponding laying environment conditions, and a set of cable temperature-load change series corresponding to multiple laying environment conditions is obtained; The laying environmental conditions include cable trench laying, laying exposed to the air, and direct underground laying. The influencing factors of the cable trench laying method include whether the cable is immersed in water and whether the cables are stacked. The influencing factors of the direct underground laying method include soil thermal conductivity and interfacial soil thermal conductivity. The influencing factors of the laying method exposed to the air include direct sunlight time, rain time, and water immersion time. Fit each set of cable temperature-load change series to obtain a temperature-load fitting change curve, and calculate the characteristic values ​​of the cable temperature in each set of temperature-load fitting change curves, including mean, mean square value, root mean square amplitude, standard deviation, peak value, peak-to-valley value, crest factor, form factor, variance, and skewness; Constructing a GRNN generalized regression neural network, and optimizing network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain an optimized GRNN generalized regression neural network; Based on the optimized GRNN generalized regression neural network, the characteristic value of cable temperature is used as input and the cable load corresponding to the temperature data is used as output to perform iterative training to obtain the trained GRNN generalized regression neural network. Inputting the cable temperature measured at a future moment into the trained GRNN generalized regression neural network, and outputting the corresponding cable load prediction value at the future moment; The dynamic current carrying capacity of the three-core cable is calculated based on the cable load forecast value at a future moment and the preset maximum load limit of the three-core cable.

2. The method for predicting the dynamic current carrying capacity of a three-core cable based on multi-dimensional correlation according to claim 1, characterized in that: Before constructing a GRNN generalized regression neural network and optimizing the network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain the optimized GRNN generalized regression neural network, the method further includes: The factor analysis method is used to perform dimensionality reduction processing on the eigenvalues, specifically including: Convert all the eigenvalues ​​to The order matrix is: Formula 1 In formula 1, Y is rank matrix, q is the number of eigenvalues, m is the number of eigenvalue samples, is the qth eigenvalue; Will Defined as the original variable, factor analysis is used to decompose each original variable to obtain: ； Formula 2 In formula 2, u q is the sample mean of the original variable, k qn is the factor loading of the original variable, n is the number of factors extracted from the original variable, n<q, is the common factor vector matrix, G n is the nth element in the common factor vector matrix, is a special factor vector matrix; Simplify Equation 2 to: Formula 3 Convert Equation 3 into a parameter sequence: Formula 4 In formula 4, is the factor loading matrix, where ； SPSS statistical software was used to solve Formula 4 to obtain the common factor vector, and a scree plot of the common factor vector sequence was generated based on the common factor vector. The slope between two adjacent common factor vectors is calculated according to the scree plot of the common factor vector sequence, and each common factor vector is traversed in sequence from left to right to determine whether the slope is greater than a preset slope threshold. If the slope is greater than the preset slope threshold, the traversed common factor vector is intercepted as the eigenvalue after dimensionality reduction.

3. The method for predicting the dynamic current carrying capacity of a three-core cable based on multi-dimensional correlation according to claim 1, characterized in that: The steps of constructing a GRNN generalized regression neural network and optimizing the network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain the optimized GRNN generalized regression neural network specifically include: Constructing a GRNN generalized regression neural network, wherein the GRNN generalized regression neural network includes an input layer, a pattern layer, a summation layer, and an output layer connected in sequence; The input layer is used to receive input and pass the input to the model layer; The transfer function of the mode layer is: Formula 5 In formula 5, is the output of neuron i, X is the input, is the training sample of the i-th neuron, is the smoothing factor, n is the number of neurons; The summation layer includes two types of neurons, wherein the first type of neurons sums the output values ​​of all neurons in the pattern layer, and the summation formula is: Formula 6 In formula 6, S D is the sum value, T is the transpose symbol; The second type of neurons assign weights to the output values ​​of all neurons in the pattern layer, thereby performing weighted summation. The weighted summation formula is: Formula 7 In formula 7, S Nj is the weighted sum value, y ij is the output sample y i The jth element of Y i is the i-th training sample, j is the number of elements, and k is the total number of elements; The number of neurons in the output layer is the dimension of the output vector. The output layer representation function is obtained by dividing the weighted sum of the neurons in each layer by the unweighted sum: Formula 8 In formula 8, y i is the i-th output; The artificial bee colony algorithm is used to optimize the smoothing factor, including: Randomly generate a population and initialize it. The initialization formula is: Formula 9 In Equation 9, s is the sth particle in the population, w is the dimension, is a random number between 0 and 1, is the location of the w-th dimension food source of the s-th solution in the population, are the maximum value and the minimum value of the food source location respectively; Hire bees to find food sources and update the location of the food sources to: Formula 10 In formula 10, is the location of the food source after the location update, a is a randomly generated value, It is a random number between -1 and 1; After the food source is mined, observe the location of the bees towards the food source The probability of selection is calculated as follows: Formula 11 Formula 12 In formulas 11 and 12, S is the number of particles, is the fitness function value of the food source, is the objective function of the food source, which is the error function between the calculated output of the GRNN generalized regression neural network and the training sample value under a specific smooth factor condition; Find the position of the observer bee relative to the food source The smoothing factor corresponding to the maximum probability is selected as the optimal smoothing factor, and the optimal smoothing factor is applied to the GRNN generalized neural network.

4. The method for predicting dynamic current carrying capacity of a three-core cable based on multi-dimensional correlation according to claim 1, characterized in that: The steps for calculating the dynamic current carrying capacity of the three-core cable based on the cable load prediction value at a future time and the preset maximum load limit of the three-core cable include: The difference between the predicted cable load value at a future moment and the preset maximum load limit of the three-core cable is calculated, and the difference result is the dynamic current carrying capacity of the three-core cable.

5. A multi-dimensional correlation three-core cable dynamic current carrying capacity prediction system, characterized in that: include: A data acquisition module is used to obtain the cable temperature and corresponding cable load of the three-core cable under different laying environment conditions; A classification module is used to classify the cable temperature and its corresponding cable load according to the corresponding laying environment conditions, and obtain a set of cable temperature-load change series corresponding to multiple laying environment conditions; The laying environmental conditions include cable trench laying, laying exposed to the air, and direct underground laying. The influencing factors of the cable trench laying method include whether the cable is immersed in water and whether the cables are stacked. The influencing factors of the direct underground laying method include soil thermal conductivity and interfacial soil thermal conductivity. The influencing factors of the laying method exposed to the air include direct sunlight time, rain time, and water immersion time. The characteristic value module is used to fit each set of cable temperature-load change series to obtain a temperature-load fitting change curve, and calculate the characteristic values ​​of the cable temperature in each set of temperature-load fitting change curves. The characteristic values ​​include mean, mean square value, root mean square amplitude, standard deviation, peak value, peak-to-valley value, crest factor, shape factor, variance and skewness; A network construction module is used to construct a GRNN generalized regression neural network, and optimize the network parameters of the GRNN generalized regression neural network using an artificial bee colony algorithm to obtain an optimized GRNN generalized regression neural network; The training module is used to perform iterative training based on the optimized GRNN generalized regression neural network, taking the characteristic value of the cable temperature as the input and the cable load corresponding to the temperature data as the output, to obtain a trained GRNN generalized regression neural network; A load prediction module is used to input the cable temperature measured at a future time into the trained GRNN generalized regression neural network and output a corresponding cable load prediction value at a future time; The current carrying capacity prediction module is used to calculate the dynamic current carrying capacity of the three-core cable based on the cable load prediction value at a future moment and the preset maximum load limit of the three-core cable.

6. The multi-dimensional correlation three-core cable dynamic current carrying capacity prediction system according to claim 5, characterized in that: Also includes: The dimensionality reduction module is used to perform dimensionality reduction processing on the eigenvalues ​​using factor analysis, specifically including: Convert all the eigenvalues ​​to The order matrix is: Formula 1 In formula 1, Y is rank matrix, q is the number of eigenvalues, m is the number of eigenvalue samples, is the qth eigenvalue; Will Defined as the original variable, factor analysis is used to decompose each original variable to obtain: Formula 2 In formula 2, u q is the sample mean of the original variable, k qn is the factor loading of the original variable, n is the number of factors extracted from the original variable, n<q, is the common factor vector matrix, G n is the nth element in the common factor vector matrix, is a special factor vector matrix; Simplify Equation 2 to: Formula 3 Convert Equation 3 into a parameter sequence: Formula 4 In formula 4, is the factor loading matrix, where ； SPSS statistical software was used to solve Formula 4 to obtain the common factor vector, and a scree plot of the common factor vector sequence was generated based on the common factor vector. The slope between two adjacent common factor vectors is calculated according to the scree plot of the common factor vector sequence, and each common factor vector is traversed in sequence from left to right to determine whether the slope is greater than a preset slope threshold. If the slope is greater than the preset slope threshold, the traversed common factor vector is intercepted as the eigenvalue after dimensionality reduction.

7. The multi-dimensional correlation three-core cable dynamic current carrying capacity prediction system according to claim 5, characterized in that: The network construction module specifically includes: A network initialization module is used to construct a GRNN generalized regression neural network, wherein the GRNN generalized regression neural network includes an input layer, a pattern layer, a summation layer, and an output layer connected in sequence; The input layer is used to receive input and pass the input to the model layer; The transfer function of the mode layer is: Formula 5 In formula 5, is the output of neuron i, X is the input, is the training sample of the i-th neuron, is the smoothing factor, n is the number of neurons; The summation layer includes two types of neurons, wherein the first type of neurons sums the output values ​​of all neurons in the pattern layer, and the summation formula is: Formula 6 In formula 6, S D is the sum value, T is the transpose symbol; The second type of neurons assign weights to the output values ​​of all neurons in the pattern layer, thereby performing weighted summation. The weighted summation formula is: Formula 7 In formula 7, S Nj is the weighted sum value, y ij is the output sample y i The jth element of Y i is the i-th training sample, j is the number of elements, and k is the total number of elements; The number of neurons in the output layer is the dimension of the output vector. The output layer representation function is obtained by dividing the weighted sum of the neurons in each layer by the unweighted sum: Formula 8 In formula 8, y i is the i-th output; The network optimization module is used to optimize the smoothing factor using the artificial bee colony algorithm, specifically including: Randomly generate a population and initialize it. The initialization formula is: Formula 9 In Equation 9, s is the sth particle in the population, w is the dimension, is a random number between 0 and 1, is the location of the w-th dimension food source of the s-th solution in the population, are the maximum value and the minimum value of the food source location respectively; Hire bees to find food sources and update the location of the food sources to: Formula 10 In formula 10, is the location of the food source after the location update, a is a randomly generated value, It is a random number between -1 and 1; After the food source is mined, observe the location of the bees towards the food source The probability of selection is calculated as follows: Formula 11 Formula 12 In formulas 11 and 12, S is the number of particles, is the fitness function value of the food source, is the objective function of the food source, which is the error function between the calculated output of the GRNN generalized regression neural network and the training sample value under a specific smooth factor condition; Find the position of the observer bee relative to the food source The smoothing factor corresponding to the maximum probability is selected as the optimal smoothing factor, and the optimal smoothing factor is applied to the GRNN generalized neural network.

8. The multi-dimensional correlation three-core cable dynamic current carrying capacity prediction system according to claim 5, characterized in that: The current carrying capacity prediction module is specifically used to calculate the difference between the cable load prediction value at a future moment and the preset maximum load limit of the three-core cable, and the difference result is the dynamic current carrying capacity of the three-core cable.

Images