A method for calculating cable parameters taking into account the frequency and temperature characteristics of materials.

Abstract

This invention discloses a method for calculating cable parameters considering the frequency and temperature variations of materials. The steps include: 1. Measuring the temperature and frequency variations of the cable material in the laboratory; 2. Establishing the correspondence between the impedance and admittance of the cable material and temperature and frequency; 3. Setting up a neural network model, training the neural network using the measured data from step 1, and obtaining the correspondence between impedance and admittance and temperature and frequency from step 2; 4. Modeling the cable system and writing the voltage and current equations along the cable; 5. Using the neural network trained in step 3 to obtain the impedance and admittance of each cable material at the current temperature and frequency, and substituting them into the voltage and current equations to solve the equations; 6. For new cable structures, obtaining the impedance matrix and admittance matrix of the corresponding cable according to step 4. This invention comprehensively considers the influence of frequency and temperature on the series impedance and parallel admittance matrices of the cable, making the calculation results more realistic and possessing great engineering value.

Description

Technical Field

[0001] This invention relates to the field of cable parameter calculation technology, and specifically to a method for calculating cable parameters that takes into account the frequency and temperature characteristics of materials. Background Technology

[0002] In recent years, with the rapid development of China's economy, the urbanization process has also progressed rapidly. Consequently, the scale of urban power grids, an important indicator of urbanization, has expanded rapidly, and the usage of power cables, the main force for power transmission in urban power grids, has naturally increased significantly. Compared with overhead lines, power cables have advantages such as saving space, being less affected by natural disasters, having higher operating efficiency, and stronger safety. Currently, power cables have become the main medium for power transmission in urban power grids.

[0003] With the advancement of technology and the growth of electricity demand, the requirements for cable performance are becoming increasingly stringent. Research on cable parameter calculation methods, especially those considering the frequency and temperature variations of materials, is crucial for improving the accuracy of cable design and the stability of power systems. First, the material properties of cables change with frequency and temperature, directly affecting their electrical parameters such as resistance, inductance, and capacitance. Traditional cable parameter calculation methods often assume these parameters are constants, but in reality, they are variables. Therefore, researching cable parameter matrix calculation methods that consider the frequency and temperature variations of materials can more accurately reflect the performance of cables under different operating conditions. Second, as power systems develop towards higher frequencies and temperatures, the reliability of cables in high-frequency and high-temperature environments has become a focus of attention. For example, in flexible DC power grids, cables need to withstand wider frequency bands and higher temperatures. This requires cable parameter calculation methods that accurately consider the frequency and temperature variations of materials. Third, accurate cable parameter matrix calculation methods are also crucial for cable fault location and diagnosis. Accurate calculation of cable parameters allows for faster fault location, reduced power outage time, and improved power supply reliability.

[0004] Currently, for calculating voltage and current along cables, scholars and engineers both domestically and internationally primarily employ formula-based methods and simulation software. For detailed calculations involving temperature and frequency variations, simulation software such as EMTP-ATP and COMSOL are preferred for modeling and analysis. This method requires first establishing a simulation model of the cable's electric and magnetic fields, then iteratively solving the problem by setting temperatures and frequencies. Repeating the solution process for different temperatures and frequencies is cumbersome and demands significant computational power. As for formula-based methods, this involves deriving and calculating the cable's impedance and admittance matrices using formulas, and then solving the voltage and current equations to obtain the cable's operating state. In formula-based solutions, the cable's material properties are often replaced with parameters at room temperature and power frequency for rough calculations. There are few research examples, both domestically and internationally, of correcting the cable's distributed parameter matrix according to the material's temperature and frequency characteristics in formula-based methods. Summary of the Invention

[0005] To address the problems existing in the prior art, the purpose of this invention is to propose a method for calculating cable parameters that takes into account the frequency and temperature characteristics of materials. This method accurately calculates the series impedance and parallel admittance matrix of the cable by considering the temperature and frequency characteristics of the cable material, making the calculation results more consistent with reality.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for calculating cable parameters that takes into account the frequency and temperature variation characteristics of materials, comprising the following steps:

[0008] Step 1: Laboratory testing of the temperature and frequency characteristics of the cable material;

[0009] Step 2: Based on the measured data, taking into account the effects of temperature and frequency, establish the corresponding relationship between the impedance and admittance of the cable material and temperature and frequency: Z = f(T,ω), Y = f(T,ω);

[0010] Step 3: Set up the neural network model and train the neural network using the measured data from Step 1 to obtain the correspondence between impedance and admittance and temperature and frequency in Step 2.

[0011] Step 4: Model the cable system according to the cable structure and laying method; based on the coupling relationship between the electric field and the magnetic field, treat all cable materials as conductors, calculate the self-impedance and mutual impedance, self-admittance and mutual admittance of the cable, and write the voltage and current equations along the cable, i.e., the impedance matrix and admittance matrix.

[0012] Step 5: Use the neural network trained in Step 3 to obtain the impedance and admittance of each cable material at the current temperature and frequency, substitute them into the voltage and current equations, and then the voltage and current equations can be solved.

[0013] Step 6: For the new cable structure, obtain the impedance matrix and admittance matrix of the corresponding cable according to Step 4. During the calculation, the corresponding temperature and frequency, impedance and admittance of each material are obtained by relying on the neural network trained in Step 3, so as to accurately solve the voltage and current distribution along the cable.

[0014] Step 1 involves laboratory measurements of the temperature and frequency characteristics of cable materials, including: the impedance and admittance of the conductor, insulation layer, sheath, shielding layer, and filler material as a function of temperature and frequency; and the measurement of the impedance and admittance of the cable material as a function of frequency at a given temperature and the impedance and admittance per unit length of the cable material as a function of temperature at a given frequency.

[0015] In step 2, based on the measured data and taking into account the effects of temperature and frequency, a neural network is used to fit the impedance and admittance curves of each cable material, and to establish the corresponding relationship between the impedance and admittance of the cable material and temperature and frequency, Z = f(T,ω) and Y = f(T,ω).

[0016] Preferably, in step 3, the neural network model is set to an input layer, a hidden layer, and an output layer; the number of hidden layers is set to 1, and the number of neurons is set to 20; the activation function of the hidden layer is selected as ReLU, and no activation function is used for the output layer; the loss function is selected as MSE loss function.

[0017] Compared with existing technologies, this invention proposes a method for calculating cable parameter matrices that takes into account the frequency and temperature characteristics of materials. It comprehensively considers the influence of frequency and temperature on the series impedance and parallel admittance matrix of cables, making the calculation results more realistic and having great engineering value. Attached Figure Description

[0018] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0019] Figure 1 A flowchart is shown for a method of calculating the cable parameter matrix that takes into account the frequency and temperature characteristics of materials.

[0020] Figure 2 A schematic diagram of the neural network model training process is shown. Detailed Implementation

[0021] The purpose of this invention is to propose a method for calculating cable parameter matrices that takes into account the frequency and temperature characteristics of materials. This method accurately calculates the series impedance and parallel admittance matrices of cables by considering the temperature and frequency characteristics of cable materials, making the calculation results more consistent with reality.

[0022] like Figure 1As shown, the present invention provides a method for calculating cable parameters that takes into account the frequency and temperature variations of materials. The method includes the following steps:

[0023] 1. In the laboratory, the temperature and frequency characteristics of cable materials are measured, such as the impedance and admittance of conductors, insulation layers, sheaths, shielding layers, and filler materials as a function of temperature and frequency. The impedance and admittance of cable materials are measured at a given temperature as a function of frequency, and at a given frequency, the impedance and admittance per unit length of cable material are measured as a function of temperature, resulting in a dataset of impedance variations for various cable materials as a function of temperature and frequency.

[0024] 2. Based on measured data, and considering the effects of temperature and frequency, a neural network is used to fit the impedance and admittance curves for each cable material. The relationship between impedance and temperature and frequency is a bivariate function Z = f(T, ω), and the training objective of the neural network is to find the corresponding relationship f. The input to the neural network is the point set (T, ω) → Z, representing a certain impedance value corresponding to a certain temperature and frequency. Similarly, the admittance of the cable material corresponds to the relationship Y = f(T, ω).

[0025] 3. For each cable material, train a neural network to fit the relationship between its impedance and admittance and temperature and frequency. (Following...) Figure 2 For each cable material, the dataset obtained in step 1 is divided into training set X1 and test set X2 in a 7:3 ratio. A neural network model is set up with an input layer, hidden layers, and an output layer. The hidden layer has 1 layer and 20 neurons. The activation function for the hidden layer is ReLU, and no activation function is used for the output layer. The loss function is MSE loss, i.e., The neural network is trained using a training set and its predictive performance is validated using a test set. Training ends when the MSE (Mean Separation of Output Y′2) between the model and the true value Y2 is less than n*(1e-2), where n is the dimension of the test set X2. If the MSE between Y2 and Y′2 is greater than n*(1e-2), the model should be retrained by adjusting the number of iterations or increasing the training set X1. This yields the corresponding models for impedance, temperature, and frequency. For the admittance of each cable material, a corresponding neural network model is built using the same method. After all networks for various cable materials are trained, they are integrated to automatically calculate the impedance and admittance based on the input cable material type, temperature, and frequency.

[0026] 4. Based on the cable structure and laying method, model the cable system, considering the effects of frequency and temperature, and establish the cable's impedance and admittance matrices. For example, for a single-circuit single-core cable, from the inside out, the components are the conductor, main insulation, metallic sheath, sheath insulation, armor, and outer sheath. Treating all these structures as conductors, there are a total of 18 conductors coupling across the three phases. The relationship between voltage drop and current loss along the cable, i.e., the impedance matrix, can be written as:

[0027]

[0028] Among them, the self-impedance Z XX and mutual impedance Z XY The admittance matrix of the cable can be obtained from the coupling relationship between the electric and magnetic fields based on the cable's structure; these are functions of temperature T and frequency ω. Similarly, the admittance matrix of the cable can be written:

[0029]

[0030] For single-circuit cable systems with different arrangements, the self-impedance, mutual impedance, self-admittance, and mutual admittance should be calculated separately according to the actual laying conditions. For multi-circuit systems, in addition to calculating the self-impedance, mutual impedance, self-admittance, and mutual admittance according to the actual situation, the number of unknowns, voltage U and current I, should also be increased accordingly.

[0031] 5. Solve the impedance and admittance equations to obtain the voltage and current distribution along the cable. During the solution process, the neural network trained in step 3 is used. By inputting the type of cable material, the current temperature, and the frequency combination T and ω, the impedance and admittance of each material at the current location can be obtained. These values ​​are then substituted into the voltage and current equations to solve the problem. This utilizes the temperature and frequency variations of the cable material to obtain a more accurate distribution of voltage and current along the cable.

[0032] 6. For the new cable structure, the corresponding impedance matrix and admittance matrix are obtained according to step 4. During the calculation process, the corresponding temperature and frequency, impedance and admittance of each material are obtained by relying on the neural network in step 3. The voltage and current distribution along the cable can be accurately solved, thereby improving the accuracy of solving the cable voltage and current using the series impedance matrix and parallel admittance matrix.

Claims

1. A method for calculating cable parameters considering the frequency and temperature variations of materials, characterized in that: Includes the following steps: Step 1: The temperature and frequency characteristics of the cable material are measured in the laboratory to obtain a dataset of the impedance and admittance of the cable material as a function of temperature and frequency. The cable material includes conductor, insulation layer, sheath, shielding layer and filling material. The measurement includes measuring the change of impedance and admittance of the cable material with frequency at a certain temperature, and measuring the change of impedance and admittance per unit length of the cable material with temperature at a certain frequency. Step 2: Based on the measured data, take temperature into consideration. and frequency The influence of the impedance of the cable material and admittance With temperature and frequency Correspondence , And a neural network was used to fit the impedance and admittance curves of each cable material; Step 3: Train and fit the impedance and admittance with temperature for each cable material. ,frequency The neural network for the relationship divides the dataset into a training set and a test set in a 7:3 ratio. The training set is used to train the neural network, and the test set is used to verify the predictive performance of the neural network. Step 4: Model the cable system based on the cable structure and laying method; Based on the coupling relationship between electric and magnetic fields, all cable materials are considered as conductors. The self-impedance, mutual impedance, self-admittance, and mutual admittance of the cable are calculated. The voltage and current equations along the cable are written, and the impedance matrix and admittance matrix are obtained. Step 5: Solve the voltage and current equations along the cable to obtain the voltage and current distributed along the cable. During the solution process, the neural network trained in Step 3 is used, with the cable material type and current temperature as input. and frequency To obtain the impedance of the cable material at the current temperature and frequency. and admittance Substitute the impedance matrix and admittance matrix into the given values ​​to solve the problem; Step 6: For the new cable structure, obtain the impedance matrix and admittance matrix of the corresponding cable according to Step 4, and accurately solve the voltage and current distribution along the cable according to Step S5.

2. The method for calculating cable parameters considering the frequency and temperature variations of materials according to claim 1, characterized in that: In step 3, when the mean square error (MSE) between the output obtained from the model input test set and the true value is less than 100%, the mean square error (MSE) is calculated. The training ended at that time, among which The dimension of the test set; when the MSE is greater than If necessary, retraining can be performed by adjusting the number of model iterations or increasing the training set.

3. The method for calculating cable parameters considering the frequency and temperature characteristics of materials according to claim 1, characterized in that: In step 3, the neural network model is set to an input layer, a hidden layer, and an output layer; the number of hidden layers is set to 1, and the number of neurons is set to 20; the activation function for the hidden layer is ReLU, and no activation function is used for the output layer; the loss function is MSE loss function.

Images