Method and device for evaluating harmonic loss of cable
By optimizing the neural network model and using historical cable data for training and updating, the problem of low accuracy in cable harmonic loss assessment was solved, achieving higher assessment accuracy.
Patent Information
- Application Number
- CN202510668479.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-23
- Publication Date
- 2025-09-23
AI Technical Summary
The existing cable harmonic loss assessment method ignores the temperature, aging and environmental factors changes of the cable during operation, resulting in low assessment accuracy.
A neural network model is used to obtain historical data samples of the cable, divide the data sets into training and test data sets, use back-propagation neural network for training and updating, and optimize the number of neural network layers to improve evaluation accuracy.
The influence of cable operating status and environmental factors is effectively considered, which significantly improves the accuracy of harmonic loss assessment and avoids excessive dependence of the neural network model on training data.
Smart Images
Figure CN120687931A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of power systems, and in particular to a method and device for evaluating cable harmonic loss. Background Art
[0002] In recent years, cables have been widely used in power grids. The large-scale integration of renewable energy generation and the growing number of power electronic loads such as charging stations and electrified railways have led to a further increase in harmonics in power grid cables. Harmonics can cause cable temperature rise, aging, insulation defects, and even failure. Cable losses caused by harmonic currents are closely related to cable temperature rise and aging, necessitating an assessment of cable harmonic losses.
[0003] Harmonic loss assessments provided by related technologies primarily use parameters such as cable impedance and conductor diameter as inputs, using formulas to calculate cable loss at different harmonic orders. However, during operation, cables are affected by environmental factors such as temperature, aging, the soil in which the cables are laid, and mechanical stress, causing the cable's impedance parameters to change over time. Consequently, the cable harmonic loss calculated using these formulas differs significantly from the actual cable harmonic loss, resulting in low accuracy in cable harmonic loss assessments. Summary of the Invention
[0004] In order to solve the problems in the prior art, the present application provides a method and device for evaluating cable harmonic loss.
[0005] In a first aspect, the present application provides a method for evaluating cable harmonic loss, which may include: obtaining historical data samples of the cable, dividing the historical data samples into a training data sample set and a test data sample set. Solving a pre-constructed neural network model based on the test data sample set to obtain the harmonic loss of the cable. The neural network model is constructed based on the training data sample set. Calculating the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set, and evaluating the cable harmonic loss based on the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set.
[0006] Optionally, the historical data samples include the fundamental current amplitude, each harmonic current amplitude, and the harmonic loss of the cable. Therefore, it is conceivable that both the training data sample set and the test data sample set may include the fundamental current amplitude, each harmonic current amplitude, and the harmonic loss of the cable.
[0007] In some possible implementations, the construction of the neural network model includes: taking the fundamental current amplitude and each harmonic current amplitude in the training data sample set as input, and taking the harmonic loss of the cable as output, and using the Levenberg-Marquardt method to train the back propagation neural network to obtain the neural network model.
[0008] In other possible implementations, a pre-built neural network model is solved based on a test data sample set to obtain the harmonic loss of the cable, including: inputting the fundamental current amplitude and each harmonic current amplitude in the test data sample set into the neural network model, and outputting the harmonic loss of the cable.
[0009] For example, the root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set satisfies:
[0010]
[0011] Among them, RMSE represents the root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set, N represents the number of test data samples, m i represents the harmonic loss of the i-th test data sample in the test data sample set, p i represents the harmonic loss of the cable in the i-th test data sample output by the first neural network model.
[0012] In some further possible implementations, the cable harmonic loss is evaluated based on a root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set, including:
[0013] If the root mean square error between the harmonic loss of the cable output by the neural network model and the loss of the cable in the test data sample set is less than or equal to a preset error threshold, the harmonic loss of the cable output by the neural network model is used as the evaluation result.
[0014] If the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set is greater than a preset error threshold, the neural network model is updated and the harmonic loss of the cable is evaluated based on the updated neural network model.
[0015] Furthermore, the neural network model is updated, including:
[0016] The number of neural network layers of the back-propagation neural network is increased. The fundamental current amplitude and each harmonic current amplitude in the training data sample set are used as input, and the harmonic loss of the cable is used as output. The back-propagation neural network is retrained using the Levenberg-Marquardt method to obtain an updated neural network model.
[0017] Optionally, the cable harmonic loss is evaluated based on the updated neural network model, including:
[0018] The fundamental current amplitude and each harmonic current amplitude in the test data sample set are input into the updated neural network model to output the cable's harmonic loss. The root mean square error (RMS) between the harmonic loss output by the updated neural network model and the cable loss in the test data sample set is calculated. The cable's harmonic loss is evaluated based on the RMS error between the harmonic loss output by the updated neural network model and the cable loss in the test data sample set.
[0019] Furthermore, the harmonic loss of the cable is evaluated based on the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set, including:
[0020] If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, the harmonic loss of the cable output by the updated neural network model is used as the evaluation result.
[0021] If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is greater than the preset error threshold, the number of neural network layers of the back propagation neural network will continue to be increased, the neural network model will continue to be updated, and the harmonic loss of the cable will be evaluated based on the updated neural network model until the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, and the evaluation result is obtained.
[0022] In a second aspect, the present application provides a device for evaluating cable harmonic loss, comprising:
[0023] The acquisition module is used to obtain historical data samples of the cable and divide the historical data samples into a training data sample set and a test data sample set.
[0024] The solving module is used to solve the pre-built neural network model based on the test data sample set to obtain the harmonic loss of the cable. The neural network model is built based on the training data sample set.
[0025] The evaluation module is used to calculate the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set, and evaluate the cable harmonic loss based on the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set.
[0026] Optionally, the historical data samples include the fundamental current amplitude flowing through the cable, the amplitudes of each harmonic current, and the harmonic loss of the cable.
[0027] In some possible implementations, the evaluation device also includes a modeling module, which is used to: use the fundamental current amplitude and each harmonic current amplitude in the training data sample set as input, and the harmonic loss of the cable as output, and use the Levenberg-Marquardt method to train the back propagation neural network to obtain a neural network model.
[0028] In some other possible implementations, the solution module is specifically used to: input the fundamental current amplitude and each harmonic current amplitude in the test data sample set into the neural network model, and output the harmonic loss of the cable.
[0029] Exemplarily, the evaluation module specifically calculates the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set according to the following formula:
[0030]
[0031] Among them, RMSE represents the root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set, N represents the number of test data samples, m i represents the harmonic loss of the i-th test data sample in the test data sample set, p i represents the harmonic loss of the cable in the i-th test data sample output by the first neural network model.
[0032] Optionally, the evaluation module is specifically used to:
[0033] If the root mean square error between the harmonic loss of the cable output by the neural network model and the loss of the cable in the test data sample set is less than or equal to a preset error threshold, the harmonic loss of the cable output by the neural network model is used as the evaluation result.
[0034] If the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set is greater than a preset error threshold, the neural network model is updated and the harmonic loss of the cable is evaluated based on the updated neural network model.
[0035] Furthermore, the evaluation module is specifically configured to increase the number of neural network layers in the back-propagation neural network, use the fundamental current amplitude and each harmonic current amplitude in the training data sample set as input, and use the cable harmonic loss as output. The back-propagation neural network is retrained using the Levenberg-Marquardt method to obtain an updated neural network model. As can be seen, the evaluation module can update the neural network model according to the above process.
[0036] Exemplarily, the evaluation module is specifically configured to: input the fundamental current amplitude and each harmonic current amplitude in the test data sample set into an updated neural network model, and output the harmonic loss of the cable. Calculate the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set. Evaluate the harmonic loss of the cable based on the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set. It can be seen that the evaluation module can evaluate the harmonic loss of the cable based on the updated neural network model according to the above process.
[0037] Furthermore, the evaluation module is specifically used to:
[0038] If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, the harmonic loss of the cable output by the updated neural network model is used as the evaluation result.
[0039] If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is greater than the preset error threshold, the number of neural network layers of the back propagation neural network will continue to be increased, the neural network model will continue to be updated, and the harmonic loss of the cable will be evaluated based on the updated neural network model until the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, and the evaluation result is obtained.
[0040] It can be seen that the evaluation module can evaluate the harmonic loss of the cable according to the above process based on the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set.
[0041] On the other hand, the present application also provides a computer device, including: one or more processors.
[0042] A processor is used to execute one or more programs.
[0043] When one or more programs are executed by one or more processors, the above-described evaluation method is implemented.
[0044] In another aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the above-mentioned evaluation method.
[0045] Compared with the prior art, the present invention has the following advantages:
[0046] In the cable harmonic loss evaluation method provided by the present application, historical data samples of the cable are obtained, and the historical data samples are divided into a training data sample set and a test data sample set. According to the test data sample set, a pre-constructed neural network model (constructed according to the training data sample set) is solved to obtain the harmonic loss of the cable. The root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set is calculated, and the harmonic loss of the cable is evaluated based on the root mean square error. It can be seen that the present application can effectively take into account the influence of the operating status of the cable and environmental factors on the harmonic loss, and can greatly improve the accuracy of the evaluation.
[0047] This method continuously adjusts the number of neural network layers according to the difference in root mean square error under different numbers of neural network layers to update the neural network model, thereby avoiding excessive dependence of the neural network model on the training data sample set caused by improper setting of the number of neural network layers and achieving optimization of the neural network model. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0049] Figure 1 This is a schematic flow chart of a method for evaluating cable harmonic loss in an embodiment of the present application;
[0050] Figure 2 Schematic diagram of harmonic loss of cables in an embodiment of the present application;
[0051] Figure 3 This is a schematic diagram comparing the harmonic loss of the cable output by the neural network model in an embodiment of the present application with the actual harmonic loss of the cable;
[0052] Figure 4 This is a schematic structural diagram of a device for evaluating cable harmonic loss in an embodiment of the present application. DETAILED DESCRIPTION
[0053] The technical solution in this application will be described below with reference to the accompanying drawings.
[0054] The terms "first," "second," and the like in the description, embodiments, claims, and drawings of this application are used solely for descriptive purposes and are not to be construed as indicating or implying relative importance or order. Furthermore, the terms "including," "having," and any variations thereof are intended to cover non-exclusive inclusions, such as, for example, inclusion of a series of steps or units. A method, system, product, or apparatus is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0055] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.
[0056] Example 1:
[0057] The embodiment of the present application provides a method for evaluating cable harmonic loss. Figure 1 As shown, the evaluation method 100 may include the following steps:
[0058] Step S1: Obtain historical data samples of the cable, and divide the historical data samples into a training data sample set and a test data sample set according to a preset division ratio (such as 7:3).
[0059] Step S2: Solve the pre-built neural network model based on the test data sample set to obtain the harmonic loss of the cable. The neural network model is built based on the training data sample set.
[0060] Step S3: Calculate the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set, and evaluate the cable harmonic loss based on the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set.
[0061] Optionally, the historical data samples in step S1 include the fundamental current amplitude, each harmonic current amplitude, and the harmonic loss of the cable. Therefore, it is conceivable that both the training data sample set and the test data sample set may include the fundamental current amplitude, each harmonic current amplitude, and the harmonic loss of the cable.
[0062] In some possible implementations, the construction of the neural network model in step S2 includes: taking the fundamental current amplitude and each harmonic current amplitude (which can be 2 to 50 times) in the training data sample set as input, and taking the harmonic loss of the cable as output, and using the Levenberg-Marquardt method (which can be abbreviated as LM method) to train the back propagation neural network (which can be called BP neural network) to obtain the neural network model.
[0063] In other possible implementations, in step S2, a pre-constructed neural network model is solved according to the test data sample set to obtain the harmonic loss of the cable, including: inputting the fundamental current amplitude and each harmonic current amplitude in the test data sample set into the neural network model, and outputting the harmonic loss of the cable.
[0064] Exemplarily, the root mean square error between the cable harmonic loss output by the neural network model in step S3 and the loss of the cable in the test data sample set satisfies:
[0065]
[0066] Among them, RMSE represents the root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set, N represents the number of test data samples, m i represents the harmonic loss of the i-th test data sample in the test data sample set, p i It represents the harmonic loss of the cable in the i-th test data sample output by the first neural network model. It can be understood that the smaller the root mean square error, the better the evaluation effect.
[0067] In some further possible implementations, step S3 evaluates the harmonic loss of the cable based on a root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set, including:
[0068] If the root mean square error between the harmonic loss of the cable output by the neural network model and the loss of the cable in the test data sample set is less than or equal to a preset error threshold, the harmonic loss of the cable output by the neural network model is used as the evaluation result.
[0069] If the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set is greater than a preset error threshold, the neural network model is updated and the harmonic loss of the cable is evaluated based on the updated neural network model.
[0070] Furthermore, the specific process of updating the neural network model may include:
[0071] The number of neural network layers of the back-propagation neural network is increased. The fundamental current amplitude and each harmonic current amplitude in the training data sample set are used as input, and the harmonic loss of the cable is used as output. The back-propagation neural network is retrained using the Levenberg-Marquardt method to obtain an updated neural network model.
[0072] Optionally, the evaluation of the harmonic loss of the cable based on the updated neural network model may include:
[0073] The fundamental current amplitude and each harmonic current amplitude in the test data sample set are input into the updated neural network model to output the harmonic loss of the cable.
[0074] Calculate the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set.
[0075] The cable harmonic loss is evaluated based on the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set. Specifically, it includes:
[0076] If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, the harmonic loss of the cable output by the updated neural network model is used as the evaluation result.
[0077] If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is greater than the preset error threshold, the number of neural network layers of the back propagation neural network will continue to be increased, the neural network model will continue to be updated, and the harmonic loss of the cable will be evaluated based on the updated neural network model until the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, and the evaluation result is obtained.
[0078] In order to verify the effectiveness of the evaluation method provided in the embodiment of the present application, the evaluation of cable harmonic loss is realized through finite element simulation. The model of the cable is YJV-8.7 / 15-3×150, and the frequency range is ultra-high harmonics between 2kHz and 150kHz. The simulation gives multiple sets of cable power losses caused by harmonic current frequencies and amplitudes. The historical data samples can be divided into training data sample sets and test data sample sets. The training data sample set is mainly for training parameters in the BP neural network. During training, the fundamental current amplitude and the amplitude of each harmonic current are used as input, and the harmonic loss of the cable is used as output. After obtaining the neural network model, the test data sample set is used to test the effectiveness of the neural network model in harmonic loss evaluation.
[0079] Figure 2 The graph shows the change curve of harmonic loss under different single harmonic current frequency and amplitude conditions obtained in the finite element simulation. Figure 3 The comparison between the harmonic loss of the cable output by the neural network model and the actual harmonic loss of the cable is shown. It can be seen that the differences between different test data samples are obvious. The evaluation method provided by the embodiment of the application is highly accurate and can fully reflect the differences between different test data samples. The actual error between each test data sample is small.
[0080] Example 2:
[0081] Based on the same inventive concept, the embodiment of the present application also provides a device for evaluating cable harmonic loss. Figure 4 As shown, the evaluation device 200 may include:
[0082] The acquisition module 201 is used to acquire historical data samples of the cable and divide the historical data samples into a training data sample set and a test data sample set.
[0083] The solving module 202 is used to solve the pre-built neural network model based on the test data sample set to obtain the harmonic loss of the cable. The neural network model is built based on the training data sample set.
[0084] The evaluation module 203 is used to calculate the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set, and evaluate the harmonic loss of the cable based on the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set.
[0085] Optionally, the historical data samples acquired by the acquisition module 201 include the fundamental current amplitude flowing through the cable, the amplitudes of each harmonic current, and the harmonic loss of the cable. Therefore, it is conceivable that both the training data sample set and the test data sample set may include the fundamental current amplitude flowing through the cable, the amplitudes of each harmonic current, and the harmonic loss of the cable.
[0086] In some possible implementations, the evaluation device 200 further includes a modeling module 204 .
[0087] The modeling module 204 is used to: use the fundamental current amplitude and each harmonic current amplitude in the training data sample set as input, and the harmonic loss of the cable as output, and use the Levenberg-Marquardt method to train the back propagation neural network to obtain a neural network model.
[0088] In some other possible implementations, the solution module 202 is specifically configured to input the fundamental current amplitude and each harmonic current amplitude in the test data sample set into the neural network model, and output the harmonic loss of the cable.
[0089] Exemplarily, the evaluation module 203 specifically calculates the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set according to the following formula:
[0090]
[0091] Among them, RMSE represents the root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set, N represents the number of test data samples, m i represents the harmonic loss of the i-th test data sample in the test data sample set, p i represents the harmonic loss of the cable in the i-th test data sample output by the first neural network model.
[0092] Optionally, the evaluation module 203 is specifically configured to:
[0093] If the root mean square error between the harmonic loss of the cable output by the neural network model and the loss of the cable in the test data sample set is less than or equal to a preset error threshold, the harmonic loss of the cable output by the neural network model is used as the evaluation result.
[0094] If the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set is greater than a preset error threshold, the neural network model is updated and the harmonic loss of the cable is evaluated based on the updated neural network model.
[0095] Furthermore, evaluation module 203 is specifically configured to increase the number of neural network layers in the back-propagation neural network, use the fundamental current amplitude and each harmonic current amplitude in the training data sample set as input, and use the cable harmonic loss as output, and retrain the back-propagation neural network using the Levenberg-Marquardt method to obtain an updated neural network model. It can be seen that evaluation module 203 can update the neural network model according to the above process.
[0096] Exemplarily, the evaluation module 203 is specifically configured to: input the fundamental current amplitude and each harmonic current amplitude in the test data sample set into the updated neural network model to output the harmonic loss of the cable. Calculate the root mean square error (RMSE) between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set. Evaluate the harmonic loss of the cable based on the RMS error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set. It can be seen that the evaluation module 203 can evaluate the harmonic loss of the cable based on the updated neural network model according to the above process.
[0097] Furthermore, the evaluation module 203 is specifically configured to:
[0098] If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, the harmonic loss of the cable output by the updated neural network model is used as the evaluation result.
[0099] If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is greater than the preset error threshold, the number of neural network layers of the back propagation neural network will continue to be increased, the neural network model will continue to be updated, and the harmonic loss of the cable will be evaluated based on the updated neural network model until the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, and the evaluation result is obtained.
[0100] It can be seen that the evaluation module 203 can evaluate the harmonic loss of the cable according to the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set according to the above process.
[0101] Example 3:
[0102] Based on the same inventive concept, an embodiment of the present application further provides a computer device, comprising a processor and a memory, the memory being used to store a computer program, the computer program comprising program instructions, and the processor being used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, and are suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a computer storage medium to implement a corresponding method flow or corresponding function, so as to implement the steps of the evaluation method provided in the above embodiment.
[0103] Example 4:
[0104] Based on the same inventive concept, an embodiment of the present application further provides a computer-readable storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium herein may include both a built-in storage medium in a computer device and, of course, an extended storage medium supported by the computer device. The computer-readable storage medium provides a storage space that stores the operating system of the terminal. Furthermore, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions may be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium herein may be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor may load and execute one or more instructions stored in the computer-readable storage medium to implement the steps of the evaluation method provided in the above embodiment.
[0105] Those skilled in the art will appreciate that embodiments of the application may be provided as methods, systems, or computer program products. Thus, the application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0106] The application is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0107] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.
[0108] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.
[0109] The above are merely embodiments of the application and are not intended to limit the application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the application are included in the scope of the claims of the pending application.
Claims
1. A method for evaluating cable harmonic loss, characterized in that: include: Acquire historical data samples of the cable, and divide the historical data samples into a training data sample set and a test data sample set; Solving a pre-built neural network model based on the test data sample set to obtain the harmonic loss of the cable; wherein the neural network model is built based on the training data sample set; Calculate the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set, and evaluate the harmonic loss of the cable based on the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set.
2. The evaluation method according to claim 1, wherein: The historical data samples include the fundamental current amplitude flowing through the cable, the amplitudes of each harmonic current, and the harmonic loss of the cable.
3. The evaluation method according to claim 2, wherein: The construction of the neural network model includes: The fundamental current amplitude and each harmonic current amplitude in the training data sample set are used as input, and the harmonic loss of the cable is used as output. The back propagation neural network is trained using the Levenberg-Marquardt method to obtain the neural network model.
4. The evaluation method according to claim 2, wherein: Solving a pre-built neural network model based on a test data sample set to obtain harmonic loss of the cable includes: The fundamental current amplitude and each harmonic current amplitude in the test data sample set are input into the neural network model, and the harmonic loss of the cable is output.
5. The evaluation method according to claim 1, wherein: The root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set satisfies: Wherein, RMSE represents the root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set, N represents the number of test data samples, m i represents the harmonic loss of the i-th test data sample in the test data sample set, p i represents the harmonic loss of the cable in the i-th test data sample output by the first neural network model.
6. The evaluation method according to claim 2, wherein: The evaluating the harmonic loss of the cable according to the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set includes: If the root mean square error between the harmonic loss of the cable output by the neural network model and the loss of the cable in the test data sample set is less than or equal to a preset error threshold, the harmonic loss of the cable output by the neural network model is used as the evaluation result; If the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set is greater than a preset error threshold, the neural network model is updated and the harmonic loss of the cable is evaluated based on the updated neural network model.
7. The evaluation method according to claim 6, characterized in that The updating of the neural network model comprises: The number of neural network layers of the back-propagation neural network is increased, the fundamental current amplitude and each harmonic current amplitude in the training data sample set are used as input, and the harmonic loss of the cable is used as output, and the back-propagation neural network is retrained using the Levenberg-Marquardt method to obtain an updated neural network model.
8. The evaluation method according to claim 6, wherein: The evaluating the harmonic loss of the cable according to the updated neural network model includes: Inputting the fundamental current amplitude and each harmonic current amplitude in the test data sample set into the updated neural network model, and outputting the harmonic loss of the cable; Calculating the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set; The harmonic loss of the cable is evaluated based on the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set.
9. The evaluation method according to claim 8, characterized in that The evaluating the harmonic loss of the cable according to the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set includes: If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to a preset error threshold, the harmonic loss of the cable output by the updated neural network model is used as the evaluation result; If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is greater than a preset error threshold, the number of neural network layers of the back propagation neural network continues to be increased, the neural network model continues to be updated, and the harmonic loss of the cable is evaluated according to the updated neural network model until the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, thereby obtaining the evaluation result.
10. A device for evaluating cable harmonic loss, characterized in that: include: An acquisition module is used to acquire historical data samples of the cable and divide the historical data samples into a training data sample set and a test data sample set; A solution module, configured to solve a pre-built neural network model based on the test data sample set to obtain the harmonic loss of the cable; wherein the neural network model is built based on the training data sample set; An evaluation module is used to calculate the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set, and to evaluate the harmonic loss of the cable based on the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set.
11. The evaluation device according to claim 10, characterized in that The historical data samples include the fundamental current amplitude flowing through the cable, the amplitudes of each harmonic current, and the harmonic loss of the cable.
12. The evaluation device according to claim 11, characterized in that The evaluation device further comprises a modeling module, wherein the modeling module is configured to: The fundamental current amplitude and each harmonic current amplitude in the training data sample set are used as input, and the harmonic loss of the cable is used as output. The back propagation neural network is trained using the Levenberg-Marquardt method to obtain the neural network model.
13. The evaluation device according to claim 11, characterized in that The solution module is specifically used for: The fundamental current amplitude and each harmonic current amplitude in the test data sample set are input into the neural network model, and the harmonic loss of the cable is output.
14. The evaluation device according to claim 10, characterized in that The evaluation module specifically calculates the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set according to the following formula: Wherein, RMSE represents the root mean square error between the cable harmonic loss output by the neural network model and the cable loss in the test data sample set, N represents the number of test data samples, m i represents the harmonic loss of the i-th test data sample in the test data sample set, p i represents the harmonic loss of the cable in the i-th test data sample output by the first neural network model.
15. The evaluation device according to claim 11, characterized in that The evaluation module is specifically used for: If the root mean square error between the harmonic loss of the cable output by the neural network model and the loss of the cable in the test data sample set is less than or equal to a preset error threshold, the harmonic loss of the cable output by the neural network model is used as the evaluation result; If the root mean square error between the cable harmonic loss output by the neural network model and the loss of the cable in the test data sample set is greater than a preset error threshold, the neural network model is updated and the harmonic loss of the cable is evaluated based on the updated neural network model.
16. The evaluation device according to claim 15, characterized in that The evaluation module is specifically used for: The number of neural network layers of the back-propagation neural network is increased, the fundamental current amplitude and each harmonic current amplitude in the training data sample set are used as input, and the harmonic loss of the cable is used as output, and the back-propagation neural network is retrained using the Levenberg-Marquardt method to obtain an updated neural network model.
17. The evaluation device according to claim 15, characterized in that The evaluation module is specifically used for: Inputting the fundamental current amplitude and each harmonic current amplitude in the test data sample set into the updated neural network model, and outputting the harmonic loss of the cable; Calculating the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set; The harmonic loss of the cable is evaluated based on the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set.
18. The evaluation device according to claim 17, characterized in that The evaluation module is specifically used for: If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to a preset error threshold, the harmonic loss of the cable output by the updated neural network model is used as the evaluation result; If the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is greater than a preset error threshold, the number of neural network layers of the back propagation neural network continues to be increased, the neural network model continues to be updated, and the harmonic loss of the cable is evaluated according to the updated neural network model until the root mean square error between the harmonic loss of the cable output by the updated neural network model and the loss of the cable in the test data sample set is less than or equal to the preset error threshold, thereby obtaining the evaluation result.
19. A computer device, characterized in that: include: one or more processors; The processor is configured to store one or more programs; When the one or more programs are executed by the one or more processors, the evaluation method according to any one of claims 1 to 9 is implemented.
20. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed, the evaluation method according to any one of claims 1 to 9 is implemented.