Correction method for induced electricity calculation

Abstract

The invention relates to a correction method for induced electricity calculation, and relates to the field of neural networks. The method comprises the following steps: acquiring a plurality of original to-be-corrected parameters, and performing sensitivity analysis and screening on the original to-be-corrected parameters to obtain screened to-be-corrected parameters; and obtaining an induced electricity test result and inputting the result into the trained neural network model to obtain a design parameter value of each screened correction parameter. According to the method, the multiple original to-be-corrected parameters are obtained and subjected to sensitivity degree analysis, the key parameters are screened out, the induced electricity test result is corrected in combination with the neural network model, errors caused by initial value deviation or parameter inaccuracy in a traditional method can be effectively reduced, and the accuracy of induced electricity calculation is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of neural networks, and particularly to a correction method for induced electricity calculation. Background Art

[0002] The high-voltage power grid in China has developed vigorously, resulting in dense transmission line corridors. This leads to a huge induced electromagnetic field generated by the transmission lines. On the one hand, the transmission lines can generate a large induced voltage and induced current on the lines running in parallel below them. On the other hand, when one circuit of a multi-circuit transmission line on the same tower is in operation and the other circuit is out of service for maintenance, the operating line will generate an induced voltage and induced current on the maintenance line. However, the operation safety regulations issued by the power grid company remain at qualitative analysis, and the regulations are relatively general. Although many scholars have realized the harm of induced electricity and carried out simulation analysis on it, the accuracy of the simulation model is insufficient and there are relatively large defects. Generally speaking, the initial values of the induced electricity calculation of the transmission lines, such as the operating voltage and operating current, will have a greater impact on the calculation of the induced electricity. Usually, it is difficult for researchers to manually measure the operating voltage and load current, resulting in certain deviations when substituting the initial values. Summary of the Invention

[0003] In view of this, the present invention aims to propose a correction method for induced electricity calculation to solve the problem of large deviation in the induced electricity calculation during the simulation process in the prior art.

[0004] To achieve the above object, the technical solution of the present invention is realized as follows:

[0005] The first aspect of the present invention provides a correction method for induced electricity calculation, including:

[0006] Obtain a plurality of original parameters to be corrected, perform sensitivity analysis on each original parameter to be corrected and screen them to obtain the screened parameters to be corrected;

[0007] Obtain the induced electricity test results and input them into the trained neural network model to obtain the design parameter values of each screened correction parameter.

[0008] Further, the plurality of original parameters to be corrected include:

[0009] Test parameters, atmospheric pressure, environmental humidity, environmental wind speed, environmental wind direction, operating voltage, operating current, frequency fluctuation, line sag, line suspension height, phase spacing, pole spacing, soil resistivity, and grass and tree influence coefficient.

[0010] Further, the sensitivity analysis of each original parameter to be corrected includes:

[0011] Perform sensitivity analysis on each original parameter to be corrected based on the principle of induced electricity.

[0012] Furthermore, the parameters to be corrected after screening include:

[0013] Operating voltage, operating current, line suspension height, phase spacing, and dielectric constant.

[0014] Furthermore, the method for constructing the neural network model includes:

[0015] Construct a BP neural network model, and based on the PSO algorithm, adjust the initial weights and thresholds in the BP neural network model.

[0016] Furthermore, the activation functions of the hidden layer and the output layer of the BP neural network model are the Sigmoid function and the Tanh function, respectively.

[0017] Furthermore, the method for training the neural network model includes:

[0018] Obtain sample data;

[0019] Divide the sample data into a training set and a test set;

[0020] Input the training set into the neural network model to obtain a trained neural network model.

[0021] Furthermore, the sample data includes induced voltage sample data, induced current sample data, operating voltage sample data, operating current sample data, line suspension height sample data, environmental parameters, and phase spacing sample data.

[0022] Furthermore, after obtaining the sample data, the method further includes:

[0023] Select sample points for the parameters to be corrected after screening in the sample data through the orthogonal design method and the uniform design method.

[0024] The second aspect of the present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the method according to any one of the first aspect when executing the program.

[0025] Compared with the prior art, the present invention has the following advantages:

[0026] By obtaining multiple original parameters to be corrected and analyzing their sensitivity, the present invention screens out key parameters, and combines a neural network model to correct the induced electricity test results, which can effectively reduce the errors caused by initial value deviation or inaccurate parameters in the traditional method, and significantly improve the accuracy of induced electricity calculation.

[0027] By performing a sensitivity analysis on multiple original parameters to be corrected, key parameters that have a greater impact on the calculation results (such as operating voltage, operating current, etc.) are screened out, avoiding the interference of redundant parameters, reducing the computational complexity, and improving the computational efficiency.

[0028] In traditional methods, a large number of initial value parameters need to be manually measured and input, while in this method, the neural network model automatically processes the induced electricity test results and outputs the corrected parameter values, reducing manual intervention and the possibility of human error.

[0029] This method corrects the induced electricity test results through a neural network model, can adapt to the computational requirements under different environments and conditions, has strong applicability and flexibility, and can be widely applied in fields such as power systems and engineering design.

[0030] The corrected parameter values output by the neural network model provide more accurate input data for subsequent induced electricity calculations, further improving the accuracy of the overall calculation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0032] Figure 1 is a flowchart of the correction method for induced electricity calculation of the present invention;

[0033] Figure 2 is a schematic structural diagram of the neural network model of the present invention;

[0034] Figure 3 is an overall implementation flowchart of the present invention;

[0035] Figure 4 is a schematic structural diagram of the computer system of the device provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0037] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", "inner", "back", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0038] In addition, in the description of the present invention, unless otherwise clearly defined, the terms "installation", "connection", "connection", and "connector" shall be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood in combination with specific situations.

[0039] Next, the present invention will be described in detail with reference to the Figures 1 to 4 accompanying drawings and in conjunction with embodiments.

[0040] Overall, the design concept of the present invention is as follows;

[0041] Since there are usually certain errors in the electromagnetic simulation of transmission lines using the finite element method, the initial values such as the operating voltage and current of the transmission line are uncertain, so usually only approximate values can be obtained, and there may also be certain deviations in the measurement of the line suspension height, which makes the simulation results have certain errors.

[0042] When performing finite element method simulation, due to many non-linear effects among the finite element initial value parameters, a neural network can be used to correct its model. The model correction of the finite element method can be summarized into the following optimization problem:

[0043]

[0044] Among them, p is the design parameter (such as operating voltage, operating current, dielectric constant, line suspension height, and phase spacing); {f A (p)} is the eigenvalue of the simulation, and {f E} is the eigenvalue of the test; R(p) is the eigenvalue residual; V UB is the upper limit of the design space, and V LB is the lower limit of the design space.

[0045] The model correction method such as the design parameter type is a method for correcting design parameters and belongs to the inverse problem of finite element unit deduction. The correction based on the inverse problem performs multiple high-dimensional and non-linear iterative solutions at the design points of the finite element model, which is time-consuming and laborious, and is prone to falling into local optimal solutions in the case of large errors. Therefore, to avoid a large number of iterations and complex non-linear optimization calculations, the present disclosure takes the design parameter as the dependent variable and the characteristic quantity as the independent variable, solves the mapping function, and uses the neural network to transform the model correction problem into a forward problem for solution. The core of the neural network method is to represent the design parameter p and the characteristic quantity y in the form of an inverse function as:

[0046] p = f -1 (y)

[0047] The mapping relationship f is obtained by fitting using a neural network -1 , the actual response is input into the trained network, and the design parameter p is directly output. The basic steps of the finite element model correction method based on the Backpropagation Neural Network (BP) of the Particle Swarm Optimization (PSO) algorithm are as follows:

[0048] Embodiment 1 of the present invention provides a correction method for induced electricity calculation, as Figure 1 shown, including:

[0049] Step S1: Obtain multiple original parameters to be corrected, perform sensitivity analysis on each original parameter to be corrected and screen them to obtain the screened parameters to be corrected;

[0050] Step S2: Obtain the induced electricity test results and input them into the trained neural network model to obtain the design parameter values of each screened correction parameter.

[0051] In one embodiment, according to Step S1, it can be seen that the present invention first needs to select the parameters to be corrected. Sensitivity analysis is performed among many design parameters to select the parameters to be corrected. Among them, the multiple original parameters to be corrected include:

[0052] Test parameters, atmospheric pressure, environmental humidity, environmental wind speed, environmental wind direction, operating voltage, operating current, frequency fluctuation, line sag, line suspension height, phase spacing, pole spacing, soil resistivity, and grass and wood influence coefficient, etc.

[0053] The sensitivity analysis of each original parameter to be corrected includes:

[0054] Perform sensitivity analysis on each original parameter to be corrected based on the principle of induced electricity. The essence of model correction is to reverse infer the simulation initial value based on the measured data, so as to correct the design parameters and finally establish an accurate finite element model that can comprehensively and correctly reflect the structural characteristics through experimental verification. The main physical parameters involved in the finite element modeling of transmission line induced electricity calculation are the above-mentioned multiple original parameters to be corrected. It can be seen from the principle of induced electricity that among many physical parameters, the five parameters of operating voltage, operating current, dielectric constant, line suspension height, and phase spacing have higher sensitivity to the induced voltage. Therefore, the present invention selects these five parameters as the parameters to be corrected.

[0055] In one embodiment, the screened parameters to be corrected include:

[0056] Operating voltage, operating current, line suspension height, and phase spacing.

[0057] In one embodiment, the method for constructing the neural network model includes:

[0058] Construct a BP neural network model, and based on the PSO algorithm, adjust the initial weights and thresholds in the BP neural network model.

[0059] The activation functions of the hidden layer and the output layer of the BP neural network model are the Sigmoid function and the Tanh function respectively.

[0060] In this embodiment, as Figure 2 shown, using the known induced electricity test results as the input quantity and the 5 screened change amounts of the parameters to be corrected as the output quantity, a neural network model with a 3-layer 2-input 4-output structure is constructed. According to the Kolmogorov theorem, the number of hidden layer neuron nodes is taken as 7, and the neuron nodes are fully connected, that is, each node in the hidden layer is connected to all the nodes in the front and back layers, and the Sigmoid function (Logistic Sigmoid function) and the hyperbolic tangent function (Hyperbolic Tangent Function, Tanh function) are respectively selected as the activation functions of the hidden layer and the output layer to construct a BP neural network model.

[0061] More specifically, in the present invention, the number of particles in the PSO algorithm is set to 50, the maximum speed of the particles is 1, and the maximum number of iterations is 500; the maximum and minimum values of the weights of the operating voltage and current are 0.9 and 0.5 respectively; the fitness curve drops rapidly in the initial stage of iteration and then performs precise search within the extreme point range in the later stage, and the convergence speed slows down. The parameter results predicted by the PSO-BP neural network are compared with the initial model parameters in Table 1, and the prediction errors of each parameter are within 4.5%, meeting the prediction accuracy requirements. The errors of the corrected operating voltage and operating current are 4.1% and 2.6% respectively, and the correction range is within 20%, meeting the requirements of the physical meaning of the parameters. The geometric parameters such as the line suspension height and the phase spacing are optimized in the same way and implemented in the geometric model for subsequent simulation.

[0062] Table 1

[0063] Parameter Initial value Predicted value Error / % Operating voltage 220 kV 229 kV 4.1％ Operating current 150A 154A 2.6％

[0064] In one embodiment, the method for training the neural network model includes:

[0065] Obtain sample data;

[0066] Divide the sample data into a training set and a test set;

[0067] Input the training set into the neural network model to obtain a trained neural network model.

[0068] In one embodiment, the sample data includes induced voltage sample data, induced current sample data, operating voltage sample data, operating current sample data, line suspension height sample data, environmental parameters, and phase spacing sample data.

[0069] In one embodiment, after obtaining the sample data, the method further includes:

[0070] Select sample points for the screened parameters to be corrected in the sample data by using the orthogonal design method and the uniform design method, and ensure the uniform distribution of the sample points determined by each correction parameter under a finite number of times.

[0071] In a specific embodiment, as Figure 3 shown, the overall implementation process of the present invention is as follows:

[0072] Select the parameters to be corrected to determine the parameters to be corrected. Conduct a sensitivity analysis among many design parameters and select the parameters to be corrected.

[0073] Construct a BP neural network, design a PSO algorithm, and optimize the initial weights and thresholds of the network.

[0074] Obtain sample data and train the BP neural network. When the error is less than 5%, that is, after meeting the error requirement, the desired mapping relationship can be regarded as obtained.

[0075] According to the network generalization characteristics, input the actual response characteristic quantity (i.e., the induced electricity test result) into the trained network to obtain the corrected design parameter value.

[0076] The present invention inversely deduces the initial value of the induced electricity calculation through a neural network model, effectively corrects the error caused by the initial value deviation in the traditional method, and significantly improves the calculation accuracy. This method can dynamically adjust the model parameters to ensure that the calculation results are closer to the actual measurement data.

[0077] This accurate calculation model of induced electricity is not only applicable to the power system, but also can be widely applied to many fields such as scientific research, engineering design, and financial analysis, with broad applicability and promotion prospects. Its versatility and flexibility enable it to adapt to the calculation requirements of different scenarios.

[0078] In the traditional method, researchers need to manually measure and input a large number of initial value parameters (such as operating voltage, load current, etc.), while this method reduces manual intervention through automated data collection and neural network processing, significantly reducing the workload of researchers and improving the calculation efficiency at the same time.

[0079] Through the automatic correction and error correction functions of the induced electricity technology, the model can adjust the parameters in real time during the calculation process, reduce human errors and systematic errors, thereby greatly improving the reliability and stability of the calculation and reducing the possibility of calculation errors.

[0080] By performing a sensitivity analysis on multiple original parameters to be corrected, key parameters (such as operating voltage, operating current, line suspension height, and phase spacing) are screened out, reducing the interference of redundant data and improving the training efficiency and calculation accuracy of the model. At the same time, the PSO algorithm is used to optimize the initial weights and thresholds of the BP neural network, further improving the convergence speed and prediction accuracy of the model.

[0081] The sample data is optimized and selected through the orthogonal design method and the uniform design method to ensure that the sample points are evenly distributed and representative, improving the training efficiency and generalization ability of the model and avoiding overfitting or underfitting problems.

[0082] The entire correction process realizes automation and intelligence. From data collection, parameter screening to model training and result output, it reduces human intervention and improves the intelligence level of the calculation process, and is applicable to large-scale data processing and applications in complex scenarios.

[0083] The present invention realizes the accurate calculation of the induced electricity magnitude of the transmission line by introducing a comprehensive model combining neural network and finite element method, providing a new tool for power grid safety management and optimization.

[0084] As Figure 4 shown, a computer system suitable for implementing the correction method for induced electricity calculation provided in the above embodiments includes a central processing module (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) or the program loaded from the storage part into the random access memory (RAM). In the RAM, various programs and data required for the operation of the computer system are also stored. The CPU, ROM, and RAM are connected to each other through a bus. The input / output (I / O) interface is also connected to the bus.

[0085] The following components are connected to the I / O interface: an input part including a keyboard, a mouse, etc.; an output part including a liquid crystal display (LCD), etc. and a speaker, etc.; a storage part including a hard disk, etc.; and a communication part including a network interface card such as a LAN card, a modem, etc. The communication part performs communication processing via a network such as the Internet. The drive is also connected to the I / O interface as needed. A removable medium, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive as needed so that the computer program read from it can be installed into the storage part as needed.

[0086] In particular, according to this embodiment, the process described in the above flowchart can be implemented as a computer software program. For example, this embodiment includes a computer program product that tangibly includes a computer program on a computer-readable medium, and the above computer program includes program code for performing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium.

[0087] The flowcharts and schematic diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of the system, method, and computer program product of this embodiment. In this regard, each block in the flowchart or schematic diagram may represent a module, a program segment, or a part of code, and the above module, program segment, or part of code includes one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order from that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the schematic diagram and / or flowchart, as well as the combination of blocks in the schematic and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0088] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A correction method for induction electricity calculation, characterized in that: Acquire multiple original parameters to be corrected, perform sensitivity analysis and screening on each of the original parameters to be corrected, and obtain screened parameters to be corrected; The induction electrical test results are obtained and input into the trained neural network model to obtain the design parameter values ​​of each screened correction parameter.

2. The correction method for induction electricity calculation according to claim 1, characterized in that: The multiple original parameters to be corrected include: Test parameters, atmospheric pressure, ambient humidity, ambient wind speed, ambient wind direction, operating voltage, operating current, frequency fluctuation, line sag, line suspension height, phase spacing, pole spacing, soil resistivity, and vegetation influence coefficient.

3. The correction method for induction electricity calculation according to claim 1, characterized in that: The sensitivity analysis of each original parameter to be corrected includes: Based on the principle of induction electricity, the sensitivity of each original parameter to be corrected is analyzed.

4. The correction method for induction electricity calculation according to claim 3, characterized in that: The parameters to be modified after screening include: Operating voltage, operating current, line suspension height, phase spacing and dielectric constant.

5. The correction method for induction electricity calculation according to claim 1, characterized in that: The method for constructing the neural network model includes: A BP neural network model is constructed, and the initial weights and thresholds in the BP neural network model are adjusted based on the PSO algorithm.

6. The correction method for induction electricity calculation according to claim 5, characterized in that: The activation functions of the hidden layer and the output layer of the BP neural network model are Sigmoid function and Tanh function respectively.

7. The correction method for induction electricity calculation according to claim 5, characterized in that: The training method of the neural network model includes: Get sample data; Dividing the sample data into a training set and a test set; The training set is input into the neural network model to obtain a trained neural network model.

8. The correction method for induction electricity calculation according to claim 7, characterized in that: The sample data include induced voltage sample data, induced current sample data, operating voltage sample data, operating current sample data, line suspension height sample data, environmental parameters and phase spacing sample data.

9. The correction method for induction electricity calculation according to claim 7, characterized in that: After obtaining the sample data, the method further includes: The sample points of the parameters to be corrected after screening in the sample data are selected through orthogonal design method and uniform design method.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 9 is implemented.

Images