GIS High-Voltage Live Display Threshold Calibration Method, System, Device and Medium

Abstract

The present invention relates to the technical field of high-voltage power transmission, and discloses a method, system, device and medium for threshold verification of a GIS high-voltage live display. By collecting in real time the voltage output waveform output by the sensor of the live display during the operation of the GIS substation, and after being processed by an improved wavelet denoising algorithm, a voltage processed waveform is obtained; when the denoising result of the voltage processed waveform meets the preset index, an improved BP neural network model that adjusts the neuron weights by the momentum method is used to fit the voltage processed waveform to generate a test waveform function; the test waveform function is sampled at a high frequency, and the sampling result is input into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result through the working condition of the status display lamp in the GIS high-voltage live display; realizing accurate threshold verification of the GIS high-voltage live display, and having good application value in the function detection and threshold calibration of the GIS high-voltage live display.

Description

Technical Field

[0001] The present invention relates to the technical field of high-voltage power transmission, and particularly to a method, system, device and medium for calibrating the threshold of a GIS high-voltage live display device. Background Art

[0002] Due to the increasingly intense contradiction between the rapidly growing power demand and the increasingly scarce land, most regions adopt Gas Insulated Switchgear (GIS) to reduce the floor area of substations; as an important part of GIS equipment, the threshold adjustment function of the GIS high-voltage live display device can ensure the reliable operation of the equipment.

[0003] Currently, the calibration of the threshold of the GIS high-voltage live display device is mainly carried out by simulating the live state of on-site equipment in the manufacturing plant, and using high-voltage test equipment to apply a voltage test. Then, the threshold of the GIS high-voltage live display device is adjusted so that when the simulated line is energized, the GIS high-voltage live display device can display that the line is live, thereby setting the factory threshold; however, due to actual complex working conditions such as line induced electricity and aging of electronic components, the voltage signals output by the sensor of the GIS high-voltage live display device and its signal cable in the actual substation are quite different from the factory test state. Therefore, the accuracy of the existing technology for calibrating the threshold of the GIS high-voltage live display device is limited.

[0004] It can be seen that how to improve the calibration accuracy of the threshold of the GIS high-voltage live display device has become an urgent technical problem to be solved by those skilled in the art. Summary of the Invention

[0005] The present invention provides a method, system, device and medium for calibrating the threshold of a GIS high-voltage live display device, so as to solve the problem of how to improve the calibration accuracy of the threshold of the GIS high-voltage live display device.

[0006] To solve the above technical problem, the first aspect of the present invention provides a method for calibrating the threshold of a GIS high-voltage live display device, including:

[0007] When the GIS substation is in operation, the voltage output waveform output by the live display sensor is collected in real time, and the improved wavelet denoising algorithm is used to process the voltage output waveform to obtain a voltage processed waveform;

[0008] Evaluate the denoising result of the voltage processed waveform, and when the denoising result meets the preset index, input the voltage processed waveform into the improved BP neural network model for fitting to obtain a test waveform function; wherein, the neuron weights in the improved BP neural network model are adjusted by the momentum method;

[0009] Perform high-frequency sampling on the test waveform function, and input the sampling result into the GIS high-voltage live display through an arbitrary waveform generator to determine the threshold verification result of the GIS high-voltage live display based on the working condition of the status indicator light in the GIS high-voltage live display.

[0010] As one of the preferred solutions, the improved wavelet denoising algorithm is used to process the voltage output waveform to obtain a voltage processed waveform, including:

[0011] Use the wavelet transform method to perform multi-scale decomposition on the voltage output waveform to obtain multi-layer decomposition coefficients;

[0012] Calculate the noise standard deviation of each decomposition coefficient, and quantify the first threshold of each decomposition coefficient through a first threshold function based on each calculation result;

[0013] Based on each first threshold, apply a second threshold function to each decomposition coefficient to obtain the second threshold of each decomposition coefficient;

[0014] According to each second threshold, perform wavelet inverse transform on each decomposition coefficient to obtain a voltage processed waveform.

[0015] As one of the preferred solutions, the first threshold function is expressed by the following formula:

[0016]

[0017] In the formula, is the first threshold; is the noise standard deviation; is the signal length; is the wavelet decomposition layer number;

[0018] The second threshold function is expressed by the following formula:

[0019]

[0020] In the formula, is the second threshold; is the sign function; is the sampling period; is the wavelet coefficient.

[0021] As one of the preferred solutions, inputting the voltage processed waveform into an improved BP neural network model for fitting to obtain a test waveform function includes:

[0022] Perform Fourier transform on the voltage processed waveform to obtain a voltage waveform function;

[0023] Construct the activation function of neurons and neuron weights in the improved BP neural network model based on the voltage waveform function;

[0024] Extract the historical voltage processing waveform and historical test waveform function of the GIS substation in the past preset period to construct training samples;

[0025] Use the training samples to iteratively train the improved BP neural network model, and introduce the momentum method during the training process to adjust the parameters of the activation function and the neuron weights until the output error of the improved BP neural network model is less than the preset error threshold, and then end the training to obtain the weight vector and basis function;

[0026] Based on the weight vector and the basis function, input the voltage processing waveform into the trained improved BP neural network model for fitting to obtain the test waveform function.

[0027] As one of the preferred solutions, the introduction of the momentum method during the training process to adjust the parameters of the activation function and the neuron weights includes:

[0028] Determine a number of training parameters based on the activation function and the neuron weights;

[0029] In the initial iteration stage, initialize the learning rate and momentum factor corresponding to each training parameter, and set the parameter limit range corresponding to each learning rate and each momentum factor;

[0030] Based on the initialized learning rates and the initialized momentum factors, calculate the gradient of the historical voltage processing waveform with respect to each training parameter, and detect the consistency of the signs of the gradients for three consecutive times. Update the learning rate and momentum factor corresponding to each training parameter according to the consistency detection result;

[0031] Perform a clipping operation on the updated learning rates and the updated momentum factors through the respective parameter limit ranges to obtain the clipped learning rates and the clipped momentum factors to adjust the corresponding training parameters.

[0032] As one of the preferred solutions, the update of the learning rate and momentum factor corresponding to each training parameter according to the consistency detection result includes:

[0033] If the consistency detection result is the first result, update the learning rate and momentum factor corresponding to each training parameter using the first ratio;

[0034] If the consistency detection result is the second result, update the learning rate and momentum factor corresponding to each training parameter using the second ratio.

[0035] As one of the preferred solutions, high-frequency sampling is performed on the test waveform function, and the sampling result is input into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status indicator lamp in the GIS high-voltage live display, including:

[0036] High-frequency sampling is performed on the test waveform function through a preset period to obtain a plurality of voltage waveform data for importing into the arbitrary waveform generator for processing to generate test voltage data;

[0037] The test voltage data is input into the GIS high-voltage live display, and the voltage amplitude is adjusted to determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status indicator lamp in the GIS high-voltage live display.

[0038] The second aspect of the present invention provides a GIS high-voltage live display threshold verification system, including:

[0039] A waveform processing module, configured to, when the GIS substation is operating, collect the voltage output waveform output by the live display sensor in real time, and process the voltage output waveform by using an improved wavelet denoising algorithm to obtain a voltage processed waveform;

[0040] A waveform fitting module, configured to evaluate the denoising result of the voltage processed waveform, and when the denoising result meets a preset index, input the voltage processed waveform into an improved BP neural network model for fitting to obtain a test waveform function; wherein, the neuron weights in the improved BP neural network model are adjusted by the momentum method;

[0041] A threshold verification module, configured to perform high-frequency sampling on the test waveform function, and input the sampling result into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status indicator lamp in the GIS high-voltage live display.

[0042] The third aspect of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the GIS high-voltage live display threshold verification method as described above is implemented.

[0043] A fourth aspect of the present invention provides a computer-readable storage medium, which includes a stored computer program. When the device where the computer-readable storage medium is located executes the computer program, the GIS high-voltage live display threshold verification method as described above is implemented.

[0044] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following:

[0045] (1) By using an improved wavelet denoising algorithm to perform denoising preprocessing on the voltage signal output by the GIS high-voltage live display sensor, noise can be effectively filtered, the signal quality can be improved, and further the influence of sampling noise on the generation of subsequent test voltage waveforms can be effectively suppressed;

[0046] (2) By using an improved BP neural network to fit the voltage processing waveform and generating a GIS high-voltage live display test waveform with higher accuracy through sampling processing. Different from the fitting of the traditional BP neural network, the neuron weights are dynamically adjusted by the momentum method, which can improve the accuracy of the fitting result and accelerate the training and fitting convergence speed;

[0047] (3) The threshold verification of the GIS high-voltage live display can be realized without high-voltage test equipment, with higher safety, lower human and material costs, improved work efficiency, and reduced operation and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for implementation will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0049] Figure 1 is a flowchart of a GIS high-voltage live display threshold verification method provided by an embodiment of the present invention;

[0050] Figure 2 is a comparison diagram of the actual sampling value of the voltage signal of the GIS high-voltage live display and the fitted waveform function provided by an embodiment of the present invention;

[0051] Figure 3 is a structural diagram of a GIS high-voltage live display threshold verification system provided by an embodiment of the present invention;

[0052] Figure 4 is a structural diagram of an electronic device provided by an embodiment of the present invention;

[0053] Reference numerals:

[0054] Among them, 10 is a waveform processing module; 20 is a waveform fitting module; 30 is a threshold verification module; 5000 is an electronic device; 5001 is a processor; 5002 is a bus; 5003 is a memory; 5004 is a transceiver. Specific implementation manners

[0055] The following describes clearly and completely the technical solutions in the embodiments of the present invention with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only a part rather than all of the embodiments of the present invention. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making creative efforts belong to the scope of protection of the present invention.

[0056] In the description of this application, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first", "second", "third", etc. may explicitly or implicitly include one or more of such features. In the description of this application, unless otherwise stated, the meaning of "a plurality" is two or more.

[0057] In the description of this application, it should be noted that unless otherwise clearly defined and limited, the terms "installed", "connected", "connected" should 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. The terms "vertical", "horizontal", "left", "right", "up", "down" and similar expressions used herein are only for the purpose of illustration and do not indicate or imply that the system or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation of the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0058] In the description of this application, it should be noted that unless otherwise defined, all the technical and scientific terms used in the present invention have the same meanings as those commonly understood by those of ordinary skill in the technical field of the present invention. The terms used in the specification of the present invention are only for describing specific embodiments and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0059] In one embodiment, as Figure 1As shown in the figure, the first aspect of the present invention provides a method for verifying the threshold of a GIS high-voltage live display, including:

[0060] S1. When the GIS substation is in operation, the voltage output waveform output by the live display sensor is collected in real time, and the improved wavelet denoising algorithm is used to process the voltage output waveform to obtain a voltage processed waveform; specifically, when the GIS substation is in operation, the present invention obtains, through oscilloscope sampling, the real-time voltage signal waveform obtained by the non-contact detection of the live display sensor outside the safe distance by using the electric field coupling principle between the high-voltage electric field and the sensor within the operating interval of the GIS substation, that is, the voltage output waveform; among them, the voltage signal waveform can also be collected in real time by using a voltage transformer, a data acquisition card, etc.

[0061] In one embodiment, the step of using the improved wavelet denoising algorithm to process the voltage output waveform to obtain a voltage processed waveform includes:

[0062] Performing multi-scale decomposition on the voltage output waveform by using the wavelet transform method to obtain multi-layer decomposition coefficients;

[0063] Calculating the noise standard deviation of each decomposition coefficient, and quantifying the first threshold of each decomposition coefficient through the first threshold function based on each calculation result;

[0064] Applying the second threshold function to each decomposition coefficient based on each first threshold to obtain the second threshold of each decomposition coefficient;

[0065] Performing inverse wavelet transform on each decomposition coefficient according to each second threshold to obtain a voltage processed waveform.

[0066] Specifically, the present invention selects a suitable wavelet basis function according to the characteristics of the voltage output waveform (such as medium and high-order dbN wavelets such as b4, db5, db6, etc., which can reduce the computational complexity while maintaining good time-frequency characteristics), and determines the decomposition layer according to the signal sampling rate and the noise frequency range to ensure that the high-frequency noise is concentrated in the top-layer coefficients; the signal length method can also be used to determine the decomposition layer, which is represented by the following formula:

[0067]

[0068] In the formula, is the wavelet decomposition layer; is the signal length; is a constant, usually taking 3 to 5 to avoid loss of details caused by over-decomposition.

[0069] Based on this, perform discrete wavelet transform on the voltage output waveform to perform multi-scale decomposition on the signal and obtain multi-layer decomposition coefficients.

[0070] Next, calculate the noise standard deviation of the decomposition coefficients of each decomposition layer, and based on the noise standard deviation of each decomposition layer, quantify the first threshold of the decomposition coefficients of each decomposition layer through the first threshold function. The first threshold can be used to preliminarily quantify the noise intensity and roughly screen out significant noise, thereby distinguishing noise from effective signals and laying a foundation for the dynamic adjustment of the second threshold. The first threshold function is expressed by the following formula:

[0071]

[0072]

[0073] In the formula, is the first threshold; is the noise standard deviation; is the median of the absolute value of the wavelet coefficients;

[0074] Among them, As the number of wavelet decomposition layers increases, the first threshold decreases layer by layer. The threshold of the low layer (high frequency) is larger to suppress noise and retain the fundamental wave component, and the threshold of the high layer (low frequency) is smaller to suppress random noise and retain details. The first threshold calculation formula uses as the denominator, which can ensure that the first threshold is adjusted non-linearly with the number of layers and avoid over-smoothing.

[0075] Then, according to the first threshold of each decomposition layer, apply the second threshold function to the decomposition coefficients of each decomposition layer to obtain the second threshold of the decomposition coefficients of each decomposition layer. The second threshold function is expressed by the following formula:

[0076]

[0077] In the formula, is the second threshold; is the sign function; is the sampling period; is the wavelet coefficient.

[0078] When the absolute value of the wavelet coefficient of each decomposition layer is not less than the first threshold, the term approaches 1, and the exponential term approaches , but since the first threshold decreases with the number of layers, the actual contraction amplitude is controlled by to retain the significant signal components. When the absolute value of the wavelet coefficient of each decomposition layer is less than the first threshold, the term and the exponential term act together to achieve a smooth transition to avoid the discontinuity of the hard threshold. By introducing the function and the exponential function, the present invention reduces the signal distortion caused by the traditional soft threshold and is especially suitable for non-stationary noise environments.

[0079] Finally, adjust the amplitudes of the decomposition coefficients of each decomposition layer according to the second threshold of each decomposition layer to retain the effective signals and suppress the noise, and then perform inverse wavelet transform on the adjusted decomposition coefficients, that is, merge the decomposition coefficients layer by layer in the reverse order of decomposition, and then reconstruct the voltage processing waveform. It should be noted that the voltage waveform reconstruction process is the same as the reconstruction process in the original wavelet denoising algorithm. Here, only this reconstruction process is used and it is not improved, so it is not elaborated as the focus.

[0080] The present invention combines multi-scale decomposition with dynamic threshold quantization to effectively suppress noise while retaining the key features (such as mutation points and harmonic components) of the voltage waveform, avoiding signal distortion; by adaptively calculating the threshold of the noise standard deviation, it can adapt to the noise level in different scenarios and improve the system robustness; the second threshold function is optimized through a non-linear function to reduce signal distortion, retain transient features, and optimize the processing speed of high-frequency coefficients. Combined with the fast algorithm of inverse wavelet transform, it meets the real-time monitoring requirements of substations; the voltage waveform after denoising can more accurately reflect the true energized state of GIS equipment, reduce the risk of misjudgment, and improve the decision-making accuracy of condition-based maintenance; algorithm parameters (such as wavelet basis type and decomposition level) can be adjusted according to the actual scenario to adapt to the characteristics of different models of GIS equipment and sensors.

[0081] S2. Evaluate the denoising result of the voltage processing waveform, and when the denoising result meets the preset index, input the voltage processing waveform into the improved BP neural network model for fitting to obtain a test waveform function; wherein, the neuron weights in the improved BP neural network model are adjusted by the momentum method; specifically, the present invention uses the signal-to-noise ratio index as the preset index and evaluates the denoising result of the voltage processing waveform through the signal-to-noise ratio calculation formula; the specific calculation process of the signal-to-noise ratio is not elaborated as the focus here. But when it is determined that the denoising result meets the preset index, input the voltage processing waveform into the improved BP neural network model for fitting, otherwise, repeat the improved wavelet denoising algorithm processing process on the voltage processing waveform until the index requirements are met.

[0082] In an embodiment, the inputting the voltage processing waveform into the improved BP neural network model for fitting to obtain a test waveform function includes:

[0083] Perform Fourier transform on the voltage processing waveform to obtain a voltage waveform function;

[0084] Construct the activation function and neuron weights of the neurons in the improved BP neural network model based on the voltage waveform function;

[0085] Extract the historical voltage processing waveforms and historical test waveform functions of the GIS substation within a past preset time period to construct a training sample;

[0086] Use the training samples to iteratively train the improved BP neural network model, and introduce the momentum method during the training process to adjust the parameters of the activation function and the neuron weights until the output error of the improved BP neural network model is less than the preset error threshold, and then end the training to obtain the weight vector and the basis function;

[0087] Based on the weight vector and the basis function, input the voltage processed waveform into the trained improved BP neural network model for fitting to obtain the test waveform function.

[0088] Specifically, the present invention performs a Fourier transform on the voltage processed waveform to obtain a voltage waveform function, which can be expressed by the following formula:

[0089]

[0090] In the formula, is the voltage waveform function; is the waveform amplitude; n is the order of Fourier expansion; is the multiple of the reference frequency; is the fundamental angular frequency, which is set to 100π; is the initial phase angle.

[0091] Then, based on in the voltage waveform function, construct the activation function of the neurons in the improved BP neural network model, and based on construct the neuron weights, which are expressed by the following formula:

[0092]

[0093]

[0094] In the formula, is the activation function; are respectively the multiple of the reference frequency and the phase (that is, in each order of Fourier expansion in the voltage waveform function, ); where, is the number of sampling points; is the neuron weight, are the components of the neuron weight; where, = 、 = 、 = , and so on.

[0095] Extract the historical voltage processed waveform within the past three months sampled by the instrument (obtained by processing the historical voltage waveform data through an improved wavelet denoising algorithm) and the corresponding historical test waveform function measured by a high-precision instrument; wherein, the sampling frequency is 1 kHz, the length of each waveform is 1 s (1000 points), and these collected data are normalized and segmented to construct a sample set for training an improved BP neural network model; wherein, the improved BP neural network model takes the historical voltage processed waveform as the input and the corresponding historical test waveform function as the output.

[0096] During the iterative training process of the improved BP neural network model, the present invention introduces the momentum method to adjust the parameters of the activation function and the neuron weights, which specifically includes:

[0097] Determine a number of training parameters based on the activation function and the neuron weights;

[0098] In the initial iteration stage, initialize the learning rate and momentum factor corresponding to each of the training parameters, and set the parameter limit ranges corresponding to each of the learning rates and each of the momentum factors;

[0099] Based on the initialized learning rates and the initialized momentum factors, calculate the gradients of the historical voltage processed waveform with respect to each of the training parameters, and detect the consistency of the signs of the gradients for three consecutive times. Update the learning rate and momentum factor corresponding to each of the training parameters according to the consistency detection result;

[0100] Perform a clipping operation on the updated learning rates and the updated momentum factors through each of the parameter limit ranges to obtain the clipped learning rates and the clipped momentum factors to adjust the corresponding training parameters.

[0101] The present invention determines a number of training parameters based on the activation function and the neuron weights, which specifically includes a first training parameter , a second training parameter and a third training parameter , and in the initial iteration stage (that is, the first two iterations), initialize the learning rate and momentum factor corresponding to each training parameter, and set the parameter limit ranges of each learning rate and each momentum factor, that is, the minimum and maximum values of the learning rate are 0.2 and 0.5 respectively, and the minimum and maximum values of the momentum factor are 0.5 and 0.9 respectively. That is to say, when the value of the learning rate or momentum factor in the calculation process is less than its minimum value, the corresponding minimum value is taken, and when the value of the learning rate or momentum factor in the calculation process is greater than its maximum value, the corresponding maximum value is taken.

[0102] Starting from the third iteration, calculate the gradients of the historical voltage processing waveform with respect to each training parameter, and detect the consistency of the signs of the gradients for three consecutive times. Update the learning rate and momentum factor corresponding to each of the training parameters according to the consistency detection result. If the consistency detection result is the first result, update the learning rate and momentum factor corresponding to each training parameter using the first ratio; if the consistency detection result is the second result, update the learning rate and momentum factor corresponding to each training parameter using the second ratio. The present invention takes the first training parameter as an example to illustrate the update process of the learning rate and momentum factor corresponding to the training parameter, which specifically includes:

[0103] At the i-th iteration, where i = 1, 2, that is, in the first two iteration processes, set , where is the learning rate of the first training parameter, and set , where is the momentum factor of the first training parameter; while when i > 2, calculate the gradient of the first training parameter through the following formula:

[0104]

[0105] In the formula, is the gradient of the first training parameter at the i-th iteration; is the historical voltage processing waveform.

[0106] Then extract the sign of the gradient of the first training parameter in three consecutive iteration processes and use it for consistency detection. If the consistency detection result is the first result, that is, the positive and negative of these three gradient signs are the same, then update the learning rate and momentum factor corresponding to the first training parameter through the following formula:

[0107]

[0108]

[0109] In the formula, is the first ratio, and its value is greater than 1. Generally, 1.1 is taken in the example.

[0110] If the consistency detection result is the second result, that is, the positive and negative of these three gradient signs are different, then update the learning rate and momentum factor corresponding to the first training parameter through the following formula:

[0111]

[0112]

[0113] In the formula, is the second ratio, and its value is , generally take 0.5 in the example.

[0114] In the present invention, when the consistency detection result shows that the gradient directions are consistent, the first ratio accelerates convergence and enhances the update inertia by increasing the learning rate and the momentum factor; when the consistency detection result shows that the gradient directions oscillate (i.e., are different), the second ratio suppresses the oscillation and improves the flexibility of direction adjustment by decreasing the learning rate and the momentum factor; the training speed and stability are dynamically balanced through ratio adjustment to improve the performance of the model in complex optimization scenarios.

[0115] Perform a clipping operation on each updated learning rate and each updated momentum factor according to the initially set parameter limit range to obtain each clipped learning rate and each clipped momentum factor; then adjust the corresponding training parameters according to each clipped learning rate and each clipped momentum factor. Taking the first training parameter as an example, its specific adjustment process is represented by the following formula:

[0116]

[0117] In the formula, and are respectively the adjustment amounts of the first training parameter in the and th iteration processes; and are respectively the learning rate and the momentum factor of the first training parameter after the clipping operation.

[0118] It should be noted that the adjustment processes of the learning rates and the momentum factors corresponding to the second training parameter and the third training parameter are the same as those of the learning rate and the momentum factor corresponding to the first training parameter; and the adjustment processes of the second training parameter and the third training parameter are also the same as those of the first training parameter, which will not be elaborated here.

[0119] Based on the adjusted training parameters, set the parameters of the improved BP neural network model so that the improved BP neural network model after parameter setting outputs the historical test waveform function, which can be represented by the following formula:

[0120]

[0121] In the formula, is the historical test waveform function.

[0122] The output error of the improved BP neural network model can be represented by the following formula:

[0123]

[0124] In the formula, is the output error; It is the ideal voltage signal in the historical voltage processing waveform.

[0125] The preset error threshold can be expressed by the following formula:

[0126]

[0127] In the formula, is the preset error threshold.

[0128] Judge the relationship between the output error of the improved BP neural network model and the preset error threshold. If the output error of the improved BP neural network model is not less than the preset error threshold, it means that the model accuracy is insufficient and further training is required; if the output error of the improved BP neural network model is less than the preset error threshold, it means that the model accuracy is sufficient, stop training, and obtain the weight vector and the basis function 、 . Finally, based on the weight vector and the basis function, input the voltage processing waveform into the trained improved BP neural network model for fitting to obtain the test waveform function.

[0129] The present invention extracts the waveform frequency domain features through Fourier transform, combines dynamic excitation function and momentum method optimization, enables the improved BP neural network to accurately fit complex voltage waveforms, especially suitable for GIS device signals with nonlinear and time-varying characteristics, and is different from the fitting of traditional BP neural networks. It dynamically adjusts the fundamental frequency multiple as the neuron weight value to reduce the interference of non-integer multiples of the fundamental wave frequency component on the fitting result, and speeds up the training and fitting convergence speed by the momentum method and setting the range of the learning rate and momentum factor; introduces historical data training and momentum method optimization to avoid overfitting, improves the prediction ability of the model for unknown working conditions, and reduces the risk of misjudgment; uses the momentum method to accelerate convergence, reduces the number of iterations, and combines Fourier transform to reduce the input dimension, significantly shortening the model training time.

[0130] S3. Perform high-frequency sampling on the test waveform function, and input the sampling result into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status display lamp in the GIS high-voltage live display;

[0131] In one embodiment, step S3 includes:

[0132] Perform high-frequency sampling on the test waveform function through a preset period to obtain a number of voltage waveform data for importing into the arbitrary waveform generator for processing to generate test voltage data;

[0133] Input the test voltage data into the GIS high-voltage live display, and determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status indicator light in the GIS high-voltage live display by adjusting the voltage amplitude.

[0134] Specifically, the present invention performs high-frequency sampling on the test waveform function through a preset period (preferably 20 ), obtains 1000 voltage waveform data within a single period for import into an arbitrary waveform generator for processing, generates test voltage data for output to the signal input end of the GIS high-voltage live display to be measured. Determine that the input amplitude of the input test voltage data exceeds the locking trigger threshold based on the constant illumination of the indicator light indicating live in the status indicator light in the GIS high-voltage live display. Each time, reduce the amplitude of the test voltage data by 5% for testing until the indicator light indicating no electricity in the live status indicator light of the GIS high-voltage live display is constantly illuminated. Then, take the amplitude of the test voltage data when the live status indicator light was constantly illuminated last time as the verification value of the current threshold; adjust the threshold knob position of the GIS high-voltage live display, and so on, repeat the above test process until the voltage amplitude setting values of all threshold positions are obtained, and the threshold verification result of the GIS high-voltage live display to be measured can be obtained.

[0135] Compare the actual sampled value of the voltage signal of the GIS high-voltage live display with the fitted waveform function as Figure 2 shown. It can be seen from Figure 2 that the improved wavelet denoising algorithm adopted by the present invention effectively eliminates noise. The fitting result shown in Figure 2 obtained after the improved BP neural network is iteratively trained 500 times has an error less than 0.001%, indicating that the solution of the present invention can greatly improve the verification accuracy of the threshold of the GIS high-voltage live display.

[0136] In the embodiments of the present application, in view of the problem of how to improve the calibration accuracy of the threshold of the GIS high-voltage live display, a method for calibrating the threshold of the GIS high-voltage live display is designed. By collecting the voltage output waveform output by the live display sensor during the operation of the GIS substation in real time, after processing with an improved wavelet denoising algorithm, a voltage processed waveform is obtained; when the denoising result of the voltage processed waveform meets the preset index, an improved BP neural network model that adjusts the neuron weights by the momentum method is used to fit the voltage processed waveform to generate a test waveform function; the test waveform function is sampled at a high frequency, and the sampling result is input into the GIS high-voltage live display through an arbitrary waveform generator to determine the threshold calibration result of the GIS high-voltage live display based on the working condition of the status indicator light in the GIS high-voltage live display. The present invention obtains the actual on-site voltage signal waveform through sampling, and obtains a good analog test signal waveform through algorithm denoising and oversampling processing, accurately and efficiently realizing the threshold calibration of the GIS high-voltage live display, and having good application value in the function detection and threshold calibration of the GIS high-voltage live display.

[0137] It should be noted that although the steps in the above flowcharts are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise clearly stated in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders.

[0138] In another embodiment, as Figure 3 shown, the second aspect of the present invention provides a system for calibrating the threshold of a GIS high-voltage live display, including:

[0139] A waveform processing module 10, configured to collect the voltage output waveform output by the live display sensor in real time during the operation of the GIS substation, and process the voltage output waveform with an improved wavelet denoising algorithm to obtain a voltage processed waveform;

[0140] A waveform fitting module 20, configured to evaluate the denoising result of the voltage processed waveform, and when the denoising result meets the preset index, input the voltage processed waveform into an improved BP neural network model for fitting to obtain a test waveform function; wherein, the neuron weights in the improved BP neural network model are adjusted by the momentum method;

[0141] A threshold calibration module 30, configured to sample the test waveform function at a high frequency, and input the sampling result into the GIS high-voltage live display through an arbitrary waveform generator to determine the threshold calibration result of the GIS high-voltage live display based on the working condition of the status indicator light in the GIS high-voltage live display.

[0142] It should be noted that each module in the above GIS high-voltage live display threshold calibration system can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above modules can be embedded in the processor in the computer device in hardware form or independent of the processor, or stored in the memory in the computer device in software form, so as to facilitate the processor to call and execute the operations corresponding to each of the above modules. For the specific limitations of a GIS high-voltage live display threshold calibration system, refer to the limitations of a GIS high-voltage live display threshold calibration method in the above text. The two have the same functions and effects, and will not be elaborated here.

[0143] The third aspect of the present invention provides an electronic device, which includes:

[0144] A processor, a memory, and a bus;

[0145] The bus is used to connect the processor and the memory;

[0146] The memory is used to store operation instructions;

[0147] The processor is used to execute the operations corresponding to a GIS high-voltage live display threshold calibration method as shown in the first aspect of the present application by calling the operation instructions.

[0148] In an optional embodiment, an electronic device is provided, as Figure 4 shown, Figure 4 The electronic device 5000 shown includes a processor 5001 and a memory 5003. Among them, the processor 5001 and the memory 5003 are connected, such as connected through a bus 5002. Optionally, the electronic device 5000 may further include a transceiver 5004. It should be noted that in practical applications, the transceiver 5004 is not limited to one, and the structure of the electronic device 5000 does not constitute a limitation to the embodiments of the present application.

[0149] The processor 5001 can be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logic blocks, modules, and circuits described in combination with the disclosed content of the present application. The processor 5001 can also be a combination that realizes computing functions, such as a combination including one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0150] The bus 5002 may include a path for transmitting information between the above components. The bus 5002 can be a PCI bus or an EISA bus, etc. The bus 5002 can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 4It is represented only by a thick line, but it does not mean that there is only one bus or one type of bus.

[0151] The memory 5003 can be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or it can also be an EEPROM, a CD-ROM or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0152] The memory 5003 is used to store the application program code for implementing the solution of this application, and is controlled and executed by the processor 5001. The processor 5001 is used to execute the application program code stored in the memory 5003 to implement the content shown in any of the foregoing method embodiments.

[0153] Among them, the electronic device includes but is not limited to: mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc.

[0154] The fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements a GIS high-voltage live display threshold verification method shown in the first aspect of this application.

[0155] Another embodiment of this application provides a computer-readable storage medium, on which a computer program is stored, and when it runs on a computer, it enables the computer to execute the corresponding content in the foregoing method embodiments.

[0156] In addition, an embodiment of the present invention also proposes a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the steps of the above method.

[0157] In summary, the present invention relates to the technical field of high-voltage power transmission, and discloses a method, system, device and medium for threshold verification of a GIS high-voltage live display. By collecting the voltage output waveform output by the live display sensor during the operation of the GIS substation in real time, and after processing with an improved wavelet denoising algorithm, a voltage processed waveform is obtained; when the denoising result of the voltage processed waveform meets the preset index, an improved BP neural network model that adjusts the neuron weights by the momentum method is used to fit the voltage processed waveform to generate a test waveform function; the test waveform function is sampled at a high frequency, and the sampling result is input into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result through the working condition of the status indicator light in the GIS high-voltage live display; the threshold verification of the GIS high-voltage live display is accurately and efficiently realized, and it has good application value in the function detection and threshold calibration of the GIS high-voltage live display.

[0158] Each embodiment in this specification is described in a progressive manner. For the parts that are the same or similar in each embodiment, reference can be made to each other directly. The key point of each embodiment is to illustrate the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0159] The above-described embodiments only represent several preferred implementation manners of the present application, and the description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be pointed out that for those of ordinary skill in the art in this technical field, without departing from the technical principle of the present invention, several improvements and substitutions can still be made, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the protection scope of the claimed rights.

Claims

1. A threshold verification method for GIS high-voltage live display, characterized in that, Including: When the GIS substation is in operation, the voltage output waveform output by the live display sensor is collected in real time, and the improved wavelet denoising algorithm is used to process the voltage output waveform to obtain the voltage processed waveform; Evaluate the denoising result of the voltage processed waveform, and when the denoising result meets the preset index, input the voltage processed waveform into the improved BP neural network model for fitting to obtain the test waveform function; wherein, the neuron weights in the improved BP neural network model are adjusted by the momentum method; Perform high-frequency sampling on the test waveform function, and input the sampling result into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status display lamp in the GIS high-voltage live display; The performing high-frequency sampling on the test waveform function, and inputting the sampling result into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status display lamp in the GIS high-voltage live display, includes: Perform high-frequency sampling on the test waveform function through a preset period, obtain 1000 voltage waveform data within a single period to be imported into an arbitrary waveform generator for processing, generate test voltage data to be output to the signal input end of the measured GIS high-voltage live display, and judge that the input amplitude of the input test voltage data exceeds the locking trigger threshold according to the fact that the indicator light showing live in the status display lamp of the GIS high-voltage live display is always on. Each time, reduce the amplitude of the test voltage data by 5% for testing until the indicator light showing no power in the live status indicator light of the GIS high-voltage live display is always on. Then, take the amplitude of the test voltage data when the live status indicator light was always on last time as the verification value of the current threshold; adjust the threshold knob gear of the GIS high-voltage live display, and so on, repeat the above test process until the voltage amplitude setting values of all threshold gears are obtained, and the threshold verification result of the measured GIS high-voltage live display can be obtained.

2. The GIS high-voltage live display threshold calibration method according to claim 1, characterized in that The using the improved wavelet denoising algorithm to process the voltage output waveform to obtain the voltage processed waveform, includes: Use the wavelet transform method to perform multi-scale decomposition on the voltage output waveform to obtain multi-layer decomposition coefficients; Calculate the noise standard deviation of each decomposition coefficient, and quantify the first threshold of each decomposition coefficient through the first threshold function based on each calculation result; Based on each first threshold, apply the second threshold function to each decomposition coefficient to obtain the second threshold of each decomposition coefficient; According to each second threshold, perform inverse wavelet transform on each decomposition coefficient to obtain the voltage processed waveform.

3. A threshold verification method for a GIS high-voltage live display according to claim 2, characterized in that, The first threshold function is expressed by the following formula: In the formula, is the first threshold; is the noise standard deviation; is the signal length; is the wavelet decomposition level; The second threshold function is expressed by the following formula: In the formula, is the second threshold; is the sign function; is the sampling period; are the wavelet coefficients.

4. A method for calibrating the threshold of a GIS high-voltage live display according to claim 1, characterized in that, The inputting the voltage processed waveform into the improved BP neural network model for fitting to obtain the test waveform function, includes: Perform Fourier transform on the voltage processed waveform to obtain the voltage waveform function; Construct the activation function of neurons and neuron weights in the improved BP neural network model based on the voltage waveform function; Extract the historical voltage processing waveform and historical test waveform function of the GIS substation within a preset period in the past to construct a training sample; Use the training sample to perform iterative training on the improved BP neural network model, and introduce the momentum method during the training process to adjust the parameters of the activation function and the neuron weights until the output error of the improved BP neural network model is less than the preset error threshold, then end the training to obtain the weight vector and basis function; Based on the weight vector and the basis function, input the voltage processing waveform into the trained improved BP neural network model for fitting to obtain the test waveform function.

5. A threshold verification method for a GIS high-voltage live display according to claim 4, characterized in that, The introduction of the momentum method during the training process to adjust the parameters of the activation function and the neuron weights includes: Determine a number of training parameters based on the activation function and the neuron weights; In the initial iteration stage, initialize the learning rate and momentum factor corresponding to each of the training parameters, and set the parameter limit ranges corresponding to each of the learning rates and each of the momentum factors; Based on the initialized learning rates and the initialized momentum factors, calculate the gradient of the historical voltage processing waveform with respect to each of the training parameters, and detect the consistency of the signs of the gradients for three consecutive times. Update the learning rate and momentum factor corresponding to each of the training parameters according to the consistency detection result; Perform a clipping operation on the updated learning rates and the updated momentum factors through the respective parameter limit ranges to obtain the clipped learning rates and the clipped momentum factors to adjust the corresponding training parameters.

6. A threshold verification method for a GIS high-voltage live display according to claim 5, characterized in that The updating of the learning rate and momentum factor corresponding to each of the training parameters according to the consistency detection result includes: If the consistency detection result is the first result, update the learning rate and momentum factor corresponding to each of the training parameters using the first ratio; If the consistency detection result is the second result, update the learning rate and momentum factor corresponding to each of the training parameters using the second ratio.

7. A threshold verification method for a GIS high-voltage live display according to claim 1, characterized in that The high-frequency sampling of the test waveform function and inputting the sampling result into the GIS high-voltage live display through an arbitrary waveform generator to determine the threshold verification result of the GIS high-voltage live display based on the working condition of the status display lamp in the GIS high-voltage live display includes: Perform high-frequency sampling on the test waveform function through a preset period to obtain a number of voltage waveform data and import them into the arbitrary waveform generator for processing to generate test voltage data; Input the test voltage data into the GIS high-voltage live display, and adjust the voltage amplitude to determine the threshold verification result of the GIS high-voltage live display based on the working condition of the status display lamp in the GIS high-voltage live display.

8. A threshold verification system for GIS high-voltage live display, characterized in that, Include: A waveform processing module, which is used to collect the voltage output waveform output by the live display sensor in real time when the GIS substation is running, and process the voltage output waveform by using an improved wavelet denoising algorithm to obtain a voltage processed waveform; A waveform fitting module, which is used to evaluate the denoising result of the voltage processed waveform, and when the denoising result meets the preset index, input the voltage processed waveform into an improved BP neural network model for fitting to obtain a test waveform function; wherein, the neuron weights in the improved BP neural network model are adjusted by the momentum method; A threshold verification module, which is used to perform high-frequency sampling on the test waveform function, and input the sampling result into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status display lamp in the GIS high-voltage live display; The performing high-frequency sampling on the test waveform function, and inputting the sampling result into the GIS high-voltage live display through an arbitrary waveform generator, so as to determine the threshold verification result of the GIS high-voltage live display according to the working condition of the status display lamp in the GIS high-voltage live display, includes: Performing high-frequency sampling on the test waveform function through a preset period to obtain 1000 voltage waveform data within a single period for importing into an arbitrary waveform generator for processing, generating test voltage data for outputting to the signal input end of the measured GIS high-voltage live display, judging that the input amplitude of the input test voltage data exceeds the locking trigger threshold according to the fact that the indicator light showing live in the status display lamp in the GIS high-voltage live display is always on, reducing the amplitude of the test voltage data by 5% each time for testing until the indicator light showing no power in the live status indicator light of the GIS high-voltage live display is always on, then taking the amplitude of the test voltage data when the previous live status indicator light was always on as the verification value of the current threshold; adjusting the threshold knob gear of the GIS high-voltage live display, and so on, repeating the above test process until the voltage amplitude setting values of all threshold gears are obtained, then the threshold verification result of the measured GIS high-voltage live display can be obtained.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the GIS high-voltage live display threshold verification method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program. When the device where the computer-readable storage medium is located executes the computer program, it implements the GIS high-voltage live display threshold verification method according to any one of claims 1 to 7.

Images