A method and system for monitoring partial discharge of insulation inside a switch cabinet

Abstract

The present invention relates to the technical field of power equipment monitoring, and discloses a method and system for monitoring partial discharge of insulation inside a switch cabinet. The present invention uses a preprocessed feature sample set to construct a target partial discharge monitoring model for monitoring partial discharge inside the switch cabinet, and monitors the preprocessed monitoring feature signal through the target partial discharge monitoring model. The preprocessed feature signal can effectively restore the important features of the original feature signal, improve the quality of the feature signal, and remove the interference of unnecessary signals, thereby accurately monitoring partial discharge of insulation inside the switch cabinet. The present invention solves the technical problem of insufficient monitoring accuracy in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment monitoring, and in particular to a method and system for monitoring partial discharge of insulation inside a switch cabinet. Background Art

[0002] As power systems gradually increase their requirements for power supply reliability, higher demands are being placed on the importance of electrical equipment status monitoring. Switchgear is an important control device in the power distribution link of the power system. The equipment in the switchgear is exposed to strong magnetic fields, high currents, and high voltages for a long time. Monitoring and protecting its normal operating life cycle is crucial. During long-term operation, the insulation inside the switchgear deteriorates due to the high currents and strong magnetic fields. In severe cases, partial discharge occurs in the insulation components, which can endanger the normal operation of the power system. There are a large number of insulating support components inside the switchgear. If each insulating component is monitored one by one, it will not only greatly increase the complexity of the equipment and the engineering cost, but also increase the failure rate of the equipment.

[0003] During normal operation of the power system, switchgear is exposed to strong magnetic fields, high field strengths, and high currents for long periods of time. This can cause various defects in the internal insulation, leading to partial discharge within the switchgear, further degrading the insulation. If the abnormality is not detected and eliminated at this time, it may lead to insulation breakdown in the electrical equipment. When partial discharge occurs in electrical insulation, it generates acoustic, optical, electrical, and magnetic fault signals. In the initial stages of a fault, the light generated by the discharge is often very weak, and due to obstruction by objects, it is difficult to detect except in a dark environment. Therefore, finding a more convenient and accurate monitoring method is of great significance to the intelligent diagnosis and decision-making support system for internal insulation component faults in switchgear equipment. Summary of the Invention

[0004] The present invention provides a method and system for monitoring partial discharge of insulation inside a switch cabinet, which solves the technical problem of insufficient monitoring accuracy in the prior art.

[0005] A first aspect of the present invention provides a method for monitoring partial discharge of internal insulation of a switch cabinet, comprising:

[0006] Preprocessing the sample characteristic signals received from the switchgear to generate a sample set;

[0007] Using the sample set as input to a preset initial partial discharge monitoring model for training, to construct a target partial discharge monitoring model;

[0008] The received monitoring characteristic signal is preprocessed, and the preprocessed monitoring characteristic signal is monitored by the target partial discharge monitoring model to generate a monitoring report.

[0009] Optionally, the preprocessing includes:

[0010] Perform wavelet transform on each characteristic signal;

[0011] Performing noise reduction on the characteristic signal after the wavelet transformation to generate a plurality of noise-reduced characteristic signals;

[0012] Reconstruct the plurality of noise reduction feature signals to generate a target feature signal.

[0013] Optionally, the performing denoising on the characteristic signal after the wavelet transformation to generate a plurality of denoised characteristic signals includes:

[0014] injecting white noise into the characteristic signal after wavelet transformation to generate a white noise characteristic signal;

[0015] The white noise characteristic signal is decomposed, and when a preset decomposition iteration stopping condition is met, a plurality of noise reduction characteristic signals are obtained.

[0016] Optionally, the preset decomposition iteration stopping condition is that the standard deviation value between the IMF mean components in two adjacent noise reduction feature signals is within a preset standard deviation value interval.

[0017] Optionally, reconstructing the plurality of noise reduction characteristic signals to generate a target characteristic signal includes:

[0018] Sort the noise reduction feature signals according to the order of the IMF mean components from high to low;

[0019] Calculating the IMF correlation coefficient of each of the IMF mean components;

[0020] The noise reduction feature signal associated with the IMF correlation coefficient that meets the preset elimination condition is eliminated to generate a target feature signal.

[0021] Optionally, the preset elimination condition is specifically that the IMF correlation coefficient is less than a preset coefficient threshold.

[0022] Optionally, a genetic algorithm is used to optimize parameters of the preset initial partial discharge monitoring model.

[0023] A second aspect of the present invention provides a switch cabinet internal insulation partial discharge monitoring system, comprising a sensor module, a partial discharge monitoring module and a computer terminal that are communicatively connected in sequence;

[0024] The sensor module is used to collect characteristic signals;

[0025] The partial discharge monitoring module is used to pre-process the characteristic signal, build a target partial discharge monitoring model, perform partial discharge monitoring, and display monitoring results;

[0026] The computer terminal is used to output the monitoring report.

[0027] Optionally, the partial discharge monitoring module includes an oscilloscope and a partial discharge monitor that are communicatively connected to each other;

[0028] The partial discharge monitor is used to pre-process the characteristic signal, construct a target partial discharge monitoring model, and perform partial discharge monitoring;

[0029] The oscilloscope is used to display monitoring characteristic signals and monitoring results.

[0030] Optionally, the sensor module and the partial discharge monitoring module are communicatively connected via an RS485 communication circuit.

[0031] It can be seen from the above technical solutions that the present invention has the following advantages:

[0032] The present invention uses a preprocessed feature sample set to construct a target partial discharge monitoring model for partial discharge monitoring inside a switch cabinet, and monitors the preprocessed monitoring feature signal through the target partial discharge monitoring model. The preprocessed feature signal can effectively restore the important features of the original feature signal, improve the quality of the feature signal, and remove the interference of unnecessary signals, thereby accurately monitoring the partial discharge of the insulation inside the switch cabinet. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0034] Figure 1 A flowchart of the steps of a method for monitoring partial discharge of internal insulation of a switch cabinet provided by an embodiment of the present invention;

[0035] Figure 2 A flowchart of another step of a method for monitoring partial discharge of internal insulation of a switch cabinet provided by an embodiment of the present invention;

[0036] Figure 3 A schematic diagram of a flow chart of an adaptive ensemble empirical mode decomposition algorithm provided in an embodiment of the present invention;

[0037] Figure 4 The collected original signal waveform provided by the embodiment of the present invention;

[0038] Figure 5 A schematic diagram of the results of noise reduction processing using a wavelet denoising algorithm provided in an embodiment of the present invention;

[0039] Figure 6 A schematic diagram of the noise reduction results using the EEMD denoising algorithm provided in an embodiment of the present invention;

[0040] Figure 7 A schematic diagram of the noise reduction results using the AEEMD denoising algorithm provided in an embodiment of the present invention;

[0041] Figure 8 A structural diagram of a switch cabinet internal insulation partial discharge monitoring system provided by an embodiment of the present invention;

[0042] Figure 9 This is a schematic diagram of an RS485 communication circuit provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The embodiments of the present invention provide a method and system for monitoring partial discharge of insulation inside a switch cabinet, which are used to solve the technical problem of insufficient monitoring accuracy in the prior art.

[0044] In order to make the purpose, features, and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0045] See also Figure 1-Figure 2 , Figure 1 A flowchart of the steps of a method for monitoring partial discharge of internal insulation of a switch cabinet provided by an embodiment of the present invention.

[0046] The present invention provides a method for monitoring partial discharge of insulation inside a switch cabinet, comprising:

[0047] Step 101: Preprocess the sample characteristic signals received from the switch cabinet to generate a sample set.

[0048] In an embodiment of the present invention, the sample characteristic signals received from the switchgear are preprocessed, and the preprocessed characteristic signals are matched with known partial discharge degree characteristic signals to obtain a high-voltage switchgear partial discharge fault characteristic data sample set.

[0049] Step 102: Using the sample set to input a preset initial partial discharge monitoring model for training, and constructing a target partial discharge monitoring model.

[0050] It should be noted that the preset initial partial discharge monitoring model uses a BP neural network model. The BP neural network is a multi-layer feedforward network that propagates backward based on deviations. It acquires standard knowledge from training input and output data, memorizing it in the network's weights. It excels in self-learning and self-adaptation, enabling it to achieve excellent recognition results based on a large amount of training data, even when the partial discharge measurement system and environment change.

[0051] In an embodiment of the present invention, part of the high-voltage switchgear partial discharge fault characteristic data sample set is input as a training set, and then part of the data set is taken as a test set to compare the prediction accuracy and construct a target partial discharge monitoring model.

[0052] Furthermore, a genetic algorithm is used to optimize the parameters of the preset initial partial discharge monitoring model. The specific process is as follows:

[0053] 1. Import the sample set of partial discharge fault characteristic data of high-voltage switchgear and perform normalization processing;

[0054] 2. Test different hidden nodes, design the optimal BP neural network topology, and establish relevant important parameters of the genetic algorithm, including population operation scale, iteration frequency, crossover probability, mutation probability, etc.

[0055] 3. Arbitrarily transform the neural network into initial weights and thresholds, and the genetic algorithm encodes the original weights and thresholds of the neural network into the initial population.

[0056] 4. Fitness function selection. The fitness function calculates the fitness value, which is an indicator of the effectiveness of the genetic algorithm. Therefore, the selection of the fitness function is crucial. The trained BP neural network is tested, the output is predicted, and the deviation between the neural network output and the expected output is calculated. The fitness function is calculated from the deviation function, and the fitness value is the inverse of the sum of squared deviations.

[0057] 5. Genetic manipulation:

[0058] (1) Selection operation. Calculate the fitness of each population. After calculating the probability based on the fitness, perform crossover and mutation operations.

[0059] (2) Crossover operation. Using arithmetic crossover, two chromosomes are arranged in a row according to the crossover probability, creating a new chromosome to replace the original chromosome, and directly copying the non-crossover chromosome.

[0060] (3) Mutation operation: Using non-uniform mutation, the genetic gene is arbitrarily perturbed according to the mutation probability, and the result of the perturbation transformation is used as the value of the new genetic gene.

[0061] 6. Calculate the individual fitness values of the new population. Assign the optimal weights and thresholds obtained by the genetic algorithm to the original weights and thresholds of the neural network to replace the arbitrarily selected weights and thresholds.

[0062] 7. BP neural network training. The training process will terminate when the model measurement error meets the preset standard error value. Otherwise, the weight value and threshold value will be adjusted inversely until the deviation accuracy or training frequency meets the requirements, and the target partial discharge monitoring model will be output.

[0063] It should be noted that the use of genetic algorithms to optimize model parameters here is a conventional model optimization technology and will not be described in detail here.

[0064] Step 103 : pre-process the received monitoring characteristic signal, monitor the pre-processed monitoring characteristic signal through the target partial discharge monitoring model, and generate a monitoring report.

[0065] Furthermore, the preprocessing includes preprocessing the sample characteristic signal of the switch cabinet and preprocessing the monitoring characteristic signal, and the preprocessing processes are consistent.

[0066] Preprocessing includes:

[0067] S11, performing wavelet transform on each characteristic signal;

[0068] S12, performing noise reduction on the characteristic signal after wavelet transformation to generate multiple noise reduction characteristic signals;

[0069] S13. Reconstruct the multiple noise reduction feature signals to generate a target feature signal.

[0070] It should be noted that the characteristic signals include transient voltage signals and ultrasonic signals.

[0071] It should be noted that the present invention describes the basis of the selected denoising algorithm. First, the basis of the selected denoising algorithm is described. As the basic algorithm of the improved algorithm, the empirical mode decomposition (EMD) algorithm has good time-frequency characteristics and strong adaptability, but due to severe modal aliasing, it is not conducive to signal denoising analysis. When the EEMD algorithm is used to decompose the signal, only the first-order mode of the signal is decomposed until the remaining part of the signal reaches the EEMD decomposition termination condition. After removing the first mode, noise is added to the residual signal, and the residual signal is decomposed in the next stage. The algorithm eliminates the modal aliasing phenomenon to a large extent, but the complexity of the method and the versatility of calculation are also increased. In order to solve this problem, the present application proposes a new denoising algorithm-adaptive ensemble empirical mode decomposition algorithm (AEEMD) to denoise the characteristic signal after wavelet transformation.

[0072] It is worth mentioning that the adaptive ensemble empirical mode decomposition (AEEMD) algorithm is an improvement on the ensemble empirical mode decomposition (EEMD) algorithm, and the EEMD algorithm is an improvement on the empirical mode decomposition (EMD) algorithm.

[0073] AEEMD algorithm advantages:

[0074] (1) It avoids the disadvantages of the Fourier transform (FFT) algorithm being unable to apply to PD non-stationary signals and the wavelet threshold denoising incorrectly identifying multiple signals.

[0075] (2) Reduce the complexity of the EEMD algorithm and the multifaceted nature of the calculations, further improve the recognition accuracy and speed, and be more conducive to signal denoising analysis.

[0076] In an embodiment of the present invention, a characteristic signal is first collected by a sensor, and a wavelet transform is performed on the characteristic signal to remove signals unrelated to the local discharge acoustic and electrical signals; then, an improved AEEMD denoising algorithm is used to denoise the characteristic signal after the wavelet transform to generate a denoised characteristic signal; finally, the denoised characteristic signal is reconstructed to generate a target characteristic signal.

[0077] Furthermore, based on the adaptive ensemble empirical mode decomposition algorithm, the characteristic signal after wavelet transformation is denoised to generate multiple denoised characteristic signals, including:

[0078] S121, injecting white noise into the characteristic signal after wavelet transformation to generate a white noise characteristic signal;

[0079] S122 , decomposing the white noise characteristic signal, and obtaining a plurality of noise reduction characteristic signals when a preset decomposition iteration stopping condition is met.

[0080] It should be noted that the characteristic signals after wavelet transformation include transient voltage signals and ultrasonic signals, and both types of signals are decomposed by the adaptive ensemble empirical mode decomposition algorithm.

[0081] For easier understanding, see Figure 3 , the following is the process of using the adaptive ensemble empirical mode decomposition algorithm to reduce the noise of the characteristic signal after wavelet transformation:

[0082] Step 1: Input the original signal x(t), i.e., transient voltage signal or ultrasonic signal;

[0083] Step 2: Add Gaussian white noise signal n to the original signal x(t) i (t), and get the noise elimination signal x i (t);

[0084] x i (t) = x(t) + n i (t)

[0085] Step 3: Eliminate the noise signal x i (t) performs AEEMD decomposition to obtain the first-order IMF component c1(t);

[0086]

[0087] Among them, all the maximum points of the original signal x(t) are obtained, and the maximum envelope e is fitted by the cubic spline function. max (t); Similarly, find all the minimum points of the original signal x(t), and fit the minimum envelope of the signal e by the cubic spline function min (t);

[0088] c i,1 (t) = x i (t)-m1(t)

[0089] Since the number of times Gaussian white noise is added is N, the first-order IMF component of AEEMD decomposition after adding Gaussian white noise is c i,1 (t), therefore, the first-order mode is decomposed and averaged, and the results are as follows:

[0090]

[0091] Step 4: Calculate the margin r1(t) of the first-order IMF mean component c1(t) after decomposition;

[0092] r1(t)=x i (t)-c1(t)

[0093] Step 5: Decompose the added Gaussian white noise signal n i (t), i = 1, 2, ... N, and the IMF component group D1 (n i (t));

[0094] It should be noted that the decomposition method is the same as step 3. Define the function D1(n i (t)) represents a set of IMF components after the empirical mode decomposition of the signal. By decomposing n i The noise in (t) obtains the first-order IMF component.

[0095]

[0096] Step 6: Use the margin r1(t) and the IMF component group D1(n i (t)), construct a new feature signal And adopt new characteristic signals Jump to step 3-step 5 until the preset decomposition iteration stop condition is met, and multiple noise reduction feature signals are obtained (t) to

[0097]

[0098] For ease of understanding, the new characteristic signal in step 6 is Perform AEEMD decomposition:

[0099] Taking the second-order decomposition as an example, the new characteristic signal Perform AEEMD decomposition to obtain the second-order IMF mean component c2(t), and calculate the margin r2(t) of the second-order IMF mean component c2(t) after decomposition, and decompose the added Gaussian white noise signal n i (t), and the IMF component group D2(n i (t)), the margin r2(t) and multi-order IMF components are used to form the IMF component group D2(n i (t)) Construct a new feature signal x new2 (t);

[0100]

[0101] r2(t)=x i (t)-r1(t)

[0102]

[0103]

[0104] By analogy, we can get the j-order IMF mean component c j (t), construct a new feature signal x newj (t).

[0105]

[0106]

[0107]

[0108] r j+l (t) = x i (t)-c j+1 (t)

[0109] It should be noted that the preset decomposition iteration stopping condition is that the standard deviation value between two adjacent IMF component signals is within a preset standard deviation value interval. For example, the standard deviation value between the first-order IMF component c1(t) and the second-order IMF component c2(t) is within a preset standard deviation interval, and the preset standard deviation interval is (0.2, 0.3).

[0110] It is worth mentioning that the AEEMD algorithm cleverly exploits the uniform transmission characteristics of the white noise spectrum. This allows the signal to be projected onto the white noise background and propagated throughout the entire time-frequency space. The margin of the AEEMD algorithm does not rely on the independent decomposition after each noise addition. Instead, it only considers the margin of the last decomposition. By analyzing the corresponding noise performance evaluation indicators, it determines whether the IMF components at each stage contain noise or signal. The noise components are then removed, and this cycle is repeated to ultimately obtain a de-noised signal, thus achieving adaptive decomposition of the partial discharge signal.

[0111] Furthermore, the preset decomposition iteration stopping condition is that the standard deviation value between the IMF mean components in two adjacent noise reduction feature signals is within a preset standard deviation value interval;

[0112] In a specific implementation, to facilitate the implementation of the method, the above process can be converted into a formula encapsulation form, where the standard deviation value can be calculated as follows:

[0113]

[0114] In the formula, SD represents the standard deviation, C j(t) represents the j-order IMF mean component, C j+1 (t) represents the j+1 order IMF mean component;

[0115] It should be noted that the iteration stopping condition is SD∈(δ1,δ2), δ1 is 0.2, and δ2 is 0.3.

[0116] Furthermore, reconstructing the multiple noise reduction feature signals to generate the target feature signal specifically includes the following sub-steps:

[0117] S131 , sorting the IMF mean components in the noise reduction feature signal from high to low according to their orders.

[0118] S132. Calculate the IMF correlation coefficient of each IMF mean component.

[0119] S133: Eliminate the noise reduction feature signals associated with the IMF correlation coefficients that meet the preset elimination conditions to generate target feature signals.

[0120] It should be noted that the preset elimination condition is specifically that the IMF correlation coefficient is less than the preset coefficient threshold.

[0121] In this embodiment, since the decomposition is performed by the adaptive set empirical mode decomposition algorithm, a plurality of noise reduction feature signals corresponding to the feature signals after each wavelet transform are obtained. Therefore, it is necessary to sort the noise reduction feature signals from high to low according to the order of the IMF mean component in each noise reduction feature signal to obtain a noise reduction feature signal group; then the IMF correlation coefficient of the IMF mean component associated with each noise reduction feature signal in the noise reduction feature signal group is calculated, and then the noise reduction feature signals associated with the IMF correlation coefficient that meets the elimination condition are eliminated, and the remaining noise reduction feature signals in the noise reduction feature signal group constitute the target feature signal.

[0122] In a specific implementation, to facilitate the implementation of the method, the above process can be converted into a formula encapsulation form, where the IMF correlation coefficient can be calculated as follows:

[0123]

[0124] Where r j represents the IMF correlation coefficient of the j-order IMF mean component, t represents the signal sampling point, t=1...m, x(t) represents the original signal, x represents the mean of the original signal, represents the mean of the IMF mean component;

[0125] In a specific implementation, to facilitate the implementation of the method, the above process can be converted into a formula encapsulation form, wherein the calculation method of the preset coefficient threshold can be as follows:

[0126]

[0127] Where TH represents the preset coefficient threshold, r represents the mean of the IMF correlation coefficient of the IMF mean component, j represents the order of the IMF mean component, j = 1...n;

[0128] When r j <TH, then there is a noise signal in the j-order IMF mean component.

[0129] See also Figure 4-Figure 7 ,The following comparison is made by using the wavelet denoising algorithm, the EEMD denoising algorithm and the AEEMD denoising algorithm to process the ,feature signal;

[0130] Figure 4 is the collected original signal waveform, from Figure 5 It can be seen that the wavelet denoising method has basically completed the denoising of the partial discharge signal, but there are still some noise glitches after denoising. Figure 6 and Figure 7 It can be seen that both the EEMD-based denoising method and the AEEMD-based denoising method have successfully denoised the partial discharge signal and have largely suppressed the generation of noise glitches after signal denoising. However, after comparison, it can be found that Figure 7 The treatment effect is better than Figure 6 In general, the above figure shows the feasibility and effectiveness of the three methods for denoising partial discharge signals. All three methods process the narrowband interference signal and white noise of the partial discharge signal. In comparison, the measurement graph after denoising by the AEEMD algorithm is basically consistent with the measured original signal graph, which also reflects the superiority of the AEEMD algorithm.

[0131] See also Figure 8-Figure 9 , Figure 8 This is a structural diagram of a switch cabinet internal insulation partial discharge monitoring system provided by an embodiment of the present invention.

[0132] The present invention provides a switch cabinet internal insulation partial discharge monitoring system, comprising a sensor module, a partial discharge monitoring module and a computer terminal that are communicatively connected in sequence;

[0133] A sensor module, used for collecting characteristic signals;

[0134] The partial discharge monitoring module is used to pre-process characteristic signals, build target partial discharge monitoring models, monitor partial discharges, and display monitoring results;

[0135] Computer terminal for outputting monitoring reports.

[0136] The partial discharge monitoring module includes an oscilloscope and a partial discharge monitor that are communicatively connected to each other;

[0137] Partial discharge monitor, used for pre-processing characteristic signals, building target partial discharge monitoring models, and partial discharge monitoring;

[0138] Oscilloscope, used to display monitoring characteristic signals and monitoring results.

[0139] In an embodiment of the present invention, the sensor module includes a transient ground voltage sensor and an ultrasonic sensor;

[0140] The relevant technical parameters of the transient ground voltage sensor are as follows Table 1

[0141] type Monitoring bandwidth Measuring range Installation Capacitive 3~100MHz 40~60dBmv Magnetic adsorption

[0142] The relevant technical parameters of ultrasonic sensors are shown in Table 2.

[0143] Center frequency Sensitivity Operating temperature Monitoring range Output format 50±1.0KHz 78±3dBmin 20～80℃ 0.2～15m Differential output

[0144] TEV and ultrasonic sensors were selected for signal acquisition. The TEV sensor was a capacitive sensor with a monitoring bandwidth of 3 to 100 MHz and a measurement range of 40 to 60 dB mV. It was mounted using magnetic adsorption. Ultrasonic sensor parameters included a center frequency of 50 ± 1.0 kHz, an operating temperature of 20 to 80°C, a measurement range of 0.2 to 15 m, a sensitivity of 78 ± 3 dB min, and a differential output signal.

[0145] It's worth mentioning that the system also includes a power module. Because the switchgear is often exposed to high voltages, such as 10kV and 35kV, when receiving and distributing electrical energy, the power module is required to be able to output an AC voltage source of 5kV to 20kV.

[0146] Furthermore, the sensor module and the partial discharge monitoring module are communicatively connected via an RS485 communication circuit.

[0147] It should be noted that the sensor module and the partial discharge monitoring module are connected through the RS485 communication circuit. The RS485 communication circuit adopts balanced transmission and differential reception technology, which can reduce the interference of external environmental factors on signal transmission and has high sensitivity.

[0148] See also Figure 9 The core component of the RS485 communication circuit is the RS485 chip, which mainly has the following pins to realize the conversion and transmission of communication signals.

[0149] RO: receiver output;

[0150] RE: Receiver output enable (active low);

[0151] DE: Driver output enable (high level valid);

[0152] DI: driver input;

[0153] GND: connect to ground;

[0154] A: Driver output / receiver input (in phase);

[0155] B: Driver output / receiver input (inverting);

[0156] VCC: chip power supply;

[0157] TXD: TXD (Transmit Data) is the pin for sending data;

[0158] RXD: RXD (Receive Data) is the pin for receiving data;

[0159] DC: DC power interface;

[0160] PG: Power good pin;

[0161] FS: fuse that limits current;

[0162] P6KE6.8CA TVS: Transient Voltage Suppressor Diode.

[0163] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0164] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0165] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0166] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for monitoring partial discharge of insulation inside a switch cabinet, characterized in that: include: Preprocessing the sample characteristic signals received from the switchgear to generate a sample set; Using the sample set as input to a preset initial partial discharge monitoring model for training, to construct a target partial discharge monitoring model; Preprocessing the received monitoring characteristic signal, and monitoring the preprocessed monitoring characteristic signal through the target partial discharge monitoring model to generate a monitoring report; The pretreatment includes: Perform wavelet transform on each characteristic signal; Performing noise reduction on the characteristic signal after the wavelet transformation to generate a plurality of noise-reduced characteristic signals; Reconstructing the plurality of noise reduction characteristic signals to generate a target characteristic signal; The step of performing noise reduction on the characteristic signal after the wavelet transformation to generate a plurality of noise-reduced characteristic signals includes: injecting white noise into the characteristic signal after wavelet transformation to generate a white noise characteristic signal; Decomposing the white noise characteristic signal to obtain a plurality of noise reduction characteristic signals when a preset decomposition iteration stop condition is met; The characteristic signal after the wavelet transformation includes a transient voltage signal and an ultrasonic signal, and the transient voltage signal and the ultrasonic signal are decomposed using an adaptive ensemble empirical mode decomposition algorithm. The process of using the adaptive ensemble empirical mode decomposition algorithm to reduce the noise of the characteristic signal after the wavelet transformation is as follows: Step 1. Input the original signal , that is, transient voltage signal or ultrasonic signal; Step 2: Add the original signal Add Gaussian white noise signal , get the noise cancellation signal ; ； Step 3: Eliminate the noise signal Perform AEEMD decomposition to obtain the first-order IMF component ; ； Among them, get the original signal All the maximum points of , the maximum envelope is fitted by the cubic spline function ; Similarly, find the original signal All the minimum points of the signal are fitted with the minimum envelope of the signal through the cubic spline function , express The average value of the maximum envelope and the minimum envelope of ; ； Since the number of times Gaussian white noise is added is N, the first-order IMF component of AEEMD decomposition after adding Gaussian white noise is Therefore, the N first-order IMF components obtained by AEEMD decomposition after adding Gaussian white noise N times are averaged, and the results are as follows: ； Step 4: Calculate the first-order IMF mean component after decomposition Margin ; ； Step 5: Decompose the added Gaussian white noise signal , i=1, 2, ...N, and the IMF component group after empirical mode decomposition is obtained by multi-order IMF components ; Defining a function Represents a set of IMF components after the empirical mode decomposition of the signal, where By decomposition The noise in obtains the first-order IMF component; ； Step 6: Use margin and IMF component groups , construct a new feature signal , and adopt new characteristic signals , jump to step 3-step 5 until the preset decomposition iteration stop condition is met, and multiple noise reduction feature signals are obtained to ; 。 2. The method for monitoring partial discharge of internal insulation of a switch cabinet according to claim 1, characterized in that: The preset decomposition iteration stopping condition is that the standard deviation value between the IMF mean components in two adjacent noise reduction feature signals is within a preset standard deviation value range.

3. The method for monitoring partial discharge of internal insulation of a switch cabinet according to claim 1, characterized in that: The reconstructing the plurality of noise reduction characteristic signals to generate a target characteristic signal includes: Sort the noise reduction feature signals according to the order of the IMF mean components from high to low; Calculating the IMF correlation coefficient of each of the IMF mean components; The noise reduction feature signal associated with the IMF correlation coefficient that meets the preset elimination condition is eliminated to generate a target feature signal.

4. The method for monitoring partial discharge of internal insulation of a switch cabinet according to claim 3, characterized in that: The preset elimination condition is specifically that the IMF correlation coefficient is less than a preset coefficient threshold.

5. The method for monitoring partial discharge of internal insulation of a switch cabinet according to claim 1, characterized in that: A genetic algorithm is used to optimize the parameters of the preset initial partial discharge monitoring model.

6. A switch cabinet internal insulation partial discharge monitoring system, characterized in that: Executing a method for monitoring partial discharge of internal insulation of a switch cabinet as claimed in claim 1, comprising a sensor module, a partial discharge monitoring module and a computer terminal that are communicatively connected in sequence; The sensor module is used to collect characteristic signals; The partial discharge monitoring module is used to pre-process the characteristic signal, build a target partial discharge monitoring model, perform partial discharge monitoring, and display monitoring results; The computer terminal is used to output the monitoring report.

7. The switch cabinet internal insulation partial discharge monitoring system according to claim 6, characterized in that: The partial discharge monitoring module includes an oscilloscope and a partial discharge monitor that are communicatively connected to each other; The partial discharge monitor is used to pre-process the characteristic signal, construct a target partial discharge monitoring model, and perform partial discharge monitoring; The oscilloscope is used to display monitoring characteristic signals and monitoring results.

8. The switch cabinet internal insulation partial discharge monitoring system according to claim 7, characterized in that: The sensor module and the partial discharge monitoring module are communicatively connected via an RS485 communication circuit.

Images