Circuit system for measuring thyristors and capacitors in series inverter circuits

Abstract

Disclosed is a circuit system for measuring thyristors and capacitors in a series inverter circuit. It includes: a power supply circuit; a measuring circuit, wherein the measuring circuit includes a signal processor and an amplifier circuit powered by the power supply circuit; the signal processor is used to transmit a pulse detection signal, and the amplifier circuit is used to amplify the captured pulse detection signal to obtain an amplified pulse detection signal and transmit it back to the signal processor; wherein the signal processor is used to determine the fault type based on the amplified pulse detection signal. In addition, an intelligent algorithm is further introduced into the signal processor to extract waveform semantic features based on feature multi-scale scanning of the amplified pulse detection signal in the circuit, and the full time domain features of the signal waveform are represented through a self-supervision mechanism, so as to realize intelligent fault diagnosis.

Description

Technical Field

[0001] The present application relates to the field of measuring circuits, and more specifically, to a circuit system for measuring thyristors and capacitors in a series inverter circuit. Background Art

[0002] The series inverter half-bridge medium frequency power supply can improve production efficiency with its high power factor and energy saving advantages. However, in the prior art, the measurement and fault diagnosis of damaged thyristors and capacitors are usually performed manually by technicians.

[0003] This method has the following problems: due to the compact design of the equipment, it needs to be disassembled one by one and measured with a multimeter. Detecting damaged thyristors and capacitors is time-consuming and laborious, and lacks intelligence.

[0004] Therefore, an optimized circuit system for measuring thyristors and capacitors in a series inverter circuit is desired. Summary of the invention

[0005] In view of this, the present application proposes a circuit system for measuring thyristors and capacitors in a series inverter circuit.

[0006] According to one aspect of the present application, a circuit system for measuring a thyristor and a capacitor in a series inverter circuit is provided, comprising:

[0007] Power supply circuit;

[0008] A measuring circuit, wherein the measuring circuit comprises a signal processor and an amplifying circuit powered by the power supply circuit; the signal processor is used to transmit a pulse detection signal, and the amplifying circuit is used to amplify the captured pulse detection signal to obtain an amplified pulse detection signal and transmit it back to the signal processor;

[0009] Wherein, the signal processor is used to determine the fault type based on the amplified pulse detection signal.

[0010] In the above circuit system for measuring the thyristor and capacitor in the series inverter circuit, the signal processor comprises:

[0011] A signal segmentation module, used for performing signal segmentation on the amplified pulse detection signal to obtain a sequence of pulse detection signal segments;

[0012] A multi-scale waveform semantic feature extraction module, used for performing multi-scale waveform semantic feature extraction on the sequence of pulse detection signal segments to obtain a sequence of pulse detection signal segment waveform semantic feature vectors;

[0013] A fusion module, used for inputting the sequence of semantic feature vectors of the pulse detection signal segment waveform into a self-supervised feature sequence significant fusion module to obtain a significant fusion representation vector of the full time domain waveform of the pulse detection signal;

[0014] A fault diagnosis module is used to determine a fault diagnosis result based on a significant fusion representation vector of the full time domain waveform of the pulse detection signal.

[0015] In the above-mentioned circuit system for measuring thyristors and capacitors in a series inverter circuit, the multi-scale waveform semantic feature extraction module includes:

[0016] A signal waveform feature extraction unit, used for passing each pulse detection signal segment in the sequence of pulse detection signal segments through a signal waveform feature extractor based on a dilated convolutional neural network model to obtain a sequence of pulse detection signal segment waveform semantic feature maps;

[0017] The multi-scale encoding unit is used to input each pulse detection signal segment waveform semantic feature map in the sequence of pulse detection signal segment waveform semantic feature maps into a feature multi-scale scanning module to obtain a sequence of pulse detection signal segment waveform semantic feature vectors.

[0018] In the above circuit system for measuring thyristors and capacitors in a series inverter circuit, the multi-scale encoding unit comprises:

[0019] A semantic feature extraction subunit, used for performing signal segment global semantic feature extraction and signal segment local semantic feature extraction on the pulse detection signal segment waveform semantic feature graph to obtain a signal segment waveform global semantic feature vector and a signal segment waveform local semantic feature vector;

[0020] A position-by-position element-by-element addition subunit is used to calculate the position-by-element addition of the global semantic feature vector of the signal segment waveform and the local semantic feature vector of the signal segment waveform to obtain a multi-scale semantic fusion feature vector of the signal segment waveform;

[0021] The activation subunit is used to perform activation processing on the multi-scale semantic fusion feature vector of the signal segment waveform based on the sigmoid function to obtain the semantic feature vector of the pulse detection signal segment waveform.

[0022] In the above circuit system for measuring thyristors and capacitors in a series inverter circuit, the semantic feature extraction subunit is used to:

[0023] Upsampling the pulse detection signal segment waveform semantic feature map to obtain an upsampled pulse detection signal segment waveform semantic feature map;

[0024] Using a signal segment global semantic feature scanner to process the upsampled pulse detection signal segment waveform semantic feature map to obtain a signal segment waveform global scanning semantic feature vector;

[0025] Performing dimension adjustment on the signal segment waveform global scanning semantic feature vector to obtain the signal segment waveform global semantic feature vector;

[0026] Processing the upsampled pulse detection signal segment waveform semantic feature map using a signal segment local semantic feature scanner to obtain a signal segment waveform local scanning semantic feature vector;

[0027] The dimension of the local scanning semantic feature vector of the signal segment waveform is adjusted to obtain the local semantic feature vector of the signal segment waveform.

[0028] In the above-mentioned circuit system for measuring thyristors and capacitors in a series inverter circuit, the signal segment global semantic feature scanner includes a global average pooling layer, a first point convolution layer, a first batch normalization layer and a first activation layer, and the signal segment local semantic feature scanner includes a second point convolution layer, a second batch normalization layer and a second activation layer.

[0029] In the above circuit system for measuring thyristors and capacitors in a series inverter circuit, the fusion module comprises:

[0030] A semantically significant correlation evaluation unit, used for performing a semantically significant correlation evaluation based on self-clustering on the sequence of semantic feature vectors of the pulse detection signal segment waveform to obtain a sequence of self-supervisory signal segment correlation weights;

[0031] A weighting unit is used to calculate the vector-by-vector weighted sum of the sequence of semantic feature vectors of the pulse detection signal segment waveform using the sequence of the self-supervisory signal segment correlation weights as the weight sequence to obtain a significant fusion representation vector of the full time domain waveform of the pulse detection signal.

[0032] In the above-mentioned circuit system for measuring thyristors and capacitors in a series inverter circuit, the semantically significant correlation assessment unit comprises:

[0033] A mean calculation subunit, used for calculating the mean vector of the sequence of semantic feature vectors of the pulse detection signal segment waveform as the cluster center of the semantic distribution of the pulse detection signal segment sequence;

[0034] A waveform semantic matching association subunit, used for calculating the waveform semantic matching association degree between each pulse detection signal segment waveform semantic feature vector in the sequence of the pulse detection signal segment waveform semantic feature vector and the pulse detection signal segment sequence semantic distribution cluster center to obtain a sequence of waveform semantic matching association degrees;

[0035] The normalization subunit is used to input the sequence of waveform semantic matching associations into a softmax function for soft maximum normalization processing to obtain a sequence of self-supervisory signal segment correlation weights.

[0036] In the above circuit system for measuring thyristors and capacitors in a series inverter circuit, the waveform semantic matching association subunit is used to:

[0037] Encoding the pulse detection signal segment waveform semantic feature vector through a fully connected layer to obtain a pulse detection signal segment waveform semantic encoding feature vector;

[0038] The transposed vector of the pulse detection signal segment sequence semantic distribution cluster center is multiplied by the pulse detection signal segment waveform semantic coding feature vector to obtain the waveform semantic matching association degree.

[0039] In the above circuit system for measuring thyristors and capacitors in a series inverter circuit, the fault diagnosis module is used to:

[0040] The full time domain waveform of the pulse detection signal is significantly fused into a representation vector and passed through a classifier-based fault diagnostic device to obtain the fault diagnosis result, which is used to indicate that the resonant capacitor is damaged or the inverter thyristor is damaged.

[0041] In this application, an intelligent algorithm is introduced into the signal processor to extract waveform semantic features of the amplified pulse detection signal in the circuit based on feature multi-scale scanning, and the full time domain features of the signal waveform are represented through a self-supervision mechanism, thereby realizing intelligent fault diagnosis. In this way, through the combination of circuit system design and intelligent algorithms, the traditional manual detection is replaced by an automated measurement and diagnosis process, thereby improving the intelligence level of measuring thyristors and capacitors in series inverter circuits.

[0042] Other features and aspects of the present application will become apparent from the following detailed description of the present application with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the present application and, together with the description, serve to explain the principles of the present application.

[0044] Figure 1 A block diagram of a circuit system for measuring a thyristor and a capacitor in a series inverter circuit according to an embodiment of the present application is shown.

[0045] Figure 2 A schematic diagram of a key circuit according to an embodiment of the present application is shown.

[0046] Figure 3 A schematic diagram of a debug port circuit according to an embodiment of the present application is shown.

[0047] Figure 4 A block diagram of the signal processor in the circuit system for measuring a thyristor and a capacitor in a series inverter circuit according to an embodiment of the present application is shown.

[0048] Figure 5 A block diagram of the multi-scale waveform semantic feature extraction module in a circuit system for measuring thyristors and capacitors in a series inverter circuit according to an embodiment of the present application is shown.

[0049] Figure 6 A block diagram of the multi-scale encoding unit in a circuit system for measuring thyristors and capacitors in a series inverter circuit according to an embodiment of the present application is shown.

[0050] Figure 7 A block diagram of the fusion module in a circuit system for measuring thyristors and capacitors in a series inverter circuit according to an embodiment of the present application is shown.

[0051] Figure 8 A block diagram of the semantically significant correlation assessment unit in a circuit system for measuring a thyristor and a capacitor in a series inverter circuit according to an embodiment of the present application is shown.

[0052] Fig. 9 The flowchart shows a method for measuring a thyristor and a capacitor in a series inverter circuit according to an embodiment of the present application.

[0053] Fig.10 A structural block diagram of a power supply circuit according to an embodiment of the present application is shown.

[0054] Fig.11 A schematic diagram of a charging circuit according to an embodiment of the present application is shown.

[0055] Fig.12 A schematic diagram of a power supply circuit according to an embodiment of the present application is shown.

[0056] Fig.13 A schematic diagram of an adjustable power supply according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0058] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0059] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0060] In addition, in order to better illustrate the present application, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present application can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present application.

[0061] The present application provides a circuit system for measuring thyristors and capacitors in a series inverter circuit, such as Figure 1 As shown, it includes: a power supply circuit 10; a measuring circuit 20, wherein the measuring circuit 20 includes a signal processor 100 and an amplifying circuit 200 powered by the power supply circuit 10; the signal processor 100 is used to transmit a pulse detection signal, and the amplifying circuit 200 is used to amplify the captured pulse detection signal to obtain an amplified pulse detection signal and transmit it back to the signal processor 100; wherein the signal processor 100 is used to determine the fault type based on the amplified pulse detection signal.

[0062] Furthermore, it also includes a key circuit 30 connected to the input end of the measurement circuit 20 and a debug port circuit 40 connected to the input end of the measurement circuit 20. Specifically, Figure 2 A schematic diagram of a key circuit according to an embodiment of the present application is shown. Figure 3 A schematic diagram of a debug port circuit according to an embodiment of the present application is shown.

[0063] Specifically, in view of the technical problem that the manual measurement of damaged thyristors and capacitors in the prior art is time-consuming and labor-intensive and lacks intelligence, the technical concept of this application is to introduce an intelligent algorithm into the signal processor, extract waveform semantic features based on feature multi-scale scanning of the amplified pulse detection signal in the circuit, and represent the full time domain features of the signal waveform through a self-supervision mechanism, so as to achieve intelligent fault diagnosis. In this way, through the combination of circuit system design and intelligent algorithms, the traditional manual detection is replaced by an automated measurement and diagnosis process, thereby improving the intelligence level of measuring thyristors and capacitors in series inverter circuits.

[0064] Accordingly, if Figure 4 As shown, the signal processor 100 includes: a signal segmentation module 110, used to perform signal segmentation on the amplified pulse detection signal to obtain a sequence of pulse detection signal segments; a multi-scale waveform semantic feature extraction module 120, used to perform multi-scale waveform semantic feature extraction on the sequence of pulse detection signal segments to obtain a sequence of pulse detection signal segment waveform semantic feature vectors; a fusion module 130, used to input the sequence of pulse detection signal segment waveform semantic feature vectors into a self-supervised feature sequence significant fusion module to obtain a pulse detection signal full time domain waveform significant fusion representation vector; a fault diagnosis module 140, used to determine a fault diagnosis result based on the pulse detection signal full time domain waveform significant fusion representation vector.

[0065] Based on this, in the technical solution of the present application, the specific implementation process of the signal processor includes: first, performing signal segmentation on the amplified pulse detection signal to obtain a sequence of pulse detection signal segments. Here, the entire amplified pulse detection signal can be decomposed into signal segments with smaller time spans by signal segmentation, guiding the model to perform a more detailed analysis of each local signal segment, so that local anomalies that may not be obvious in the overall signal can be more easily captured.

[0066] Then, each pulse detection signal segment in the sequence of pulse detection signal segments is passed through a signal waveform feature extractor based on a dilated convolutional neural network model to obtain a sequence of pulse detection signal segment waveform semantic feature maps. Here, compared with the traditional convolutional neural network, the dilated convolution can increase the receptive field without increasing the number of parameters, which means that a wider range of contextual information can be extracted while maintaining computational efficiency, which helps to capture richer fault features in each pulse detection signal segment.

[0067] Take into account that each pulse detection signal segment may have different waveform feature distribution patterns at different scales (such as time scales). Specifically, the pulse detection signal may have a transient response behavior, and may also contain a long-term complex waveform pattern with periodicity. In order to be able to capture this multi-scale waveform semantic feature, in the technical solution of the present application, each pulse detection signal segment waveform semantic feature map in the sequence of the pulse detection signal segment waveform semantic feature map is further input into the feature multi-scale scanning module to obtain a sequence of pulse detection signal segment waveform semantic feature vectors. In the technical solution of the present application, the core of the feature multi-scale scanning module being able to capture the multi-scale waveform semantic features is that it designs a signal segment global semantic feature scanner and a signal segment local semantic feature scanner to respectively capture the multi-scale waveform semantic features in each of the pulse detection signal segment waveform semantic feature maps.

[0068] In particular, in an embodiment of the present application, the difference between the signal segment global semantic feature scanner and the signal segment local semantic feature scanner is that the signal segment global semantic feature scanner includes a global average pooling layer, which can refine the global waveform feature information in the pulse detection signal segment waveform semantic feature map to retain such global waveform feature information.

[0069] Accordingly, if Figure 5 As shown, the multi-scale waveform semantic feature extraction module 120 includes: a signal waveform feature extraction unit 121, which is used to pass each pulse detection signal segment in the sequence of pulse detection signal segments through a signal waveform feature extractor based on a void convolutional neural network model to obtain a sequence of pulse detection signal segment waveform semantic feature maps; a multi-scale encoding unit 122, which is used to input each pulse detection signal segment waveform semantic feature map in the sequence of pulse detection signal segment waveform semantic feature maps into a feature multi-scale scanning module to obtain a sequence of pulse detection signal segment waveform semantic feature vectors.

[0070] Among them, Figure 6As shown, the multi-scale encoding unit 122 includes: a semantic feature extraction subunit 1221, which is used to perform signal segment global semantic feature extraction and signal segment local semantic feature extraction on the pulse detection signal segment waveform semantic feature map to obtain a signal segment waveform global semantic feature vector and a signal segment waveform local semantic feature vector; a position-based element-by-element addition subunit 1222, which is used to calculate the position-based element-by-element addition of the signal segment waveform global semantic feature vector and the signal segment waveform local semantic feature vector to obtain a signal segment waveform multi-scale semantic fusion feature vector; an activation subunit 1223, which is used to perform activation processing based on a sigmoid function on the signal segment waveform multi-scale semantic fusion feature vector to obtain the pulse detection signal segment waveform semantic feature vector.

[0071] More specifically, the semantic feature extraction subunit 1221 is used to: upsample the pulse detection signal segment waveform semantic feature map to obtain an upsampled pulse detection signal segment waveform semantic feature map; use a signal segment global semantic feature scanner to process the upsampled pulse detection signal segment waveform semantic feature map to obtain a signal segment waveform global scanning semantic feature vector; adjust the dimension of the signal segment waveform global scanning semantic feature vector to obtain the signal segment waveform global semantic feature vector; use a signal segment local semantic feature scanner to process the upsampled pulse detection signal segment waveform semantic feature map to obtain a signal segment waveform local scanning semantic feature vector; adjust the dimension of the signal segment waveform local scanning semantic feature vector to obtain the signal segment waveform local semantic feature vector. Wherein, the signal segment global semantic feature scanner includes a global average pooling layer, a first point convolution layer, a first batch normalization layer and a first activation layer, and the signal segment local semantic feature scanner includes a second point convolution layer, a second batch normalization layer and a second activation layer.

[0072] Taking into account the way in which the amplified pulse detection signal is subjected to signal segmentation, hole convolution processing, and processing through a feature multi-scale scanning module, the multi-scale signal segment waveform semantic features in each pulse detection signal segment can be deeply characterized and portrayed. However, since each pulse detection signal segment does not exist independently, if only the waveform feature representation of the local signal is considered, and the mutual correlation between the local signals is ignored, it may not be possible to fully understand the overall behavior and waveform distribution pattern of the amplified pulse detection signal, thereby affecting the ability to capture systematic fault characteristics. Therefore, in the technical solution of the present application, the sequence of the pulse detection signal segment waveform semantic feature vectors is input into the self-supervised feature sequence significant fusion module to obtain the pulse detection signal full time domain waveform significant fusion representation vector.

[0073] In a specific example of the present application, the implementation process of inputting the sequence of semantic feature vectors of the pulse detection signal segment waveform into the self-supervised feature sequence significant fusion module to obtain the significant fusion representation vector of the pulse detection signal full time domain waveform, includes: performing a self-clustering-based semantic significant correlation assessment on the sequence of semantic feature vectors of the pulse detection signal segment waveform to obtain a sequence of self-supervised signal segment correlation weights, and then using the sequence of self-supervised signal segment correlation weights as a weight sequence to calculate the vector-by-vector weighted sum of the sequence of semantic feature vectors of the pulse detection signal segment waveform to obtain the significant fusion representation vector of the pulse detection signal full time domain waveform.

[0074] Among them, the process of calculating the self-supervisory signal segment correlation weight is to first learn the local signal segment semantic features of each pulse detection signal segment waveform semantic feature vector through a fully connected layer. Specifically, this local signal segment semantic feature is automatically learned through a (trained) weight matrix and a (trained) bias vector. Then, the transposed vector of the pulse detection signal segment sequence semantic distribution cluster center (for example, the mean vector of the sequence of the pulse detection signal segment waveform semantic feature vector) is multiplied with the local signal segment semantic features of each pulse detection signal segment waveform semantic feature vector to characterize and model the complex association relationship and correlation between the overall clustered waveform semantic features of the amplified pulse detection signal and the semantic features of each local signal segment, thereby obtaining a sequence of waveform semantic matching correlations. The overall correlation of each waveform semantic matching correlation is quantified using soft maximum normalization processing to obtain a sequence of self-supervisory signal segment correlation weights. Then, the sequence of self-supervisory signal segment correlation weights is distributed to each pulse detection signal segment waveform semantic feature vector through weighted addition processing to achieve self-supervised full-time domain waveform feature fusion.

[0075] Accordingly, if Figure 7 As shown, the fusion module 130 includes: a semantically significant correlation assessment unit 131, which is used to perform a semantically significant correlation assessment on the sequence of semantic feature vectors of the pulse detection signal segment waveform based on self-clustering to obtain a sequence of self-supervisory signal segment correlation weights; a weighting unit 132, which is used to calculate the vector-by-vector weighted sum of the sequence of semantic feature vectors of the pulse detection signal segment waveform using the sequence of self-supervisory signal segment correlation weights as a weight sequence to obtain a significant fusion representation vector of the pulse detection signal full time domain waveform.

[0076] Among them, Figure 8As shown, the semantic significant correlation assessment unit 131 includes: a mean calculation subunit 1311, which is used to calculate the mean vector of the sequence of waveform semantic feature vectors of the pulse detection signal segment as the semantic distribution cluster center of the pulse detection signal segment sequence; a waveform semantic matching association subunit 1312, which is used to calculate the waveform semantic matching association between each pulse detection signal segment waveform semantic feature vector in the sequence of waveform semantic feature vectors of the pulse detection signal segment and the semantic distribution cluster center of the pulse detection signal segment sequence to obtain a sequence of waveform semantic matching associations; a normalization subunit 1313, which is used to input the sequence of waveform semantic matching associations into a softmax function for normalization processing based on soft maximum value to obtain a sequence of self-supervised signal segment correlation weights.

[0077] Specifically, the waveform semantic matching association subunit 1312 is used to: encode the pulse detection signal segment waveform semantic feature vector through a fully connected layer to obtain the pulse detection signal segment waveform semantic coding feature vector; multiply the transposed vector of the pulse detection signal segment sequence semantic distribution cluster center with the pulse detection signal segment waveform semantic coding feature vector to obtain the waveform semantic matching association degree.

[0078] In one example, the fusion module 130 is used to: input the sequence of semantic feature vectors of the pulse detection signal segment waveform into the self-supervised feature sequence significant fusion module for processing using the following self-supervised feature sequence significant fusion formula to obtain the pulse detection signal full time domain waveform significant fusion representation vector; wherein the self-supervised feature sequence significant fusion formula is:

[0079]

[0080] Among them, v i is the semantic feature vector of the waveform of the ith pulse detection signal segment, L is the number of semantic feature vectors of the waveform of the pulse detection signal segment, v r is the semantic distribution cluster center of the pulse detection signal segment sequence, is the transposed vector of the semantic distribution cluster center of the pulse detection signal segment sequence, sigmoid is the sigmoid function, W i and B i are the i-th weight matrix and the i-th bias vector, respectively, e i is the semantic matching association degree of the ith waveform, softmax is the softmax function, * represents dot product, and A is the significant fusion representation vector of the full time domain waveform of the pulse detection signal.

[0081] Furthermore, the full time domain waveform of the pulse detection signal is significantly fused into a representation vector and passed through a classifier-based fault diagnoser to obtain a fault diagnosis result, and the fault diagnosis result is used to indicate that the resonant capacitor is damaged or the inverter thyristor is damaged.

[0082] Correspondingly, the fault diagnosis module 140 is used to: significantly fuse the representation vector of the full time domain waveform of the pulse detection signal through a classifier-based fault diagnoser to obtain the fault diagnosis result, and the fault diagnosis result is used to indicate that the resonant capacitor is damaged or the inverter thyristor is damaged.

[0083] It should be understood that the role of the classifier is to use the given categories and known training data to learn classification rules and classifiers, and then classify (or predict) unknown data. Logistic regression and SVM are often used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are required to form a multi-classification. However, this is prone to errors and is not efficient. Commonly used multi-classification methods include the Softmax classification function.

[0084] Furthermore, in the technical solution of the present application, the signal processor also includes a training module for training the signal waveform feature extractor based on the void convolutional neural network model, the feature multi-scale scanning module, the self-supervised feature sequence significant fusion module and the classifier-based fault diagnostic device; wherein the training module includes: a training data acquisition unit for acquiring training data, the training data including the training amplified pulse detection signal and the real value of the resonant capacitor damage or the inverter thyristor damage; a training signal segmentation unit for performing signal segmentation on the training amplified pulse detection signal to obtain a sequence of training pulse detection signal segments; a training signal waveform feature extraction unit for passing each training pulse detection signal segment in the sequence of the training pulse detection signal segments through the signal waveform feature extractor based on the void convolutional neural network model to obtain a sequence of training pulse detection signal segment waveform semantic feature maps; a training multi-scale encoding unit for encoding each training pulse detection signal in the sequence of the training pulse detection signal segment waveform semantic feature maps. The semantic feature maps of the segment waveforms are respectively input into the feature multi-scale scanning module to obtain a sequence of semantic feature vectors of the segment waveforms of the training pulse detection signal; a training fusion unit is used to input the sequence of semantic feature vectors of the segment waveforms of the training pulse detection signal into the self-supervised feature sequence significant fusion module to obtain a significant fusion representation vector of the full time domain waveform of the training pulse detection signal; a loss function value calculation unit is used to pass the significant fusion representation vector of the full time domain waveform of the training pulse detection signal through the classifier-based fault diagnostic device to obtain a resonant capacitor damage probability value corresponding to the resonant capacitor damage, an inverter thyristor damage probability value corresponding to the inverter thyristor damage, and a classification loss function value; a loss training unit is used to train the signal waveform feature extractor based on the void convolutional neural network model, the feature multi-scale scanning module, the self-supervised feature sequence significant fusion module and the classifier-based fault diagnostic device with the classification loss function value, wherein in each round of iteration of the training, the significant fusion representation vector of the full time domain waveform of the training pulse detection signal is optimized and iterated.

[0085] When the local image semantic space domain self-supervised significant fusion is performed on the multi-scale image semantic features of the training amplified pulse detection signal in the local image semantic space domain for the significant fusion representation vector of the full time domain waveform of the training pulse detection signal in the global image semantic space, due to the distribution self-supervised fusion difference of the image semantic features in the local-global image semantic space domain, the local distribution difference of the significant fusion representation vector of the full time domain waveform of the training pulse detection signal is caused, which affects the classification iteration effect.

[0086] In a preferred embodiment, based on the resonant capacitor damage probability value corresponding to the resonant capacitor damage and the inverter thyristor damage probability value corresponding to the inverter thyristor damage, the training pulse detection signal full time domain waveform significant fusion representation vector is optimized and iterated. Wherein, the specific process of performing optimization iteration includes the following steps: determining the first eigenvalue and the second eigenvalue of any two different positions of the training pulse detection signal full time domain waveform significant fusion representation vector; dividing the resonant capacitor damage probability value by the inverter thyristor damage probability value to obtain the pulse detection signal full time domain waveform significant fusion relative class probability distribution value; multiplying the resonant capacitor damage probability value by the inverter thyristor probability value to obtain the pulse detection signal full time domain waveform significant fusion collaborative class probability distribution value; subtracting the pulse detection signal full time domain waveform significant fusion relative class probability distribution value from the first eigenvalue and the second eigenvalue respectively, and then multiplying them to obtain the pulse detection signal full time domain waveform significant fusion relative joint representation; multiplying the first eigenvalue and the The second eigenvalues ​​are added and then divided by the probability distribution value of the pulse detection signal full time domain waveform significant fusion collaborative class to obtain the pulse detection signal full time domain waveform significant fusion system response representation; the pulse detection signal full time domain waveform significant fusion relative joint representation and the pulse detection signal full time domain waveform significant fusion system response representation are added and used as the matrix value at the coordinates corresponding to the different positions of the first eigenvalue and the second eigenvalue to obtain the pulse detection signal full time domain waveform significant fusion correction matrix; the inner product of each row vector of the pulse detection signal full time domain waveform significant fusion correction matrix and the training pulse detection signal full time domain waveform significant fusion representation vector is calculated to obtain the optimized training pulse detection signal full time domain waveform significant fusion representation vector.

[0087] That is, the above preferred embodiment takes the eigenvalue of the training pulse detection signal full time domain waveform significant fusion representation vector as a unit, and performs a relativity-based joint characterization and a synergy-based response characterization on the eigenvalue pairs of the training pulse detection signal full time domain waveform significant fusion representation vector as a feature set, and obtains the relative class probability distribution form and the synergy-based probability distribution form of the resonant capacitor damage class probability and the inverter thyristor damage class probability through a classifier-based fault diagnostic device, so as to avoid the class induction bias caused by the corresponding local distribution differences between the eigenvalues ​​of the training pulse detection signal full time domain waveform significant fusion representation vector, so as to establish a robust understanding paradigm for class cognition of the training pulse detection signal full time domain waveform significant fusion representation vector as a feature set as a whole, thereby improving the iteration effect of the training pulse detection signal full time domain waveform significant fusion representation vector as a feature set in the classification process, that is, improving the classification training speed and the accuracy of the classification results of the training pulse detection signal full time domain waveform significant fusion representation vector when it is classified through a classifier-based fault diagnostic device.

[0088] As described above, the circuit system for measuring thyristors and capacitors in a series inverter circuit according to the embodiment of the present application can be implemented in various terminal devices, such as a server having a measurement algorithm for thyristors and capacitors in a series inverter circuit. In one example, the circuit system for measuring thyristors and capacitors in a series inverter circuit can be integrated into the terminal device as a software module and / or a hardware module. For example, the circuit system for measuring thyristors and capacitors in a series inverter circuit can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the circuit system for measuring thyristors and capacitors in a series inverter circuit can also be one of the many hardware modules of the terminal device.

[0089] Alternatively, in another example, the circuit system for measuring thyristors and capacitors in the series inverter circuit and the terminal device may also be discrete devices, and the circuit system for measuring thyristors and capacitors in the series inverter circuit may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0090] Fig. 9 A flow chart showing a method for measuring a thyristor and a capacitor in a series inverter circuit according to an embodiment of the present application is shown. Fig. 9As shown, according to the measurement method of thyristors and capacitors in a series inverter circuit of an embodiment of the present application, it includes: S110, signal segmentation of the amplified pulse detection signal to obtain a sequence of pulse detection signal segments; S120, multi-scale waveform semantic feature extraction of the sequence of pulse detection signal segments to obtain a sequence of pulse detection signal segment waveform semantic feature vectors; S130, inputting the sequence of pulse detection signal segment waveform semantic feature vectors into a self-supervised feature sequence significant fusion module to obtain a pulse detection signal full time domain waveform significant fusion representation vector; S140, determining a fault diagnosis result based on the pulse detection signal full time domain waveform significant fusion representation vector.

[0091] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned method for measuring the thyristor and the capacitor in the series inverter circuit have been described in the above reference. Figures 4 to 8 The description of the circuit system for measuring the thyristor and capacitor in the series inverter circuit has been described in detail, and therefore, its repeated description will be omitted.

[0092] Furthermore, if Fig.10 As shown, the power supply circuit 10 further includes a charging circuit 11, a power supply circuit 12 and an adjustable power supply circuit 13. Specifically, the input end of the power supply circuit 12 is connected to the output end of the charging circuit 11; the input end of the adjustable power supply circuit 13 is connected to the output end of the power supply circuit 12. Specifically, Figures 11 to 13 Schematic diagrams of a charging circuit 11 , a power supply circuit 12 and an adjustable power supply circuit 13 are shown respectively.

[0093] In an exemplary embodiment, there is also provided a non-volatile computer-readable storage medium, such as a memory including computer program instructions, which can be executed by a processing component of an apparatus to perform the above method.

[0094] The present application may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present application.

[0095] Various aspects of the present application are described herein with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0096] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0097] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0098] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of the module, program segment or instruction includes one or more executable instructions for realizing the logical function of the specification. In some alternative implementations, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can also be executed in the opposite order sometimes, depending on the function involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be realized by a dedicated hardware-based system that performs the function or action of the specification, or can be realized by a combination of special-purpose hardware and computer instructions.

[0099] The embodiments of the present application have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A circuit system for measuring thyristors and capacitors in a series inverter circuit, characterized in that: include: Power supply circuit; A measuring circuit, wherein the measuring circuit comprises a signal processor and an amplifying circuit powered by the power supply circuit; the signal processor is used to transmit a pulse detection signal, and the amplifying circuit is used to amplify the captured pulse detection signal to obtain an amplified pulse detection signal and transmit it back to the signal processor; wherein the signal processor is used to determine the fault type based on the amplified pulse detection signal; Wherein, the signal processor comprises: A signal segmentation module, used for performing signal segmentation on the amplified pulse detection signal to obtain a sequence of pulse detection signal segments; A multi-scale waveform semantic feature extraction module, used for performing multi-scale waveform semantic feature extraction on the sequence of pulse detection signal segments to obtain a sequence of pulse detection signal segment waveform semantic feature vectors; A fusion module, used for inputting the sequence of semantic feature vectors of the pulse detection signal segment waveform into a self-supervised feature sequence significant fusion module to obtain a significant fusion representation vector of the full time domain waveform of the pulse detection signal; A fault diagnosis module, used to determine a fault diagnosis result based on a significant fusion representation vector of the full time domain waveform of the pulse detection signal; Wherein, the multi-scale waveform semantic feature extraction module includes: A signal waveform feature extraction unit, used for passing each pulse detection signal segment in the sequence of pulse detection signal segments through a signal waveform feature extractor based on a dilated convolutional neural network model to obtain a sequence of pulse detection signal segment waveform semantic feature maps; A multi-scale encoding unit, used for inputting each pulse detection signal segment waveform semantic feature map in the sequence of pulse detection signal segment waveform semantic feature maps into a feature multi-scale scanning module to obtain a sequence of pulse detection signal segment waveform semantic feature vectors; Wherein, the multi-scale coding unit comprises: A semantic feature extraction subunit, used for performing signal segment global semantic feature extraction and signal segment local semantic feature extraction on the pulse detection signal segment waveform semantic feature graph to obtain a signal segment waveform global semantic feature vector and a signal segment waveform local semantic feature vector; A position-by-position element-by-element addition subunit is used to calculate the position-by-element addition of the global semantic feature vector of the signal segment waveform and the local semantic feature vector of the signal segment waveform to obtain a multi-scale semantic fusion feature vector of the signal segment waveform; The activation subunit is used to perform activation processing on the multi-scale semantic fusion feature vector of the signal segment waveform based on the sigmoid function to obtain the semantic feature vector of the pulse detection signal segment waveform.

2. The circuit system for measuring thyristors and capacitors in a series inverter circuit according to claim 1, characterized in that: The semantic feature extraction subunit is used to: Upsampling the pulse detection signal segment waveform semantic feature map to obtain an upsampled pulse detection signal segment waveform semantic feature map; Using a signal segment global semantic feature scanner to process the upsampled pulse detection signal segment waveform semantic feature map to obtain a signal segment waveform global scanning semantic feature vector; Performing dimension adjustment on the signal segment waveform global scanning semantic feature vector to obtain the signal segment waveform global semantic feature vector; Processing the upsampled pulse detection signal segment waveform semantic feature map using a signal segment local semantic feature scanner to obtain a signal segment waveform local scanning semantic feature vector; The dimension of the local scanning semantic feature vector of the signal segment waveform is adjusted to obtain the local semantic feature vector of the signal segment waveform.

3. The circuit system for measuring thyristors and capacitors in a series inverter circuit according to claim 2, characterized in that: The signal segment global semantic feature scanner includes a global average pooling layer, a first point convolution layer, a first batch normalization layer and a first activation layer, and the signal segment local semantic feature scanner includes a second point convolution layer, a second batch normalization layer and a second activation layer.

4. The circuit system for measuring thyristors and capacitors in a series inverter circuit according to claim 3, characterized in that: The fusion module comprises: A semantically significant correlation evaluation unit, used for performing a semantically significant correlation evaluation based on self-clustering on the sequence of semantic feature vectors of the pulse detection signal segment waveform to obtain a sequence of self-supervisory signal segment correlation weights; A weighting unit is used to calculate the vector-by-vector weighted sum of the sequence of semantic feature vectors of the pulse detection signal segment waveform using the sequence of the self-supervisory signal segment correlation weights as the weight sequence to obtain a significant fusion representation vector of the full time domain waveform of the pulse detection signal.

5. The circuit system for measuring thyristors and capacitors in a series inverter circuit according to claim 4, characterized in that: The semantic significant relevance assessment unit comprises: A mean calculation subunit, used for calculating the mean vector of the sequence of semantic feature vectors of the pulse detection signal segment waveform as the cluster center of the semantic distribution of the pulse detection signal segment sequence; A waveform semantic matching association subunit, used for calculating the waveform semantic matching association degree between each pulse detection signal segment waveform semantic feature vector in the sequence of the pulse detection signal segment waveform semantic feature vector and the pulse detection signal segment sequence semantic distribution cluster center to obtain a sequence of waveform semantic matching association degrees; The normalization subunit is used to input the sequence of waveform semantic matching associations into a softmax function for soft maximum normalization processing to obtain a sequence of self-supervisory signal segment correlation weights.

6. The circuit system for measuring thyristors and capacitors in a series inverter circuit according to claim 5, characterized in that: The waveform semantic matching association subunit is used to: Encoding the pulse detection signal segment waveform semantic feature vector through a fully connected layer to obtain a pulse detection signal segment waveform semantic encoding feature vector; The transposed vector of the pulse detection signal segment sequence semantic distribution cluster center is multiplied by the pulse detection signal segment waveform semantic coding feature vector to obtain the waveform semantic matching association degree.

7. The circuit system for measuring thyristors and capacitors in a series inverter circuit according to claim 6, characterized in that: The fault diagnosis module is used for: The full time domain waveform of the pulse detection signal is significantly fused into a representation vector and passed through a classifier-based fault diagnostic device to obtain the fault diagnosis result, which is used to indicate that the resonant capacitor is damaged or the inverter thyristor is damaged.

Images