A fault detection method, apparatus, device, medium and program product

Abstract

The application discloses a fault detection method and device, equipment, medium and program product, and relates to the technical field of fault diagnosis. The method comprises the following steps: acquiring a first voltage signal and a first current signal input to a converter, and a second voltage signal of a DC bus; determining whether a voltage variation represented by the second voltage signal is greater than a preset variation, and obtaining a first determination result; if the first determination result indicates that the voltage variation represented by the second voltage signal is not greater than the preset variation, fusing the first voltage signal and the first current signal to obtain a fused signal; and processing the fused signal to obtain a fault position of the converter. The method effectively avoids the problem that the current fault position and the voltage fault position do not correspond in the traditional detection method, and improves the efficiency and accuracy of fault position determination.

Description

Technical Field

[0001] This invention belongs to the field of fault diagnosis technology, and particularly relates to a fault detection method, device, equipment, medium and program product. Background Technology

[0002] In the field of wind power generation, wind power converters play a crucial role in ensuring the stable and efficient operation of the power generation system. These converters are responsible for converting the frequency- and voltage-changing electrical energy generated by wind turbines into stable grid-connected electrical energy. However, due to the complex operating environment of wind turbines, wind power converters are prone to various faults. Timely and accurate detection of these faults is essential for avoiding power outages, reducing maintenance costs, and improving the overall reliability of wind power generation systems.

[0003] In wind power converter fault detection, traditional methods typically rely on signal analysis techniques. In recent years, with the development of artificial intelligence, the Transformer model has shown great potential in processing sequential data. However, it faces significant challenges in practical applications of wind power converter fault detection, primarily the difficulty of data integration. Wind power converters generate various types of signals, including voltage, current, and DC bus voltage. Previously, when using the Transformer model for fault detection, mismatches often occurred between current and voltage fault locations, making it impossible to accurately pinpoint fault locations under multiple signal types and severely impacting the ability to respond quickly to faults. Summary of the Invention

[0004] This application provides a fault detection method, apparatus, equipment, medium, and program product, which can accurately and quickly identify the fault location and further improve the ability to respond quickly to faults.

[0005] To achieve the above objectives, this application adopts the following technical solution:

[0006] Firstly, this application provides a fault detection method, including:

[0007] Acquire the first voltage signal and the first current signal input to the converter, as well as the second voltage signal of the DC bus;

[0008] Determine whether the voltage change represented by the second voltage signal is greater than a preset change amount to obtain a first determination result;

[0009] If the first judgment result indicates that the voltage change represented by the second voltage signal is not greater than a preset change, the first voltage signal and the first current signal are fused to obtain a fused signal;

[0010] The fault location of the converter is obtained by processing the fused signal.

[0011] In some optional implementations, the first voltage signal and the first current signal are fused, including:

[0012] Feature extraction is performed on the first voltage signal to obtain the first voltage matrix; feature extraction is performed on the first current signal to obtain the first current matrix.

[0013] Based on the voltage weight matrix and the first voltage matrix, a first voltage attention matrix is ​​determined; based on the current weight matrix and the first current matrix, a first current attention matrix is ​​determined.

[0014] The first voltage attention matrix and the first current attention matrix are fused.

[0015] In some alternative implementations, the first voltage attention matrix is ​​determined based on the voltage weight matrix and the first voltage matrix, including:

[0016] Based on the voltage weight matrix and the first voltage matrix, determine the first voltage query matrix, the first voltage key matrix, and the first voltage value matrix;

[0017] The first voltage attention weight matrix is ​​determined based on the first voltage query matrix and the first voltage key matrix;

[0018] The first voltage attention matrix is ​​determined based on the first voltage attention weight matrix and the first voltage value matrix;

[0019] Based on the current weight matrix and the first current matrix, the first current attention matrix is ​​determined, including:

[0020] Based on the current weight matrix and the first current matrix, the first current query matrix, the first current key matrix, and the first current value matrix are determined respectively.

[0021] The first current attention weight matrix is ​​determined based on the first current query matrix and the first current key matrix;

[0022] The first current attention matrix is ​​determined based on the first current attention weight matrix and the first current value matrix.

[0023] In some optional implementations, the voltage weight matrix includes a voltage lookup weight matrix, a voltage key weight matrix, and a voltage value weight matrix. Based on the voltage weight matrix and the first voltage matrix, the first voltage lookup matrix, the first voltage key matrix, and the first voltage value matrix are determined, including:

[0024]

[0025] in, This is the first voltage lookup matrix. It is the first voltage bond matrix. First voltage value matrix, It is the first voltage matrix. It is a voltage lookup weight matrix. It is the voltage bond weight matrix. It is a voltage value weight matrix;

[0026] Based on the first voltage query matrix and the first voltage key matrix, the first voltage attention weight matrix is ​​determined as follows:

[0027]

[0028] in, A V It is the first voltage attention weight matrix. It is the transpose of the first voltage matrix; The dimension of the key vector;

[0029] The first voltage attention matrix is ​​determined based on the first voltage attention weight matrix and the first voltage value matrix, including:

[0030]

[0031] in, It is the first voltage attention matrix;

[0032] The current weight matrix includes a current query weight matrix, a current key weight matrix, and a current value weight matrix. Based on the current weight matrix and the first current matrix, the first current query matrix, the first current key matrix, and the first current value matrix are determined respectively as follows:

[0033]

[0034] in, This is the first current lookup matrix. It is the first current bond matrix. First current value matrix, It is the first current matrix. It is a current query weight matrix. It is the current bond weight matrix. It is a current value weight matrix;

[0035] Based on the first current query matrix and the first current key matrix, the first current attention weight matrix is ​​determined as follows:

[0036]

[0037] in, It is the first voltage attention weight matrix. It is the transpose of the first current matrix; The dimension of the key vector;

[0038] Based on the first current attention weight matrix and the first current value matrix, the first current attention matrix is ​​determined, including:

[0039]

[0040] in, It is the first current attention matrix.

[0041] In some optional implementations, the step of extracting features from the first voltage signal to obtain a first voltage matrix and extracting features from the first current signal to obtain a first current matrix includes:

[0042] Obtain the first voltage signal input to the converter;

[0043] The first voltage change is obtained based on the first voltage signal;

[0044] Determine whether the first voltage change is greater than the preset first voltage change to obtain a second determination result;

[0045] If the second judgment result indicates that the first voltage change is greater than the first voltage preset change, a first voltage abnormality signal is output.

[0046] The first voltage matrix is ​​obtained based on the first voltage anomaly signal and the time corresponding to the first voltage anomaly signal;

[0047] Obtain the first current signal input to the converter;

[0048] Based on the first current signal, the change in the first current is obtained;

[0049] Determine whether the change in the first current is greater than the preset change in the first current, and obtain the third determination result;

[0050] If the third judgment result indicates that the change in the first current is greater than the preset change in the first current, the first current abnormal signal is output.

[0051] The first current matrix is ​​obtained based on the first current anomaly signal and the time corresponding to the first current anomaly signal.

[0052] In some optional implementations, if the first judgment result indicates that the voltage change represented by the second voltage signal is greater than a preset change, the system is stopped.

[0053] Secondly, this application provides a fault detection device, the device comprising:

[0054] The acquisition module is used to acquire the first voltage signal and the first current signal input to the converter, as well as the second voltage signal of the DC bus;

[0055] The judgment module is used to determine whether the voltage change represented by the second voltage signal is greater than a preset change, and to obtain a first judgment result;

[0056] The fusion module is used to fuse the first voltage signal and the first current signal to obtain a fused signal if the first judgment result indicates that the voltage change represented by the second voltage signal is not greater than a preset change amount.

[0057] The processing module is used to process the fused signal to obtain the fault location of the converter.

[0058] Thirdly, this application provides a computing device, including one or more processors and a memory;

[0059] The memory stores computer-readable instructions that, when executed by the one or more processors, perform the steps of the fault detection method as described in any of the first aspects.

[0060] Fourthly, this application provides a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the fault detection method as described in any of the first aspects.

[0061] Fifthly, this application provides a computer program product, characterized in that the computer program product includes one or more computer instructions, which, when executed by a computer, perform the method as described in any of the first aspects.

[0062] As can be seen from the above technical solution, this application has at least the following beneficial effects:

[0063] In this application, a first voltage signal and a first current signal input to the converter, as well as a second voltage signal of the DC bus, are first acquired; then, it is determined whether the voltage change represented by the second voltage signal is greater than a preset change, and a first determination result is obtained; if the first determination result indicates that the voltage change represented by the second voltage signal is not greater than the preset change, the first voltage signal and the first current signal are fused to obtain a fused signal; the fused signal is processed to obtain the fault location of the converter.

[0064] Among these, an abnormal change in the second voltage of the DC bus corresponds to an extremely dangerous fault that requires a rapid response. Therefore, prioritizing the determination of whether the voltage change represented by the second voltage signal of the DC bus exceeds a preset change amount can greatly improve the timeliness and accuracy of fault detection.

[0065] The fault detection method proposed in this application not only reduces the amount of data computation, but also effectively avoids the common problem of mismatch between current fault location and voltage fault location in traditional detection methods. By comprehensively utilizing voltage, current, and DC bus voltage signals, it ultimately outputs an accurate fault location, greatly improving the efficiency and accuracy of fault location determination. Attached Figure Description

[0066] Figure 1 A flowchart of a fault detection method provided in an embodiment of this application;

[0067] Figure 2 A schematic diagram of a fault detection device provided in an embodiment of this application;

[0068] Figure 3 This is a schematic diagram of a computing device provided in an embodiment of this application. Detailed Implementation

[0069] The terms "first," "second," and "third," etc., used in this application specification and accompanying drawings are used to distinguish different objects, not to limit a specific order.

[0070] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0071] To ensure clarity and conciseness in the description of the following embodiments, a brief introduction to the related technologies is given first:

[0072] Currently, wind turbine converter fault detection relies on Transformer models for data processing. However, wind turbine converters generate various types of signals, including voltage, current, and DC bus voltage. Each Transformer model can only perform fault analysis for one type of signal. If both voltage and current signals are analyzed simultaneously, two Transformer models are needed, resulting in a large computational load. Furthermore, if two Transformer models analyzing current and voltage both identify a fault, but the fault locations are in different phases, a direct judgment cannot be made, requiring additional judgment procedures. This significantly slows down the judgment speed and reduces accuracy.

[0073] In view of this, this application provides a fault detection method. The method first acquires a first voltage signal and a first current signal input to the converter, as well as a second voltage signal from the DC bus. Then, it determines whether the voltage change represented by the second voltage signal is greater than a preset change, obtaining a first judgment result. An abnormal change in the second voltage of the DC bus corresponds to an extremely dangerous fault requiring rapid response. Therefore, prioritizing the determination of whether the voltage change represented by the second voltage signal of the DC bus is greater than a preset change can greatly improve the timeliness and accuracy of fault detection. Once an abnormal change in the second voltage of the DC bus is detected, the system can quickly trigger a protection mechanism, cutting off relevant circuits in a very short time to prevent serious faults in the converter and even the entire power system caused by sudden changes in DC bus voltage. This effectively avoids catastrophic consequences such as equipment damage and fires, thereby ensuring the safe and stable operation of the entire power system, significantly improving system reliability and availability, and reducing potential economic losses and safety risks.

[0074] If the first judgment result indicates that the voltage change represented by the second voltage signal is not greater than a preset change, the first voltage signal and the first current signal are fused to obtain a fused signal; the fused signal is processed to obtain the fault location of the converter.

[0075] Among them, converters include wind power converters.

[0076] The first voltage signal and the first current signal input to the converter, as well as the second voltage signal of the DC bus, are obtained by the acquisition module sampling at a preset sampling position. The first voltage signal and the first current signal, as well as the second voltage signal of the DC bus, are sampled multiple times evenly according to a set time interval within each signal cycle.

[0077] The time interval is set according to the application scenario of the fault detection method described in this application.

[0078] The first voltage anomaly signal is a combination of the sampled voltage value in the signal period before the anomaly and the sampled voltage value in the signal period after the anomaly. The first current anomaly signal is a combination of the sampled current value in the signal period before the anomaly and the sampled current value in the signal period after the anomaly.

[0079] Compared to traditional methods that involve complex calculations and independent analysis of multiple signals, this approach, which integrates voltage, current, and DC bus voltage signals during fault detection, significantly reduces the amount of data computation. The system avoids lengthy and tedious calculations for each signal individually, instead cleverly utilizing signal fusion to simplify the data processing flow and save substantial computational resources and time. Most importantly, multi-signal fusion analysis effectively avoids the common problem in traditional detection methods where the locations of current and voltage faults do not correspond. By fusing multiple signals, this method comprehensively considers the system's operating status and analyzes faults from multiple dimensions, greatly improving the efficiency and accuracy of fault location determination. This makes fault diagnosis more efficient and precise, providing a more reliable guarantee for the stable operation of the power system.

[0080] To make the technical solution of this application clearer and easier to understand, a fault detection method provided by an embodiment of this application is described below. Figure 1 As shown, this figure is a flowchart of a fault detection method provided in an embodiment of this application.

[0081] In this embodiment of the application, the fault detection method includes:

[0082] S1. Obtain the first voltage signal and the first current signal input to the converter, as well as the second voltage signal of the DC bus.

[0083] In this process, a low-pass filter is used to filter out signal distortion caused by electromagnetic interference or transmission errors from the first voltage signal, the first current signal and the second voltage signal of the DC bus, so as to obtain the filtered first voltage signal, the first current signal and the second voltage signal of the DC bus for subsequent data processing.

[0084] The frequency response H( of the low-pass filter) f The formula is:

[0085]

[0086] Where H is the frequency response function of the filter, representing the gain of the filter for signals of different frequencies;

[0087] j is the imaginary unit, satisfying j 2 =−1;

[0088] R is the resistance value in the filter, and the unit is ohms (Ω).

[0089] C is the capacitance value in the filter, and the unit is farad (F).

[0090] The cutoff frequency is:

[0091]

[0092] Where fc is the cutoff frequency, and the unit is Hertz (Hz).

[0093] R is the resistance value in the filter, and the unit is ohms (Ω).

[0094] C is the capacitance value in the filter, measured in farads (F); R and C determine the filter's cutoff frequency.

[0095] It uses a simple RC filter, which has a simple structure, fast response speed, and is fully applicable to the frequency range of wind power converters.

[0096] S2. Determine whether the voltage change represented by the second voltage signal is greater than the preset change amount, and obtain the first judgment result.

[0097] The calculation method for the preset change is as follows:

[0098] The second voltage signal in the sample data is decomposed into a first-order value, that is, the difference between the second voltage signals of adjacent sampling points is calculated to obtain the change in the second voltage. .

[0099] Then based on the second voltage change The mean value of the second voltage change was calculated. Standard deviation of the second voltage change According to the mean and standard deviation Calculate the preset change amount ,include:

[0100]

[0101] Here, K is a preset constant that will change in real time according to the rotational speed of the wind turbine to adapt to different mean and standard deviation at different rotational speeds.

[0102] The sample data is a dataset of the second voltage signal of a DC bus without any anomalies.

[0103] The voltage change represented by the second voltage signal is the absolute value of the second voltage change obtained by first-order decomposition of the second voltage signal acquired during fault detection. .

[0104] Determine whether the voltage change represented by the second voltage signal is greater than a preset change amount.

[0105] Determine whether > .

[0106] S3. If the first judgment result indicates that the voltage change represented by the second voltage signal is not greater than a preset change, then the first voltage signal is decomposed into first-order components to obtain the first voltage change. .

[0107] include , and Where a represents phase a, b represents phase b, and c represents phase c.

[0108] Determine whether the change in the first voltage is greater than the preset change in the first voltage. ,Right now , or If one or more of the conditions are met, it means that the change in the first voltage is greater than the preset change in the first voltage. Thus, the second judgment result was obtained.

[0109] If the second judgment result indicates that the change in the first voltage is greater than the preset change in the first voltage, a first voltage abnormality signal is output, and the first voltage abnormality signal is concatenated with the time corresponding to the first voltage abnormality signal to obtain a first voltage matrix. ,

[0110] =

[0111] Where a represents phase a, b represents phase b, c represents phase c, and t represents time.

[0112] For example, the signal period is 10 seconds, and 100 points are sampled in each signal period. The first voltage change is greater than the first voltage preset change in the 10th second. The output first voltage abnormality signal is the voltage signal of 200 points sampled within 1 to 20 seconds.

[0113] The first voltage anomaly signal is concatenated with the corresponding time intervals to obtain the first voltage matrix. for:

[0114] =

[0115] The first current signal is decomposed into first-order components to obtain the first current change. , include , and Where a represents phase a, b represents phase b, and c represents phase c.

[0116] Determine whether the change in the first current is greater than the preset change in the first current. ,Right now , or If one or more of the conditions are met, it means that the change in the first current is greater than the preset change in the first current. The third judgment result is obtained.

[0117] If the third judgment result indicates that the change in the first current is greater than the preset change in the first current, a first current abnormal signal is output, and the first current abnormal signal is concatenated with the time corresponding to the first current abnormal signal to obtain a first current matrix. ,

[0118] =

[0119] Where a represents phase a, b represents phase b, c represents phase c, and t represents time.

[0120] For example, the signal period is 10 seconds, and 100 points are sampled in each signal period. The first current change is greater than the first current preset change when it is the 10th second. The first voltage abnormality signal output is the voltage signal of 200 points sampled within 1 to 20 seconds.

[0121] The first current anomaly signal is concatenated with the corresponding time intervals to obtain the first current matrix. for:

[0122] =

[0123] The voltage weight matrix includes the voltage lookup weight matrix, the voltage key weight matrix, and the voltage value weight matrix.

[0124] Based on the voltage weight matrix and the first voltage matrix, the first voltage query matrix, the first voltage key matrix, and the first voltage value matrix are determined, including:

[0125]

[0126] in, This is the first voltage lookup matrix. It is the first voltage bond matrix. First voltage value matrix, It is the first voltage matrix. It is a voltage lookup weight matrix. It is the voltage bond weight matrix. It is a voltage value weight matrix.

[0127] Among them, the first voltage preset change amount The calculation method is as follows:

[0128] The first voltage signal in the sample data is decomposed into first-order components, that is, the difference between the first voltage signals of adjacent sampling points in the sample data is calculated to obtain the change in the first voltage of the sample. ;

[0129] Then take the first voltage change of the sample Calculate the mean of the first voltage change in the sample. and the standard deviation of the first voltage change in the sample According to the mean and standard deviation Calculate the preset change in the first voltage ,include:

[0130]

[0131] Here, K is a preset constant that will change in real time according to the rotational speed of the wind turbine to adapt to different mean and standard deviation at different rotational speeds.

[0132] The sample data is a dataset of the first voltage signal input to the converter without any anomalies.

[0133] The calculation method for the preset change in the first current is as follows:

[0134] The first current signal in the sample data is decomposed into first-order components, that is, the difference between the first current signals of adjacent sampling points in the sample data is calculated to obtain the change in the first current of the sample. ;

[0135] Then take the first current change of the sample Calculate the mean of the first current change in the sample. and the standard deviation of the first current change in the sample According to the mean and standard deviation Calculate the preset change in the first current ,include:

[0136]

[0137] Here, K is a preset constant that will change in real time according to the rotational speed of the wind turbine to adapt to different mean and standard deviation at different rotational speeds.

[0138] The sample data is a dataset of the first current signal input to the converter without any abnormalities.

[0139] Based on the first voltage query matrix and the first voltage key matrix, the first voltage attention weight matrix is ​​determined as follows:

[0140]

[0141] in, A V It is the first voltage attention weight matrix. It is the transpose of the first voltage matrix; The dimension of the key vector;

[0142] The step of determining the first voltage attention matrix based on the first voltage attention weight matrix and the first voltage value matrix includes:

[0143]

[0144] in, It is the first voltage attention matrix;

[0145] The current weight matrix includes a current query weight matrix, a current key weight matrix, and a current value weight matrix. Based on the current weight matrix and the first current matrix, the first current query matrix, the first current key matrix, and the first current value matrix are determined respectively as follows:

[0146]

[0147] in, This is the first current lookup matrix. It is the first current bond matrix. First current value matrix, It is the first current matrix. It is a current query weight matrix. It is the current bond weight matrix. It is a current value weight matrix;

[0148] Based on the first current query matrix and the first current key matrix, the first current attention weight matrix is ​​determined as follows:

[0149]

[0150] in, It is the first voltage attention weight matrix. It is the transpose of the first current matrix; The dimension of the key vector;

[0151] The first current attention matrix is ​​determined based on the first current attention weight matrix and the first current value matrix, including:

[0152]

[0153] in, It is the first current attention matrix.

[0154] The first voltage attention matrix and the first current attention matrix are fused by element-level multiplication to obtain the fused signal matrix C, which includes:

[0155] ⊙

[0156] Furthermore, if the obtained matrix C contains two identical data points in the same column, meaning that the weights of two different phases are the same at the same time step, this indicates an error occurred during the previous fusion process. When performing fault diagnosis, it becomes impossible to determine which phase is faulty and thus the specific fault location cannot be determined. In this case, S3 needs to be repeated until a fused signal matrix with no identical data points in the same column is obtained. .

[0157] S4 will fuse the signal matrix or fused signal matrix Input the data into the Transformer model to obtain the fault location of the converter.

[0158] In some possible implementations, if the first judgment result indicates that the voltage change represented by the second voltage signal is greater than a preset change, the system is stopped.

[0159] In this invention, an abnormal change in the second voltage of the DC bus corresponds to an extremely dangerous fault requiring a rapid response. Therefore, prioritizing the determination of whether the voltage change represented by the second voltage signal of the DC bus exceeds a preset value can greatly improve the timeliness and accuracy of fault detection. Once an abnormal change in the second voltage of the DC bus is detected, the system can quickly trigger a protection mechanism, cutting off the relevant circuits in a very short time. This prevents serious faults in the converter and even the entire power system caused by sudden changes in DC bus voltage, effectively avoiding catastrophic consequences such as equipment damage and fires. This ensures the safe and stable operation of the entire power system, significantly improves system reliability and availability, and reduces potential economic losses and safety risks.

[0160] This application also provides a fault detection device, such as... Figure 2 As shown in the figure, this is a schematic diagram of a wind power converter fault detection device 200 provided in an embodiment of this application. The device includes:

[0161] The acquisition module 201 is used to acquire the first voltage signal and the first current signal input to the converter, as well as the second voltage signal of the DC bus;

[0162] The judgment module 202 is used to determine whether the voltage change represented by the second voltage signal is greater than a preset change, and to obtain a first judgment result;

[0163] Fusion module 203 is used to fuse the first voltage signal and the first current signal to obtain a fused signal;

[0164] The processing module 204 is used to process the fused signal to obtain the fault location of the converter.

[0165] Optionally, the fusion module 203 is specifically used to extract features from the first voltage signal to obtain a first voltage matrix, extract features from the first current signal to obtain a first current matrix, determine a first voltage attention matrix based on the voltage weight matrix and the first voltage matrix, determine a first current attention matrix based on the current weight matrix and the first current matrix, and fuse the first voltage attention matrix and the first current attention matrix.

[0166] Optionally, the fusion module 203 is specifically used to determine the first voltage query matrix, the first voltage key matrix, and the first voltage value matrix based on the voltage weight matrix and the first voltage matrix, respectively, including:

[0167]

[0168] in, This is the first voltage lookup matrix. It is the first voltage bond matrix. First voltage value matrix, It is the first voltage matrix. It is a voltage lookup weight matrix. It is the voltage bond weight matrix. It is a voltage value weight matrix; the voltage weight matrix includes the voltage lookup weight matrix, the voltage key weight matrix, and the voltage value weight matrix;

[0169] Based on the first voltage query matrix and the first voltage key matrix, the first voltage attention weight matrix is ​​determined, including:

[0170]

[0171] in, A V It is the first voltage attention weight matrix. It is the transpose of the first voltage matrix; The dimension of the key vector;

[0172] Based on the first voltage attention weight matrix and the first voltage value matrix, the first voltage attention matrix is ​​determined, including:

[0173]

[0174] in, It is the first voltage attention matrix;

[0175] Based on the current weight matrix and the first current matrix, the first current query matrix, the first current key matrix, and the first current value matrix are determined respectively, including:

[0176]

[0177] in, This is the first current lookup matrix. It is the first current bond matrix. First current value matrix, It is the first current matrix. It is a current query weight matrix. It is the current bond weight matrix. It is a current value weight matrix;

[0178] Based on the first current query matrix and the first current key matrix, the first current attention weight matrix is ​​determined as follows:

[0179]

[0180] in, It is the first voltage attention weight matrix. It is the transpose of the first current matrix; The dimension of the key vector;

[0181] Based on the first current attention weight matrix and the first current value matrix, the first current attention matrix is ​​determined, including:

[0182]

[0183] in, It is the first current attention matrix.

[0184] Optionally, the fusion module 203 is specifically used to acquire the first voltage signal input to the converter;

[0185] The first voltage change is obtained based on the first voltage signal;

[0186] Determine whether the first voltage change is greater than the preset first voltage change to obtain a second determination result;

[0187] If the second judgment result indicates that the first voltage change is greater than the first voltage preset change, a first voltage abnormality signal is output.

[0188] The first voltage matrix is ​​obtained based on the first voltage anomaly signal and the time corresponding to the first voltage anomaly signal;

[0189] Obtain the first current signal input to the converter;

[0190] Based on the first current signal, the change in the first current is obtained;

[0191] Determine whether the change in the first current is greater than the preset change in the first current, and obtain the third determination result;

[0192] If the third judgment result indicates that the change in the first current is greater than the preset change in the first current, the first current abnormal signal is output.

[0193] The first current matrix is ​​obtained based on the first current anomaly signal and the time corresponding to the first current anomaly signal.

[0194] The fault detection device further includes a control execution module, specifically used to stop the machine if the first judgment result indicates that the voltage change represented by the second voltage signal is greater than a preset change.

[0195] This application also provides a computing device. For example... Figure 3 As shown in the figure, this is a schematic diagram of a computing device provided in an embodiment of this application. The computing device 300 includes a bus 301, a processor 302, a communication interface 303, and a memory 304. The processor 302, the memory 304, and the communication interface 303 communicate with each other via the bus 301.

[0196] Bus 301 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0197] The processor 302 can be any one or more of the following processors: central processing unit (CPU), graphics processing unit (GPU), microprocessor (MP), or digital signal processor (DSP).

[0198] Communication interface 303 is used for communication with external devices.

[0199] Memory 304 may include volatile memory, such as random access memory (RAM). Memory 304 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0200] The memory 304 stores executable code, and the processor 302 executes the executable code to perform the aforementioned fault detection method.

[0201] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium capable of being stored by a computing device, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to perform the above-described method.

[0202] This application also provides a computer program product comprising one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in this application are generated.

[0203] The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, or data center to another website, computer, or data center via wired (e.g., coaxial cable, fiber optic) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0204] When the computer program product is executed by a computer, the computer performs any of the aforementioned fault detection methods. The computer program product can be a software installation package; when any of the aforementioned fault detection methods is required, the computer program product can be downloaded and executed on the computer.

[0205] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0206] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be covered within the scope of protection of this application.

Claims

1. A fault detection method characterized by, The method comprises: obtaining a first voltage signal and a first current signal input to the converter, and a second voltage signal of a DC bus; determining whether a voltage variation represented by the second voltage signal is greater than a preset variation, to obtain a first determination result; if the first determination result indicates that the voltage variation represented by the second voltage signal is not greater than the preset variation, fusing the first voltage signal and the first current signal to obtain a fused signal; processing the fused signal to obtain a fault position of the converter; the fusing of the first voltage signal and the first current signal comprises: performing feature extraction on the first voltage signal to obtain a first voltage matrix, and performing feature extraction on the first current signal to obtain a first current matrix; determining a first voltage attention matrix according to a voltage weight matrix and the first voltage matrix, and determining a first current attention matrix according to a current weight matrix and the first current matrix; fusing the first voltage attention matrix and the first current attention matrix through element-level multiplication to obtain a fused signal matrix C, comprising: ⊙ 。 2. The method of claim 1, wherein, the determination of the first voltage attention matrix according to the voltage weight matrix and the first voltage matrix comprises: determining a first voltage query matrix, a first voltage key matrix and a first voltage value matrix respectively according to the voltage weight matrix and the first voltage matrix; determining a first voltage attention weight matrix according to the first voltage query matrix and the first voltage key matrix; determining the first voltage attention matrix according to the first voltage attention weight matrix and the first voltage value matrix; the determination of the first current attention matrix according to the current weight matrix and the first current matrix comprises: determining a first current query matrix, a first current key matrix and a first current value matrix respectively according to the current weight matrix and the first current matrix; determining a first current attention weight matrix according to the first current query matrix and the first current key matrix; determining the first current attention matrix according to the first current attention weight matrix and the first current value matrix.

3. The method of claim 2, wherein, The voltage weight matrix comprises a voltage query weight matrix, a voltage key weight matrix and a voltage value weight matrix, and the determination of the first voltage query matrix, the first voltage key matrix and the first voltage value matrix respectively according to the voltage weight matrix and the first voltage matrix comprises: wherein, is a first voltage query matrix, is a first voltage key matrix, is a first voltage value matrix, is a first voltage matrix, is a voltage query weight matrix, is a voltage key weight matrix, is a voltage value weight matrix; determining the first voltage attention weight matrix according to the first voltage query matrix and the first voltage key matrix comprises: wherein, A V is a first voltage attention weight matrix, is a transpose of the first voltage matrix; denotes a dimension of the key vector. the determination of the first voltage attention matrix according to the first voltage attention weight matrix and the first voltage value matrix comprises: wherein, is a first voltage attention matrix; The current weight matrix comprises a current query weight matrix, a current key weight matrix and a current value weight matrix, and the determination of the first current query matrix, the first current key matrix and the first current value matrix respectively according to the current weight matrix and the first current matrix comprises: wherein, is a first current query matrix, is a first current key matrix, is a first current value matrix, is a first current matrix, is a current query weight matrix, is a current key weight matrix, is a current value weight matrix; determining the first current attention weight matrix according to the first current query matrix and the first current key matrix comprises: wherein, is a first current attention weight matrix, is a transpose of the first current matrix; denotes a dimension of the key vector; The step of determining the first current attention matrix based on the first current attention weight matrix and the first current value matrix includes: wherein, is a first current attention matrix.

4. The method of claim 1, wherein, The step of extracting features from the first voltage signal to obtain a first voltage matrix and extracting features from the first current signal to obtain a first current matrix includes: Obtain the first voltage signal input to the converter; The first voltage change is obtained based on the first voltage signal; Determine whether the first voltage change is greater than the preset first voltage change to obtain a second determination result; If the second judgment result indicates that the first voltage change is greater than the first voltage preset change, a first voltage abnormality signal is output. The first voltage matrix is ​​obtained based on the first voltage anomaly signal and the time corresponding to the first voltage anomaly signal; Obtain the first current signal input to the converter; Based on the first current signal, the change in the first current is obtained; Determine whether the change in the first current is greater than the preset change in the first current, and obtain the third determination result; If the third judgment result indicates that the change in the first current is greater than the preset change in the first current, the first current abnormal signal is output. The first current matrix is ​​obtained based on the first current anomaly signal and the time corresponding to the first current anomaly signal.

5. The method of claim 1, wherein, The method further includes: If the first judgment result indicates that the voltage change represented by the second voltage signal is greater than the preset change amount, the system will be stopped.

6. A fault detection apparatus characterized by comprising: The device includes: The acquisition module is used to acquire the first voltage signal and the first current signal input to the converter, as well as the second voltage signal of the DC bus; The judgment module is used to determine whether the voltage change represented by the second voltage signal is greater than a preset change amount, and to obtain a first judgment result; The fusion module is used to fuse the first voltage signal and the first current signal to obtain a fused signal if the first judgment result indicates that the voltage change represented by the second voltage signal is not greater than a preset change amount. The fusion of the first voltage signal and the first current signal includes: Feature extraction is performed on the first voltage signal to obtain a first voltage matrix, and feature extraction is performed on the first current signal to obtain a first current matrix; A first voltage attention matrix is ​​determined based on the voltage weight matrix and the first voltage matrix; a first current attention matrix is ​​determined based on the current weight matrix and the first current matrix. The first voltage attention matrix and the first current attention matrix are fused by element-level multiplication to obtain the fused signal matrix C, which includes: ⊙ ； The processing module is used to process the fused signal to obtain the fault location of the converter.

7. A computing device, comprising: Includes one or more processors, and memory; The memory stores computer-readable instructions, which, when executed by the one or more processors, perform the steps of the fault detection method as described in any one of claims 1 to 5.

8. A storage medium, characterized by The storage medium stores computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the fault detection method as described in any one of claims 1 to 5.

9. A computer program product, characterised in that, The computer program product comprises one or more computer instructions, which, when executed by a computer, cause the computer to perform the method of any one of claims 1-5.

Images