Bidirectional power supply intelligent fault diagnosis method, electronic equipment and storage medium

Abstract

The invention relates to the technical field of power electronic monitoring, and discloses a bidirectional power supply intelligent fault diagnosis method, electronic equipment and a storage medium, and the method comprises the steps: carrying out the hierarchical decomposition of a sample signal under the condition of obtaining the sample signal of a bidirectional power supply under a historical operation condition, and obtaining a plurality of sub-band signals of different frequency bands; carrying out feature extraction on each sub-band signal based on multi-level entropy to obtain complexity features, and carrying out data fusion on all the complexity features to obtain integrated features; and constructing a fault diagnosis model based on the integrated features in combination with a preset machine learning algorithm, so as to carry out online fault diagnosis on the real-time operation condition of the bidirectional power supply through the fault diagnosis model. According to the method, the multi-frequency-band fault information of the sample signals is mined from different angles by integrating the multi-level entropy and combining the hierarchical decomposition strategy, the richness and diversity of the obtained fault information are ensured, and the accuracy of bidirectional power supply fault diagnosis is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power electronic monitoring, and in particular to a bidirectional power supply intelligent fault diagnosis method, electronic equipment and computer-readable storage medium. Background Art

[0002] With the continuous development of the new energy industry, bidirectional power supplies account for an increasingly higher proportion in power systems. Considering that any fault in a bidirectional power supply may cause power outages or voltage fluctuations, thereby affecting the normal operation of user-end equipment, it is necessary to diagnose and identify its faults in real time to improve overall reliability and safety. Currently, the main method used for fault diagnosis of bidirectional power supplies is based on time domain analysis. However, the fault signals of complex equipment may be distributed at different angles. A single time domain analysis method may not be able to fully identify the fault characteristics, resulting in a decrease in the accuracy of fault diagnosis. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems in the related art to a certain extent. To this end, the present invention provides a bidirectional power supply intelligent fault diagnosis method, electronic device and storage medium, which can improve the accuracy of bidirectional power supply fault diagnosis.

[0004] In a first aspect, an embodiment of the present invention provides a method for intelligent fault diagnosis of a bidirectional power supply, comprising: When a sample signal of the bidirectional power supply under historical operating conditions is obtained, the sample signal is hierarchically decomposed to obtain a plurality of sub-band signals of different frequency bands; Extracting features from each of the sub-band signals based on multi-level entropy to obtain complexity features, and fusing all of the complexity features to obtain integrated features; A fault diagnosis model is constructed based on the integrated features in combination with a preset machine learning algorithm, so as to perform online fault diagnosis on the real-time operating conditions of the bidirectional power supply through the fault diagnosis model.

[0005] Optionally, in one embodiment of the present invention, the hierarchical decomposition of the sample signal to obtain a plurality of sub-band signals in different frequency bands includes: Determining a sample sequence corresponding to the sample signal based on the sample signal, and constructing a low-frequency operator and a high-frequency operator according to the sample sequence; Constructing a hierarchical vector of corresponding dimension according to a predetermined number of hierarchical levels, wherein the hierarchical node indicated by the hierarchical vector is associated with the number of hierarchical levels; According to the low-frequency operator, the high-frequency operator, the layered vector and the layered node, the layered components corresponding to each layer are determined respectively, thereby obtaining a plurality of sub-band signals of different frequency bands.

[0006] Optionally, in one embodiment of the present invention, extracting features from each of the sub-band signals based on multi-level entropy to obtain complexity features includes: respectively constructing a first sub-learner for performing fuzzy entropy operation, a second sub-learner for performing spread entropy operation, and a third sub-learner for performing diversity entropy operation; Based on the first sub-learner, the second sub-learner and the third sub-learner, entropy values ​​of the hierarchical components corresponding to each layer are calculated to obtain complexity features.

[0007] Optionally, in one embodiment of the present invention, the performing entropy calculation on the hierarchical components corresponding to each layer based on the first sub-learner, the second sub-learner, and the third sub-learner to obtain the complexity feature includes: Pre-constructing a plurality of sub-characteristic signal groups, wherein the plurality of sub-characteristic signal groups include a first sub-characteristic signal group, a second sub-characteristic signal group, and a third sub-characteristic signal group; According to the number of layers of each hierarchical component, each hierarchical component is divided into a corresponding one of the sub-feature signal groups; Inputting each of the hierarchical components in the first sub-feature signal group into the first sub-learner to obtain a plurality of first entropy values, inputting each of the hierarchical components in the second sub-feature signal group into the second sub-learner to obtain a plurality of second entropy values, and inputting each of the hierarchical components in the third sub-feature signal group into the third sub-learner to obtain a plurality of third entropy values; All of the first entropy values, all of the second entropy values, and all of the third entropy values ​​are used as complexity features.

[0008] Optionally, in one embodiment of the present invention, constructing a fault diagnosis model based on the integrated features in combination with a preset machine learning algorithm includes: Performing loss minimization learning on the initial extreme learning machine based on the integrated features to obtain the extreme learning machine in an intermediate state; Inputting the integrated features into the extreme learning machine in the intermediate state for training to obtain the pre-trained extreme learning machine; The pre-trained extreme learning machine is used as the corresponding fault diagnosis model.

[0009] Optionally, in one embodiment of the present invention, the step of inputting each of the hierarchical components in the second sub-feature signal group into the second sub-learner to obtain a plurality of second entropy values ​​comprises: For each of the hierarchical components input into the second sub-learner, mapping the hierarchical components into a first signal sequence by using a normal distribution probability cumulative distribution function; For each characteristic element in the first signal sequence, assigning a corresponding dispersion category number from a plurality of preconfigured dispersion category numbers to the characteristic element by a linear classification algorithm, and constructing a feature embedding vector corresponding to the characteristic element based on the assigned dispersion category number and a preset embedding dimension and delay factor; Calculating all dispersion modes corresponding to the feature embedding vector, and respectively calculating the probability of each dispersion mode; A second entropy value of the hierarchical component is obtained based on the information entropy combined with the probabilities of the various dispersion modes.

[0010] Optionally, in one embodiment of the present invention, the step of inputting the hierarchical components in the third sub-feature signal group into the third sub-learner to obtain a plurality of third entropy values ​​comprises: For each of the hierarchical components input into the third sub-learner, performing phase space reconstruction based on the hierarchical components to obtain multiple trajectory sequences; respectively calculating the cosine similarities between any two adjacent trajectory sequences, and generating a similarity vector based on all the cosine similarities; Divide the target interval into multiple subintervals, and use the histcounts function to count the similarity vectors in each subinterval according to the multiple subintervals, to obtain the statistical probability of the similarity vector in each subinterval; A third entropy value of the hierarchical component is obtained based on the information entropy and the statistical probability of the similarity vector in each of the subintervals.

[0011] In a second aspect, an embodiment of the present invention provides an electronic device, including: at least one processor; at least one memory for storing at least one program; When at least one of the programs is executed by at least one of the processors, the bidirectional power supply intelligent fault diagnosis method as described in the first aspect is implemented.

[0012] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium storing a program executable by a processor, wherein the program executable by the processor is used to implement the bidirectional power supply intelligent fault diagnosis method as described in the first aspect when executed by the processor.

[0013] The present invention proposes a bidirectional power supply intelligent fault diagnosis method, electronic device and storage medium, which decomposes the sample signal into sub-band signals of different frequency bands through a hierarchical decomposition strategy, and then extracts features of each sub-band signal based on multi-level entropy, which can ensure the richness and diversity of the fault information obtained, and performs data fusion on all the said complexity features to obtain an integrated feature with higher accuracy and capable of characterizing the working conditions of the bidirectional power supply, which is conducive to improving the accuracy of bidirectional power supply fault diagnosis. On this basis, a fault diagnosis model is constructed based on the integrated features combined with a preset machine learning algorithm, so that the real-time operating conditions of the bidirectional power supply can be diagnosed online through the fault diagnosis model, further improving the accuracy of bidirectional power supply fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a flow chart of a bidirectional power supply intelligent fault diagnosis method provided by one embodiment of the present invention; Figure 2 1 is a schematic diagram of the execution steps of a bidirectional power supply intelligent fault diagnosis method provided by one embodiment of the present invention; Figure 3 yes Figure 1 Flowchart of step S1 in FIG. Figure 4 yes Figure 1 Partial flow chart of the step S2 in step "extracting features from each sub-band signal based on multi-level entropy to obtain complexity features"; Figure 5 yes Figure 4 Flowchart of step S22 in FIG. Figure 6 yes Figure 4 Schematic diagram of the execution flow of step S22 in FIG. Figure 7 yes Figure 5 A partial flow chart of the step S223 of step “respectively inputting each hierarchical component in the second sub-feature signal group into the second sub-learner to obtain a plurality of second entropy values”; Figure 8 yes Figure 5 A partial flow chart of the step S223 of step “respectively inputting each hierarchical component in the third sub-feature signal group into the third sub-learner to obtain a plurality of third entropy values”; Figure 9 yes Figure 8 Schematic diagram of the execution steps; Figure 10 1 is a schematic diagram of the execution steps of a bidirectional power supply intelligent fault diagnosis method provided by another embodiment of the present invention; Figure 11 This is a structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0015] Figure 1 This is a flow chart of a bidirectional power supply intelligent fault diagnosis method provided by an embodiment of the present invention. Figure 1 As shown, the bidirectional power supply intelligent fault diagnosis method may include but is not limited to steps S1 to S3.

[0016] Step S1: When a sample signal of the bidirectional power supply under historical operating conditions is obtained, the sample signal is hierarchically decomposed to obtain a plurality of sub-band signals of different frequency bands. The historical operating conditions represent the corresponding working state of the bidirectional power supply under historical operating conditions. Therefore, the sample signal collected is the historical operating signal of the bidirectional power supply. The specific value and type of this part of the sample signal can be set accordingly according to the specific scenario, for example, referring to Figure 2 , can be but not limited to sampling voltage signals, current signals, frequency signals and temperature signals under different historical operating conditions, that is, for the split capacity scenario, in other words, the operating data of the bidirectional power supply under the split capacity scenario can be immediately stored in a pre-set scenario database, so that when sampling signals are needed, they can be directly obtained from the scenario database; Step S2: extracting features from each sub-band signal based on multi-level entropy to obtain complexity features, and performing data fusion on all complexity features to obtain integrated features; Step S3: construct a fault diagnosis model based on the integrated features in combination with a preset machine learning algorithm, so as to perform online fault diagnosis on the real-time operating conditions of the bidirectional power supply through the fault diagnosis model.

[0017] In this step, the sample signal is decomposed into sub-band signals of different frequency bands through a hierarchical decomposition strategy, and then the features of each sub-band signal are extracted based on multi-level entropy, which can ensure the richness and diversity of the fault information obtained, and perform data fusion on all complexity features to obtain an integrated feature with higher accuracy that can characterize the working conditions of the bidirectional power supply, which is conducive to improving the accuracy of bidirectional power supply fault diagnosis. On this basis, a fault diagnosis model is constructed based on the integrated features combined with a preset machine learning algorithm, so that the real-time operating conditions of the bidirectional power supply can be diagnosed online through the fault diagnosis model, further improving the accuracy of bidirectional power supply fault diagnosis.

[0018] In one embodiment, before performing hierarchical decomposition on the sample signal, the sample signal may also be subjected to, but not limited to, denoising and de-cluttering processing, in order to remove redundant features in the sample signal and reduce the impact of noise on feature extraction, thereby improving the accuracy of feature extraction.

[0019] like Figure 3 As shown, in one embodiment of the present invention, step S1 may include but is not limited to the following steps: Step S11: determining a sample sequence corresponding to the sample signal based on the sample signal, and constructing a low-frequency operator and a high-frequency operator according to the sample sequence; Step S12: constructing a hierarchical vector of corresponding dimension according to a predetermined number of hierarchical levels, wherein the hierarchical node indicated by the hierarchical vector is associated with the number of hierarchical levels; Step S13: Determine the layered components corresponding to each layer according to the low-frequency operator, the high-frequency operator, the layered vector, and the layered node, thereby obtaining a plurality of sub-band signals of different frequency bands.

[0020] In this step, considering that under complex fault conditions, fault information may exist not only in the low-frequency components of the signal in a multi-scale manner, but also in the high-frequency components, and the time domain analysis method in the related prior art mainly uses dimensionless parameters for feature extraction, which cannot mine the high-frequency and low-frequency components that exist at the same time. Figure 2 As shown in the figure, a hierarchical decomposition strategy is adopted to divide the collected sample sequence, that is, firstly, a low-frequency operator and a high-frequency operator are constructed according to the sample sequence, and then a hierarchical vector of corresponding dimension is constructed according to a predetermined number of hierarchical layers. On this basis, the hierarchical components corresponding to each layer can be determined respectively by combining the low-frequency operator, the high-frequency operator, the hierarchical vector and the hierarchical node, where each layer corresponds to a different frequency band, thereby obtaining sub-band signals of multiple different frequency bands.

[0021] Specifically, the hierarchical decomposition principle can be summarized as follows, but is not limited to: Assume that the sample sequence is of length 2 n sequence , then the low-frequency operators can be constructed separately and high-frequency operators , can be but not limited to ; Or, the operator , , can also be represented by the following matrix: ; Then, select the number of layers , Indicates the maximum number of layers that can be layered, and constructs dimensional vector ,in, 0 or , That is, the hierarchical node, which can be calculated by the following formula: ; for dimensional vector , repeat the above steps to get Layer node The corresponding layered component, that is, The sub-band signal of the frequency band corresponding to the layer is as follows: .

[0022] like Figure 4 As shown, in one embodiment of the present invention, the steps in step S2, which extract features from each sub-band signal based on multi-level entropy to obtain complexity features, may include but are not limited to the following steps: Step S21, respectively constructing a first sub-learner for performing fuzzy entropy operation, a second sub-learner for performing spread entropy operation, and a third sub-learner for performing diversity entropy operation; Step S22: Based on the first sub-learner, the second sub-learner and the third sub-learner, entropy values ​​of the hierarchical components corresponding to each layer are calculated to obtain complexity features.

[0023] In this step, the complexity of the sub-band signal is evaluated from the perspective of entropy. By constructing multiple sub-learners to extract the entropy features of different sub-band signals, data fusion of the entropy features of different frequency bands can be performed. Compared with the signal analysis and processing methods in related technologies, the entropy algorithm has lower requirements on the signal. For nonlinear and non-stationary vibration signals, the entropy algorithm can better extract the characteristic information of the signal. When a fault occurs, the internal complexity of the signal will change. At this time, the change in the entropy value of the signal can indicate the change in the system state. Therefore, based on the first sub-learner, the second sub-learner and the third sub-learner, the entropy value of the hierarchical components corresponding to each layer is calculated, and the complexity characteristics under the bidirectional power supply working condition can be extracted from multiple angles.

[0024] like Figure 5 As shown, in one embodiment of the present invention, step S22 may include, but is not limited to, the following steps: Step S221: pre-constructing a plurality of sub-characteristic signal groups, wherein the plurality of sub-characteristic signal groups include a first sub-characteristic signal group, a second sub-characteristic signal group, and a third sub-characteristic signal group; Step S222: Divide each hierarchical component into a corresponding sub-feature signal group according to the number of layers of each hierarchical component; Step S223: inputting each hierarchical component in the first sub-feature signal group into a first sub-learner to obtain a plurality of first entropy values, inputting each hierarchical component in the second sub-feature signal group into a second sub-learner to obtain a plurality of second entropy values, and inputting each hierarchical component in the third sub-feature signal group into a third sub-learner to obtain a plurality of third entropy values; Step S224: Use all first entropy values, all second entropy values, and all third entropy values ​​as complexity features.

[0025] In this step, fuzzy entropy measures the signal complexity by the probability of generating new patterns within the signal. Dispersion entropy (DE), also known as dispersion entropy, uses the probability of reconstructing the internal arrangement order of the subsequence as the basis for entropy calculation. Diversity entropy (DivE) uses the statistical probability of cosine similarity between different tracks of the signal to calculate information entropy and measures the signal complexity by the change of the internal angle of the signal. All three have the characteristics of fast calculation and good stability. Therefore, Figure 2 As shown in the figure, on the basis of the hierarchical decomposition strategy, fuzzy entropy, spread entropy and diversity entropy are used as three sub-learners for feature integration to obtain the first sub-learner, the second sub-learner and the third sub-learner respectively, and each hierarchical component is adaptively grouped, so that each hierarchical component can be input into the corresponding sub-learner for entropy calculation to obtain the corresponding entropy value. The larger the entropy value, the higher the signal complexity. Overall, the health of the bidirectional power supply can be measured from three aspects: the probability of signal mutation, the internal arrangement order of the signal, and the angle information. That is, all the first entropy values, all the second entropy values ​​and all the third entropy values ​​are used as the complexity characteristics of the bidirectional power supply.

[0026] In one embodiment, each hierarchical component is divided into a corresponding one of the sub-feature signal groups according to the number of layers of each hierarchical component. This can be implemented according to the actual application scenario, that is, it depends on the division of the first sub-feature signal group, the second sub-feature signal group and the third sub-feature signal group. For example, one or more hierarchical thresholds are set, and multiple threshold intervals are divided by the set hierarchical thresholds, each threshold interval corresponds to one of the sub-feature signal groups; the number of layers of each hierarchical component is traversed, and the number of layers of each hierarchical component is divided into a corresponding one of the sub-feature signal groups according to whether the number of layers is within the threshold interval.

[0027] Specifically, if Figure 6 As shown, the hierarchical thresholds are set to 1, 2, and 3, and when calculating the hierarchical entropy, When , the sample signal will be divided into two sub-components, high frequency and low frequency. At this time, the calculated entropy value represents the initial state of the initial signal and the overall fluctuation condition. In this case, the first sub-learner is used to calculate the entropy value; when When , the sample signal is decomposed into 2 high-frequency components and 2 low-frequency components. The entropy value calculated at this time represents the detailed information of the initial signal, and the second sub-learner is selected to extract the 4 component features; when Similarly, Taking 3 as an example, the sample signal is decomposed into 4 high-frequency components and 4 low-frequency components, and the third sub-learner is selected to calculate the entropy value; finally, the complex features extracted in the above embodiment are fused using the Conact data fusion method, that is, by weighing the fault information from different angles to enhance the accuracy of the judgment of the fault problem result, it can be known that When the corresponding feature length is 2, When the corresponding feature length is 4, The corresponding feature length extracted is 8, so the feature length finally extracted based on the Conact data fusion method is 14.

[0028] like Figure 7 As shown, in one embodiment of the present invention, the steps in step S223, respectively inputting each hierarchical component in the second sub-feature signal group into the second sub-learner to obtain multiple second entropy values, may include but are not limited to the following steps: Step S2231: for each hierarchical component input into the second sub-learner, map the hierarchical component into a first signal sequence using a normal distribution probability cumulative distribution function; Step S2232: For each characteristic element in the first signal sequence, assign a corresponding dispersion category number from a plurality of pre-configured dispersion category numbers to the characteristic element using a linear classification algorithm, and construct a feature embedding vector corresponding to the characteristic element based on the assigned dispersion category number and a preset embedding dimension and delay factor. Step S2233: Calculate all dispersion modes corresponding to the feature embedding vector, and calculate the probability of each dispersion mode respectively; Step S2234: Obtain a second entropy value of the hierarchical component based on the information entropy and the probabilities of various dispersion modes.

[0029] Specifically, for a univariate signal of length N , the calculation process of its diffusion entropy is as follows: First, Mapping to signal ,Right now ; in, express The operations, express The variance of express The mean of Furthermore, the maximum number of dispersion categories is selected , embedding dimension and delay factor , for each mapped element , through the linear classification algorithm in related technology (which can be implemented by but not limited to linear classifier, training the model through features and labels, and finding a linear segmentation surface to divide the data points into two categories) from 1 to The element Assign a corresponding level and then construct the embedding vector ; Furthermore, the dispersion mode is calculated ,make , … , we can know that The corresponding dispersion mode is ,because Related to digits, and each digit exists The corresponding dispersion mode is indivual; Furthermore, calculation Probability , specifically: ; in, express Map to The number of Representation Map to The number of The quotient of the number of elements in ; Finally, the dispersion entropy is calculated based on the information entropy and the probability of the dispersion mode calculated above. is as follows: .

[0030] like Figure 8 As shown, in one embodiment of the present invention, the steps in step S223, respectively inputting each hierarchical component in the third sub-feature signal group into the third sub-learner to obtain multiple third entropy values, may include but are not limited to the following steps: Step S2235: For each hierarchical component input into the third sub-learner, perform phase space reconstruction based on the hierarchical component to obtain multiple trajectory sequences; Step S2236: Calculate the cosine similarity between any two adjacent trajectory sequences respectively, and generate a similarity vector based on all cosine similarities; Step S2237: Divide the target interval into multiple subintervals, and use the histcounts function to count the similarity vectors in each subinterval according to the multiple subintervals to obtain the statistical probability of the similarity vector in each subinterval; Step S2238: Obtain a third entropy value of the hierarchical component based on the information entropy combined with the statistical probability of the similarity vector in each subinterval.

[0031] Specifically, if Figure 9 As shown, for a sequence of length N , the calculation process of its diversity entropy is as follows: First, the sequence is reconstructed in phase space. The purpose of reconstruction is to mine more information of the entire sequence and find another dynamic system equivalent to the original sequence. The dynamic characteristics of the system are studied by reconstructing the phase space trajectory. According to the phase space embedding theory, based on the given embedding dimension and delay factor, the sequence can be reconstructed into a series of tracks, which can be expressed as ; Then, calculate the adjacent orbital sequence and Cosine similarity of , and combine all the calculation results to get the similarity vector , specifically, cosine similarity for: ; Furthermore, the target interval corresponding to the similarity is selected as [-1,1], and the interval is divided into subintervals, which can be expressed as , and then use the histcounts function (which is well known to those skilled in the art and will not be described here) to count the similarity vectors in each subinterval and calculate its statistical probability in each subinterval. ; Finally, according to the definition of information entropy, the statistical probability of the similarity vector in each subinterval is calculated and normalized to obtain the diversity entropy: ; From the above calculation process, we can see that the core idea of ​​diversity entropy is to measure the complexity of the signal by calculating the statistical probability of the cosine similarity between the internal tracks of the signal. Determines the degree of refinement in the entropy calculation process.

[0032] In one embodiment of the present invention, step S3 may include, but is not limited to, the following steps: Step S31: performing loss minimization learning on the initial extreme learning machine based on the integrated features to obtain an extreme learning machine in an intermediate state; Step S32: input the integrated features into the extreme learning machine in the intermediate state for training to obtain a pre-trained extreme learning machine; Step S33: Use the pre-trained extreme learning machine as the corresponding fault diagnosis model.

[0033] Considering the high requirement for fault response speed, the Extreme Learning Machine (ELM) avoids iterative adjustments to network weights compared to traditional machine learning algorithms, saving time and effectively addressing the challenges of high dimensionality, large data, and training difficulty. The ELM is a single-hidden-layer feedforward neural network. It first randomly generates input weights and biases for hidden-layer neurons and then calculates the output matrix of the single-hidden-layer feedforward neural network using regularized linear regression. Even with randomly generated hidden-layer weights, the ELM maintains the universal approximation capabilities of a single-hidden-layer feedforward neural network. By randomly generating parameters for hidden-layer nodes, the ELM achieves fast training and high efficiency in processing large amounts of data. It also boasts rapid learning and excellent performance in handling nonlinear relationships and complex problems. The principles behind this approach are well known to those skilled in the art and are not elaborated on here. For reference, see https: / / cloud.tencent.com / developer / article / 2113705 and https: / / blog.csdn.net / weixin_32747681 / article / details / 149856503.

[0034] In this step, if Figure 2 、 Figure 10As shown, in the offline case, the initial extreme learning machine can be used as a training set based on the integrated features in the feature pool to minimize the loss learning, so that the initial extreme learning machine can achieve the minimum loss output, and obtain the extreme learning machine model of the intermediate state. Then, the integrated features are input to the extreme learning machine in the intermediate state for training, so as to obtain the trained related parameters. Based on this, the pre-trained extreme learning machine can be determined, which is the corresponding required fault diagnosis model; after determining the fault diagnosis model, return to the online case, sample real-time data when the bidirectional power supply is running, and perform hierarchical integrated entropy feature extraction on the collected data (the principle is the same as the above embodiments, the only difference is that the base is Feature extraction is performed on real-time sampling data, so it will not be described in detail here), and then input into the extreme learning machine as a test set, and the extreme learning machine is used to perform fault classification to output the corresponding fault type. If the output is an unknown fault type, the classification samples and fault detection results can be saved again in the scenario database corresponding to the chemical composition capacity history data, that is, used as new chemical composition capacity history data, so as to facilitate the update of the extreme learning machine parameters in an offline situation based on the updated chemical composition capacity history data. It can be understood that even if the output is a known fault type, the corresponding classification data can be saved again in the scenario database corresponding to the chemical composition capacity history data, and there is no restriction here.

[0035] Figure 11 FIG. 1 is a schematic diagram of the structure of an electronic device 1000 provided by an embodiment of the present invention. Figure 11 As shown, the electronic device 1000 includes a memory 1100 and a processor 1200. The number of the memory 1100 and the processor 1200 can be one or more. Figure 11 In the embodiment, a memory 1100 and a processor 1200 are taken as an example; the memory 1100 and the processor 1200 in the device can be connected via a bus or other means. Figure 11 The bus connection is taken as an example.

[0036] Memory 1100, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the bidirectional power supply intelligent fault diagnosis method provided in any embodiment of the present invention. Processor 1200 implements the bidirectional power supply intelligent fault diagnosis method by executing the software programs, instructions, and modules stored in memory 1100.

[0037] The memory 1100 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function. In addition, the memory 1100 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 1100 may further include a memory remotely located relative to the processor 1200, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0038] An embodiment of the present invention further provides a computer-readable storage medium storing computer-executable instructions for executing the bidirectional power supply intelligent fault diagnosis method provided in any embodiment of the present invention.

[0039] An embodiment of the present invention also provides a computer program product, including a computer program or computer instructions, which are stored in a computer-readable storage medium. A processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, so that the computer device executes the bidirectional power supply intelligent fault diagnosis method provided in any embodiment of the present invention.

[0040] The electronic devices and application scenarios described in the embodiments of the present invention are intended to more clearly illustrate the technical solutions of the embodiments of the present invention, and do not constitute a limitation on the technical solutions provided by the embodiments of the present invention. Those skilled in the art will appreciate that with the evolution of electronic devices and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present invention are equally applicable to similar technical problems.

[0041] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0042] In hardware implementations, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division between physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all of the physical components may be implemented as software executed by a processor (such as a central processing unit, digital signal processor, or microprocessor), or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on computer-readable media, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is well known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVDs) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. Furthermore, as is well known to those skilled in the art, communication media typically embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0043] As used in this specification, the terms "component," "module," "system," and the like are used to refer to computer-related entities, hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program, or a computer. By way of illustration, both an application running on a computing device and the computing device can be a component. One or more components can reside in a process or execution thread, and a component can be located on a single computer or distributed between two or more computers. In addition, these components can be executed from various computer-readable media having various data structures stored thereon. Components can communicate, for example, via local or remote processes based on signals having one or more data packets (e.g., data from two components interacting with another component on a local system, a distributed system, or a network, such as the Internet interacting with other systems via signals).

Claims

1. A bidirectional power supply intelligent fault diagnosis method, characterized in that: include: When a sample signal of the bidirectional power supply under historical operating conditions is obtained, the sample signal is hierarchically decomposed to obtain a plurality of sub-band signals of different frequency bands; Extracting features from each of the sub-band signals based on multi-level entropy to obtain complexity features, and fusing all of the complexity features to obtain integrated features; A fault diagnosis model is constructed based on the integrated features in combination with a preset machine learning algorithm, so as to perform online fault diagnosis on the real-time operating conditions of the bidirectional power supply through the fault diagnosis model.

2. The bidirectional power supply intelligent fault diagnosis method according to claim 1, characterized in that: The step of performing hierarchical decomposition on the sample signal to obtain a plurality of sub-band signals in different frequency bands includes: Determining a sample sequence corresponding to the sample signal based on the sample signal, and constructing a low-frequency operator and a high-frequency operator according to the sample sequence; Constructing a hierarchical vector of corresponding dimension according to a predetermined number of hierarchical levels, wherein the hierarchical node indicated by the hierarchical vector is associated with the number of hierarchical levels; According to the low-frequency operator, the high-frequency operator, the layered vector and the layered node, the layered components corresponding to each layer are determined respectively, thereby obtaining a plurality of sub-band signals of different frequency bands.

3. The bidirectional power supply intelligent fault diagnosis method according to claim 2, characterized in that: The extracting features of each of the sub-band signals based on multi-level entropy to obtain complexity features includes: respectively constructing a first sub-learner for performing fuzzy entropy operation, a second sub-learner for performing spread entropy operation, and a third sub-learner for performing diversity entropy operation; Based on the first sub-learner, the second sub-learner and the third sub-learner, entropy values ​​of the hierarchical components corresponding to each layer are calculated to obtain complexity features.

4. The bidirectional power supply intelligent fault diagnosis method according to claim 3, characterized in that: The step of calculating the entropy value of the hierarchical components corresponding to each layer based on the first sub-learner, the second sub-learner, and the third sub-learner to obtain a complexity feature includes: Pre-constructing a plurality of sub-characteristic signal groups, wherein the plurality of sub-characteristic signal groups include a first sub-characteristic signal group, a second sub-characteristic signal group, and a third sub-characteristic signal group; According to the number of layers of each hierarchical component, each hierarchical component is divided into a corresponding one of the sub-feature signal groups; Inputting each of the hierarchical components in the first sub-feature signal group into the first sub-learner to obtain a plurality of first entropy values, inputting each of the hierarchical components in the second sub-feature signal group into the second sub-learner to obtain a plurality of second entropy values, and inputting each of the hierarchical components in the third sub-feature signal group into the third sub-learner to obtain a plurality of third entropy values; All of the first entropy values, all of the second entropy values, and all of the third entropy values ​​are used as complexity features.

5. The bidirectional power supply intelligent fault diagnosis method according to claim 1, characterized in that: The method of constructing a fault diagnosis model based on the integrated features in combination with a preset machine learning algorithm includes: Performing loss minimization learning on the initial extreme learning machine based on the integrated features to obtain the extreme learning machine in an intermediate state; Inputting the integrated features into the extreme learning machine in the intermediate state for training to obtain the pre-trained extreme learning machine; The pre-trained extreme learning machine is used as the corresponding fault diagnosis model.

6. The bidirectional power supply intelligent fault diagnosis method according to claim 4, characterized in that: The step of inputting the hierarchical components in the second sub-feature signal group into the second sub-learner to obtain a plurality of second entropy values ​​comprises: For each of the hierarchical components input into the second sub-learner, mapping the hierarchical components into a first signal sequence by using a normal distribution probability cumulative distribution function; For each characteristic element in the first signal sequence, assigning a corresponding dispersion category number from a plurality of preconfigured dispersion category numbers to the characteristic element by a linear classification algorithm, and constructing a feature embedding vector corresponding to the characteristic element based on the assigned dispersion category number and a preset embedding dimension and delay factor; Calculating all dispersion modes corresponding to the feature embedding vector, and respectively calculating the probability of each dispersion mode; A second entropy value of the hierarchical component is obtained based on the information entropy combined with the probabilities of the various dispersion modes.

7. The bidirectional power supply intelligent fault diagnosis method according to claim 4, characterized in that: The step of inputting the hierarchical components in the third sub-feature signal group into the third sub-learner to obtain a plurality of third entropy values ​​comprises: For each of the hierarchical components input into the third sub-learner, performing phase space reconstruction based on the hierarchical components to obtain multiple trajectory sequences; respectively calculating the cosine similarities between any two adjacent trajectory sequences, and generating a similarity vector based on all the cosine similarities; Divide the target interval into multiple subintervals, and use the histcounts function to count the similarity vectors in each subinterval according to the multiple subintervals, to obtain the statistical probability of the similarity vector in each subinterval; A third entropy value of the hierarchical component is obtained based on the information entropy and the statistical probability of the similarity vector in each of the subintervals.

8. An electronic device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When at least one of the programs is executed by at least one of the processors, the bidirectional power supply intelligent fault diagnosis method according to any one of claims 1 to 7 is implemented.

9. A computer-readable storage medium, characterized in that A program executable by a processor is stored therein, and when the program executable by the processor is executed by the processor, it is used to implement the bidirectional power supply intelligent fault diagnosis method according to any one of claims 1 to 7.

Images