Method, device and equipment for determining abnormity of transformer of electric shovel and storage medium

Abstract

The invention provides an electric shovel transformer abnormity determination method, device and equipment and a storage medium, and the method comprises the steps: collecting the original voiceprint data and cable current waveform of an electric shovel transformer through a voiceprint monitoring system; preprocessing and framing are carried out on the original voiceprint data, a fundamental wave phase is tracked based on the original voiceprint data to generate a coefficient of a wave trap, fundamental waves and harmonic waves in frame voiceprints are suppressed and distinguished through the wave trap, and a reduced voiceprint signal is obtained; predicting and offsetting residual harmonic energy in the reduced voiceprint signal by using an adaptive filter and taking a cable current waveform as a reference to obtain a primary noise reduction signal; performing feature extraction based on the primary noise reduction signal to obtain a feature vector; and identifying the current working condition of the electric shovel transformer through the lightweight decision model, performing normalization processing on the feature vector through the mean variance weight table corresponding to the current working condition, and inputting the normalized vector into the transformer normal voiceprint Gaussian mixture model corresponding to the current working condition to obtain a judgment result of whether the transformer is abnormal or not.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault diagnosis, and in particular to a method, device and equipment for determining abnormality of an electric shovel transformer and a storage medium. BACKGROUND

[0002] As the core equipment in mine operation, the electric shovel undertakes a large amount of material excavation and transportation tasks, and its running state is directly related to the efficiency and safety of the entire mine production. In order to ensure the stable operation of the electric shovel and reduce downtime caused by faults, it is particularly important to monitor and judge whether there is an abnormality in the key components such as the electric shovel transformer in real time.

[0003] In the related art, firstly, in the soundprint processing of the electric shovel transformer, the traditional method relies on single soundprint collection and simple processing, lacks effective dynamic suppression means for harmonic interference under the complex working conditions of the electric shovel, and is difficult to accurately extract the transformer mechanical vibration main frequency signal, resulting in low purity and unobvious features of the obtained soundprint information, which affects the accuracy of subsequent fault diagnosis. Secondly, the existing technology has limited feature dimension, which is mainly concentrated in the conventional time-frequency or frequency domain features, ignores the coupling relationship between the soundprint and the load current, and cannot fully represent the running state of the transformer, resulting in insufficient distinguishability of the feature vector for different working conditions and fault types, which is difficult to meet the demand of accurate diagnosis. In addition, the existing technology rarely combines the electric shovel working condition to judge the soundprint information, resulting in poor recognition accuracy due to different working conditions, and lacks a mechanism for dynamically updating the normalization parameters according to the working conditions, so that the feature amplitude difference under different working conditions is not properly eliminated, the comparability of the feature vector is poor, and the reliability of the abnormality detection is further reduced.

[0004] In summary of the above analysis of the development status of the technical field, the existing technology lacks a scheme that combines trap wave suppression and adaptive filter processing of harmonic interference, considers load modulation features in the extraction process of the feature vector, and sets up a normalization mechanism and an abnormality judgment mechanism under different working conditions. SUMMARY

[0005] The present application relates to the technical field of fault diagnosis, and in particular to a method, device and equipment for determining abnormality of an electric shovel transformer and a storage medium.

[0006] According to a first aspect of an embodiment of the present application, a method for determining abnormality of an electric shovel transformer is provided, comprising: acquiring original soundprint data of the electric shovel transformer and cable current waveform through a soundprint monitoring system; The original voiceprint data is preprocessed and framed, coefficients of a notch filter are generated based on tracking of fundamental wave phases of the original voiceprint data, fundamental waves and harmonics in the frame voiceprint are suppressed by the notch filter, a restored voiceprint signal is obtained, and a first denoising signal is obtained by using an adaptive filter to predict and offset residual harmonic energy in the restored voiceprint signal with the cable current waveform as a reference. Feature extraction is performed based on the first denoising signal, and a feature vector is obtained. A current working condition of the electric shovel transformer is identified through a lightweight decision model, the feature vector is normalized through a mean-variance weight table corresponding to the current working condition, the normalized vector is input into a transformer normal voiceprint Gaussian mixture model corresponding to the current working condition, and a judgment result of whether it is abnormal is obtained.

[0007] According to a second aspect of the embodiment of the present application, a device for determining an electric shovel transformer abnormality is provided, comprising: A data acquisition module is configured to acquire original voiceprint data of the electric shovel transformer and a cable current waveform through a voiceprint monitoring system. A filter processing module is configured to preprocess and frame the original voiceprint data, generate coefficients of a notch filter based on tracking of fundamental wave phases of the original voiceprint data, suppress fundamental waves and harmonics in the frame voiceprint through the notch filter, obtain a restored voiceprint signal, and obtain a first denoising signal by using an adaptive filter to predict and offset residual harmonic energy in the restored voiceprint signal with the cable current waveform as a reference. A feature extraction module is configured to perform feature extraction based on the first denoising signal, and obtain a feature vector. A judgment module is configured to identify a current working condition of the electric shovel transformer through a lightweight decision model, normalize the feature vector through a mean-variance weight table corresponding to the current working condition, input the normalized vector into a transformer normal voiceprint Gaussian mixture model corresponding to the current working condition, and obtain a judgment result of whether it is abnormal.

[0008] According to a third aspect of the embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program is executed by the processor to implement the steps of the electric shovel transformer abnormality determination method provided in the first aspect of the present application.

[0009] According to a fourth aspect of the embodiment of the present application, a computer readable storage medium is provided, which stores an information transmission implementation program, wherein the program is executed by a processor to implement the steps of the electric shovel transformer abnormality determination method provided in the first aspect of the present application.

[0010] The technical solution provided by the embodiments of the present invention includes the following beneficial effects: using an adaptive filter with the cable current waveform as a reference, the remaining harmonic energy in the restored acoustic signature signal is predicted and canceled, significantly improving the signal purity and laying a solid foundation for subsequent analysis; the current operating condition of the electric shovel transformer is identified through a lightweight decision model, and the corresponding mean-variance weight table and normal acoustic signature Gaussian mixture model are called to ensure the fairness and accuracy of feature comparison under all operating conditions.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in one or more embodiments of this specification or in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart of the method for determining abnormalities in an electric shovel transformer according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the electric shovel transformer anomaly determination device according to an embodiment of the present invention; Figure 3 This is a schematic diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0014] To enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this document.

[0015] Method Implementation Examples According to an embodiment of the present invention, a method for determining abnormalities in an electric shovel transformer is provided. Figure 1 This is a flowchart of the method for determining transformer anomalies in an electric shovel according to an embodiment of the present invention, as follows: Figure 1 As shown, the method for determining transformer anomalies in an electric shovel according to an embodiment of the present invention specifically includes: In step S110, the original acoustic signature data of the electric shovel transformer and the cable current waveform are collected through the acoustic signature monitoring system, specifically including: In this embodiment of the invention, the method for determining the anomaly of the electric shovel transformer relies on a voiceprint monitoring system. This system not only collects the voiceprint information of the transformer inside the electric shovel equipment, storing the voiceprint information in various forms such as waveform diagrams, time domain diagrams, and spectrum diagrams, but also simultaneously records the instantaneous current and waveform characteristics of the power supply cable. All data is stored in a cloud database and interacts in real time with multiple platforms such as big data base and data middleware to achieve unified management and service of data.

[0016] Before data collection, it is necessary to identify the cable-powered electric shovel to be monitored, as its transformer is as important as the heart and needs to be monitored closely. The original acoustic signature data of the electric shovel transformer is obtained through an acoustic signature monitoring system, and the cable current waveform is collected simultaneously.

[0017] In step S120, the original acoustic data is preprocessed and framed. Based on the original acoustic data, the fundamental phase is tracked to generate the coefficients of a notch filter. The notch filter suppresses and distinguishes the fundamental and harmonic waves in the frame acoustic data, resulting in a restored acoustic signal. An adaptive filter, using the cable current waveform as a reference, predicts and cancels the remaining harmonic energy in the restored acoustic signal, resulting in a first-stage noise reduction signal. Specifically, this includes: The raw voiceprint data was preprocessed and framed using the Hanning window method; A phase-locked loop is used to track the fundamental phase and generate IIR notch filter coefficients. IIR stands for Infinite Impulse Response. The coefficients are configured in the notch filter, and the fundamental, third, fifth and seventh harmonics are dynamically suppressed by the notch filter to obtain the restored acoustic signature signal.

[0018] After the first stage of dynamic suppression, other harmonics or non-periodic noise related to the load current may still remain in the signal. Using the LMS adaptive filter with the cable current waveform as a reference template (LMS stands for minimum mean square), the output analog signal is made to infinitely approximate the remaining harmonic energy in the restored acoustic signal, thus canceling out the remaining harmonic energy from the restored acoustic signal. In other words, the part that matches the main frequency of the transformer's mechanical vibration is finally extracted as the first noise reduction signal.

[0019] In step S130, feature extraction is performed based on the first denoised signal to obtain a feature vector, specifically including: Extract time-domain features, frequency-domain features, and load modulation features from the first noise-reduced signal as feature vectors; In this embodiment of the invention, the time-domain features include short-time energy, signal kurtosis, and zero crossover rate; the frequency-domain features include the energy ratio of each segment within the 50-550Hz range; and the load modulation features include the current-acoustic cross-spectral coefficient, with the feature vector being a 16-dimensional feature vector, comprehensively characterizing the time-frequency characteristics of the signal and its coupling relationship with the load state.

[0020] In step S140, the current operating condition of the electric shovel transformer is identified using a lightweight decision model. The feature vector is normalized using a mean-variance weight table corresponding to the current operating condition. The normalized vector is then input into the normal sound signature Gaussian mixture model of the transformer corresponding to the current operating condition to obtain a judgment result on whether it is abnormal. Specifically, this includes: The motor current RMS, gear signal, and hydraulic pump pressure of the electric shovel transformer are obtained through the CAN bus as decision data. RMS refers to the root mean square. A decision tree model is used as a lightweight decision model. The decision data is used to create a lightweight decision model. The lightweight decision model is used to determine the current working condition code. The working condition code is divided into light load, medium load and heavy load. The specific code for light load is G01, the specific code for medium load is G02, and the specific code for heavy load is G03.

[0021] Obtain the corresponding mean-variance weight table, calculate the difference between the eigenvalues ​​of the original eigenvector and the corresponding mean, and divide the difference by the corresponding variance to obtain the normalized vector, i.e. (original eigenvalues ​​- corresponding working condition mean) / corresponding working condition variance. Normalization eliminates the difference in eigenvalue amplitude under different working conditions. Preferably, the mean-variance weight table is updated every 30 seconds according to the RMS of the motor current to ensure that the parameters reflect the working status of the electric shovel in real time.

[0022] The normalized vector is input into the normal soundprint Gaussian mixture model of the transformer. The Gaussian mixture model consists of K Gaussian distribution components. The probability of the distribution component to which it belongs is calculated. The distribution component that is most likely to belong is obtained, which is the maximum probability distribution component. The Mahalanobis distance between the normalized vector and the mean vector in the maximum probability distribution component is calculated. If the Mahalanobis distance exceeds the preset threshold, it is marked as abnormal; otherwise, it is marked as normal. Preferably, to avoid accidental misjudgment, a multiple continuous detection and judgment mechanism is adopted. If n consecutive feature vectors are all marked as abnormal candidates, the relevant information is sent to the operation terminal to prompt the operator that abnormal data has been generated.

[0023] In summary, addressing the existing problems, this invention presents a method for determining anomalies in electric shovel transformers. It sequentially uses a notch filter to suppress and differentiate the fundamental and harmonic frequencies in the frame acoustic signature. Further, an adaptive filter, referencing the cable current waveform, predicts and cancels the remaining harmonic energy in the restored acoustic signature signal, significantly improving signal purity and laying a solid foundation for subsequent analysis. During feature extraction, a feature vector encompassing time, frequency, and load modulation features is constructed to comprehensively capture the deep coupling relationship between acoustic signatures and load status, resulting in richer feature representation and more accurate diagnosis. A lightweight decision model identifies the current operating condition of the electric shovel transformer and utilizes the corresponding mean-variance weighting table and a normal acoustic signature Gaussian mixture model to ensure the fairness and accuracy of feature comparisons across all operating conditions. Overall, through refined signal processing, feature extraction, and normalization, combined with a statistical model, the method achieves rapid and accurate determination of abnormal data in electric shovel transformers, effectively improving the efficiency and accuracy of electric shovel fault diagnosis and ensuring the safe and stable operation of electric shovels.

[0024] Device Examples According to an embodiment of the present invention, a device for determining abnormalities in an electric shovel transformer is provided. Figure 2 This is a schematic diagram of the electric shovel transformer anomaly determination device according to an embodiment of the present invention, as shown below. Figure 2 As shown, the electric shovel transformer anomaly determination device according to an embodiment of the present invention specifically includes: Data acquisition module 20 is used to collect the original acoustic signature data and cable current waveform of the electric shovel transformer through the acoustic signature monitoring system; The filtering module 22 is used to preprocess and frame the original acoustic data. Based on the original acoustic data, it tracks the fundamental phase to generate the coefficients of a notch filter. The notch filter suppresses and distinguishes the fundamental and harmonic waves in the framed acoustic data to obtain the restored acoustic signal. An adaptive filter, using the cable current waveform as a reference, predicts and cancels the remaining harmonic energy in the restored acoustic signal to obtain a primary noise reduction signal. Specifically, this is used for: The original acoustic print data is preprocessed and framed. Based on the original acoustic print data, the fundamental phase is tracked to generate the coefficients of the notch filter. The notch filter is used to suppress and distinguish the fundamental and harmonic waves in the frame acoustic print to obtain the restored acoustic print signal. Specifically, this includes: The raw voiceprint data was preprocessed and framed using the Hanning window method; The fundamental phase is tracked using a phase-locked loop to generate IIR notch filter coefficients. The fundamental, third, fifth, and seventh harmonics are dynamically suppressed by the notch filter to obtain the restored acoustic signature signal.

[0025] Using an LMS adaptive filter with the cable current waveform as a reference template, an analog signal is output to make the analog signal infinitely close to the remaining harmonic energy in the restored acoustic signal, thereby canceling out the remaining harmonic energy from the restored acoustic signal and obtaining a noise-reduced signal.

[0026] Feature extraction module 24 is used to extract features from the first-stage denoised signal to obtain feature vectors, specifically for: The time-domain features, frequency-domain features, and load modulation features are extracted from the noise-reduced signal as feature vectors.

[0027] The judgment module 26 is used to identify the current operating condition of the electric shovel transformer through a lightweight decision model. It normalizes the feature vector using a mean-variance weight table corresponding to the current operating condition, and then inputs the normalized vector into the transformer's normal acoustic signature Gaussian mixture model corresponding to the current operating condition to obtain a judgment result on whether it is abnormal. Specifically, it is used for: The motor current RMS, gear signal, and hydraulic pump pressure of the electric shovel transformer are obtained as decision data, and a decision tree model is used as a lightweight decision model. The decision data is used to create a lightweight decision model. This lightweight decision model determines the current operating condition code, which is divided into light load, medium load, and heavy load.

[0028] Obtain the corresponding mean-variance-weight table, calculate the difference between the eigenvalues ​​of the original eigenvector and the corresponding mean, and divide the difference by the corresponding variance to obtain the normalized vector.

[0029] Input the normalized vector into the normal sound signature Gaussian mixture model of the transformer, calculate the probability of the distribution component to which it belongs, and calculate the Mahalanobis distance between the normalized vector and the mean vector in the component with the highest probability distribution. If the Mahalanobis distance exceeds a preset threshold, it is marked as abnormal; otherwise, it is marked as normal.

[0030] In summary, addressing the existing problems, this invention, an electric shovel transformer anomaly determination device, sequentially uses a notch filter to suppress and differentiate the fundamental and harmonic frequencies in the frame acoustic signature. Further, an adaptive filter, using the cable current waveform as a reference, predicts and cancels the remaining harmonic energy in the restored acoustic signature signal, significantly improving signal purity and laying a solid foundation for subsequent analysis. During feature extraction, a feature vector encompassing time, frequency, and load modulation features is constructed to comprehensively capture the deep coupling relationship between acoustic signature and load status, resulting in richer feature representation and more accurate diagnosis. A lightweight decision model identifies the current operating condition of the electric shovel transformer and calls the corresponding mean-variance weight table and normal acoustic signature Gaussian mixture model to ensure the fairness and accuracy of feature comparison under all operating conditions. Overall, through refined signal processing, feature extraction, and normalization, combined with a statistical model, the device achieves rapid and accurate determination of electric shovel transformer anomaly data, effectively improving the efficiency and accuracy of electric shovel fault diagnosis and ensuring the safe and stable operation of the electric shovel.

[0031] Electronic device examples Figure 3This is a schematic diagram of an electronic device according to an embodiment of the present invention. The electronic device 300 may include at least one processor 310 and a memory 320. The processor 310 can execute instructions stored in the memory 320. The processor 310 is communicatively connected to the memory 320 via a data bus. In addition to the memory 320, the processor 310 can also be communicatively connected to an input device 330, an output device 340, and a communication device 350 via the data bus.

[0032] Processor 310 can be any conventional processor, such as a commercially available CPU. Processors may also include graphics processing units (GPUs), field-programmable gate arrays (FPGAs), systems-on-chips (SoCs), application-specific integrated circuits (ASICs), or combinations thereof.

[0033] The memory 320 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0034] In this embodiment of the present disclosure, the memory 320 stores executable instructions, and the processor 310 can read the executable instructions from the memory 320 and execute the instructions to implement all or part of the steps of the electric shovel transformer anomaly determination method in any of the exemplary embodiments described above.

[0035] Computer-readable storage medium embodiments In addition to the methods and apparatus described above, exemplary embodiments of this disclosure may also be a computer program product or a computer-readable storage medium storing the computer program product, the computer product including computer program instructions that can be executed by a processor to implement all or part of the steps described in any of the above exemplary embodiments of the electric shovel transformer anomaly determination method.

[0036] Computer program products can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. Programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages, and scripting languages ​​(e.g., Python). The program code can be executed entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0037] Computer-readable storage media may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example,, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media include: static random access memory (SRAM) having one or more electrically connected wires; electrically erasable programmable read-only memory (EEPROM); erasable programmable read-only memory (EPROM); programmable read-only memory (PROM); read-only memory (ROM); magnetic storage; flash memory; magnetic disk or optical disk; or any suitable combination thereof.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method of shovel transformer anomaly determination, the method comprising: The method comprises the following steps: Collecting original acoustic fingerprint data of the electric shovel transformer and cable current waveform through an acoustic fingerprint monitoring system; Preprocessing and framing the original acoustic fingerprint data, generating coefficients of a notch filter based on tracking of fundamental wave phase of the original acoustic fingerprint data, suppressing fundamental wave and harmonic wave in the framed acoustic fingerprint through the notch filter, and obtaining a restored acoustic fingerprint signal; Extracting features based on the first denoising signal to obtain a feature vector; Identifying the current working condition of the electric shovel transformer through a lightweight decision model, normalizing the feature vector through a mean-variance weight table corresponding to the current working condition, inputting the normalized vector into a transformer normal acoustic fingerprint Gaussian mixture model corresponding to the current working condition, and obtaining a judgment result of whether it is abnormal.

2. The method of claim 1, wherein, The preprocessing and framing of the original acoustic fingerprint data, the generation of coefficients of a notch filter based on tracking of fundamental wave phase of the original acoustic fingerprint data, the suppression of fundamental wave and harmonic wave in the framed acoustic fingerprint through the notch filter, and the obtaining of a restored acoustic fingerprint signal specifically comprise: Using the Hamming window method to preprocess and frame the original acoustic fingerprint data; Tracking the fundamental wave phase using a phase-locked loop and generating IIR notch filter coefficients, dynamically suppressing the fundamental wave, 3rd harmonic wave, 5th harmonic wave and 7th harmonic wave through the notch filter, and obtaining the restored acoustic fingerprint signal.

3. The method of claim 1, wherein, The use of an adaptive filter to predict and cancel the remaining harmonic energy in the restored acoustic fingerprint signal based on the cable current waveform specifically comprises: Using an LMS adaptive filter to use the cable current waveform as a reference template, outputting an analog signal to make the analog signal infinitely approach the remaining harmonic energy in the restored acoustic fingerprint signal, canceling the remaining harmonic energy from the restored acoustic fingerprint signal, and obtaining the first denoising signal.

4. The method of claim 1, wherein, The feature extraction based on the first denoising signal to obtain a feature vector specifically comprises: Extracting time domain features, frequency domain features and load modulation features from the first denoising signal as the feature vector.

5. The method of claim 1, wherein, The identification of the current working condition of the electric shovel transformer through a lightweight decision model specifically comprises: Obtaining motor current RMS, gear signal and hydraulic pump pressure of the electric shovel transformer as decision data, and using a decision tree model as the lightweight decision model; Inputting the decision data into the lightweight decision model to determine the working condition code of the current working condition through the lightweight decision model, wherein the working condition code is divided into light load, medium load and heavy load.

6. The method of claim 1, wherein, The normalization of the feature vector through the mean-variance weight table corresponding to the current working condition specifically comprises: Obtaining the corresponding mean-variance weight table, calculating the difference between the feature value of the original feature vector and the corresponding mean value, and dividing the difference by the corresponding variance to obtain the normalized vector.

7. The method of claim 1, wherein, The inputting of the normalized vector into the transformer normal acoustic fingerprint Gaussian mixture model corresponding to the current working condition to obtain a judgment result of whether it is abnormal specifically comprises: input the normalized vector into the transformer normal voiceprint Gaussian mixture model, calculate the probability of belonging to the distribution component, calculate the Mahalanobis distance between the normalized vector and the mean vector in the maximum probability distribution component; if the Mahalanobis distance exceeds a preset threshold, it is marked as abnormal, otherwise it is marked as normal.

8. A dragline transformer abnormality determination device characterized by comprising: It comprises: a data acquisition module for collecting original voiceprint data and cable current waveform of the electric shovel transformer through a voiceprint monitoring system; a filtering processing module for pre-processing and framing the original voiceprint data, generating the coefficients of the notch filter based on the tracking of the fundamental wave phase of the original voiceprint data, suppressing and distinguishing the fundamental wave and harmonic wave in the frame voiceprint through the notch filter, obtaining a restored voiceprint signal; using an adaptive filter to take the cable current waveform as a reference, predicting and canceling the residual harmonic energy in the restored voiceprint signal, obtaining a first noise reduction signal; a feature extraction module for feature extraction based on the first noise reduction signal, obtaining a feature vector; a judgment module for identifying the current working condition of the electric shovel transformer through a lightweight decision model, normalizing the feature vector through the mean variance weight table corresponding to the current working condition, inputting the normalized vector into the transformer normal voiceprint Gaussian mixture model corresponding to the current working condition, and obtaining the judgment result of whether it is abnormal.

9. An electronic device, comprising: It comprises: a memory, a processor and a computer program stored on the memory and executable on the processor, which implements the steps of the electric shovel transformer abnormality determination method according to any one of claims 1 to 7 when executed by the processor.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores an information transmission implementation program, and the program is executed by the processor to implement the steps of the electric shovel transformer abnormality determination method according to any one of claims 1 to 7.

Images