Switch cabinet fault voiceprint detection method and device, medium and program product

Abstract

The invention discloses a switch cabinet fault voiceprint detection method and device, a medium and a program product, and relates to the field of switch cabinets, and the method comprises the steps: obtaining an initial sound signal and current environment information during the operation of a to-be-detected switch cabinet; adjusting the initial sound signal according to the current environment information to obtain an adjusted sound signal; extracting voiceprint features based on the adjusted sound signal; inputting the voiceprint features into a preset voiceprint fault detection model to judge whether the switch cabinet has a fault or not; and if yes, determining the fault type of the switch cabinet through the voiceprint fault detection model. According to the invention, the sound signals are intelligently adjusted in combination with the environmental parameters, the influence of environmental changes on sound propagation characteristics is compensated, rapid and accurate diagnosis of switch cabinet faults is realized, and the automation level and reliability of fault detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of switch cabinets, and in particular to a switch cabinet fault voiceprint detection method, device, medium and program product. Background Art

[0002] Switchgear is a common and very important electrical equipment. Its main function is to open and close, control and protect electrical equipment in the process of power generation, transmission, distribution and power conversion in the power system. The components in the switchgear mainly include circuit breakers, disconnectors, load switches, operating mechanisms, transformers and other components. These devices require very precise control. Partial discharge and switch jamming will have a great impact on the normal operation and safety of the switchgear. Therefore, it is necessary to inspect the switchgear to promptly discover and eliminate potential faults.

[0003] Existing troubleshooting has begun to use voiceprint detection technology to replace traditional manual detection methods to improve detection accuracy and efficiency. However, current voiceprint detection technology usually ignores the impact of environmental factors on sound signals, resulting in unreliable detection accuracy. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a switch cabinet fault voiceprint detection method, device, electronic device and medium to solve the problem that the existing voiceprint detection ignores the influence of environmental factors on the sound signal, resulting in unreliable detection accuracy.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a switch cabinet fault voiceprint detection method, which is applied to a control processing unit of a detection device, wherein the detection device further includes a voiceprint collection unit and an environment perception unit, wherein the voiceprint collection unit and the environment perception unit are respectively connected to the control processing unit, and the method includes:

[0009] Obtain the initial sound signal and current environment information when the switch cabinet to be tested is in operation;

[0010] Adjusting the initial sound signal according to the current environment information to obtain an adjusted sound signal;

[0011] A voiceprint feature is extracted based on the adjusted sound signal, and the voiceprint feature is input into a preset voiceprint fault detection model to detect a fault condition of the switch cabinet.

[0012] As a preferred solution of the switch cabinet fault soundprint detection method of the present invention, the current environmental information includes environmental temperature and air humidity, and the initial sound signal is adjusted according to the current environmental information to obtain the adjusted sound signal, including:

[0013] Determining, according to the ambient temperature and the air humidity, a compensated ambient sound velocity, a compensated signal frequency, and a compensated signal amplitude of the initial sound signal;

[0014] Determining a signal time delay according to the compensated ambient sound speed and a preset reference sound speed;

[0015] The initial sound signal is adjusted according to the time delay, the compensation signal frequency, and the compensation signal amplitude to obtain an adjusted sound signal.

[0016] As a preferred solution of the switch cabinet fault soundprint detection method of the present invention, wherein: the adjusting the initial sound signal according to the time delay, the compensation signal frequency and the compensation signal amplitude to obtain the adjusted sound signal includes: adjusting the initial sound signal by a compensation algorithm to obtain the adjusted sound signal; the compensation algorithm is expressed as:

[0017]

[0018] Among them, S comp (t) is the adjusted sound signal as a function of time t, A adj is the compensation signal amplitude, S(t) is the initial sound signal as a function of time t, Δt is the time delay, j is an imaginary unit, and f adj is the compensation signal frequency.

[0019] As a preferred solution of the switch cabinet fault soundprint detection method of the present invention, wherein: the determining of the compensated ambient sound velocity, the compensated signal frequency and the compensated signal amplitude of the initial sound signal according to the ambient temperature and the air humidity comprises:

[0020] Determining the compensated ambient sound speed according to the ambient temperature and the air humidity;

[0021] determining the compensation signal frequency according to the air temperature;

[0022] The compensation signal amplitude is determined according to the air humidity.

[0023] As a preferred solution of the switch cabinet fault soundprint detection method of the present invention, wherein: the determining the compensated ambient sound speed according to the ambient temperature and the air humidity includes: compensating and calibrating the sound speed of the initial sound signal based on the ambient sound speed influence to obtain the compensated ambient sound speed; the ambient sound speed influence is expressed as:

[0024] c(T)=c0+k T (T-T0)

[0025] c RH =c(T)·(1+k RH RH)

[0026] Where T is the ambient temperature, T0 is the reference temperature, c(T) is the speed of sound corresponding to the ambient temperature, c0 is the preset reference speed of sound, and k T is the temperature influence coefficient, RH is the air humidity, k RH is the humidity influence coefficient, c RH To compensate for the ambient sound speed.

[0027] As a preferred solution of the switch cabinet fault voiceprint detection method described in the present invention, the voiceprint fault detection model adopts a lightweight neural network structure, and data compression processing is performed through the voiceprint fault detection model, and the compression processing includes parameter clipping or low-rank decomposition.

[0028] As a preferred solution of the switch cabinet fault voiceprint detection method described in the present invention, wherein: extracting voiceprint features based on the adjusted sound signal also includes: performing dimensionality reduction processing on the extracted voiceprint features to reduce the number of features in the voiceprint features.

[0029] In a second aspect, the present invention provides a switch cabinet fault voiceprint detection device, comprising: a control processing unit, a voiceprint collection unit and an environment perception unit, wherein the voiceprint collection unit and the environment perception unit are respectively connected to the control processing unit; the voiceprint collection unit is used to collect initial sound signals, and the environment perception unit is used to collect current environment information; the control processing unit includes one or more processors and memories;

[0030] The memory is coupled to the one or more processors, and the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and the one or more processors call the computer instructions to enable the control processing unit to execute the steps of the switch cabinet fault soundprint detection method as described in any one of claims 1 to 7.

[0031] In a third aspect, the present invention provides a computing device, comprising:

[0032] Memory and processor;

[0033] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the switch cabinet fault voiceprint detection method are implemented.

[0034] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the switch cabinet fault voiceprint detection method.

[0035] Compared with the prior art, the present invention has the following beneficial effects: the present invention integrates an environmental sensing unit to monitor and compensate for the impact of ambient temperature and humidity on sound signals in real time, thereby significantly improving the accuracy of voiceprint detection. The compensation algorithm that uses environmental parameters to adjust the sound signal ensures the reliability of voiceprint feature extraction under different environmental conditions, thereby making the fault detection results more accurate and reducing the possibility of false alarms and missed alarms. The use of a lightweight neural network structure, combined with model compression techniques such as parameter clipping and low-rank decomposition, achieves a high degree of optimization of the voiceprint fault detection model, which not only reduces the model's computing requirements and storage space, but also enables the model to run efficiently on edge devices, meeting the needs of modern industry for rapid response and processing, while reducing dependence on central computing resources.

[0036] The present invention realizes a fully automated process from sound signal collection, environmental parameter compensation, feature extraction, dimensionality reduction processing to intelligent diagnosis. Through the preset voiceprint fault detection model, the system can automatically identify and classify the fault type of the switch cabinet, improve the operation and maintenance efficiency, reduce the cost and time of manual detection, and improve the speed of fault response and the accuracy of maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0038] Figure 1 A method flow chart of a switch cabinet fault voiceprint detection method, device, electronic device and medium according to an embodiment of the present invention;

[0039] Figure 2 A schematic diagram of the system architecture of a switch cabinet fault voiceprint detection method, device, electronic equipment and medium according to an embodiment of the present invention;

[0040] Figure 3A schematic diagram of an electronic device architecture of a switch cabinet fault voiceprint detection method, device, electronic device, and medium according to an embodiment of the present invention;

[0041] Figure 4 A schematic diagram of an application scenario of a switch cabinet fault voiceprint detection method, device, electronic equipment and medium according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0043] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0045] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0046] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0047] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0048] Example 1

[0049] Reference Figure 1 , is an embodiment of the present invention, which provides a switch cabinet fault voiceprint detection method, which is applied to the control processing unit 101 of the detection device, and the detection device also includes a voiceprint collection unit 102 and an environment perception unit 103, and the voiceprint collection unit 102 and the environment perception unit 103 are respectively connected to the control processing unit 101, such as Figure 1 As shown, the method includes:

[0050] S1: Obtaining the initial sound signal and current environment information of the switch cabinet to be tested during operation;

[0051] It should be noted that voiceprint detection technology has shown the potential to replace traditional manual detection methods in switch cabinet fault detection, especially in improving the accuracy and efficiency of detection, but the current related technologies still have some defects and shortcomings. The main problem is that related technologies often ignore the important impact of environmental factors on sound signals.

[0052] For example, changes in ambient temperature and humidity will affect the propagation speed and attenuation characteristics of sound waves in the air, thereby changing the arrival time, frequency components and amplitude of the sound signal. If these environmental factors are not properly considered and compensated, the feature extraction and fault diagnosis of the voiceprint signal may produce errors, resulting in unreliable detection accuracy.

[0053] The embodiment of the present application realizes real-time monitoring of the operating status of the switch cabinet and accurate fault diagnosis by combining the voiceprint collection unit and the environmental perception unit; by collecting the sound signal of the switch cabinet and intelligently adjusting the sound signal according to the temperature and humidity information provided by the environmental perception unit, it effectively compensates for the impact of environmental changes on the sound propagation characteristics.

[0054] Furthermore, by extracting key voiceprint features from the adjusted sound signal and inputting them into the preset fault detection model, it is possible to more accurately determine whether a fault exists and determine the specific type of fault, thereby automating the detection process, improving the efficiency and accuracy of fault detection, and providing strong technical support for improving the reliability and operation and maintenance efficiency of the power system.

[0055] S2: adjusting the initial sound signal according to the current environment information to obtain an adjusted sound signal;

[0056] In an embodiment of the present application, the current environmental information includes ambient temperature and air humidity; the initial sound signal is adjusted according to the current environmental information to obtain an adjusted sound signal, including: determining the compensated ambient sound speed, compensated signal frequency and compensated signal amplitude of the initial sound signal according to the ambient temperature and the air humidity; determining the signal time delay according to the compensated ambient sound speed and a preset reference sound speed; adjusting the initial sound signal according to the time delay, the compensated signal frequency and the compensated signal amplitude to obtain an adjusted sound signal.

[0057] It should be noted that, by adopting the above embodiment, by determining the compensation environment sound speed, compensation signal frequency and compensation signal amplitude, this embodiment ensures that the sound signal can accurately reflect the actual operating status of the switch cabinet before feature extraction, thereby improving the accuracy and reliability of fault detection.

[0058] In the embodiment of the present application, adjusting the initial sound signal according to the time delay, the compensation signal frequency and the compensation signal amplitude to obtain the adjusted sound signal includes: adjusting the initial sound signal by a compensation algorithm to obtain the adjusted sound signal, and the compensation algorithm is expressed as:

[0059]

[0060] Among them, S comp (t) is the adjusted sound signal as a function of time t, A adj is the compensation signal amplitude, S(t) is the initial sound signal as a function of time t, Δt is the time delay, j is the imaginary unit, and f adj To compensate the signal frequency.

[0061] It should be noted that the above embodiment uses a specific compensation algorithm formula to accurately adjust the time and amplitude of the sound signal. Not only the frequency change of the signal is taken into account, but also the time delay of signal propagation is taken into account, ensuring the accuracy of signal processing and providing high-quality data for subsequent fault feature analysis.

[0062] In an embodiment of the present application, determining the compensated ambient sound velocity, compensated signal frequency and compensated signal amplitude of the initial sound signal according to the ambient temperature and air humidity includes: determining the compensated ambient sound velocity according to the ambient temperature and air humidity; determining the compensated signal frequency according to the air temperature; determining the compensated signal amplitude according to the air humidity.

[0063] It should be noted that the above-mentioned embodiment defines in detail how to determine the compensation ambient sound speed, signal frequency and signal amplitude based on environmental parameters, and provides a systematic method to quantify the specific impact of environmental factors on sound signals, thereby making the compensation process more accurate and operational.

[0064] In the embodiment of the present application, the sound velocity of the initial sound signal is compensated and calibrated according to the ambient sound velocity influence formula to obtain the compensated ambient sound velocity. The ambient sound velocity influence is expressed as:

[0065] c(T)=c0+k T (T-T0)

[0066] c RH =c(T)·(1+k RH RH)

[0067] Where T is the ambient temperature, T0 is the reference temperature, c(T) is the speed of sound corresponding to the ambient temperature, c0 is the preset reference speed of sound, and k T is the temperature influence coefficient, RH is the air humidity, k RH is the humidity influence coefficient, c RH To compensate for the ambient sound speed.

[0068] It should be noted that, by adopting the above-mentioned embodiment, the sound velocity is compensated and calibrated by the environmental sound velocity influence formula, which directly affects the time delay calculation of the sound signal and can ensure the time synchronization and accuracy of fault detection.

[0069] S3: extracting voiceprint features based on the adjusted sound signal, and inputting the voiceprint features into a preset voiceprint fault detection model to detect the fault condition of the switch cabinet;

[0070] In an embodiment of the present application, the voiceprint feature is input into a preset voiceprint fault detection model to determine whether the switch cabinet has a fault; if so, the fault type of the switch cabinet is determined by the voiceprint fault detection model.

[0071] In an embodiment of the present application, the voiceprint fault detection model adopts a lightweight neural network structure, and data compression processing is performed through the voiceprint fault detection model, and the compression processing includes parameter clipping or low-rank decomposition.

[0072] It should be noted that the above-mentioned embodiment adopts a lightweight neural network structure and undergoes model compression processing, so that the voiceprint fault detection model not only maintains efficient fault identification capabilities, but also can run efficiently on resource-constrained devices, thereby achieving miniaturization and low power consumption of the fault detection system.

[0073] In an embodiment of the present application, extracting voiceprint features based on the adjusted sound signal further includes: performing dimensionality reduction processing on the extracted voiceprint features to reduce the number of features in the voiceprint features.

[0074] It should be noted that, by adopting the above embodiment, the voiceprint features are reduced in dimension through principal component analysis technology. This step optimizes the feature selection process, removes redundant features, improves the training efficiency and prediction performance of the model, and simplifies the complexity of the fault detection model.

[0075] Example 2

[0076] Reference Figure 2 to Figure 4 , is an embodiment of the present invention. This embodiment is different from the first embodiment in that it provides a switch cabinet fault voiceprint detection device 100, such as Figure 2 As shown, it can be directly installed on the outer casing of the switch cabinet to realize the fault soundprint detection of the switch cabinet. The detection results and data can be transmitted to the back-end system for staff to view and process.

[0077] The detection device 100 of this embodiment can be used in a switch cabinet fault voiceprint detection method, such as Figure 3 As shown, the detection device 100 includes a control processing unit 101, a voiceprint collection unit 102 and an environment perception unit 103. The voiceprint collection unit 102 and the environment perception unit 103 are connected to the control processing unit 101 respectively.

[0078] The voiceprint collection unit 102 is composed of a sound sensor 1021 and a signal conditioning module 1022. The sound sensor 1021 can use a high-precision MEMS (Micro-Electro-Mechanical System) microphone to meet the 20Hz to 48kHz audible and ultrasonic frequency band requirements of the detection device. Since the sound signal received by the sound sensor 1021 may not be strong enough, the signal conditioning module 1022 is required to condition the sound signal collected by the sound sensor through impedance matching, bandpass filter, gain amplification and other parts to complete the matching, filtering and amplification of small signals.

[0079] Specifically, various noise interferences often exist in the signals collected by the sound sensor 1021, and are particularly obvious under the effect of the amplification circuit, so a filter circuit in the signal conditioning module 1022 is used to remove some of the noise. After being pre-processed by the signal conditioning module 1022, the sound signal is output to the main processing unit via the TDM or I2S interface.

[0080] The environmental sensing unit 103 includes a temperature sensor 1031 and a humidity sensor 1032. The temperature sensor 1031 is responsible for real-time detection of the ambient temperature, which is used to capture the possible impact of temperature changes on the propagation speed of sound waves and the performance of switchgear materials; while the humidity sensor 1032 is used to measure the moisture content in the air, which is crucial for evaluating the potential impact of humidity on the propagation of sound signals and the insulation performance of equipment. The data collected by the two sensors will be sent to the control processing unit and combined with the sound signal obtained by the voiceprint acquisition unit 102 to provide accurate environmental context information for the fault detection model, ensuring that the accuracy of voiceprint feature extraction and fault diagnosis is not affected by environmental factors, thereby improving the reliability and effectiveness of the entire switchgear fault detection device.

[0081] The control processing unit 101 is the executor of the switch cabinet fault soundprint detection method of this embodiment and the core of the entire detection device. Its internal processor reads and executes instructions from the memory, performs data processing such as preprocessing and noise reduction on the sound data, communicates with external devices through the serial port and other interfaces, and performs debugging and testing through the JTAG interface.

[0082] In this embodiment, the control processing unit 101 can select some low-power single-chip microcomputers as the system computing core. This embodiment does not specifically limit the specific model of the single-chip microcomputer, and the single-chip microcomputers commonly found in the market can be considered for selection. For example, the low-power single-chip microcomputer of the STM32H750 model series (STM32H750VBT6) uses the ARM Cortex-M7 core, has an operating frequency of up to 480MHz and a 1280KB Flash memory.

[0083] In some embodiments, the detection device further includes a data transmission module 104 for receiving the voiceprint detection result output by the control processing unit 101 and transmitting it to the backend system via RS485 or 100M network transmission. The staff can view the voiceprint detection result in real time through the backend system.

[0084] The detection device of this embodiment achieves the goals of miniaturization, low power consumption and low cost by adopting a series of targeted design strategies. In terms of miniaturization, the system volume and the number of components are reduced by optimizing the circuit board design, selecting low-cost plug-in PCB terminals, and adopting a sensor collection integrated structure. In terms of low-power design, a low-power single-chip microcomputer STM32H750 is selected as the main processor, and measures such as sleep mode, low-power chip selection, and reasonable power management are taken in the hardware design. At the same time, algorithm optimization is also performed on the software level to reduce the computing load. Regarding low-cost design, the material and production costs are reduced by the low-cost design principle of the circuit board, the selection of cost-effective components and simplified structural design, and the design that is easy to install and maintain also reduces the operation and maintenance costs. These comprehensive measures enable the detection device to achieve cost-effectiveness and wide applicability while maintaining high performance.

[0085] In this embodiment, the detection device 100 can be applied to a switch cabinet fault soundprint detection method, which may include the following steps:

[0086] Obtain the initial sound signal and current environment information of the switch cabinet to be tested during operation.

[0087] The initial sound signal is collected by the voiceprint collection unit 102, and the current environment information is collected by the environment perception unit 103. Further, the current environment information may include but is not limited to the environment temperature and air humidity.

[0088] Specifically, the control processing unit 101, as the core of the detection device, realizes the acquisition of the initial sound signal and current environmental information of the switch cabinet during operation through communication connection with the voiceprint collection unit 102 and the environmental perception unit 103. The voiceprint collection unit 102 is equipped with a high-precision sound sensor, which is responsible for capturing the sound signals generated by the mechanical operation of the switch cabinet, and converting these analog signals into digital signals and transmitting them to the control processing unit. At the same time, the environmental perception unit 103 includes a temperature sensor and a humidity sensor, which can monitor and record the temperature and humidity conditions of the surrounding environment in real time, and also sends the measurement results of these environmental parameters to the control processing unit in digital form.

[0089] In some embodiments, the voiceprint collection unit 102 and the environment perception unit 103 will also pre-process the collected signals and data to remove possible noise and irrelevant signal components to improve the accuracy of subsequent analysis. Pre-processing may include filtering, denoising, normalization and other operations to ensure the quality of the signal.

[0090] The initial sound signal is adjusted according to the current environment information to obtain an adjusted sound signal.

[0091] Specifically, the control processing unit 101 first receives the initial sound signal of the switch cabinet during operation captured by the voiceprint collection unit 102, and the current environmental information collected by the environmental perception unit 103. Subsequently, the control processing unit 101 adjusts the initial sound signal according to these environmental parameters, such as ambient temperature and air humidity. The adjustment process may include the application of a compensation algorithm to the sound signal to adapt to the impact of environmental changes on sound speed, signal attenuation and frequency characteristics.

[0092] For example, the control processing unit 101 adjusts the frequency response and time delay of the sound signal, as well as possible amplitude changes, according to environmental parameters to ensure that the accuracy of the signal is not interfered by environmental factors. Through adjustment, the control processing unit 101 ultimately generates an adjusted sound signal that can more accurately reflect the operating status of the switch cabinet, providing a reliable data basis for subsequent voiceprint feature extraction and fault detection.

[0093] It should be noted that it is necessary for this embodiment to adjust the initial sound signal based on the current environmental information, because environmental factors such as temperature and humidity will have a certain impact on the propagation characteristics of sound waves in the air. Changes in temperature and humidity will change the speed of sound, causing the propagation time and phase of the sound signal to shift, and will also affect the attenuation rate and frequency characteristics of the sound signal. If these environmental changes are not compensated, the collected sound signal will not accurately reflect the actual operating status of the switch cabinet, which may lead to inaccurate extraction of voiceprint features, and thus affect the accuracy and reliability of fault detection.

[0094] Therefore, by adjusting the sound signal according to the environmental information, the accuracy of signal processing and analysis can be ensured, the accuracy of fault detection can be improved, false alarms and missed alarms can be reduced, and the safe operation and timely maintenance of the switch cabinet can be ensured.

[0095] Extracting voiceprint features based on the adjusted sound signal, and inputting the voiceprint features into a preset voiceprint fault detection model to detect fault conditions of the switch cabinet;

[0096] Specifically, the control processing unit 101 first receives and analyzes the adjusted sound signal that has been calibrated by environmental information compensation to ensure the accuracy and reliability of the signal.

[0097] Next, the control processing unit 101 performs time-frequency analysis on the clean sound signal, usually using methods such as short-time Fourier transform (STFT) or wavelet transform to decompose the signal into time-frequency representation, so as to better understand the frequency characteristics of the signal in different time periods. This step is particularly important for capturing transient fault characteristics.

[0098] After obtaining the time-frequency representation, the control processing unit 101 further applies feature extraction technology to identify key information in the sound signal. In this embodiment, the feature extraction technology may include but is not limited to statistical features such as Mel-frequency cepstral coefficients (MFCC), spectral centroid, spectral bandwidth, energy and entropy. MFCC is a commonly used voiceprint feature that simulates the human auditory system and maps the spectrum to the Mel scale to obtain a more sensitive representation of the sound signal.

[0099] The extracted features are further optimized and selected to remove redundancy and retain the most representative features. This may involve principal component analysis (PCA) or other dimensionality reduction techniques to reduce the dimensionality of the feature space while maximizing the information of the original data. Dimensionality reduction not only improves processing speed, but also helps avoid model overfitting.

[0100] The control processing unit 101 may also use a more advanced feature selection method, such as model-based feature importance assessment, to select the most distinguishing features for the fault type. In addition, the control processing unit 101 may use a machine learning algorithm to further abstract and combine features to form a more advanced feature representation to enhance the generalization ability of the model.

[0101] In some embodiments, after extracting the voiceprint features from the adjusted sound signal, the voiceprint features may be subjected to dimensionality reduction processing by principal component analysis technology to reduce the number of features in the voiceprint features.

[0102] Among them, the purpose of dimensionality reduction is to reduce the number of features while retaining the most important information as much as possible to improve the performance of the model and reduce computational complexity. Principal component analysis (PCA) is a data dimensionality reduction technique that projects the original data into a new coordinate system through a linear transformation. In this embodiment, the new system is defined by the main eigenvectors (i.e., principal components) of the data. The principal components are sorted according to the variance contribution rate, the first principal component has the largest variance (i.e., the data variability is the largest), and the subsequent principal components are reduced in turn. PCA can remove redundancy and noise in the data while retaining important information of the data, which is convenient for data visualization, analysis, and subsequent modeling.

[0103] In this embodiment, the control processing unit 101 needs to receive a voiceprint feature set extracted from a sound signal. The voiceprint feature may include Mel Frequency Cepstral Coefficients (MFCC), energy, pitch, etc., which describe the characteristics of the sound signal from different angles. Since the original feature set may contain a large number of features, it not only increases the computational burden, but may also cause overfitting of the model.

[0104] In the PCA dimensionality reduction process, the control processing unit 101 first performs a centralization process on the feature set to ensure that the mean of each feature is zero. Then, the covariance matrix of the feature set is calculated to determine the correlation between the features. The covariance matrix reflects the distribution of data in the feature space.

[0105] The control processing unit 101 performs eigenvalue decomposition on the covariance matrix to obtain a set of eigenvalues ​​and corresponding eigenvectors. The eigenvalues ​​are sorted by size, and each eigenvalue corresponds to an eigenvector, which describes the main changes in the data in that direction. In this embodiment, the control processing unit 101 selects the first few largest eigenvalues ​​and their corresponding eigenvectors. The eigenvectors are called principal components, which represent the main change trends in the data.

[0106] The control processing unit 101 also projects the original feature set onto these principal components to generate a new feature set whose dimension is much lower than the original feature set. The new feature set retains most of the variance in the original data set while removing redundancy and noise, thereby achieving dimensionality reduction.

[0107] Through PCA dimensionality reduction processing, the control processing unit 101 not only reduces the number of voiceprint features, but also extracts the most representative features, providing a more refined and efficient input for the voiceprint fault detection model, which helps to improve the training speed and prediction performance of the model, while reducing the model's sensitivity to noise and irrelevant features. In addition, dimensionality reduction also helps to improve the generalization ability of the model, so that it can still maintain a high accuracy rate when facing new and unseen data. In short, PCA, as an effective dimensionality reduction technology, plays a vital role in the voiceprint fault detection device and provides strong support for achieving efficient and accurate fault detection.

[0108] The voiceprint feature is input into a preset voiceprint fault detection model to determine whether the switch cabinet has a fault.

[0109] The voiceprint fault detection model is constructed using open source machine learning algorithms, especially lightweight neural network structures optimized for voiceprint signals. The voiceprint fault detection model has been fully trained and can identify and distinguish the voiceprint features of the switch cabinet during normal operation and different types of faults. The control processing unit inputs the voiceprint features into the voiceprint fault detection model, which can perform nonlinear transformation and decision-making on the input features through forward propagation calculations, using the weights and biases learned during the training process, and finally output the possibility and type of fault.

[0110] The mathematics and pattern recognition techniques involved in this process are understandable, but transparent to the user, who only needs to focus on the final diagnosis provided by the model.

[0111] If it is determined that the switch cabinet has a fault, the fault type of the switch cabinet is determined by the voiceprint fault detection model. If there is no fault, the fault detection task of the switch cabinet is completed.

[0112] In this embodiment, the training of the voiceprint fault detection model involves multiple stages, aiming to enable the model to accurately identify and classify normal modes and various fault types in the voiceprint signals of the switch cabinet. The training process of the model can be as follows:

[0113] 1) Data collection: First, a large amount of voiceprint signal data is collected from the switch cabinet. The voiceprint signal data includes the sound generated when the switch cabinet is operating under different conditions, as well as the sound signals in known fault conditions.

[0114] 2) Data preprocessing: The collected raw sound signals are preprocessed to improve the data quality. Preprocessing may include denoising, filtering, normalization and other steps to ensure that the signal is clear and suitable for analysis.

[0115] 3) Feature extraction: Extract useful features from the preprocessed sound signal, such as Mel-frequency cepstral coefficients (MFCC), spectral features, time domain features, etc., to characterize the key attributes of the sound signal.

[0116] 4) Label assignment: Assign the correct label to each sample in the training dataset. The label indicates the category to which the sample belongs, such as normal operation or a specific type of failure.

[0117] 5) Model selection: Choose a suitable machine learning model as the basis for fault detection. In voiceprint detection, deep learning models such as convolutional neural networks (CNN) or recurrent neural networks (RNN) are usually selected.

[0118] 6) Model training: The model is trained using the labeled feature vectors and the corresponding sound signal samples; to minimize the difference between the predicted output and the actual label, this embodiment may use techniques such as cross-validation to evaluate the performance of the model and prevent overfitting.

[0119] 7) Hyperparameter Tuning: During the training process, the model’s hyperparameters, such as learning rate, batch size, number of layers, and number of neurons, are adjusted to improve the model’s accuracy and generalization ability.

[0120] 8) Model validation and testing: Use the validation set to evaluate the performance of the model and fine-tune the model based on the validation results. After training, use an independent test set to perform a final evaluation of the model to ensure that the model has good generalization ability to unseen data.

[0121] 9) Model optimization: Based on the training and testing results, adjust the model structure or training process to further improve the performance.

[0122] 10) Model deployment: Once the model has completed training and performs well on the test set, it can be deployed in the control processing unit for real-time monitoring and fault detection.

[0123] Through the above training process, the voiceprint fault detection model can learn to distinguish the characteristics of normal and abnormal sound signals, and provide accurate fault diagnosis information for the operation and maintenance of the switch cabinet.

[0124] The fault type of the switch cabinet is determined through the voiceprint fault detection model.

[0125] Among them, the fault types include but are not limited to partial discharge and switch jamming. The voiceprint fault detection model can identify and distinguish multiple fault modes including partial discharge and switch jamming, which is achieved by using a large amount of voiceprint data marked with specific fault types in the model training stage.

[0126] Partial discharge usually generates ultra-high frequency sound wave signals, while switch jamming may produce abnormal mechanical noise; the model learns these fault characteristics so that in actual applications, when the control processing unit 101 receives the sound signal transmitted by the voiceprint collection unit 102 and the environmental parameters provided by the environmental perception unit 103, it can accurately adjust and analyze these signals, extract key features, and evaluate whether the sound signal matches the known fault mode through the model's prediction mechanism.

[0127] Ultimately, the prediction results output by the model will indicate whether a fault exists and, if so, further determine the type of fault, such as partial discharge or switch jam, providing accurate diagnostic information for operation and maintenance personnel so that they can take appropriate maintenance or repair measures in a timely manner. This fault detection method based on voiceprint analysis not only improves the accuracy of fault identification, but also significantly improves the operation and maintenance efficiency and reliability of the switchgear through an automated diagnostic process.

[0128] In some embodiments, the voiceprint fault detection model adopts a lightweight neural network structure and undergoes a model compression process, wherein the model compression process includes parameter pruning or low-rank decomposition.

[0129] The voiceprint fault detection model uses a lightweight neural network structure, which can meet the needs of efficient operation on resource-constrained devices. Lightweight neural networks reduce the complexity and computational load of the model by simplifying the network structure, such as reducing the number of neural layers or the number of neurons in each layer.

[0130] In addition, the control processing unit 101 can adopt an edge computing device. In order to further optimize the model to adapt to the edge computing device, the model undergoes model compression processing, which usually includes two key technologies: parameter clipping and low-rank decomposition.

[0131] Parameter pruning reduces the number of parameters of a model by removing unimportant weights in a neural network without causing a significant drop in model performance, thereby reducing the model's storage requirements and computational complexity.

[0132] Low-rank decomposition is a technique that decomposes a large weight matrix into several small matrix products. Since the small matrix takes up less space while maintaining the main characteristics of the original weight matrix, the model can significantly reduce its parameter amount while still being able to approximate the behavior of the original model.

[0133] The voiceprint fault detection model processed by compression technology not only reduces the model size while maintaining relatively high accuracy, but also improves the energy efficiency when running on edge devices, enabling the fault detection model to perform switch cabinet fault detection tasks in real time and efficiently on low-power, low-cost hardware.

[0134] In some embodiments, adjusting the initial sound signal according to the current environment information to obtain the adjusted sound signal may specifically include the following steps:

[0135] (1) Determine the compensated ambient sound velocity, compensated signal frequency and compensated signal amplitude of the initial sound signal according to the ambient temperature and the air humidity.

[0136] Specifically, when determining the compensation parameters of the initial sound signal including the compensation ambient sound velocity, compensation signal frequency and compensation signal amplitude, the current ambient temperature and humidity are firstly accurately measured by the ambient sensing unit 103. The ambient information has a certain influence on the propagation characteristics of the sound wave in the air.

[0137] For example, an increase in temperature increases the speed of air molecules, thereby increasing the speed of sound; while an increase in humidity may change the density and viscosity of the air, which in turn affects the speed of sound and the attenuation of sound waves. Based on the relevant physical acoustic formulas and coefficients determined experimentally, the control processing unit calculates the compensated ambient sound speed to ensure that the time domain analysis of the sound signal accurately reflects the actual sound wave propagation time. At the same time, the control processing unit adjusts the frequency and amplitude of the sound signal according to the environmental parameters, and uses a compensation algorithm to correct the frequency and compensate the amplitude of the signal, taking into account the frequency offset and amplitude attenuation that may be caused by temperature and humidity, to ensure that the frequency content and energy level of the signal remain authentic during the analysis. The compensation operation enables the sound signal to be used more accurately for subsequent fault feature extraction and analysis.

[0138] In some embodiments, this step may specifically include:

[0139] (11) Determine the compensated ambient sound speed based on the ambient temperature and air humidity.

[0140] Among them, the process of determining the compensated ambient sound speed involves real-time measurement of ambient temperature and air humidity, and these two parameters have a direct impact on the speed at which sound waves propagate in the medium. The control processing unit 101 obtains the temperature and humidity data of the current environment through the environmental sensing unit 103. Based on the physical acoustic principle, the speed of sound increases with the increase of temperature, and the change of humidity will also affect the speed of sound. The control processing unit 101 uses these data and combines the relevant formula of the relationship between the speed of sound and the environmental parameters to calculate the speed of sound in the current environment. The calculation result is called the compensated ambient sound speed, which reflects the actual propagation speed of sound waves under specific environmental conditions. In this way, the control processing unit 101 can accurately calibrate the time of the sound signal to ensure the accuracy of signal processing and fault detection.

[0141] In some embodiments, this step may perform compensation calibration on the sound velocity of the initial sound signal according to the ambient sound velocity influence formula to obtain a compensated ambient sound velocity, and the ambient sound velocity influence is expressed as:

[0142] c(T)=c0+k T (T-T0)

[0143] c RH =c(T)·(1+k RH RH)

[0144] Where T is the ambient temperature, T0 is the reference temperature, c(T) is the speed of sound corresponding to the ambient temperature, c0 is the preset reference speed of sound, and k T is the temperature influence coefficient, which can be 0.2 to 0.8 m / s °C, RH is the air humidity, k RH is the humidity influence coefficient, c RH To compensate for the ambient sound velocity, the value can be set to 0.1 to 0.3, c RH To compensate for the ambient sound speed.

[0145] This embodiment has found through a large number of experiments that the speed of sound is affected by ambient temperature and humidity, and this influence is similar to the influence of positive correlation. Therefore, the ambient speed of sound influence formula is designed in combination with a large number of test data.

[0146] First, the formula c(T) = c0 + k T (T-T0) describes the relationship between the speed of sound and temperature. Where c0 is the standard speed of sound at the reference temperature T0, for example, the reference temperature T0 is set to 20°C; k T It can represent the influence coefficient of temperature on the speed of sound, and is used to quantify the specific influence of temperature change on the speed of sound. It is obtained through the analysis of a large number of experimental data. In this embodiment, k T The value of can be 0.35~0.68 / ℃, which shows that temperature has a positive correlation with the speed of sound.

[0147] Next, formula c RH =c(T)·(1+k RH RH) introduces the effect of humidity on the speed of sound based on the consideration of temperature, where c(T) is the speed of sound adjusted according to temperature, k RH is the humidity influence coefficient, RH is a measure of humidity. The compensated ambient sound velocity formula shows that an increase in humidity will lead to a further adjustment in the sound velocity. For every unit increase in humidity, the sound velocity will increase by approximately k RH times the ratio.

[0148] These two formulas can be combined to accurately calculate the speed of sound under specific temperature and humidity conditions, that is, the compensated ambient sound speed c RH This calculation is crucial for voiceprint detection devices because it ensures that the time delay and propagation characteristics of the sound signal can be accurately compensated under different environmental conditions, thereby ensuring the accuracy and reliability of fault detection. In this way, the system can adapt to environmental changes and maintain the consistency of voiceprint feature extraction and analysis.

[0149] (12) Determine the compensation signal frequency according to the air temperature.

[0150] The determination of the compensation signal frequency is based on the effect of air temperature on the frequency characteristics of sound waves. Changes in ambient temperature not only affect the speed of sound, but may also cause changes in the frequency of the sound signal. The control processing unit calculates the frequency compensation value based on the collected air temperature data and the physical model of the effect of temperature on sound frequency. In this embodiment, the compensation takes into account the changes in the energy distribution of the sound waves when the temperature rises, as well as the possible frequency shift. By adjusting the frequency components of the sound signal, the control processing unit 101 can ensure that the frequency domain analysis of the signal is not interfered by temperature changes, thereby more accurately identifying and analyzing fault characteristics.

[0151] Specifically, the compensation signal frequency can be calculated by a frequency compensation formula, which includes:

[0152]

[0153] Among them, f adj is the frequency of the adjusted compensation signal, f is the frequency of the initial sound signal, T0 is the reference temperature, T is the current ambient temperature, k f is the frequency temperature coefficient.

[0154] Specifically, k fUsed to quantify the degree of impact of temperature changes on sound frequency, this coefficient is determined based on the physical properties of a specific medium (in this case, air) and experimental data. Through the frequency compensation formula, the frequency offset caused by temperature changes can be accurately predicted and compensated, ensuring the accuracy and reliability of the voiceprint signal in feature extraction and fault detection. It is essential to achieve accurate voiceprint fault detection, especially in an environment with temperature fluctuations.

[0155] The T+273.15 in the frequency compensation formula converts the temperature T to Kelvin to accommodate the physical model of the relationship between the speed of sound and temperature. The denominator 273.15+T0 ensures that the reference temperature is also converted to Kelvin, making the formula consistent when comparing temperature effects. The exponential term kf is used to adjust the actual effect of temperature changes on frequency. This coefficient is determined based on experimental data and material properties.

[0156] The design of the entire formula takes into account the change in the energy distribution of the sound wave when the temperature rises, and how this change affects the frequency of the sound signal. In this way, the sound signal can be accurately compensated for frequency, ensuring that the analysis results of the voiceprint detection device remain accurate even in an environment with temperature fluctuations. The frequency compensation of this embodiment is crucial for identifying and classifying switch cabinet faults, and helps maintain the consistency and reliability of frequency characteristics during signal processing.

[0157] (13) Determine the compensation signal amplitude according to the air humidity.

[0158] Among them, the determination of the compensation signal amplitude is related to the air humidity. Changes in humidity will affect the density of the air and the attenuation characteristics of the sound wave, and thus affect the propagation intensity of the sound signal. The control processing unit 101 measures the current humidity level and calculates the necessary amplitude compensation factor based on the influence of humidity on the propagation characteristics of the sound wave. The compensation in this embodiment helps to correct the signal attenuation or enhancement caused by humidity changes, ensuring that the amplitude information of the signal remains accurate during fault detection. Through amplitude compensation, the control processing unit can more accurately evaluate the energy level of the sound signal, providing a reliable data basis for the extraction of voiceprint features and the judgment of fault types.

[0159] In some embodiments, the compensation signal amplitude may be calculated using an amplitude compensation formula, the formula comprising:

[0160]

[0161] Among them, A adj is the adjusted amplitude, A is the amplitude of the initial sound signal, k amp is the coefficient of humidity on the amplitude, and RH is a measure of humidity.

[0162] In the amplitude compensation formula, the use of the exponential function e (the base of the natural logarithm) reflects the nonlinear effect of humidity on the amplitude of the sound signal. When the relative humidity RH increases, the density and viscosity of the air also increase, which will cause the energy attenuation of the sound wave during propagation to increase, so the amplitude of the signal will decrease. The coefficient k amp Quantifies the effect of humidity changes on signal amplitude attenuation. It is an experimentally determined value that reflects the expected change in signal amplitude for each unit change in humidity under a specific environment.

[0163] By applying the amplitude compensation formula, the sound signal can be properly adjusted to compensate for changes in signal strength due to changes in ambient humidity. Since amplitude compensation allows the system to maintain consistent performance under different humidity conditions, it can reduce false alarms and missed alarms and improve the reliability of fault detection.

[0164] (2) Determine the signal time delay based on the compensated ambient sound speed and the preset reference sound speed.

[0165] Among them, the step of determining the signal time delay is a key link in compensating for changes in the propagation time of sound waves. The control processing unit 101 determines the time delay of signal propagation based on the difference between the calculated compensated ambient sound speed and the preset reference sound speed. Changes in the speed of sound directly affect the time it takes for sound waves to reach the receiver, thereby affecting the time domain characteristics of the signal. For example, at higher temperatures, the speed of sound increases, and the time required for sound waves to propagate the same distance decreases, so the signal needs to be shifted forward in time to compensate for this change. The control processing unit accurately calculates the time delay and applies a time correction algorithm to time-shift the original sound signal to ensure that the time domain waveform of the signal is synchronized with the actual sound source event, providing a solid foundation for accurate signal analysis and fault detection.

[0166] In some embodiments, the signal time delay may be calculated by a delay formula, which is expressed as:

[0167]

[0168] Where Δt is the signal time delay, d is the distance from the sound source to the sound sensor, and c is the distance from the sound source to the sound sensor. RH To compensate for the ambient sound velocity, c0 is the standard sound velocity at the reference temperature T0.

[0169] The time delay formula of this embodiment takes into account the difference in the propagation speed of sound waves in different media, which is affected by environmental conditions such as temperature and humidity. By comparing the propagation time of sound waves at the compensated environmental sound speed and the standard sound speed, the time delay difference caused by environmental changes is determined. When sound waves propagate in the medium, the change in sound speed directly affects the time it takes for the sound waves to reach the sensor.

[0170] By calculating the signal time delay through the time delay formula, we can accurately calculate the impact of environmental factors on the propagation time of sound waves, and then make necessary time corrections to the sound signal, which can ensure the accuracy of the signal processing and analysis of the detection device, as well as the reliability and accuracy of fault detection.

[0171] (3) The initial sound signal is adjusted according to the time delay, the compensation signal frequency and the compensation signal amplitude to obtain an adjusted sound signal.

[0172] Specifically, after obtaining the compensated ambient sound velocity, compensated signal frequency, and compensated signal amplitude, the control processing unit 101 performs comprehensive adjustments on the initial sound signal to obtain an adjusted sound signal. This step involves multi-dimensional calibration of the signal, including applying time delay correction to ensure that the time domain waveform of the signal accurately reflects the sound source event; adjusting the signal frequency to compensate for the frequency offset caused by environmental parameters to ensure that the frequency domain characteristics of the signal are correct; and adjusting the signal amplitude to compensate for the sound wave attenuation caused by changes in air density and viscosity. Through adjustment, the control processing unit 101 can eliminate the influence of environmental factors on the sound signal, extract clearer and more accurate voiceprint features, and provide high-quality input data for subsequent fault detection models, thereby improving the accuracy and reliability of fault diagnosis.

[0173] In some embodiments, this step may adjust the initial sound signal by a compensation algorithm formula to obtain an adjusted sound signal, and the compensation algorithm formula is expressed as:

[0174]

[0175] Among them, S comp (t) is the adjusted sound signal as a function of time t, A adj is the compensation signal amplitude, S(t) is the initial sound signal as a function of time t, Δt is the time delay, j is the imaginary unit, and f adj is the compensation signal frequency.

[0176] Specifically, S. comp (t) is the compensated voiceprint signal as a function of time t. It is the result of the original signal after compensation for environmental factors. adj is the adjusted signal amplitude, which is adjusted according to the effect of humidity changes on the sound signal amplitude. adj is the original amplitude A multiplied by a humidity-dependent amplification factor. S(t) is the original voiceprint signal as a function of time t. Δt is the time delay due to the change in the speed of sound. The speed of sound changes with temperature, so the time it takes for the signal to propagate is affected. Δt is the time compensation calculated based on the speed of sound adjustment. This is a complex exponential term used to describe the phase change of the signal. Where j is the imaginary unit, f adj is the signal frequency adjusted according to temperature, 2πf adj represents the phase changing with time. adj It is the adjusted signal frequency, which is adjusted according to the effect of temperature changes on the sound signal frequency.

[0177] The compensation algorithm formula of this embodiment combines the amplitude, time delay and phase adjustment of the original signal to compensate for the impact of environmental factors on the voiceprint signal. It can maintain the accuracy of voiceprint signal analysis even when environmental conditions change. In practical applications, the compensation adjustment needs to be accurately calculated based on specific environmental parameters and the characteristics of the voiceprint signal.

[0178] This embodiment provides a comprehensive switch cabinet fault detection method, which improves the accuracy, efficiency and reliability of fault detection, realizes accurate detection of switch cabinet faults, and provides a solid guarantee for the safe and stable operation of the power system.

[0179] First, this embodiment realizes real-time monitoring of the operating status of the switch cabinet by integrating the sound sensor and the environmental sensor. The sound sensor is responsible for collecting the sound signal of the switch cabinet, while the environmental sensor monitors the environmental parameters that affect the propagation of sound waves, such as temperature and humidity, to ensure that the detection system can take into account the impact of environmental factors on the sound signal, thereby improving the accuracy of the detection results.

[0180] Secondly, the control processing unit of this embodiment uses advanced signal processing technology to pre-process and extract features of the collected sound signals. By compensating for environmental parameters, the control processing unit can adjust the sound signals to eliminate the impact of environmental changes and ensure that the accuracy of signal analysis is not affected by environmental fluctuations.

[0181] Next, principal component analysis technology was used to reduce the dimensionality of the extracted voiceprint features, reducing the number of features while retaining the most important information. This not only improved the efficiency of data processing, but also enhanced the model's ability to recognize fault features.

[0182] In addition, this embodiment uses a lightweight neural network structure as a voiceprint fault detection model. This model is processed by model compression techniques such as parameter pruning or low-rank decomposition to adapt to resource-constrained control processing units. The lightweight model reduces the demand for computing resources while maintaining high accuracy, making fault detection both fast and economical.

[0183] This embodiment also pays special attention to the training process of the model, and trains the model with a large amount of labeled sound data so that it can identify different types of switch cabinet faults, such as partial discharge and switch jamming, etc. The high generalization ability of the model ensures its reliability in practical applications.

[0184] Finally, this embodiment forms an efficient and automated fault detection system by comprehensively considering sound signal processing, environmental parameter compensation, feature dimension reduction, lightweight model design and training. This system not only reduces the labor intensity and subjectivity of manual detection, but also improves the speed and accuracy of fault detection, providing strong technical support for the maintenance and operation of switchgear.

[0185] The method provided in the above embodiment can be executed by a control processing unit of the detection device, and the control processing unit can be an electronic device. The electronic device in the embodiment of the present invention is described below from the perspective of hardware processing. Figure 4 , is a schematic diagram of a physical device structure of an electronic device in an embodiment of the present invention.

[0186] It should be noted that Figure 4 The structure of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0187] like Figure 4 As shown, the electronic device includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 to the random access memory (RAM) 403, such as executing the method in the above embodiment. In RAM 403, various programs and data required for system operation are also stored. CPU 401, ROM 402 and RAM 403 are connected to each other through bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0188] The following components are connected to the I / O interface 405: an input section 406 including an audio input device, a button switch, etc.; an output section 407 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed so that a computer program read therefrom is installed into the storage section 408 as needed.

[0189] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 409, and / or installed from the removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, various functions defined in the present invention are performed.

[0190] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0191] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0192] Specifically, the electronic device of this embodiment includes a processor and a memory, the memory is coupled to one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the electronic device to execute the method provided by the above embodiment.

[0193] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the electronic device, the electronic device implements the method provided in the above embodiment.

[0194] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

[0195] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0196] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

[0197] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A switch cabinet fault voiceprint detection method, characterized in that: A control processing unit (101) applied to a detection device, wherein the detection device further comprises a voiceprint collection unit (102) and an environment perception unit (103), wherein the voiceprint collection unit (102) and the environment perception unit (103) are respectively connected to the control processing unit (101), and the method comprises: Obtain the initial sound signal and current environment information of the switch cabinet to be tested during operation; Adjusting the initial sound signal according to the current environment information to obtain an adjusted sound signal; A voiceprint feature is extracted based on the adjusted sound signal, and the voiceprint feature is input into a preset voiceprint fault detection model to detect a fault condition of the switch cabinet.

2. The switch cabinet fault voiceprint detection method according to claim 1, characterized in that: The current environment information includes environment temperature and air humidity, and the adjusting the initial sound signal according to the current environment information to obtain the adjusted sound signal includes: Determining, according to the ambient temperature and the air humidity, a compensated ambient sound velocity, a compensated signal frequency, and a compensated signal amplitude of the initial sound signal; Determining a signal time delay according to the compensated ambient sound speed and a preset reference sound speed; The initial sound signal is adjusted according to the time delay, the compensation signal frequency, and the compensation signal amplitude to obtain an adjusted sound signal.

3. The switch cabinet fault voiceprint detection method according to claim 2, characterized in that: The adjusting the initial sound signal according to the time delay, the compensation signal frequency and the compensation signal amplitude to obtain the adjusted sound signal comprises: adjusting the initial sound signal by a compensation algorithm to obtain the adjusted sound signal; the compensation algorithm is expressed as: Among them, S comp (t) is the adjusted sound signal as a function of time t, A adj is the compensation signal amplitude, S(t) is the initial sound signal as a function of time t, Δt is the time delay, j is an imaginary unit, and f adj is the compensation signal frequency.

4. The switch cabinet fault voiceprint detection method according to claim 2, characterized in that: The determining, according to the ambient temperature and the air humidity, the compensated ambient sound velocity, the compensated signal frequency, and the compensated signal amplitude of the initial sound signal comprises: Determining the compensated ambient sound speed according to the ambient temperature and the air humidity; determining the compensation signal frequency according to the air temperature; The compensation signal amplitude is determined according to the air humidity.

5. The switch cabinet fault voiceprint detection method according to claim 4, characterized in that: The determining the compensated ambient sound speed according to the ambient temperature and the air humidity comprises: performing compensation calibration on the sound speed of the initial sound signal based on the ambient sound speed influence to obtain the compensated ambient sound speed; the ambient sound speed influence is expressed as: c(T)=c0+k T ·(T-T0) c RH =c(T)·(1+k RH ·RH) Where T is the ambient temperature, T0 is the reference temperature, c(T) is the speed of sound corresponding to the ambient temperature, c0 is the preset reference speed of sound, and k T is the temperature influence coefficient, RH is the air humidity, k RH is the humidity influence coefficient, c RH To compensate for the ambient sound speed.

6. The switch cabinet fault voiceprint detection method according to any one of claims 1 to 5, characterized in that: include: The voiceprint fault detection model adopts a lightweight neural network structure, and data compression processing is performed through the voiceprint fault detection model, and the compression processing includes parameter clipping or low-rank decomposition.

7. The switch cabinet fault voiceprint detection method according to claim 6, characterized in that: Extracting voiceprint features based on the adjusted sound signal also includes: performing dimensionality reduction processing on the extracted voiceprint features to reduce the number of features in the voiceprint features.

8. A switch cabinet fault voiceprint detection device, characterized in that: include: A control processing unit (101), a voiceprint collection unit (102) and an environment perception unit (103), wherein the voiceprint collection unit (102) and the environment perception unit (103) are respectively connected to the control processing unit (101); the voiceprint collection unit (102) is used to collect initial sound signals, and the environment perception unit (103) is used to collect current environment information; the control processing unit (101) includes one or more processors and memories; The memory is coupled to the one or more processors, and the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and the one or more processors call the computer instructions to enable the control processing unit to execute the steps of the switch cabinet fault soundprint detection method as described in any one of claims 1 to 7.

9. A computer-readable storage medium storing computer instructions, characterized in that: When the computer instructions are executed on the control processing unit of the detection device, the control processing unit is caused to execute the steps of the switch cabinet fault soundprint detection method as claimed in any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product runs on the control processing unit of the detection device, the control processing unit is enabled to execute the steps of the switch cabinet fault voiceprint detection method according to any one of claims 1 to 7.

Images