Rebar processing equipment fault voiceprint identification and early warning method and device

Abstract

The application discloses a steel bar processing equipment fault voiceprint identification and early warning method and device, relates to the technical field of equipment state monitoring, and comprises the following steps: collecting original audio signals of equipment operation through an acoustic sensor; processing the signals to filter out periodic impact noise and background noise, and obtaining a pure time domain signal; performing time-frequency analysis on the pure signal to generate a two-dimensional time-frequency graph; extracting frequency spectrum features from the time-frequency graph and inputting the frequency spectrum features into a classification model for pattern recognition to obtain equipment state classification results; and finally generating an early warning according to the identified fault information. Through the above method, the real-time, non-contact monitoring and early fault warning of the running state of the steel bar processing equipment are realized, and the timeliness and intelligent level of equipment maintenance are improved.

Description

Technical Field

[0001] This application relates to the field of equipment status monitoring technology, and in particular to a method and device for soundprint recognition and early warning of faults in steel bar processing equipment. Background Technology

[0002] Rebar processing equipment (such as rebar bending machines, cutting machines, and straightening machines) is an indispensable key piece of equipment in modern industrialized construction production. Its operational status directly affects project quality, construction efficiency, and production costs. Currently, traditional methods for fault diagnosis and condition monitoring of such rotating machinery mainly rely on periodic manual inspections, vibration analysis, or temperature monitoring. Manual inspections suffer from drawbacks such as strong subjectivity, inability to respond in real time, and high missed detection rates; while vibration analysis typically requires the installation of sensors at specific locations on the equipment, which is costly and inconvenient to implement in some complex or enclosed structures.

[0003] Compared to vibration signals, acoustic signals contain richer equipment status information and enable non-contact measurement, offering advantages such as flexible deployment and large information capacity. However, applying voiceprint recognition technology to noisy industrial environments, especially in steel bar processing scenarios, faces significant challenges. Strong, periodic impact noise generated during equipment operation (such as the impact sound of steel bars striking molds) and changing environmental background noise (such as the sound of other equipment operating) can severely interfere with or even drown out the weak acoustic features characterizing equipment malfunctions, resulting in low accuracy and high false alarm rates with traditional sound analysis methods.

[0004] In addition, existing voiceprint recognition methods often rely on general audio features in the feature extraction stage, failing to fully consider the unique mechanical structure and operating mechanism of steel bar processing equipment (such as spindle speed and transmission ratio), resulting in weak correlation between the extracted features and the equipment status, and insufficient model generalization ability and timely warning.

[0005] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0006] The main purpose of this application is to provide a method and device for soundprint recognition and early warning of equipment failures in steel bar processing equipment, which aims to solve the technical problems of untimely and inaccurate equipment failure detection in the prior art.

[0007] To achieve the above objectives, this application provides a method for acoustic recognition and early warning of faults in steel bar processing equipment, the method comprising: Acoustic sensors are placed at specific locations on the rebar processing equipment to collect raw audio signals during the equipment's operation. The original audio signal is processed to obtain a clean time-domain signal; Time-frequency analysis was performed on the pure time-domain signal to obtain a two-dimensional time-frequency diagram; Feature extraction and pattern recognition are performed on the two-dimensional time-frequency graph to obtain the equipment status classification result; Fault information is determined based on the equipment status classification results, and early warning information is generated based on the fault information.

[0008] In one embodiment, the step of processing the original audio signal to obtain a clean time-domain signal includes: Extract the spectral characteristics of the periodic impact noise generated by the steel bar processing equipment during operation; Collect background noise samples of the steel bar processing equipment under no-load or load conditions; A digital band-stop filter is constructed based on the spectral characteristics, and the original audio signal is filtered using the digital band-stop filter to obtain a filtered audio signal. The target background noise is determined from the background noise sample based on the processing status of the steel bar processing equipment, and the filtered audio signal is enhanced by spectral enhancement based on the target background noise to obtain a clean time-domain signal.

[0009] In one embodiment, the step of determining the target background noise from the background noise sample based on the processing status of the rebar processing equipment, and performing spectral enhancement on the filtered audio signal based on the target background noise to obtain a clean time-domain signal includes: The real-time processing parameters of the steel bar processing equipment are obtained, and the target background noise is determined from the background noise samples according to the correspondence between the real-time processing parameters and the background noise samples. The filter weights are determined based on the target background noise, and the filtered audio signal is enhanced by spectral lines based on the filter weights to remove the background noise component in the filtered audio signal and obtain a clean time-domain signal.

[0010] In one embodiment, the step of performing time-frequency analysis on the pure time-domain signal to obtain a two-dimensional time-frequency graph includes: The length of the window function is determined based on the main motor speed and gearbox ratio of the steel bar processing equipment, and the time window is determined based on the length of the window function. Based on the time window, time-frequency analysis is performed on the clean time-domain signal with a preset step size to determine the spectral sequence in the time domain; The spectral sequence is sorted along the time dimension to obtain a two-dimensional time-frequency diagram.

[0011] In one embodiment, the step of performing time-frequency analysis on the clean time-domain signal based on the time window with a preset step size to determine the spectral sequence in the time domain includes: The time window is slid along the time axis of the pure time domain signal with the preset step size. After each slide, the time window covers a new signal segment. The signal segments within the time window are weighted to obtain weighted signal segments; The weighted signal segment is transformed from the time domain to the frequency domain to obtain the spectrum corresponding to the weighted signal segment; The spectra of all weighted signal segments are collected in chronological order to obtain a spectrum sequence.

[0012] In one embodiment, the step of performing feature extraction and pattern recognition on the two-dimensional time-frequency graph to obtain the device status classification result includes: The spectrum center, spectrum bandwidth, and zero-crossing rate are extracted from the two-dimensional time-frequency graph, and the spectrum center, spectrum bandwidth, and zero-crossing rate are combined into a feature vector. The feature vector is input into a classification model for pattern recognition to obtain the posterior probability of the feature vector for each device state category. The device state category with the highest posterior probability is taken as the device state classification result.

[0013] In one embodiment, the step of extracting the spectrum center, spectrum bandwidth, and zero-crossing rate from the two-dimensional time-frequency graph, and combining the spectrum center, spectrum bandwidth, and zero-crossing rate into a feature vector includes: Based on the amplitude of each frequency component in the two-dimensional time-frequency graph, calculate the weighted average value of each frequency, and determine the spectral center based on the weighted average value; Calculate the degree of dispersion of spectral energy around the spectral center based on the spectral center, and determine the spectral bandwidth based on the degree of dispersion. The number of times the time-domain signal waveform crosses the zero level within a preset time window is counted, and the zero-crossing rate is obtained based on the number of times. The spectral center, the spectral bandwidth, and the zero-crossing rate are combined into a feature vector.

[0014] In one embodiment, the step of inputting the feature vector into a classification model for pattern recognition to obtain the posterior probability of the feature vector for each device state category further includes: The feature vector is input into a classification model for pattern recognition. The feature vector is compared with the feature vector center of various device states determined by the historical data stored in the classification model to obtain the Mahalanobis distance between the feature vector and the feature vector center. The Mahalanobis distance is converted into the posterior probability of each device state category.

[0015] In one embodiment, the step of determining fault information based on the device status classification result and generating early warning information based on the fault information includes: When the equipment status classification result is a fault status, the fault type, fault location, and severity level are determined based on the equipment status classification result; Determine the faulty device number and fault timestamp; Early warning information is generated based on the faulty device number, the fault timestamp, the fault type, the fault location, and the severity level.

[0016] Furthermore, to achieve the above objectives, this application also proposes a soundprint recognition and early warning device for steel bar processing equipment malfunctions, which includes: The audio acquisition module is used to acquire the raw audio signals during the operation of the equipment based on acoustic sensors arranged in specific parts of the steel bar processing equipment. The signal processing module is used to process the original audio signal to obtain a clean time-domain signal; The signal analysis module is used to perform time-frequency analysis on the pure time-domain signal to obtain a two-dimensional time-frequency diagram; The voiceprint recognition module is used to extract features and recognize patterns from the two-dimensional time-frequency graph to obtain the device status classification result. The fault warning module is used to determine fault information based on the equipment status classification results and generate warning information based on the fault information.

[0017] In addition, to achieve the above objectives, this application also proposes a soundprint recognition and early warning device for steel bar processing equipment faults. The soundprint recognition and early warning device for steel bar processing equipment faults includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. The computer program is configured to implement the steps of the soundprint recognition and early warning method for steel bar processing equipment faults as described above.

[0018] In addition, to achieve the above objectives, the present invention also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the above-described method for identifying and warning of faults in steel bar processing equipment by soundprint.

[0019] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the above-described method for identifying and warning of faults in steel bar processing equipment using soundprint recognition.

[0020] This application provides a method for acoustic recognition and early warning of faults in rebar processing equipment. The method involves acquiring the raw audio signal of the equipment operation using an acoustic sensor; processing the signal to filter out periodic impact noise and background noise to obtain a clean time-domain signal; performing time-frequency analysis on the clean signal to generate a two-dimensional time-frequency graph; extracting spectral features from the time-frequency graph and inputting them into a classification model for pattern recognition to obtain the equipment status classification result; and finally generating an early warning based on the identified fault information. Through this method, real-time, non-contact monitoring of the rebar processing equipment's operating status and early fault warning are achieved, improving the timeliness and intelligence of equipment maintenance. Attached Figure Description

[0021] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0022] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating an embodiment of the method for soundprint recognition and early warning of steel bar processing equipment malfunctions in this application. Figure 2 This is a schematic diagram of sensor installation for an embodiment of the method for soundprint recognition and early warning of steel bar processing equipment faults in this application; Figure 3 This is a schematic diagram of the module structure of the soundprint recognition and early warning device for steel bar processing equipment failure according to an embodiment of this application; Figure 4 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the soundprint recognition and early warning method for steel bar processing equipment faults in the embodiments of this application.

[0024] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0025] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0026] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0027] The main solution of this application embodiment is: to collect the original audio signal during the operation of the equipment based on acoustic sensors arranged in specific parts of the steel bar processing equipment; The original audio signal is processed to obtain a clean time-domain signal; Time-frequency analysis was performed on the pure time-domain signal to obtain a two-dimensional time-frequency diagram; Feature extraction and pattern recognition are performed on the two-dimensional time-frequency graph to obtain the equipment status classification result; Fault information is determined based on the equipment status classification results, and early warning information is generated based on the fault information.

[0028] Currently, the reliable operation of steel bar processing equipment is crucial in industrialized construction production. However, existing fault monitoring technologies, such as manual inspection and vibration analysis, suffer from problems such as poor real-time performance, high cost, or inconvenient deployment. Although voiceprint recognition technology provides a non-contact solution, in actual noisy industrial sites, strong periodic impact noise and changing background noise can seriously interfere with the extraction of fault features, resulting in low recognition accuracy. Furthermore, general acoustic features are difficult to adapt to the specific operating mechanism of steel bar processing equipment, which restricts the effective application of this technology.

[0029] This application provides a solution that acquires raw audio signals from equipment operation using acoustic sensors; processes these signals to filter out periodic impact noise and background noise, obtaining a clean time-domain signal; performs time-frequency analysis on the clean signal to generate a two-dimensional time-frequency graph; extracts spectral features from the time-frequency graph and inputs them into a classification model for pattern recognition, obtaining equipment status classification results; and finally generates early warnings based on the identified fault information. Through this method, real-time, non-contact monitoring of the operating status of rebar processing equipment and early fault warnings are achieved, improving the timeliness and intelligence of equipment maintenance.

[0030] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, such as a steel bar processing equipment fault soundprint recognition and early warning device. This embodiment does not specifically limit it in this regard. The following uses a steel bar processing equipment fault soundprint recognition and early warning device as an example to describe this embodiment and the following embodiments.

[0031] This application provides a method for soundprint recognition and early warning of faults in rebar processing equipment, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the method for soundprint recognition and early warning of steel bar processing equipment failure in this application.

[0032] In this embodiment, the method for soundprint recognition and early warning of steel bar processing equipment faults includes steps S10~S50: Step S10: Acoustic sensors located at specific locations on the rebar processing equipment are used to collect raw audio signals during equipment operation.

[0033] It should be noted that acoustic sensors are devices that can convert sound signals into electrical signals, such as microphones or accelerometers. Specific locations refer to the vicinity of critical moving parts of the equipment or locations prone to failure, such as the spindle bearing housing or gearbox housing.

[0034] It should be understood that traditional contact sensors are complex to install and may affect equipment operation, while acoustic signals contain rich equipment status information and can achieve continuous monitoring of equipment operation status without contact.

[0035] Understandably, by fixing a high-fidelity acoustic sensor to a selected monitoring point and connecting it to a data acquisition card, the sound signal is continuously recorded at a sampling rate much higher than the device's operating characteristic frequency, while ensuring sufficient amplitude range to avoid signal clipping, thus ensuring that the acquired original audio signal can completely preserve the acoustic characteristics of the device's operation.

[0036] In practical implementation, the first step is to select key locations that best characterize the equipment's operating state and install acoustic sensors based on its structural mechanism and common failure modes. Examples include the bearing housing at the spindle drive end or the outer surface of the gear meshing area. Figure 2 The sensor is securely mounted via a magnetic base or threads to ensure close contact with the equipment surface to reduce signal attenuation. A suitable orientation is chosen to maximize the pickup of the target sound source. Simultaneously, it is connected to a data acquisition module housed in the control cabinet via shielded cables to eliminate the impact of on-site electromagnetic interference on signal quality.

[0037] The data acquisition module, based on the Nyquist sampling theorem, sets a sampling rate at more than twice the highest characteristic frequency of the device's operation for continuous recording to ensure distortion-free signal transmission. An appropriate range is also set to capture a range from stable sounds during normal operation to weak impact signals during abnormal events. The acquired raw audio signals are buffered in time segments of specific durations, and timestamped and identified by the device, providing a high-quality data source for subsequent signal preprocessing and feature extraction modules.

[0038] Step S20: Process the original audio signal to obtain a clean time-domain signal.

[0039] It should be noted that the original audio signal refers to a mixed signal obtained directly from acoustic sensors, which includes components such as equipment operating sound, environmental noise, and electromagnetic interference. The pure time-domain signal, on the other hand, refers to a signal obtained through a series of signal processing methods, which mainly contains the sound components generated by the operation of the equipment itself and whose interference has been effectively suppressed.

[0040] Understandably, the original audio signal is first pre-emphasized to enhance high-frequency components. Then, a reference model based on the sound samples of the device in normal operation is used to estimate and cancel the periodic background noise in real time through an adaptive filtering algorithm. The residual signal is then bandpass filtered to remove interference outside the operating frequency band, ultimately resulting in a clean time-domain signal that highlights the device's state characteristics.

[0041] In one feasible implementation, the step of processing the original audio signal to obtain a clean time-domain signal includes: Extract the spectral characteristics of the periodic impact noise generated by the steel bar processing equipment during operation; Collect background noise samples of the steel bar processing equipment under no-load or load conditions; A digital band-stop filter is constructed based on the spectral characteristics, and the original audio signal is filtered using the digital band-stop filter to obtain a filtered audio signal. Based on the processing status of the steel bar processing equipment, the target background noise is determined from the background noise sample, and the filtered audio signal is enhanced by spectral line enhancement based on the target background noise to obtain a clean time domain signal. The step of determining the target background noise from the background noise sample based on the processing status of the steel bar processing equipment, and performing spectral enhancement on the filtered audio signal based on the target background noise to obtain a clean time-domain signal includes: The real-time processing parameters of the steel bar processing equipment are obtained, and the target background noise is determined from the background noise samples according to the correspondence between the real-time processing parameters and the background noise samples. The filter weights are determined based on the target background noise, and the filtered audio signal is enhanced by spectral lines based on the filter weights to remove the background noise component in the filtered audio signal and obtain a clean time-domain signal.

[0042] It should be noted that periodic impact noise refers to transient acoustic events generated by periodic impacts caused by specific mechanical actions of equipment, such as shearing or bending steel bars. Its spectral characteristics refer to the peak distribution of the noise in the frequency domain. A digital bandstop filter is a filter that sets a stopband at a specific center frequency to attenuate the peak energy.

[0043] In practical implementation, the average power spectral density of the corresponding audio segments is calculated by analyzing the multiple impacts of the device within a typical operating cycle to stably extract the spectral characteristics of the periodic impact noise, identifying the narrowband peak frequencies with significantly higher energy than the background. Based on these characteristic frequencies, a digital bandstop filter is constructed, whose transfer function can typically be expressed as a notch filter based on a second-order cascade. :

[0044] in, It is the normalized digital angular frequency corresponding to the peak frequency to be filtered out. It is a complex number operator representing a time shift operation. and Indicates a time delay of one unit and two units. It is the pole radius parameter that controls the stopband width and depth. This filter can effectively reduce the prominent impulse noise line spectrum in the original signal to obtain the filtered audio signal.

[0045] Next, the system reads the equipment's processing parameters in real time, such as processing speed and rebar specifications, and selects the most matching target background noise sample based on a pre-established mapping relationship between parameters and background noise samples. Utilizing the statistical characteristics of this sample, an adaptive algorithm calculates the optimal filter weights to estimate and correct the short-time spectrum of the filtered signal, enhancing the harmonic structure related to equipment operation. The core of this process is solving the Wiener filter and its frequency domain response. for:

[0046] in, It is the estimated signal power spectrum. It is the power spectrum of the target background noise. This process effectively suppresses the residual background noise components and finally outputs a clean time-domain signal.

[0047] Step S30: Perform time-frequency analysis on the pure time-domain signal to obtain a two-dimensional time-frequency diagram.

[0048] It should be noted that time-frequency analysis is an analytical method that unfolds a one-dimensional time-domain signal into a joint distribution of time and frequency. It is used to observe the pattern of frequency component changes of the signal over time. The two-dimensional time-frequency graph is a visual representation of the analysis results. Its horizontal axis represents time, the vertical axis represents frequency, and the brightness or color of the image pixels represents the energy intensity of the frequency component at that moment.

[0049] Understandably, the short-time Fourier transform method is usually used. Its principle is to extract the pure time-domain signal with a sliding window of fixed time length, perform Fourier transform on the signal segment in each window to obtain the local spectrum of that time period, and arrange the spectrum results of all time windows along the time axis to finally generate a two-dimensional grayscale or pseudo-color image that can reflect the time-frequency characteristics of the signal.

[0050] In one feasible implementation, the step of performing time-frequency analysis on the pure time-domain signal to obtain a two-dimensional time-frequency diagram includes: The length of the window function is determined based on the main motor speed and gearbox ratio of the steel bar processing equipment, and the time window is determined based on the length of the window function. Based on the time window, time-frequency analysis is performed on the clean time-domain signal with a preset step size to determine the spectral sequence in the time domain; The spectral sequence is sorted according to the time dimension to obtain a two-dimensional time-frequency diagram; The step of performing time-frequency analysis on the clean time-domain signal based on the time window with a preset step size to determine the spectral sequence in the time domain includes: The time window is slid along the time axis of the pure time domain signal with the preset step size. After each slide, the time window covers a new signal segment. The signal segments within the time window are weighted to obtain weighted signal segments; The weighted signal segment is transformed from the time domain to the frequency domain to obtain the spectrum corresponding to the weighted signal segment; The spectra of all weighted signal segments are collected in chronological order to obtain a spectrum sequence.

[0051] It should be noted that a window function is a function used in signal processing to weight truncated signal segments. Its length directly determines the trade-off between time resolution and frequency resolution in time-frequency analysis. The main motor speed and gearbox transmission ratio are key parameters used to calculate the fundamental frequency of the equipment's mechanical movements to match the analysis scale. The preset step size refers to the distance the time window slides on the signal time axis each time. It affects the smoothness of the time axis of the time-frequency graph and the amount of computation. Weighting processing refers to the process of using a window function to adjust the amplitude of the signal segment to reduce spectral leakage. A spectral sequence refers to a collection of multiple local spectra arranged in chronological order.

[0052] In practical implementation, determining the window function length is crucial for balancing time and frequency resolution. The main motor speed of the equipment... (Unit: Hz) and transmission ratio Together they determine the fundamental frequency of mechanical impact. To ensure the time window covers at least one complete impact cycle for accurate feature capture, the window length... It is usually set to an integer multiple of the fundamental frequency period, for example ,in It is an integer (usually between 1 and 3). This depends on the sampling frequency. This will convert the window length to points. This determines the time window used to capture the signal.

[0053] After each sliding time window, the signal segment within the window needs to be... Apply window functions Weighting is applied to reduce spectral leakage caused by signal truncation. The windowed signal segment is... Commonly used window functions include the Hanning window, whose function form is: This can effectively reduce spectral sidelobes. Here, N is the number of points corresponding to the window length. It is the index of the sampling points within the window.

[0054] Next, for each weighted signal segment Performing a Fast Fourier Transform (FFT) converts the frequency spectrum from the time domain to the frequency domain, yielding a complex spectrum. :

[0055] in, It is a frequency index ( The corresponding actual frequency is , It is the imaginary unit. This spectrum. It characterizes the frequency components, amplitude, and phase of the signal within this time window.

[0056] Finally, the spectrum corresponding to each time window is... Arrange them chronologically to form a three-dimensional dataset (time, frequency, amplitude). The amplitude at each frequency point (usually taken as...) is then... or The pixel brightness is mapped to a grayscale or pseudo-color image and arranged with time as the horizontal axis and frequency as the vertical axis, resulting in a final two-dimensional time-frequency graph. This image visually illustrates the evolution of signal frequency components over time. The preset step size is usually smaller than the window length (i.e., there is overlap) to increase the smoothness of the time axis.

[0057] Step S40: Perform feature extraction and pattern recognition on the two-dimensional time-frequency graph to obtain the device status classification result.

[0058] It should be noted that the equipment status classification result refers to the final judgment conclusion on the current operating status of the equipment obtained through pattern recognition, such as normal, minor fault, serious fault, etc.

[0059] Understandably, image processing techniques such as gray-level co-occurrence matrix or signal feature algorithms such as calculating energy in a specific frequency band are first used to extract numerical feature vectors from the time-frequency map. These feature vectors are then input into a pre-trained classification model (such as support vector machine, convolutional neural network, etc.), and the model outputs the corresponding device status category label based on the learned decision boundary or feature pattern.

[0060] In one feasible implementation, the step of performing feature extraction and pattern recognition on the two-dimensional time-frequency graph to obtain the device status classification result includes: The spectrum center, spectrum bandwidth, and zero-crossing rate are extracted from the two-dimensional time-frequency graph, and the spectrum center, spectrum bandwidth, and zero-crossing rate are combined into a feature vector. The feature vector is input into a classification model for pattern recognition to obtain the posterior probability of the feature vector for each device state category. The device state category with the highest posterior probability is taken as the device state classification result; The step of extracting the spectrum center, spectrum bandwidth, and zero-crossing rate from the two-dimensional time-frequency graph, and combining the spectrum center, spectrum bandwidth, and zero-crossing rate into a feature vector includes: Based on the amplitude of each frequency component in the two-dimensional time-frequency graph, calculate the weighted average value of each frequency, and determine the spectral center based on the weighted average value; Calculate the degree of dispersion of spectral energy around the spectral center based on the spectral center, and determine the spectral bandwidth based on the degree of dispersion. The number of times the time-domain signal waveform crosses the zero level within a preset time window is counted, and the zero-crossing rate is obtained based on the number of times. The spectrum center, the spectrum bandwidth, and the zero-crossing rate are combined into a feature vector; The step of inputting the feature vector into a classification model for pattern recognition to obtain the posterior probability of the feature vector for each device state category further includes: The feature vector is input into a classification model for pattern recognition. The feature vector is compared with the feature vector center of various device states determined by the historical data stored in the classification model to obtain the Mahalanobis distance between the feature vector and the feature vector center. The Mahalanobis distance is converted into the posterior probability of each device state category.

[0061] It should be noted that the spectral center refers to the average frequency position of the spectral energy distribution, reflecting the concentrated frequency band of signal energy; the spectral bandwidth refers to an indicator that quantifies the degree of diffusion or dispersion of spectral energy around the center frequency; the zero-crossing rate refers to the number of times the signal amplitude crosses the zero level per unit time; the feature vector refers to a mathematical representation for machine learning composed of multiple feature indicators; and the classification model refers to a mathematical model trained based on historical data for identifying and distinguishing different device states.

[0062] In practice, the formula for calculating the spectrum center (SC) is as follows:

[0063] in, It is an index of frequency points. It is the total number of frequency points. It is the first The actual frequency value (unit: Hz) corresponding to each frequency point. It is the amplitude (or energy) at that frequency point in the time-frequency graph.

[0064] The spectral bandwidth (SBW) is calculated based on the spectrum center, using the following formula:

[0065] Here, SC represents the spectral center, and the other parameters have the same meaning as in the formula above. A larger bandwidth value indicates a more dispersed energy distribution in the frequency domain.

[0066] The zero-crossing rate (ZCR) is calculated directly on the time-domain signal, and the formula is:

[0067] in, It is the sampling point index of the time-domain signal. It is the total number of sampling points within the preset time window. It is the signal amplitude at time t. It is an indicator function that returns 1 if adjacent sample points have different signs, and 0 otherwise.

[0068] After obtaining the above three features, they are combined into a feature vector. The feature vector is then input into the classification model. Assume there are C device status categories, and each category has a pre-calculated average feature vector based on historical data, which is the feature vector center of that category. (in Calculate the current feature vector. To each category center Mahalanobis distance :

[0069] in, It is the inverse matrix of the covariance matrix of the historical feature vector of the i-th category.

[0070] Finally, the Mahalanobis distance is converted into the posterior probability of belonging to each class i. This can be done through a soft assignment based on the distance, for example:

[0071] in It is a smoothing parameter. It ensures that the sum of the posterior probabilities of all classes is 1. The classification result is to make the posterior probability... The largest category .

[0072] Step S50: Determine fault information based on the equipment status classification result, and generate early warning information based on the fault information.

[0073] It should be noted that fault information refers to a detailed description of the specific fault type, severity, and possible location further analyzed based on the equipment status classification results, while warning information is a notification or alarm generated by the system to alert operators or trigger maintenance procedures.

[0074] Understandably, when generating early warning information based on fault information, a knowledge base or rule engine that maps status to fault information is usually established. After the classification model outputs the device status category, the system will query the knowledge base to obtain the corresponding fault description, suggested measures and early warning level, and then automatically generate and send an early warning message containing key information according to the preset communication protocol.

[0075] In one feasible implementation, the step of determining fault information based on the equipment status classification result and generating early warning information based on the fault information includes: When the equipment status classification result is a fault status, the fault type, fault location, and severity level are determined based on the equipment status classification result; Determine the faulty device number and fault timestamp; Early warning information is generated based on the faulty device number, the fault timestamp, the fault type, the fault location, and the severity level.

[0076] It should be noted that the fault device number refers to a string or code that uniquely identifies the specific device that malfunctioned; the fault timestamp refers to the precise date and time information of the fault being identified and recorded by the system, usually including year, month, day, hour, minute, and second; the fault type refers to the classification description of the fault phenomenon or mechanism; the fault location refers to the specific component or subsystem of the equipment that malfunctioned; and the severity level refers to the level of the fault consequences or urgency based on predetermined standards.

[0077] In its implementation, the system maintains an internal database or configuration file that maps device status to fault details. When the classification model determines that a device is in a faulty state, the system uses this status category as a key index to retrieve the corresponding fault type, fault location, and severity level determined by a combination of factors such as algorithm confidence or vibration amplitude from the database. Simultaneously, the system obtains a unique faulty device ID from the real-time data stream or device registration information and records the current fault timestamp from the system clock.

[0078] The system then assembles these structured data elements according to predefined templates or communication protocols to generate the final alert information. This process typically involves data formatting, such as populating the information into JSON or XML message bodies, or generating log entries in a specific format. The generated alert information is then sent to message queues, a central monitoring platform, or the terminals of maintenance personnel to ensure that relevant personnel can promptly learn the details of the fault and take appropriate action.

[0079] This embodiment provides a method for acoustic recognition and early warning of faults in rebar processing equipment. The method involves acquiring the raw audio signal of the equipment operation using an acoustic sensor; processing the signal to filter out periodic impact noise and background noise to obtain a clean time-domain signal; performing time-frequency analysis on the clean signal to generate a two-dimensional time-frequency graph; extracting spectral features from the time-frequency graph and inputting them into a classification model for pattern recognition to obtain the equipment status classification result; and finally generating an early warning based on the identified fault information. Through this method, real-time, non-contact monitoring of the rebar processing equipment's operating status and early fault warning are achieved, improving the timeliness and intelligence of equipment maintenance.

[0080] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the method for soundprint identification and early warning of steel bar processing equipment failure in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0081] This application also provides a soundprint recognition and early warning device for steel bar processing equipment malfunctions. Please refer to [link / reference]. Figure 3 The soundprint recognition and early warning device for steel bar processing equipment malfunctions includes: The audio acquisition module 10 is used to acquire the original audio signals during the operation of the equipment based on acoustic sensors arranged in specific parts of the steel bar processing equipment. Signal processing module 20 is used to process the original audio signal to obtain a clean time-domain signal; Signal analysis module 30 is used to perform time-frequency analysis on the pure time-domain signal to obtain a two-dimensional time-frequency diagram; Voiceprint recognition module 40 is used to extract features and recognize patterns from the two-dimensional time-frequency graph to obtain device status classification results; The fault warning module 50 is used to determine fault information based on the equipment status classification results and generate warning information based on the fault information.

[0082] In one feasible implementation, the signal processing module 20 is further configured to extract the spectral characteristics of the periodic impact noise generated by the rebar processing equipment during operation. Collect background noise samples of the steel bar processing equipment under no-load or load conditions; A digital band-stop filter is constructed based on the spectral characteristics, and the original audio signal is filtered using the digital band-stop filter to obtain a filtered audio signal. The target background noise is determined from the background noise sample based on the processing status of the steel bar processing equipment, and the filtered audio signal is enhanced by spectral enhancement based on the target background noise to obtain a clean time-domain signal.

[0083] In one feasible implementation, the signal processing module 20 is further configured to acquire the real-time processing parameters of the steel bar processing equipment, and determine the target background noise from the background noise samples based on the correspondence between the real-time processing parameters and the background noise samples; The filter weights are determined based on the target background noise, and the filtered audio signal is enhanced by spectral lines based on the filter weights to remove the background noise component in the filtered audio signal and obtain a clean time-domain signal.

[0084] In one feasible implementation, the signal analysis module 30 is further configured to determine the length of the window function based on the main motor speed and transmission ratio of the rebar processing equipment, and to determine a time window based on the length of the window function; Based on the time window, time-frequency analysis is performed on the clean time-domain signal with a preset step size to determine the spectral sequence in the time domain; The spectral sequence is sorted along the time dimension to obtain a two-dimensional time-frequency diagram.

[0085] In one feasible implementation, the signal analysis module 30 is further configured to slide the time window along the time axis of the pure time domain signal with the preset step size, and after each slide, the time window covers a new signal segment; The signal segments within the time window are weighted to obtain weighted signal segments; The weighted signal segment is transformed from the time domain to the frequency domain to obtain the spectrum corresponding to the weighted signal segment; The spectra of all weighted signal segments are collected in chronological order to obtain a spectrum sequence.

[0086] In one feasible implementation, the voiceprint recognition module 40 is further configured to extract the spectrum center, spectrum bandwidth and zero-crossing rate from the two-dimensional time-frequency graph, and combine the spectrum center, spectrum bandwidth and zero-crossing rate into a feature vector; The feature vector is input into a classification model for pattern recognition to obtain the posterior probability of the feature vector for each device state category. The device state category with the highest posterior probability is taken as the device state classification result.

[0087] In one feasible implementation, the voiceprint recognition module 40 is further configured to calculate the weighted average value of each frequency based on the amplitude of each frequency component in the two-dimensional time-frequency graph, and determine the spectrum center based on the weighted average value. Calculate the degree of dispersion of spectral energy around the spectral center based on the spectral center, and determine the spectral bandwidth based on the degree of dispersion. The number of times the time-domain signal waveform crosses the zero level within a preset time window is counted, and the zero-crossing rate is obtained based on the number of times. The spectral center, the spectral bandwidth, and the zero-crossing rate are combined into a feature vector.

[0088] In one feasible implementation, the voiceprint recognition module 40 is further configured to input the feature vector into a classification model for pattern recognition, compare the feature vector with the feature vector center of various device states determined by the historical data stored in the classification model, and obtain the Mahalanobis distance between the feature vector and the feature vector center. The Mahalanobis distance is converted into the posterior probability of each device state category.

[0089] In one feasible implementation, the fault warning module 50 is further configured to determine the fault type, fault location and severity level based on the equipment status classification result when the equipment status classification result is a fault status. Determine the faulty device number and fault timestamp; Early warning information is generated based on the faulty device number, the fault timestamp, the fault type, the fault location, and the severity level.

[0090] The rebar processing equipment fault acoustic recognition and early warning device provided in this application adopts the rebar processing equipment fault acoustic recognition and early warning method in the above embodiments, which can solve the technical problems of untimely and inaccurate equipment fault detection. Compared with the prior art, the beneficial effects of the rebar processing equipment fault acoustic recognition and early warning device provided in this application are the same as the beneficial effects of the rebar processing equipment fault acoustic recognition and early warning method provided in the above embodiments, and other technical features in the rebar processing equipment fault acoustic recognition and early warning device are the same as the features disclosed in the above embodiments, and will not be repeated here.

[0091] This application provides a soundprint recognition and early warning device for steel bar processing equipment faults. The soundprint recognition and early warning device for steel bar processing equipment faults includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the soundprint recognition and early warning method for steel bar processing equipment faults in the above embodiment 1.

[0092] The following is for reference. Figure 4 The diagram illustrates a structural schematic suitable for implementing a voiceprint recognition and early warning device for rebar processing equipment faults in the embodiments of this application. The voiceprint recognition and early warning device for rebar processing equipment faults in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), vehicle terminals (e.g., vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 4 The illustrated steel bar processing equipment fault soundprint recognition and early warning device is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0093] like Figure 4 As shown, the rebar processing equipment fault soundprint recognition and early warning device may include a processing unit 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to the program stored in ROM (Read Only Memory) 1002 or the program loaded from storage device 1003 into RAM (Random Access Memory) 1004. RAM 1004 also stores various programs and data required for the operation of the rebar processing equipment fault soundprint recognition and early warning device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. Input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, LCDs (Liquid Crystal Displays), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the rebar processing equipment fault acoustic recognition and early warning device to exchange data wirelessly or via wired communication with other devices. Although the figure shows rebar processing equipment fault acoustic recognition and early warning devices with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0094] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0095] The rebar processing equipment fault acoustic recognition and early warning device provided in this application adopts the rebar processing equipment fault acoustic recognition and early warning method in the above embodiments, which can solve the technical problem of acoustic recognition and early warning of rebar processing equipment faults. Compared with the prior art, the beneficial effects of the rebar processing equipment fault acoustic recognition and early warning device provided in this application are the same as the beneficial effects of the rebar processing equipment fault acoustic recognition and early warning method provided in the above embodiments, and other technical features in the rebar processing equipment fault acoustic recognition and early warning device are the same as the features disclosed in the previous embodiment method, and will not be repeated here.

[0096] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0097] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0098] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the steel bar processing equipment fault soundprint recognition and early warning method in the above embodiments.

[0099] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable Read Only Memory or Flash Memory), optical fibers, CD-ROM (CD-Read Only Memory), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0100] The aforementioned computer-readable storage medium may be included in the soundprint recognition and early warning device for steel bar processing equipment faults; or it may exist independently and not be assembled into the soundprint recognition and early warning device for steel bar processing equipment faults.

[0101] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the steel bar processing equipment fault acoustic recognition and early warning device, cause the steel bar processing equipment fault acoustic recognition and early warning device to: collect the original audio signals during equipment operation based on acoustic sensors arranged in specific parts of the steel bar processing equipment; The original audio signal is processed to obtain a clean time-domain signal; Time-frequency analysis was performed on the pure time-domain signal to obtain a two-dimensional time-frequency diagram; Feature extraction and pattern recognition are performed on the two-dimensional time-frequency graph to obtain the equipment status classification result; Fault information is determined based on the equipment status classification results, and early warning information is generated based on the fault information.

[0102] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including LAN (Local Area Network) or WAN (Wide Area Network)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0103] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0104] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0105] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described method for identifying and warning of faults in rebar processing equipment using acoustic signatures. This method can solve the technical problem of identifying and warning of faults in rebar processing equipment using acoustic signatures. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the method for identifying and warning of faults in rebar processing equipment provided in the above embodiments, and will not be elaborated upon here.

[0106] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for soundprint recognition and early warning of faults in steel bar processing equipment.

[0107] The computer program product provided in this application can solve the technical problems of untimely and inaccurate equipment fault detection. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the steel bar processing equipment fault soundprint recognition and early warning method provided in the above embodiments, and will not be repeated here.

[0108] The above are only some embodiments of this application and do not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for soundprint identification and early warning of faults in steel bar processing equipment, characterized in that, The method for soundprint recognition and early warning of steel bar processing equipment malfunctions includes: Acoustic sensors are placed at specific locations on the rebar processing equipment to collect raw audio signals during the equipment's operation. Extract the spectral characteristics of the periodic impact noise generated by the steel bar processing equipment during operation; Collect background noise samples of the steel bar processing equipment under no-load or load conditions; A digital band-stop filter is constructed based on the spectral characteristics, and the original audio signal is filtered using the digital band-stop filter to obtain a filtered audio signal. Based on the processing status of the steel bar processing equipment, the target background noise is determined from the background noise sample, and the filtered audio signal is enhanced by spectral line enhancement based on the target background noise to obtain a clean time domain signal. Time-frequency analysis was performed on the pure time-domain signal to obtain a two-dimensional time-frequency diagram; Feature extraction and pattern recognition are performed on the two-dimensional time-frequency graph to obtain the equipment status classification result; Fault information is determined based on the equipment status classification results, and early warning information is generated based on the fault information.

2. The method as described in claim 1, characterized in that, The step of determining the target background noise from the background noise sample based on the processing status of the steel bar processing equipment, and performing spectral enhancement on the filtered audio signal based on the target background noise to obtain a clean time-domain signal includes: The real-time processing parameters of the steel bar processing equipment are obtained, and the target background noise is determined from the background noise samples according to the correspondence between the real-time processing parameters and the background noise samples. The filter weights are determined based on the target background noise, and the filtered audio signal is enhanced by spectral lines based on the filter weights to remove the background noise component in the filtered audio signal and obtain a clean time-domain signal.

3. The method as described in claim 1, characterized in that, The step of performing time-frequency analysis on the pure time-domain signal to obtain a two-dimensional time-frequency graph includes: The length of the window function is determined based on the main motor speed and gearbox ratio of the steel bar processing equipment, and the time window is determined based on the length of the window function. Based on the time window, time-frequency analysis is performed on the clean time-domain signal with a preset step size to determine the spectral sequence in the time domain; The spectral sequence is sorted along the time dimension to obtain a two-dimensional time-frequency diagram.

4. The method as described in claim 3, characterized in that, The step of performing time-frequency analysis on the clean time-domain signal based on the time window with a preset step size to determine the spectral sequence in the time domain includes: The time window is slid along the time axis of the pure time domain signal with the preset step size. After each slide, the time window covers a new signal segment. The signal segments within the time window are weighted to obtain weighted signal segments; The weighted signal segment is transformed from the time domain to the frequency domain to obtain the spectrum corresponding to the weighted signal segment; The spectra of all weighted signal segments are collected in chronological order to obtain a spectrum sequence.

5. The method as described in claim 1, characterized in that, The steps of performing feature extraction and pattern recognition on the two-dimensional time-frequency graph to obtain the device status classification result include: Based on the amplitude of each frequency component in the two-dimensional time-frequency graph, calculate the weighted average value of each frequency, and determine the spectral center based on the weighted average value; Calculate the degree of dispersion of spectral energy around the spectral center based on the spectral center, and determine the spectral bandwidth based on the degree of dispersion. The number of times the time-domain signal waveform crosses the zero level within a preset time window is counted, and the zero-crossing rate is obtained based on the number of times. The spectrum center, the spectrum bandwidth, and the zero-crossing rate are combined into a feature vector; The feature vector is input into a classification model for pattern recognition to obtain the posterior probability of the feature vector for each device state category. The device state category with the highest posterior probability is taken as the device state classification result.

6. The method as described in claim 5, characterized in that, The step of inputting the feature vector into a classification model for pattern recognition to obtain the posterior probability of the feature vector for each device state category further includes: The feature vector is input into a classification model for pattern recognition. The feature vector is compared with the feature vector center of various device states determined by the historical data stored in the classification model to obtain the Mahalanobis distance between the feature vector and the feature vector center. The Mahalanobis distance is converted into the posterior probability of each device state category.

7. The method as described in claim 1, characterized in that, The step of determining fault information based on the equipment status classification result and generating early warning information based on the fault information includes: When the equipment status classification result is a fault status, the fault type, fault location, and severity level are determined based on the equipment status classification result; Determine the faulty device number and fault timestamp; Early warning information is generated based on the faulty device number, the fault timestamp, the fault type, the fault location, and the severity level.

8. A soundprint recognition and early warning device for steel bar processing equipment malfunctions, characterized in that, The steel bar processing equipment fault acoustic recognition and early warning device includes: The audio acquisition module is used to acquire the raw audio signals during the operation of the equipment based on acoustic sensors arranged in specific parts of the steel bar processing equipment. The signal processing module is used to extract the spectral characteristics of the periodic impact noise generated by the steel bar processing equipment during operation. The signal processing module is also used to collect background noise samples of the steel bar processing equipment under no-load or load conditions; The signal processing module is also used to construct a digital band-stop filter based on the spectral characteristics, and to filter the original audio signal based on the digital band-stop filter to obtain a filtered audio signal; The signal processing module is also used to determine the target background noise from the background noise sample according to the processing status of the steel bar processing equipment, and to perform spectral enhancement on the filtered audio signal based on the target background noise to obtain a clean time-domain signal; The signal analysis module is used to perform time-frequency analysis on the pure time-domain signal to obtain a two-dimensional time-frequency diagram; The voiceprint recognition module is used to extract features and recognize patterns from the two-dimensional time-frequency graph to obtain the device status classification result. The fault warning module is used to determine fault information based on the equipment status classification results and generate warning information based on the fault information.

Images