Coal mine water pump cavitation intelligent early warning method and system based on all-intrinsic-safety voiceprint collection

Abstract

The invention provides a coal mine water pump cavitation intelligent early warning method and system based on all-intrinsic-safety voiceprint collection, and relates to the technical field of voiceprint monitoring, and the method comprises the steps: collecting an original voiceprint signal of a coal mine water pump in real time through an intrinsic-safety voiceprint sensor array, carrying out the intrinsic-safety signal conditioning, obtaining a preprocessed digital signal, and transmitting the preprocessed digital signal to a computer; and carrying out noise reduction processing on the preprocessed digital signal to obtain a target digital signal, inputting the target digital signal into a voiceprint feature extraction model to obtain a cavitation early voiceprint feature vector, inputting the cavitation early voiceprint feature vector into a cavitation recognition model to obtain a cavitation confidence coefficient, and based on the cavitation confidence coefficient, carrying out noise reduction processing on the cavitation early voiceprint feature vector to obtain a cavitation early voiceprint feature vector. The technical problem that the accuracy of cavitation fault recognition is low due to the fact that an existing coal mine water pump cavitation monitoring scheme is generally difficult to achieve early discovery and accurate judgment of water pump cavitation faults in a complex underground environment is solved, and the water pump cavitation fault recognition method has the effect of improving the accuracy of water pump cavitation fault recognition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of voiceprint monitoring, and in particular to a coal mine water pump cavitation intelligent early warning method and system based on full intrinsic safety voiceprint collection. BACKGROUND

[0002] In coal mine underground operation, the stable operation of the drainage system is crucial, and the normal operation of the coal mine water pump as the core equipment of the drainage system is directly related to the safety production of the coal mine. With the increase of the coal mining depth and the expansion of the mining scale, the underground operation environment becomes more complex, and higher requirements are put forward for the performance and reliability of the water pump. The cavitation phenomenon of the water pump is one of the key problems affecting its normal operation. Cavitation not only causes the efficiency of the water pump to decrease, but also causes the equipment to be damaged in severe cases, affecting the stability of the entire drainage system, and further threatening the safety production of the coal mine. Therefore, the monitoring and early warning of the coal mine water pump cavitation become an important link to ensure the safety production of the coal mine.

[0003] At present, the coal mine water pump cavitation monitoring technology mainly includes the following: first, a flameproof vibration monitoring device is used. The device uses a heavy flameproof housing to package a vibration sensor, and judges cavitation by monitoring the vibration acceleration. However, the device is heavy and complex to install, and can only detect the middle and late stages of cavitation that have caused damage to the blade surface, with serious lag in early warning. Second, an intrinsic safety pressure sensing system is set up to identify cavitation by outlet pressure fluctuations. However, it is easy to be blocked in high-sand water quality, and cannot effectively distinguish between cavitation and pipeline blockage, resulting in a high false alarm rate. Third, a non-intrinsic safety voiceprint collection device is used. Although attempts have been made to use voiceprint signals for monitoring, the device cannot be applied to underground environments due to the lack of intrinsic safety design, and the voiceprint recognition algorithm is not optimized for cavitation characteristics, resulting in low recognition accuracy.

[0004] Based on the above deficiencies, the existing coal mine water pump cavitation monitoring scheme is generally difficult to achieve early detection and accurate judgment of water pump cavitation faults in complex underground environments, resulting in low accuracy of cavitation fault identification. SUMMARY

[0005] In view of the above deficiencies in the prior art, the present application aims to provide a coal mine water pump cavitation intelligent early warning method based on full intrinsic safety voiceprint collection, which has the characteristics of improving the accuracy of water pump cavitation fault identification.

[0006] The above-mentioned first application of the present application is achieved by the following technical scheme: A coal mine water pump cavitation intelligent early warning method based on full intrinsic safety voiceprint collection, comprising: An intrinsic safety voiceprint sensor array is used to collect the original voiceprint signals of the coal mine water pump in real time, and the original voiceprint signals are subjected to intrinsic safety signal conditioning to obtain digital signals after preprocessing. The preprocessed digital signal is subjected to noise reduction processing to obtain a target digital signal; The target digital signal is input into a voiceprint feature extraction model to obtain a cavitation early-stage voiceprint feature vector; The cavitation early-stage voiceprint feature vector is input into a cavitation recognition model to obtain a cavitation confidence level; Based on the cavitation confidence level, a corresponding hierarchical early-warning strategy is matched and executed.

[0007] By adopting the above technical solution, the water pump operation voiceprint is collected based on a full intrinsic safety voiceprint sensor array, intrinsic safety signal conditioning and noise reduction processing are performed, the strong electromagnetic and environmental noise interference in the well is effectively suppressed, the signal quality is ensured, the multi-dimensional voiceprint feature vector unique to the early stage of cavitation is extracted, precise analysis is performed in combination with a lightweight cavitation recognition model, early-stage recognition of cavitation in the micro-bubble stage is realized, a hierarchical early-warning strategy is matched based on the cavitation confidence level, and a gradual response from local alarm to linkage frequency reduction is realized, thereby ensuring the safety of the water pump while maximizing the maintenance of production continuity, so as to improve the timeliness, accuracy and reliability of the coal mine water pump cavitation early-warning.

[0008] Preferably, the intrinsic safety signal conditioning of the original voiceprint signal to obtain a preprocessed digital signal comprises: The original voiceprint signal is subjected to low-noise amplification processing to obtain an initial voiceprint signal; The initial voiceprint signal is subjected to band-pass filtering processing to obtain an intermediate voiceprint signal; The intermediate voiceprint signal is subjected to analog-to-digital conversion to obtain the preprocessed digital signal.

[0009] By adopting the above technical solution, low-noise amplification processing is adopted to effectively improve the amplitude of the weak voiceprint signal, then band-pass filtering processing is used to accurately retain the cavitation characteristic frequency band while effectively suppressing motor low-frequency vibration noise and high-frequency electromagnetic interference, and then analog voiceprint signals are converted into high-precision digital signals through analog-to-digital conversion, thereby ensuring that the signals meet the intrinsic safety energy limit and realizing high-fidelity digitization of the voiceprint signal.

[0010] Preferably, the noise reduction processing of the preprocessed digital signal to obtain a target digital signal comprises: The preprocessed digital signal is subjected to wavelet packet decomposition to obtain multi-level wavelet packet coefficients; The multi-level wavelet packet coefficients are subjected to threshold processing based on an adaptive threshold to obtain noise-reduced wavelet packet coefficients; The noise-reduced wavelet packet coefficients are subjected to wavelet packet reconstruction to obtain the target digital signal.

[0011] By adopting the technical scheme, fine division of the voiceprint signal in the full-band range is realized by using wavelet packet decomposition processing, then the wavelet packet coefficients dominated by noise are accurately separated and removed by using threshold processing based on an adaptive threshold, while the cavitation voiceprint characteristic coefficients are retained, then the coefficients after noise reduction are synthesized into a pure voiceprint signal with high signal-to-noise ratio by wavelet packet reconstruction processing, so that the early cavitation characteristics submerged in strong background noise are effectively extracted.

[0012] Preferably, the voiceprint feature vector comprises a frequency domain feature, a time-frequency feature and a statistical feature.

[0013] By adopting the technical scheme, the frequency domain feature of power spectrum entropy is extracted, the disorder of sound wave energy distribution caused by the collapse of micro-bubble cavitation is effectively quantified, the time-frequency feature of wavelet ridge instantaneous frequency is extracted, the high-frequency pulse transient characteristics generated in the dynamic collapse process of micro-bubbles are accurately captured, the statistical feature of pulse count is extracted, the activity and intensity of cavitation events are objectively characterized, and finally the comprehensive feature vector constructed by fusing the multi-dimensional features is used to comprehensively and stably depict the unique voiceprint fingerprint of early cavitation, thereby effectively solving the problem of low sensitivity to weak early cavitation characteristics and low recognition accuracy in the prior art under complex noise.

[0014] Preferably, based on the cavitation confidence, a corresponding hierarchical early warning strategy is matched and executed, comprising: determining a warning level corresponding to the cavitation confidence based on the cavitation confidence; if the warning level is level one, issuing a warning and generating a risk report; if the warning level is level two, adjusting the rotation speed of the coal mine water pump and generating a risk report.

[0015] By adopting the technical scheme, the continuous cavitation confidence is mapped to discrete warning levels, when it is determined to be level one warning, early discovery of cavitation risk and artificial intervention reminder are realized by issuing a warning and generating a report, thereby effectively solving the problem of lagging warning in the traditional method, when it is determined to be level two warning, automatic adjustment of the rotation speed of the water pump as a direct intervention measure is used to actively inhibit the further development of cavitation, thereby realizing the leap from passive alarm to active protection, so that the safety of equipment is ensured, non-scheduled downtime is minimized, and the recognition accuracy and reliability of cavitation fault of the water pump are improved.

[0016] Preferably, based on the cavitation confidence, the warning level corresponding to the cavitation confidence is determined, comprising: performing first-level comparison on the cavitation confidence based on a first pre-set confidence threshold to obtain a first comparison result; performing second-level comparison on the cavitation confidence based on a second pre-set confidence threshold to obtain a second-level comparison result. generate an early warning level based on the first comparison result and the second comparison result.

[0017] By adopting the technical scheme, the early cavitation risk is preliminarily identified by setting the first preset confidence threshold, the sensitivity to the initial fault is effectively improved, the accurate determination of the medium-term cavitation stage is ensured by setting the second preset confidence threshold higher than the first threshold, the misjudgment or omission caused by the single threshold is avoided, and finally the discrete early warning level is generated based on the logic synthesis of the two-stage comparison results, so that the judgment result of the cavitation state is more clear and operable.

[0018] The second object of the present application is to provide a coal mine water pump cavitation intelligent early warning system based on full intrinsic safety voiceprint collection, which has the characteristics of improving the accuracy of water pump cavitation fault identification.

[0019] The above-mentioned second object of the present application is achieved by the following technical scheme: A coal mine water pump cavitation intelligent early warning system based on full intrinsic safety voiceprint collection, comprising: a full intrinsic safety voiceprint collection device, an intrinsic safety monitoring substation, a control center and a water pump frequency conversion cabinet; The full intrinsic safety voiceprint collection device is installed on the surface of the coal mine water pump, and the full intrinsic safety voiceprint collection device comprises an intrinsic safety voiceprint sensor array, an intrinsic safety signal conditioning module and a communication unit. The intrinsic safety voiceprint sensor array is used to collect the original voiceprint signal of the coal mine water pump in real time, and transmit the original voiceprint signal to the intrinsic safety signal conditioning module. The intrinsic safety signal conditioning module is used for low-noise amplification, band-pass filtering and analog-to-digital conversion processing of the original voiceprint signal, to obtain a preprocessed digital signal, and transmit the digital signal to the communication unit. The communication unit is used for transmitting the digital signal to the intrinsic safety monitoring substation through an intrinsic safety communication link. The intrinsic safety monitoring substation is used for receiving and forwarding the digital signal to the control center. The control center is used for extracting voiceprint features and identifying cavitation from the received digital signal, generating a cavitation confidence level, and generating a hierarchical early warning instruction based on the cavitation confidence level, The water pump frequency conversion cabinet is in communication connection with the control center, and is used for receiving the hierarchical early warning instruction and performing corresponding water pump speed adjustment operation.

[0020] By adopting the technical scheme, the water pump operation voiceprint is directly collected through the intrinsic safety voiceprint sensor array, non-contact monitoring is realized, the problems of easy blockage and short service life of the traditional pressure sensor are avoided, the original signal is amplified, filtered and digitized through the intrinsic safety signal conditioning module, the signal quality is ensured under the condition of meeting the intrinsic safety requirements of the underground, the cavitation erosion features are extracted from the complex noise environment and the confidence is generated by using the intelligent algorithm of the control center, the early and accurate identification of the cavitation erosion fault is realized, finally, the differential early warning strategy is triggered based on the confidence level, the closed-loop management from local warning to equipment linkage control is realized, the problems of low accuracy of water pump cavitation erosion fault identification caused by the phenomena of early warning lag, high false alarm rate and lack of effective control means in the traditional monitoring scheme are solved, and the reliability and intelligent level of system operation are improved.

[0021] Preferably, the full intrinsic safety voiceprint collection device further comprises a power module. The power module is used to provide intrinsic safety power supply for the intrinsic safety voiceprint sensor array, the intrinsic safety signal conditioning module and the communication unit.

[0022] By adopting the technical scheme, a dedicated intrinsic safety power module is integrated to provide stable power supply meeting the intrinsic safety standard for each component of the voiceprint collection device, the absolute safety of the entire collection process in the combustible and explosive environment of the coal mine underground is ensured, the device does not need to rely on external non-intrinsic safety power supply or heavy explosion-proof power supply box, the light weight and integration of the collection device are realized, the installation adaptability in the narrow pump cabin is improved, meanwhile, the intrinsic safety power supply eliminates the circuit spark risk from the source, so that the voiceprint collection device can be directly installed on the surface of the water pump shell to facilitate continuous and reliable cavitation erosion monitoring.

[0023] The third object of the present application is to provide an electronic device with the characteristic of improving the accuracy of water pump cavitation erosion fault identification.

[0024] The third object of the present application is achieved by the following technical scheme: An electronic device includes a memory and a processor, and the memory stores a computer program capable of being loaded and executed by the processor to execute any one of the above-mentioned coal mine water pump cavitation intelligent early warning methods based on full intrinsic safety voiceprint collection.

[0025] The fourth object of the present application is to provide a computer storage medium capable of storing a corresponding program, with the characteristic of facilitating the improvement of the accuracy of water pump cavitation erosion fault identification.

[0026] The fourth object of the present application is achieved by the following technical scheme: A computer-readable storage medium storing a computer program that can be loaded by a processor and executed any of the above-mentioned intelligent early warning methods for coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition.

[0027] In summary, the present invention has at least one of the following beneficial technical effects: 1. This invention is based on intrinsically safe acoustic fingerprint acquisition and processing technology, which effectively suppresses complex downhole noise interference and ensures high-quality signal acquisition. By extracting multi-dimensional acoustic fingerprint features and using a lightweight intelligent model for identification, it can complete early diagnosis at the microbubble stage of cavitation. Furthermore, combined with a graded early warning mechanism, it realizes a progressive response from local alarm to equipment linkage speed adjustment. While ensuring the safe operation of equipment, it maximizes the maintenance of production continuity and improves the timeliness, accuracy and system reliability of cavitation early warning. 2. This invention maps cavitation confidence levels into a graded early warning system. In the early risk stage, timely human intervention is achieved through warnings and reports. In the mid-stage of development, automatic speed adjustment is used to actively suppress cavitation deterioration. This effectively solves the problem of delayed early warning in traditional systems and improves the accuracy of cavitation fault identification while ensuring equipment safety. 3. This invention constructs an acquisition device that integrates an intrinsically safe acoustic sensor array, an intrinsically safe signal conditioning module, and an intrinsically safe power supply module, eliminating the need for an explosion-proof enclosure and achieving lightweight, miniaturized, and intrinsically safe devices. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating the steps of an intelligent early warning method for cavitation in coal mine water pumps based on intrinsically safe acoustic fingerprint acquisition, as provided in Embodiment 1 of the present invention.

[0029] Figure 2 This is a flowchart illustrating the steps of another intelligent early warning method for coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition, provided in Embodiment 2 of the present invention.

[0030] Figure 3 This is a three-view diagram of the structure of a coal mine water pump cavitation intelligent early warning system based on intrinsically safe acoustic fingerprint acquisition, provided in Embodiment 3 of the present invention.

[0031] Figure 4 This is a schematic diagram of the structure of a coal mine water pump cavitation intelligent early warning system based on intrinsically safe acoustic fingerprint acquisition, provided in Embodiment 3 of the present invention. Detailed Implementation

[0032] This invention provides an intelligent early warning method and system for coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition. It solves the technical problem that existing coal mine water pump cavitation monitoring schemes are generally difficult to achieve early detection and accurate judgment of water pump cavitation faults in complex underground environments, resulting in low accuracy of cavitation fault identification. It has the technical effect of improving the accuracy of water pump cavitation fault identification.

[0033] It should be noted that the "intrinsically safe" value in the embodiments of the present invention refers to intrinsic safety.

[0034] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0035] It should be noted that, in the embodiments of this invention, the data related to object information, when applied to specific products or technologies, requires the permission or consent of the object. Furthermore, the collection, use, and processing of this data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. In other words, if the embodiments of this invention involve data related to an object, it must be obtained with the object's authorization and consent, the authorization and consent of relevant departments, and in accordance with the relevant laws, regulations, and standards of the country and region. If the embodiments involve personal information, the acquisition of all personal information requires the individual's consent. If sensitive information is involved, the separate consent of the information subject is required. The embodiments also need to be implemented with the object's authorization and consent.

[0036] It should be noted that the terms "first," "second," etc., used in this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The implementations described in the following exemplary embodiments do not represent all implementations consistent with this disclosure.

[0037] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0038] Example 1: Please see Figure 1 The present invention provides an intelligent early warning method for cavitation of coal mine water pumps based on intrinsically safe acoustic fingerprint acquisition, comprising: Step 101: Collect the original acoustic signature signal of the coal mine water pump in real time using an intrinsically safe acoustic signature sensor array, and perform intrinsically safe signal conditioning on the original acoustic signature signal to obtain the preprocessed digital signal.

[0039] Intrinsically safe acoustic sensor arrays refer to acquisition units composed of multiple acoustic sensors that meet the intrinsic safety standards (such as Ex ia I level) in underground coal mines, arranged in a specific geometric structure. Their function is to perform multi-point acquisition through spatial distribution and to enhance the target sound source and suppress spatial interference noise using array signal processing technology.

[0040] The original acoustic signal refers to the original analog acoustic signal containing the operating status information of the coal mine water pump, which is directly acquired by the intrinsically safe acoustic sensor array. This signal is a complex mixed signal, whose components mainly include characteristic sound waves generated by cavitation phenomena in the pump impeller, electromagnetic noise of the motor, mechanical vibration noise, water flow impact noise, and background noise of the underground environment.

[0041] Intrinsically safe signal conditioning refers to a series of front-end processing steps performed on the original analog signal under the premise of meeting the intrinsically safe circuit energy limit, with the aim of improving signal quality. The intrinsically safe circuit energy limit means that the electrical spark or thermal effect generated by the circuit under any fault state cannot ignite a specified explosive gas mixture.

[0042] The preprocessed digital signal refers to the high-precision, discrete-time digital sequence obtained by converting the original voiceprint signal through an intrinsically safe signal conditioning link for amplification, filtering, and other processing, and then through an analog-to-digital converter.

[0043] It is worth mentioning that, in the above-described feasible embodiments, the arrival time difference of the microphone array can also be used to directionally enhance the acoustic signal in the direction of the water pump housing. Specifically, after analog-to-digital conversion, a directional enhancement algorithm is executed on the three filtered digital signals. The algorithm estimates in real time the arrival time difference of the sound wave from the preset sound source direction of the water pump housing (i.e., the normal direction of the array) to the three microphones. Based on the calculated arrival time difference, the three signals are compensated for the corresponding time delay so that the sound wave signal from the target direction is completely synchronized (in phase) in the three channels, and then a weighted sum is performed.

[0044] In this embodiment of the invention, an intrinsically safe acoustic sensor array is used to collect the acoustic signatures of coal mine water pumps. The signals are then processed through an intrinsically safe signal conditioning link that includes low-noise amplification, bandpass filtering, and analog-to-digital conversion to finally output a high-fidelity pre-processed digital signal. This effectively improves the amplitude of weak cavitation acoustic signatures, overcomes the signal attenuation problem, and improves the signal-to-noise ratio and early identification accuracy of the entire early warning method.

[0045] Step 102: Perform noise reduction processing on the preprocessed digital signal to obtain the target digital signal.

[0046] Noise reduction processing refers to the use of algorithms to reduce noise in preprocessed digital signals, thereby separating and eliminating various environmental and mechanical noises mixed in the cavitation characteristic acoustic text, thus extracting a pure acoustic text signal with a higher signal-to-noise ratio. This process aims to preserve and enhance the weak features that characterize the early stage of cavitation from complex mixed signals to the greatest extent possible.

[0047] The target digital signal refers to the voiceprint digital sequence with a high signal-to-noise ratio obtained after noise reduction processing.

[0048] In this embodiment of the invention, an algorithm is used to perform noise reduction processing on the preprocessed digital signal to obtain a voiceprint digital sequence.

[0049] Step 103: Input the target digital signal into the acoustic signature feature extraction model to obtain the acoustic signature feature vector for early cavitation.

[0050] The voiceprint feature extraction model refers to a trained, lightweight deep learning network whose core uses an improved MobileNetV2 architecture to extract multi-dimensional features from target digital signals that constitute the voiceprint feature vector in the early stages of cavitation.

[0051] Early-stage acoustic signature feature vector refers to a high-dimensional, deep feature representation automatically learned and output from the target digital signal by the MobileNetV2 acoustic signature feature extraction model.

[0052] Understandably, the target digital signal is first converted into a time-frequency map as input to the model. Then, the time-frequency map is input to the MobileNetV2 voiceprint feature extraction model, which removes the classification layer at the top of the original network and uses the output of its last-stage global pooling layer as the early voiceprint feature vector.

[0053] It is worth mentioning that the training process of this voiceprint feature extraction model adopts a feature guidance strategy based on physical mechanisms to improve the model's sensitivity and generalization ability to early cavitation features. Specifically, during the model training phase, a corresponding multi-dimensional physical feature annotation map is generated for each voiceprint time-frequency map used as a training sample. This annotation map is generated by calculating and visualizing three key physical features of the voiceprint signal. 1. Power spectral entropy (frequency domain characteristics)

[0054] In the formula, Here, N represents the power spectral entropy value, and N is the total number of frequency bands in the spectrum division. Let be the normalized power percentage of the i-th frequency band. Let i be the i-th frequency unit.

[0055] It should be noted that the value is ≥0.75 under cavitation conditions and ≤0.5 under normal operation. This feature is used to indicate the disorder of energy distribution.

[0056] 2. Instantaneous frequency (time-frequency characteristics)

[0057] In the formula, f(t) is the instantaneous frequency, which is a function of time t. Let be the instantaneous phase function of the signal. The rate of change of phase reflects the transient fluctuations in the signal frequency.

[0058] It should be noted that the instantaneous frequency fluctuation amplitude during cavitation is ≥200Hz. This feature is used to mark the high-frequency transients caused by microbubble collapse.

[0059] 3. Pulse counting (statistical characteristics)

[0060] In the formula, For pulse counting, Let M be the signal amplitude value at time i, M be the number of sampling points within time window T, and T be the length of the statistical time window.

[0061] It should be noted that during cavitation, N P ≥8 / 100ms, this feature is used to indicate the activity level of cavitation events.

[0062] Effectively inject the physical prior knowledge of cavitation into the deep learning model, so that the model can extract the early acoustic feature vector of cavitation. Here, the acoustic feature extraction model is a machine learning model trained on a large amount of data. It is not limited here, as long as it can achieve the corresponding function.

[0063] In this embodiment of the invention, the target digital signal is input into the acoustic signature feature extraction model to obtain the acoustic signature feature vector of the early stage of cavitation.

[0064] Step 104: Input the early acoustic signature feature vector of cavitation into the cavitation identification model to obtain the cavitation confidence level.

[0065] The cavitation identification model refers to an LSTM model trained with a large number of samples. Its core function is to perform pattern recognition and state discrimination on the input early cavitation acoustic feature vector.

[0066] Cavitation confidence score refers to a quantitative probability value output by the cavitation identification model that indicates the current operating state of a coal mine water pump is prone to cavitation failure. It is a continuous value between 0 and 1, and its magnitude directly reflects the severity of the cavitation risk.

[0067] It should be noted that the cavitation identification model is a machine learning model trained on a large number of samples of acoustic feature vectors under various working conditions, including normal state, early cavitation, and mid-term cavitation. It is not limited here, as long as it can accurately establish the complex correspondence between feature combinations and cavitation state and realize the function of outputting cavitation confidence.

[0068] In this embodiment of the invention, the early acoustic signature feature vector of cavitation is input into the cavitation identification model to obtain the cavitation confidence level.

[0069] Step 105: Based on the cavitation confidence level, match and execute the corresponding graded early warning strategy.

[0070] A graded early warning strategy refers to a set of predefined response mechanisms and handling procedures corresponding to different cavitation risk levels.

[0071] It is worth mentioning that the tiered early warning strategy maps continuous cavitation confidence to discrete early warning levels and specifies clear response actions for each level, thereby achieving progressive and precise early warning from risk alerts to equipment intervention, ensuring equipment safety while maintaining production continuity to the greatest extent.

[0072] In this embodiment of the invention, by accurately comparing continuous cavitation confidence with preset thresholds to determine discrete early warning levels, refined and hierarchical management of cavitation risk from early perception to mid-term alarm is achieved. This overcomes the coarseness of traditional binary alarms, enabling the system to trigger manual intervention through warnings and reports at the first-level warning stage, providing a critical time window for preventive maintenance. When the risk escalates, it can also actively suppress it by automatically reducing the frequency, thus avoiding serious equipment damage while maximizing the continuity of production. This achieves the optimal balance between early warning accuracy, equipment safety, and production stability.

[0073] Example 2: Please see Figure 2 Another intelligent early warning method for cavitation of coal mine water pumps based on intrinsically safe acoustic fingerprint acquisition provided by the present invention includes: Step 201: Collect the original acoustic signature signal of the coal mine water pump in real time through an intrinsically safe acoustic signature sensor array and perform intrinsically safe signal conditioning on the original acoustic signature signal to obtain the preprocessed digital signal.

[0074] Preferably, the intrinsically safe signal conditioning of the original voiceprint signal in step 201 to obtain the preprocessed digital signal may include the following sub-steps: S11. Perform low-noise amplification processing on the original voiceprint signal to obtain the initial voiceprint signal.

[0075] S12. Perform bandpass filtering on the initial voiceprint signal to obtain the intermediate voiceprint signal.

[0076] S13. Perform analog-to-digital conversion on the intermediate voiceprint signal to obtain the preprocessed digital signal.

[0077] Low-noise amplification refers to the process of using an amplifier with a low noise figure to boost weak original acoustic signals while meeting the energy limitations of intrinsically safe circuits.

[0078] The initial voiceprint signal refers to the analog voiceprint signal whose amplitude has been increased after low-noise amplification, but which still contains noise across the entire frequency band.

[0079] Bandpass filtering refers to the process of using a filter to allow signals within a specific frequency band to pass through, while attenuating or suppressing frequency components outside that frequency band.

[0080] Intermediate acoustic signature signal refers to an analog acoustic signature signal that, after bandpass filtering, highlights the cavitation characteristic frequency band while effectively suppressing the main out-of-band interference.

[0081] Analog-to-digital conversion refers to the process of converting an intermediate audio signal in analog form into a discrete digital sequence.

[0082] In this embodiment of the invention, a low-noise preamplifier is used to amplify the original voiceprint signal to obtain an initial voiceprint signal. Then, a second-order Butterworth bandpass filter with a passband frequency of 1kHz to 8kHz is used to filter out 50Hz / 100Hz power frequency vibration harmonics and high-frequency electromagnetic interference above 10kHz, outputting an intermediate voiceprint signal. Finally, a 24-bit high-precision analog-to-digital converter is used to digitize the intermediate voiceprint signal at a sampling rate of 22.05kHz to obtain a preprocessed digital signal.

[0083] Step 202: Perform noise reduction processing on the preprocessed digital signal to obtain the target digital signal.

[0084] Preferably, step 202 may include the following sub-steps: S21. Perform wavelet packet decomposition on the preprocessed digital signal to obtain multi-level wavelet packet coefficients.

[0085] S22. Perform thresholding based on adaptive thresholding on the multi-level wavelet packet coefficients to obtain the denoised wavelet packet coefficients.

[0086] S23. Perform wavelet packet reconstruction on the denoised wavelet packet coefficients to obtain the target digital signal.

[0087] Wavelet packet decomposition is a time-frequency analysis method that can perform multi-level and fine-grained division of signals across the entire frequency band. It expands the signal simultaneously in the time and frequency domains using a set of orthogonal or bioorthogonal wavelet basis functions, achieving uniform decomposition of the high-frequency and low-frequency components of the signal.

[0088] Multi-level wavelet packet coefficients refer to the set of expansion coefficients corresponding to each decomposition level and frequency sub-band after the signal has undergone wavelet packet decomposition.

[0089] Adaptive thresholding refers to a denoising method that dynamically adjusts the threshold value based on the characteristics of the signal itself or the noise level of different sub-bands.

[0090] The wavelet packet coefficients after noise reduction refer to the set of wavelet packet coefficients that are considered to be mainly composed of effective signals after adaptive thresholding.

[0091] Wavelet packet reconstruction refers to the process of gradually synthesizing and restoring the time-domain signal by using the wavelet packet coefficients retained after noise reduction, following the inverse process of wavelet packet decomposition.

[0092] It should be noted that this embodiment uses an improved wavelet packet thresholding denoising algorithm for signal purification. Specifically: First, the Db4 wavelet basis function is selected to perform 3-level wavelet packet decomposition on the preprocessed digital signal, which finely divides the signal frequency band into 8 sub-frequency bands, thereby obtaining multi-level wavelet packet coefficients covering the entire frequency band.

[0093] Then, the obtained wavelet packet coefficients undergo thresholding based on an adaptive threshold. The adaptive threshold is dynamically calculated according to the noise level of the specific wavelet packet subband being processed. For a wavelet packet subband containing N coefficients, the squared values ​​of all coefficients are first sorted in ascending order to obtain an ordered sequence. A risk value sequence is then constructed based on this sequence. This risk value sequence comprehensively evaluates the trade-off between signal preservation and noise removal under different potential thresholds. The risk value with the smallest value is found in this risk value sequence, and the adaptive threshold is taken as the square root of the coefficient amplitude corresponding to this smallest risk value.

[0094] After obtaining the adaptive threshold, the coefficients are processed using an improved wavelet packet thresholding function defined as follows:

[0095] In the formula, Let the original wavelet packet coefficients be the values ​​at the i-th level decomposition and the j-th node. For adaptive threshold, The sign function of the coefficients, These are the processed coefficients.

[0096] This threshold function can effectively alleviate the signal discontinuity caused by the traditional hard threshold function, while avoiding the constant deviation caused by the soft threshold function. Thus, while removing noise, it can better preserve the high-frequency pulse information that characterizes cavitation features. After this processing, the denoised wavelet packet coefficients are obtained.

[0097] Finally, the wavelet packet coefficients after noise reduction are used to reconstruct the wavelet packet and synthesize the target digital signal.

[0098] In this embodiment of the invention, an adaptive threshold selection mechanism can accurately match the time-varying characteristics of complex downhole noise environments, thereby enabling effective extraction of weak acoustic features of early cavitation under strong background interference. At the same time, an improved threshold function is used to denoise the digital signal to obtain the target digital signal, ensuring the integrity and authenticity of the high-frequency pulse information of cavitation features.

[0099] Step 203: Input the target digital signal into the acoustic signature feature extraction model to obtain the acoustic signature feature vector for the early stage of cavitation.

[0100] Preferably, the acoustic signature feature vector in the early stage of cavitation includes frequency domain features, time-frequency features, and statistical features.

[0101] Frequency domain characteristics refer to parameters that quantify the characteristics of a signal from the perspective of its frequency distribution, and are used to characterize the distribution of sound wave energy at different frequency components.

[0102] Time-frequency characteristics refer to parameters that combine time and frequency dimensions to describe the local properties of a signal, and are used to capture the dynamic behavior of the frequency components of a signal as they change over time.

[0103] Statistical characteristics refer to parameters extracted from the statistical distribution of signal amplitude, used to quantify the impact and dispersion of signal waveforms.

[0104] In this embodiment of the invention, the target digital signal obtained after noise reduction is input into the acoustic fingerprint feature extraction model to obtain an early acoustic fingerprint feature vector containing frequency domain features, time-frequency features, and statistics. By integrating the three complementary feature dimensions of frequency domain features, time-frequency features, and statistical features into a single vector of early cavitation acoustic fingerprint feature vector, not only is the physical meaning of various features preserved, but also an acoustic fingerprint that can comprehensively and stably characterize the early cavitation state is formed, effectively overcoming the limitation of insufficient characterization ability of a single feature under complex working conditions.

[0105] Step 204: Input the early acoustic signature feature vector of cavitation into the cavitation identification model to obtain the cavitation confidence level.

[0106] In this embodiment of the invention, the early acoustic signature feature vector of cavitation is input into the cavitation identification model to obtain the cavitation confidence level.

[0107] The specific implementation process of step 204 is similar to that of step 104, and will not be repeated here.

[0108] Step 205: Based on the cavitation confidence level, determine the early warning level corresponding to the cavitation confidence level.

[0109] Preferably, step 205 may include the following sub-steps: S31. Based on the first preset confidence threshold, perform a first-level comparison of the cavitation confidence to obtain the first comparison result.

[0110] S32. Perform a second-level comparison of the cavitation confidence level using the second preset confidence threshold to obtain the second-level comparison result.

[0111] S33. Based on the first comparison result and the second comparison result, generate the warning level.

[0112] The first preset confidence threshold refers to the critical probability value used to initially identify cavitation risk and classify the risk-free state and the early risk state. The first preset confidence threshold can be determined as needed, and in this embodiment, the specific value is 0.65.

[0113] The first comparison result refers to the logical judgment result obtained by comparing the cavitation confidence level with the first preset confidence threshold, which is used to identify whether the early risk threshold has been reached or exceeded.

[0114] The second preset confidence threshold refers to the critical probability value used to identify medium-term cavitation risk and distinguish between early and medium-term risks. The second preset confidence threshold can be determined as needed, and in this embodiment, it is specifically set to 0.85.

[0115] The second-level comparison result refers to the logical judgment result obtained by comparing the cavitation confidence level with the second preset confidence threshold, which is used to identify whether the medium-term risk threshold has been reached or exceeded.

[0116] The warning level refers to the final risk level determined based on the logical combination of the results of the two-level comparison.

[0117] In this embodiment of the invention, the output cavitation confidence level is compared with a first preset confidence threshold. If the cavitation confidence level is less than the first preset confidence threshold, the first comparison result is "not exceeded," and it is initially determined that there is no significant risk. If the cavitation confidence level is greater than or equal to the first preset confidence threshold, the first comparison result is "exceeded," and it is initially determined that there is a risk of cavitation. Then, the cavitation confidence level is compared with a higher second preset confidence threshold: if the cavitation confidence level is less than the second preset confidence threshold, the second comparison result is "not exceeded," and if the cavitation confidence level is greater than or equal to the second preset confidence threshold, the second comparison result is "exceeded." Finally, a final warning level is generated based on the two comparison results: when the first comparison result is "not exceeded," a "no warning" level is generated; when the first comparison result is "exceeded" and the second comparison result is "not exceeded," a "Level 1 Warning" level is generated; when both the first and second comparison results are "exceeded," a "Level 2 Warning" level is generated.

[0118] Step 206: If the warning level is Level 1, issue an alert and generate a risk report.

[0119] Level 1 warning refers to the warning level corresponding to the early stage of cavitation risk.

[0120] Risk reports refer to the creation of analytical documents that include cavitation risk indices, timestamps, and key characteristic parameters.

[0121] In this embodiment of the invention, when the warning level is determined to be Level 1, the locally installed intrinsically safe LED indicator is immediately triggered to emit a solid red light or slow flashing warning. Its power consumption is strictly controlled to ≤0.02W. At the same time, a risk report is generated, which is sent to the monitoring center through an intrinsically safe communication link. The report clearly indicates that the "cavitation risk index" is in the range of 65-80, and includes the instantaneous values ​​of key characteristic parameters such as time, location information, power spectral entropy, and instantaneous frequency, providing decision support for operation and maintenance personnel and suggesting that they adjust operating parameters such as water pump flow rate.

[0122] Step 207: If the warning level is Level II, adjust the coal mine water pump speed and generate a risk report.

[0123] Level 2 warning refers to the warning level corresponding to the medium-term risk of cavitation.

[0124] In this embodiment of the invention, when the system determines the warning level to be Level II, in addition to executing the local LED red flashing warning and generating a risk report, it will also automatically send a frequency reduction control command to the water pump frequency converter cabinet through the intrinsically safe communication unit. This command controls the water pump frequency converter cabinet to adjust the operating frequency of the water pump drive motor, reducing the water pump speed to 80% of the rated speed. Through proactive intervention, the operating point of the water pump can be quickly changed, moving it away from the operating condition area prone to severe cavitation, thereby effectively delaying the development of cavitation and preventing equipment damage. The status and parameters of the entire frequency reduction process are synchronously recorded in the risk report.

[0125] Example 3: Please see Figures 3-4 The present invention provides an intelligent early warning system for cavitation of coal mine water pumps based on intrinsically safe acoustic fingerprint acquisition, comprising: an intrinsically safe acoustic fingerprint acquisition device 1, an intrinsically safe monitoring substation 2, a control center 3, and a water pump frequency converter cabinet 4. The intrinsically safe acoustic fingerprint acquisition device 1 is installed on the surface of a coal mine water pump. The intrinsically safe acoustic fingerprint acquisition device 1 includes an intrinsically safe acoustic fingerprint sensor array 11, an intrinsically safe signal conditioning module 12, and a communication unit 13. The intrinsically safe acoustic sensor array 11 is used to collect the raw acoustic signal of the coal mine water pump in real time and transmit the raw acoustic signal to the intrinsically safe signal conditioning module 12. The intrinsically safe signal conditioning module 12 is used to perform low-noise amplification, bandpass filtering and analog-to-digital conversion on the original voiceprint signal to obtain a pre-processed digital signal, and then transmit the digital signal to the communication unit 13. Communication unit 13 is used to transmit digital signals to intrinsically safe monitoring substation 2 via intrinsically safe communication link; Intrinsic monitoring substation 2 is used to receive and forward digital signals to control center 3; Control center 3 is used to extract voiceprint features and identify cavitation from the received digital signals, generate cavitation confidence scores, and generate graded early warning commands based on the cavitation confidence scores. The pump frequency converter cabinet 4 is connected to the control center 3 and is used to receive graded early warning commands and execute corresponding pump speed adjustment operations.

[0126] It should be noted that the intrinsically safe acoustic fingerprint acquisition device 1 is fixedly installed on the surface of the coal mine water pump housing via a magnetic base with an adsorption force ≥50N. It combines convenient installation with operational stability in underground vibration environments. The device adopts an integrated lightweight design, with the outer shell made of magnesium-aluminum alloy die-casting. Its external dimensions are 120mm×80mm×50mm, volume is 0.48L, and weight is ≤1.2kg. The outer shell surface is coated with polytetrafluoroethylene, which has excellent resistance to coal mine water corrosion. The overall protection level reaches IP67, and the impact resistance level is IK08.

[0127] The intrinsically safe acoustic sensor array 11 is arranged at the front end of the housing of the intrinsically safe acoustic sensor 1, close to the housing and preferably facing the impeller of the water pump, so as to best capture the high-frequency acoustic emission signal generated by the collapse of microbubbles in the early stage of cavitation. The array consists of three Exia Class I electret microphones arranged in an isosceles triangle with a spacing of 3cm. Each microphone has a built-in metal mesh dust cover with a diameter of 0.5mm in front of it, which effectively blocks coal dust intrusion, prevents sensitivity attenuation and device damage, and ensures that the sensor frequency response range fully covers the cavitation characteristic frequency band of 500Hz–10kHz, and maintains high sensitivity of -40dB±2dB and working stability in the strong electromagnetic interference environment downhole.

[0128] It is worth mentioning that the microphone model is BM-20A.

[0129] The intrinsically safe signal conditioning module 12, designed with fully intrinsically safe circuitry, is used to condition the raw acoustic signature signal acquired by the sensor array. Specifically, it includes: A low-noise preamplifier with a gain range of 20-40dB is used to initially amplify weak signals and increase their amplitude. The module is powered by 3.3V and has a quiescent current of ≤5mA to ensure that the output power meets intrinsic safety limits. A second-order Butterworth filter with a passband of 1kHz–8kHz is used to effectively suppress low-frequency motor vibration and high-frequency electromagnetic noise, highlighting the cavitation characteristic frequency band. A 24-bit high-precision ADC is used to convert analog signals into digital signals at a sampling rate of 22.05kHz with a power consumption of ≤0.05W, meeting the Nyquist sampling requirements and preserving signal details.

[0130] Preferably, the gain is 30dB.

[0131] Communication unit 13 is based on MSP430F5529 microcontroller and integrates Exia Class I RS485 communication module with a transmission rate of 9600bps and a communication current of ≤8mA. It transmits the conditioned digital signal to intrinsically safe monitoring substation 2 through intrinsically safe cable and anti-loosening thread interface.

[0132] Preferably, the intrinsically safe cable is model MYQ-0.38 / 0.66.

[0133] Preferably, the intrinsically safe voiceprint acquisition device 1 further includes a power supply module, which provides intrinsically safe power to the intrinsically safe voiceprint sensor array 11, the intrinsically safe signal conditioning module 12, and the communication unit 13.

[0134] It is worth mentioning that the power module is connected to the intrinsically safe power supply box installed in the underground roadway via an intrinsically safe cable, receives the DC 12V power output from the box, and converts it to DC 3.3V / 500mA through the LP2985 LDO voltage regulator chip. This power supply powers the intrinsically safe acoustic sensor array 11, the intrinsically safe signal conditioning module 12, and the communication unit 13. The power module has overcurrent and overvoltage protection functions, ensuring that the total power consumption is ≤1.65W and the output current is ≤30mA, fully meeting the GB3836.4 intrinsic safety standard.

[0135] Preferably, the intrinsically safe voiceprint acquisition device 1 also includes an intrinsically safe LED indicator light, which is disposed on the outer surface of the intrinsically safe protective housing and is communicatively connected to the control center 3 to receive the graded early warning command and provide local light alarm prompts.

[0136] Preferably, the intrinsically safe voiceprint acquisition device 1 also includes a data management unit, which is used to allocate fixed storage space for key information such as voiceprint feature data and early warning event logs. When the storage is full, it automatically overwrites the oldest data to ensure that only the most recent 7 days of running data are always retained in the Flash. At the same time, it can also compress each voiceprint data.

[0137] It should be noted that the implementation principle of the above system is consistent with a method for intelligent early warning of cavitation in coal mine water pumps based on intrinsically safe acoustic fingerprint acquisition. For the sake of convenience and brevity, those skilled in the art can clearly understand that the specific working process of the system and modules described above can be referred to the corresponding process in the aforementioned method embodiments, and will not be repeated here.

[0138] Example 4: An electronic device according to an embodiment of the present invention includes: a memory and a processor, wherein the memory stores a computer program; when the computer program is executed by the processor, the processor performs the intelligent early warning method for coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition as described in any of the above embodiments.

[0139] The memory can be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. The memory has storage space for program code used to perform any of the method steps described above. For example, the storage space for program code may include individual program codes for implementing the various steps in the methods described above. This program code can be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, compact discs (CDs), memory cards, or floppy disks. The program code may be compressed, for example, in a suitable form. When run by a computing processing device, this code causes the computing processing device to perform the various steps in the methods described above.

[0140] Example 5: This invention provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed, it implements the intelligent early warning method for coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition, as described in any of the above embodiments.

[0141] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0142] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0143] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0144] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0145] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0146] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent early warning of cavitation in coal mine water pumps based on intrinsically safe acoustic fingerprint acquisition, characterized in that, include: The original acoustic signature signal of the coal mine water pump is acquired in real time by an intrinsically safe acoustic signature sensor array, and the original acoustic signature signal is intrinsically safe signal conditioned to obtain a pre-processed digital signal. The preprocessed digital signal is subjected to noise reduction processing to obtain the target digital signal; The target digital signal is input into the acoustic signature feature extraction model to obtain the acoustic signature feature vector of the early stage of cavitation. The early acoustic signature feature vector of cavitation is input into the cavitation identification model to obtain the cavitation confidence level; Based on the cavitation confidence level, a corresponding graded early warning strategy is matched and executed.

2. The intelligent early warning method for coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition according to claim 1, characterized in that, The intrinsically safe signal conditioning of the original voiceprint signal to obtain the preprocessed digital signal includes: The original voiceprint signal is amplified with low noise to obtain the initial voiceprint signal; The initial voiceprint signal is bandpass filtered to obtain the intermediate voiceprint signal; The intermediate voiceprint signal is converted from analog to digital to obtain the preprocessed digital signal.

3. The intelligent early warning method for coal mine water pump cavitation based on intrinsically safe acoustic signature acquisition according to claim 1, characterized in that, The step of performing noise reduction processing on the preprocessed digital signal to obtain the target digital signal includes: The preprocessed digital signal is subjected to wavelet packet decomposition to obtain multi-level wavelet packet coefficients; The multi-level wavelet packet coefficients are subjected to thresholding based on adaptive thresholds to obtain the denoised wavelet packet coefficients. The target digital signal is obtained by reconstructing the wavelet packet coefficients after noise reduction using wavelet packet reconstruction.

4. The intelligent early warning method for coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition according to claim 1, characterized in that, The voiceprint feature vector includes frequency domain features, time-frequency features, and statistical features.

5. The intelligent early warning method for coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition according to claim 1, characterized in that, The step of matching and executing a corresponding graded early warning strategy based on the cavitation confidence level includes: Based on the cavitation confidence level, the warning level corresponding to the cavitation confidence level is determined; If the warning level is Level 1, an alert will be issued and a risk report will be generated; If the warning level is Level II, the speed of the coal mine water pump will be adjusted and a risk report will be generated.

6. The intelligent early warning method for coal mine water pump cavitation based on intrinsically safe acoustic signature acquisition according to claim 5, characterized in that, The step of determining the early warning level corresponding to the cavitation confidence level includes: A first-level comparison is performed on the cavitation confidence based on a first preset confidence threshold to obtain a first comparison result; A second comparison is performed on the cavitation confidence level using a second preset confidence threshold to obtain a second comparison result. Based on the first comparison result and the second comparison result, an early warning level is generated.

7. A coal mine water pump cavitation intelligent early warning system based on intrinsically safe acoustic signature acquisition, characterized in that, include: Intrinsically safe acoustic fingerprint acquisition device, intrinsically safe monitoring substation, control center, and water pump frequency converter cabinet; The intrinsically safe acoustic fingerprint acquisition device is installed on the surface of the coal mine water pump. The intrinsically safe acoustic fingerprint acquisition device includes an intrinsically safe acoustic fingerprint sensor array, an intrinsically safe signal conditioning module, and a communication unit. The intrinsically safe acoustic sensor array is used to collect the original acoustic signal of the coal mine water pump in real time and transmit the original acoustic signal to the intrinsically safe signal conditioning module. The intrinsically safe signal conditioning module is used to perform low-noise amplification, bandpass filtering and analog-to-digital conversion on the original voiceprint signal to obtain a pre-processed digital signal, and then transmit the digital signal to the communication unit. The communication unit is used to transmit the digital signal to the intrinsically safe monitoring substation via an intrinsically safe communication link; The intrinsically safe monitoring substation is used to receive and forward the digital signals to the control center; The control center is used to extract voiceprint features and identify cavitation from the received digital signals, generate cavitation confidence scores, and generate graded early warning commands based on the cavitation confidence scores. The pump frequency converter cabinet is communicatively connected to the control center and is used to receive the graded early warning commands and execute the corresponding pump speed adjustment operations.

8. The intelligent early warning system for coal mine water pump cavitation based on intrinsically safe acoustic signature acquisition according to claim 7, characterized in that, The intrinsically safe voiceprint acquisition device also includes: a power module; The power module is used to provide intrinsically safe power to the intrinsically safe acoustic sensor array, the intrinsically safe signal conditioning module and the communication unit.

9. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and executed as described in any one of claims 1 to 6, which is a method for intelligent early warning of coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition.

10. A computer-readable storage medium, characterized in that, The computer program is stored and can be loaded by a processor and executed as described in any one of claims 1 to 6, which is a method for intelligent early warning of coal mine water pump cavitation based on intrinsically safe acoustic fingerprint acquisition.

Images