An intelligent early warning method and system for abnormal state of high-pressure spray dust removal system

Abstract

The application provides an abnormal state intelligent early warning method and system of a high-pressure spray dust removal system, applied to the technical field of high-pressure spray dust removal systems, obtains original vibration signals, interface state parameters and feedback response signals, determines the signal transmission degradation degree of the sensor according to the information, then compensates the original vibration signals in the frequency domain to obtain calibrated vibration signals, and finally extracts abnormal characteristic frequency components for wear early warning, which can effectively solve the problem of inaccurate early warning caused by sensor signal transmission degradation and internal performance drift, and has the advantages of improving the reliability and accuracy of the early warning system.

Description

Technical Field

[0001] This application relates to the technical field of high-pressure spray dust removal systems, and in particular to an intelligent early warning method and system for abnormal states of high-pressure spray dust removal systems. Background Technology

[0002] In modern industrial production, high-pressure spray dust suppression systems are commonly used to suppress dust. Rotating equipment, such as high-pressure pumps, fans, and compressors, are the core power components for industrial dust removal, and the health of their bearings directly affects the safety and efficiency of the entire production line.

[0003] In existing technologies, typical bearing wear early warning methods include the following steps: installing vibration sensors on the bearing housing of rotating equipment to continuously collect vibration signals; converting the time-domain signal into a spectrum using Fourier transform; extracting characteristic frequency components related to bearing faults; and issuing an early warning message when the amplitude or energy of the characteristic frequency exceeds a preset threshold. For example, Chinese invention patent CN104155108A discloses a rolling bearing fault diagnosis method based on vibration time-frequency analysis, which achieves early warning by calculating the energy accumulation at the fault characteristic frequency. Furthermore, some existing technologies have introduced intelligent algorithms such as adaptive thresholds and neural networks to improve the accuracy of fault identification.

[0004] However, all of the aforementioned existing technologies are based on a common premise: the vibration sensor itself has ideal and constant performance, and the signal transmission path between the sensor and the device is stable and unchanging. Under this assumption, all changes in vibration signals are assumed to originate from changes in the state of the monitored bearing, and the reliability of the early warning system depends entirely on the accuracy of the data collected by the sensor.

[0005] In practical industrial applications, especially in harsh environments with high dust, high humidity, and slurry accumulation, the above assumptions often fail to hold true. In real-world environments such as cement production, high dust levels combine with water mist to form a viscous slurry, which gradually accumulates on the sensor surface and mounting interface, creating an unstable damping layer. This alters the mechanical impedance between the sensor and the measured structure, causing weak high-frequency fault signals to attenuate, become distorted, or be contaminated with false frequencies during transmission. Simultaneously, high-humidity and corrosive environments can cause sensor packaging aging and internal component drift, further contaminating the data. When contaminated vibration signals are input into analysis methods trained under ideal conditions, the method struggles to distinguish between genuine bearing wear and spectral changes caused by sensor connection degradation: wear signals may be missed due to attenuation, while abnormal frequencies generated by loose connections or circuit faults may be misjudged as bearing failures, triggering false alarms.

[0006] Therefore, how to accurately distinguish between actual bearing failures and signal anomalies caused by sensor degradation in harsh environments, and how to actively compensate for sensor performance degradation to reduce false alarms and missed alarms, is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0007] In view of the shortcomings of the prior art, this application provides an intelligent early warning method and system for abnormal states of a high-pressure spray dust removal system, which can effectively solve the problem of inaccurate early warning caused by sensor signal transmission degradation and internal performance drift, and has the advantage of improving the reliability and accuracy of the early warning system.

[0008] In a first aspect, a method for intelligent early warning of abnormal states in a high-pressure spray dust suppression system is provided. The high-pressure spray dust suppression system includes a rotating device, and the method includes the following steps:

[0009] S1: Acquire the raw vibration signal collected by the sensor installed on the rotating equipment;

[0010] S2: Obtain the interface state parameters between the sensor and the mounting interface of the rotating device;

[0011] S3: Apply a preset excitation signal to the sensor and acquire the feedback response signal generated by the sensor in response to the excitation signal;

[0012] S4: Determine the degree of signal transmission degradation of the sensor based on the interface state parameters and the feedback response signal;

[0013] S5: Based on the degree of signal transmission degradation, a preset compensation model is invoked to perform frequency domain compensation on the original vibration signal to obtain a calibrated vibration signal;

[0014] S6: Extract the abnormal characteristic frequency components from the calibrated vibration signal, and provide wear warning based on the amplitude of the abnormal characteristic frequency components.

[0015] Furthermore, step S2 includes:

[0016] S21: Apply an AC test signal to the interface by means of at least a pair of microelectrodes pre-set at the contact interface between the sensor and the rotating device;

[0017] S22: Measure the response current or voltage of the AC test signal at the interface, and calculate the equivalent complex impedance of the interface, wherein the equivalent complex impedance includes a resistive component and a capacitive component.

[0018] S23: Compare the equivalent complex impedance with the pre-stored health reference impedance value to calculate the impedance deviation; the interface state parameters include the equivalent complex impedance and the impedance deviation.

[0019] Furthermore, step S3 includes:

[0020] S31: Monitor the operating condition of the rotating equipment. When it is determined that the rotating equipment is in a stable operating condition of no load or low load, trigger an excitation action according to a preset time interval.

[0021] S32: By using a miniature self-excited pulse generator integrated inside the sensor, a preset self-excited pulse signal with a known frequency and known amplitude is applied to the sensitive element or mounting base of the sensor as the preset excitation signal.

[0022] S33: Acquire the real-time response waveform generated by the sensor under the action of the self-excited pulse signal;

[0023] S34: Perform time-domain and frequency-domain analysis on the real-time response waveform, and extract at least one characteristic parameter from the amplitude attenuation percentage, frequency drift, total waveform distortion rate, and signal-to-noise ratio change of the real-time response waveform relative to the preset standard response waveform, as the feedback response signal.

[0024] Furthermore, step S2 includes the following:

[0025] S24: Obtain the impedance-attenuation mapping relationship that has been pre-calibrated through experiments. The mapping relationship records the corresponding relationship between impedance deviation and interface coupling attenuation factor at different frequencies.

[0026] S25: Input the currently calculated impedance deviation into the mapping relationship to obtain the initial attenuation factor corresponding to the current impedance deviation;

[0027] S26: Obtain the frequency components of the real-time rotational speed of the rotating device and determine the frequency band range of interest;

[0028] S27: The initial attenuation factor is frequency-corrected according to the frequency band range to obtain a frequency-dependent attenuation factor that varies with frequency. The frequency-dependent attenuation factor is output as the interface coupling attenuation factor, which is used to characterize the degree of vibration signal transmission loss at the interface between the sensor and the rotating equipment.

[0029] Furthermore, step S3 includes the following:

[0030] S35: Calculate the ratio of the amplitude attenuation percentage to the preset amplitude attenuation threshold, and use it as the amplitude drift factor;

[0031] S36: Calculate the ratio of the frequency drift amount to the preset frequency tolerance, and use it as the frequency drift factor;

[0032] S37: The amplitude drift factor and frequency drift factor are weighted and synthesized to obtain the internal performance drift amount, which is used to characterize the degree of signal distortion caused by aging of internal components or changes in response characteristics of the sensor.

[0033] Furthermore, step S4 includes:

[0034] S41: Extract the interface coupling attenuation factor from the interface state parameters, and extract the internal performance drift from the feedback response signal;

[0035] S42: Normalize the interface coupling attenuation factor and the internal performance drift respectively to obtain the dimensionless interface degradation index and internal degradation index.

[0036] S43: The interface degradation index and the internal degradation index are weighted and summed according to preset weighting coefficients to calculate the comprehensive degradation index, which is used as the degree of signal transmission degradation.

[0037] Furthermore, step S5 includes:

[0038] S51: Perform a Fourier transform on the original vibration signal to obtain the original vibration spectrum, which contains amplitude information corresponding to different frequency components;

[0039] S52: Based on the comprehensive degradation index, call the pre-calibrated compensation function to determine the amplitude transfer function of the sensor for different frequency signals under the current degradation state, wherein the compensation function is the compensation model, which records the reciprocal of amplitude attenuation or gain coefficient at each frequency point under different degradation indices.

[0040] S53: Divide the amplitude of each frequency point in the original vibration spectrum by the transfer coefficient of the corresponding frequency point in the amplitude transfer function to obtain the compensated amplitude, thus forming the calibration spectrum;

[0041] S54: Perform an inverse Fourier transform on the calibration spectrum to obtain the calibrated vibration signal in the time domain.

[0042] Furthermore, in step S52, the pre-calibration step of the compensation function includes:

[0043] S521: Before the sensor leaves the factory, place the sensor on a standard vibration table, apply standard excitation signals of different frequencies and amplitudes, and measure the sensor's response in a healthy state as a reference.

[0044] S522: Simulate different degrees of degradation and measure the sensor's response to standard excitations at various frequencies under each degree of degradation; among which, simulating different degrees of degradation includes, but is not limited to, adding slurry layers of different thicknesses to the mounting interface, changing bolt torque, and subjecting the sensor to accelerated aging treatment;

[0045] S523: Calculate the transmission coefficient at each frequency under each degree of degradation, and use its reciprocal as the compensation gain to fit the compensation function.

[0046] Furthermore, step S6 includes:

[0047] S61: Obtain the geometric parameters and real-time rotational speed of the bearing on the rotating equipment, wherein the geometric parameters include the bearing pitch circle diameter, rolling element diameter, number of rolling elements, and contact angle;

[0048] S62: Calculate at least one bearing fault characteristic frequency based on the geometric parameters and real-time rotational speed, wherein the fault characteristic frequency includes one or more of the following: inner ring fault frequency, outer ring fault frequency, rolling element fault frequency, and cage fault frequency;

[0049] S63: Perform spectrum analysis on the calibrated vibration signal to extract the amplitude at the fault characteristic frequency and its integer multiple harmonic frequencies;

[0050] S64: Compare the extracted amplitude with a preset warning threshold. When the amplitude or energy value exceeds the warning threshold, output wear warning information and attach a sensor status report to the warning information. The sensor status report includes at least the current degree of signal transmission degradation of the sensor and the compensation parameters used for frequency domain compensation.

[0051] Secondly, an intelligent early warning system for abnormal states of a high-pressure spray dust removal system, the system being used to implement the steps of any of the methods described above, the system comprising:

[0052] First acquisition module: acquires the raw vibration signals collected by the sensors installed on the rotating equipment;

[0053] Second acquisition module: acquires interface status parameters between the sensor and the installation interface of the rotating device;

[0054] Excitation application module: applies a preset excitation signal to the sensor and acquires the feedback response signal generated by the sensor in response to the excitation signal;

[0055] Degradation determination module: Determines the degree of signal transmission degradation of the sensor based on the interface state parameters and the feedback response signal;

[0056] Compensation module: Based on the degree of signal transmission degradation, it calls a preset compensation model to perform frequency domain compensation on the original vibration signal to obtain a calibrated vibration signal;

[0057] Early warning module: Extracts abnormal characteristic frequency components from the calibrated vibration signal and provides wear warning based on the amplitude of the abnormal characteristic frequency components.

[0058] Beneficial effects: The intelligent early warning method and system for abnormal states of high-pressure spray dust removal system proposed in this application acquires the original vibration signal, interface state parameters and feedback response signal, and determines the degree of signal transmission degradation of the sensor based on this information. Then, frequency domain compensation is performed on the original vibration signal to obtain the calibrated vibration signal. Finally, abnormal characteristic frequency components are extracted for wear early warning. This method can effectively solve the problem of inaccurate early warning caused by sensor signal transmission degradation and internal performance drift, and has the advantage of improving the reliability and accuracy of the early warning system. Attached Figure Description

[0059] Figure 1 This is a flowchart of an intelligent early warning method for abnormal states in a high-pressure spray dust removal system proposed in this application.

[0060] Figure 2 This is a structural diagram of an intelligent early warning system for abnormal states in a high-pressure spray dust removal system proposed in this application.

[0061] Figure 3 This is a schematic diagram of an intelligent early warning system for abnormal states of a high-pressure spray dust removal system proposed in this application.

[0062] Labeling Explanation: 201, First Acquisition Module; 202, Second Acquisition Module; 203, Excitation Application Module; 204, Deterioration Determination Module; 205, Compensation Module; 206, Early Warning Module. Detailed Implementation

[0063] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. The components of the embodiments of this application described and marked in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0064] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0065] Please refer to Figure 1 A method for intelligent early warning of abnormal states in a high-pressure spray dust suppression system, the high-pressure spray dust suppression system including a rotating device, the method including the following steps:

[0066] S1: Acquire the raw vibration signal collected by the sensor installed on the rotating equipment;

[0067] S2: Obtain the interface status parameters between the sensor and the rotating equipment mounting interface;

[0068] S3: Apply a preset excitation signal to the sensor and obtain the feedback response signal generated by the sensor in response to the excitation signal;

[0069] S4: Determine the degree of signal transmission degradation of the sensor based on the interface status parameters and feedback response signals;

[0070] S5: Based on the degree of signal transmission degradation, the preset compensation model is invoked to perform frequency domain compensation on the original vibration signal to obtain the calibrated vibration signal;

[0071] S6: Extract abnormal characteristic frequency components from the calibrated vibration signal and provide wear warning based on the amplitude of the abnormal characteristic frequency components.

[0072] The core working principle of this application is as follows: Instead of viewing the vibration sensor as a consistently accurate and reliable measuring tool, it is treated as a dynamically changing object requiring monitoring and diagnosis. By endowing the sensor with self-sensing and self-calibration capabilities, a comprehensive health check of the sensor itself is performed before conducting a health assessment of rotating equipment. This check includes two dimensions: first, checking the physical connection between the sensor and the rotating equipment to ensure its robustness and the unobstructed signal transmission path; second, checking whether the sensor's internal sensing elements and processing circuits maintain their factory performance. Once sensor degradation is detected, whether due to external connection problems or internal performance decline, the method activates a compensation mechanism. Based on the quantified degree of degradation, the method performs real-time frequency domain correction and calibration of the currently acquired raw vibration signal. Only after this purification and calibration is the vibration signal sent to the subsequent wear warning analysis stage. By ensuring accuracy from the data source, this method effectively eliminates interference introduced by sensor degradation, thereby accurately identifying weak signals truly caused by equipment wear in harsh environments, significantly reducing false alarms and missed alarms, and improving the reliability of intelligent early warning devices and the trust of maintenance personnel.

[0073] First, accelerometers deployed on rotating equipment, such as the bearing housing of a high-pressure pump, continuously capture vibration waveforms during equipment operation to obtain raw vibration signals. However, as mentioned earlier, the reliability of this raw vibration signal is questionable under harsh operating conditions.

[0074] Next, the interface state parameters between the sensor and the rotating equipment mounting interface are acquired. This step is the first step in achieving sensor self-diagnosis, and its purpose is to quantify the mechanical coupling state between the sensor and the structure under test. An unstable mounting interface is a major external cause of vibration signal attenuation and distortion.

[0075] Next, a preset excitation signal is applied to the sensor, and the sensor's feedback response signal to the excitation signal is acquired. This step is the second step of the sensor's self-diagnosis, and its purpose is to assess whether the performance of the sensor's internal components has drifted or aged. By applying a known standard input and observing its actual output, it is possible to accurately determine whether the sensor's own transfer function has changed.

[0076] Subsequently, based on the interface status parameters and feedback response signals, the degree of signal transmission degradation of the sensor is determined. This step is a comprehensive evaluation of the diagnostic results of the first two steps, integrating the degradation of the external mounting interface and the performance degradation of internal components into a quantitative index to comprehensively describe the degree of distortion in the entire link from receiving vibration to outputting an electrical signal.

[0077] Next, once the degree of sensor degradation is determined, a pre-established mathematical model is used to reverse-correct the original vibration signal, which is full of noise and distortion. This compensation operation is performed in the frequency domain, allowing different compensation amounts to be applied to signals with different frequency components, thereby achieving accurate signal recovery.

[0078] As a specific implementation method, the compensation process is executed in the edge computing gateway at the cement production site. In situations where the ambient noise sensed by the sensors is high and dynamically changing, the compensation model is a pre-calibrated compensation function. By calling this pre-calibrated compensation function, the compensation function records the reciprocal of the amplitude attenuation or the gain coefficient at each frequency point under different degradation indices.

[0079] When the high-pressure pump is in the variable frequency speed control stage, speed fluctuations will cause a main frequency offset. The compensation model automatically adjusts the weighting coefficients of the high-frequency and low-frequency bands based on the real-time detected main shaft frequency. When performing compensation in the frequency domain, the Fourier transform spectrum data is divided into several characteristic frequency domain intervals. For specific intervals containing bearing fault characteristic frequencies, a higher-precision compensation function is applied for refinement and correction.

[0080] If the overall degradation index indicates extremely severe interface contamination, the compensation model will automatically trigger gain limiting protection to prevent spurious spectral interference caused by overcompensation. The compensated frequency domain data is then restored to the time domain signal through inverse Fourier transform. Compared to the original signal, this calibrated waveform retains more subtle impact characteristics while eliminating low-frequency vibration energy caused by loose sensor mounting, thus restoring the time domain and pulse indicators to their true levels.

[0081] Finally, abnormal characteristic frequency components are extracted from the calibrated vibration signal, and wear warnings are issued based on the amplitude of these components. After compensation and calibration, and with the vibration signal restored to its true state, traditional bearing wear characteristic analysis is then performed. Because the accuracy of the input data is ensured, the reliability of the warning results is significantly improved.

[0082] Furthermore, step S2 includes:

[0083] S21: Apply an AC test signal to the interface through at least one pair of microelectrodes pre-set at the contact interface between the sensor and the rotating device;

[0084] S22: Measure the response current or voltage of the AC test signal at the interface and calculate the equivalent complex impedance of the interface, which includes resistive and capacitive components.

[0085] S23: Compare the equivalent complex impedance with the pre-stored health reference impedance value and calculate the impedance deviation; the interface state parameters include the equivalent complex impedance and the impedance deviation.

[0086] First, an AC test signal is applied to the interface via at least one pair of microelectrodes pre-installed at the contact interface between the sensor and the rotating device. During sensor manufacturing, one or more pairs of concentric ring-shaped or interdigitated microelectrodes can be integrated into the bottom surface of the base where the sensor contacts the device using micromachining techniques. This pair of electrodes is connected to a signal processing unit inside the sensor. During operation, the signal processing unit generates a weak AC voltage signal with a known frequency and amplitude, which is applied between the pair of electrodes.

[0087] Secondly, the response current or voltage of the AC test signal at the interface is measured to calculate the equivalent complex impedance of the interface, which includes resistive and capacitive components. When an AC voltage is applied to the electrodes, an electric field is generated between the sensor base, the interface medium, and the surface of the rotating device, producing a response current. By simultaneously measuring the amplitude and phase of this response current and comparing it with the input voltage signal, the equivalent complex impedance of the interface can be accurately calculated according to the complex form of Ohm's law. This complex impedance value contains two parts: a resistive component, which mainly reflects the conductivity and contact resistance of the interface material; and a capacitive component, which mainly reflects the dielectric constant and equivalent capacitance of the interface material.

[0088] Finally, the equivalent complex impedance is compared with the pre-stored health reference impedance value to calculate the impedance deviation. Interface state parameters include the equivalent complex impedance and the impedance deviation. When the sensor is initially installed and in good connection condition, a set of complex impedance values ​​at this time is measured and stored as a health reference. During subsequent operation, when slurry enters the installation interface, the dielectric constant and conductivity of the slurry are significantly different from those of air or a clean metal surface, causing significant changes in the measured equivalent capacitance and resistance. Similarly, if the connecting bolts loosen, resulting in reduced contact pressure and increased contact resistance, this will also cause changes in the complex impedance. By vector subtracting the real-time measured complex impedance from the health reference value, a complex form of impedance deviation can be obtained. The magnitude and phase angle of this deviation can accurately quantify the degree of degradation of the installation interface.

[0089] Furthermore, step S3 includes:

[0090] S31: Monitor the operating condition of the rotating equipment. When it is determined that the rotating equipment is in a stable operating condition of no load or low load, trigger the excitation to apply the action according to the preset time interval.

[0091] S32: By using a miniature self-excited pulse generator integrated inside the sensor, a preset self-excited pulse signal with a known frequency and known amplitude is applied to the sensor's sensitive element or mounting base as a preset excitation signal.

[0092] S33: Acquire the real-time response waveform generated by the sensor under the action of a self-excited pulse signal;

[0093] S34: Perform time-domain and frequency-domain analysis on the real-time response waveform, and extract at least one characteristic parameter from the following relative to the preset standard response waveform: amplitude attenuation percentage, frequency drift, total waveform distortion rate, and signal-to-noise ratio change, as the feedback response signal.

[0094] First, monitor the operating conditions of the rotating equipment. When the equipment is determined to be in a stable, unloaded or low-load condition, trigger an excitation action at preset time intervals. During normal operation, the rotating equipment generates strong vibrations, which can become background noise for evaluating sensor performance. To obtain a clean sensor self-test signal, a suitable timing needs to be chosen. By monitoring the equipment's rotational speed, load current, or other operating parameters, it's possible to determine when the equipment is in a stable operating state—stopped, idling, or under extremely low load. Within these time windows, the equipment's own vibration interference is minimal, making it an ideal time for sensor self-testing. Self-testing can be automatically triggered at preset time cycles, such as once per hour or per shift.

[0095] Secondly, a pre-set self-excited pulse signal with a known frequency and amplitude is applied to the sensor's sensitive element or mounting base via a miniature self-excited pulse generator integrated within the sensor, serving as the pre-set excitation signal. Inside the sensor, a miniature actuator, such as a piezoelectric ceramic or a miniature electromagnetic coil, is integrated adjacent to its core piezoelectric sensitive element or directly fixed to its mounting base. Through precise control of the drive circuit by an internal microcontroller, this actuator can generate a standardized, repeatable mechanical vibration pulse when needed. The waveform, frequency components, and amplitude of this pulse are pre-set and calibrated, for example, a sinusoidal envelope pulse with a center frequency of 10kHz and a peak acceleration of 0.1g.

[0096] As one specific implementation method, under the control of the microcontroller unit inside the sensor, a separate piezoelectric excitation circuit periodically generates high-frequency narrow pulse excitation. During periods of low background noise when the pump unit is shut down for self-testing or running under no-load, the excitation circuit generates a set of swept-frequency signals, with a frequency coverage range encompassing three times the frequency of common bearing faults.

[0097] Next, the real-time response waveform generated by the sensor under the action of the self-excited pulse signal is acquired. When the self-excited pulse occurs, this standard mechanical vibration is directly transmitted to the sensor's sensitive element, which, like in response to external vibration, converts it into an electrical signal output. The data acquisition circuit inside the sensor captures this complete response process at a high sampling rate, obtaining a real-time response waveform data. The sensor receives the real-time response waveform generated by this frequency sweep signal.

[0098] Finally, time-domain and frequency-domain analyses are performed on the real-time response waveform to extract at least one characteristic parameter from among the following relative to the preset standard response waveform: amplitude attenuation percentage, frequency drift, total waveform distortion rate, and signal-to-noise ratio change. This parameter serves as the feedback response signal. The acquired real-time response waveform is then compared in detail with the standard response waveform measured by the sensor before it leaves the factory or during initial installation, under healthy conditions.

[0099] The comparison dimensions can include: in the time domain, calculating the peak or root mean square value of the response waveform and comparing it with a standard value to obtain the percentage of amplitude attenuation; in the frequency domain, performing a Fourier transform on the response waveform to analyze whether its dominant frequency has shifted and calculating the frequency drift; simultaneously, analyzing whether unexpected harmonics or noise appear in the spectrum besides the dominant frequency, and calculating the total harmonic distortion rate or total waveform distortion rate; it can also analyze the energy ratio of useful components to noise components in the signal and calculate the change in signal-to-noise ratio. These quantified parameters together constitute the feedback response signal, comprehensively reflecting the performance status of the entire link within the sensor from mechanical vibration to electrical signal conversion.

[0100] If the internal components of the sensor become damp, age, or develop microcracks, the response to this standard self-excited pulse will inevitably deteriorate, manifesting as amplitude attenuation, frequency inaccuracy, or waveform distortion. The processing unit compares the envelope of the current response waveform with the factory-stored standard envelope to identify whether there is a response delay or phase lag.

[0101] In extreme high-temperature environments, the sensitivity of the piezoelectric elements inside the sensor may temporarily drift. By calculating the energy centroid offset of the real-time response waveform, a precise feedback response signal is obtained. This active detection method does not rely on external mechanical vibration and can assess the sensor's hearing status before the equipment is officially started. If the feedback response signal shows severe nonlinear distortion in the signal waveform, the confidence weight of the sensor is reduced in the warning result, and a prompt is issued to the operator to maintain the sensor hardware.

[0102] Furthermore, step S2 includes the following:

[0103] S24: Obtain the impedance-attenuation mapping relationship that has been pre-calibrated through experiments. The mapping relationship is recorded as the correspondence between impedance deviation and interface coupling attenuation factor at different frequencies.

[0104] S25: Input the calculated impedance deviation into the mapping relationship to obtain the initial attenuation factor corresponding to the current impedance deviation;

[0105] S26: Obtain the frequency components of the real-time rotational speed of the rotating equipment and determine the frequency band of interest.

[0106] S27: The initial attenuation factor is frequency-corrected according to the frequency band range to obtain the frequency-dependent attenuation factor that varies with the frequency. The frequency-dependent attenuation factor is output as the interface coupling attenuation factor. The interface coupling attenuation factor is used to characterize the degree of vibration signal transmission loss at the installation interface between the sensor and the rotating equipment.

[0107] First, an impedance-attenuation mapping relationship, pre-calibrated through experiments, is obtained. This mapping relationship records the correspondence between impedance deviation and interface coupling attenuation factor at different frequencies. During the sensor development phase, numerous calibration experiments are conducted on a standard vibration table in the laboratory. By artificially introducing different degrees of degradation at the sensor mounting interface—for example, adding simulated slurry layers of different thicknesses and materials, or tightening bolts with different torques—various interface conditions that may occur in the field are simulated. Under each degradation condition, on the one hand, the interface complex impedance is measured to calculate the impedance deviation; on the other hand, a series of standard sinusoidal vibration signals of different frequencies are applied to the vibration table, and the ratio of the actual output signal amplitude of the sensor to the standard input amplitude is measured, thus obtaining the vibration transmission attenuation factor at each frequency point under that degradation state. Through regression analysis and fitting of a large amount of experimental data, a multi-dimensional database or mathematical model, namely the impedance-attenuation mapping relationship, can be established. This model can directly query or calculate the corresponding interface coupling attenuation factor based on the input impedance deviation and the frequency of interest.

[0108] The interface coupling attenuation factor is used to quantify the mechanical energy loss at the mounting interface caused by factors such as foreign matter filling and loosening. When establishing the mapping relationship, the proportion of energy loss conducted from the sensor mounting base to the sensitive element is experimentally measured under different impedance deviations. The impedance deviation includes not only changes in the resistive component but also fluctuations in the capacitive component caused by the dielectric properties of the slurry. This mapping relationship is achieved by constructing a frequency-dependent attenuation matrix. For each set of impedance deviation data, the mapping matrix outputs a set of frequency response correction values. This ensures accurate restoration of the signal transmittance at the interface under different levels of contamination.

[0109] Secondly, the calculated impedance deviation is input into the mapping relationship to obtain the initial attenuation factor corresponding to the current impedance deviation. During field operation, once the real-time impedance deviation is obtained through electrical measurements, this pre-established mapping relationship can be used to quickly find an initial attenuation factor that matches the current interface state.

[0110] Next, the frequency components of the real-time rotational speed of the rotating equipment are obtained to determine the frequency band of interest. Since the fault characteristic frequencies of rotating components such as bearings are usually closely related to their rotational speed, it is necessary to obtain the real-time rotational speed of the equipment. Based on the rotational speed and the bearing's geometric parameters, the frequency band containing the fault characteristic frequencies that need to be monitored can be calculated. For example, the inner ring fault frequency is usually located in the mid-to-high frequency band, while the cage fault frequency is located in the low frequency band.

[0111] Finally, the initial attenuation factor is frequency-corrected based on the frequency band range to obtain a frequency-dependent attenuation factor that varies with frequency. This frequency-dependent attenuation factor is then output as the interface coupling attenuation factor. The interface coupling attenuation factor characterizes the degree of vibration signal transmission loss at the sensor-rotating equipment mounting interface. Since the attenuation effect of interface degradation on vibration signals of different frequencies is not uniform, high-frequency signals are generally more easily attenuated. Therefore, the initial attenuation factor obtained from the mapping relationship needs to be corrected according to the current frequency band of interest. For example, if the current focus is on inner-ring faults in the high-frequency band, a frequency correction factor greater than 1 may be needed to multiply a base attenuation factor. The frequency-dependent attenuation factor obtained after frequency correction more accurately reflects the degree of transmission loss of the specific fault signal of interest at the current mounting interface.

[0112] Furthermore, step S3 includes the following:

[0113] S35: Calculate the ratio of the amplitude attenuation percentage to the preset amplitude attenuation threshold, and use it as the amplitude drift factor;

[0114] S36: Calculate the ratio of the frequency drift to the preset frequency tolerance, and use it as the frequency drift factor;

[0115] S37: Weighted synthesis of amplitude drift factor and frequency drift factor to obtain internal performance drift, which is used to characterize the degree of signal distortion caused by aging of internal components or changes in response characteristics of the sensor.

[0116] First, the ratio of the amplitude attenuation percentage to a preset amplitude attenuation threshold is calculated as the amplitude drift factor. The percentage attenuation of the response signal amplitude obtained from the self-excitation test is compared to an acceptable maximum attenuation threshold determined based on sensor design specifications and aging experiments. For example, if the threshold is 20% and the current attenuation is 10%, then the amplitude drift factor is 10% ÷ 20% = 0.5. This factor visually represents the degree of degradation in the sensor's amplitude response.

[0117] Secondly, the ratio of the frequency drift to a preset frequency tolerance is calculated as the frequency drift factor. Similarly, the drift of the measured main frequency of the response signal is compared with a preset frequency accuracy tolerance. For example, if the tolerance is 5Hz and the current drift is 2Hz, then the frequency drift factor is 2÷5=0.4. This factor reflects the degree of degradation in the sensor's frequency response accuracy.

[0118] Finally, the amplitude drift factor and frequency drift factor are weighted and synthesized to obtain the internal performance drift. The internal performance drift is used to characterize the degree of signal distortion caused by aging of internal sensor components or changes in response characteristics. Considering that the degradation of amplitude response and frequency response may have different effects on the final warning result, different weight coefficients can be assigned to these two drift factors. For example, the weight of amplitude drift can be set to 0.6, and the weight of frequency drift can be set to 0.4.

[0119] By weighted summing of the two factors, for example, 0.5 × 0.6 + 0.4 × 0.4 = 0.46, a comprehensive, dimensionless internal performance drift can be obtained to fully assess the sensor's internal health. The internal performance drift reflects errors caused by aging of the sensor's internal piezoelectric ceramics or circuitry. In the weighted synthesis process, the amplitude drift factor represents the overall attenuation of signal strength, while the frequency drift factor represents the nonlinear distortion of phase and frequency response. Different weighting coefficients are assigned according to the sensitivity weights determined in the sensor's hardware specifications. When a frequency offset exceeding a preset range is detected, it indicates a change in the internal resonant frequency. In this case, the weighting ratio of the frequency drift factor is increased to reflect the significant impact of signal frequency distortion on the final analysis.

[0120] Furthermore, step S4 includes:

[0121] S41: Extract the interface coupling attenuation factor from the interface state parameters and extract the internal performance drift from the feedback response signal;

[0122] S42: Normalize the interface coupling attenuation factor and the internal performance drift respectively to obtain the dimensionless interface degradation index and internal degradation index.

[0123] S43: The interface degradation index and the internal degradation index are weighted and summed according to the preset weight coefficients to calculate the comprehensive degradation index, which is used as the degree of signal transmission degradation.

[0124] First, the interface coupling attenuation factor is extracted from the interface state parameters, and the internal performance drift is extracted from the feedback response signal. This is the first step in data fusion, using the two key indicators calculated earlier—the interface coupling attenuation factor characterizing external influences and the internal performance drift characterizing internal influences—as inputs.

[0125] Secondly, the interface coupling attenuation factor and internal performance drift are normalized to obtain dimensionless interface degradation index and internal degradation index. Since these two indices may have different physical meanings and numerical ranges, they need to be converted to a unified scale for comparison and synthesis. Through normalization, for example, mapping them to the interval between 0 and 1, where 0 represents healthy and without degradation, and 1 represents complete failure, two directly comparable dimensionless indices can be obtained. Normalization transforms parameters with different physical dimensions into dimensionless values ​​between zero and one using a linear mapping function. The interface degradation index is calculated by subtracting the healthy baseline value from the real-time impedance deviation and then dividing by a preset limit deviation. The internal degradation index compares the comprehensive performance offset with the failure threshold. This approach eliminates the computational obstacles caused by different units such as current, voltage, and frequency, allowing for weighted fusion of the external installation environment and internal hardware status within the same evaluation system.

[0126] Finally, the interface degradation index and the internal degradation index are weighted and summed according to preset weighting coefficients to calculate the comprehensive degradation index, which serves as the indicator of signal transmission degradation. Based on an understanding of specific application scenarios, different importance weights can be assigned to interface degradation and internal degradation. For example, in environments with extremely severe slurry contamination, interface degradation may be the primary concern, and its weight can be set higher, such as 0.7. Conversely, in a relatively clean but highly humid environment, internal moisture aging of the sensor may be the main problem, and its weight can be set even higher. Through weighted summation, a single, comprehensive comprehensive degradation index is obtained. This index accurately quantifies the degree of degradation in the overall signal transmission capability of the sensor compared to its healthy state, providing a direct and reliable basis for subsequent signal compensation.

[0127] Furthermore, step S5 includes:

[0128] S51: Perform a Fourier transform on the original vibration signal to obtain the original vibration spectrum, which contains amplitude information corresponding to different frequency components.

[0129] S52: Based on the comprehensive degradation index, call the pre-calibrated compensation function to determine the amplitude transfer function of the sensor for different frequency signals under the current degradation state. The compensation function is a compensation model that records the reciprocal of amplitude attenuation or gain coefficient at each frequency point under different degradation indices.

[0130] S53: Divide the amplitude of each frequency point in the original vibration spectrum by the transfer coefficient of the corresponding frequency point in the amplitude transfer function to obtain the compensated amplitude, thus forming the calibration spectrum;

[0131] S54: Perform an inverse Fourier transform on the calibration spectrum to obtain the calibrated vibration signal in the time domain.

[0132] First, a Fourier transform is performed on the original vibration signal to obtain the original vibration spectrum, which contains amplitude information corresponding to different frequency components. The acquired time-domain original vibration signal is then decomposed into a superposition of a series of sine waves of different frequencies using a Fast Fourier Transform algorithm. The transformed result is the original vibration spectrum, where the horizontal axis represents frequency and the vertical axis represents the amplitude corresponding to each frequency component.

[0133] Secondly, based on the comprehensive degradation index, a pre-calibrated compensation function is invoked to determine the amplitude transfer function of the sensor for different frequency signals under the current degradation state. The compensation function is a compensation model that records the reciprocal of the amplitude attenuation or the gain coefficient at each frequency point under different degradation indices. The core here is a pre-established compensation model, which is essentially a function.

[0134] Its input is the current overall degradation index, and its output is a curve, namely the amplitude transfer function. This curve describes the sensor's response to vibration signals of different frequencies under the current degradation state. For example, the curve might show a signal transfer coefficient of 0.9 at 100Hz, while the transfer coefficient is only 0.6 at 5000Hz, indicating more severe attenuation at higher frequencies. The compensation model typically stores the reciprocal of this transfer coefficient, i.e., the compensation gain coefficient. The amplitude transfer function describes the sensor's actual gain performance for input signals at various frequencies under a specific degradation index. This function is derived by interpolating the response values ​​at discrete frequency points in the original spectrum with the coefficient library in the compensation model. The compensation gain coefficient is the reciprocal of the transfer coefficient. When a signal in a certain frequency band is attenuated by 50% due to interface degradation, the gain coefficient generated by the compensation function at that frequency point is 2, thus compensating for the lost vibration amplitude.

[0135] Next, the amplitude at each frequency point in the original vibration spectrum is divided by the transfer coefficient of the corresponding frequency point in the amplitude transfer function to obtain the compensated amplitude, forming the calibration spectrum. This is the core of the compensation operation. For each frequency point in the original vibration spectrum, the amplitude corresponding to that frequency point is read, and then the transfer coefficient of the corresponding frequency point is found in the amplitude transfer function obtained in the previous step. Dividing the original amplitude by this transfer coefficient yields the compensated and corrected true amplitude. For example, if the amplitude measured at 5000 Hz is 0.03g, and the transfer function shows a transfer coefficient of 0.6 for that frequency, then the compensated true amplitude should be 0.03g ÷ 0.6 = 0.05g. After performing this operation on all frequency points, a completely new, calibrated spectrum is obtained.

[0136] Finally, an inverse Fourier transform is performed on the calibration spectrum to obtain the calibrated vibration signal in the time domain. The amplitude-corrected calibration spectrum is then resynthesized into a vibration waveform in the time domain using an inverse fast Fourier transform algorithm. This final time-domain signal, after eliminating the effects of sensor degradation, is the closest to the actual vibration state of the rotating equipment and can be reliably used for subsequent fault diagnosis and early warning.

[0137] Furthermore, in step S52, the pre-calibration step of the compensation function includes:

[0138] S521: Before the sensor leaves the factory, place the sensor on a standard vibration table, apply standard excitation signals of different frequencies and amplitudes, and measure the sensor's response in a healthy state as a reference.

[0139] S522: Simulate different degrees of degradation and measure the sensor's response to standard excitations at various frequencies under each degree of degradation; among which, simulating different degrees of degradation includes, but is not limited to, adding slurry layers of different thicknesses to the mounting interface, changing bolt torque, and subjecting the sensor to accelerated aging treatment;

[0140] S523: Calculate the transfer coefficient at each frequency under each degree of degradation, and use its reciprocal as the compensation gain to fit a compensation function.

[0141] To ensure the accuracy of compensation, the pre-calibration process of the compensation function is crucial. This calibration step is typically completed before the sensor leaves the factory. A standard vibration table is a device capable of generating precisely known and repeatable vibration signals. A brand-new, fully functional sensor is mounted on the vibration table, and the table generates a series of standard sinusoidal vibrations of different frequencies and amplitudes covering the sensor's operating frequency range. Simultaneously, the sensor's output signal is recorded. By comparing the input and output, the reference transfer function of the sensor at various frequencies under healthy conditions can be obtained; theoretically, the transfer coefficient should be close to 1.

[0142] Secondly, different degrees of degradation were simulated, and the sensor's response to standard excitations at various frequencies was measured under each degradation level. Simulating different degradation levels included, but was not limited to, adding slurry layers of varying thicknesses to the mounting interface, changing bolt torque, and subjecting the sensor to accelerated aging. This is crucial for calibration. To simulate mounting interface degradation, silicone sheets or adhesive slurries of varying thicknesses could be added between the sensor and the vibration table. To simulate bolt loosening, a torque wrench could be used to tighten the sensor to a torque lower than the standard value. To simulate internal component aging, the sensor could be placed in a high-temperature, high-humidity environmental chamber for accelerated aging testing. Under each simulated degradation condition, the first step—applying a standard excitation across the entire frequency band—was repeated, and the sensor's response was measured.

[0143] Finally, the transfer coefficients at each frequency under each degree of degradation are calculated, and their reciprocals are used as the compensation gain to fit a compensation function. For each degree of degradation, the measured frequency response amplitudes are compared with the baseline response amplitudes under healthy conditions to calculate the transfer coefficient at each frequency point under that degradation state. Then, the reciprocals of these transfer coefficients are taken to obtain the compensation gain. Finally, these degradation degree indices and the corresponding frequency compensation gain data are input into a multilayer perceptron neural network model for training. The neural network model contains an input layer, multiple hidden layers, and an output layer. The input layer receives the quantified degradation degree indices, and the output layer outputs a curve containing the compensation gain at each frequency point. Through backpropagation and gradient descent optimization, the weights and biases of the neural network are adjusted until the model can accurately predict the corresponding frequency compensation gain curve based on the input degradation indices. The finally trained neural network model serves as the compensation function and is stored in the sensor's firmware.

[0144] Furthermore, step S6 includes:

[0145] S61: Obtain the geometric parameters and real-time rotational speed of the bearing on the rotating equipment. The geometric parameters include the bearing pitch circle diameter, rolling element diameter, number of rolling elements, and contact angle.

[0146] S62: Calculate at least one bearing failure characteristic frequency based on geometric parameters and real-time rotational speed. The failure characteristic frequency includes one or more of the following: inner ring failure frequency, outer ring failure frequency, rolling element failure frequency, and cage failure frequency.

[0147] S63: Perform spectrum analysis on the calibrated vibration signal to extract the fault characteristic frequency and the amplitude at its integer multiple harmonic frequencies;

[0148] S64: Compare the extracted amplitude with the preset warning threshold. When the amplitude or energy value exceeds the warning threshold, output wear warning information and attach a sensor status report to the warning information. The sensor status report includes at least the current degree of signal transmission degradation of the sensor and the compensation parameters used for frequency domain compensation.

[0149] Among these parameters, the bearing pitch circle diameter, rolling element diameter, number of rolling elements, and contact angle are the theoretical basis for calculating the characteristic frequencies of bearing failures, and can usually be found in the equipment manual or bearing model. Real-time rotational speed, on the other hand, needs to be obtained from a speed sensor or from the frequency signal of the motor driver.

[0150] Secondly, by utilizing mature rolling bearing fault diagnosis theories and formulas, and substituting the aforementioned parameters, the characteristic frequencies generated in the vibration signal when different components of the bearing experience localized damage can be accurately calculated. Specifically, when calculating the bearing fault characteristic frequencies, the processing unit first reads the geometric parameters of the target bearing from the memory, including the bearing pitch circle diameter D, rolling element diameter d, number of rolling elements Z, and contact angle α. Simultaneously, the rotational speed f_r of the high-pressure pump shaft is acquired in real time via a speed sensor. Subsequently, the processing unit substitutes these parameters into the following rolling bearing fault characteristic frequency calculation formula:

[0151] Outer ring fault frequency BPFO=(Z / 2)×f_r×(1-(d / D)×cos(α))

[0152] Inner ring fault frequency BPFI=(Z / 2)×f_r×(1+(d / D)×cos(α))

[0153] The rolling element spin frequency BSF = (D / (2d)) × f_r × (1 - (d / D)^2 × cos(α)^2)

[0154] Cage failure frequency FTF = (1 / 2) × f_r × (1 - (d / D) × cos(α)).

[0155] These calculations yield specific values ​​for the outer race fault frequency, inner race fault frequency, rolling element spin frequency, and cage fault frequency at the current rotational speed. These results are then passed to the spectrum analysis module as a reference for characteristic frequency extraction in the calibrated vibration spectrum.

[0156] Next, a Fourier transform is performed on the compensated and calibrated vibration signal to obtain the calibration spectrum. Then, on this spectrum, the positions of each fault characteristic frequency and its second, third, and other harmonic frequencies calculated in the previous step are precisely located, and the amplitudes corresponding to these frequency points are extracted.

[0157] Finally, the extracted amplitude is compared with a preset warning threshold. When the amplitude exceeds the warning threshold, a wear warning message is output, along with a sensor status report. The sensor status report includes at least the current degree of signal transmission degradation of the sensor and the compensation parameters used for frequency domain compensation. The warning threshold is typically set based on historical data, health benchmarks of similar equipment, and maintenance experience. Once the extracted characteristic frequency amplitude exceeds this threshold, it is determined that the bearing is at risk of wear, and a warning message is generated.

[0158] This warning message goes beyond a simple alarm; it includes a detailed sensor status report. The report clearly informs maintenance personnel of the degree of sensor degradation (e.g., an overall degradation index of 35%) and the parameters used to calculate the warning. This transparency significantly enhances the reliability of the warning, helping maintenance personnel accurately pinpoint the root cause of the problem and make informed maintenance decisions.

[0159] Please refer to Figure 2 , Figure 3 An intelligent early warning system for abnormal states of a high-pressure spray dust suppression system, the system being used to implement the steps of any of the above methods, the system comprising:

[0160] First acquisition module 201: Acquires the raw vibration signal collected by the sensor installed on the rotating equipment;

[0161] Second acquisition module 202: Acquires interface status parameters between the sensor and the rotating equipment installation interface;

[0162] Excitation application module 203: applies a preset excitation signal to the sensor and acquires the feedback response signal generated by the sensor in response to the excitation signal;

[0163] Degradation determination module 204: Determines the degree of signal transmission degradation of the sensor based on the interface status parameters and feedback response signals;

[0164] Compensation module 205: Based on the degree of signal transmission degradation, it calls a preset compensation model to perform frequency domain compensation on the original vibration signal to obtain a calibrated vibration signal;

[0165] Early warning module 206: Extracts abnormal characteristic frequency components from the calibrated vibration signal and provides wear warning based on the amplitude of the abnormal characteristic frequency components.

[0166] The first acquisition module 201 can be a data acquisition unit, including an analog-to-digital converter (ADC) and a data storage device, for converting the analog vibration signal output by the sensor into a digital signal and storing it. For example, the module can be configured to receive signals from a piezoelectric accelerometer, an eddy current sensor, or a fiber optic sensor via a wired connection (such as a BNC interface) or a wireless connection (such as Bluetooth or Wi-Fi).

[0167] The second acquisition module 202 can be a standalone impedance measurement unit, connected to the mounting interface via a microelectrode array, applying a test signal and measuring the response to calculate the interface impedance. Alternatively, this module can be a micro pressure sensor array integrated on a sensor base for real-time monitoring of interface contact pressure. This module can also be an ultrasonic transceiver, evaluating interface coupling quality by transmitting and receiving ultrasonic waves.

[0168] The excitation application module 203 can be a combination of a signal generator and an exciter. For example, a programmable digital signal generator can be connected to a piezoelectric or electromagnetic exciter to apply an excitation signal to the sensor body or its mounting base. Alternatively, this module can integrate a miniature self-excited pulse generator inside the sensor, which is triggered by the control unit. The feedback response signal is acquired and processed by the data acquisition unit through the sensor's own signal output channel.

[0169] The degradation determination module 204 can be an embedded processor or a host computer software module, receiving interface state parameters from the second acquisition module and feedback response signals from the excitation application module. This module has a pre-set algorithm to analyze these parameters, for example, by comparing the interface impedance with a reference value, analyzing the amplitude attenuation, frequency drift, or waveform distortion of the feedback response signal, etc., to comprehensively calculate the degree of signal transmission degradation and output a comprehensive degradation index.

[0170] The compensation module 205 can be a digital signal processor (DSP) or a high-performance microcontroller, which internally stores a preset compensation model. After receiving the degradation degree information output by the degradation determination module, the compensation module 205 retrieves or calculates the corresponding compensation parameters from the compensation model based on the information, performs a Fourier transform on the original vibration signal acquired by the first acquisition module, corrects the amplitude or phase in the frequency domain, and then performs an inverse Fourier transform to output the calibrated vibration signal.

[0171] The early warning module 206 can be an analysis software module running on an embedded system or industrial PC. It receives the calibrated vibration signal from the compensation module and performs spectral analysis on it. This module has a built-in algorithm for calculating bearing fault characteristic frequencies, calculating the fault characteristic frequencies based on the input equipment geometry parameters and rotational speed. Subsequently, the module extracts the amplitudes of these characteristic frequencies and their harmonics and compares them with preset early warning thresholds. Once the amplitude exceeds the threshold, the module generates and outputs wear warning information, which may include a sensor status report for maintenance personnel to diagnose the problem.

[0172] Unlike existing technologies that directly analyze the original vibration signal, the compensation module in this application determines the degree of signal transmission degradation based on the degradation level of the module output and performs frequency domain compensation on the original vibration signal to obtain a calibrated vibration signal. This compensation process effectively eliminates signal attenuation and distortion caused by the degradation of the sensor and mounting interface, making the calibrated vibration signal more realistically reflect the actual vibration state of the rotating equipment.

[0173] Therefore, the early warning module 206 extracts abnormal characteristic frequency components based on the calibrated vibration signal and performs wear early warning, significantly improving its accuracy and reliability. The system of this application can effectively distinguish between actual bearing wear and spectral changes caused by sensor connection degradation, avoiding the common problems of missed detections and false alarms in traditional methods. In summary, the early warning system of this application overcomes the shortcomings of insufficient early warning accuracy in harsh industrial environments by introducing an intelligent assessment and frequency domain compensation mechanism for the degree of sensor signal transmission degradation, significantly improving the reliability and effectiveness of intelligent early warning of abnormal states in high-pressure spray dust removal systems.

[0174] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for intelligent early warning of abnormal states in a high-pressure spray dust suppression system, the high-pressure spray dust suppression system including a rotating device, characterized in that, The method includes the following steps: S1: Acquire the raw vibration signal collected by the sensor installed on the rotating equipment; S2: Obtain the interface state parameters between the sensor and the mounting interface of the rotating device; S3: Apply a preset excitation signal to the sensor and acquire the feedback response signal generated by the sensor in response to the excitation signal; S4: Determine the degree of signal transmission degradation of the sensor based on the interface state parameters and the feedback response signal; S5: Based on the degree of signal transmission degradation, a preset compensation model is invoked to perform frequency domain compensation on the original vibration signal to obtain a calibrated vibration signal; S6: Extract the abnormal characteristic frequency components from the calibrated vibration signal, and provide wear warning based on the amplitude of the abnormal characteristic frequency components.

2. The intelligent early warning method for abnormal states of a high-pressure spray dust removal system according to claim 1, characterized in that, Step S2 includes: S21: Apply an AC test signal to the interface by means of at least a pair of microelectrodes pre-set at the contact interface between the sensor and the rotating device; S22: Measure the response current or voltage of the AC test signal at the interface, and calculate the equivalent complex impedance of the interface, wherein the equivalent complex impedance includes a resistive component and a capacitive component. S23: Compare the equivalent complex impedance with the pre-stored health reference impedance value to calculate the impedance deviation; the interface state parameters include the equivalent complex impedance and the impedance deviation.

3. The intelligent early warning method for abnormal states of a high-pressure spray dust removal system according to claim 1, characterized in that, Step S3 includes: S31: Monitor the operating condition of the rotating equipment. When it is determined that the rotating equipment is in a stable operating condition of no load or low load, trigger an excitation action according to a preset time interval. S32: By using a miniature self-excited pulse generator integrated inside the sensor, a preset self-excited pulse signal with a known frequency and known amplitude is applied to the sensitive element or mounting base of the sensor as the preset excitation signal. S33: Acquire the real-time response waveform generated by the sensor under the action of the self-excited pulse signal; S34: Perform time-domain and frequency-domain analysis on the real-time response waveform, and extract at least one characteristic parameter from the amplitude attenuation percentage, frequency drift, total waveform distortion rate, and signal-to-noise ratio change of the real-time response waveform relative to the preset standard response waveform, as the feedback response signal.

4. The intelligent early warning method for abnormal states of a high-pressure spray dust removal system according to claim 2, characterized in that, Step S2 is followed by: S24: Obtain the impedance-attenuation mapping relationship that has been pre-calibrated through experiments. The mapping relationship records the corresponding relationship between impedance deviation and interface coupling attenuation factor at different frequencies. S25: Input the currently calculated impedance deviation into the mapping relationship to obtain the initial attenuation factor corresponding to the current impedance deviation; S26: Obtain the frequency components of the real-time rotational speed of the rotating device and determine the frequency band range of interest; S27: The initial attenuation factor is frequency-corrected according to the frequency band range to obtain a frequency-dependent attenuation factor that varies with frequency. The frequency-dependent attenuation factor is output as the interface coupling attenuation factor, which is used to characterize the degree of vibration signal transmission loss at the interface between the sensor and the rotating equipment.

5. The intelligent early warning method for abnormal states of a high-pressure spray dust removal system according to claim 3, characterized in that, Step S3 is followed by: S35: Calculate the ratio of the amplitude attenuation percentage to the preset amplitude attenuation threshold, and use it as the amplitude drift factor; S36: Calculate the ratio of the frequency drift to the preset frequency tolerance, and use it as the frequency drift factor; S37: The amplitude drift factor and frequency drift factor are weighted and synthesized to obtain the internal performance drift amount, which is used to characterize the degree of signal distortion caused by aging of internal components or changes in response characteristics of the sensor.

6. The intelligent early warning method for abnormal states of a high-pressure spray dust removal system according to claim 1, characterized in that, Step S4 includes: S41: Extract the interface coupling attenuation factor from the interface state parameters, and extract the internal performance drift from the feedback response signal; S42: Normalize the interface coupling attenuation factor and the internal performance drift respectively to obtain the dimensionless interface degradation index and internal degradation index. S43: The interface degradation index and the internal degradation index are weighted and summed according to preset weighting coefficients to calculate the comprehensive degradation index, which is used as the degree of signal transmission degradation.

7. The intelligent early warning method for abnormal states of a high-pressure spray dust removal system according to claim 6, characterized in that, Step S5 includes: S51: Perform a Fourier transform on the original vibration signal to obtain the original vibration spectrum, which contains amplitude information corresponding to different frequency components; S52: Based on the comprehensive degradation index, call the pre-calibrated compensation function to determine the amplitude transfer function of the sensor for different frequency signals under the current degradation state, wherein the compensation function is the compensation model, which records the reciprocal of amplitude attenuation or gain coefficient at each frequency point under different degradation indices. S53: Divide the amplitude of each frequency point in the original vibration spectrum by the transfer coefficient of the corresponding frequency point in the amplitude transfer function to obtain the compensated amplitude, thus forming the calibration spectrum; S54: Perform an inverse Fourier transform on the calibration spectrum to obtain the calibrated vibration signal in the time domain.

8. The intelligent early warning method for abnormal states of a high-pressure spray dust removal system according to claim 7, characterized in that, In step S52, the pre-calibration step of the compensation function includes: S521: Before the sensor leaves the factory, place the sensor on a standard vibration table, apply standard excitation signals of different frequencies and amplitudes, and measure the sensor's response in a healthy state as a reference. S522: Simulate different degrees of degradation and measure the sensor's response to standard excitations at various frequencies under each degree of degradation; among which, simulating different degrees of degradation includes, but is not limited to, adding slurry layers of different thicknesses to the mounting interface, changing bolt torque, and subjecting the sensor to accelerated aging treatment; S523: Calculate the transmission coefficient at each frequency under each degree of degradation, and use its reciprocal as the compensation gain to fit the compensation function.

9. The intelligent early warning method for abnormal states of a high-pressure spray dust removal system according to claim 1, characterized in that, Step S6 includes: S61: Obtain the geometric parameters and real-time rotational speed of the bearing on the rotating equipment, wherein the geometric parameters include the bearing pitch circle diameter, rolling element diameter, number of rolling elements, and contact angle; S62: Calculate at least one bearing fault characteristic frequency based on the geometric parameters and real-time rotational speed, wherein the fault characteristic frequency includes one or more of the following: inner ring fault frequency, outer ring fault frequency, rolling element fault frequency, and cage fault frequency; S63: Perform spectrum analysis on the calibrated vibration signal to extract the amplitude at the fault characteristic frequency and its integer multiple harmonic frequencies; S64: Compare the extracted amplitude with a preset warning threshold. When the amplitude or energy value exceeds the warning threshold, output wear warning information and attach a sensor status report to the warning information. The sensor status report includes at least the current degree of signal transmission degradation of the sensor and the compensation parameters used for frequency domain compensation.

10. An intelligent early warning system for abnormal states of a high-pressure spray dust removal system, characterized in that, The system is used to implement the steps of the method according to any one of claims 1-9, and the system includes: First acquisition module: acquires the raw vibration signals collected by the sensors installed on the rotating equipment; Second acquisition module: acquires interface status parameters between the sensor and the installation interface of the rotating device; Excitation application module: applies a preset excitation signal to the sensor and acquires the feedback response signal generated by the sensor in response to the excitation signal; Degradation determination module: Determines the degree of signal transmission degradation of the sensor based on the interface state parameters and the feedback response signal; Compensation module: Based on the degree of signal transmission degradation, it calls a preset compensation model to perform frequency domain compensation on the original vibration signal to obtain a calibrated vibration signal; Early warning module: Extracts abnormal characteristic frequency components from the calibrated vibration signal and provides wear warning based on the amplitude of the abnormal characteristic frequency components.

Images