Environment dust suppression effect monitoring system and method based on piezoelectric array

Abstract

The application relates to the technical field of monitoring, and provides an environment dust suppression effect monitoring system and method based on a piezoelectric array, wherein a piezoelectric sensing array is unevenly arranged through a sensing module, passive response signals generated in response to dust particle impacts, and active response signals generated in response to frequency-modulated excitation signals; a calibration module compares a preset electromechanical impedance spectrum baseline by applying the frequency-modulated excitation signals, generates a sensitivity factor of the piezoelectric sensing array, and improves the stability and monitoring consistency of the system in a complex and harsh environment; a processing module corrects the passive response signals through the sensitivity factor, and extracts effective impact signals through frequency domain decomposition; an evaluation module reconstructs a dust mass distribution field and converts the dust mass distribution field into an equivalent kinetic energy distribution field through an inverse problem reconstruction of the effective impact signals through a compressed sensing algorithm, compares the equivalent kinetic energy distribution fields before and after dust suppression operation, calculates dust suppression efficiency, and identifies and locates dust suppression blind areas, so that the intelligentization and refinement level of dust suppression operation are improved.

Description

Technical Field

[0001] This application relates to the field of monitoring technology, and in particular to an environmental dust suppression effect monitoring system and method based on piezoelectric arrays. Background Technology

[0002] In industrial production, mining, construction and bulk material storage and transportation, dust diffusion has become an important factor affecting environmental quality and operational safety. Dust suppression monitoring is gradually developing from manual sampling to automated monitoring. Among the existing technologies, the more mature monitoring methods mainly include optical concentration monitoring, gravity settling monitoring, filter weighing monitoring, and indirect monitoring based on vibration sensing.

[0003] Optical monitoring equipment can quickly collect dust concentration data, but it is easily affected by water mist, light, and obstruction, making it difficult to work stably for a long time in high humidity and high dust environments. Gravity settling and filter weighing methods have high accuracy, but they have slow response, cannot achieve dynamic monitoring, and cannot reflect the spatial distribution and movement of dust. Some technologies introduce piezoelectric sensing elements, which use the vibration signals generated by dust impacts to achieve monitoring. They have certain advantages in environmental adaptability and real-time performance, and have become an important development direction in the field of dust suppression effect monitoring.

[0004] Although existing dust suppression monitoring technologies have basic monitoring capabilities, they have significant limitations in complex dust suppression scenarios. Most solutions only focus on dust concentration values ​​and do not fully consider the impact of dust particle kinetic energy on the evaluation of dust suppression effects. Simply judging dust suppression efficiency based on concentration changes makes it difficult to truly reflect the effect of dust suppression operations on inhibiting dust diffusion trends. Piezoelectric sensing units are easily affected by changes in temperature and humidity, dust deposition, and component aging, causing continuous drift in sensing sensitivity and resulting in prominent signal distortion problems.

[0005] However, passively acquired dust impact signals are easily mixed with noise such as environmental vibration and electromagnetic interference, which can easily cause errors in subsequent analysis. To address the shortcomings of the existing technology, the technical problem to be solved in this application is how to achieve dynamic, high-precision and comprehensive monitoring of environmental dust suppression effect based on piezoelectric array under complex working conditions. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this application provides a method and system for monitoring environmental dust suppression effects based on piezoelectric arrays.

[0007] In the first aspect, this application provides an environmental dust suppression effect monitoring system based on piezoelectric array, which includes: a sensing module, a calibration module, a processing module and an evaluation module;

[0008] The sensing module includes a piezoelectric sensing array that is non-uniformly distributed in the dust suppression area, used to respond to passive response signals generated by dust particle impact, and to receive active response signals generated by frequency modulation excitation signals.

[0009] The calibration module is used to apply a frequency-modulated excitation signal to the piezoelectric sensing array, compare the active response signal with a preset mechanical impedance spectrum baseline, and generate a sensitivity factor for the piezoelectric sensing array.

[0010] The processing module is used to correct the passive response signal by a sensitivity factor, and to perform frequency domain decomposition on the corrected passive response signal to extract the effective impact signal.

[0011] The evaluation module is used to reconstruct the effective impact signal by performing an inverse problem using a compressed sensing algorithm, so as to reconstruct the dust mass distribution field in the dust suppression area and convert it into an equivalent kinetic energy distribution field. It also compares the equivalent kinetic energy distribution field before and after the dust suppression operation, calculates the dust suppression efficiency, and identifies the dust suppression blind zone.

[0012] As an optional implementation, the sensing module is specifically used for:

[0013] The dust suppression area is divided into multiple regional units according to the dust diffusion gradient of the dust suppression area. The piezoelectric sensing array is non-uniformly distributed in different regional units, and the spatial coordinates of different piezoelectric sensing units are calibrated.

[0014] The working cycle of the sensing module is divided into independent calibration time slots and monitoring time slots. In the calibration time slot, the frequency modulation excitation signal is received and the corresponding active response signal is generated. In the monitoring time slot, the impact of dust particles is responded to and the corresponding passive response signal is generated. The calibration time slot and the monitoring time slot are executed alternately periodically.

[0015] As an optional implementation, the preset electromechanical impedance spectrum baseline is obtained in advance under standard clean operating conditions, extreme temperature and humidity operating conditions, dust deposition operating conditions, and piezoelectric element aging operating conditions. The amplitude, phase, resonant frequency offset, real part and imaginary part of impedance are extracted under different operating conditions through wavelet denoising and principal component analysis. The impedance spectrum reference features corresponding to the operating conditions are constructed respectively. Then, all impedance spectrum reference features are weighted and fused to form an electromechanical impedance spectrum baseline adapted to different operating conditions.

[0016] As an optional implementation, the sensitivity factor of the generated piezoelectric sensing array includes:

[0017] A frequency-modulated excitation signal is applied to the piezoelectric sensing array during the calibration time slot to obtain the response parameters of the active response signals of different piezoelectric sensing units. The response parameters are then aligned with the impedance spectrum reference characteristics in the preset mechanical impedance spectrum baseline on the frequency axis, and the deviation of the response parameters is calculated.

[0018] Based on the real-time temperature and humidity, dust deposition amount and piezoelectric element working time, weight coefficients for different response parameter deviations are dynamically assigned. After weighted fusion and normalization, the sensitivity factor of the piezoelectric sensing unit is obtained. After weighted averaging, the sensitivity factor of the piezoelectric sensing array is generated.

[0019] As an optional implementation, the extraction of the effective impact signal includes:

[0020] The passive response signal acquired in the monitoring time slot is corrected unit by unit by the sensitivity factor of the piezoelectric sensing array.

[0021] The corrected passive response signal is preprocessed and frequency domain decomposition is performed to filter signal segments with frequency fluctuations within a preset range to obtain the initial effective signal.

[0022] The initial valid signal is verified and the signal is interpolated and completed based on the spatial correlation of the piezoelectric sensor array, and the valid impact signal is output.

[0023] As an optional implementation, the frequency band decomposition refers to dynamically dividing the frequency band intervals corresponding to different dust particle sizes based on the impedance spectrum reference characteristics and the amplitude distribution of the corrected passive response signal, and then performing frequency domain decomposition by dividing the frequency bands using Fourier transform.

[0024] As an optional implementation, reconstructing the dust mass distribution field includes:

[0025] Based on the spatial coordinates and non-uniform density of the piezoelectric sensor array, a compressed sensing observation matrix matching the dust suppression area is constructed.

[0026] The effective impact signal is split according to the dust particle size, and the dust sub-mass distribution field of different dust particle sizes is reconstructed by the compressed sensing observation matrix;

[0027] The dust submass distribution field is weighted and corrected based on the sensitivity factor and impedance spectrum reference characteristics of the piezoelectric sensor array, and the corrected dust submass distribution is fused according to the real-time proportion of dust particle size to generate the dust mass distribution field of the dust suppression area.

[0028] As an optional implementation, converting to the equivalent kinetic energy distribution field includes:

[0029] Based on the piezoelectric impact response characteristics of different dust particle sizes in corresponding frequency bands, the dynamic impact velocities of different dust particle sizes are determined in order to calculate the initial sub-kinetic energy distribution of different dust sub-mass distribution fields.

[0030] The initial subkinetic energy distribution is corrected for operating conditions by using the sensitivity factor and impedance spectrum reference characteristics of the piezoelectric sensing array, and spatial gridding is performed by combining the spatial coordinates of the piezoelectric sensing array with the regional cells.

[0031] The equivalent kinetic energy distribution field of the dust suppression area is generated by fusing and correcting the initial sub-kinetic energy distribution based on the real-time proportion of dust particle size.

[0032] As an optional implementation, the calculation of dust suppression efficiency and identification of dust suppression blind spots includes:

[0033] Obtain the equivalent kinetic energy distribution field before and after dust suppression operation, calculate the single-dimensional dust suppression efficiency by region unit and dust particle size, and perform working condition correction by combining the sensitivity factor and the impedance spectrum reference characteristics.

[0034] The dust suppression efficiency of the dust suppression area is obtained by weighting the dust diffusion gradient and the real-time proportion of dust particle size, and then weighting and fusing the single-dimensional dust suppression efficiency.

[0035] The initial dust suppression blind zone is screened based on the dynamic threshold determined by the dust diffusion gradient, and the dust suppression blind zone is identified by verifying the spatial correlation of the kinetic energy of the surrounding area.

[0036] Secondly, this application provides a method for monitoring the environmental dust suppression effect based on a piezoelectric array. The method includes: non-uniformly distributing a piezoelectric sensor array in the dust suppression area, generating a passive response signal in response to the impact of dust particles, and generating an active response signal by receiving a frequency modulation excitation signal.

[0037] A frequency-modulated excitation signal is applied to the piezoelectric sensing array, and the active response signal is compared with a preset mechanical impedance spectrum baseline to generate a sensitivity factor for the piezoelectric sensing array.

[0038] The passive response signal is corrected by a sensitivity factor, and the corrected passive response signal is then decomposed in the frequency domain to extract the effective impact signal.

[0039] The effective impact signal is reconstructed by performing an inverse problem using a compressed sensing algorithm to rebuild the dust mass distribution field of the dust suppression area and convert it into an equivalent kinetic energy distribution field. The equivalent kinetic energy distribution field before and after the dust suppression operation is compared to calculate the dust suppression efficiency and identify the dust suppression blind zone.

[0040] Compared with the prior art, the beneficial effects of this application are: by non-uniformly arranging the piezoelectric sensor array in the sensing module, it responds to the passive response signal generated by the impact of dust particles and receives the active response signal generated by the frequency modulation excitation signal; the calibration module generates the sensitivity factor of the piezoelectric sensor array by applying the frequency modulation excitation signal and comparing it with the preset mechanical impedance spectrum baseline, thereby offsetting the sensing drift caused by temperature and humidity changes, dust deposition and component aging, and improving the long-term stability and monitoring consistency of the system in complex environments.

[0041] The processing module corrects the passive response signal through a sensitivity factor, then extracts the effective impact signal through frequency domain decomposition, eliminates environmental interference and system errors, and improves the purity and signal-to-noise ratio of the dust characteristic signal, providing a data foundation for subsequent distribution field reconstruction and avoiding reconstruction distortion caused by noise signals. The evaluation module reconstructs the effective impact signal through a compressed sensing algorithm to accurately reconstruct the dust mass distribution field and convert it into an equivalent kinetic energy distribution field, achieving an upgrade from concentration characterization to kinetic energy characterization, which is more in line with the physical mechanism of dust diffusion and dust suppression, making the evaluation of dust suppression effect scientific and comprehensive.

[0042] By comparing the equivalent kinetic energy distribution field before and after dust suppression operations, the dust suppression efficiency can be quantitatively calculated, and areas where the dust suppression effect does not meet expectations can be located. This allows for the accurate identification of dust suppression blind spots, providing data support for the scheduling of dust suppression equipment and the optimization of operating parameters, thereby improving the overall intelligence and precision of dust suppression operations. This application achieves a closed-loop process of signal acquisition, online calibration, signal processing, field reconstruction, and effect evaluation. It has the advantages of strong environmental adaptability, high monitoring accuracy, scientific evaluation system, and the ability to locate blind spots, meeting the needs for continuous, stable, and efficient dust suppression effect monitoring in complex scenarios. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0044] Figure 1 This is a system flowchart of the environmental dust suppression effect monitoring system based on piezoelectric array provided in the embodiments of this application;

[0045] Figure 2 This is a logic flowchart of the reconstructed dust mass distribution field provided in an embodiment of this application;

[0046] Figure 3 This is a flowchart illustrating the environmental dust suppression effect monitoring method based on a piezoelectric array provided in an embodiment of this application. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of this application more apparent and understandable, 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 some embodiments of this application, and not all embodiments.

[0048] Example 1:

[0049] like Figure 1The diagram shown illustrates a system flowchart for an environmental dust suppression effect monitoring system based on a piezoelectric array, which is provided in this application embodiment. The system includes a sensing module, a calibration module, a processing module, and an evaluation module.

[0050] The sensing module includes a piezoelectric sensor array that is non-uniformly distributed in the dust suppression area, used to respond to passive response signals generated by dust particle impacts, and active response signals generated by receiving frequency modulation excitation signals.

[0051] The sensing module is specifically used for:

[0052] Based on the dust diffusion gradient of the dust suppression area, the dust suppression area is divided into multiple regional units. Piezoelectric sensor arrays are non-uniformly distributed in different regional units, and the spatial coordinates of different piezoelectric sensor units are calibrated.

[0053] The working cycle of the sensing module is divided into independent calibration time slots and monitoring time slots. In the calibration time slot, the frequency modulation excitation signal is received and the corresponding active response signal is generated. In the monitoring time slot, the impact of dust particles is responded to and the corresponding passive response signal is generated. The calibration time slot and the monitoring time slot are executed alternately periodically.

[0054] The spatial distribution of dust within the dust suppression area follows the laws of fluid diffusion, exhibiting a non-uniform concentration gradient. Traditional piezoelectric sensor arrays, when uniformly deployed, cannot match this distribution characteristic, leading to insufficient coverage of monitoring units in high-dust-concentration areas and redundant monitoring units in low-concentration areas. Furthermore, the lack of calibration of the spatial coordinates of the sensor units prevents subsequent spatial correlation analysis of signals, directly impacting the overall monitoring accuracy of the system. The dust diffusion gradient is used to characterize the spatial rate of change of dust concentration within the dust suppression area. By conducting multi-point dust concentration measurements on the target dust suppression area and combining the attenuation characteristics of dust spatial diffusion, the entire dust suppression area is divided into regional units corresponding to different dust gradients. Taking an industrial raw material storage dust suppression area as an example, the total area of ​​this area is 4000㎡, which is divided into a high dust gradient core area of ​​1200㎡, a medium dust gradient transition area of ​​1800㎡, and a low dust gradient edge area of ​​1000㎡ based on actual measurements.

[0055] The non-uniform deployment density of the piezoelectric sensor array is determined based on the matching relationship between dust gradient and dust impact signal acquisition requirements. The higher the dust gradient, the higher the frequency of dust impact on the piezoelectric sensor unit, and the greater the required density of piezoelectric sensor units. The deployment parameters are: one piezoelectric sensor unit per 4㎡ in the core area, one piezoelectric sensor unit per 9㎡ in the transition area, and one piezoelectric sensor unit per 20㎡ in the edge area. The spatial coordinate calibration adopts a three-dimensional coordinate system, with the fixed corner of the dust suppression area as the coordinate origin. The X-axis and Y-axis correspond to the planar dimensions of the area, and the Z-axis corresponds to the installation height of the piezoelectric sensor unit. The installation height is uniformly set to 1.1m. Each piezoelectric sensor unit is assigned three-dimensional coordinates. For example, the coordinates of a sensor unit in the core area are (X=15m, Y=12m, Z=1.1m). All coordinate data are stored in correspondence with the sensor unit number.

[0056] This enables the adaptation of monitoring resources to the spatial distribution characteristics of dust, avoids the problems of monitoring redundancy and blind spots, provides a positioning basis for subsequent spatial signal processing, and ensures the effectiveness of spatial correlation of monitoring data; it also affects subsequent time slot division and signal sampling operations. Spatial coordinate data will be used for signal sampling management of calibration time slots and monitoring time slots, providing basic parameters for subsequent processing.

[0057] If the piezoelectric sensing unit simultaneously receives both the frequency modulation excitation signal and the passive response signal generated by dust impact, the two types of signals will cause crosstalk in the time domain, resulting in a significant reduction in the signal-to-noise ratio of the active response signal and the passive response signal, which in turn affects the accuracy of subsequent sensitivity factor calculation and effective impact signal extraction. The division between calibration time slots and monitoring time slots is based on the electromechanical response time characteristics of the piezoelectric sensing unit and the time-domain continuity characteristics of the dust impact signal. The total working cycle of the sensing module is set to 10 seconds, which is determined based on the real-time requirements of dust suppression monitoring and the response stability of the piezoelectric sensing unit. The calibration time slot is 2 seconds long and the monitoring time slot is 8 seconds long. The two types of time slots are executed continuously and periodically in a fixed sequence.

[0058] During the calibration time slot, the sensing module only turns on the receiving channel of the frequency modulation excitation signal and the output channel of the active response signal, while simultaneously turning off the sampling channel of the dust impact passive response signal. During the monitoring time slot, the input path of the frequency modulation excitation signal is cut off, the active response signal sampling channel is turned off, and only the sampling channel of the passive response signal triggered by dust particle impact is enabled. The channel switching logic follows the design specifications for anti-crosstalk in signal acquisition to ensure that there is no time domain overlap in the sampling process of the two types of signals.

[0059] This eliminates crosstalk between active excitation signals and passive impact signals, improves the purity of the two types of original signals, meets the requirements of continuous monitoring in dust suppression scenarios, and ensures the continuity and stability of signal sampling. The active response signal acquired in the calibration time slot is the data for generating the sensitivity factor, while the passive response signal acquired in the monitoring time slot lays the data foundation for signal correction and frequency domain decomposition. The timing logic of the alternating time slots connects the working rhythm of the calibration module and the processing module, ensuring the timing consistency of the monitoring system.

[0060] The calibration module is used to apply a wideband frequency-modulated excitation signal to the piezoelectric sensing array, compare the active response signal with a preset mechanical impedance spectrum baseline, and generate the sensitivity factor of the piezoelectric sensing array.

[0061] It should be explained that the preset electromechanical impedance spectrum baseline is the electromechanical response data obtained in advance under standard clean operating conditions, extreme temperature and humidity operating conditions, dust deposition operating conditions, and piezoelectric element aging operating conditions. The amplitude, phase, resonant frequency offset, real part and imaginary part of impedance are extracted under different operating conditions through wavelet denoising and principal component analysis. The impedance spectrum reference features corresponding to the operating conditions are constructed respectively. Then, all impedance spectrum reference features are weighted and fused to form an electromechanical impedance spectrum baseline adapted to different operating conditions.

[0062] During dust suppression monitoring, the electromechanical response characteristics of piezoelectric sensor arrays can drift due to changes in environmental conditions and their own aging. If a single fixed electromechanical impedance spectrum baseline is used as the calibration benchmark, the comparison deviation between the active response signal and the baseline will increase, thereby affecting the generation accuracy of the sensitivity factor and ultimately reducing the correction effect of the passive response signal. Standard cleanroom conditions, extreme temperature and humidity conditions, dust deposition conditions, and aging conditions of piezoelectric elements are all typical conditions that piezoelectric sensor arrays may face in dust suppression scenarios. These conditions are classified according to the actual range of changes in the dust suppression environment and the working life characteristics of the piezoelectric elements.

[0063] Standard cleanroom conditions refer to baseline conditions where there is no dust, no significant temperature and humidity fluctuations, and the piezoelectric elements are in brand new condition, used to provide basic impedance spectrum characteristics; extreme temperature and humidity conditions refer to the extreme temperature and humidity ranges that may occur in the dust suppression area, set according to the site climate conditions, used to cover the impact of environmental temperature and humidity fluctuations on piezoelectric elements; dust deposition conditions refer to conditions where different thicknesses of dust adhere to the surface of the piezoelectric sensing unit, used to adapt to impedance spectrum drift caused by dust deposition; piezoelectric element aging conditions refer to the state of the piezoelectric element after accumulating a certain amount of operation, used to compensate for changes in response characteristics caused by element aging.

[0064] A wideband frequency-modulated excitation signal is applied to the piezoelectric sensor array. Based on the resonant frequency characteristics of the piezoelectric elements and the signal response range required for dust suppression monitoring, the corresponding frequency range is determined to be 0.1MHz to 10MHz. The amplitude of the excitation signal is set to 5Vpp to ensure that the piezoelectric elements can generate a stable electromechanical response. Electromechanical response data of each piezoelectric sensor unit is acquired under various operating conditions. For example, under standard cleanroom conditions, with the ambient temperature controlled at 25℃ and the relative humidity controlled at 50%, and the piezoelectric elements in a brand-new, unused state, the data for this condition is acquired. Electromechanical response data of all piezoelectric sensing units under various conditions were obtained. The extreme temperature and humidity conditions were divided into two sub-conditions: high temperature and high humidity (45℃, 85%RH) and low temperature and low humidity (-5℃, 15%RH), and electromechanical response data under the corresponding conditions were obtained. The dust deposition condition was divided into dust deposition thicknesses of 0.1mm, 0.3mm, and 0.5mm, and electromechanical response data were obtained after simulating dust deposition. The aging condition of piezoelectric elements was obtained by simulating the aging state of the piezoelectric elements by continuously powering them on for 1000 hours.

[0065] Wavelet denoising and principal component analysis (PCA) were used to remove noise interference from electromechanical response data and extract core features. Wavelet denoising employed a db4 wavelet basis with a decomposition level of 5. By performing wavelet decomposition, thresholding, and inverse wavelet transform on the acquired electromechanical response signal, redundant signals caused by environmental electromagnetic interference and circuit noise were removed. Principal component analysis was used to reduce the dimensionality of the denoised signal data, retaining core parameters that reflect the impedance spectrum characteristics of the piezoelectric element, including amplitude, phase, resonant frequency offset, real part of impedance, and imaginary part of impedance, while eliminating irrelevant redundant features and reducing the complexity of subsequent feature processing. That is, calculate the eigenvalues ​​and cumulative variance contribution rates of each principal component, retain the top p principal components with a cumulative variance contribution rate ≥ 85%, and simultaneously satisfy that the eigenvalues ​​of the retained principal components are ≥ 1. After the above dual screening conditions, finally retain 5 principal components, which correspond exactly to the amplitude, phase, resonant frequency offset, real part of impedance, and imaginary part of impedance. For example, perform principal component analysis on the response signal after noise reduction under standard clean working conditions, extract the amplitude and phase data of each piezoelectric sensing unit at different frequency points, determine the resonant frequency offset, that is, the offset value relative to the standard resonant frequency, and calculate the values ​​of the real part of impedance and the imaginary part of impedance.

[0066] The five types of features extracted above are categorized and organized according to operating conditions to form the impedance spectrum reference features corresponding to each type of operating condition. The impedance spectrum reference features of each type of operating condition correspond to the response characteristics of the piezoelectric sensing unit. For example, the impedance spectrum reference features of the standard clean operating condition are obtained by averaging the features of all piezoelectric sensing units under this operating condition, ensuring the representativeness of the reference features. The impedance spectrum reference features of the extreme temperature and humidity operating conditions are obtained by weighted fusion of the features of the two sub-operating conditions of high temperature and high humidity and low temperature and low humidity, respectively, with the weights set according to the probability of occurrence of the two extreme operating conditions. The impedance spectrum reference features of the dust deposition operating condition and the piezoelectric element aging operating condition are obtained by weighted fusion of the features of the corresponding operating condition level, ensuring that the reference features can cover the full range of changes in the operating conditions.

[0067] Weights were assigned based on the actual frequency of various operating conditions in dust suppression monitoring. Continuous monitoring of these conditions was conducted over a 30-day period in the target dust suppression area, with a statistical period of 1 hour, resulting in 720 statistical periods. The percentage of periods each operating condition actually occurred was calculated: standard cleanroom conditions occurred approximately 288 times (40%), extreme temperature and humidity conditions occurred approximately 144 times (20%), dust deposition conditions occurred approximately 180 times (25%), and piezoelectric element aging conditions occurred approximately 108 times (15%). The impedance spectrum baselines of the four operating conditions are integrated into a dynamic baseline. The baseline is weighted according to the real-time changes in the operating conditions, with a total weight of 1. The baseline is then fused to form an impedance spectrum baseline that is adapted to different operating conditions. For example, the standard clean operating condition has the highest probability of occurrence and is assigned a weight of 0.4; the temperature and humidity extreme operating condition has a lower probability of occurrence and is assigned a weight of 0.2; the dust deposition operating condition is assigned a weight of 0.25; and the piezoelectric element aging operating condition is assigned a weight of 0.15. Through a weighted fusion algorithm, the impedance spectrum reference characteristics of the four operating conditions are integrated into a dynamic baseline. This baseline is matched with the reference characteristics of the corresponding operating condition according to the real-time changes in the operating conditions to ensure the accuracy of the calibration.

[0068] This overcomes the limitations of a single fixed baseline that cannot adapt to changes in operating conditions, ensuring that the baseline accurately matches the response characteristics of piezoelectric elements under various operating conditions in dust suppression scenarios. It also counteracts the impedance spectrum drift caused by fluctuations in ambient temperature and humidity, dust deposition, and element aging, ensuring the reliability of the impedance spectrum reference characteristics. Furthermore, it provides a stable and accurate calibration benchmark for subsequent sensitivity factor generation, avoiding calibration deviations caused by a single operating condition baseline and ensuring the accuracy of the entire calibration process. Finally, it provides a basis for weight allocation under subsequent real-time operating conditions, ensuring that the sensitivity factor can adapt to changes in real-time operating conditions and laying the foundation for the correction of the passive response signal of the processing module.

[0069] Specifically, the sensitivity factors for generating the piezoelectric sensing array include:

[0070] A wideband frequency-modulated excitation signal is applied to the piezoelectric sensing array during the calibration time slot to obtain the response parameters of the active response signals of different piezoelectric sensing units. The response parameters are then aligned with the impedance spectrum reference characteristics in the preset mechanical impedance spectrum baseline on the frequency axis, and the deviation of the response parameters is calculated.

[0071] Based on the real-time temperature and humidity, dust deposition amount and piezoelectric element working time, weight coefficients for different response parameter deviations are dynamically assigned. After weighted fusion and normalization, the sensitivity factor of the piezoelectric sensing unit is obtained. After weighted averaging, the sensitivity factor of the piezoelectric sensing array is generated.

[0072] Manufacturing tolerances in piezoelectric sensor arrays lead to differences in initial characteristics, and real-time operating conditions continuously alter the electromechanical response sensitivity of the piezoelectric sensor units. A fixed sensitivity coefficient cannot correct for dynamically drifting passive response signals, affecting the extraction accuracy of dust impact signals. The broadband frequency-modulated excitation signal applied during the calibration time slot must be completely consistent with the excitation parameters used during baseline construction. Only active response signals under the same excitation conditions can be effectively compared with baseline characteristics. Inconsistent excitation parameters will cause the response parameters to lose their basis for comparison. The response parameters of the active response signal correspond to the baseline characteristics, ensuring the comprehensiveness of subsequent deviation calculations.

[0073] In real-time operation, the resonant frequency of a piezoelectric sensing unit will shift, leading to frequency misalignment during direct comparison. Frequency axis alignment employs cubic spline interpolation, using the baseline frequency axis as a reference grid. The frequency-amplitude and frequency-phase response parameters of the real-time active response signal are mapped onto the baseline frequency axis through cubic spline interpolation, ensuring point-by-point comparison of the two types of data at the same frequency sampling points. The deviation of the active response signal from the reference characteristics is quantified, i.e., the response parameter deviation. This deviation corresponds to the drift amplitude of the piezoelectric sensing unit's sensitivity. For example, if the real-time response amplitude of a piezoelectric sensing unit is 8V, and the corresponding amplitude in the baseline is 7.5V, the amplitude deviation is 0.5V. Similarly, if the real-time impedance real part is 1000Ω, and the corresponding baseline parameter is 980Ω, the impedance real part deviation is 20Ω. The corresponding response parameter deviation of the piezoelectric sensing unit is determined accordingly. A larger deviation value indicates a more significant deviation of the unit's sensitivity from the reference state, intuitively reflecting the offset of physical quantities and conforming to the linear characterization law of piezoelectric sensitivity drift.

[0074] The dynamic weighting coefficients are not randomly assigned, but determined based on the physical mechanisms by which each response parameter is affected by real-time operating conditions. Changes in temperature and humidity alter the dielectric constant and elastic modulus of the piezoelectric material, primarily affecting phase and resonant frequency shifts. When temperature and humidity deviate from standard operating conditions, the weights for phase and resonant frequency shifts are increased. Dust deposition forms an additional mass load on the surface of the piezoelectric sensing unit, directly altering the response amplitude and the real part of the impedance; therefore, the weights are increased when the amount of dust deposition rises. Long-term service of piezoelectric elements causes piezoelectric coefficient decay, primarily affecting impedance. The imaginary part of impedance is the primary component. Therefore, the longer the operating time, the higher the weight of the imaginary part deviation. The core logic of weight allocation is to allow the sensitivity factor to reflect the drift of parameters that are more sensitive to changes in operating conditions. For example, if the real-time operating condition is high temperature and high humidity, the weights are allocated as follows: phase deviation 0.3, resonant frequency offset deviation 0.25, amplitude deviation 0.15, impedance real part deviation 0.15, and impedance imaginary part deviation 0.15. If the amount of dust deposition exceeds the standard, the amplitude deviation weight is adjusted to 0.3 and the impedance real part deviation weight is adjusted to 0.25, and the weights of other parameter deviations are reduced accordingly. The total weight is 1.

[0075] Multi-dimensional parameter deviations and their corresponding weighting coefficients are integrated into a single sensitivity drift index because a single parameter deviation cannot fully characterize the sensitivity change of a piezoelectric sensing unit. Multi-parameter fusion can comprehensively reflect all influencing factors under operating conditions. At the same time, different parameters have different dimensions and numerical ranges. The fused drift value can be converted into a dimensionless relative coefficient by using the min-max normalization method, which maps it to the range of 0 to 1. That is, the smaller the sensitivity factor, the greater the sensitivity drift of the unit, and the greater the correction amplitude needs to be. This coefficient can be used for subsequent passive response signal correction to avoid the failure of correction logic due to inconsistent dimensions.

[0076] The sensitivity factor of the piezoelectric sensor array is generated by weighted averaging of the sensitivity factors of each piezoelectric sensor unit. Based on the non-uniform distribution density, the piezoelectric sensor units in the core area have a greater impact on the monitoring accuracy. The sensitivity drift of these piezoelectric sensor units has a more significant impact on the monitoring results, and they are assigned higher weighting coefficients. For example, the average sensitivity factor of 300 piezoelectric sensor units in the core area is 0.82, the average sensitivity factor of 200 piezoelectric sensor units in the transition area is 0.78, and the average sensitivity factor of 75 piezoelectric sensor units in the edge area is 0.75. By weighting the core area with 0.4, the transition area with 0.35, and the edge area with 0.25, the sensitivity factor of the piezoelectric sensor array is 0.8.

[0077] This enables the sensitivity factor to adaptively match the differences between real-time operating conditions and piezoelectric sensing units, eliminating the defects of fixed coefficient calibration. The accuracy of the sensitivity factor determines the signal correction effect and affects the overall accuracy of subsequent effective impact signal extraction and dust distribution field reconstruction, thus providing a guarantee for quantitative analysis of the monitoring system.

[0078] The processing module is used to correct the passive response signal by means of a sensitivity factor, and to perform frequency domain decomposition on the corrected passive response signal to extract the effective impact signal.

[0079] Specifically, extracting the effective impact signal includes:

[0080] The passive response signal acquired in the monitoring time slot is corrected unit by unit by the sensitivity factor of the piezoelectric sensing array.

[0081] The corrected passive response signal is preprocessed and frequency domain decomposition is performed to filter signal segments with frequency fluctuations within a preset range to obtain the initial effective signal.

[0082] The initial valid signal is verified and the signal is interpolated and completed based on the spatial correlation of the piezoelectric sensor array, and the valid impact signal is output.

[0083] Among them, frequency band decomposition refers to dynamically dividing frequency band intervals corresponding to different dust particle sizes based on the impedance spectrum reference characteristics and the amplitude distribution of the corrected passive response signal, and then performing frequency domain decomposition by dividing the frequency bands through Fourier transform.

[0084] The passive response signal acquired during the monitoring time slot is affected by individual differences in piezoelectric sensing units, real-time operating conditions, and sensitivity drift, resulting in issues such as signal amplitude deviation and phase shift. This leads to deviations in the extraction of effective impact signals and fails to accurately reflect the impact characteristics of dust particles. The passive response signal of each piezoelectric sensing unit is corrected by using its sensitivity factor. The correction logic matches the generation logic of the sensitivity factor to ensure the accuracy of the correction. The passive response signal acquired during the monitoring time slot is the electrical signal generated by the piezoelectric sensing unit after being impacted by dust particles. It mainly includes core parameters such as amplitude and phase. The correction process only adjusts these signal parameters and does not change the time-domain and frequency characteristics of the signal.

[0085] The sensitivity factor of each piezoelectric sensing unit is calculated by combining it with the passive response signal of that unit during the monitoring time slot. Specifically, the corrected passive response signal amplitude equals the original passive response signal amplitude multiplied by the sensitivity compensation coefficient. The specific derivation of the sensitivity compensation coefficient is as follows: Under standard cleanroom conditions, the reference sensitivity factor S0 of a brand-new piezoelectric sensing unit is 1, and the real-time sensitivity factor is S... i (0) i ≤1), sensitivity factor S i ​The sensitivity retention ratio of the i-th piezoelectric sensing unit relative to the reference state, i.e., S i The smaller the value, the more severe the sensitivity attenuation. The amplitude of the original passive response signal is compressed proportionally due to the sensitivity attenuation. Therefore, the sensitivity compensation coefficient C = S0 / S i =1 / S i This is to restore the actual signal intensity of dust particle impact; for example, if the sensitivity factor of a certain piezoelectric sensing unit is 0.85, it indicates that the unit has sensitivity drift, and its corresponding sensitivity compensation coefficient is 1÷0.85≈1.18. If the original passive response signal amplitude obtained by the unit is 2.5V, the signal amplitude after correction is 2.5V×1.18=2.95V, which reflects the actual signal intensity of dust particle impact on the unit.

[0086] During the correction process, ensure that the sensitivity factor of each piezoelectric sensing unit corresponds to the passive response signal it acquires, avoid cross-use of sensitivity factors of different piezoelectric sensing units, and retain the signal data before and after correction for subsequent tracking and verification of abnormal signals. For the overall correction of the piezoelectric sensing array, the corrected signals of all units can be recalibrated by combining the array's sensitivity factor to ensure signal consistency at the array level, adapt to the needs of subsequent spatial signal processing, and achieve targeted compensation for signal deviations.

[0087] This compensates for the sensitivity drift of each piezoelectric sensing unit caused by individual differences and changes in operating conditions, eliminates the amplitude deviation and phase shift of the passive response signal, and ensures that the corrected passive response signal truly and accurately reflects the actual characteristics of dust particle impact; it avoids deviation in the extraction of subsequent effective impact signals due to uncorrected signals, provides a signal basis for subsequent steps, and provides support for system fault diagnosis and signal accuracy verification.

[0088] The corrected passive response signal still includes irrelevant noise signals such as environmental electromagnetic interference, airflow disturbance, and equipment vibration. These noise signals will overlap with the effective signals generated by dust particle impact. At the same time, the frequency characteristics of the passive response signals corresponding to different particle sizes are different. If frequency band decomposition is not performed, the effective signals of dust particles of different sizes will interfere with each other, making it impossible to filter out the real dust impact signals. The corrected passive response signal is preprocessed by spatiotemporal joint denoising to remove irrelevant noise interference. The temporal denoising uses a wavelet denoising algorithm with a db4 wavelet basis and a decomposition level of 5. By performing wavelet decomposition, thresholding, and inverse wavelet transform on the corrected passive response signal, high-frequency noise such as environmental electromagnetic interference and circuit noise is removed. The spatial denoising uses the spatial correlation of the piezoelectric sensing array and the spatial coordinates of each piezoelectric sensing unit to compare the signals of adjacent piezoelectric sensing units at the same time, removing isolated noise signals and retaining the signals that occur synchronously in adjacent units.

[0089] Based on the impedance spectrum reference characteristics and the amplitude distribution of the corrected passive response signal, frequency bands corresponding to different dust particle sizes are dynamically divided. Fourier transform is then used for frequency domain decomposition of these bands. The amplitude distribution of the corrected passive response signal reflects the signal intensity at different frequency points. The combination of these two factors accurately divides frequency bands suitable for different dust particle sizes. When dust particles of different sizes collide with the piezoelectric sensing unit, they generate passive response signals of different frequencies. The larger the particle size, the lower the frequency of the impact signal, and vice versa. This division is based on the dust impact dynamics and the frequency response characteristics of the piezoelectric element.

[0090] Specifically, based on the resonant frequency range (0.1MHz~10MHz) of the impedance spectrum reference characteristics and the peak amplitude distribution of the corrected passive response signal, three frequency bands are divided, each corresponding to a different dust particle size. The low-frequency band is [0.1MHz, 0.3MHz], the mid-frequency band is [0.3MHz, 1MHz], and the high-frequency band is [1MHz, 5MHz]. There is no frequency overlap or ambiguity between the frequency bands. The portion greater than 5MHz is the piezoelectric high-frequency resonant redundancy range, which only contains circuit noise and has no effective dust impact response, and is not included in the frequency band corresponding to the particle size. The high-frequency band corresponds to fine-sized dust (particle size ≤10μm), the mid-frequency band corresponds to medium-sized dust (10μm < particle size ≤50μm), and the low-frequency band corresponds to coarse-sized dust (particle size >50μm). If the amplitude distribution of the corrected passive response signal changes, or the impedance spectrum reference characteristics drift due to adjustments in operating conditions, the boundaries of the frequency bands will be adjusted synchronously to ensure that the correspondence between the frequency bands and the dust particle size remains accurate.

[0091] The preprocessed time-domain passive response signal is converted into a frequency-domain signal to achieve frequency band decomposition. A Fast Fourier Transform (FFT) algorithm is used, with a sampling frequency set to 20MHz and 1024 transform points to ensure the accuracy of the frequency domain decomposition and clearly distinguish signals from different frequency bands. After frequency band decomposition, signal segments with frequency fluctuations within a preset range are selected for each frequency band. This preset range is determined based on the frequency stability of the impedance spectrum reference characteristics. Specifically, under standard cleanroom conditions, the background noise signal of the piezoelectric sensing unit without dust impact is analyzed in the frequency domain. The standard deviation σ of the frequency fluctuation of the background noise signal is statistically analyzed, and 3σ (covering the 99.7% confidence interval) is taken as the boundary for frequency stability judgment. Statistical analysis of multiple sets of experimental data shows that the standard deviation of the frequency fluctuation of the background noise is approximately 1% of the center frequency. Therefore, 3σ corresponds to approximately ±3%. That is, if the frequency fluctuation amplitude of a signal segment does not exceed ±3% of the center frequency of that frequency band, it is determined to be a frequency-stable signal segment. Such signal segments are likely effective signals generated by dust particle impact. If the frequency fluctuation exceeds this range, it is determined to be a stray signal segment and is discarded.

[0092] For example, if the center frequency of a high-frequency band is 3MHz, and the frequency of a certain signal segment fluctuates between 2.91MHz and 3.09MHz with a fluctuation range of ±3%, which is within the preset range, it is determined to be an initial valid signal segment; if the frequency of a certain signal segment fluctuates between 2.8MHz and 3.2MHz with a fluctuation range exceeding ±3%, it is discarded; through the above screening, all signal segments that meet the requirements are integrated to obtain the initial valid signal.

[0093] This eliminates irrelevant noise interference in the corrected passive response signal, preventing noise signals from overlapping with the effective signal; it enables precise separation of dust impact signals of different particle sizes, avoiding mutual interference between signals of different particle sizes, adapting to changes in signal amplitude distribution and operating conditions, and ensuring the accuracy of frequency band division; it eliminates unstable noise signal segments, providing support for subsequent signal verification and completion, and reconstruction of dust mass distribution field.

[0094] Even after screening, the initial valid signals may still contain some isolated stray signals that have not been completely eliminated, or due to the non-uniform layout of the piezoelectric array and local monitoring blind spots, some areas may have missing valid signals. These may not be genuine dust impact signals, or they may result in incomplete signals, affecting the accuracy of subsequent dust mass distribution field reconstruction. By utilizing the spatial coordinates, non-uniform layout, and regional unit division characteristics of the piezoelectric sensing array, the initial valid signals can be verified and supplemented to ensure the authenticity and integrity of the signals. The spatial correlation of the piezoelectric sensing array refers to the fact that signals acquired by adjacent piezoelectric sensing units at the same time will exhibit similar signal characteristics if they are genuine dust impact signals due to the spatial continuity of dust diffusion; however, isolated stray signals will not exhibit such spatial correlation.

[0095] First, based on the divided regional units, the initial valid signals are classified according to the regional units. For each initial valid signal within a regional unit, the spatial coordinates of the corresponding piezoelectric sensing unit are extracted, and the correlation between the signal and the signals of adjacent piezoelectric sensing units is analyzed. Specifically, for a piezoelectric sensing unit corresponding to a certain initial valid signal, the Euclidean distance between the piezoelectric sensing unit and all other piezoelectric sensing units in the same regional unit is calculated based on the calibrated spatial coordinates. The three closest units are selected as adjacent units in ascending order of distance. When there are fewer than three usable piezoelectric sensing units in the same regional unit, the selection is expanded to adjacent regional units to make up the difference. That is, three adjacent piezoelectric sensing units are selected, and the signal characteristics of the unit and its adjacent units at the same time are compared. If at least two adjacent units have initial valid signals with the same signal characteristics, the signal is determined to be a real dust impact valid signal. If there are no corresponding signals in the surrounding adjacent units, or the signal characteristics are significantly different, the signal is determined to be an isolated noise signal and is discarded.

[0096] For example, within a certain area, a piezoelectric sensing unit with coordinates (X=20m, Y=15m, Z=1.1m) acquires an initial valid signal with a frequency of 0.5MHz, which is in the mid-frequency band and corresponds to medium-sized dust particles. The amplitude trend is a slow increase followed by a decrease. Comparing the signals of its three adjacent units, if two of them acquire initial valid signals with frequencies of 0.48~0.52MHz and consistent amplitude trends at the same time, then the signal is determined to be a genuine valid signal. If none of the three adjacent units have a signal with the corresponding frequency, then it is determined to be an isolated noise signal and is discarded.

[0097] For signal missing regions discovered after verification, i.e., dust impact signals exist in a certain area of ​​the cell, but some piezoelectric sensing units do not collect effective signals, a linear interpolation algorithm is used to complete the signal by combining the effective signal characteristics and spatial coordinates of adjacent units in that area. Because the dust distribution in the same area of ​​the cell has spatial continuity and the effective signal characteristics of adjacent piezoelectric sensing units are correlated, the signal parameters of the missing area can be deduced from the signal parameters of adjacent units.

[0098] For example, within a certain area, the piezoelectric sensing unit at coordinates (X=25m, Y=20m, Z=1.1m) did not acquire a valid signal. However, the two adjacent units (X=24m, Y=20m, Z=1.1m) and (X=26m, Y=20m, Z=1.1m) both acquired valid signals in the mid-frequency band, with amplitudes of 3.2V and 3.4V respectively. Through a linear interpolation algorithm, the effective signal amplitude of the missing unit was calculated to be 3.3V, and the frequency was consistent with that of the adjacent units, thus completing the signal completion. The completed signal must be consistent with the effective signal characteristics of the area to avoid signal distortion and ensure the reliability of the completed signal.

[0099] The verified valid signals are integrated with the completed signals, and sorted by region unit and dust particle size to form valid impact signals. The spatial coordinates, frequency and amplitude of each signal corresponding to the piezoelectric sensing unit are output synchronously to ensure that the output valid impact signals are complete and accurate for subsequent dust mass distribution field reconstruction in the evaluation module. All unqualified isolated stray vibration signals are removed to ensure the purity of the output signals, while the identifier of the completed signals is retained for subsequent data verification.

[0100] This method accurately removes isolated stray signals from the initial valid signals, avoiding interference from stray signals in subsequent processing and ensuring that all valid impact signals are genuine dust impact signals. It also solves the problem of signal loss caused by non-uniform deployment of piezoelectric sensor arrays and monitoring blind spots, avoiding deviations in subsequent dust mass distribution field reconstruction due to signal loss. Furthermore, it provides support for the spatial positioning and reconstruction of the dust mass distribution field, realizing a closed loop of valid impact signals from acquisition, correction, screening, and completion, and ensuring the accuracy of dust suppression effect evaluation.

[0101] The evaluation module is used to reconstruct the effective impact signal by performing an inverse problem using a compressed sensing algorithm, so as to reconstruct the dust mass distribution field in the dust suppression area and convert it into an equivalent kinetic energy distribution field. It also compares the equivalent kinetic energy distribution field before and after the dust suppression operation, calculates the dust suppression efficiency, and identifies the dust suppression blind zone.

[0102] Furthermore, such as Figure 2 As shown, the reconstructed dust mass distribution field includes:

[0103] Based on the spatial coordinates and non-uniform density of the piezoelectric sensor array, a compressed sensing observation matrix matching the dust suppression area is constructed.

[0104] The effective impact signal is split according to the dust particle size, and the dust sub-mass distribution field of different dust particle sizes is reconstructed by the compressed sensing observation matrix;

[0105] The dust submass distribution field is weighted and corrected based on the sensitivity factor and impedance spectrum reference characteristics of the piezoelectric sensor array, and the corrected dust submass distribution is fused according to the real-time proportion of dust particle size to generate the dust mass distribution field of the dust suppression area.

[0106] Dust mass distribution field reconstruction involves inferring the spatial distribution of dust in the dust suppression area from effective impact signals. However, piezoelectric sensor arrays employ a non-uniform deployment method, and their spatial coordinates and deployment density determine the spatial coverage characteristics of the signal acquisition. Traditional compressed sensing observation matrices are mostly fixed uniform matrices, which cannot adapt to the non-uniform deployment characteristics of piezoelectric sensor arrays. This leads to signal redundancy in high-density areas and insufficient signal in low-density areas during the reconstruction process, resulting in distortion of the dust mass distribution field reconstruction and an increase in blind spots.

[0107] The spatial location and non-uniform deployment density of the piezoelectric sensing array are integrated into the construction of the compressed sensing observation matrix, enabling the matrix to accurately adapt to the signal acquisition characteristics of the non-uniform array and meet the requirements of the inverse problem reconstruction in compressed sensing. The spatial coordinates of the piezoelectric sensing array are calibrated three-dimensional rectangular coordinates, and the coordinates of each piezoelectric sensing unit correspond to the unit number. The deployment density is the ratio of the number of piezoelectric sensing units in each region to the area of ​​the region, with the highest deployment density in the core area, followed by the transition area, and the lowest in the edge area.

[0108] Based on the spatial gridding of the dust suppression area, and combined with the spatial coordinates and deployment density of the piezoelectric sensing units, the row and column dimensions and element values ​​of the matrix are determined. First, the dust suppression area is divided into uniform spatial grids according to the preset grid size. The grid size is determined according to the array deployment density. The higher the deployment density, the smaller the grid. This ensures that each grid contains at least one piezoelectric sensing unit. For example, the core area has a high deployment density, so the grid size is set to 1m×1m, the transition area is set to 2m×2m, and the edge area is set to 3m×3m. After the grid is divided, a spatial identifier is assigned to each grid, which corresponds to the coordinates of the piezoelectric sensing unit.

[0109] The number of rows in the observation matrix corresponds to the total number of units in the piezoelectric sensor array, and the number of columns corresponds to the total number of spatial grids in the dust suppression area. The matrix elements are determined by the product of the spatial correlation between the piezoelectric sensor unit and the spatial grid, and the weight of the deployment density. For the f-th piezoelectric sensor unit and the n-th spatial grid, the spatial correlation is defined as g(f,n)=exp(-d 2 (f,n) / (2σ g 2 ), where d(f,n) is the Euclidean distance between the center of the f-th piezoelectric sensing unit and the center of the n-th spatial grid, σ g To establish the correlation attenuation constant, it is set to half the average spacing of the piezoelectric sensing units in this region. exp(·) represents an exponential function with the natural constant as its base. When g(f,n) ≥ 0.1, the signal coverage of this unit is determined to include the grid, and the matrix element ϕ(f,n) = ρ × g(f,n), where ρ is the density weight of the unit in this region (ρ = 0.8 for the core area, ρ = 0.5 for the transition area, and ρ = 0.3 for the edge area). When g(f,n) < 0.1, the matrix element ϕ(f,n) = 0. The higher the density, the greater the weight, and vice versa. If the piezoelectric sensing unit and the grid are not related... If there is spatial correlation, the matrix element is 0. For example, the dust suppression area is divided into 1000 spatial grids, and the piezoelectric sensor array has a total of 575 units. A 575×1000 observation matrix is ​​constructed. At the grid corresponding to a piezoelectric sensor unit in the core area and the surrounding grids, the spatial correlation degree g is 1 and the deployment density weight ρ is 0.8, so the matrix element is 0.8. At the grid corresponding to a piezoelectric sensor unit in the transition area, g is 1 and ρ is 0.5, so the matrix element is 0.5. At the grid corresponding to a piezoelectric sensor unit in the edge area, g is 1 and ρ is 0.3, so the matrix element is 0.3. All other uncorrelated matrix elements are 0.

[0110] During matrix construction, the matrix elements are optimized and adjusted based on the dust diffusion gradient of the regional units. The higher the dust diffusion gradient, the more concentrated the weight distribution of the matrix elements, ensuring the reconstruction accuracy of high dust areas. At the same time, the matrix must meet the constraint equidistant condition in compressed sensing theory. Through orthogonalization, the matrix is ​​ensured to accurately capture the spatial distribution characteristics of effective impact signals, avoiding signal aliasing and distortion during reconstruction. The spatial coordinates, deployment density, and spatial characteristics of the dust suppression area of ​​the observation matrix are matched with those of the piezoelectric sensing units, providing observational support for the subsequent reconstruction of the dust submass distribution field by particle size.

[0111] This overcomes the limitation that traditional fixed uniform matrices cannot adapt to non-uniform arrays, enabling the observation matrix to accurately match the spatial characteristics of the dust suppression area and the array layout pattern, solving the signal redundancy in high-density areas and the signal insufficiency in low-density areas; it provides an observation basis for reconstructing the dust mass distribution field of particle sizes, avoids reconstruction distortion caused by mismatched observation matrices, and improves the reliability of the entire reconstruction process.

[0112] The effective impact signal includes signals generated by the impact of dust particles of different sizes. The spatial diffusion characteristics and mass distribution patterns of dust particles of different sizes are significantly different. If the effective impact signals of all particle sizes are mixed and reconstructed, the mass distribution of dust particles of different sizes will interfere with each other, making it impossible to accurately distinguish the spatial distribution state of dust particles of different sizes, thus affecting the accuracy of the overall dust mass distribution field. The effective impact signal is split by combining the frequency domain decomposition results of the frequency bands, and then the sub-mass distribution field of each particle size is reconstructed by the observation matrix through the compressed sensing algorithm. The effective impact signal has been divided into high frequency band, medium frequency band and low frequency band according to the frequency band, which correspond to fine particle size, medium particle size and coarse particle size dust, respectively.

[0113] During the splitting process, ensure that each signal subset contains only dust impact signals of the corresponding particle size, remove the overlapping parts of signals from different frequency bands, and verify the purity of the signal subset by combining the selected frequency-stable signal segments. For example, the high-frequency signal subset corresponds to fine-particle-size dust, and only retain the effective impact signals within the frequency fluctuation range of ±3% in this frequency band, the mid-frequency band corresponds to medium-particle-size dust, and the low-frequency band corresponds to coarse-particle-size dust. After the signal is split, label the particle size range for each signal subset.

[0114] By using the compressed sensing algorithm, the effective impact signal subsets of each particle size after splitting are coupled with the constructed compressed sensing observation matrix to infer the spatial mass distribution of dust particles of each size in the dust suppression area, and obtain the corresponding dust sub-mass distribution field; the orthogonal matching pursuit algorithm is adopted to adapt to the characteristics of the non-uniform observation matrix. The specific iterative process is as follows: (1) Initialize the residual r0=y, where y is the effective impact signal vector, the support set Ω0=∅, and the iteration count t=0; (2) Calculate the residual r t With each column of the observation matrix ϕ jThe inner product is used to select the column index λ with the largest absolute value of the inner product. t+1 =argmax j |⟨r t ,ϕ j >|, will λ t+1 Add support set Ω t+1 =Ω t ∪{λ t+1};(3)Use the least squares method to solve for the current estimated value x on the support set. t+1 =(ϕ_Ω t+1 T ϕ_Ω t+1 ) -1 ϕ_Ω t+1 T y, ϕ_Ω t+1 Ω represents the support set in the observation matrix. t+1 The corresponding set of column vectors; (4) Update the residual r t+1 =y-ϕ_Ω t+1 x t+1 Let t = t + 1; (5) Determine the termination condition: when the residual energy ||r t || 2 ≤ε×‖y‖ 2 (ε is the residual decay threshold, with a value of 1×10) -3 The iteration terminates when the number of iterations t reaches the preset sparsity K (K is 5%~10% of the total number of spatial grids), and the reconstruction result x=x is output. t Otherwise, return to step (2) to continue iterating; after convergence, obtain the mass distribution parameters of the dust particle size in each spatial grid to construct the dust sub-mass distribution field of the particle size.

[0115] For example, after inputting the effective impact signal subset corresponding to fine-sized dust, the mass distribution of fine-sized dust in each spatial grid is inferred by compressing the sensing observation matrix. The concentration of fine-sized dust in the core grid is higher, while the concentration in the edge grid is lower, forming a fine-sized dust sub-mass distribution field. Similarly, the corresponding medium-sized and coarse-sized dust sub-mass distribution fields are obtained. Each sub-mass distribution field includes the spatial distribution information of the corresponding particle size dust in the dust suppression area, which corresponds to the spatial coordinates of the piezoelectric sensing array and the grid division of the observation matrix, ensuring the accurate spatial positioning of the sub-mass distribution field.

[0116] This enables the separation of impact signals from dust particles of different sizes, avoiding the interference caused by the aliasing of signals from different particle sizes, and providing high-quality input data for particle size reconstruction. By making full use of the spatial characteristics of the effective impact signals, the spatial mass distribution of dust particles of different sizes can be inferred, improving the reconstruction accuracy of the sub-mass distribution field and providing support for subsequent analysis of the dust suppression effect of dust particles of different sizes.

[0117] The reconstructed sub-mass distribution field of dust particles by size is still affected by the sensitivity drift of the piezoelectric sensing unit and changes in operating conditions, resulting in distribution deviations. Direct fusion would lead to insufficient accuracy of the overall dust mass distribution field. At the same time, the proportion of dust particles of different sizes in the dust suppression area varies. If equal-weighted fusion is used, it cannot truly reflect the actual distribution ratio of dust particles of different sizes, affecting the accuracy of subsequent equivalent kinetic energy distribution field conversion and dust suppression effect evaluation. Based on the aforementioned generated sensitivity factor and impedance spectrum reference characteristics, the sub-mass distribution field is corrected, and then fusion is completed by combining the real-time proportion of particle size.

[0118] Based on the sensitivity factor and impedance spectrum reference characteristics of the piezoelectric sensing array, the sensitivity factor reflects the real-time response state and sensitivity drift of the piezoelectric sensing unit, while the impedance spectrum reference characteristics reflect the electromechanical response reference of the piezoelectric sensing unit under different operating conditions. The combination of these two characteristics offsets the influence of operating conditions and sensitivity drift on the submass distribution field. Specifically, the sensitivity factor of the piezoelectric sensing unit in each region is calculated with the corresponding submass distribution field parameters. The smaller the sensitivity factor, the greater the sensitivity drift of the unit in that region, and the greater the correction weight for the submass distribution field. Simultaneously, the impedance spectrum reference characteristics are used to... Based on the frequency and amplitude characteristics of the operating conditions, the deviation of the sub-mass distribution field is compensated to ensure that the corrected sub-mass distribution field truly reflects the actual mass distribution of dust. For example, if the sensitivity factor of a certain unit in the core area is 0.85, it indicates that there is sensitivity drift in that unit. The correction weight of the sub-mass distribution field in that area is increased. Combined with the deviation between the standard operating conditions and the real-time operating conditions in the impedance spectrum reference characteristics, the concentration parameters of the sub-mass distribution field are adjusted to offset the deviation caused by the drift. The sensitivity factor in the edge area is relatively stable, and the correction weight is relatively low to ensure that the corrected sub-mass distribution field is consistent with the actual dust distribution.

[0119] The real-time proportion of dust particle size is determined by statistically analyzing the signal intensity of the effective impact signal subsets after splitting. The higher the signal intensity, the greater the quantity of dust particles of the corresponding size, and the higher the proportion. The dust impact signal intensity is positively correlated with the quantity of dust. Specifically, the total signal amplitude of each particle size's effective impact signal subset is statistically analyzed, and the proportion of each particle size's signal amplitude to the total amplitude of all particle size signals is calculated as the real-time proportion of that particle size. For example, the proportion of fine particle size signals is 40%, medium particle size is 45%, and coarse particle size is 15%, which are used as the weighting coefficients for subsequent sub-mass distribution field fusion, with a total weight of 1.

[0120] The corrected dust sub-mass distribution fields of fine, medium, and coarse particle sizes are weighted and fused with the real-time proportion weights of the corresponding particle sizes to obtain the global dust mass distribution field of the dust suppression area. During the fusion process, the dust mass distribution parameter of each spatial grid is the sum of the products of the corresponding grid parameter of each particle size sub-mass distribution field and the corresponding particle size proportion weight, and is a dimensionless value to ensure that the fused distribution field truly reflects the actual distribution ratio of different particle sizes. For example, in a certain spatial grid, the fine particle size sub-mass distribution parameter is 0.4, the medium particle size is 0.45, and the coarse particle size is 0.45. With a diameter of 0.15, corresponding to particle size proportions of 0.4, 0.45, and 0.15, the dust mass distribution parameters of the fused grid are 0.4×0.4+0.45×0.45+0.15×0.15, thus obtaining the dust mass distribution parameters of the grid. After fusion, the distribution parameters of all spatial grids are integrated to generate the dust mass distribution field of the dust suppression area, which includes the dust mass distribution information of each spatial location within the dust suppression area. This information corresponds to the spatial coordinates and regional unit division of the piezoelectric sensing array, ensuring accurate spatial positioning of the distribution field.

[0121] This offsets the effects of changes in operating conditions and piezoelectric sensing unit sensitivity drift on the sub-mass distribution field, closely matches the actual dust distribution state, and improves the accuracy of the sub-particle size distribution field; it avoids the problem that equal-weighted fusion cannot reflect the actual particle size distribution ratio, ensuring that the dust mass distribution field truly and comprehensively reflects the overall dust distribution characteristics of the dust suppression area, and ensuring the logical closed-loop continuity of the entire system's monitoring and evaluation.

[0122] Specifically, the conversion to an equivalent kinetic energy distribution field includes:

[0123] Based on the piezoelectric impact response characteristics of different dust particle sizes in corresponding frequency bands, the dynamic impact velocities of different dust particle sizes are determined in order to calculate the initial sub-kinetic energy distribution of different dust sub-mass distribution fields.

[0124] The initial subkinetic energy distribution is corrected for operating conditions by using the sensitivity factor and impedance spectrum reference characteristics of the piezoelectric sensing array, and spatial gridding is performed by combining the spatial coordinates of the piezoelectric sensing array with the regional cells.

[0125] The equivalent kinetic energy distribution field of the dust suppression area is generated by fusing and correcting the initial sub-kinetic energy distribution based on the real-time proportion of dust particle size.

[0126] The equivalent kinetic energy distribution field reflects the motion characteristics of dust particles through the correlation between dust mass and impact velocity, thereby accurately assessing the dust suppression effect. Dust particles of different sizes have significant differences in physical properties, and the velocity of impacting the piezoelectric sensing unit also shows dynamic changes. If a fixed impact velocity is used to calculate the kinetic energy, the initial sub-kinetic energy distribution will deviate significantly from the actual dust motion state, and cannot truly reflect the impact intensity and kinetic energy distribution of dust particles of different sizes. Based on the aforementioned correspondence between frequency bands and particle size, and the piezoelectric impact response characteristics, the dynamic velocity is determined and the initial sub-kinetic energy is calculated.

[0127] The piezoelectric impact response characteristics of different dust particle sizes in corresponding frequency bands refer to the amplitude, frequency stability, and other characteristics of the effective impact signal in each frequency band after frequency domain decomposition. These characteristics are positively correlated with the dust impact velocity. The greater the dust particle impact velocity, the higher the amplitude of the response signal generated by the piezoelectric sensing unit and the better the frequency stability. By using this correlation and combining the piezoelectric impact response characteristics of each frequency band, impact velocities that change in real time are assigned to dust particles of different sizes to ensure that the velocity matches the actual impact state.

[0128] The amplitude parameters of the piezoelectric impact response characteristics of the effective impact signal in each frequency band were extracted. The dynamic impact velocity of different dust particle sizes was solved using a linear fitting formula: v = v0 + k × (A - A0), where v is the dynamic impact velocity of the dust particle, v0 is the reference impact velocity for the corresponding dust particle size, k is the amplitude velocity proportionality coefficient, A is the real-time signal amplitude, and A0 is the particle size reference calibration amplitude, which was predetermined through dust impact calibration experiments. In a sealed calibration chamber, particles with diameters ≤10 μm and 1 μm were prepared respectively. Dust samples with three standard particle sizes (0~50μm and >50μm) were used. A variable-speed airflow generator controlled the dust particles to impact a calibration piezoelectric sensing unit at different velocities (1~15m / s, step size 0.5m / s). A high-speed camera simultaneously recorded the actual impact velocity *v* of the dust particles and acquired the output signal amplitude *A* of the piezoelectric sensing unit. At least 200 sets of (v, A) data were collected for each particle size. A linear fit was performed using the least squares method: v = v0 + k × (A - A0). The goodness of fit R0 was measured. 2 All are greater than 0.95; among them, fine-particle dust corresponds to a reference impact velocity v0=5m / s, a proportionality coefficient k=2m / (s·V), and a reference calibration amplitude A0=2V; medium-particle dust corresponds to a reference impact velocity v0=3m / s, a proportionality coefficient k=2m / (s·V), and a reference calibration amplitude A0=1V; coarse-particle dust corresponds to a reference impact velocity v0=1.5m / s, a proportionality coefficient k=1m / (s·V), and a reference calibration amplitude A0=1V.

[0129] For example, for fine-particle dust, the higher the amplitude of the high-frequency signal, the greater the impact velocity. When the amplitude of the high-frequency signal is 3V, the dynamic impact velocity is 7m / s; when the amplitude drops to 2.5V, the dynamic impact velocity is adaptively adjusted to 6m / s, achieving dynamic speed adaptation. For medium-particle dust, the dynamic impact velocity is determined to be 5m / s when the amplitude is 2V, and the velocity is adjusted synchronously when the amplitude fluctuates. Coarse-particle dust has a larger mass and a relatively lower impact velocity. When the amplitude of the low-frequency signal is 1.5V, the dynamic impact velocity is determined to be 2m / s, ensuring that the dynamic impact velocity of different particle sizes is accurately matched with the piezoelectric impact response characteristics.

[0130] Based on the theory of kinetic energy calculation, and combining the reconstructed particle size distribution field of dust sub-mass with the dynamic impact velocity determined above, the sub-kinetic energy is calculated, that is, the initial sub-kinetic energy in each spatial grid is E=α×M×v. 2 Where E is the initial sub-kinetic energy of a single-particle-size dust particle within the spatial grid; α is the working condition correction coefficient, which is selected under standard clean working conditions using standard dust particles with known mass and velocity (particle size 30 μm, mass approximately 4.2 × 10⁻⁶). -11 The piezoelectric sensing unit was calibrated by impacting particles of different sizes (1~10m / s) with a mass of kg. The mass distribution parameter M output by the piezoelectric sensing unit was then fitted with the known dust mass. The reciprocal of the fitting slope was α, which was determined to be 0.05 after multiple calibration experiments. M is the dimensionless mass distribution parameter of the corresponding particle size dust in the spatial grid. v is the dynamic impact velocity of the corresponding dust particle size. The calculation process follows the correlation law between kinetic energy, mass, and velocity to ensure that the impact kinetic energy of the corresponding particle size dust in the grid is accurately reflected.

[0131] For example, in the sub-mass distribution field of fine-particle dust, the mass distribution parameter of a certain spatial grid is 0.4, corresponding to a dynamic impact velocity of 8 m / s. The initial sub-kinetic energy of the fine-particle dust in this grid is obtained through calculation. The mass distribution parameter of the corresponding grid for medium-particle dust is 0.45, and the dynamic impact velocity is 5 m / s, so the initial sub-kinetic energy of the medium-particle dust is obtained. The mass distribution parameter of the corresponding grid for coarse-particle dust is 0.15, and the dynamic impact velocity is 3 m / s, so the initial sub-kinetic energy of the coarse-particle dust is calculated. The initial sub-kinetic energy calculations for all spatial grids are completed in sequence to form the initial sub-kinetic energy distribution fields corresponding to fine-particle, medium-particle, and coarse-particle dust.

[0132] This breakthrough overcomes the limitations of traditional fixed-velocity calculation of kinetic energy, enabling the impact velocity to adapt in real time to the actual impact state of dust particles of different sizes, thus avoiding calculation deviations caused by fixed velocity; it achieves precise decomposition of the kinetic energy distribution of dust particles of different sizes, ensuring that the initial sub-kinetic energy distribution reflects the motion characteristics and impact intensity of dust particles of that size; and it ensures the conversion accuracy of the equivalent kinetic energy distribution field, providing support for subsequent analysis of dust suppression effects by particle size.

[0133] The initial sub-kinetic energy distribution is still affected by the sensitivity drift of the piezoelectric sensor array and real-time operating conditions, resulting in deviations in kinetic energy parameters. If it is directly used for fusion, the equivalent kinetic energy distribution field will not be able to truly reflect the actual dust kinetic energy state. The correspondence between the initial sub-kinetic energy distribution and the spatial grid and regional unit division of the dust suppression area is not close enough, lacking a clear spatial positioning association, and cannot meet the spatial analysis requirements of subsequent dust suppression efficiency calculation and blind zone identification. Based on the aforementioned sensitivity factor and impedance spectrum reference characteristics to offset operating condition interference, and based on spatial coordinates and regional unit division, the spatial mapping of the sub-kinetic energy distribution is achieved.

[0134] The initial sub-kinetic energy distribution is corrected based on the sensitivity factor and impedance spectrum reference characteristics of the piezoelectric sensing array. The sensitivity factor reflects the real-time response state and sensitivity drift of the piezoelectric sensing unit, while the impedance spectrum reference characteristics reflect the electromechanical response reference of the piezoelectric sensing unit under different operating conditions. The combination of the two can accurately identify and offset the influence of operating condition changes and sensitivity drift on the initial sub-kinetic energy distribution. Specifically, the sensitivity factor of the piezoelectric sensing unit of each region is calculated with the initial sub-kinetic energy distribution parameters of the corresponding region. The smaller the sensitivity factor, the greater the sensitivity drift of the region, and the greater the correction weight for the initial sub-kinetic energy distribution. The deviation of kinetic energy parameters caused by drift is offset by weight adjustment.

[0135] Simultaneously, the frequency and amplitude characteristics corresponding to the real-time operating conditions in the impedance spectrum reference features are compared with the frequency and amplitude parameters of the initial sub-kinetic energy distribution. The deviation is compensated to ensure that the corrected sub-kinetic energy distribution reflects the actual kinetic energy state of the dust. For example, if the sensitivity factor of a certain unit in the core area is 0.85, it indicates that there is a significant sensitivity drift in that unit. The correction weight of the initial sub-kinetic energy distribution in that area is increased. Combined with the response deviation between the real-time operating conditions and the standard operating conditions in the impedance spectrum reference features, the initial sub-kinetic energy parameters of all spatial grids in that area are adjusted to offset the drift deviation. The sensitivity factor in the edge area is relatively stable, and the correction weight is relatively low to ensure that the corrected sub-kinetic energy distribution is consistent with the actual dust kinetic energy distribution.

[0136] The corrected initial sub-kinetic energy distribution is precisely correlated with the spatial coordinates of the regional unit, the piezoelectric sensing unit, and the spatial grid to achieve spatial positioning standardization of the sub-kinetic energy distribution. Specifically, based on the constructed spatial grid, the corrected initial sub-kinetic energy distributions of fine, medium, and coarse particles are mapped to their respective spatial grids. The sub-kinetic energy parameters of each spatial grid correspond to the spatial coordinates of that grid and the regional unit to which it belongs. At the same time, combined with the spatial coordinates of the piezoelectric sensing array, the kinetic energy parameters corresponding to the effective impact signals acquired by each piezoelectric sensing unit are precisely correlated with the spatial grid and regional unit to which it belongs, ensuring that the spatial positioning of the sub-kinetic energy distribution is consistent with the monitoring position of the piezoelectric sensing array.

[0137] For example, a piezoelectric sensing unit with coordinates (X=20m, Y=15m, Z=1.1m) maps its corresponding effective impact signal, after correction, to the sub-kinetic energy parameters, onto the core area spatial grid (1m×1m) where the unit is located, and labels the core area regional unit to which the grid belongs, thus achieving precise binding between sub-kinetic energy distribution and spatial location and regional unit; during the mapping process, it is ensured that the sub-kinetic energy parameters of each spatial grid are unique and accurate, avoiding spatial misalignment or parameter confusion, and ensuring the spatial coherence of sub-kinetic energy distribution.

[0138] This offsets the impact of changes in operating conditions and piezoelectric sensing unit sensitivity drift on the initial sub-kinetic energy distribution, improving the accuracy and reliability of the sub-kinetic energy distribution; it solves the problem of insufficient spatial correlation of the initial sub-kinetic energy distribution, providing spatially accurate and parameter-reliable input data for the subsequent generation of the equivalent kinetic energy distribution field, and providing a spatial adaptation basis for comparing the kinetic energy distribution before and after subsequent dust suppression operations.

[0139] The corrected and mapped initial sub-kinetic energy distribution is still a kinetic energy distribution field based on particle size, which cannot reflect the overall state of dust kinetic energy in the dust suppression area. The real-time proportion of dust particles of different sizes in the dust suppression area varies. If equal-weighted fusion is used, it cannot reflect the actual kinetic energy contribution ratio of dust particles of different sizes, affecting the accuracy of subsequent dust suppression efficiency calculation and dust suppression blind zone identification. Based on the real-time proportion of dust particle size, the corrected and mapped initial sub-kinetic energy distribution based on particle size is weighted and fused to generate an equivalent kinetic energy distribution field.

[0140] The percentage of the total effective impact signal amplitude corresponding to different particle sizes is the real-time percentage of that particle size. The higher the signal strength, the more dust particles of the corresponding size there are, and the greater their contribution to the overall kinetic energy. The principle is the same as the above. Specifically, the previously determined real-time percentage is used. For example, the real-time percentage of fine-sized dust is 40%, medium-sized dust is 45%, and coarse-sized dust is 15%. This is the weighting coefficient in the fusion process. The total weight is 1, which ensures that the kinetic energy distribution after fusion can truly reflect the actual contribution ratio of dust particles of different sizes.

[0141] Using spatial grids as units, the corrected initial sub-kinetic energy parameters for fine, medium, and coarse grains within each spatial grid are calculated with the corresponding real-time proportion weights of the grain size. The results are then summed to obtain the equivalent kinetic energy parameters for that spatial grid. The fusion logic follows the principle of accumulating contributions by grain size to ensure that the kinetic energy contribution of each grain size is accurately reflected. For example, in a spatial grid, the corrected initial sub-kinetic energy parameters for fine grains are 0.32, for medium grains 0.225, and for coarse grains 0.045, with corresponding real-time proportions of 40%, 45%, and 15%, respectively. Then, the equivalent kinetic energy parameters for that spatial grid are 0.32×40%+0.225×45%+0.045×15%. The equivalent kinetic energy parameters for each spatial grid are obtained through this calculation.

[0142] After fusion, the equivalent kinetic energy parameters of all spatial grids are integrated to generate the equivalent kinetic energy distribution field of the dust suppression area. This distribution field contains the equivalent kinetic energy parameters of each spatial grid within the dust suppression area, which correspond to the spatial coordinates, regional unit division, and spatial grid of the piezoelectric sensing array. It clearly presents the spatial distribution characteristics of dust kinetic energy within the dust suppression area. The equivalent kinetic energy parameters are higher in the core area of ​​high dust, followed by the transition area, and lower in the edge area, which is consistent with the actual dust distribution pattern. Simultaneously, the equivalent kinetic energy parameters, the regional unit to which each spatial grid belongs, and the kinetic energy contribution ratio of the corresponding particle size are labeled to ensure the integrity and traceability of the data.

[0143] This avoids the limitation that equal-weighted fusion cannot reflect the actual kinetic energy contribution of dust particles of different sizes, ensuring that the equivalent kinetic energy distribution field reflects the overall state of dust kinetic energy in the dust suppression area. It retains the detailed characteristics of the kinetic energy distribution by particle size while achieving precise integration of kinetic energy distribution. It solves the problem that the kinetic energy distribution by particle size cannot reflect the overall state, accurately matches the needs of subsequent dust suppression efficiency calculation and dust suppression blind zone identification, provides data support for the efficient implementation of the entire evaluation module, and ensures the accuracy and scientific nature of dust suppression effect evaluation.

[0144] Specifically, calculating dust suppression efficiency and identifying dust suppression blind spots includes:

[0145] Obtain the equivalent kinetic energy distribution field before and after dust suppression operation, calculate the single-dimensional dust suppression efficiency by region unit and dust particle size, and perform working condition correction by combining sensitivity factor and impedance spectrum reference characteristics.

[0146] The dust suppression efficiency of the dust suppression area is obtained by weighting the dust diffusion gradient and the real-time proportion of dust particle size, and then weighting and fusing the single-dimensional dust suppression efficiency.

[0147] Initial dust suppression blind zones are screened based on the dynamic threshold determined by the dust diffusion gradient, and the spatial correlation of kinetic energy in the surrounding area is verified to identify the dust suppression blind zones.

[0148] Dust suppression efficiency is quantified by comparing the changes in dust kinetic energy before and after dust suppression operations. However, there are significant differences in dust concentration and the motion characteristics of dust particles of different sizes in different areas within the dust suppression region. If the overall average of the entire area is used to calculate the dust suppression efficiency, it cannot reflect the differences in local dust suppression effects, nor can it accurately reflect the dust suppression effectiveness of dust particles of different sizes. At the same time, the equivalent kinetic energy distribution field before and after dust suppression operations will still be affected by the sensitivity drift of the piezoelectric sensing unit and the changes in real-time operating conditions, resulting in deviations in kinetic energy parameters. If it is directly used for efficiency calculation, it will lead to distorted efficiency results. Based on the generated equivalent kinetic energy distribution field, combined with the aforementioned regional unit division, dust particle size classification, and calibration parameters, the single-dimensional dust suppression efficiency calculation and correction are completed.

[0149] The equivalent kinetic energy distribution fields are obtained before the start of dust suppression operation (dust stable state) and after the dust suppression operation has lasted for a preset time (dust suppression effect stable state). The acquisition conditions, spatial grid division, and regional unit division of the two distribution fields are completely consistent to ensure the effectiveness of the comparison. The equivalent kinetic energy distribution field corresponding to the initial state of dust is obtained before the operation, and the equivalent kinetic energy distribution field after the dust suppression equipment has been running stably is obtained after the operation. For example, after the dust suppression operation lasts for 30 minutes, the dust state tends to be stable. At this time, the equivalent kinetic energy distribution field after the operation is obtained to ensure that the comparison of the two distribution fields can truly reflect the actual effect of the dust suppression operation.

[0150] The dust suppression efficiency is calculated by dividing the area into units and by dust particle size, specifically η, which represents the dust suppression efficiency for each unit and each type of dust particle size. ab =(E ab _pre-E ab _post) / E ab _pre×100%, where η ab E represents the single-dimensional dust suppression efficiency of dust of particle size b in the a-th region unit. ab _pre represents the average equivalent kinetic energy parameter of the spatial grid corresponding to this particle size in this area before dust suppression operations, E ab _post represents the mean of the equivalent kinetic energy parameters after dust suppression operations; when E ab When _pre=0, the kinetic energy of dust particles of that size in the area is zero before the operation, indicating no dust diffusion, and the dust suppression efficiency is defined as 100%. For example, for fine-sized dust in the core area, the equivalent kinetic energy parameter of a certain spatial grid is 0.4 before the operation and 0.12 after the operation. The single-dimensional dust suppression efficiency of fine-sized particles in this area is calculated to be (0.4-0.12) / 0.4×100%=70%. Similarly, the single-dimensional dust suppression efficiency of particles of different sizes in the core area, coarse particles, transition area, and edge area is calculated to form the dust suppression efficiency of all dimensions.

[0151] By combining sensitivity factor and impedance spectrum reference characteristics, operating condition correction is performed to offset the impact of changes in operating conditions and sensitivity drift before and after dust suppression operations on the single-dimensional dust suppression efficiency. Specifically, for the single-dimensional dust suppression efficiency of each regional unit, the efficiency parameter is adjusted based on the sensitivity factor of the piezoelectric sensing unit of that regional unit. That is, the smaller the sensitivity factor, the greater the sensitivity drift of that regional unit, and the greater the efficiency correction weight. The deviation caused by the drift is offset by adjusting the weight. At the same time, the operating condition differences before and after the operation are combined with the impedance spectrum reference characteristics, such as changes in temperature and humidity, and changes in dust deposition. The parameter deviation of the equivalent kinetic energy distribution field before and after the operation is compared to compensate for the single-dimensional dust suppression efficiency, ensuring that the corrected single-dimensional dust suppression efficiency truly reflects the actual effect of the dust suppression operation, rather than the kinetic energy change caused by operating condition interference. For example, the sensitivity factor of a certain regional unit in the core area is 0.85, and there are significant fluctuations in temperature and humidity before and after the operation. Based on the response deviation of the corresponding operating conditions in the impedance spectrum reference characteristics, the single-dimensional dust suppression efficiency of each particle size of that regional unit is adjusted to offset the influence of operating conditions and sensitivity drift, thus obtaining the corrected single-dimensional dust suppression efficiency.

[0152] This breaks through the limitations of traditional overall average calculation, captures the differences in dust suppression effects in different regions and with different particle sizes, and avoids the local dust suppression effect being masked; it offsets the interference of operating condition changes and sensitivity drift on efficiency calculation, and improves the accuracy and reliability of single-dimensional dust suppression efficiency; it provides input data for subsequent weighted fusion to generate dust suppression efficiency, and provides accurate basis for optimizing dust suppression operations by region and particle size.

[0153] The corrected single-dimensional dust suppression efficiency cannot reflect the dust suppression effect of the dust suppression area. Furthermore, the dust diffusion gradient of the regional unit is different, the weight of its influence on the dust suppression effect is different, the real-time proportion of dust particles is different, and their contribution to the dust suppression effect is different. If equal weight fusion is adopted, the overall dust suppression efficiency will not be able to reflect the actual dust suppression effect, nor will it be able to highlight the dust suppression effect of high dust areas and high proportion of particle sizes. Based on the dust diffusion gradient and the real-time proportion of dust particle size, a reasonable weighting coefficient is assigned to complete the fusion of single-dimensional dust suppression efficiency.

[0154] The weight of the dust diffusion gradient is allocated according to the dust gradient of the divided regional units. The higher the dust diffusion gradient of the regional unit (core area), the greater the impact on the dust suppression effect, and the higher the weight is allocated; the lower the gradient (edge ​​area), the lower the weight, which is in line with the dust suppression monitoring needs, that is, to focus on the dust suppression effect of high dust areas. For example, the dust diffusion gradient of the core area is the highest, and the weight is allocated to 0.5; the gradient of the transition area is medium, and the weight is allocated to 0.3; the gradient of the edge area is the lowest, and the weight is allocated to 0.2, with a total regional weight of 1. The weight of the real-time proportion of dust particle size is allocated according to the real-time proportion of different particle sizes. The higher the real-time proportion of a particle size, the greater its contribution to the overall dust suppression effect, and the higher the weight is allocated, to ensure that the fused overall efficiency can truly reflect the dust suppression contribution of different particle sizes. For example, the real-time proportion of fine particle size is 40%, and the weight is allocated to 0.4; the proportion of medium particle size is 45%, and the weight is allocated to 0.45; the proportion of coarse particle size is 15%, and the weight is allocated to 0.15, with a total particle size weight of 1.

[0155] Each single-dimensional dust suppression efficiency is calculated by combining the dust diffusion gradient weight of the corresponding regional unit and the real-time proportion weight of the corresponding particle size. All calculation results are then summed to obtain the dust suppression efficiency of the dust suppression area. The fusion process follows the principle of dual weighting of regional weight and particle size weight to ensure that the contribution of each single-dimensional dust suppression efficiency is accurately reflected. For example, the single-dimensional dust suppression efficiency of fine particles in the core area is 0.7 after working condition correction, with a corresponding regional weight of 0.5 and a particle size weight of 0.4. The contribution value of this single-dimensional dust suppression efficiency is 0.7 × 0.5 × 0.4. Similarly, the contribution values ​​of all single-dimensional dust suppression efficiencies are calculated sequentially and summed to obtain the dust suppression efficiency. The contribution proportion of each single-dimensional dust suppression efficiency is retained during the fusion process to facilitate subsequent analysis of the impact of different regions and different particle sizes on the overall dust suppression effect, ensuring the traceability and practicality of the dust suppression efficiency.

[0156] This avoids the limitations of equal-weighted fusion in reflecting regional differences and particle size contributions, ensuring that weighted fusion closely matches the actual dust distribution and dust suppression effect, and ensuring that dust suppression efficiency fully reflects the dust suppression effect of the dust suppression area; it also avoids the inability of single-dimensional dust suppression efficiency to reflect the overall state, providing quantitative indicators for the overall evaluation of dust suppression effect, providing a basis for subsequent dust suppression blind spot identification, and facilitating the judgment of the overall effectiveness and weak links of dust suppression operations.

[0157] Dust suppression blind zones are areas where dust suppression is ineffective and dust kinetic energy remains at a high level. If a fixed threshold is used to screen blind zones, it cannot adapt to the differences in dust diffusion gradients in different regional units. The reasonable kinetic energy threshold for high dust core areas should be higher than that for low dust edge areas. Fixed thresholds may lead to misjudgment of blind zones in core areas or omission of blind zones in edge areas. Initially screened blind zones may experience instantaneous fluctuations in operating conditions, leading to misjudgments. By setting dynamic thresholds based on dust diffusion gradients and combining them with the spatial characteristics of the equivalent kinetic energy distribution field, blind zone screening and verification can be completed.

[0158] The higher the dust diffusion gradient, the lower the dust suppression efficiency threshold, meaning the dust suppression effect requirements for high-dust areas are relatively lenient, avoiding misjudgments; conversely, the lower the dust diffusion gradient, the higher the dust suppression efficiency threshold, meaning the dust suppression effect requirements for low-dust areas are stricter, avoiding missed judgments. The kinetic energy threshold is positively correlated with the baseline mean. Under the baseline condition without dust suppression operations, the mean and standard deviation of the equivalent kinetic energy parameters for each regional unit are statistically analyzed. The dust suppression efficiency threshold is taken as the 25th percentile of the historical dust suppression efficiency of that regional unit, meaning 75% of historical monitoring data are better than this value. The kinetic energy threshold is the sum of the mean and standard deviation, covering approximately 84% of the normal operating range. Areas exceeding this kinetic energy threshold are judged as having abnormal dust suppression effects. Taking the core area as an example, under the baseline condition, the equivalent kinetic energy parameters are... The value is 0.15 and the standard deviation is 0.05, so the kinetic energy threshold is 0.15 + 0.05 = 0.2. The 25th percentile of the historical dust suppression efficiency in the core area is about 0.6, so the dust suppression efficiency threshold is 0.6. The transition area and the edge area follow the same principle, so the dust suppression efficiency threshold set for the transition area is 0.7 and the kinetic energy threshold is 0.15; the dust suppression efficiency threshold set for the edge area is 0.8 and the kinetic energy threshold is 0.1. Only when the single-dimensional dust suppression efficiency of a certain area unit is lower than the dust suppression efficiency threshold of the corresponding area, and the equivalent kinetic energy parameter of the area is higher than the kinetic energy threshold of the corresponding area, is it determined to be an initial dust suppression blind zone. If the single-dimensional dust suppression efficiency of a certain area unit in the core area is 0.55 and the kinetic energy parameter is 0.22, it is determined to be an initial dust suppression blind zone.

[0159] Based on the corrected single-dimensional dust suppression efficiency and dynamic threshold, for each region unit and each particle size, the single-dimensional dust suppression efficiency is compared with the dust suppression efficiency threshold of the corresponding region, and the kinetic energy parameter is compared with the kinetic energy threshold of the corresponding region. Region units that simultaneously meet the conditions of dust suppression efficiency lower than the dust suppression efficiency threshold and kinetic energy higher than the kinetic energy threshold are selected as the initial dust suppression blind zone. During the selection process, the selection is classified according to particle size. For example, if a certain region unit with fine particle size in the core area meets the threshold condition, it is determined to be the initial blind zone corresponding to fine particle size. If the particle size and coarse particle size in the region unit also meet the conditions, they are integrated into the initial dust suppression blind zone of the region unit to ensure that the initial dust suppression blind zone corresponds to a specific particle size and region.

[0160] Based on the spatial coordinates of the piezoelectric sensor array, the division of regional units, and the spatial correlation of the equivalent kinetic energy distribution field, the authenticity of the initial dust suppression blind zone is verified, and blind zones misjudged due to instantaneous operating condition fluctuations are eliminated. Because the dust kinetic energy distribution in a true dust suppression blind zone has spatial continuity, the kinetic energy parameters of the surrounding area will also be higher and the dust suppression efficiency will be lower. If the blind zone is misjudged due to instantaneous operating condition fluctuations, the kinetic energy parameters of its surrounding area are normal and the dust suppression efficiency meets the standard, and there is no spatial continuity. For each initial dust suppression blind zone, the equivalent kinetic energy parameters and single-dimensional dust suppression efficiency of its 3-5 adjacent regional units are extracted. If the kinetic energy parameters of at least 2 adjacent regional units are close to those of the initial dust suppression blind zone and the dust suppression efficiency is lower than the dust suppression efficiency threshold of the corresponding area, it indicates that the initial dust suppression blind zone has spatial continuity and is determined to be a true dust suppression blind zone. If the kinetic energy parameters of the surrounding adjacent regional units are normal and the dust suppression efficiency meets the standard, and there is no spatial correlation, it is determined to be a misjudged blind zone and is eliminated.

[0161] For example, if the kinetic energy parameters of three adjacent area units in a certain initial dust suppression blind zone in the core area are all higher than the core area's kinetic energy threshold, and the dust suppression efficiency is lower than the core area's dust suppression efficiency threshold, it indicates that the blind zone has spatial continuity and is determined to be a true dust suppression blind zone. If the kinetic energy parameters of the adjacent area units are normal and the efficiency meets the standard, then the initial dust suppression blind zone is removed. All verified true dust suppression blind zones are integrated, classified and labeled according to area units and dust particle size, and the spatial coordinates, area unit, corresponding particle size, kinetic energy, and efficiency parameters of each blind zone are clearly defined to form a dust suppression blind zone, thus completing the accurate identification of dust suppression blind zones.

[0162] This allows for precise adaptation to the dust characteristics of different regional units, avoiding misjudgments and omissions in blind spots, and improving the accuracy of initial dust suppression blind spot screening; ensuring that the identified dust suppression blind spots are all real inefficient dust suppression areas, thus improving the reliability of blind spot identification; providing spatial positioning and particle size orientation for dust suppression equipment layout optimization and dust suppression parameter adjustment, ensuring that dust suppression operations can solve the dust problem in inefficient areas and improve the overall dust suppression effect; and providing benchmark data for subsequent continuous monitoring and iterative optimization of dust suppression effects, ensuring that the entire environmental dust suppression effect monitoring system forms a closed-loop operation logic of monitoring, evaluation, and optimization.

[0163] Example 2:

[0164] like Figure 3 The diagram shows a flowchart of a method for monitoring environmental dust suppression effects based on a piezoelectric array, as provided in this application embodiment. The method includes:

[0165] A piezoelectric sensor array is non-uniformly deployed in the dust suppression area to respond to passive response signals generated by dust particle impacts and active response signals generated by receiving frequency modulation excitation signals.

[0166] A wideband frequency-modulated excitation signal is applied to the piezoelectric sensing array, and the active response signal is compared with a preset mechanical impedance spectrum baseline to generate a sensitivity factor for the piezoelectric sensing array.

[0167] The passive response signal is corrected by a sensitivity factor, and the corrected passive response signal is then decomposed in the frequency domain to extract the effective impact signal.

[0168] The effective impact signal is reconstructed by performing an inverse problem using a compressed sensing algorithm to rebuild the dust mass distribution field of the dust suppression area and convert it into an equivalent kinetic energy distribution field. The equivalent kinetic energy distribution field before and after the dust suppression operation is compared to calculate the dust suppression efficiency and identify the dust suppression blind zone.

[0169] Since the principle of the method in this application embodiment is similar to that of the system described in this application embodiment, the implementation of the method is the same as that of the system, and the repeated parts will not be described again.

Claims

1. An environmental dust suppression effect monitoring system based on piezoelectric array, characterized in that, include: Sensing module, calibration module, processing module, and evaluation module; The sensing module includes a piezoelectric sensing array that is non-uniformly distributed in the dust suppression area, used to respond to passive response signals generated by dust particle impact and active response signals generated by receiving frequency modulation excitation signals; the dust suppression area is divided into multiple regional units according to the dust diffusion gradient of the dust suppression area, the piezoelectric sensing array is non-uniformly distributed in different regional units, and the spatial coordinates of different piezoelectric sensing units are calibrated. The calibration module is used to apply a frequency-modulated excitation signal to the piezoelectric sensor array, compare the active response signal with a preset electromechanical impedance spectrum baseline, and generate a sensitivity factor for the piezoelectric sensor array. The preset electromechanical impedance spectrum baseline is the electromechanical response data under pre-acquired standard clean operating conditions, extreme temperature and humidity operating conditions, dust deposition operating conditions, and aging operating conditions of piezoelectric elements. The amplitude, phase, resonant frequency offset, real part and imaginary part of impedance are extracted under different operating conditions through wavelet denoising and principal component analysis. The impedance spectrum reference features corresponding to the operating conditions are constructed respectively. Then, all impedance spectrum reference features are weighted and fused to form an electromechanical impedance spectrum baseline adapted to different operating conditions. The sensitivity factor of the generated piezoelectric sensing array includes: A frequency-modulated excitation signal is applied to the piezoelectric sensing array during the calibration time slot to obtain the response parameters of the active response signals of different piezoelectric sensing units. The response parameters are then aligned with the impedance spectrum reference characteristics in the preset mechanical impedance spectrum baseline on the frequency axis, and the deviation of the response parameters is calculated. Based on the real-time temperature and humidity, dust deposition amount and piezoelectric element working time, the weight coefficients of different response parameter deviations are dynamically allocated. After weighted fusion and normalization, the sensitivity factor of the piezoelectric sensing unit is obtained. After weighted averaging, the sensitivity factor of the piezoelectric sensing array is generated. The processing module is used to correct the passive response signal by a sensitivity factor, and to perform frequency domain decomposition on the corrected passive response signal to extract the effective impact signal. The evaluation module is used to reconstruct the effective impact signal using a compressed sensing algorithm, thereby rebuilding the dust mass distribution field in the dust suppression area and converting it into an equivalent kinetic energy distribution field. It then compares the equivalent kinetic energy distribution fields before and after dust suppression operations to calculate the dust suppression efficiency and identify dust suppression blind spots. The conversion into the equivalent kinetic energy distribution field includes: Based on the piezoelectric impact response characteristics of different dust particle sizes in corresponding frequency bands, the dynamic impact velocities of different dust particle sizes are determined in order to calculate the initial sub-kinetic energy distribution of different dust sub-mass distribution fields. The initial subkinetic energy distribution is corrected for operating conditions by using the sensitivity factor and impedance spectrum reference characteristics of the piezoelectric sensing array, and spatial gridding is performed by combining the spatial coordinates of the piezoelectric sensing array with the regional unit. The equivalent kinetic energy distribution field of the dust suppression area is generated by fusing and correcting the initial sub-kinetic energy distribution based on the real-time proportion of dust particle size.

2. The environmental dust suppression effect monitoring system based on piezoelectric array as described in claim 1, characterized in that, The sensing module is specifically used for: The working cycle of the sensing module is divided into independent calibration time slots and monitoring time slots. In the calibration time slot, the frequency modulation excitation signal is received and the corresponding active response signal is generated. In the monitoring time slot, the impact of dust particles is responded to and the corresponding passive response signal is generated. The calibration time slot and the monitoring time slot are executed alternately periodically.

3. The environmental dust suppression effect monitoring system based on piezoelectric array as described in claim 2, characterized in that, The extraction of the effective impact signal includes: The passive response signal acquired in the monitoring time slot is corrected unit by unit by the sensitivity factor of the piezoelectric sensing array. The corrected passive response signal is preprocessed and frequency domain decomposition is performed to filter signal segments with frequency fluctuations within a preset range to obtain the initial effective signal. The initial valid signal is verified and the signal is interpolated and completed based on the spatial correlation of the piezoelectric sensor array, and the valid impact signal is output.

4. The environmental dust suppression effect monitoring system based on piezoelectric array as described in claim 3, characterized in that, The frequency band decomposition refers to dynamically dividing frequency bands corresponding to different dust particle sizes based on the impedance spectrum reference characteristics and the amplitude distribution of the corrected passive response signal, and then performing frequency domain decomposition by dividing the frequency bands using Fourier transform.

5. The environmental dust suppression effect monitoring system based on piezoelectric array as described in claim 4, characterized in that, Reconstructing the dust mass distribution field includes: Based on the spatial coordinates and non-uniform density of the piezoelectric sensor array, a compressed sensing observation matrix matching the dust suppression area is constructed. The effective impact signal is split according to the dust particle size, and the dust sub-mass distribution field of different dust particle sizes is reconstructed by the compressed sensing observation matrix; The dust submass distribution field is weighted and corrected based on the sensitivity factor and impedance spectrum reference characteristics of the piezoelectric sensor array, and the corrected dust submass distribution is fused according to the real-time proportion of dust particle size to generate the dust mass distribution field of the dust suppression area.

6. The environmental dust suppression effect monitoring system based on piezoelectric array as described in claim 5, characterized in that, The calculation of dust suppression efficiency and identification of dust suppression blind spots includes: Obtain the equivalent kinetic energy distribution field before and after dust suppression operation, calculate the single-dimensional dust suppression efficiency by region unit and dust particle size, and perform working condition correction by combining the sensitivity factor and the impedance spectrum reference characteristics. The dust suppression efficiency of the dust suppression area is obtained by weighting the dust diffusion gradient and the real-time proportion of dust particle size, and then weighting and fusing the single-dimensional dust suppression efficiency. The initial dust suppression blind zone is screened based on the dynamic threshold determined by the dust diffusion gradient, and the dust suppression blind zone is identified by verifying the spatial correlation of the kinetic energy of the surrounding area.

7. A method for monitoring environmental dust suppression effects based on piezoelectric arrays, implemented based on the environmental dust suppression effect monitoring system based on piezoelectric arrays as described in any one of claims 1-6, characterized in that, include: A piezoelectric sensor array is non-uniformly deployed in the dust suppression area to respond to passive response signals generated by dust particle impacts and active response signals generated by receiving frequency modulation excitation signals. A frequency-modulated excitation signal is applied to the piezoelectric sensing array, and the active response signal is compared with a preset mechanical impedance spectrum baseline to generate a sensitivity factor for the piezoelectric sensing array. The passive response signal is corrected by a sensitivity factor, and the corrected passive response signal is then decomposed in the frequency domain to extract the effective impact signal. The effective impact signal is reconstructed by performing an inverse problem using a compressed sensing algorithm to rebuild the dust mass distribution field of the dust suppression area and convert it into an equivalent kinetic energy distribution field. The equivalent kinetic energy distribution field before and after the dust suppression operation is compared to calculate the dust suppression efficiency and identify the dust suppression blind zone.

Images