Acoustic emission ore pulp granularity detection system and method based on Hertz collision mode

Abstract

The invention provides an acoustic emission ore pulp granularity online detection system and method based on a Hertz collision mode, and belongs to the technical field of ore dressing process monitoring and control. The system comprises an ore pulp bypass circulation module, a particle size calculation module and a particle size measurement control module. A built-in acoustic emission sensor is used for collecting acoustic emission signals generated by Hertz elastic collision between signal ore pulp particles and a probe; multi-band energy features and time domain features are extracted in combination with wavelet packet decomposition, and a fusion feature vector is constructed by fusing ore pulp flow velocity and density information; the fusion feature vector is input into a BP neural network model after dimension reduction processing, and the proportion of particles with the particle size smaller than a set threshold value in the ore pulp is predicted online. Full-process management of sampling, preparing, measuring and liquid discharging is achieved through automatic control, and the system has the advantages of being high in measuring precision and automation degree, low in cost, high in anti-interference capacity and the like, is suitable for continuous online monitoring of the ore pulp granularity on the ore dressing site and provides data support for flotation process optimization.

Description

Technical Field

[0001] This invention belongs to the field of mineral processing monitoring and control technology, specifically relating to an acoustic emission slurry particle size detection system and method based on Hertzian collision mode. Background Technology

[0002] Particle size distribution is a crucial parameter characterizing the dimensional composition of powder particles. It is typically expressed as the proportion of particles of a specific size or size range within the total material and significantly impacts the physical, chemical, and application properties of materials. In the mineral processing industry, the particle size distribution of slurry directly affects the efficiency and product quality of subsequent processes such as flotation and gravity separation. Proper particle size control can significantly improve metal recovery rates and reduce energy consumption and resource waste. If the slurry particle size is too large, it will lead to insufficient liberation of ore monomers, affecting mineral recovery; if the particle size is too small, it may cause excessive mudding, compromising froth stability and increasing reagent consumption. Therefore, achieving real-time and accurate monitoring of slurry particle size is of great significance for optimizing mineral processing flows and realizing intelligent control.

[0003] Traditional particle size analysis methods include sieving and sedimentation. While these methods are simple to operate, they lack online real-time detection capabilities. To meet the demands of modern industrial automation, online particle size analysis technologies such as mechanical displacement, laser diffraction, and ultrasonic attenuation have been developed. Although these methods have improved in measurement accuracy and response speed, they still suffer from problems such as complex structure, difficult maintenance, high cost, and poor adaptability, limiting their widespread application in high-concentration and complex media such as slurries.

[0004] In recent years, acoustic emission technology has become a new direction in particle size detection research due to its advantages such as non-contact operation, high response, and strong anti-interference ability. Acoustic emission signals originate from elastic waves generated by Hertzian contact collisions between particles in a slurry and the probe. The waveform characteristics are closely related to the particle size. By analyzing the characteristic parameters of the elastic waves in the time and frequency domains, such as peak value, energy spectral density, and spectral distribution, a mapping model between these parameters and particle size can be established. To improve detection accuracy, researchers have proposed various acoustic emission probes based on Hertzian collision theory, piezoelectric thin film detection, and non-invasive installation structures. Simultaneously, machine learning methods are used to extract and model high-dimensional features of complex signals, achieving good results in various particle systems such as coal powder, biomass, and plastic particles. However, in the context of a slurry environment, existing acoustic emission particle size detection systems still face many challenges in high-concentration, highly interfering slurry environments, mainly in the following aspects:

[0005] 1. Low signal-to-noise ratio: The slurry medium itself has strong randomness and heterogeneity. During the slurry transportation or stirring process, the solid particles in the slurry are diverse in type, have a wide particle size distribution, and fluctuate frequently in concentration, resulting in a large amount of background noise and non-target signals superimposed on the acoustic emission signal. This complex physical state makes it difficult to distinguish between effective signals and interference signals, thus significantly reducing the signal-to-noise ratio.

[0006] 2. Difficulty in feature extraction: During the flow of slurry, a large number of mechanical and hydrodynamic disturbances are generated, such as pipe vibration, pump resonance, bubble bursting, and particle collision with pipe wall. These non-target sound sources will continuously interfere with the acoustic emission sensor, masking the useful acoustic emission signals caused by the collision or breakage between particles, further increasing the complexity of feature extraction.

[0007] 3. Insufficient modeling accuracy: The significant inter-particle shielding and energy dissipation effects in high-concentration slurries cause some weak acoustic emission signals to attenuate or even be completely lost during propagation. This not only results in unstable signal strength but also prevents the complete capture of certain key particle size features, affecting subsequent accurate modeling based on acoustic features.

[0008] 4. Weak model generalization ability: Due to the complex noise background and multi-source interference, the effectiveness of the model training data is reduced, making it difficult for the granular prediction model built based on traditional signal processing and machine learning methods to generalize. In practical applications, it exhibits poor robustness and prediction accuracy.

[0009] Traditional methods for extracting energy features using wavelet packet analysis suffer from multicollinearity, a problem that arises when strong linear relationships exist between certain independent variables. The impact is twofold: if redundant variables with strong linear correlations are not removed, the complexity of the neural network increases significantly, and complex neural networks require more training data to maintain measurement accuracy; during neural network training, the similar effects of strongly linearly correlated variables on measurement results can lead to abnormal weights in the trained neurons. Summary of the Invention

[0010] The purpose of this invention is to address the problems of low signal-to-noise ratio, difficulty in feature extraction, insufficient modeling accuracy, and weak model generalization ability in existing acoustic emission particle detection technologies under high-concentration and highly interfering mineral slurry environments. This invention provides an acoustic emission slurry particle size detection system and method based on Hertzian collision modes. To address the multicollinearity problem in energy features extracted by traditional wavelet packet analysis, this invention introduces slurry velocity and density information collected by a Coriolis mass flow meter. Principal component analysis (PCA) is used to reduce the dimensionality of the multidimensional feature vectors and fuse them to construct a slurry particle size feature vector. Considering the complex nonlinear mapping relationship between this feature vector and particle size, a particle size prediction model based on a BP neural network is established to achieve accurate particle size estimation. Based on acoustic emission measurement requirements and field environmental requirements, an online slurry particle size detection system is constructed and tested under actual working conditions in a mineral processing plant, verifying the effectiveness, stability, and engineering application value of the method.

[0011] To achieve the above objectives, the technical solution provided by the present invention is: to provide an online particle size detection system for acoustic emission slurry based on Hertzian collision mode, including a slurry bypass circulation module, a particle size calculation module, and a particle size measurement and control module;

[0012] The slurry bypass circulation module is installed on the slurry gravity flow pipeline between the grinding workshop and the flotation workshop, and includes a slurry sampling unit, a slurry sample circulation unit and an acoustic emission measurement unit.

[0013] The slurry sampling unit is used to collect slurry samples from the slurry gravity flow pipeline near the outlet of the grinding workshop at regular intervals and in quantitative quantities, and transport them to the slurry sample circulation unit.

[0014] The slurry sample circulation unit is used to adjust the slurry sample flow rate, detect the density and mass flow rate of the slurry in real time, and feed back to the particle size calculation module and the particle size measurement and control module.

[0015] The acoustic emission measurement unit includes an industrial control computer, an acoustic emission probe, and an acoustic emission signal acquisition card. The industrial control computer is used to control the acoustic emission signal acquisition card to collect the elastic wave generated by the collision between the slurry and the acoustic emission probe when the slurry flow rate reaches a preset target value, according to the control instructions of the particle size measurement control module, and convert the elastic wave into an acoustic emission signal according to a preset sampling frequency.

[0016] The particle size calculation module is used to fuse the acoustic emission signal with the slurry density and mass flow rate when the slurry flow rate reaches a preset target value to obtain a dimension-reduced fused feature vector; the particle size calculation module is also equipped with a particle size prediction model based on a BP neural network, which is used to estimate the slurry particle size according to the dimension-reduced fused feature vector and output the proportion of particles with a particle size smaller than a set threshold in the slurry.

[0017] The particle size measurement and control module is used to control the sampling of the slurry sampling unit, and to control the slurry flow rate and acoustic emission signal acquisition of the slurry sample circulation unit.

[0018] Furthermore, the slurry sampling unit includes:

[0019] An imported gate valve is connected to the slurry gravity flow pipeline and is located near the outlet of the grinding workshop. It is used to manually control the opening and closing of the slurry sampling pipeline.

[0020] An imported electrically controlled valve is connected to the output pipe of the imported gate valve and is used to adjust the valve opening according to the control command of the particle size measurement control module in order to control the sampling timing and the slurry inlet flow rate.

[0021] The first flushing solenoid valve has its inlet end connected to an external clean water source via a water supply pipeline, and its outlet end connected to the main pipeline between the inlet gate valve and the inlet solenoid valve via a connecting pipeline. The first flushing solenoid valve is used to inject clean water into the main pipeline according to the control command of the particle size measurement and control module in order to unclog and clean the pipeline.

[0022] Furthermore, the slurry sample circulation unit includes a storage tank, a level radar, a pipeline flushing branch, a first circulation branch, a second circulation branch, and a slurry output branch;

[0023] The top of the storage tank has a slurry inlet, which is connected to the output pipe of the inlet electrically controlled valve to store slurry samples;

[0024] The liquid level radar is used to monitor the slurry level in the storage tank in real time.

[0025] The first circulation branch includes a first circulation pipe, and a circulation pump and a pressure transmitter arranged sequentially along the slurry circulation direction on the first circulation pipe; the input port of the first circulation pipe is connected to the slurry outlet at the bottom of the storage tank; the acoustic emission measurement unit is arranged on the first circulation pipe and is located on the output side of the pressure transmitter.

[0026] The circulating pump is used to circulate the slurry at a low flow rate during the slurry sampling stage, and to adjust the slurry flow rate to a preset target value when the slurry sampling volume reaches the set single measurement capacity.

[0027] The second circulation branch includes a second circulation pipe, and a circulation electrically controlled valve and a Coriolis mass flow meter sequentially arranged along the slurry circulation direction on the second circulation pipe; the input port of the second circulation pipe is connected to the output port of the first circulation pipe, and the output port of the second circulation pipe is connected to the slurry circulation inlet located at the upper end of the storage tank; the Coriolis mass flow meter is used to measure the density and mass flow rate of the slurry in real time.

[0028] The slurry output branch includes a third circulation pipe, and an outlet electrically controlled valve and an outlet gate valve arranged sequentially along the slurry flow direction on the third circulation pipe; the inlet of the third circulation pipe is connected to the connection between the first circulation pipe and the second circulation pipe; the outlet of the third circulation pipe is connected to the slurry gravity flow pipe and is close to the inlet of the flotation workshop.

[0029] The pipeline flushing branch includes a second flushing solenoid valve, whose inlet end is connected to an external clean water source through a water supply pipe, and whose outlet end is connected to the main pipeline between the circulating pump and the pressure transmitter through a connecting pipe.

[0030] Furthermore, the acoustic emission probe includes a housing, a mounting base, a support rod, a probe head, a sensor, and a cover plate assembly;

[0031] The housing is in the shape of a rotating body and has a central through hole for mounting the probe and allowing slurry to pass through the circulation pipe; its side wall is provided with a mounting area that matches the mounting base;

[0032] The mounting base is fixed in the mounting area and has a mounting through hole that communicates with the central through hole of the housing;

[0033] The probe head is disposed in the central through hole of the housing along the axial direction of the housing, and has an inner cavity for accommodating the sensor; and the head of the probe head has a spherical structure to reduce the impact wear of slurry particles on the probe head head and the intensity of eddy currents caused by obstruction of slurry fluid.

[0034] The body of the support rod is coaxially disposed in the mounting through hole of the mounting base and is interference-fitted with the mounting through hole; one end of the support rod extends out of the mounting through hole, and the other end is vertically fixedly connected to the side wall of the probe head, and the central through hole of the support rod communicates with the inner cavity of the probe head.

[0035] The sensor is fixedly installed in the inner cavity of the probe head, and its signal line is led out through the central through hole of the support rod;

[0036] The cover plate assembly is used to close the opening end of the inner cavity after the sensor is installed in the inner cavity of the probe head, and the outer surface of the cover plate assembly is flush with the outer wall surface of the probe head.

[0037] Furthermore, the particle size calculation module and the particle size measurement and control module are integrated in the control cabinet;

[0038] The control cabinet is equipped with a human-machine interaction module, which is used to receive user operation commands and dynamically display slurry sampling data and particle size estimation results.

[0039] Furthermore, the human-computer interaction module is also configured to: in response to a user-triggered cyclic measurement command, send a periodic sampling signal to the granularity calculation module and synchronously update the displayed data.

[0040] Furthermore, the human-computer interaction module includes:

[0041] The touch screen is used to receive user-triggered cyclic measurement commands and dynamically display the slurry sample volume, slurry flow rate, slurry density and mass flow rate, as well as the particle size estimation results of the slurry particles.

[0042] The physical button area includes a measurement start / stop button.

[0043] This invention also provides an online particle size detection method for slurry based on the above-mentioned detection system, comprising the following steps:

[0044] Step 1: Based on PLC control, online acquisition of elastic waves generated by the collision of mineral slurry particles with the acoustic emission probe during the flow process is carried out, and the elastic waves are converted into acoustic emission signals;

[0045] Step 2: Extract the time-domain features from the acoustic emission signal and construct an original feature vector using these features; fuse the original feature vector with the detected slurry flow velocity and density to obtain a multi-dimensional fused feature vector; and perform dimensionality reduction processing on the multi-dimensional fused feature vector to obtain a dimensionality-reduced fused feature vector. :

[0046]

[0047]

[0048] In the formula, , These are the energy and root mean square value of the acoustic emission signal, respectively. The number of sub-band signals obtained from the decomposition. to These are the energy proportions of sub-band signals in each sub-frequency band after the acoustic emission signal is decomposed using wavelet packet decomposition; and These are the slurry density and slurry flow rate measured by the Coriolis mass flow meter, respectively.

[0049] Step 3: Use a particle size prediction model based on BP neural network to estimate the particle size of the slurry online and output the proportion of particles with a diameter smaller than a set threshold in the slurry sample.

[0050] Step 4: After testing, discharge the slurry sample into the gravity flow pipe near the entrance of the flotation workshop.

[0051] Furthermore, step 1 includes the following process:

[0052] Step 1.1: Slurry Sample Infeed Control Stage:

[0053] Open the inlet gate and inlet electric control valve of the slurry sampling unit, as well as the circulation control valve of the slurry sample circulation unit, so that the slurry enters the storage tank of the slurry sample circulation unit from the slurry gravity flow pipe; during the liquid inlet process, monitor the liquid level in the storage tank in real time and feed it back to the particle size measurement control module, and at the same time adjust the slurry flow rate to make it circulate at a low flow rate.

[0054] When the liquid level in the storage tank reaches the preset sampling capacity, close the inlet gate and the inlet electric control valve;

[0055] Step 1.2: Preparation stage for adjusting slurry flow rate:

[0056] Adjust the slurry flow rate to a preset target value and maintain the slurry in circulation at the preset target value; during the adjustment process, monitor the density and mass flow rate of the slurry in real time and feed them back to the particle size measurement and control module;

[0057] Step 1.3: Acoustic emission signal acquisition stage:

[0058] When the density and mass flow rate of the slurry remain stable, the particle size measurement control module outputs a measurement trigger command to the industrial control computer of the acoustic emission measurement unit. The industrial control computer controls the acoustic emission probe acquisition card to collect the elastic wave generated by the Hertz collision between the slurry and the acoustic emission probe, and converts the elastic wave into an acoustic emission signal according to a preset sampling frequency and feeds it back to the particle size calculation module.

[0059] Furthermore, step 2 includes the following process:

[0060] Step 2.1: Calculate the root mean square value of the acoustic emission signal according to the formula. and total energy The time-domain characteristics of acoustic emission signals:

[0061]

[0062]

[0063] In the formula, For acoustic emission signals; The duration of the acoustic emission period;

[0064] Step 2.2: Use wavelet packet decomposition to decompose the acoustic emission signal Decomposed into Layer frequency bands, and extract the energy of each sub-band signal:

[0065]

[0066] in, This represents the number of wavelet packet decomposition levels. For the first Layer Each one carries a signal; The first wavelet packet decomposition Layer The energy value of the signal in each sub-carrier; This represents the number of sampling points for the sub-band signal.

[0067] Step 2.3: After normalizing the energy of the sub-band signals in all sub-bands, construct the original feature vector. :

[0068] Z 0 = [ e 1 , e 2 , … e l ] = [ E j ， 1 / E , E j , 2 / E , … , E j , l / E ]

[0069] In the formula, The total energy of the acoustic emission signal; They represent the first The first in the layer The energy that carries signals The number of sub-band signals obtained from the decomposition; to This represents the energy percentage of each subband signal after decomposition.

[0070] Step 2.4: Convert the original feature vector Total energy of acoustic emission signal Root mean square value and the detected slurry flow rate With pulp density Data fusion is performed to obtain a multi-dimensional fused feature vector:

[0071]

[0072] in: and These are the total energy and root mean square value of the acoustic emission signal, respectively; and These are the slurry density and slurry flow rate measured by the Coriolis mass flow meter, respectively.

[0073] Step 2.5: Principal component analysis is used to reduce the dimension of the multidimensional fused feature vector to obtain the dimension-reduced fused feature vector. .

[0074] The advantages of this invention are:

[0075] 1. This invention relates to an online particle size detection system for slurry based on Hertzian collision mode acoustic emission, comprising a slurry bypass circulation module, a particle size calculation module, and a particle size measurement and control module. The slurry bypass circulation module incorporates a Coriolis mass flow meter to collect slurry flow velocity and density information, ensuring the accuracy of acoustic emission signal acquisition. Based on and considering the complex nonlinear mapping relationship between the acoustic emission signal energy feature vector and particle size, a BP neural network-based particle size prediction model is established to achieve fully automated online particle size detection. This detection system is installed in a bypass manner on the gravity-flow slurry pipeline between the concentrator and flotation workshop, without affecting the normal slurry production process. It has advantages such as high automation, low cost, and strong anti-interference capability, and is suitable for continuous online monitoring of slurry particle size in concentrators, providing data support for flotation process optimization.

[0076] 2. The acoustic emission probe designed in this invention has the advantages of compact structure and low invasiveness of installation. The sensor is installed inside the probe, which reduces signal attenuation and composition changes during signal propagation. Furthermore, the front end of the probe adopts a spherical structure design, which not only helps to reduce impact wear, but also effectively reduces the intensity of eddy currents caused by fluid obstruction, ensuring the accuracy of acoustic emission signal acquisition.

[0077] 3. The detection method of this invention fully considers the comprehensive influence of particle size, flow velocity and density on the dynamic characteristics of acoustic emission signals (such as energy distribution and spectral structure). By fusing the obtained original energy characteristics with the slurry flow meter mass flow rate data, and performing dimensionality reduction processing on the obtained multidimensional fused characteristics, the multicollinearity problem of energy characteristics extracted by wavelet packet analysis in traditional methods is avoided.

[0078] 4. To verify the detection effect of the system of the present invention, the applicant built the online particle size detection system of the slurry of the present invention under the actual working conditions of the mineral processing plant, and carried out sampling and testing based on the actual acoustic emission measurement requirements and the on-site environment. The test results show that the system has the best measurement accuracy when the slurry flow rate is 23 m³ / h, and the maximum relative error is controlled within 8%, which verifies the effectiveness, stability and engineering application value of the method of the present invention.

[0079] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0080] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0081] Figure 1 This is a block diagram of the acoustic emission slurry particle size online detection system of the present invention;

[0082] Figure 2 This is a schematic diagram of the slurry circulation flow direction in this invention;

[0083] Figure 3 This is a structural diagram of the slurry bypass circulation module in this invention;

[0084] Figure 4 This is a schematic diagram of the acoustic emission probe structure in this invention;

[0085] Figure 5 This is a schematic diagram of the physical main body of the acoustic emission slurry particle size online detection system of the present invention;

[0086] Figure 6 This is a schematic diagram illustrating the principle of particle size measurement using acoustic emission phenomena.

[0087] Figure 7 This is a schematic diagram of the working cycle of the particle size detection system of the present invention.

[0088] In the diagram: 1-Slurry gravity flow pipe, 2-Inlet gate valve, 3-Inlet solenoid valve, 4-Level radar, 5-Flushing solenoid valve, 6-Circulation pump, 7-Storage tank, 8-Pressure transmitter, 9-Coriolis mass flow meter, 10-Long straight pipe, 11-Acoustic emission probe, 1101-Housing, 1102-Mounting base, 1103-Support rod, 1104-Detector head, 1105-Bottom cover, 1106-Sensor, 1107- Screws, 1108-End Cap, 1109- Countersunk screws, 1110- Screw, 12-Circulation solenoid valve, 13-Outlet solenoid valve, 14-Outlet gate valve, 15-Slurry particles. Detailed Implementation

[0089] The embodiments of the present invention are described in detail below. These embodiments are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0090] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless explicitly specified.

[0091] Reference Figures 1-3This invention proposes an online particle size detection system for ore slurry based on Hertzian collision mode acoustic emission, comprising a slurry bypass circulation module, a particle size calculation module, and a particle size measurement and control module. The slurry bypass circulation module is installed on the gravity-flow slurry pipeline 1 between the grinding workshop and the flotation workshop, and includes a slurry sampling unit, a slurry sample circulation unit, and an acoustic emission measurement unit. The particle size calculation module and the particle size measurement and control module are integrated in a control cabinet.

[0092] The slurry sampling unit includes an inlet gate valve 2, an inlet electrically controlled valve 3, and a flushing solenoid valve 5. The inlet gate valve 2 and the inlet electrically controlled valve 3 form a flushing valve group, responsible for periodically and quantitatively collecting slurry samples from the slurry gravity flow pipeline 1. These samples are then used by the slurry sample circulation unit and the acoustic emission measurement unit for slurry acoustic emission signal measurement and analysis. Specifically: the inlet gate valve 2 is connected to the slurry gravity flow pipeline and is located near the grinding workshop outlet, used for manually controlling the opening and closing of the slurry sampling pipeline. The inlet electrically controlled valve 3 is installed on the pipeline at the output end of the inlet gate valve, used to adjust the valve opening according to the control commands of the particle size measurement control module to control the sampling timing and slurry inflow rate. The inlet end of the flushing solenoid valve 5 is connected to an external clean water source through a water supply pipeline, and the outlet end is connected to the main pipeline between the inlet gate valve 2 and the inlet electrically controlled valve 3 through a connecting pipeline. The flushing solenoid valve 5 is used to inject clean water into the main pipeline according to the control commands of the particle size measurement control module to unclog and clean the pipeline.

[0093] Because the slurry flow rate in the concentrator's pipeline is low, it cannot generate a sufficiently strong acoustic emission signal, thus affecting the particle size measurement effect. To solve this problem, this invention designs a slurry sample circulation unit to increase the local slurry flow rate by adjusting the slurry flow path, thereby meeting the basic requirements of acoustic emission measurement for signal amplitude. Simultaneously, the slurry sample circulation unit integrates a Coriolis mass flow meter for real-time measurement of the slurry density and mass flow rate in the circulation pipeline, and uses the measurement data for slurry particle size calculation and slurry circulation measurement control. Specifically, the slurry sample circulation unit includes a storage tank 7, a level radar 4, a pipeline flushing branch, a first circulation branch, a second circulation branch, and a slurry output branch. Because the sampled slurry flow rate is too small to be directly circulated in the circulation bypass, a storage tank is needed to increase the total amount of slurry in a single measurement process. Specifically:

[0094] The top inlet of the storage tank 7 is connected to the output pipe of the inlet solenoid valve 3 for storing slurry samples. A level radar 4 is installed at the top of the storage tank to monitor the slurry level in real time, controlled by the particle size measurement control module. The first circulation branch includes a first circulation pipe, and a circulation pump 6 and a pressure transmitter 8 sequentially arranged along the slurry circulation direction. The input port of the first circulation pipe is connected to the slurry outlet at the bottom of the storage tank. The acoustic emission measurement unit is located on the first circulation pipe, on the output side of the pressure transmitter 8. In addition to providing the target flow rate required for acoustic emission measurement of the slurry, the circulation pump 6 also ensures low-speed circulation of the slurry during the sampling inlet stage, preventing pipe blockage caused by slurry sedimentation from causing abnormal measurements by the Coriolis mass flow meter. The pressure transmitter 8 is used to detect the pressure in the circulation pipe during the slurry circulation process in real time and feeds it back to the particle size measurement control module.

[0095] The second circulation branch includes a second circulation pipe, and a circulation solenoid valve 12 and a Coriolis mass flow meter 9 sequentially arranged along the slurry circulation direction on the second circulation pipe. The input port of the second circulation pipe is the output port of the first circulation pipe, and the output port of the second circulation pipe is connected to the slurry circulation inlet located at the upper end of the storage tank. The Coriolis mass flow meter 9 is used to measure the density and mass flow rate of the slurry in real time. It can not only calculate the flow velocity of the slurry in the pipe for particle size measurement, but also calculate the particle mass content for adjusting grinding process parameters. The circulation solenoid valve 12 is used to control the bypass circulation of the slurry sample until the slurry flow rate stabilizes to the point where the acoustic emission signal meets the acquisition requirements.

[0096] The slurry output branch includes a third circulation pipe, and an outlet electrically controlled valve 13 and an outlet gate valve 14 sequentially arranged along the slurry flow direction on the third circulation pipe. These are used to control the discharge of the measured slurry sample from the circulation pipe so that it flows into the slurry gravity flow pipe 1. The inlet of the third circulation pipe connects to the junction of the first and second circulation pipes, and the outlet of the third circulation pipe connects to the slurry gravity flow pipe 1 and is located near the inlet of the flotation workshop. The pipeline flushing branch includes another flushing solenoid valve, whose inlet is connected to an external clean water source via a water supply pipe, and whose outlet is connected to a long straight pipe 10 between the circulation pump 6 and the pressure transmitter 8 via a connecting pipe, for unblocking and cleaning the circulation pipe.

[0097] Based on the above-designed slurry bypass circulation module, one path of slurry from the grinding workshop directly enters the flotation workshop to ensure normal production operation; the other path enters the storage tank, which provides sufficient slurry samples to ensure the normal operation of the bypass circulation, thereby enabling online measurement of slurry particle size.

[0098] The acoustic emission measurement unit includes an industrial control computer, an acoustic emission probe 11, and an acoustic emission signal acquisition card. The acoustic emission probe and the acoustic emission signal acquisition card convert the elastic waves generated when the slurry collides with the probe into electrical signals, and then convert them into digital signals according to a preset sampling frequency, feeding them back to the particle size calculation module for particle size calculation. The industrial control computer, based on the control instructions from the particle size measurement control module, controls the operation of the acoustic emission probe and the acoustic emission signal acquisition card when the slurry flow rate reaches the preset target value.

[0099] Reference Figure 4 The acoustic emission probe designed in this invention is installed on the slurry circulation pipeline of the slurry sample circulation unit, and includes a housing 1101, a mounting base 1102, a support rod 1103, a probe head 1104, a bottom cover 1105, and a sensor 1106. Screw 1107, End Cap 1108 Countersunk screw 1109 and Screw 1110. The housing 1101 is a rotating body, with one end flange sealingly connected to the circulation pipe on the output side of the pressure transmitter 8, and the other end flange sealingly connected to the circulation pipe connected to the circulation solenoid valve 12; the housing sidewall has a mounting groove that matches the mounting base 1102. The housing 1101 has a large wall thickness, providing greater installation rigidity for the probe head 1104, reducing external interference while avoiding the influence of installation stress on the acoustic emission signal propagation process. The mounting base 1102 is connected via... Screw 1110 is fixed in the mounting groove on the outer wall of housing 1101. The mounting base has a radially arranged through hole that communicates with the central through hole of the housing, allowing the support rod 1103 to pass through. The mounting base 1102 and the support rod 1103 are interference-fitted. The support rod 1103 has a central through hole, and its lower end is vertically welded to the probe head 1104, communicating with the inner cavity of the probe head 1104. This ensures the structural integrity of the support rod and the probe head and provides space for cable installation. The support rod passes through multiple... The countersunk screw 1109 is axially fixed to the probe head 1104. The sensor 1106 is connected via... Screw 1107 is installed in the inner cavity of probe head 1104, and the gap between it and the mounting surface of probe head is filled with acoustic emission coupling agent to eliminate the influence of gaps and air bubbles on signal reception and ensure the integrity of acoustic emission signal acquisition. The signal line of sensor 1106 is led out through the central through hole of support rod. Probe head 1104 is axially disposed in the central through hole of housing; the front end of probe head is designed with a spherical structure, which not only helps to reduce the impact wear of slurry particles on the end of probe head, but also effectively reduces the intensity of eddy currents caused by fluid obstruction. In addition, to facilitate sensor installation, the bottom of the inner cavity of probe head has a first opening, the size of which is larger than the size of sensor, and the tail end of inner cavity (i.e., the tail end of probe head) has a second opening along the axial direction to ensure the integrity of probe head shape. In this embodiment, bottom cover 1105 and end cover 1108 are designed to seal with the first and second openings respectively, and the surface of bottom cover is flush with the outer wall surface of the corresponding position of probe head. In this embodiment, the sensor is installed inside the probe head cavity, which reduces signal attenuation and composition changes during signal propagation and ensures the accuracy of acoustic signal acquisition.

[0100] The particle size calculation module designed in this invention performs wavelet packet decomposition on acoustic emission signals and extracts multi-band energy features. These energy features are then fused with the slurry density and mass flow rate when the slurry flow rate reaches a preset target value to obtain a dimension-reduced fused feature vector. The particle size calculation module also includes a BP neural network-based particle size prediction model, used to estimate the slurry particle size based on the obtained dimension-reduced fused feature vector and output the proportion of particles with a diameter smaller than a set threshold in the slurry. The construction process of the BP neural network-based particle size prediction model is as follows: first, the BP neural network model architecture is constructed, and model parameters are set; then, acoustic emission data of standard quartz sand samples with known particle sizes are collected as a training set to train the constructed model, optimizing the network structure and parameter configuration to improve prediction accuracy; finally, the trained BP neural network is deployed in the particle size calculation module to achieve online particle size estimation of real-time acquired data.

[0101] The particle size measurement control module is responsible for coordinating and controlling the stable operation of the entire measurement process. Specifically, it includes controlling the sampling of the slurry sampling unit, regulating the slurry flow rate of the slurry sample circulation unit, and controlling the acquisition of acoustic emission signals. It also has fault detection and handling functions to ensure system reliability and measurement accuracy. The particle size calculation module and the particle size measurement control module are integrated in the control cabinet. The control cabinet has a human-machine interface module, including a touch screen and a physical button area. The touch screen receives user operation commands and dynamically displays the slurry sample volume, slurry flow rate, slurry density and mass flow rate, and the particle size estimation results. The physical button area has a measurement start / stop button to control the start and stop of the detection system.

[0102] Reference Figure 5 The physical body of the acoustic emission slurry particle size online detection system of the present invention has been applied in actual mineral processing sites, realizing continuous online monitoring of slurry particle size in mineral processing sites, and providing data support for subsequent flotation process optimization.

[0103] Reference Figure 6 and Figure 7 Based on the above-designed online particle size detection system for slurry, and fully considering the comprehensive influence of particle size, flow velocity, and density on the dynamic characteristics of acoustic emission signals (such as energy distribution and spectral structure), this invention provides a method for detecting the particle size of acoustic emission slurry based on Hertzian collision modes, comprising the following steps:

[0104] Step 1: Based on PLC control, online acquisition of elastic waves generated by the collision of slurry particles with the acoustic emission probe during the flow process is performed, and the elastic waves are converted into acoustic emission signals. Specifically, this includes:

[0105] (1) Liquid inlet stage

[0106] The liquid inlet stage aims to complete slurry sampling and provide sufficient slurry samples for particle size measurement. Before measurement begins, inlet gate valve 2 and inlet solenoid valve 3 must be opened to prepare the detection equipment. Under the control of the particle size measurement control module (PLC), inlet solenoid valve 3 reaches its maximum opening, and the slurry flows from slurry gravity pipe 1 through the inlet gate valve and inlet solenoid valve into storage tank 7. Radar level gauge 4 monitors the liquid level in the storage tank in real time and transmits the data to the PLC for monitoring the operating status and controlling the measurement process.

[0107] During field testing, it takes approximately 10 minutes for the slurry in storage tank 7 to be injected from an empty tank to the required measurement level. During this process, particles in the slurry can settle at the bottom of the storage tank, clogging the inlet of circulation pump 6. To avoid the impact of clogging on the measurement process during the slurry injection phase, circulation pump 6 operates at 30% of its rated power. This not only helps prevent clogging but also reduces wear on the equipment caused by the high-speed flow of the slurry. When the slurry level in the storage tank reaches the required height (at which point the volume of slurry in the tank must be greater than the minimum volume required for circulation, and there must be no air bubbles in the slurry), the inlet gate and inlet solenoid valve are closed, ending the slurry injection phase.

[0108] Step 1.2: Preparation stage for adjusting slurry flow rate.

[0109] During acoustic emission measurements, the flow field within the circulating pipeline needs to be kept stable to reduce the influence of eddies, bubbles, and other contaminants on the measurement results. Therefore, after the liquid inlet stage, the PLC controls the inlet solenoid valve to completely close and increases the power of the circulating pump, raising the slurry flow rate in the circulating pipeline to the required flow rate for acoustic emission measurements and maintaining it thereafter.

[0110] The Coriolis mass flow meter monitors the density and flow rate of the slurry in the circulating pipeline in real time and sends this data to the PLC. The preparation stage is completed when the Coriolis mass flow meter 9 detects that the slurry density tends to stabilize and the slurry flow rate reaches the flow rate required for acoustic emission measurement.

[0111] Step 1.3: Acoustic emission signal acquisition stage.

[0112] After the preparation phase is completed and the slurry flow field in the circulating pipeline reaches stability, the PLC sends a measurement trigger command to the industrial control computer of the acoustic emission measurement unit. Upon receiving the command, the industrial control computer controls the acoustic emission signal acquisition card to sample the elastic waves generated by the collision between slurry particles and the acoustic emission probe, and converts the elastic waves into acoustic emission signals according to the preset sampling frequency, and feeds them back to the particle size calculation module.

[0113] like Figure 6 As shown, slurry particles 15 impact the surface of the probe head 1103 of the acoustic emission probe during flow, generating elastic waves, which are converted into electrical signals by the acoustic emission sensor 1102. The formation process can be represented as follows:

[0114] (1)

[0115] in, Here is the acoustic emission source function; Let be the propagation function of the elastic wave; For sensor functions; This represents convolution.

[0116] After signal acquisition is complete, the industrial control computer writes the result into the designated PLC register. Once the PLC detects a change in the register value, it displays the result on the touch screen.

[0117] Step 2: Acoustic emission signal feature extraction and fusion processing.

[0118] Sub-step 2.1: Calculate the root mean square value of the acoustic emission signal obtained in step 1. and total energy As a time-domain feature, the calculation formula is as follows:

[0119] (2)

[0120] (3)

[0121] In the formula, For acoustic emission signals; The duration of the acoustic emission cycle.

[0122] Sub-step 2.2: Use wavelet packet decomposition to decompose the acoustic emission signal Decomposed into The frequency bands are layered, and the energy of each sub-band signal in all sub-bands is extracted. The calculation formula is as follows:

[0123] (4)

[0124] in, The first wavelet packet decomposition Layer The size carries a signal. For the first Layer The energy value of the individual signal This represents the number of sampling points for the sub-band signal. The number of decomposition layers needs to be determined based on the actual signal energy distribution. In this embodiment of the invention, to balance feature extraction effectiveness and computational complexity, the number of decomposition layers is... Take 9.

[0125] Sub-step 2.3: After normalizing the sub-band signal energy of all sub-frequency bands, construct the original feature vector. , represented as:

[0126] Z 0 = [ e 1 , e 2 , … e l ] = [ E j ， 1 / E , E j , 2 / E , … , E j , l / E ] (5)

[0127] In the formula, The total energy of the acoustic emission signal is calculated using equation (3); wavelet packet decomposition uses a bisection method to divide sub-frequency bands, therefore the number of decomposition layers is... When the value is 9, a total of 512 sub-band signals are generated, that is The value is 512. This part is common knowledge and will not be elaborated here. Therefore, the energies of the signals in the 1st to 512th sub-bands of the 9th layer are respectively... .

[0128] Sub-step 2.4: Due to the original feature vector Besides the particle size of the slurry, the slurry flow rate and density also have a significant impact on the parameters. Therefore, the original feature vector is... and total energy Root mean square value and slurry flow rate collected from Coriolis mass flow meter With pulp density Perform data fusion to construct a multi-dimensional fusion feature vector. :

[0129] (6)

[0130] in, and These are the total energy and root mean square value of the acoustic emission signal, respectively; to The energy percentage of each sub-band after wavelet 9-level decomposition; and These are the slurry density and slurry flow rate measured by the Coriolis mass flow meter, respectively.

[0131] Sub-step 2.5: The dimensions of the multidimensional fusion feature vector obtained in step 2.4 include energy, root mean square value, 512 energy distribution values, slurry density, and slurry flow rate, totaling 516.

[0132] Excessive dimensionality can lead to multicollinearity issues in the energy features extracted using traditional wavelet packet analysis (multicollinearity refers to a strong linear relationship between certain independent variables). If redundant variables with strong linear correlations are not removed, the complexity of the neural network increases significantly, and complex neural networks require more training data to maintain measurement accuracy. Furthermore, during neural network training, the similar impact of linearly correlated variables on measurement results can cause abnormal weights in the trained neurons. Therefore, this embodiment of the invention incorporates slurry velocity and density information collected by a Coriolis mass flow meter and uses Principal Component Analysis (PCA) to simplify the dimensionality of the high-dimensional feature vector obtained in step 2.4, achieving feature fusion. When using PCA to process the features, 516 principal components need to be calculated, each corresponding to a feature vector. Typically, principal components with eigenvalues ​​greater than 1 are selected, and the contribution of all eigenvalues ​​is calculated to ensure a cumulative contribution greater than 90% for optimal subsequent feature fusion.

[0133] Dimensionally reduced and fused feature vectors obtained through dimensionality reduction simplification Represented as:

[0134] (7)

[0135] Step 3: Perform online granularity estimation using a granularity prediction model based on a BP neural network.

[0136] The granularity calculation module calculates the fused feature vector. The input is fed into a BP neural network particle size prediction model. Based on this model, the particle size prediction of the slurry sample is estimated, and the percentage of particles smaller than 200 mesh in the slurry is output. After the measurement is completed, the particle size calculation module writes the estimation result into a designated PLC register. When the PLC detects a change in the register value, it displays the calculation result on the touch screen and controls the particle size detection system to enter the drainage stage.

[0137] Step 4: Control of slurry sample drainage.

[0138] After measurement, the slurry sample in storage tank 7 needs to be emptied and the next measurement cycle begins. The PLC closes the circulation control valve 12 while opening the outlet control valve 13 and the outlet gate valve 14. The slurry sample is circulated and discharged by the circulation pump, flowing into the slurry gravity flow pipe 1 near the flotation workshop inlet. To empty the slurry in the storage tank and pipes as much as possible, the PLC controls the circulation pump to operate at maximum power. After the radar level gauge 4 detects that the storage tank is empty, the PLC closes the outlet control valve 13 and opens the inlet gate valve 2, inlet control valve 3, and circulation control valve 12, entering the next measurement cycle.

[0139] In summary, the acoustic emission slurry particle size online detection system and method proposed in this invention, based on Hertzian collision mode, relies on the elastic wave signal generated by the collision between particles and the probe. Combined with the characteristics of the acoustic emission radio frequency domain and the slurry flow rate and density information, it achieves a four-stage process of automatic liquid feeding, speed adjustment preparation, measurement, and liquid discharge through a constructed slurry bypass circulation module, particle size calculation module, and particle size measurement control module, thus realizing stable circulation of the slurry in the measurement loop. Furthermore, it extracts multi-frequency energy features from the acoustic emission signal using wavelet packet decomposition and fuses the slurry flow rate and density information to obtain a dimension-reduced fused feature vector. Finally, based on a BP neural network, it achieves online particle size estimation and output without affecting the normal slurry production process, possessing high precision, low cost, and good industrial application value.

[0140] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the scope of the technology disclosed in the present invention, and such modifications or substitutions should all be covered within the scope of protection of the present invention.

Claims

1. An online particle size detection system for slurry based on Hertzian collision mode, characterized in that, It includes a slurry bypass circulation module, a particle size calculation module, and a particle size measurement and control module; The slurry bypass circulation module is installed on the slurry gravity flow pipeline between the grinding workshop and the flotation workshop, and includes a slurry sampling unit, a slurry sample circulation unit and an acoustic emission measurement unit. The slurry sampling unit is used to collect slurry samples from the slurry gravity flow pipeline near the outlet of the grinding workshop at regular intervals and in quantitative quantities, and transport them to the slurry sample circulation unit. The slurry sample circulation unit is used to adjust the slurry sample flow rate, detect the density and mass flow rate of the slurry in real time, and feed back to the particle size calculation module and the particle size measurement and control module. The acoustic emission measurement unit includes an industrial control computer, an acoustic emission probe, and an acoustic emission signal acquisition card. The industrial control computer is used to control the acoustic emission signal acquisition card to collect the elastic wave generated by the collision between the slurry and the acoustic emission probe when the slurry flow rate reaches a preset target value, according to the control instructions of the particle size measurement control module, and convert the elastic wave into an acoustic emission signal according to a preset sampling frequency. The particle size calculation module is used to fuse the acoustic emission signal with the slurry density and mass flow rate when the slurry flow rate reaches a preset target value to obtain a dimension-reduced fused feature vector; the particle size calculation module is also equipped with a particle size prediction model based on a BP neural network, which is used to estimate the slurry particle size according to the dimension-reduced fused feature vector and output the proportion of particles with a particle size smaller than a set threshold in the slurry. The particle size measurement and control module is used to control the sampling of the slurry sampling unit, and to control the slurry flow rate and acoustic emission signal acquisition of the slurry sample circulation unit.

2. The detection system according to claim 1, characterized in that, The slurry sampling unit includes: An imported gate valve is connected to the slurry gravity flow pipeline and is located near the outlet of the grinding workshop. It is used to manually control the opening and closing of the slurry sampling pipeline. An imported electrically controlled valve is connected to the output pipe of the imported gate valve and is used to adjust the valve opening according to the control command of the particle size measurement control module in order to control the sampling timing and the slurry inlet flow rate. The first flushing solenoid valve has its inlet end connected to an external clean water source via a water supply pipeline, and its outlet end connected to the main pipeline between the inlet gate valve and the inlet solenoid valve via a connecting pipeline. The first flushing solenoid valve is used to inject clean water into the main pipeline according to the control command of the particle size measurement and control module in order to unclog and clean the pipeline.

3. The detection system according to claim 2, characterized in that, The slurry sample circulation unit includes a storage tank, a level radar, a pipeline flushing branch, a first circulation branch, a second circulation branch, and a slurry output branch. The top of the storage tank has a slurry inlet, which is connected to the output pipe of the inlet electrically controlled valve to store slurry samples; The liquid level radar is used to monitor the slurry level in the storage tank in real time. The first circulation branch includes a first circulation pipe, and a circulation pump and a pressure transmitter arranged sequentially along the slurry circulation direction on the first circulation pipe; the input port of the first circulation pipe is connected to the slurry outlet at the bottom of the storage tank; the acoustic emission measurement unit is arranged on the first circulation pipe and is located on the output side of the pressure transmitter. The circulating pump is used to circulate the slurry at a low flow rate during the slurry sampling stage, and to adjust the slurry flow rate to a preset target value when the slurry sampling volume reaches the set single measurement capacity. The second circulation branch includes a second circulation pipe, and a circulation electrically controlled valve and a Coriolis mass flow meter sequentially arranged along the slurry circulation direction on the second circulation pipe; the input port of the second circulation pipe is connected to the output port of the first circulation pipe, and the output port of the second circulation pipe is connected to the slurry circulation inlet located at the upper end of the storage tank; the Coriolis mass flow meter is used to measure the density and mass flow rate of the slurry in real time. The slurry output branch includes a third circulation pipe, and an outlet electrically controlled valve and an outlet gate valve arranged sequentially along the slurry flow direction on the third circulation pipe; the inlet of the third circulation pipe is connected to the connection between the first circulation pipe and the second circulation pipe; the outlet of the third circulation pipe is connected to the slurry gravity flow pipe and is close to the inlet of the flotation workshop. The pipeline flushing branch includes a second flushing solenoid valve, whose inlet end is connected to an external clean water source through a water supply pipe, and whose outlet end is connected to the main pipeline between the circulating pump and the pressure transmitter through a connecting pipe.

4. The detection system according to claim 1 or 2, characterized in that, The acoustic emission probe includes a housing, a mounting base, a support rod, a probe head, a sensor, and a cover plate assembly; The housing is in the shape of a rotating body and has a central through hole for mounting the probe and allowing slurry to pass through the circulation pipe; its side wall is provided with a mounting area that matches the mounting base; The mounting base is fixed in the mounting area and has a mounting through hole that communicates with the central through hole of the housing; The probe head is disposed in the central through hole of the housing along the axial direction of the housing, and has an inner cavity for accommodating the sensor; and the head of the probe head has a spherical structure to reduce the impact wear of slurry particles on the probe head head and the intensity of eddy currents caused by obstruction of slurry fluid. The body of the support rod is coaxially disposed in the mounting through hole of the mounting base and is interference-fitted with the mounting through hole; one end of the support rod extends out of the mounting through hole, and the other end is vertically fixedly connected to the side wall of the probe head, and the central through hole of the support rod communicates with the inner cavity of the probe head. The sensor is fixedly installed in the inner cavity of the probe head, and its signal line is led out through the central through hole of the support rod; The cover plate assembly is used to close the opening end of the inner cavity after the sensor is installed in the inner cavity of the probe head, and the outer surface of the cover plate assembly is flush with the outer wall surface of the probe head.

5. The detection system according to claim 1 or 2, characterized in that, The particle size calculation module and the particle size measurement and control module are integrated in the control cabinet; The control cabinet is equipped with a human-machine interaction module, which is used to receive user operation commands and dynamically display slurry sampling data and particle size estimation results.

6. The detection system according to claim 5, characterized in that, The human-computer interaction module is also configured to: in response to a user-triggered cyclic measurement command, send a periodic sampling signal to the granularity calculation module and synchronously update the displayed data.

7. The detection system according to claim 6, characterized in that, The human-computer interaction module includes: The touch screen is used to receive user-triggered cyclic measurement commands and dynamically display the slurry sample volume, slurry flow rate, slurry density and mass flow rate, as well as the particle size estimation results of the slurry particles. The physical button area includes a measurement start / stop button.

8. A method for online particle size detection of slurry based on the detection system according to any one of claims 1-7, characterized in that, Includes the following steps: Step 1: Based on PLC control, online acquisition of elastic waves generated by the collision of mineral slurry particles with the acoustic emission probe during the flow process is carried out, and the elastic waves are converted into acoustic emission signals; Step 2: Extract the time-domain features from the acoustic emission signal and construct an original feature vector using these features; fuse the original feature vector with the detected slurry flow velocity and density to obtain a multi-dimensional fused feature vector; and perform dimensionality reduction processing on the multi-dimensional fused feature vector to obtain a dimensionality-reduced fused feature vector. : In the formula, , These are the energy and root mean square value of the acoustic emission signal, respectively. The number of sub-band signals obtained from the decomposition. to These are the energy proportions of sub-band signals in each sub-frequency band after the acoustic emission signal is decomposed using wavelet packet decomposition; and These are the slurry density and slurry flow rate measured by the Coriolis mass flow meter, respectively. Step 3: Use a particle size prediction model based on BP neural network to estimate the particle size of the slurry online and output the proportion of particles with a diameter smaller than a set threshold in the slurry sample. Step 4: After testing, discharge the slurry sample into the gravity flow pipe near the entrance of the flotation workshop.

9. The online detection method according to claim 8, characterized in that, Step 1 includes the following process: Step 1.1: Slurry Sample Infeed Control Stage: Open the inlet gate and inlet solenoid valve of the slurry sampling unit, as well as the circulation control valve of the slurry sample circulation unit, so that the slurry enters the storage tank of the slurry sample circulation unit from the slurry gravity flow pipe; during the liquid inlet process, monitor the liquid level in the storage tank in real time and feed it back to the particle size measurement control module, and at the same time adjust the slurry flow rate to make it circulate at a low flow rate. When the liquid level in the storage tank reaches the preset sampling capacity, close the inlet gate and the inlet electric control valve; Step 1.2: Preparation stage for adjusting slurry flow rate: Adjust the slurry flow rate to a preset target value and maintain the slurry in circulation at the preset target value; during the adjustment process, monitor the density and mass flow rate of the slurry in real time and feed them back to the particle size measurement and control module; Step 1.3: Acoustic emission signal acquisition stage: When the density and mass flow rate of the slurry remain stable, the particle size measurement control module outputs a measurement trigger command to the industrial control computer of the acoustic emission measurement unit. The industrial control computer controls the acoustic emission probe acquisition card to collect the elastic wave generated by the Hertz collision between the slurry and the acoustic emission probe, and converts the elastic wave into an acoustic emission signal according to a preset sampling frequency and feeds it back to the particle size calculation module.

10. The online detection method according to claim 8, characterized in that, Step 2 includes the following process: Step 2.1: Calculate the root mean square value of the acoustic emission signal according to the formula. and total energy The time-domain characteristics of acoustic emission signals: In the formula, For acoustic emission signals; The duration of the acoustic emission period; Step 2.2: Use wavelet packet decomposition to decompose the acoustic emission signal Decomposed into Layer frequency bands, and extract the energy of each sub-band signal: in, This represents the number of wavelet packet decomposition levels. For the first Layer Each one carries a signal; The first wavelet packet decomposition Layer The energy value of the signal in each sub-carrier; This represents the number of sampling points for the sub-band signal. Step 2.3: After normalizing the energy of the sub-band signals in all sub-bands, construct the original feature vector. ： In the formula, The total energy of the acoustic emission signal; They represent the first The first in the layer The energy that carries signals The number of sub-band signals obtained from the decomposition; to This represents the energy percentage of each subband signal after decomposition. Step 2.4: Convert the original feature vector Total energy of acoustic emission signal Root mean square value and the detected slurry flow rate With pulp density Data fusion is performed to obtain a multi-dimensional fused feature vector: in: and These are the total energy and root mean square value of the acoustic emission signal, respectively; and These are the slurry density and slurry flow rate measured by the Coriolis mass flow meter, respectively. Step 2.5: Principal component analysis is used to reduce the dimension of the multidimensional fused feature vector to obtain the dimension-reduced fused feature vector. .

Images