Ultrasonic dispersion power adjustment method and system of laser particle size analyzer based on particle size feedback

Abstract

The present application relates to the technical field of laser particle size detection, and discloses a laser particle size instrument ultrasonic dispersion power adjusting method based on particle size feedback, which comprises the following steps: obtaining initial particle size distribution data and smoothing to obtain a first particle size characteristic sequence; extracting the median particle size change rate, and determining agglomeration and generating a power increase instruction if the change rate exceeds a preset threshold; obtaining an initial driving current and performing stepwise increment to obtain a first target driving current; sending the current to the ultrasonic transducer through a digital signal feedback channel and obtaining a feedback second particle size characteristic sequence, and using a decision tree algorithm to classify the volume average diameter to determine the dispersion state category; if it is an over-crushing state, a power reduction instruction is generated, a second target driving current is obtained by decrement calculation, and dynamic adjustment is completed by sending. The method can solve the problem that the dispersion power cannot be self-adaptively adjusted due to the lack of real-time particle size characteristic feedback channels between the measurement end of the laser particle size instrument and the ultrasonic dispersion end.

Description

Technical Field

[0001] This invention relates to the field of laser particle size detection technology, and in particular to a method and system for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback. Background Technology

[0002] Laser particle size analyzers are core analytical tools used in industries such as materials research and development, pharmaceuticals, chemicals, and coatings to determine particle quality and guide process parameters. As a typical type of optical intelligent sensor, the accuracy and repeatability of the measurement results directly affect the reliability of product quality control.

[0003] In actual measurement processes, ultrasonic dispersion is a crucial preliminary step that determines whether a sample can present a true particle distribution. Only when agglomerates are effectively broken up without being excessively fragmented can the measured particle size distribution be of reference value. Currently, most laser particle size analyzers use a control method independent of the measurement process in the ultrasonic dispersion stage. Operators set the ultrasonic power level based on experience or a fixed process table. The key characteristic parameters of particle distribution, such as median diameter and volume average diameter, acquired by the measurement unit remain only in numerical form on the display interface. There is no signal feedback channel to the ultrasonic control mainboard. The dispersion control system is unaware of the actual particle state of the sample, creating a situation where the measurement process and the dispersion process are isolated. However, when sample types are frequently changed or there are particle size differences between batches of powder, the fixed dispersion power cannot be adjusted to keep up with the real-time changes in particle state. If the power is too low, agglomerated particles remain, resulting in a measured particle size that is too large. If the power is too high, the particles are broken up by ultrasound, resulting in a measured particle size that is too small. The results of two measurements of the same batch of samples differ significantly, making repeatability difficult to guarantee.

[0004] In summary, the existing technology lacks a real-time feedback channel based on particle size characteristics between the measuring end and the ultrasonic dispersing end of the laser particle size analyzer, resulting in the dispersion power being unable to automatically adjust according to changes in particle state. Summary of the Invention

[0005] This invention provides a method and system for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback, in order to solve the problem that the dispersion power cannot be automatically adjusted according to changes in particle state due to the lack of a real-time feedback channel based on particle size characteristics between the measurement end and the ultrasonic dispersion end of the laser particle size analyzer.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback, comprising: The initial particle size distribution data output by the measurement unit is acquired, and the initial particle size distribution data is smoothed to obtain the first particle size feature sequence. The median diameter of the particles at adjacent time points is extracted based on the first particle size feature sequence, and the difference between the median diameter values ​​is calculated to obtain the particle size change rate. If the particle size change rate is greater than a preset first rate threshold, the current particle state is determined to be a particle agglomeration state, and a first power increase command is generated. Obtain the current initial drive current value of the distributed unit, and increment the initial drive current value according to the first power increase instruction to obtain the first target drive current; The first target driving current is sent to the preset ultrasonic transducer through the digital signal feedback channel, and the second particle size feature sequence fed back by the preset ultrasonic transducer under the first target driving current is obtained. The volume average diameter in the second particle size feature sequence is classified by the decision tree algorithm to determine the particle dispersion state category. If the particle dispersion state is classified as excessively broken, a first power reduction command is generated. Based on the first power reduction command, the first target driving current is calculated to decrease, a second target driving current is determined, and the second target driving current is sent to the preset ultrasonic transducer through the digital signal feedback channel to complete the dynamic adjustment of the dispersion power.

[0007] Secondly, the present invention provides a laser particle size analyzer ultrasonic dispersion power adjustment system based on particle size feedback, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the laser particle size analyzer ultrasonic dispersion power adjustment method based on particle size feedback as described above.

[0008] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback.

[0009] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention extracts the median diameter values ​​of adjacent time nodes and calculates the difference to obtain the particle size change rate. When the change rate is greater than the preset first rate threshold, it is determined that the particles are in a state of aggregation and a first power increase command is generated. At the same time, the adjusted first target driving current is sent to the ultrasonic transducer through the digital signal feedback channel. This realizes the real-time closed-loop adjustment of the dispersed power with the particle size characteristic change rate as the feedback signal. It solves the problem that the dispersed power cannot be adaptively adjusted due to the lack of a feedback channel between the measurement end and the dispersion end in the prior art, and significantly improves the accuracy and repeatability of particle size measurement.

[0010] (2) The present invention classifies the volume average diameter in the second particle size feature sequence by using a decision tree algorithm, and determines the particle dispersion state category by combining information gain and node splitting processing. When it is determined to be an over-fragmented state, the first power reduction instruction is automatically generated and the first target driving current is reduced. This realizes bidirectional adaptive control of dispersion power, avoids the problem of primary particle over-fragmentation caused by excessive power, and ensures that the particles are always in the best dispersion state.

[0011] (3) This invention smooths the initial particle size distribution data, combines the Kalman filter algorithm to suppress measurement noise, and obtains the driving voltage and initial driving current of the dispersion unit to calculate the current impedance value to correct the increment, so that the power adjustment process is not affected by the change in material viscosity, thereby enhancing the robustness and stability of the dispersion power adjustment. Attached Figure Description

[0012] Figure 1 This is a schematic flowchart illustrating a laser particle size analyzer ultrasonic dispersion power adjustment method based on particle size feedback, provided in an embodiment of the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] Reference Figure 1 The first embodiment of the present invention provides a method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback, comprising the following steps: S11, acquire the initial particle size distribution data output by the measurement unit, smooth the initial particle size distribution data, and obtain the first particle size feature sequence; S12, extract the median diameter values ​​of adjacent time nodes according to the first particle size feature sequence, and calculate the difference between the median diameter values ​​to obtain the particle size change rate. If the particle size change rate is greater than a preset first rate threshold, determine that the current particle state is a particle agglomeration state and generate a first power increase command. S13, obtain the current initial drive current value of the distributed unit, and perform incremental calculation on the initial drive current value according to the first power increase instruction to obtain the first target drive current; S14, the first target driving current is sent to the preset ultrasonic transducer through the digital signal feedback channel, and the second particle size feature sequence fed back by the preset ultrasonic transducer under the first target driving current is obtained. The volume average diameter in the second particle size feature sequence is classified by the decision tree algorithm to determine the particle dispersion state category. S15, if the particle dispersion state is excessively broken, a first power reduction command is generated, the first target driving current is reduced according to the first power reduction command, a second target driving current is determined, and the second target driving current is sent to the preset ultrasonic transducer through the digital signal feedback channel to complete the dynamic adjustment of dispersion power.

[0015] In step S11, the initial particle size distribution data output by the measurement unit is obtained, and the initial particle size distribution data is smoothed to obtain the first particle size feature sequence.

[0016] In one implementation, smoothing the initial particle size distribution data to obtain a first particle size feature sequence includes: Based on the initial particle size distribution data, the volume percentage data of the preset particle size range is extracted to obtain the local particle size distribution sequence; The local particle size distribution sequence is smoothed using the Kalman filter algorithm to obtain the first particle size feature sequence.

[0017] It should be noted that the measurement unit is a laser particle size analyzer, which acquires the scattered light energy signal of powder sample particles in real time through the principle of light scattering. The initial particle size distribution data is output through Mie scattering inversion, and this data includes volume percentage distribution information across the entire range from submicron to millimeter. Due to mechanical vibration and dust interference in industrial settings, the data from a single measurement exhibits random fluctuations. To extract more stable particle size characteristics, the initial particle size distribution data from 10 consecutive measurements are truncated according to a time window, extracting the volume percentage data for particle sizes in the 50μm to 100μm range to obtain a local particle size distribution sequence. The preset particle size range of 50μm to 100μm is chosen because this range covers the critical range where the particle size of most industrial powders (such as nano-oxides and pigment slurries) undergoes significant changes during ultrasonic dispersion. Filtering out global data outside this range effectively suppresses high-frequency noise in irrelevant intervals across the entire range, focusing on the particle size segment sensitive to dispersion state, thus improving the targeting and computational efficiency of subsequent data processing.

[0018] The local particle size distribution sequence is smoothed using a Kalman filter algorithm to obtain the first particle size feature sequence. Specifically, the volume percentage data at each time point in the local particle size distribution sequence is used as the observation input for the Kalman filter algorithm. The diagonal elements of the process noise covariance matrix Q are set to 0.01, and the diagonal elements of the measurement noise covariance matrix R are set to 0.1. This parameter configuration is based on the calibration of the noise characteristics of the measurement unit under standard particle samples through pre-experiments, and the determination of the measurement variance by statistical calculation and maximum likelihood estimation under 50 repeated measurement conditions. The Kalman filter algorithm, through alternating prediction and update steps, uses the optimal state estimate of the previous time step and the observation value of the current time step for weighted fusion to smooth out signal spikes caused by sensor transient interference and circuit noise, and outputs a continuous and smooth first particle size feature sequence. For example, the local particle size distribution volume percentage sequence [8.2, 8.1, 8.3, 8.0, 8.5, 7.9, 8.2, 8.1, 8.0, 8.3] at 10 consecutive time points is input, and after Kalman filtering, a smooth sequence [8.15, 8.13, 8.18, 8.12, 8.21, 8.10, 8.16, 8.14, 8.11, 8.19] is output. The sequence variance is reduced from 0.028 before processing to 0.0012 after processing, and the smoothness and continuity are significantly improved.

[0019] In step S12, the median diameter values ​​of adjacent time nodes are extracted based on the first particle size feature sequence, and the difference between the median diameter values ​​is calculated to obtain the particle size change rate, including: Based on the first particle size feature sequence, the median diameter value of the particles at each time point is extracted; According to the chronological order, the median diameter values ​​of adjacent time nodes are combined into median diameter value pairs. Calculate the difference between the median diameter values ​​of the particles to obtain the median diameter difference; If the median diameter difference satisfies the preset continuity condition, the particle size change rate is calculated based on the time interval between the median diameter difference and the adjacent time nodes.

[0020] It should be noted that the first particle size feature sequence contains the cumulative distribution information of powder particles that changes continuously over time. The cumulative particle size distribution curve for each time point in the sequence is plotted with particle size on the x-axis and the percentage of particles smaller than that size on the y-axis. For each time point in the sequence, the x-axis value corresponding to 50% of the cumulative volume distribution is located on the cumulative distribution curve, and this x-axis value is taken as the median particle size value for that time point. The median particle size value is a core statistical indicator characterizing the overall fineness of powder particles; its physical meaning is that 50% of the particles in the sample have a particle size smaller than this value. For example, at the first time T1, the particle size coordinate corresponding to 50% of the cumulative volume is located on the cumulative distribution curve as 65 μm; therefore, the median particle size value at that time is extracted as 65 μm.

[0021] According to the chronological order, the median particle diameter values ​​at adjacent time points are combined into median particle diameter value pairs. Specifically, the median particle diameter value of 65 μm corresponding to the first time point T1 and the median particle diameter value of 68 μm corresponding to the immediately following second time point T2 are combined in chronological order to form a pair of median particle diameter values, denoted as (65 μm, 68 μm). The median particle diameter value of 68 μm at the second time point T2 is also extracted from the cumulative distribution curve at the second time point.

[0022] The difference between the median diameter values ​​of the particle size pairs is calculated to obtain the median diameter difference. For example, the median diameter difference is calculated by subtracting the first time value from the second time value in the median diameter pair (65μm, 68μm), resulting in a median diameter difference of 68μm-65μm=3μm.

[0023] It should be noted that the preset continuity condition is set as the absolute value of the median diameter difference being less than 15 μm. This threshold is based on statistical analysis of 1000 sets of particle size data from adjacent time points under continuous operation of a powder production line. During normal continuous evolution, 99.7% (three standard deviations) of the absolute value of the median diameter difference does not exceed 14.3 μm. This is rounded down to 15 μm to exclude abnormal gaps caused by material flow interruptions or abnormal equipment start-ups and shutdowns. The currently calculated median diameter difference of 3 μm has an absolute value of 3 μm, which is less than 15 μm, satisfying the preset continuity condition. This indicates that the powder particle size is in a normal continuous evolution state, and this set of values ​​is considered valid data and input to the next calculation stage. The time interval between adjacent time nodes T1 and T2 is obtained. This time interval is determined by the sampling period of the measurement unit. Assuming the sampling period is set to 2 seconds, the time interval Δt = 2 seconds. Using a first-order difference algorithm, the median diameter difference of 3 μm is divided by the time interval of 2 s to calculate the particle size change rate for the current time period as 3 μm ÷ 2 s = 1.5 μm / s. This rate objectively reflects the speed of particle size evolution under the current process conditions.

[0024] It should be noted that the preset first rate threshold is 2 μm / s, which was determined through the following calibration procedure. Three powder standard samples known to be in a critical agglomeration state were selected, with average particle sizes of 30 μm, 60 μm, and 100 μm, respectively. Each sample was subjected to 30 repeated measurements under standard ultrasonic dispersion conditions (ultrasonic frequency 20 kHz, dispersion power 500 W, dispersion time 60 s, sample cell volume 500 mL, medium: deionized water, temperature 25 °C). The rate of change of the median particle size during dispersion was recorded for each measurement. For each sample, the 30 measured rate values ​​were arranged in ascending order, and the value at the 95th percentile (i.e., the 29th value) was taken. This value represents the particle size change rate of most samples under normal dispersion conditions in the critical agglomeration state. The values ​​at the 95th percentile of the three samples were 2.1 μm / s, 1.9 μm / s, and 2.0 μm / s, respectively. The average value of the three was (2.1 + 1.9 + 2.0) / 3 = 2.0 μm / s, and after rounding, the first rate threshold was set to 2 μm / s.

[0025] When the calculated particle size change rate is greater than the threshold, the current particle state is determined to be a particle agglomeration state; if the particle size change rate is not greater than the threshold, the current particle state is determined to be a normal state, the current dispersion power is maintained, and monitoring continues. For example, suppose the currently obtained particle size change rate is 4 μm / s, which is greater than the preset first rate threshold of 2 μm / s, thus satisfying the agglomeration determination condition.

[0026] In one implementation, determining that the current particle state is a particle agglomeration state and generating a first power increase command includes: The power compensation benchmark is obtained by querying a preset power compensation mapping table based on the particle size change rate. Obtain the current load factor of the device, and multiply the power compensation benchmark with the current load factor of the device to obtain the target adjustment range; The current operating power of the device is obtained, the target adjustment range is added to the current operating power of the device to obtain the target operating power, and a first power increase command is generated based on the target operating power.

[0027] It should be noted that the preset power compensation mapping table is stored in the non-volatile memory of the equipment control system. The table stores the correspondence between particle size change rate ranges and power compensation benchmarks. The construction process of this mapping table is as follows: under standard test conditions, such as an ultrasonic frequency of 20kHz, a sample cell volume of 500mL, a medium of deionized water, a temperature of 25℃, and a test sample of standard alumina powder with a solid content of 30%, for different particle size change rates, the ultrasonic power is increased in increments of 50W. The minimum additional power required to completely depolymerize each grade of agglomerates is recorded. After three repeated experiments, the average value is taken as the power compensation benchmark for each grade. For example, the power compensation benchmark is 0.05 kW for particle size change rates in the range of (0,2] μm / s, 0.10 kW for the range of (2,4] μm / s, 0.20 kW for the range of (4,6] μm / s, 0.30 kW for the range of (6,8] μm / s, and 0.40 kW for the range of (8,10] μm / s. The corresponding power compensation benchmark can be obtained directly by looking up this mapping table based on the currently measured particle size change rate.

[0028] It should be noted that the current load factor of the equipment is determined by a combination of three factors: the real-time operating temperature of the ultrasonic transducer, the resonant frequency offset, and the upper limit of the duty cycle of the drive circuit. This is updated every 5 seconds by the equipment control system. Specifically, the temperature load factor η is calculated separately. T = (T max -T current ) / (T max -T ambient ), frequency load factor η F =1-|Δf| / f max and duty cycle load factor η D =1-D current / D max T max The maximum allowable operating temperature of the transducer is 80℃, T ambient At an ambient temperature of 25℃, f max For a maximum permissible frequency offset of 500Hz, D maxTo ensure the maximum duty cycle of the drive circuit is 100%, the minimum value of three factors is taken as the current load factor of the equipment. The load factor ranges from 0 to 1, where 1 represents no-load and 0 represents full-load. The target adjustment range is obtained by multiplying the power compensation benchmark with the current load factor. The power compensation benchmark represents the additional dispersion power required to completely deagglomerate the aggregates under ideal no-load conditions. However, in actual operation, the equipment is limited by factors such as increased operating temperature, frequency drift, and duty cycle margin. The actual additional power that can be provided needs to be reduced according to the current load capacity. The lower the load factor, the closer the equipment is to full load, the smaller the additional power margin, and the lower the additional power compensation value that the system can add. This ensures that the equipment always operates within a safe range while achieving effective dispersion. For example, the current temperature load factor η T =0.85, frequency load factor η F =0.90, duty cycle load factor η D =0.80, taking the minimum value yields a load factor of 0.80, representing that the current load level of the equipment is 80% of the rated power. Multiplying the power compensation benchmark of 0.2kW with the load factor of 0.80, the target adjustment range is calculated to be 0.2kW × 0.80 = 0.16kW.

[0029] Next, the current operating power of the device is obtained, and the target adjustment range is added to the current operating power to obtain the target operating power. For example, the actual operating power of the ultrasonic transducer at the current moment is read from the device power monitoring module as 1kW. The actual operating power of 1kW is summed with the target adjustment range of 0.16kW to obtain the target operating power of 1.16kW. Then, a first power increase command is generated based on the target operating power. This example uses a PID module. For example, the target operating power of 1.16kW is input as the setpoint to the proportional-integral-derivative (PID) control algorithm module, with the proportional coefficient Kp set to 1.2, the integral time Ti set to 0.8s, and the derivative time Td set to 0.2s. These PID parameters are determined during the device commissioning phase using the Ziegler-Nichols engineering tuning method. The PID module performs proportional, integral, and derivative calculations based on the deviation between the setpoint and the current actual operating power, outputting an analog power control signal from 0 to 10V, and generating a first power increase command based on this power control signal according to the device's CAN bus communication protocol.

[0030] In step S13, the current initial drive current value of the distributed unit is obtained, and the initial drive current value is incrementally calculated according to the first power increase instruction to obtain the first target drive current, including: The driving voltage value and initial driving current value of the distributed unit are obtained, and the current impedance value is calculated based on the initial driving current value and the driving voltage value. Extract the preset increment value according to the first power increase instruction; The preset increment value is corrected based on the current impedance value to obtain the target increment value; The first target driving current is obtained by summing the target increment and the initial driving current value.

[0031] It should be noted that the dispersion unit is the drive circuit module of the ultrasonic transducer, and its electrical operating parameters include drive current and drive voltage. The current initial drive current value is extracted as 5A and the drive voltage value as 220V from the current sampling resistor and voltage sampling circuit of the drive circuit, respectively. According to Ohm's law, dividing the drive voltage value of 220V by the initial drive current value of 5A yields the current impedance value of 44Ω. This impedance value comprehensively reflects the electromechanical coupling state of the piezoelectric ceramic sheet of the ultrasonic transducer, as well as the current viscosity and solid content of the powder slurry in contact with it. Higher impedance indicates greater resistance to the absorption of ultrasonic energy by the material.

[0032] The preset increment value is extracted based on the first power increase command. Specifically, the first power increase command is parsed to extract the preset increment value of 2A carried in the command data frame. The preset increment value of 2A is set based on the safe current ramp-up rate limit of the ultrasonic transducer drive circuit. That is, under rated heat dissipation conditions, the drive current should not increase by more than 2A each time to ensure that the power amplification module will not trigger protection due to instantaneous overheating. This safety limit is determined by the thermal resistance parameter (R) in the drive circuit datasheet. th(JA) =1.5℃ / W) is derived from the maximum junction temperature limit (Tjmax=150℃).

[0033] It should be noted that the absorption efficiency of ultrasonic dispersion energy varies significantly depending on the viscosity of the material. The current impedance value of 44Ω was dynamically adjusted using a preset impedance and amplitude compensation mapping relationship. The mapping relationship was obtained through the following calibration experiment: a test point was taken every 10Ω within the impedance range of 20Ω to 100Ω. The impedance value was measured using standard alumina powder slurry at different solid contents (10% to 60%), and the optimal increment was tested with a step size of 0.5A. Nine sets of impedance and optimal increment data were fitted into a linear interpolation table. For any impedance value not listed in the table, the corresponding optimal increment was calculated using linear interpolation between two adjacent known data points. When the current impedance value is 44Ω, the preset increment value of 2A was corrected to the target increment value of 3A by referring to the table and linear interpolation calculation. This correction logic takes into account the additional absorption requirements of high-impedance materials for ultrasonic energy and appropriately increases the increment to overcome viscous damping.

[0034] Then, using the initial drive current value of 5A as a reference, the initial drive current value is summed with the target increment of 3A, resulting in 5A + 3A = 8A, thus obtaining the first target drive current of 8A. This target drive current value is then sent to the digital register of the power supply control module of the dispersion unit via the SPI communication interface, for use in driving the ultrasonic transducer to perform ultrasonic dispersion at the adjusted power.

[0035] In step S14, the digital signal feedback channel is the core hardware infrastructure of the present invention. Its physical layer is based on RS-485 differential bus, and the data link layer adopts Modbus-RTU protocol. The particle size data output port of the measurement unit and the drive current control port of the dispersion unit are bidirectionally interconnected through shielded twisted pair cable, thereby breaking the information isolation between the measurement subsystem and the dispersion subsystem in the traditional laser particle size analyzer.

[0036] Next, a first target driving current is sent through the digital signal feedback channel, and a second particle size characteristic sequence is obtained. Specifically, when the ultrasonic transducer receives the 8A first target driving current determined in step S13, its internal piezoelectric ceramic component generates high-frequency mechanical vibration under the excitation of the alternating electric field, emitting ultrasonic waves to disperse the powder slurry in the sample cell. After the ultrasonic dispersion continues for a preset incremental period of 5 seconds, the measurement unit (i.e., the laser particle size analyzer) is triggered through the digital signal feedback channel to perform a second particle size measurement on the powder slurry in the sample cell. The measurement unit uses the same light scattering principle and Mie scattering inversion algorithm as in step S11 to output the initial particle size distribution data under the current dispersion power conditions. Following the same processing method as in step S11, volume percentage data of a preset particle size range of 50μm to 100μm is extracted from the initial particle size distribution data, and discretized equidistant sampling is performed according to a preset particle size interval of 0.5μm. The volume fraction percentage value corresponding to each sampling point is extracted and arranged in ascending order of particle size to generate a second particle size characteristic sequence containing 20 discrete data points. This sequence reflects the true particle size distribution of the powder slurry under the current dispersion power.

[0037] In one implementation, the step of using a decision tree algorithm to classify the volume average diameter in the second particle size feature sequence and determine the particle dispersion state category includes: The second particle size feature sequence is subjected to statistical moment analysis to obtain the initial volume average diameter value; The initial average volume diameter value is subjected to node splitting processing to obtain the target average volume diameter feature; The target volume average diameter feature is input into a preset decision tree model for traversal matching to obtain leaf node category labels. Based on the particle aggregation degree mapped by the leaf node category labels, the particle dispersion state category is determined.

[0038] Specifically, for each particle size value di and its corresponding volume fraction vi in ​​the second particle size feature sequence, the volume-weighted average value D[4,3]=Σ(vi×d) is calculated. i 4 ) / Σ(vi×d i 3 In this formula, the exponents 4 and 3 correspond to the fourth and third volumetric moments, respectively, and the calculation outputs the initial volumetric average diameter value characterizing the overall fineness of the current powder material. For example, the initial volumetric average diameter value D[4,3] is calculated to be 3 μm by statistical moment analysis.

[0039] It should be noted that the initial particle size distribution data from the previous measurement cycle (i.e., when step S11 is executed), calculated using the same method, is used as the historical volume average diameter value, stored in the system cache, and retrieved in this step. Using the initial volume average diameter value of 3 μm as a reference point, the 20 data points in the second particle size feature sequence are clustered according to their Euclidean distance from the reference point into coarse and fine particle size intervals. The values ​​of the coarse dominant diameter and the fine dominant diameter are extracted separately and combined to form the target volume average diameter feature. For example, the clustered coarse dominant diameter is 4.2 μm, and the fine dominant diameter is 1.8 μm. These two are combined into a two-dimensional feature vector (4.2, 1.8) as the target volume average diameter feature.

[0040] The target volume average diameter feature is input into a preset decision tree model for traversal matching to obtain leaf node category labels. The particle dispersion state category is determined based on the particle aggregation degree mapped to the leaf node category labels. Specifically, the preset decision tree model is a classification and regression tree algorithm model, which is obtained through the following training process: 500 sets of volume average diameter feature data for powder samples with different dispersion states (fully dispersed, slightly aggregated, moderately aggregated, heavily aggregated, and excessively broken, a total of 5 categories) are collected, with 100 sets for each category. The dominant diameters of coarse and fine particles are used as input features, and the dispersion state category is used as the output label. The Gini coefficient is used as the splitting criterion. The maximum tree depth is set to 5 layers, and the minimum number of leaf node samples is set to 10. The classification accuracy after 10-fold cross-validation is 92.4%. Starting from the root node of the decision tree, the target volume average diameter feature (4.2, 1.8) is sequentially traversed and matched with the splitting thresholds of each internal node, searching downwards layer by layer along the branch paths that meet the conditions. For example, the root node determines whether the dominant diameter of coarse particles is greater than 3.5 μm. Since 4.2 μm is greater than 3.5 μm, it enters the left subtree. The second layer determines whether the dominant diameter of fine particles is less than 2.0 μm. Since 1.8 μm is less than 2.0 μm, it enters the left subtree, and so on, until the 5th leaf node is reached. The category label bound to this node is extracted as label 3. A pre-constructed agglomeration state mapping table is called. In the mapping table, label 0 corresponds to fully dispersed, label 1 corresponds to light agglomeration, label 2 corresponds to moderate agglomeration, label 3 corresponds to heavy agglomeration, and label 4 corresponds to excessive crushing. According to the obtained label 3, the table is matched to map the particle agglomeration degree corresponding to the current powder material as moderate agglomeration. Based on the physical characterization of moderate agglomeration and combined with the process dispersion requirements of the current grinding stage, the particle dispersion state category is determined as the category to be deagglomerated.

[0041] In step S15, when step S14 determines the particle dispersion state as excessively broken through decision tree classification, the system automatically triggers the extraction process of the excessively broken degree value and generates a first power reduction instruction.

[0042] In one implementation, generating the first power reduction instruction includes: Obtain the value of the degree of excessive fragmentation corresponding to the excessive fragmentation state; The decrease in driving current is obtained by mapping the value of the excessive breakage degree. The power attenuation value is calculated based on the decrease in the drive current, and a first power reduction command is generated based on the power attenuation value.

[0043] It should be noted that when step S14 determines the particle dispersion state as excessively fragmented through decision tree classification, the system automatically triggers the extraction process for the degree of excessive fragmentation. Using the target volume average diameter feature calculated during decision tree classification as input, the dominant diameter D of the coarse particles is extracted. coarse and fine particle dominant diameter D fine Through formula N 破碎 =(D fine / D coarse The normalization factor of 200 is used to calculate the degree of excessive breakage. The normalization factor of 200 is determined as follows: Statistical analysis was performed on 50 samples with known excessive breakage states. The maximum value of the ratio of the dominant diameter of fine particles to the dominant diameter of coarse particles was 0.50. In order to map the degree of excessive breakage to the integer range of 0 to 100 for easy table lookup later, the normalization factor of 200 (200 × 0.50 = 100) was selected, so that the degree of excessive breakage values ​​are linearly distributed in the range of 0 to 100. The larger the value, the more severe the excessive breakage.

[0044] If step S14 determines that the particle dispersion state is fully dispersed, slightly aggregated, moderately aggregated, or heavily aggregated, it indicates that the current particles are not completely dispersed but have not yet reached the state of excessive fragmentation. At this time, the power reduction process in step S15 is not triggered, and the process returns to step S12 to continue monitoring whether the particle size change rate is still greater than the preset threshold so that the power adjustment process can be triggered again when the aggregated state continues to exist.

[0045] For example, using the coarse dominant diameter of 4.2 μm and the fine dominant diameter of 1.8 μm obtained in step S14 as input (assuming that this set of data corresponds to the over-fragmentation state scenario), the fine / coarse ratio is calculated to be 1.8 / 4.2≈0.43. After multiplying by the normalization coefficient of 200, the over-fragmentation degree value is output as 86.

[0046] It should be noted that there is an experimentally calibrated mapping relationship between the degree of over-fragmentation and the decrease in driving current. This mapping relationship was obtained through the following calibration experiment: Ten groups of standard powder samples with the same initial particle size were selected. Under standard ultrasonic dispersion conditions (ultrasonic frequency 20kHz, sample cell volume 500mL, medium: deionized water, temperature 25℃), the ultrasonic power was increased from 100W to 500W in 50W increments. After each dispersion treatment for 5 minutes, the degree of over-fragmentation was recorded. At the same time, the decrease in driving current required to restore the particle size to the normal range at each degree of over-fragmentation was recorded. The mapping relationship of driving current decrease (A) = 0.17 × degree of over-fragmentation was obtained by least squares linear regression fitting. The coefficient of determination R of the linear fit was... 2 =0.94. For example, based on the excessive breakage value of 86, the decrease in drive current is calculated to be 86 × 0.17 ≈ 15A.

[0047] The power attenuation value is calculated based on the decrease in drive current, and a first power reduction command is generated based on the power attenuation value. Specifically, the power attenuation value is calculated by multiplying the decrease in drive current by the rated drive voltage of the ultrasonic transducer, where the decrease in drive current is 15A, the rated drive voltage of the ultrasonic transducer is 220V, and the calculated power attenuation value is 3.3kW. According to the device's CAN bus communication protocol, the calculated power attenuation value of 3.3kW is encapsulated as the data payload field of the first power reduction command, with the command type identifier 0x02 (representing a power reduction command) and a CRC16 checksum appended to generate the complete first power reduction command.

[0048] In one implementation, determining the second target drive current by performing a decreasing calculation on the first target drive current according to the first power reduction instruction includes: The attenuation slope value is analyzed according to the first power reduction instruction, and the decreasing step size value is calculated based on the attenuation slope value and the preset drive current reference value. The decreasing step size value is mapped to obtain the transient current difference. Obtain the compensation bias current, and smooth the transient current difference and the compensation bias current to obtain the corrected decrease amplitude; The first target drive current is calculated by decreasing the correction rate to determine the second target drive current.

[0049] It should be noted that the attenuation slope value reflects the rate of decrease of the drive current during power attenuation. The first power reduction command is parsed, and the power attenuation value of 3.3kW is extracted from the data payload field. Simultaneously, the rheological damping coefficient of the current grinding material is obtained as the attenuation slope value. The rheological damping coefficient is obtained by real-time measurement of the apparent viscosity of the material by an online viscosity sensor installed in the measuring tank, conversion of the apparent viscosity, and division by the standard water medium viscosity (0.89 mPa·s) to obtain the relative damping coefficient. For example, the current relative damping coefficient is 3.0. The preset drive current reference value is 25A, which is the typical operating current calibrated at the factory for the ultrasonic transducer drive circuit. The attenuation slope value of 3.0 and the preset drive current reference value of 25A are input into the linear interpolation algorithm model. This model is based on the current and power attenuation characteristic curves calibrated by the equipment at the factory. It takes the power attenuation value of 3.3kW as input and the attenuation slope value of 3.0 as the curve slope control parameter. The linear interpolation operation is performed in the reference coordinate system with the drive current reference value of 25A, and the output decrease step size of 2A for each control cycle is 2A.

[0050] The decreasing step size is mapped to obtain the transient current difference. Specifically, the decreasing step size of 2A is input into the transducer impedance mapping matrix. The transducer impedance mapping matrix is ​​a parameter lookup table provided by the ultrasonic transducer manufacturer, which records the transient response characteristics of the transducer impedance corresponding to different current changes. Based on the transducer model, the row containing the current change of 2A in the mapping matrix is ​​looked up, and the corresponding transient current difference is found to be 5A. This difference reflects the current overshoot or undershoot caused by the release of energy stored in the stator winding inductance at the moment of current decrease.

[0051] It should be noted that the compensation bias current is determined by the bias state of the piezoelectric ceramic of the ultrasonic transducer and is acquired in real time from the bias current sampling circuit; the current acquired value is 1.5A. The transient current difference of 5A and the compensation bias current of 1.5A are input together into the Kalman filter algorithm model for smoothing. The process noise covariance Q is set to 0.05, and the measurement noise covariance R is set to 0.02. After one step of prediction and iterative updates, high-frequency harmonic interference generated by the piezoelectric ceramic resonance is filtered out, and the output correction decrease amplitude is 4A. The role of the Kalman filter here is to eliminate the random fluctuation components in the transient current difference and the measurement noise in the compensation bias current, obtaining a smooth and stable decrease amplitude value, ensuring that the amplitude is consistent in each decrease operation.

[0052] The first target drive current is calculated by decreasing the correction increment based on the stated reduction increment to determine the second target drive current. For example, the first target drive current to be executed in the current control cycle is 30A (here, 30A is the current value reached after multiple increments, distinct from the first target drive current of 8A obtained in step S13; in the actual system, the current value continuously increases with each step increment). The first target drive current of 30A is differentially calculated with the correction increment of 4A, i.e., 30A - 4A = 26A, to determine the second target drive current as 26A. This value serves as the target output current of the ultrasonic transducer drive circuit in the next control cycle.

[0053] The second target drive current is sent to the preset ultrasonic transducer through a digital signal feedback channel to achieve dynamic adjustment of the distributed power. Specifically, the second target drive current of 26A is input into the pulse width modulation coding algorithm model. The carrier frequency is set to 16kHz, which is higher than the audio range to avoid audible noise and lower than the limiting switching frequency of the power switching device. Based on the target current value of 26A, the pulse width modulation coding algorithm calculates the duty cycle according to the formula: target current value ÷ maximum output current of the drive circuit × 100%. The maximum output current of the drive circuit is 35A, so the duty cycle is 26 ÷ 35 × 100% = 74%, generating a digital drive signal sequence with a duty cycle of 74%. This digital drive signal is applied to the excitation end of the ultrasonic transducer, simultaneously activating the voltage sampling circuit and the current sampling circuit to collect transient voltage and transient current data at both ends of the ultrasonic transducer.

[0054] In summary, this invention discloses a method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback. This invention extracts the rate of change of the median particle diameter as a feedback signal. When the rate of change exceeds a preset first rate threshold, it determines that the particles are in an agglomeration state and automatically increases the driving current. A digital signal feedback channel enables bidirectional data communication between the measurement and dispersion ends. A decision tree algorithm is used to classify the volume average diameter to identify excessive fragmentation and automatically reduce the driving current. This achieves closed-loop bidirectional dynamic adjustment of the dispersion power, significantly improving the accuracy and repeatability of particle size measurement using the laser particle size analyzer.

[0055] The second embodiment of the present invention provides a laser particle size analyzer ultrasonic dispersion power adjustment system based on particle size feedback, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the laser particle size analyzer ultrasonic dispersion power adjustment method based on particle size feedback as described above.

[0056] It should be noted that the ultrasonic dispersion power adjustment system for laser particle size analyzer based on particle size feedback provided in this embodiment of the invention is used to execute all the process steps of the ultrasonic dispersion power adjustment method for laser particle size analyzer based on particle size feedback in the above embodiment. The working principle and beneficial effects of the two are one-to-one, so they will not be described again.

[0057] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0058] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback, characterized in that, include: The initial particle size distribution data output by the measurement unit is acquired, and the initial particle size distribution data is smoothed to obtain the first particle size feature sequence. The median diameter of the particles at adjacent time points is extracted based on the first particle size feature sequence, and the difference between the median diameter values ​​is calculated to obtain the particle size change rate. If the particle size change rate is greater than a preset first rate threshold, the current particle state is determined to be a particle agglomeration state, and a first power increase command is generated. Obtain the current initial drive current value of the distributed unit, and increment the initial drive current value according to the first power increase instruction to obtain the first target drive current; The first target driving current is sent to the preset ultrasonic transducer through the digital signal feedback channel, and the second particle size feature sequence fed back by the preset ultrasonic transducer under the first target driving current is obtained. The volume average diameter in the second particle size feature sequence is classified by the decision tree algorithm to determine the particle dispersion state category. If the particle dispersion state is classified as excessively broken, a first power reduction command is generated. Based on the first power reduction command, the first target driving current is calculated to decrease, a second target driving current is determined, and the second target driving current is sent to the preset ultrasonic transducer through the digital signal feedback channel to complete the dynamic adjustment of the dispersion power.

2. The method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback according to claim 1, characterized in that, The smoothing process of the initial particle size distribution data to obtain the first particle size feature sequence includes: Based on the initial particle size distribution data, the volume percentage data of the preset particle size range is extracted to obtain the local particle size distribution sequence; The local particle size distribution sequence is smoothed using the Kalman filter algorithm to obtain the first particle size feature sequence.

3. The method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback according to claim 1, characterized in that, The step of extracting the median diameter values ​​of adjacent time nodes based on the first particle size feature sequence and calculating the difference between the median diameter values ​​to obtain the particle size change rate includes: Based on the first particle size feature sequence, the median diameter value of the particles at each time point is extracted; According to the chronological order, the median diameter values ​​of adjacent time nodes are combined into median diameter value pairs. Calculate the difference between the median diameter values ​​of the particles to obtain the median diameter difference; If the median diameter difference satisfies the preset continuity condition, the particle size change rate is calculated based on the time interval between the median diameter difference and the adjacent time nodes.

4. The method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback according to claim 1, characterized in that, The step of determining that the current particle state is a particle aggregation state and generating a first power increase command includes: The power compensation benchmark is obtained by querying a preset power compensation mapping table based on the particle size change rate. Obtain the current load factor of the device, and multiply the power compensation benchmark with the current load factor of the device to obtain the target adjustment range; The current operating power of the device is obtained, the target adjustment range is added to the current operating power of the device to obtain the target operating power, and a first power increase command is generated based on the target operating power.

5. The method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback according to claim 1, characterized in that, The step of obtaining the current initial drive current value of the distributed unit and incrementally calculating the initial drive current value according to the first power increase command to obtain the first target drive current includes: The driving voltage value and initial driving current value of the distributed unit are obtained, and the current impedance value is calculated based on the initial driving current value and the driving voltage value. Extract the preset increment value according to the first power increase instruction; The preset increment value is corrected based on the current impedance value to obtain the target increment value; The first target driving current is obtained by summing the target increment and the initial driving current value.

6. The method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback according to claim 1, characterized in that, The step of using a decision tree algorithm to classify the volume average diameter in the second particle size feature sequence to determine the particle dispersion state category includes: The second particle size feature sequence is subjected to statistical moment analysis to obtain the initial volume average diameter value; The initial average volume diameter value is subjected to node splitting processing to obtain the target average volume diameter feature; The target volume average diameter feature is input into a preset decision tree model for traversal matching to obtain leaf node category labels. Based on the particle aggregation degree mapped by the leaf node category labels, the particle dispersion state category is determined.

7. The method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback according to claim 1, characterized in that, The generation of the first power reduction instruction includes: Obtain the value of the degree of excessive fragmentation corresponding to the excessive fragmentation state; The decrease in driving current is obtained by mapping the value of the excessive breakage degree. The power attenuation value is calculated based on the decrease in the drive current, and a first power reduction command is generated based on the power attenuation value.

8. The method for adjusting the ultrasonic dispersion power of a laser particle size analyzer based on particle size feedback according to claim 1, characterized in that, The step of calculating the decrease of the first target drive current according to the first power reduction instruction to determine the second target drive current includes: The attenuation slope value is analyzed according to the first power reduction instruction, and the decreasing step size value is calculated based on the attenuation slope value and the preset drive current reference value. The decreasing step size value is mapped to obtain the transient current difference. Obtain the compensation bias current, and smooth the transient current difference and the compensation bias current to obtain the corrected decrease amplitude; The first target drive current is calculated by decreasing the correction rate to determine the second target drive current.

9. A laser particle size analyzer ultrasonic dispersion power adjustment system based on particle size feedback, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the ultrasonic dispersion power adjustment method for laser particle size analyzer based on particle size feedback as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the ultrasonic dispersion power adjustment method for laser particle size analyzer based on particle size feedback as described in any one of claims 1 to 8.

Images