Particle size control method, device and equipment for power grid pressing type atomizer

Abstract

The invention relates to a particle size control method, device and equipment for a power grid pressing type atomizer. The method comprises the following steps: collecting an original data set of a voltage grid type atomizer, and determining a target particle size control interval; according to the target particle size control interval, creating an initial intermittent power supply scheme including pulse frequency, pulse width and pulse interval time; executing an initial intermittent power supply scheme, and collecting a current particle size measurement value through a laser scattering sensor; the deviation between the current particle size measurement value and the target particle size control interval is calculated, and a driving parameter adjustment instruction is generated through self-adaptive error compensation; and according to the driving parameter adjustment instruction, correcting the pulse frequency, the pulse width and the pulse interval time in the initial intermittent power supply scheme to obtain a target intermittent power supply scheme. According to the method, the optimal control parameters are formulated for the starting stage, the stable stage and the ending stage respectively, the specific particle size requirements of different atomization stages are met, and real-time high-precision monitoring of the particle size of the fog drops is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of piezoelectric mesh atomizers, and particularly to a method, device, and equipment for controlling the particle size of a piezoelectric mesh atomizer. Background Art

[0002] The piezoelectric mesh atomizer drives the liquid medicine to pass through the microporous mesh plate through piezoelectric vibration to form micron-sized droplets, realizing the directional delivery of drugs to the respiratory tract and lungs. The key technical problem faced by the current piezoelectric mesh atomizer is the insufficient accuracy of particle size control. Factors such as changes in the characteristics of the liquid medicine (viscosity, surface tension, and density), fluctuations in environmental parameters (temperature and humidity), instability of piezoelectric element parameters, and deviation of mesh hole sizes will all lead to unstable droplet size distribution. Especially the occurrence of mesh clogging faults will seriously disrupt the normal operation of the atomization system and cause a significant reduction in drug delivery efficiency. Summary of the Invention

[0003] The main object of the present invention is to provide a method, device, and equipment for controlling the particle size of a piezoelectric mesh atomizer. The present invention formulates optimal control parameters for the startup stage, stable stage, and end stage respectively, meets the specific particle size requirements of different atomization stages, and realizes real-time high-precision monitoring of droplet particle size.

[0004] To achieve the above object, the present invention provides a method for controlling the particle size of a piezoelectric mesh atomizer, including the following steps: Collect the original data set of the piezoelectric mesh atomizer and determine the target particle size control interval; Create an initial intermittent power supply scheme including pulse frequency, pulse width, and pulse interval time according to the target particle size control interval; Execute the initial intermittent power supply scheme, and collect the current particle size measurement value through a laser scattering sensor; Calculate the deviation between the current particle size measurement value and the target particle size control interval, and generate a driving parameter adjustment instruction through adaptive error compensation; According to the driving parameter adjustment instruction, correct the pulse frequency, pulse width, and pulse interval time in the initial intermittent power supply scheme to obtain the target intermittent power supply scheme.

[0005] The present invention also provides a device for controlling the particle size of a piezoelectric mesh atomizer, including: A collection module for collecting the original data set of the piezoelectric mesh atomizer and determining the target particle size control interval; A creation module for creating an initial intermittent power supply scheme including pulse frequency, pulse width, and pulse interval time according to the target particle size control interval; An execution module for executing the initial intermittent power supply scheme and collecting the current particle size measurement value through a laser scattering sensor; A calculation module, configured to calculate the deviation between the current particle size measurement value and the target particle size control range, and generate a driving parameter adjustment instruction through adaptive error compensation; A correction module, configured to correct the pulse frequency, pulse width, and pulse interval time in the initial intermittent power supply scheme according to the driving parameter adjustment instruction to obtain a target intermittent power supply scheme.

[0006] The present invention also provides a computer device, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps of the method described in any one of the above are implemented.

[0007] In summary, the technical solution provided by the present invention realizes an accurate mapping of particle size control by establishing a correspondence matrix between piezoelectric driving parameters and droplet particle sizes, provides an accurate mathematical model basis for subsequent control, and significantly improves the target positioning accuracy of particle size control. By adopting a phased control strategy, optimal control parameters are formulated for the start-up stage, stable stage, and end stage respectively, meeting the specific particle size requirements of different atomization stages, and effectively solving the technical problem that traditional single-parameter control is difficult to adapt to the whole process. The introduction of an intermittent power supply scheme realizes a rapid response to changes in liquid medicine characteristics by dynamically adjusting the pulse frequency, pulse width, and pulse interval time, enhancing the adaptability of the system to liquid medicines with different viscosities, surface tensions, and densities. The integration of high-precision laser scattering technology and a distributed specified-time observation system realizes real-time high-precision monitoring of droplet particle sizes, overcomes the defect that traditional open-loop control cannot perceive actual particle size changes, and provides a reliable feedback signal for closed-loop control. The development of an adaptive error compensation algorithm adopts a differential control strategy for different error regions, solves the problem of system parameter uncertainty, and ensures the particle size stability under various disturbance conditions. The construction of a mesh blockage detection and compensation mechanism accurately identifies the blockage state and implements corresponding parameter adjustments, effectively solving the problem of abnormal particle sizes under blockage faults and ensuring the stability of drug delivery throughout the treatment process. Description of the Drawings

[0008] Figure 1 is a schematic diagram of the steps of the particle size control method of a piezoelectric mesh nebulizer in an embodiment of the present invention; Figure 2 is a structural block diagram of the particle size control device of a piezoelectric mesh nebulizer in an embodiment of the present invention; Figure 3 is a schematic structural block diagram of a computer device in an embodiment of the present invention.

[0009] The realization, functional characteristics, and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. Detailed Embodiments

[0010] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0011] Referring to Figure 1 , this embodiment provides a method for controlling the particle size of a piezoelectric mesh nebulizer, including the following steps: S1, collect the original data set of the piezoelectric mesh nebulizer and determine the target particle size control range; Among them, a plurality of sensors are installed on the nebulizer to collect a series of important working parameters in real time. These sensors are used to obtain the piezoelectric oscillation frequency, driving voltage, and characteristic data of the liquid medicine, such as the viscosity, surface tension, and density of the liquid medicine. The sensors generate an original data set by accurately measuring these parameters. Data preprocessing is performed on the original data, including operations such as noise removal, outlier detection, and smoothing. After these steps, a standardized parameter data set is obtained. A three-dimensional parameter space is constructed based on the standardized parameter data set, and the droplet size distribution data measured under different parameter combination conditions is filled into the corresponding spatial positions. Each spatial position corresponds to a specific parameter combination condition, which is obtained by the sensors in actual measurement, and the droplet size distribution data under these conditions is recorded. In this way, a mapping relationship between the working parameters of the piezoelectric mesh nebulizer and the droplet size is initially established. The initial mapping relationship data is input into a multiple regression analysis model for coefficient fitting calculation to construct a corresponding relationship matrix between the parameters and the particle size. The atomization process is divided into three different stages: the startup stage, the stable stage, and the end stage. The target particle size requirements are different for each stage. Therefore, priorities are set for different stages to ensure that the particle size control can meet the treatment requirements to the greatest extent. In the startup stage, the target particle size should be as small as possible to ensure that the drug quickly enters the respiratory tract; in the stable stage, the particle size control needs to be maintained within a relatively stable range to ensure the uniform distribution of the drug; in the end stage, the particle size needs to be appropriately reduced to improve the deposition efficiency of the drug and reduce residues. By dividing each stage and setting a priority sequence for each stage, the system adjusts the particle size control strategy specifically to optimize the treatment effect. Based on the control priority sequence and the corresponding relationship matrix between the parameters and the particle size, the target particle sizes for different stages are calculated to form a target particle size control range. In each stage, the ideal target particle size range is calculated according to the actual parameter values and control priorities, combined with the data in the corresponding relationship matrix. The setting of the target particle size control range can provide a clear control range for the piezoelectric mesh nebulizer, ensuring that during the entire atomization process, the change in particle size can meet the treatment requirements and avoiding affecting the effect and safety of the drug due to excessive particle size fluctuations.

[0012] S2. According to the target particle size control range, create an initial intermittent power supply scheme including pulse frequency, pulse width, and pulse interval time; Specifically, after determining the control objectives of the atomization process according to the target particle size control range, the entire atomization process is divided into stages, including the start stage, the stable stage, and the end stage. In these three stages, the control requirements for particle size and the drug delivery method are different. Therefore, different target particle size values are set for each stage. By extracting the target particle size control range, a segmented control target sequence is generated to clarify the specific particle size values to be achieved in different stages. Based on the segmented control target sequence and the corresponding relationship matrix, reverse query calculations are performed to determine the initial piezoelectric drive parameter combinations required in each stage. The reverse query calculation process is based on the previously constructed corresponding relationship matrix between parameters and particle size. By substituting the target particle size value into the matrix, the corresponding piezoelectric drive parameters are obtained, mainly including parameters such as piezoelectric oscillation frequency and drive voltage. Based on the initial piezoelectric drive parameter combinations, a segmented linear response function model is constructed, which is used to describe the mapping relationship between the liquid medicine viscosity and the drive parameters. The liquid medicine viscosity is one of the important factors affecting the particle size, and the change in viscosity will directly affect the formation and size of the droplets. Therefore, constructing an accurate mapping relationship helps the system to dynamically adjust the drive parameters according to the different characteristics of the liquid medicine. By mathematically describing this mapping relationship, the relationship between the characteristic parameters of the liquid medicine and the target particle size value is transformed into a specific parameter response equation to ensure effective adjustment of particle size control under different liquid medicine characteristics. The parameter response equation is constrained and optimized to solve. By substituting the liquid medicine characteristic parameters and the target particle size value into the parameter response equation, the optimal parameter combinations including key parameters such as pulse frequency, pulse width, and pulse interval time are calculated. These optimal parameters help to accurately control the particle size in different atomization stages and ensure that the drug can be effectively delivered according to the predetermined treatment requirements. The optimal parameter combinations are simulated for the atomization effect. Through simulation, the performance of the selected parameter combinations in actual operation is predicted theoretically, and the simulation results are obtained and compared with the target particle size control range. If there is a deviation between the simulation results and the target particle size control range, a parameter correction amount is generated based on this deviation, and the parameters such as pulse frequency, pulse width, and pulse interval time in the power supply scheme are adjusted. According to the comparison between the simulation results and the target control range, the correction amount can accurately reflect the difference between the currently set parameters and the ideal state. Based on this correction amount, the initially set pulse frequency, pulse width, and pulse interval time are adjusted so that the power supply parameters used in the start stage, stable stage, and end stage can optimize the control of the particle size and achieve the expected treatment effect. The corrected power supply parameters configure the corresponding power supply parameters for each stage to form an initial intermittent power supply scheme.

[0013] S3. Execute the initial intermittent power supply scheme and collect the current particle size measurement value through a laser scattering sensor; It should be noted that the initial intermittent power supply scheme is input into the piezoelectric drive control circuit. The control circuit generates corresponding pulse signals according to these power supply parameters. These signals pass through the drive system of the piezoelectric mesh nebulizer to drive the piezoelectric element to work, enabling the liquid medicine to form droplets through the microporous mesh plate. The particle size and distribution of these droplets are affected by the physical properties of the liquid medicine and the atomization conditions. Therefore, after the drive system starts and completes the atomization process, the laser scattering sensor installed at the mist outlet of the piezoelectric mesh nebulizer is used to perform real-time scattered light measurement on the droplets. The laser scattering sensor can accurately measure the light signals scattered by the droplets. These light signals contain droplet particle size information and are the key source for obtaining particle size data. Multi-angle signal acquisition and digital filtering processing are performed on the original scattered light signals. The accuracy and reliability of the signals are enhanced through multi-angle acquisition. Digital filtering processing eliminates unnecessary noise signals and retains the effective information related to the droplet particle size, forming the initial particle size distribution data. The initial particle size distribution data is input into the distributed specified time observation system. This system uses multiple particle size observation nodes to perform parallel processing on data in different particle size ranges. Each observation node processes particle size data within a specific range. This distributed processing method can significantly improve the efficiency and real-time performance of data processing, ensuring that the entire process can obtain results in a short time. Through distributed processing, the segmented particle size characteristic data obtained reflects the distribution characteristics of droplets in different particle size ranges. Based on the non-linear state equation and the time-varying gain matrix, mathematical processing is performed on the segmented particle size characteristic data. The non-linear state equation can more accurately describe the dynamic changes in the droplet particle size distribution, while the time-varying gain matrix can dynamically adjust the observation gain according to different stages of the atomization process. As the atomization process progresses, the particle size characteristics of the droplets will change differently, especially in the start-up stage, stable stage, and end stage of atomization. The adjustment of the gain can help the system more accurately capture and process particle size information, thereby generating calibrated particle size data. Based on the calibrated particle size data, the mean value and distribution width of the particle size are calculated. The particle size mean value represents the average size of the droplets, while the distribution width describes the width of the particle size distribution. These two indicators are important parameters for evaluating the atomization effect. By calculating the particle size mean value and distribution width, the current atomization state is understood, and the power supply parameters are further adjusted based on these data to ensure that the particle size always remains within the target control range. The currently calculated particle size measurement value will be used as a feedback signal to provide a basis for the next adjustment of the system.

[0014] S4. Calculate the deviation between the currently measured particle size value and the target particle size control range, and generate a drive parameter adjustment instruction through adaptive error compensation; Specifically, calculate the difference between the current particle size measurement value and the target particle size control range to obtain the particle size error. The particle size error represents the deviation between the current droplet particle size and the desired target particle size, and measures the accuracy of the atomization effect. Through the calculation of this error, a particle size error evaluation index is generated to help the system determine whether the current particle size is within the target control range or deviates from the predetermined range. Based on the particle size error evaluation index, the error space is divided into multiple regions, including the first error region, the second error region, and the third error region. According to different degrees of error, different control strategies are adopted to adjust the power supply parameters. The first error region refers to the range with a smaller error, and the proportional control method is used to process it. Proportional control adjusts the drive parameter directly proportional to the error to ensure rapid correction of the particle size. The second error region corresponds to the range of medium-sized errors, and the proportional-integral control method is used for adjustment. The proportional-integral control method can not only adjust according to the current error, but also consider the cumulative effect of the error, thereby improving the system's response ability to medium errors. The third error region is the range with a larger error, and the proportional-integral-derivative control method is used for processing. The proportional-integral-derivative control method introduces a differential term on the basis of the proportional-integral control, which can more finely respond to the change trend of the error, quickly eliminate larger errors and stabilize the system operation. In order to effectively respond to the adjustment requirements of different error regions, a multi-mode control parameter library is constructed, which contains the control parameters required in each error region. The control parameter library specifies different control strategies for each error region to ensure that when an error occurs, the system adopts an appropriate control method according to the size and change of the error. The control parameters in the multi-mode control parameter library are not static, but are associated and analyzed with the error change trend. By analyzing the error change trend, the control parameters are dynamically adjusted so that the parameter configuration can adapt to the change of the error at any time. Input the adaptive control parameter matrix and the current error data into the error compensation controller. The error compensation controller calculates and generates an initial compensation amount according to the control algorithm and the adaptive control parameter matrix. The initial compensation amount represents the compensation amount required to adjust the particle size to the target range under the current error condition. The calculated initial compensation amount is directly used to adjust the working parameters of the atomizer to improve the deviation of the droplet particle size. During the atomization process, the droplet particle size will change over time, especially in the start-up stage, stable stage, and end stage of atomization. The particle size control requirements in different stages are different. Therefore, the initial compensation amount is optimized and adjusted according to the characteristics of the atomization stage. During the optimization process, a time-varying gain strategy is adopted, and by applying different weight coefficients in different atomization stages, the influence degree of the compensation amount is dynamically adjusted.In the startup phase, since the system is in the initial working state, the change of droplet size is more drastic, so a larger weight coefficient is needed for rapid adjustment; in the stable phase, the system is close to the target particle size range, the adjustment range should be smaller, and the weight coefficient is reduced accordingly; in the end phase, since the atomization process is nearing its end, the demand for particle size control gradually weakens, and the weight coefficient is further reduced. Through this optimization adjustment, the compensation amount can more accurately adapt to the characteristics of each atomization stage, ensuring the accuracy of particle size control at different stages. Through the adjustment of the time-varying gain strategy, the system outputs the drive parameter adjustment instruction, which contains precise parameters such as pulse frequency, pulse width and pulse interval time to ensure that the particle size control is always within the target range.

[0015] By analyzing the changing trend of the droplet size and the historical changes of the power supply parameters, the state of the atomization process is identified to determine whether the current system is in the startup stage, the stable stage or the end stage, and a stage identification signal is generated to identify the specific stage in which the current atomization process is located. According to the stage identification signal, the corresponding stage weight coefficient is extracted from the preset weight database. Each atomization stage has a specific gain value, and different gain values are set according to the characteristics of different stages to ensure that the adjustment of the compensation amount can adapt to the actual needs in the atomization process. For the startup stage, since the droplet size changes greatly in the initial stage of the atomization process, the system needs to respond quickly, and a relatively high first gain value is set for the startup stage; in the stable stage, the system is already close to the target droplet size range, so the control requirements tend to be stable, and the gain value should be set as the second gain value, which is relatively small; while in the end stage, since the atomization process is about to be completed and the demand for droplet size control gradually weakens, a third gain value is set to further reduce the gain and ensure that the droplet size can be maintained within the target range and avoid over-adjustment. Through these phased gain value adjustments, a time-varying weight vector is formed. The initial compensation amount is decomposed into components, which are decomposed into a frequency compensation component, a pulse width compensation component and a pulse interval compensation component. Through the decomposition, the adjustment of each atomization parameter can be controlled more carefully, so that each parameter can be accurately corrected according to specific needs. The decomposed compensation amount forms a compensation component matrix. Based on the time-varying weight vector, a weighted calculation is performed on the compensation component matrix to obtain the stage adaptive compensation amount. The stage adaptive compensation amount is predicted and verified through a dynamic response model. The dynamic response model can simulate the effect of the compensation amount in the actual atomization process and predict whether the adjustment of the compensation amount can achieve the expected droplet size control effect. Through verification, it is ensured that the adjustment of the compensation amount will not cause the droplet size to fluctuate beyond the control range and can keep the droplet size stable to meet the treatment requirements. After verification, the target compensation amount is generated and converted into an actual operable drive parameter adjustment value. The target compensation amount will be converted into a pulse frequency adjustment value, a pulse width adjustment value and a pulse interval adjustment value, and these adjustment values are used to drive the control circuit of the system and guide the working state of the piezoelectric element. Through these adjustment instructions, the droplet size is accurately controlled, and the drive parameters are flexibly adjusted at different atomization stages to ensure that the droplet size is always within the predetermined target range throughout the atomization process, thereby optimizing the atomization effect and improving the accuracy and stability of the treatment.

[0016] S5. According to the drive parameter adjustment instruction, correct the pulse frequency, pulse width and pulse interval time in the initial intermittent power supply scheme to obtain the target intermittent power supply scheme.

[0017] Among them, parameter update calculation is performed on the initial intermittent power supply scheme according to the drive parameter adjustment instruction, which affects the adjustment of power supply parameters. The current power supply parameters are added and subtracted with the adjustment amount to obtain the updated power supply parameter combination. The updated power supply parameters include key parameters such as pulse frequency, pulse width, and pulse interval time, ensuring that each parameter is accurately corrected according to the real-time measured particle size and error, so that the droplet particle size can be accurately controlled within the target range. However, the adjustment of power supply parameters is not just a simple numerical addition and subtraction. Some potential equipment problems need to be considered, especially the mesh blockage situation. Mesh blockage will directly affect the atomization effect, causing fluctuations in the droplet particle size and even resulting in unstable atomization effect. Detect the mesh blockage state of the updated power supply parameter combination. Detection is carried out through the Convolutional Block Attention Module (CBAM) and Convolutional Neural Network (CNN). CBAM is used to extract relevant features of mesh blockage, while CNN performs in-depth learning and analysis on the extracted features. The purpose of this analysis is to evaluate the influence degree of mesh blockage on the atomization effect and generate blockage compensation parameters. These compensation parameters are used to correct the current power supply parameters, so as to adjust the power supply strategy in case of mesh blockage and ensure that the atomizer can continue to work effectively. After detecting different degrees of mesh blockage, the updated power supply parameters are corrected according to the blockage compensation parameters. Adjust the power supply parameters according to different mesh blockage states. For example, when mild blockage is detected, slightly increase the drive voltage or adjust the pulse frequency; while when the blockage degree is severe, more adjustments are needed, such as increasing the drive voltage, increasing the pulse frequency, or reducing the pulse interval time. Such adjustments can effectively compensate for the change in atomization effect caused by blockage, so that the droplet particle size can be maintained within the target range. The adjustment process forms a blockage-adaptive power supply parameter, which can be automatically optimized according to different degrees of mesh blockage to ensure that the atomizer can achieve the best atomization effect under various working conditions. Parameter optimization and digital processing are carried out on the blockage-adaptive power supply parameters. The purpose of parameter optimization is to improve the stability and response speed of the power supply system. By optimizing the blockage-adaptive power supply parameters, the system adjusts the parameters to be more refined according to the actual operating conditions and equipment status, avoiding parameter fluctuations caused by excessive adjustment. Through digital processing, the optimized parameters are converted into discrete digital quantities executable by the equipment. By converting continuous parameter values into digital signals, each power supply link can be controlled more precisely, enabling the drive circuit to execute accurately according to the target instruction. After the above steps, the target intermittent power supply scheme is formed.

[0018] In one example, collect the original data set of the piezoelectric mesh atomizer and determine the target particle size control interval, including: Collect piezoelectric oscillation frequency, drive voltage, and liquid medicine characteristic data through multiple sensors installed on the piezoelectric mesh atomizer to generate the original data set; Perform data preprocessing on the original dataset to obtain a standardized parameter dataset; Construct a three-dimensional parameter space based on the standardized parameter dataset, and fill in the droplet size distribution data measured under different parameter combinations at the corresponding spatial positions to form initial mapping relationship data; Input the initial mapping relationship data into a multiple regression analysis model for coefficient fitting calculation to construct a corresponding relationship matrix between parameters and particle size; Divide the atomization process into a startup stage, a stable stage, and an end stage, and establish a control priority sequence for each stage; Calculate the target particle size for each stage based on the control priority sequence and the corresponding relationship matrix to form a target particle size control interval.

[0019] In this example, working parameters related to the atomization process are collected by multiple sensors installed on a piezoelectric mesh atomizer. These sensors collect key working data in real time, including piezoelectric oscillation frequency, driving voltage, and liquid medicine characteristic data. The piezoelectric oscillation frequency and driving voltage determine the working state of the piezoelectric mesh, affecting the working efficiency of the atomizer and the formation of droplets. The characteristics of the liquid medicine, such as viscosity, surface tension, density, etc., directly affect the atomization effect and particle size distribution of the liquid medicine. After summarizing these data, an original dataset is generated. Perform data preprocessing on the original dataset to eliminate problems such as noise, missing data, and inconsistent units. The steps of data preprocessing include data cleaning, standardization, and normalization, etc., so that the data of all parameters meet a unified standard and can be effectively used for subsequent modeling and analysis. The preprocessed data forms a standardized parameter dataset. Construct a three-dimensional parameter space based on the standardized parameter dataset. The construction of this parameter space is based on the relationships between multiple factors, including multiple variables such as piezoelectric driving frequency, driving voltage, and liquid medicine characteristics. By inputting these standardized parameters into the three-dimensional space, the droplet size distribution data measured under different parameter combinations are filled in the corresponding positions. The establishment of the three-dimensional parameter space helps the system to accurately understand the change trend of droplet size under the conditions of multiple variable combinations, providing an accurate mathematical model for subsequent particle size control. For example, assume there are piezoelectric oscillation frequency , driving voltage , and liquid medicine viscosity and other parameters. Map the combination of these three to a three-dimensional space, where the coordinates of each point represent a specific parameter combination, and the value of this point represents the droplet size measured under this parameter condition According to the above construction of the three-dimensional parameter space, initial mapping relationship data is obtained. This data set contains the droplet size data corresponding to each set of input parameters. The initial mapping relationship data is input into a multiple regression analysis model for coefficient fitting calculation. The multiple regression model fits a set of coefficients based on multiple input parameters (such as piezoelectric oscillation frequency, driving voltage, and liquid medicine characteristics, etc.), and these coefficients reveal the specific mathematical relationship between the parameters and the particle size. Set up a mathematical model:

[0020] where, represents the droplet size, is the piezoelectric oscillation frequency, is the driving voltage, is the viscosity of the liquid medicine, are the coefficients obtained through regression analysis, is the constant term. Through multiple regression analysis, the corresponding relationship matrix between the parameters and the particle size is accurately constructed, so as to predict the droplet size under given parameter conditions. The atomization process is divided into different stages, including the startup stage, the stable stage, and the end stage. Different stages have different control requirements. The startup stage requires a smaller particle size to quickly enter the respiratory tract; the stable stage requires maintaining a relatively stable particle size to ensure uniform distribution of the drug; the end stage requires a smaller particle size to reduce residual drugs. In order to achieve this precise control, different control priorities are set for each stage. These priorities determine how the control target of the particle size needs to be adjusted in different stages, and how to achieve the target particle size through changes in the power supply parameters. Based on the control priority sequence and the corresponding relationship matrix established previously, the target particle size is calculated for each stage. In the startup stage, the target particle size is smaller, such as 1 - 2 μm; in the stable stage, the target particle size may be maintained between 2 - 4 μm; while in the end stage, the target particle size returns to 1 - 3 μm. By calculating the target particle size value for each stage, the target particle size control interval for each stage is determined. For example, in the stable stage, the target particle size will be adjusted according to various factors in the atomization process, and the following formula is used to determine the target value of the particle size:

[0021] where, and are the set frequency and voltage, is the characteristic of the liquid medicine, and the function represents the particle size prediction model obtained through regression analysis.

[0022] In an example, according to the target particle size control interval, an initial intermittent power supply scheme including pulse frequency, pulse width, and pulse interval time is created, including: Extract the target particle size values for the start-up stage, stable stage, and end stage of the atomization process according to the target particle size control range, and generate a segmented control target sequence; Perform reverse query calculations based on the segmented control target sequence and the correspondence matrix to obtain the initial piezoelectric drive parameter combinations corresponding to each stage; Construct a piecewise linear response function model based on the initial piezoelectric drive parameter combinations to mathematically describe the mapping relationship between the liquid medicine viscosity and the drive parameters, and form a parameter response equation; Perform constrained optimization to solve the parameter response equation, substitute the liquid medicine characteristic parameters and the target particle size value into the parameter response equation, and calculate the optimal parameter combination including the pulse frequency, pulse width, and pulse interval time; Perform atomization effect simulation on the optimal parameter combination to obtain the simulation results, and compare the simulation results with the target particle size control range to generate a parameter correction amount; Adjust the pulse frequency, pulse width, and pulse interval time according to the parameter correction amount, and configure the corresponding power supply parameters for each stage to form an initial intermittent power supply scheme.

[0023] In this example, extract the target particle size values for the start-up stage, stable stage, and end stage of the atomization process according to the target particle size control range, and generate a segmented control target sequence. The target particle size control range sets a specific particle size range for each atomization stage. For example, in the start-up stage, the particle size range is set to 1 - 2 μm, in the stable stage it is set to 2 - 4 μm, and in the end stage it is set to 1 - 3 μm. Based on the target particle size values, a segmented control target sequence is generated, indicating the particle size targets to be achieved in different stages. After generating the target particle size sequence, perform reverse query calculations by comparing the segmented control target sequence with the established correspondence matrix between parameters and particle size to obtain the initial piezoelectric drive parameter combinations required in different stages. Assume that the relationship between the piezoelectric oscillation frequency 、drive voltage and the liquid medicine viscosity and the particle size has been established through experiments or numerical models, and a regression model has been obtained through regression analysis, where are the regression coefficients. Through reverse query of the target particle size and other parameters, calculate the initial parameter combinations required for each stage (start-up, stable, end). These initial parameter combinations provide a starting point for subsequent particle size control. For example, assume that the target particle size obtained through reverse query calculations in a certain stage has the initial drive parameters of , and these parameters serve as the basis for the calculation. Based on the initial piezoelectric drive parameter combinations, construct a piecewise linear response function model to describe the liquid medicine viscosity The mapping relationship with the driving parameters. The viscosity of the liquid medicine is an important factor affecting the particle size. Liquid medicines with different viscosities require different driving voltages and frequencies during the atomization process to maintain particle size control. By establishing such a response model, the driving parameters can be accurately adjusted to cope with the characteristics of different liquid medicines. For example, assume the driving frequency and the viscosity of the liquid medicine are described by a linear function, and the model form is:

[0024] where are the fitted coefficients, is the viscosity of the liquid medicine, is the driving frequency, is the driving voltage. The piecewise linear model can help the system better understand the relationship between the characteristics of the liquid medicine and the driving parameters. By performing constrained optimization on the parameter response equation, the optimal combination of driving parameters is calculated. Substitute the liquid medicine characteristic parameters and the target particle size value into the parameter response equation to solve for the optimal pulse frequency, pulse width, and pulse interval time. For example, when performing constrained optimization, set an objective function representing the deviation between the actual particle size and the target particle size, and use an optimization algorithm (such as the least squares method) to solve for the optimal parameter combination:

[0025] where is the actually measured particle size, is the target particle size value. Through constrained optimization, a set of optimal driving parameters is found to minimize the error of the particle size, thus ensuring that the atomizer reaches the predetermined particle size target at each stage. The atomization effect of the optimal parameter combination is simulated to obtain the simulation results to verify whether these parameter combinations can achieve the target particle size control in actual operation. The simulation process is carried out based on physical models or numerical simulation methods. By simulating the atomization process under different parameter combinations, the corresponding droplet size distribution results are obtained. The simulation results will be compared with the target particle size control interval to evaluate the effectiveness of the current parameter combination. If the simulation results show a large particle size deviation, the system generates corresponding parameter correction amounts according to the deviation. These correction amounts are used to adjust the driving parameters to ensure the final particle size control accuracy. For example, if the simulation shows that the particle size is too large, the driving voltage is reduced or the frequency is adjusted to reduce the droplet size. According to these parameter correction amounts, the pulse frequency, pulse width, and pulse interval time are adjusted to further optimize the particle size control. According to the adjusted parameters, the corresponding power supply parameters are configured for each stage to form an initial intermittent power supply scheme. These power supply parameters include pulse frequency, pulse width, and pulse interval time, which will be adjusted according to the characteristics of each stage. A higher frequency and a smaller pulse interval time are required in the startup stage; a lower frequency and a longer pulse interval time are required in the stable stage; and the end stage will be adjusted again according to the change of the liquid medicine. Through this process, an initial intermittent power supply scheme is generated to ensure that the atomizer achieves precise particle size control in different atomization stages, thus providing the best treatment effect.

[0026] In one example, the initial intermittent power supply scheme is executed, and the current particle size measurement values are collected through a laser scattering sensor, including: The initial intermittent power supply scheme is input into the piezoelectric drive control circuit to generate corresponding pulse signals to drive the piezoelectric mesh atomizer to work and generate liquid medicine atomization particles; Through the laser scattering sensor installed at the fog outlet of the piezoelectric mesh atomizer, the scattered light of the atomized droplets is measured in real time, and the original scattered light signal is collected; The original scattered light signal is subjected to multi-angle signal acquisition and digital filtering processing to form initial particle size distribution data; The initial particle size distribution data is input into the distributed specified time observation system. Through multiple particle size observation nodes, different particle size range data is processed in parallel to obtain segmented particle size characteristic data; Based on the nonlinear state equation and the time-varying gain matrix, the segmented particle size characteristic data is mathematically processed, and the observation gain is dynamically adjusted according to the atomization stage to generate corrected particle size data; According to the corrected particle size data, the particle size mean value and distribution width are calculated, and the current particle size measurement value is output.

[0027] In this example, the initial intermittent power supply scheme is input into the piezoelectric drive control circuit. The control circuit generates corresponding pulse signals according to parameters such as the input pulse frequency, pulse width, and pulse interval time, and makes the piezoelectric network work by driving the piezoelectric elements. Through the vibration of the piezoelectric elements, the liquid medicine is pressed out and forms droplets through the microporous mesh plate, generating atomized liquid medicine particles. The droplets generated by atomization are measured in real time by a laser scattering sensor installed at the mist outlet of the piezoelectric network atomizer. The laser scattering sensor determines the droplet size by emitting a laser beam and measuring the signal intensity generated by the scattered light of the droplets. The original scattered light signal collected by the sensor contains information about the droplet size. Since in the actual atomization process, the scattered light signal is affected by noise, interference, and measurement errors, a series of processes are performed on the original scattered light signal to ensure the accuracy and stability of the data. The processing of the original scattered light signal includes multi-angle signal acquisition. The sensor collects scattered light signals from multiple angles. Through multi-angle acquisition, the reliability and accuracy of the signal are improved, and the errors caused by the missing or distortion of the signal at a single angle are reduced. All the collected signals are processed by digital filtering to remove high-frequency noise, low-frequency interference, and other irrelevant signals, and retain the most valuable droplet size information. The signal after digital filtering is used to form the initial droplet size distribution data. The preliminary droplet size distribution data is input into the distributed specified time observation system. The system consists of multiple droplet size observation nodes. Each node processes data in different droplet size ranges and improves the data processing efficiency through parallel computing. Each observation node extracts characteristic data from different droplet size ranges and estimates the distribution of the droplet size based on these data. Through this distributed processing, the droplet size is analyzed comprehensively in real time and accurately, and more detailed and complete droplet size characteristic data is obtained. Based on the nonlinear state equation and the time-varying gain matrix, mathematical processing is performed on the segmented droplet size characteristic data. The nonlinear state equation is used to describe the dynamic process of droplet size change. This equation can more accurately simulate the change law of droplet size and take into account the influence of various factors on the droplet size, such as the properties of the liquid medicine, the supply voltage, and the ambient temperature. In addition, the time-varying gain matrix is used to adjust the observation accuracy of the droplet size measurement data. The gain matrix changes with time and can dynamically adjust the observation gain according to different stages in the atomization process. The gain is larger in the startup stage, moderate in the stable stage, and smaller in the ending stage. Ensure that during the entire atomization process, the measurement error of the droplet size data can be effectively controlled and better adapt to the disturbances in the droplet size measurement process. After the above mathematical processing, the corrected droplet size data is generated. Based on the corrected data, the mean value and distribution width of the droplet size are calculated to reflect the droplet size distribution characteristics. The mean droplet size represents the average level of the droplet size, while the distribution width indicates the degree of dispersion of the droplet size distribution. The smaller the mean droplet size, the finer the atomization effect; and the smaller the distribution width, the more consistent the droplet sizes are. By calculating these indicators, the current droplet size measurement value is obtained.

[0028] In one example, the deviation between the current particle size measurement value and the target particle size control range is calculated, and a driving parameter adjustment instruction is generated through adaptive error compensation, including: Calculate the difference between the current particle size measurement value and the target particle size control range to generate a particle size error evaluation index; Based on the particle size error evaluation index, divide the error space into a first error region, a second error region, and a third error region; Proportional control is adopted in the first error region, proportional-integral control is adopted in the second error region, and proportional-integral-derivative control is adopted in the third error region to construct a multi-mode control parameter library; Perform a correlation analysis on the control parameters in the multi-mode control parameter library and the error change trend to form an adaptive control parameter matrix; Input the adaptive control parameter matrix and the current error data into an error compensation controller, execute the control algorithm calculation of the corresponding mode, and generate an initial compensation amount; Optimize and adjust the initial compensation amount according to the characteristics of the atomization stage, and use a time-varying gain strategy to apply different weight coefficients in different atomization stages to output a driving parameter adjustment instruction.

[0029] In this example, calculate the difference between the current particle size measurement value and the target particle size control range to generate a particle size error evaluation index. The particle size error refers to the difference between the actually measured particle size and the target particle size control range. Assume that the currently measured particle size is , the lower limit of the target particle size range is , and the upper limit is , then the particle size error is expressed as:

[0030] where is the particle size error, representing the deviation between the actual particle size and the median value of the target particle size range. This error metric is used to measure whether the current atomization effect meets the treatment requirements and provides a quantitative basis for subsequent adjustments. Based on the particle size error evaluation index, the error space is divided into multiple regions to select appropriate control strategies. The error space is divided into three parts: the first error region, the second error region, and the third error region. The first error region corresponds to the case of smaller errors. At this time, proportional control is used to quickly adjust the driving parameters so that the particle size can return to the target range as soon as possible. The simplicity of proportional control makes it suitable for cases with smaller errors, and the power supply parameters are adjusted through a simple linear relationship. The second error region corresponds to the case of medium errors, and proportional-integral control is introduced. Proportional-integral control combines the current value and the historical cumulative value of the error, thereby improving the response ability to the long-term deviation of the system. The third error region corresponds to larger errors, and proportional-integral-derivative control is used to further enhance the response to the rate of change of the error, ensuring that the system can quickly and accurately adjust the parameters and prevent large fluctuations in the particle size. Through the division of the error region, the appropriate control strategy is automatically selected according to the size of the error. To achieve this function, a multi-mode control parameter library is constructed, which contains the control parameters required under different error regions. Each control strategy (proportional control, proportional-integral control, and proportional-integral-derivative control) corresponds to a specific set of control parameters, and these parameters are used to adjust the driving signal of the system to keep the particle size within the target range. For example, the parameter of proportional control is only one gain factor , the parameters of proportional-integral control include the proportional gain and the integral gain , while proportional-integral-derivative control includes the proportional gain , the integral gain and the derivative gain . Correlation analysis is carried out on the control parameters in the multi-mode control parameter library and the error change trend. The error change trend refers to the change of the error over a period of time. The system analyzes these trends to identify the rapid change or the trend of stabilizing of the error, and then adjusts the control parameters. Suppose the system analyzes the rate of change of the error , and the change trend of the error is expressed as:

[0031] where, is the rate of change of the error, is the error value at time , is the time interval. By analyzing the rate of change of the error, it is decided whether to increase the response speed of the control parameter or decrease the response to avoid over-adjustment. Based on this correlation analysis, an adaptive control parameter matrix is formed, which can flexibly adjust the parameters of each control strategy under different error states. For example, if the error changes rapidly, a larger gain parameter will be selected to accelerate the response; if the error changes slowly, a smaller gain parameter will be selected to stabilize the system. After inputting the adaptive control parameter matrix and the current error data into the error compensation controller, the controller selects an appropriate control mode according to the current error state and executes the control algorithm calculation of the corresponding mode to generate an initial compensation amount. The initial compensation amount refers to the minimum driving parameter adjustment amount required to adjust the particle size to the target interval under the current error conditions. Assume that the compensation amounts output by the control algorithm are , which correspond to the adjustment amounts of the pulse frequency, pulse width, and pulse interval time respectively. Through these adjustment amounts, the amplitude and frequency of the driving signal are adjusted to make the droplet particle size approach the target value. The initial compensation amount is optimized and adjusted according to the characteristics of the atomization stage. The atomization process is divided into a startup stage, a stable stage, and an end stage, and the particle size control requirements in each stage are different. In the startup stage, since the system is not yet stable, the change in particle size is large, and a stronger adjustment force is required; in the stable stage, the system is already close to the target particle size, so a smaller adjustment amplitude is needed; in the end stage, since the drug delivery is approaching the end, the adjustment requirement is relatively small. To adapt to different atomization stages, a time-varying gain strategy is adopted, and different weight coefficients are applied according to different stages. Assume that the gain coefficients in different stages are , , and , and these coefficients will be multiplied by the initial compensation amount to adjust the final compensation amount. In the startup stage, the gain coefficient is increased to accelerate the speed of particle size adjustment; in the stable stage, the gain coefficient is smaller to maintain stability; in the end stage, the gain coefficient is further reduced to ensure that the particle size no longer changes violently. After optimizing and adjusting the initial compensation amount through the time-varying gain strategy, drive parameter adjustment instructions are output. These instructions include the updated values of the pulse frequency, pulse width, and pulse interval time, ensuring that the droplet particle size is within the target control interval and meets the treatment requirements of each atomization stage.

[0032] In one example, the initial compensation amount is optimized and adjusted according to the characteristics of the atomization stage, a time-varying gain strategy is used to apply different weight coefficients in different atomization stages, and drive parameter adjustment instructions are output, including: By analyzing the change trend of the droplet particle size and the change history of the power supply parameters, the state of the atomization process is identified, it is determined whether the current is in the startup stage, the stable stage, or the end stage, and a stage identification signal is generated; Extract the corresponding stage weight coefficient from the preset weight database according to the stage identification signal, set the first gain value for the startup stage, the second gain value for the stable stage, and the third gain value for the end stage to form a time-varying weight vector; Perform component decomposition on the initial compensation amount, decompose the compensation amount into a frequency compensation component, a pulse width compensation component, and a pulse interval compensation component, and generate a compensation component matrix; Perform weighted calculation on the compensation component matrix based on the time-varying weight vector to form a stage adaptive compensation amount, and perform prediction verification on the stage adaptive compensation amount through a dynamic response model to generate a target compensation amount; Convert the target compensation amount into a pulse frequency adjustment value, a pulse width adjustment value, and a pulse interval adjustment value to form a drive parameter adjustment instruction.

[0033] In this example, the state of the atomization process is identified by analyzing the changing trend of the droplet size and the historical data of the power supply parameter changes. By analyzing this data, the current atomization stage - startup stage, stable stage, or end stage - is identified. The changing trend of the droplet size reflects the stability of the current atomization effect, while the historical changes of the power supply parameters reflect the speed and amplitude of the system response. To perform accurate state identification, this information is comprehensively analyzed, and the changes in droplet size and power supply parameters are modeled through specific algorithms (such as moving average method, Kalman filter, etc.) to calculate the state of the atomization process. Based on these analyses, a stage identification signal is generated, which is used to identify the current stage and provide a basis for subsequent control decisions. Assume the changing trend of the droplet size is expressed as and the historical changes of the power supply parameters are , through these data, determine which stage the current atomization process is in and generate the corresponding stage identification signal. Extract the corresponding stage weight coefficient from the preset weight database according to the stage identification signal. These weight coefficients represent the influence degree on the compensation amount adjustment in each stage. In the startup stage, the droplet size changes rapidly, so a larger gain is required to quickly adjust the droplet size; in the stable stage, the change of the droplet size tends to be stable, so the gain is relatively small; while in the end stage, since the liquid medicine delivery is approaching the end, the gain value is further reduced to prevent over-adjustment. Set the first gain value for the startup stage , set the second gain value for the stable stage , set the third gain value for the end stage . These gain values are obtained through the stage weight coefficient It is adjusted to form a time-varying weight vector, which dynamically changes according to the atomization stage, ensuring that the adjustment of the compensation amount can be finely controlled according to the requirements of different stages. The initial compensation amount is decomposed into components, including a frequency compensation component, a pulse width compensation component, and a pulse interval compensation component. The decomposed compensation amount reflects the independent adjustment effects of each control parameter during the atomization process. For example, assume the initial compensation amount is , which is decomposed into a frequency compensation amount , a pulse width compensation amount , and a pulse interval compensation amount . This decomposition process is represented by the following equations:

[0034] where represents the compensation of the pulse frequency, represents the compensation of the pulse width, and represents the compensation of the pulse interval time. These components are stored in a compensation component matrix, and the matrix is represented as:

[0035] Based on the time-varying weight vector, weighted calculations are performed on the compensation component matrix. The time-varying weight vector reflects the adjustment requirements of each stage and adjusts the compensation components through weighted calculations. For example, if the weight coefficient is during the startup stage, the corresponding weighted compensation amount is expressed as:

[0036] This weighted calculation will generate stage-adaptive compensation amounts, reflecting the differences in the adjustment of control parameters at different stages. Based on this compensation amount, a dynamic response model is used for prediction verification to verify the rationality of the compensation amount. The dynamic response model evaluates whether the compensation amount is effective enough by simulating the particle size change and ensures that the particle size can be stabilized within the target control range. Assume the predicted compensation amount is , which is calculated through the dynamic response model and compared with the actual particle size change. If the prediction error is small, it indicates that the compensation amount is effective; if the error is large, further adjustment is required. Through this step, the target compensation amount is generated. The target compensation amount is converted into an actual drive parameter adjustment instruction. According to the target compensation amount, the pulse frequency, pulse width, and pulse interval time are adjusted to form a drive parameter adjustment instruction.

[0037] In an example, according to the drive parameter adjustment instruction, the pulse frequency, pulse width, and pulse interval time in the initial intermittent power supply scheme are corrected to obtain the target intermittent power supply scheme, including: Update the parameters of the initial intermittent power supply scheme according to the drive parameter adjustment instruction, perform addition and subtraction operations on the current power supply parameters and the adjustment amount to form an updated power supply parameter combination; Detect the mesh blockage state of the updated power supply parameter combination through the convolutional block attention module and the convolutional neural network, evaluate the influence degree of the mesh blockage on the atomization effect, and generate blockage compensation parameters; Correct the updated power supply parameter combination according to the blockage compensation parameters, adjust the power supply parameters for different degrees of mesh blockage states, and form blockage-adaptive power supply parameters; Perform parameter optimization and digital processing on the blockage-adaptive power supply parameters, convert continuous parameter values into discrete digital quantities executable by the device, and generate the target intermittent power supply scheme.

[0038] In this example, update the parameters of the initial intermittent power supply scheme according to the drive parameter adjustment instruction. The initial intermittent power supply scheme includes key signal parameters such as the pulse frequency, pulse width, and pulse interval time of the piezoelectric drive. These parameters affect the atomization effect and the distribution of droplet sizes. According to the drive parameter adjustment instruction generated by the system, perform addition and subtraction operations on the current power supply parameters and the adjustment amount to obtain the updated power supply parameter combination. Assume the current power supply parameters are , and the adjustment amount is , then the updated power supply parameter combination is obtained through the following calculation:

[0039] Among them, , and represent the updated pulse frequency, pulse width, and pulse interval time respectively. Through this addition and subtraction operation, the drive signal is accurately adjusted to meet the target requirements for controlling the particle size. During the actual operation of the atomizer, the mesh will become blocked, resulting in a decrease in the atomization effect. To address this situation, detect the mesh blockage state of the updated power supply parameter combination through the convolutional block attention module (CBAM) and the convolutional neural network (CNN). CBAM and CNN can process image and signal data, extract features related to mesh blockage through convolutional operations, and screen important information through the attention mechanism. Through these deep learning methods, analyze the influence degree of the mesh blockage on the atomization effect and generate blockage compensation parameters. The convolutional neural network determines the type and degree of mesh blockage based on the data obtained from the sensor (such as droplet size, spray pattern, etc.). If a mild blockage is detected, the drive parameters need to be adjusted slightly; while for moderate or severe blockages, the adjustment amplitude will increase. Assume the blockage degree detected by the CNN is represented as a quantization value, and the higher the mesh blockage degree, the compensation parameter It will also increase. According to the clogging compensation parameter, the updated power supply parameter combination is corrected. Different degrees of mesh clogging require different compensation strategies. Generally, the more severe the clogging, the greater the compensation intensity. Assume the clogging compensation amount is , then the corrected power supply parameters are:

[0040] The corrected power supply parameter combination can be adjusted according to different mesh clogging situations to ensure that the atomizer can still work stably under mesh clogging. These adjustments ensure that the droplet size remains within the target range, improving the treatment effect. The corrected power supply parameters enter the parameter optimization process. Since the control accuracy and system stability of the atomizer have an important impact on the treatment effect. The goal of parameter optimization is to improve the system response speed and stability to ensure that the particle size always remains within the target control range under changing working environments. The optimization process involves constraining and adjusting parameters to eliminate excessive fluctuations and unnecessary energy waste. Assume an optimization algorithm (such as genetic algorithm, particle swarm algorithm, etc.) is used to optimize the power supply parameters, and the optimized power supply parameters are represented as , and these parameters will be further used for device control after optimization. At the same time, the continuous values after parameter optimization are digitized so that they can be executed by the device. The device can only accept discrete signal inputs, and the optimized continuous parameter values are converted into discrete digital quantities. The digitization process converts the continuous signal into a series of discrete numerical values through quantization, enabling the drive circuit to correctly understand and execute these signals. Through the above steps, a target intermittent power supply scheme is generated.

[0041] Referring to Figure 2 , this embodiment provides a particle size control device for a piezoelectric mesh atomizer, including: Acquisition module 1, configured to acquire the original data set of the piezoelectric mesh atomizer and determine the target particle size control range; Creation module 2, configured to create an initial intermittent power supply scheme including pulse frequency, pulse width, and pulse interval time according to the target particle size control range; Execution module 3, configured to execute the initial intermittent power supply scheme and acquire the current particle size measurement value through a laser scattering sensor; Calculation module 4, configured to calculate the deviation between the current particle size measurement value and the target particle size control range, and generate a drive parameter adjustment instruction through adaptive error compensation; Correction module 5, configured to correct the pulse frequency, pulse width, and pulse interval time in the initial intermittent power supply scheme according to the drive parameter adjustment instruction to obtain the target intermittent power supply scheme.

[0042] In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to the description in the above method embodiment, and details are not described herein again.

[0043] Refer to Figure 3 , in an embodiment of the present invention, a computer device is further provided. The computer device may be a server, and its internal structure may be as Figure 3 shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface, and a database connected through a system bus. Among them, the processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0044] Those skilled in the art can understand that Figure 3 the structure shown in

[0045] is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0046] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the description and drawings of the present invention, or directly or indirectly applied to other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A method for controlling particle size of a piezoelectric network atomizer, characterized in that: The following steps are involved: Collect the original data set of the piezoelectric grid atomizer and determine the target particle size control range; Creating an initial intermittent power supply plan including pulse frequency, pulse width and pulse interval time according to the target particle size control interval; Executing the initial intermittent power supply scheme, collecting current particle size measurement values ​​through a laser scattering sensor; Calculating the deviation between the current particle size measurement value and the target particle size control interval, and generating a drive parameter adjustment instruction through adaptive error compensation; According to the driving parameter adjustment instruction, the pulse frequency, pulse width and pulse interval time in the initial intermittent power supply scheme are corrected to obtain a target intermittent power supply scheme.

2. The particle size control method of the piezoelectric network atomizer according to claim 1, characterized in that: The method of collecting the original data set of the piezoelectric grid atomizer and determining the target particle size control range includes: The piezoelectric oscillation frequency, driving voltage and liquid characteristics data are collected by multiple sensors installed on the piezoelectric network atomizer to generate an original data set; Performing data preprocessing on the original data set to obtain a standardized parameter data set; Constructing a three-dimensional parameter space based on the standardized parameter data set, filling the droplet size distribution data measured under different parameter combination conditions into the corresponding spatial position to form initial mapping relationship data; The initial mapping relationship data is input into a multiple regression analysis model to perform coefficient fitting calculation, and a corresponding relationship matrix between parameters and particle sizes is constructed; Divide the atomization process into the startup phase, the stabilization phase and the end phase, and establish the control priority sequence for each phase; The target particle size is calculated for each stage based on the control priority sequence and the corresponding relationship matrix to form a target particle size control interval.

3. The particle size control method of the piezoelectric network atomizer according to claim 2, characterized in that: The step of creating an initial intermittent power supply scheme including pulse frequency, pulse width and pulse interval time according to the target particle size control interval includes: Extracting target particle size values ​​at the start-up phase, the stable phase and the end phase of the atomization process according to the target particle size control interval to generate a segmented control target sequence; Perform reverse query calculation according to the segmented control target sequence and the corresponding relationship matrix to obtain the initial piezoelectric driving parameter combination corresponding to each stage; Based on the initial piezoelectric driving parameter combination, a piecewise linear response function model is constructed to mathematically describe the mapping relationship between the liquid viscosity and the driving parameters to form a parameter response equation; Performing constrained optimization to solve the parameter response equation, substituting the characteristic parameters of the liquid medicine and the target particle size value into the parameter response equation, and calculating the optimal parameter combination including pulse frequency, pulse width and pulse interval time; Performing atomization effect simulation on the optimal parameter combination to obtain a simulation result, and comparing the simulation result with the target particle size control range to generate a parameter correction amount; The pulse frequency, pulse width and pulse interval are adjusted according to the parameter correction amount, and corresponding power supply parameters are configured for each stage to form an initial intermittent power supply plan.

4. The particle size control method of the piezoelectric network atomizer according to claim 1, characterized in that: The executing of the initial intermittent power supply scheme, collecting the current particle size measurement value through the laser scattering sensor, includes: The initial intermittent power supply scheme is input into the piezoelectric drive control circuit to generate a corresponding pulse signal to drive the piezoelectric network atomizer to work and generate atomized particles of the drug solution; The laser scattering sensor installed at the outlet of the piezoelectric atomizer is used to measure the scattered light of the droplets produced by atomization in real time and collect the original scattered light signal. Performing multi-angle signal acquisition and digital filtering processing on the original scattered light signal to form initial particle size distribution data; The initial particle size distribution data is input into a distributed prescribed time observation system, and data of different particle size ranges are processed in parallel by a plurality of particle size observation nodes to obtain segmented particle size characteristic data; Mathematically processing the segmented particle size characteristic data based on a nonlinear state equation and a time-varying gain matrix, dynamically adjusting the observation gain according to the atomization stage, and generating corrected particle size data; According to the corrected particle size data, the particle size mean and distribution width are calculated, and the current particle size measurement value is output.

5. The particle size control method of the piezoelectric network atomizer according to claim 1, characterized in that: The calculating the deviation between the current particle size measurement value and the target particle size control interval and generating a driving parameter adjustment instruction through adaptive error compensation includes: Calculating the difference between the current particle size measurement value and the target particle size control interval to generate a particle size error evaluation index; Based on the particle size error evaluation index, the error space is divided into a first error region, a second error region and a third error region; The first error region adopts proportional control, the second error region adopts proportional-integral control, and the third error region adopts proportional-integral-differential control to construct a multi-mode control parameter library; Performing correlation analysis on the control parameters in the multi-mode control parameter library and the error variation trend to form an adaptive control parameter matrix; Input the adaptive control parameter matrix and current error data into the error compensation controller, execute the control algorithm calculation of the corresponding mode, and generate an initial compensation amount; The initial compensation amount is optimized and adjusted according to the characteristics of the atomization stage, and different weight coefficients are applied in different atomization stages using a time-varying gain strategy, and a driving parameter adjustment instruction is output.

6. The particle size control method of the piezoelectric network atomizer according to claim 5, characterized in that: The initial compensation amount is optimized and adjusted according to the characteristics of the atomization stage, different weight coefficients are applied in different atomization stages using a time-varying gain strategy, and a driving parameter adjustment instruction is output, including: By analyzing the change trend of droplet size and the change history of power supply parameters, the state of the atomization process is identified to determine whether it is currently in the startup stage, stable stage or end stage, and a stage identification signal is generated; According to the stage identification signal, a corresponding stage weight coefficient is extracted from a preset weight database, a first gain value is set for the startup stage, a second gain value is set for the stable stage, and a third gain value is set for the end stage to form a time-varying weight vector; Decomposing the initial compensation amount into components, decomposing the compensation amount into a frequency compensation component, a pulse width compensation component and a pulse interval compensation component, and generating a compensation component matrix; Performing weighted calculation on the compensation component matrix based on the time-varying weight vector to form a stage adaptive compensation amount, and predicting and verifying the stage adaptive compensation amount through a dynamic response model to generate a target compensation amount; The target compensation amount is converted into a pulse frequency adjustment value, a pulse width adjustment value and a pulse interval adjustment value to form a driving parameter adjustment instruction.

7. The particle size control method of the piezoelectric network atomizer according to claim 1, characterized in that: The step of correcting the pulse frequency, pulse width and pulse interval time in the initial intermittent power supply scheme according to the drive parameter adjustment instruction to obtain a target intermittent power supply scheme includes: Perform parameter update calculation on the initial intermittent power supply scheme according to the drive parameter adjustment instruction, perform addition and subtraction operations on the current power supply parameter and the adjustment amount to form an updated power supply parameter combination; Performing mesh block block state detection on the updated power supply parameter combination through a convolution block attention module and a convolution neural network, evaluating the influence of mesh blockage on atomization effect, and generating blockage compensation parameters; The updated power supply parameter combination is modified according to the congestion compensation parameter, and the power supply parameter is adjusted according to different degrees of mesh congestion to form a congestion-adaptive power supply parameter; The congestion-adaptive power supply parameters are optimized and digitized, and continuous parameter values ​​are converted into discrete digital quantities executable by the device to generate a target intermittent power supply plan.

8. A particle size control device for a piezoelectric network atomizer, characterized in that: For implementing the steps of the method according to any one of claims 1 to 7, the particle size control device of the piezoelectric network atomizer comprises: An acquisition module is used to collect the original data set of the piezoelectric grid atomizer and determine the target particle size control range; A creation module, used to create an initial intermittent power supply plan including pulse frequency, pulse width and pulse interval time according to the target particle size control interval; An execution module, used for executing the initial intermittent power supply scheme, and collecting current particle size measurement values ​​through a laser scattering sensor; A calculation module, used for calculating the deviation between the current particle size measurement value and the target particle size control interval, and generating a driving parameter adjustment instruction through adaptive error compensation; The correction module is used to correct the pulse frequency, pulse width and pulse interval time in the initial intermittent power supply scheme according to the driving parameter adjustment instruction to obtain a target intermittent power supply scheme.

9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

Images