Particle size real-time detection method and system based on closed-loop multi-channel micro-fluidic homogenizer

Abstract

The application discloses a particle size real-time detection method and system based on a closed-loop multi-channel micro-jet homogenizer, and relates to the field of homogenizer particle size analysis. The method comprises the following steps: after high-particle-size materials are sheared and impacted to be low-particle-size materials, multiple channels are used to collect multiple modal parameters at the same time during discharging, and the monitoring data in a continuous time window is constructed into multi-modal time series data; the time series data is input into a multi-modal particle size analysis model pre-trained based on an LSTM network, deep time sequence features are extracted through the multi-modal particle size analysis model for comparison and analysis, and pressure regulation is performed according to the analysis result. Through multi-channel monitoring, real-time particle size analysis is performed by using a multi-modal particle size analysis model, so that the real-time performance of particle size detection is improved; through closed-loop reflux and pressure self-adaptive regulation, material pollution and finished product waste caused by offline sampling are avoided, and the material utilization rate and the control speed of the crushing process are improved.

Description

Technical Field

[0001] This application relates to the field of particle size analysis using homogenizers, and in particular to a method and system for real-time particle size detection based on a closed-loop multi-channel microjet homogenizer. Background Technology

[0002] With the development of industries such as biopharmaceuticals, nanomaterials, and food and dairy products, ultrafine grinding and homogenization emulsification technologies have become crucial factors determining product performance and quality. Especially in the preparation of high-value-added products such as liposomes, nano-suspensions, protein formulations, and functional food emulsions, the size and uniformity of particle size distribution directly affect the product's bioavailability, stability, appearance, and shelf life. Microfluidic homogenizers, with their unique grinding mechanism of high-pressure microfluidic impaction, shearing, and cavitation effects, have become core equipment for preparing submicron and even nanoscale particles. However, traditional microfluidic homogenization processes generally rely on offline sampling and monitoring, with particle size analysis often employing offline detection methods such as microscopic observation, dynamic light scattering instruments, or Coulter counters. This not only results in long detection cycles and high equipment investment, but also, due to the lag between sampling and detection, it fails to detect abnormal particle size fluctuations in a timely manner, leading to the scrapping of large quantities of material. Furthermore, offline sampling carries the risk of external contamination, and the sampled material cannot be recycled, resulting in material waste.

[0003] In summary, particle size detection in existing microfluidic homogenization processes relies on offline sampling and analysis, which suffers from severe detection lag and the inability to provide real-time feedback and control, resulting in significant material loss due to exceeding standards and a high risk of contamination. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for real-time particle size detection based on a closed-loop multi-channel microjet homogenizer. This system can improve the real-time performance of particle size detection by using multi-channel monitoring and a multi-modal particle size analysis model. By using closed-loop reflux and adaptive pressure adjustment, it avoids material contamination and waste of finished products caused by offline sampling, thereby improving material utilization and the control speed of the crushing process.

[0005] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a method for real-time particle size detection based on a closed-loop multi-channel microjet homogenizer. The closed-loop multi-channel microjet homogenizer is used to pulverize high-particle-size materials into low-particle-size materials. Simultaneously, it monitors the multimodal parameters of the low-particle-size materials through multiple channels. After particle size analysis, the pulverization process is controlled based on the analysis results. The method includes: using the closed-loop multi-channel microjet homogenizer to shear and impact high-particle-size materials obtained from an external feed hopper, pulverizing the high-particle-size materials into low-particle-size materials; simultaneously, while the low-particle-size materials are being discharged, multimodal parameters of the low-particle-size materials are monitored in real-time through multiple channels, and the monitoring results are recorded. The low-size material is recirculated and pressurized for re-crushing; the multimodal parameters include light scattering intensity, particle velocity, and electromobility; the multimodal parameters collected in a continuous time window are used to construct multimodal time-series data; the multimodal time-series data is input into a trained multimodal particle size analysis model for particle size analysis to obtain particle size analysis results; the multimodal particle size analysis model is constructed based on an LSTM network; if the particle size analysis result is less than or equal to the particle size threshold, the low-size material is kept discharged; otherwise, the discharge of low-size material is stopped, and the pressure of the closed-loop multi-channel microfluidic homogenizer is adjusted until the particle size analysis result is less than or equal to the particle size threshold.

[0006] Optionally, the multimodal parameters collected within a continuous time window are constructed into multimodal time-series data. Specifically, this includes: acquiring the original sampling sequences of the multimodal parameters within the continuous time window, and sorting the original sampling sequences of the multimodal parameters according to timestamps to obtain light scattering intensity time series, particle velocity time series, and electromobility time series, respectively; aligning the time axes of the light scattering intensity time series, particle velocity time series, and electromobility time series; and fusing the aligned light scattering intensity time series, particle velocity time series, and electromobility time series in chronological order to obtain multimodal time-series data.

[0007] Optionally, the training process of the trained multimodal particle size analysis model is as follows: Standard particle samples of different particle sizes are obtained, and time series of light scattering intensity, particle velocity, and electromobility corresponding to each standard particle sample are collected to construct an original multimodal time series database; the standard particle samples are calibrated using a particle size measurement device to obtain the true particle size value corresponding to each time series, a labeled supervised learning training sample set is constructed, and the supervised learning training sample set is divided into a training set, a validation set, and a test set according to a preset ratio; the initial multimodal particle size analysis model is iteratively trained using the training set, with the goal of minimizing the loss function, to obtain the trained multimodal particle size analysis model.

[0008] Optionally, the loss function is the mean squared error loss function or the cross-entropy loss function.

[0009] Optionally, the multimodal particle size analysis model includes an input layer, an LSTM hidden layer, a feature fusion layer, and a fully connected layer connected in sequence.

[0010] Optionally, the high-particle-size material is a material with a particle size greater than or equal to 400 μm and less than or equal to 500 μm; the low-particle-size material is a material with a particle size less than or equal to 200 nm.

[0011] Optionally, the closed-loop multi-channel microjet homogenizer includes: a pressure pump, a diamond homogenizing chamber, a cooler, a discharge unit, a discharge electric valve, a reflux electric valve, a particle size analysis unit, a first monitoring channel, a second monitoring channel, and a third monitoring channel; the inlet of the pressure pump is connected to an external feed hopper pipeline, and the outlet of the pressure pump is connected to the inlet pipeline of the diamond homogenizing chamber; the outlet of the diamond homogenizing chamber is respectively connected to the inlet of the discharge unit, the inlet of the first monitoring channel, the inlet of the second monitoring channel, and the third monitoring channel. The inlet pipe of the channel is connected, the cooling water outlet of the diamond homogenizing chamber is connected to the inlet pipe of the cooler, and the cooling water inlet of the diamond homogenizing chamber is connected to the outlet pipe of the cooler; the outlets of the first monitoring channel, the second monitoring channel, and the third monitoring channel are all connected to the inlet pipe of the pressurizing pump; the first monitoring channel is used to monitor the light scattering intensity of the material; the second monitoring channel is used to monitor the particle velocity of the material; and the third monitoring channel is used to monitor the electromobility of the material; The first outlet of the discharge unit is connected to the external storage silo pipeline, and the second outlet of the discharge unit is connected to the inlet pipeline of the pressure pump; the discharge electric valve is installed on the pipeline between the first outlet of the discharge unit and the external storage silo; the reflux electric valve is installed on the pipeline between the second outlet of the discharge unit and the inlet of the pressure pump; the first monitoring channel, the second monitoring channel, and the third monitoring channel are all electrically connected to the particle size analysis unit; the particle size analysis unit is connected to the pressure pump and the discharge electric valve respectively. The valve and the reflux electric valve are electrically connected; the particle size analysis unit is used to perform particle size analysis on low-particle-size materials using light scattering intensity, particle velocity, and electromobility as multimodal parameters to obtain analysis results; if the particle size analysis result is less than or equal to the particle size threshold, the discharge electric valve is controlled to open to maintain discharge; otherwise, the discharge electric valve is controlled to close to stop the discharge of low-particle-size materials, and the reflux electric valve is controlled to open to return the low-particle-size materials to the pressurizing pump, while the pressurizing pump is controlled to adjust the pressure until the particle size analysis result is less than or equal to the particle size threshold.

[0012] Optionally, the discharge electric valve and the return electric valve are interlocked.

[0013] Optionally, the type of the diamond homogeneous cavity includes at least one of the following: a Y-shaped cavity and a Z-shaped cavity.

[0014] In a second aspect, this application also provides a computer system, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the real-time particle size detection method based on a closed-loop multi-channel microfluidic homogenizer as described in the first aspect.

[0015] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application uses multi-channel synchronous acquisition of multimodal time-series data such as light scattering intensity, particle velocity, and electromobility, and uses a multimodal particle size analysis model for real-time particle size analysis, which improves the real-time performance of particle size detection. Through closed-loop reflux and adaptive pressure adjustment, the material is recirculated and pressurized for further crushing after monitoring, without the need to stop the discharge or separately collect the finished product. This avoids material contamination and waste of finished product caused by offline sampling, improves material utilization and control response speed of the crushing process, and reduces the amount of defective products generated. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic flowchart of a real-time particle size detection method based on a closed-loop multi-channel microfluidic homogenizer provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the structural connection of a closed-loop multi-channel microfluidic homogenizer provided in an embodiment of this application.

[0019] Figure 3 This is a schematic diagram of the internal structure of the diamond homogenizing cavity provided in an embodiment of this application.

[0020] Figure 4 This is a schematic diagram of the structure of the cooler provided in the embodiment of this application, which is disposed between the diamond homogenization chamber and the discharge unit.

[0021] Figure 5 This is an internal structure diagram of a computer system provided in an embodiment of this application.

[0022] Reference numerals: 1. Pressure pump; 2. Diamond homogenizing chamber; 3. Cooler; 4. Discharge unit; 5. Discharge electric valve; 6. Recirculation electric valve; 7. Particle size analysis unit; 8. First monitoring channel; 9. Second monitoring channel; 10. Third monitoring channel. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] Example 1, as Figure 1 As shown, this embodiment provides a real-time particle size detection method based on a closed-loop multi-channel microjet homogenizer. The closed-loop multi-channel microjet homogenizer is used to pulverize high-particle-size materials into low-particle-size materials. Simultaneously, it monitors the multimodal parameters of the low-particle-size materials through multiple channels. After particle size analysis, the pulverization process is controlled based on the particle size analysis results. The method includes: S1. The closed-loop multi-channel micro-jet homogenizer is used to shear and impact the high-particle-size material obtained from the external feed hopper, crushing the high-particle-size material into low-particle-size material.

[0026] Optionally, high-particle-size materials are those with a particle size larger than that of low-particle-size materials.

[0027] Furthermore, the high-particle-size material is a material with a particle size greater than or equal to 400 μm and less than or equal to 500 μm; the low-particle-size material is a material with a particle size less than or equal to 200 nm.

[0028] In practical applications, the high particle size material is a nano-suspension, liposome, and nano-carbon powder dispersion.

[0029] S2. While the low-particle-size material is being discharged, the low-particle-size material is monitored in real time through multiple channels for multi-modal parameters, and the monitored low-particle-size material is recirculated and pressurized for re-crushing; the multi-modal parameters include: light scattering intensity, particle velocity and electromobility.

[0030] In practical applications, the roles of each modal parameter in particle size analysis are as follows: 1) Light scattering intensity is an optical response parameter characterizing particle size. When a laser or other monochromatic beam passes through a suspension or emulsion containing particles, the particle surface and internal structure scatter the incident light. The distribution of scattered light intensity and its angular intensity have a definite functional relationship with the equivalent spherical diameter of the particle. Large particles tend to produce strong forward scattering at small angles, while small particles produce weak scattering in the backward or wide-angle directions. Therefore, by real-time detection of the scattered light intensity time-series signal at specific angles (such as forward, side, or backward), changes in particle size in the material can be captured.

[0031] 2) Particle velocity is a dynamic parameter characterizing the motion behavior of particles in a fluid, and its relationship with particle size is based on the principles of fluid mechanics and inertial separation. When homogenized material passes through the monitoring channel at a constant flow rate, particles of different sizes exhibit different following response characteristics to changes in the flow field due to differences in mass, inertia, and fluid drag.

[0032] Specifically, large-diameter particles, due to their large mass and strong inertia, will lag in velocity response when the flow field accelerates or changes direction, and there is a measurable slip velocity between their motion velocity and the continuous phase fluid; while small-diameter particles, with their light mass and low inertia, are more likely to follow the fluid motion, and their velocity is closer to the average flow velocity of the fluid.

[0033] By detecting the migration speed of particles within the monitoring channel, such as using laser Doppler velocimetry or particle image velocimetry, the equivalent hydrodynamic diameter of the particles can be indirectly inferred.

[0034] 3) Electromobility characterizes the ability of charged particles to migrate under the influence of an applied electric field. Its relationship with particle size is based on the coupled balance of Coulomb force, fluid resistance, and particle charge. After applying a stable DC or AC electric field in the monitoring channel, particles with net surface charge in the material will migrate towards the opposite polarity electrode under the drive of the electric field force. The smaller the particle size, the larger the specific surface area, and the higher the charge density, the faster the migration speed under the same electric field strength; the larger the particle size, the more linear or quadratic the increase in fluid resistance, and the lower the mobility.

[0035] By detecting the steady-state migration velocity of particles under a known electric field strength, the electromobility distribution of the particles can be obtained, and then the equivalent particle size can be calculated.

[0036] S3. Construct multimodal time series data from the multimodal parameters collected in continuous time windows.

[0037] Furthermore, step S3 specifically includes: S31. Obtain the original sampling sequence of multimodal parameters within a continuous time window, and sort the original sampling sequence of multimodal parameters according to the timestamp to obtain the light scattering intensity time series, particle velocity time series and electromobility time series respectively.

[0038] S32. Align the time axes of the light scattering intensity time series, particle velocity time series, and electromobility time series.

[0039] S33. The aligned light scattering intensity time series, particle velocity time series and electromobility time series are fused in chronological order to obtain multimodal time series data.

[0040] S4. Input the multimodal time series data into the trained multimodal particle size analysis model for particle size analysis to obtain the particle size analysis results; the multimodal particle size analysis model is constructed based on an LSTM network.

[0041] Furthermore, the training process of the trained multimodal particle size analysis model is as follows: Standard particle samples of different particle sizes are obtained, and time series of light scattering intensity, particle velocity, and electromobility corresponding to each standard particle sample are collected to construct an original multimodal time series database; the standard particle samples are calibrated using a particle size measurement device to obtain the true particle size value corresponding to each time series, and a labeled supervised learning training sample set is constructed, which is then divided into a training set, a validation set, and a test set according to a preset ratio; the initial multimodal particle size analysis model is iteratively trained using the training set, with the goal of minimizing the loss function, to obtain the trained multimodal particle size analysis model.

[0042] Furthermore, the loss function is either the mean squared error loss function or the cross-entropy loss function.

[0043] In practical applications, the multimodal particle size analysis model is a deep learning model built and trained based on an LSTM network. Utilizing the recurrent gating mechanism of LSTM, the model dynamically extracts features and models long-term dependencies from three temporal signals—light scattering intensity, particle velocity, and electromobility—within a continuous time window, capturing the temporal correlations in particle size evolution. Simultaneously, it integrates the hidden states of the LSTM through a feature fusion layer and outputs real-time particle size predictions via a fully connected layer. During runtime, the model loads pre-trained weights, receives multi-channel real-time data streams, and obtains particle size analysis results after a single forward inference, enabling real-time feedback and control of the crushing process.

[0044] In addition, the multimodal particle size analysis model can also perform particle size analysis based on light scattering intensity, particle velocity, and electromobility using the entropy weight method.

[0045] S5. If the particle size analysis result is less than or equal to the particle size threshold, continue to discharge low-particle-size material; otherwise, stop the discharge of low-particle-size material and adjust the pressure of the closed-loop multi-channel micro-jet homogenizer until the particle size analysis result is less than or equal to the particle size threshold.

[0046] Furthermore, the multimodal particle size analysis model includes an input layer, an LSTM hidden layer, a feature fusion layer, and a fully connected layer connected in sequence.

[0047] As an optional implementation method, such as Figures 2-4 As shown, the closed-loop multi-channel microjet homogenizer includes: a pressure pump 1, a diamond homogenizing chamber 2, a cooler 3, a discharge unit 4, a discharge electric valve 5, a reflux electric valve 6, a particle size analysis unit 7, a first monitoring channel 8, a second monitoring channel 9, and a third monitoring channel 10.

[0048] 1) The inlet of the pressurizing pump 1 is connected to the external feed hopper pipeline, and the outlet of the pressurizing pump 1 is connected to the inlet pipeline of the diamond homogenizing chamber 2.

[0049] 2) The outlet of the diamond homogenizing chamber 2 is connected to the inlet of the discharge unit 4, the inlet of the first monitoring channel 8, the inlet of the second monitoring channel 9 and the inlet of the third monitoring channel 10 respectively. The cooling water outlet of the diamond homogenizing chamber 2 is connected to the inlet of the cooler 3. The cooling water inlet of the diamond homogenizing chamber 2 is connected to the outlet of the cooler 3.

[0050] Furthermore, the diamond homogenizing cavity 2 includes at least one of the following types: a Y-shaped cavity and a Z-shaped cavity. The Y-shaped cavity is more suitable for processing liquid-liquid materials such as emulsions, drug encapsulation, and liposome preparation; the Z-shaped cavity is more suitable for processing solid-liquid materials such as cell disruption or nano-dispersion, deagglomeration, particle size reduction, and mechanical exfoliation. In practical applications, different cavities can be configured according to requirements.

[0051] As an optional implementation, the cooler 3 can also be installed on the pipeline between the diamond homogenizing chamber 2 and the discharge unit 4, specifically as follows: the liquid outlet of the diamond homogenizing chamber is connected to the liquid inlet pipeline of the cooler 3; the liquid outlet of the cooler 3 is connected to the liquid inlet pipeline of the discharge unit 4, the liquid inlet of the first monitoring channel 8, the liquid inlet of the second monitoring channel 9, and the liquid inlet pipeline of the third monitoring channel 10.

[0052] 3) The outlet of the first monitoring channel 8, the outlet of the second monitoring channel 9, and the outlet of the third monitoring channel 10 are all connected to the inlet pipe of the pressurizing pump 1; the first monitoring channel 8 is used to monitor the light scattering intensity of the material; the second monitoring channel 9 is used to monitor the particle velocity of the material; and the third monitoring channel 10 is used to monitor the electromobility of the material.

[0053] 4) The first liquid outlet of the discharge unit 4 is connected to the external storage silo pipeline, and the second liquid outlet of the discharge unit 4 is connected to the liquid inlet pipeline of the pressure pump 1.

[0054] 5) The discharge electric valve 5 is installed on the pipeline between the first liquid outlet of the discharge unit 4 and the external storage bin.

[0055] 6) The reflux electric valve 6 is installed on the pipeline between the second liquid outlet of the discharge unit 4 and the liquid inlet of the pressure pump 1.

[0056] 7) The first monitoring channel 8, the second monitoring channel 9, and the third monitoring channel 10 are all electrically connected to the particle size analysis unit 7; the particle size analysis unit 7 is electrically connected to the pressurizing pump 1, the discharge electric valve 5, and the return electric valve 6, respectively; the particle size analysis unit 7 is used to perform particle size analysis on low-particle-size materials using light scattering intensity, particle velocity, and electromobility as multimodal parameters to obtain analysis results; if the particle size analysis result is less than or equal to the particle size threshold, the discharge electric valve 5 is controlled to open to maintain discharge; otherwise, the discharge electric valve 5 is controlled to close to stop the discharge of low-particle-size materials, and the return electric valve 6 is controlled to open to return the low-particle-size materials to the pressurizing pump 1, while the pressurizing pump 1 is controlled to adjust the pressure until the particle size analysis result is less than or equal to the particle size threshold.

[0057] Furthermore, the discharge electric valve 5 and the return electric valve 6 are interlocked. (This ensures that only one valve is open at any given time, preventing discharge and return from occurring simultaneously.)

[0058] In practical applications, the operation process of a closed-loop multi-channel microfluidic homogenizer is as follows: 1) The inlet of the pressurizing pump 1 draws in high-particle-size material from the external feed hopper and pressurizes the high-particle-size material before conveying it to the diamond homogenizing chamber 2.

[0059] 2) Start the cooler 3 to make the cooling water circulate between the cooling water inlet and outlet of the diamond homogenizing chamber 2 and the cooler 3 to cool the homogenizing chamber.

[0060] 3) After the high-pressure material enters the diamond homogenizing chamber 2, it is crushed into low-particle size material by high-speed micro-jet impact, shearing and cavitation in the diamond homogenizing chamber 2. The crushed low-particle size material is discharged from the liquid outlet of the diamond homogenizing chamber 2 and simultaneously diverted to the discharge unit 4, the first monitoring channel 8, the second monitoring channel 9 and the third monitoring channel 10.

[0061] 4) The light scattering intensity of the low-particle-size material diverted to the first monitoring channel 8 is detected through this channel; the particle velocity of the low-particle-size material diverted to the second monitoring channel 9 is detected; and the electromobility of the low-particle-size material diverted to the third monitoring channel 10 is detected.

[0062] 5) After the materials in the first monitoring channel 8, the second monitoring channel 9 and the third monitoring channel 10 are detected, they are discharged from their respective outlets and converge into the inlet of the pressurizing pump 1 to form a backflow.

[0063] 6) The first monitoring channel 8, the second monitoring channel 9 and the third monitoring channel 10 transmit the light scattering intensity, particle velocity and electromobility signals detected by each channel to the particle size analysis unit 7 in real time.

[0064] 7) The particle size analysis unit 7 receives multimodal parameters and performs real-time particle size analysis based on the built-in multimodal particle size analysis model to obtain the particle size analysis results of the current low particle size material.

[0065] 8) The particle size analysis unit 7 compares the obtained particle size analysis results with the preset particle size threshold.

[0066] If the particle size analysis result is less than or equal to the particle size threshold, the particle size analysis unit 7 controls the discharge electric valve 5 to open, while keeping the return electric valve 6 closed, so that the low particle size material at the first liquid outlet of the discharge unit 4 is transported to the external storage silo through the pipeline, thus achieving the discharge of qualified finished products.

[0067] 9) If the particle size analysis result is greater than the particle size threshold, the particle size analysis unit 7 controls the discharge electric valve 5 to close, stops the discharge to the external storage bin, and controls the return electric valve 6 to open, so that the low particle size material at the second liquid outlet of the discharge unit 4 flows back to the liquid inlet of the pressure pump 1 through the pipeline; at the same time, the particle size analysis unit 7 sends a pressure adjustment command to the pressure pump 1 to increase the pressure of the pressure pump 1 (thereby enhancing the shearing and impact force in the diamond homogenization chamber 2).

[0068] 10) The material, after being repressurized by the pressurizing pump 1, re-enters the diamond homogenizing chamber 2 for secondary crushing and is then monitored and analyzed again.

[0069] 11) Until the particle size analysis unit 7 determines that the current particle size analysis result meets the condition of being less than or equal to the particle size threshold, the normal discharge state of low particle size material is switched back.

[0070] The technical effects of this application are as follows: This application uses multi-channel synchronous acquisition of multimodal time-series data such as light scattering intensity, particle velocity, and electromobility, and uses a multimodal particle size analysis model for real-time particle size analysis, which improves the real-time performance of particle size detection. Through closed-loop reflux and adaptive pressure adjustment, the material is recirculated and pressurized for further crushing after monitoring, without the need to stop the discharge or separately collect the finished product. This avoids material contamination and waste of finished product caused by offline sampling, improves material utilization and control response speed of the crushing process, and reduces the amount of defective products generated.

[0071] Example 2: This example provides a computer system, which can be a server or a terminal, and its internal structure diagram can be as follows. Figure 5 As shown, the computer system includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and databases. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media to run. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements the methods described above.

[0072] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer system to which the present application is applied. A specific computer system may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0073] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0074] The processors involved in the various embodiments provided in this application may be general-purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited thereto.

[0075] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0076] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer, characterized in that, The closed-loop multi-channel microjet homogenizer is used to pulverize high-particle-size materials into low-particle-size materials. Simultaneously, it monitors the multimodal parameters of the low-particle-size materials through multiple channels. After particle size analysis, the pulverization process is controlled based on the particle size analysis results. The method includes: The closed-loop multi-channel micro-jet homogenizer is used to shear and impact high-particle-size materials obtained from an external feed hopper, crushing the high-particle-size materials into low-particle-size materials. While the low-particle-size material is being discharged, its multi-modal parameters are monitored in real time through multiple channels. The monitored low-particle-size material is then recirculated and pressurized for re-crushing. The multi-modal parameters include light scattering intensity, particle velocity, and electromobility. The multimodal parameters collected in a continuous time window are used to construct multimodal time series data; Multimodal time-series data is input into a trained multimodal particle size analysis model for particle size analysis to obtain particle size analysis results; the multimodal particle size analysis model is constructed based on an LSTM network. If the particle size analysis result is less than or equal to the particle size threshold, then continue to discharge low-particle-size material; otherwise, stop discharging low-particle-size material and adjust the pressure of the closed-loop multi-channel micro-jet homogenizer until the particle size analysis result is less than or equal to the particle size threshold.

2. The method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer according to claim 1, characterized in that, The multimodal parameters collected within a continuous time window are used to construct multimodal time series data, specifically including: The original sampling sequences of multimodal parameters within a continuous time window are obtained, and the original sampling sequences of multimodal parameters are sorted according to timestamps to obtain the light scattering intensity time series, particle velocity time series and electromobility time series, respectively. Align the time axes of the light scattering intensity time series, particle velocity time series, and electromobility time series; The aligned light scattering intensity time series, particle velocity time series, and electromobility time series are fused in chronological order to obtain multimodal time series data.

3. The method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer according to claim 2, characterized in that, The training process of the trained multimodal particle size analysis model is as follows: Standard particle samples of different particle sizes were obtained, and time series of light scattering intensity, particle velocity and electromobility corresponding to each standard particle sample were collected to construct the original multimodal time series database. Standard particle samples are calibrated using particle size measurement equipment to obtain the true particle size values ​​for each time series, a labeled supervised learning training sample set is constructed, and the supervised learning training sample set is divided into training set, validation set and test set according to a preset ratio; The initial multimodal particle size analysis model is iteratively trained using the training set, with the goal of minimizing the loss function, to obtain a well-trained multimodal particle size analysis model.

4. The method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer according to claim 3, characterized in that, The loss function is either the mean squared error loss function or the cross-entropy loss function.

5. The method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer according to claim 1, characterized in that, The multimodal particle size analysis model includes an input layer, an LSTM hidden layer, a feature fusion layer, and a fully connected layer connected in sequence.

6. The method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer according to claim 1, characterized in that, The high-particle-size material is a material with a particle size greater than or equal to 400 μm and less than or equal to 500 μm; the low-particle-size material is a material with a particle size less than or equal to 200 nm.

7. The method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer according to claim 1, characterized in that, The closed-loop multi-channel micro-jet homogenizer includes: a pressure pump, a diamond homogenization chamber, a cooler, a discharge unit, a discharge electric valve, a reflux electric valve, a particle size analysis unit, a first monitoring channel, a second monitoring channel, and a third monitoring channel; The inlet of the pressurizing pump is connected to the external feed hopper pipeline, and the outlet of the pressurizing pump is connected to the inlet pipeline of the diamond homogenizing chamber. The outlet of the diamond homogenizing chamber is connected to the inlet of the discharge unit, the inlet of the first monitoring channel, the inlet of the second monitoring channel, and the inlet of the third monitoring channel, respectively. The cooling water outlet of the diamond homogenizing chamber is connected to the inlet of the cooler, and the cooling water inlet of the diamond homogenizing chamber is connected to the outlet of the cooler. The outlets of the first monitoring channel, the second monitoring channel, and the third monitoring channel are all connected to the inlet pipe of the pressurizing pump; the first monitoring channel is used to monitor the light scattering intensity of the material; the second monitoring channel is used to monitor the particle velocity of the material; and the third monitoring channel is used to monitor the electromobility of the material. The first liquid outlet of the discharge unit is connected to the external storage silo pipeline, and the second liquid outlet of the discharge unit is connected to the liquid inlet pipeline of the pressurizing pump. The discharge electric valve is installed on the pipeline between the first liquid outlet of the discharge unit and the external storage silo. The reflux electric valve is installed on the pipeline between the second liquid outlet of the discharge unit and the liquid inlet of the pressure pump; The first monitoring channel, the second monitoring channel, and the third monitoring channel are all electrically connected to the particle size analysis unit; the particle size analysis unit is electrically connected to the pressurizing pump, the discharge electric valve, and the reflux electric valve, respectively; the particle size analysis unit is used to perform particle size analysis on low-particle-size materials using light scattering intensity, particle velocity, and electromobility as multimodal parameters to obtain analysis results; if the particle size analysis result is less than or equal to the particle size threshold, the discharge electric valve is controlled to open to maintain discharge; otherwise, the discharge electric valve is controlled to close to stop the discharge of low-particle-size materials, and the reflux electric valve is controlled to open to return the low-particle-size materials to the pressurizing pump, while the pressurizing pump is controlled to adjust the pressure until the particle size analysis result is less than or equal to the particle size threshold.

8. The method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer according to claim 7, characterized in that, The discharge electric valve and the return electric valve are interlocked.

9. The method for real-time particle size detection based on a closed-loop multi-channel microfluidic homogenizer according to claim 7, characterized in that, The type of the diamond homogeneous cavity includes at least one of the following: Y-type cavity and Z-type cavity.

10. A computer system, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the real-time particle size detection method based on a closed-loop multi-channel microfluidic homogenizer as described in any one of claims 1-9.

Images