Dynamic optimization control method and system for multi-component batching of paste making machine

Abstract

The invention belongs to the technical field of industrial automation control, and particularly relates to a dynamic optimization control method and system for multi-component batching of a paste making machine, and the method comprises the steps: generating a characteristic waveform spectrum corresponding to a residue molecular structure; dynamically generating electric lubricating medium parameters containing the electroosmotic liquid concentration and the release rate; triggering a low-frequency vibration module of the material layer to generate vibration waves, and enabling the solution and the vibration waves to form a molecular-level synergistic stripping effect; calculating a nanoparticle dispersity index, and generating an inverse gradient mixing instruction when the index exceeds a threshold value; activating a hot vortex ring in a spiral flow guide convex pattern area and establishing a directional heat energy migration channel; and verifying the dynamic equilibrium state of the temperature of the paste core area and the viscoelasticity of the target formula, and triggering a phase lock control signal to terminate the parameter adjustment chain when the energy consumption rate reaches a stable interval. According to the invention, the problems that the transition period during formula switching is prolonged and the phase differences among the execution mechanisms are difficult to match in real time due to a traditional staged regulation and control mode are solved.

Description

Technical Field

[0001] The invention belongs to the technical field of industrial automation control and relates to a dynamic optimization control method and system for multi-component batching of a paste making machine. Background Art

[0002] In the process of mixing multi-component materials in the paste making machine, traditional technologies generally face the risk of cross-contamination caused by residue adhesion. Molecular-level residues are easily formed on the surface of the stirring component due to differences in material properties. The existing method uses regular shutdown and disassembly cleaning or single solvent flushing, but it is difficult to identify the residue components in real time and accurately match the cleaning medium. In addition, manual intervention in cleaning not only reduces production efficiency, but also has the problem of excessive damage to the equipment surface or incomplete cleaning due to the solidification of cleaning parameters.

[0003] Traditional solutions rely on fixed speed and mixing timing control strategies. It is difficult for the stirring parameters preset by experience to dynamically adapt to changes in the rheological properties of materials. In terms of nanoparticle dispersion control, relying solely on torque feedback or timed sampling analysis results in the inability to timely predict the risk of particle agglomeration. The lack of synergy between mechanical vibration and solution release, and the fixed heat conduction path can easily cause local temperature instability, further exacerbating uneven dispersion and viscoelastic imbalance. This type of parameter hysteresis adjustment method is difficult to meet the high-precision paste quality requirements, resulting in energy waste and reduced yield.

[0004] Based on the above problems, the traditional phased control mode leads to a longer transition period when switching recipes, and the phase difference between the actuators is difficult to match in real time. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a method and system for dynamically optimizing the multi-component batching of a paste making machine.

[0006] In the first aspect, the present invention provides a method for dynamically optimizing the multi-component batching of a paste making machine, using the following technical solutions: The method for dynamically optimizing and controlling the multi-component batching of a paste making machine comprises the following steps: S1, obtaining the electromagnetic fingerprint signal of the residue on the surface of the stirring component, collecting the current phase offset through the polarization probe array, and generating a characteristic waveform spectrum corresponding to the molecular structure of the residue; S2, dynamically generating electric lubrication medium parameters including electroosmotic fluid concentration and release rate based on matching the characteristic waveform spectrum with a preset formula switching database; S3, controlling the microporous ceramic tube to release the polar solution according to the electric lubrication medium parameters, triggering the low-frequency vibration module of the material layer to generate vibration waves, so that the solution and the vibration waves form a molecular-level synergistic exfoliation effect; S4, collecting acoustic impedance change data and stirring torque value in the mixing chamber, calculating the nanoparticle dispersion index, and generating a reverse gradient mixing instruction when the index exceeds a threshold; S5, executing the reverse gradient mixing instruction to adjust the stirring speed and the extrusion screw advancement phase difference, activate the thermal vortex ring in the spiral guide convex pattern area and establish a directional thermal energy migration channel; S6. Verify the dynamic equilibrium state between the paste core area temperature and the target formula viscoelasticity. When the energy consumption rate reaches a stable range, trigger a phase-locked control signal to terminate the parameter adjustment chain.

[0007] A further solution of the present invention generates a characteristic waveform spectrum corresponding to the molecular structure of the residue, comprising the following steps: Applying a dynamic scanning voltage to the polarization probe array and collecting a current phase offset; Perform differential preprocessing on the current phase offset to eliminate interference signals; The pre-processed phase offset is input into the Gram angular field converter for time conversion, and a two-dimensional gray matrix reflecting the molecular orientation characteristics is output; The two-dimensional grayscale matrix is ​​encoded into a pseudo-color waveform spectrum as a characteristic waveform spectrum.

[0008] A further solution of the present invention is to dynamically generate the electrolubricating medium parameters including the electroosmotic fluid concentration and the release rate, comprising the following steps: Perform multimodal decomposition on the characteristic waveform spectrum and construct a three-dimensional feature matrix including waveform slope and extreme point density; The chemical bond types of the three-dimensional feature matrix are analyzed through the pre-trained bond type recognition model, and the standard bond type vectors in the formula switching database are matched; The microfluidic chip controller is called to calculate the electroosmotic fluid concentration gradient curve and pulse release interval, and generate a control sequence of the electrolubricating medium parameters.

[0009] A further solution of the present invention generates a control sequence of electric lubrication medium parameters, comprising the following steps: Inject the parameters of the electrolubricating medium into the multi-physics coupling simulation environment to verify the molecular bond breaking efficiency of electroosmosis and mechanical vibration; When the product of the vibration energy transfer function and the solution diffusion rate reaches a critical value, the final parameters are output to the microfluidic chip controller.

[0010] A further solution of the present invention triggers the molecular-level synergistic exfoliation effect, comprising the following steps: The parameters of the electrolubricating medium are converted into a differential pressure control signal of the microporous ceramic tube to drive the polar solution to penetrate through the spiral microchannel. After a preset delay, the piezoelectric ceramic actuator array at the bottom of the material layer is activated to generate a vibration wave in the form of a composite standing wave; The concentration of detached particles is monitored by an optical turbidity sensor to determine the molecular level of exfoliation completion and feed back to the mixing chamber.

[0011] A further solution of the present invention generates an inverse gradient mixing instruction, comprising the following steps: The propagation time difference and amplitude attenuation of the sound waves in the mixing cavity are collected by a piezoelectric sensor group; The acoustic-mechanical coupling dynamic parameter model was constructed in combination with the stirring torque fluctuation data to calculate the instantaneous diffusion rate and agglomeration risk level of nanoparticles; When the diffusion rate is lower than the process standard and the agglomeration risk level exceeds the threshold, the reverse gradient mixing logic controller is triggered to output the adjustment instruction set.

[0012] A further solution of the present invention is to establish a directional heat energy migration channel, comprising the following steps: A pulse electric field is applied to the variable heat-conducting film layer on the surface of the spiral flow-guiding convex pattern to generate asymmetric heat conduction characteristics to drive a tornado-like circulation; The temperature difference distribution in the core area of ​​the paste is monitored by a thermosensitive optical fiber array, and the optimal heat conduction path is determined and the eddy current speed is fixed using a fuzzy clustering algorithm; When the extrusion screw advancement phase difference and the thermal vortex ring speed reach the synchronization critical point, the stirring speed reduction rate and the heat conduction gradient value are jointly controlled.

[0013] A further solution of the present invention verifies the dynamic equilibrium state between the paste core area temperature and the target formula viscoelasticity, comprising the following steps: The temperature data of the longitudinal section of the paste is collected through a temperature gradient sensor array and input into a convolutional neural network to generate a thermodynamic state prediction diagram; Obtain the rheological parameters of the paste and input them into the dynamic recursive equation to calculate the viscoelastic energy dissipation rate; When the energy dissipation rate changes below the set threshold for three consecutive iterations, the thermodynamic state prediction diagram is topologically matched with the inclination parameters of the spiral guide convex pattern to trigger the phase locking signal.

[0014] A further solution of the present invention triggers a phase locking signal, comprising the following steps: Verify the standard deviation decay rate of temperature gradient and the spatial coupling coefficient of viscoelastic parameters; If the matching degree exceeds the critical value, a composite phase-locked control signal including synchronization pulses and power spectrum density is sent to the PLC controller.

[0015] In a second aspect, the present invention provides a dynamic optimization control system for multi-component batching of a paste making machine, which adopts the following technical solution: A polarization probe array module, based on a ring-shaped microelectrode group embedded on the surface of the stirring assembly, is configured to collect current phase offsets by dynamically scanning voltage and generate characteristic waveform spectra that characterize the molecular structure of the residue; A formula dynamic analysis module, which is in communication with the polarization probe array module and is configured to analyze the chemical bond type of the characteristic waveform spectrum and generate electric lubrication medium parameters including electroosmotic fluid concentration and release rate based on a preset formula switching database; The solution-vibration synergistic stripping module controls the microporous ceramic tube to release polar solution according to the parameters of the electric lubrication medium, triggers the low-frequency vibration module of the material layer to generate vibration waves, so that the solution and the vibration waves form a molecular-level synergistic stripping effect; The dispersion determination module is based on the piezoelectric ceramic sensor group and torque sensor deployed in the mixing cavity, and is configured to calculate the coupled dynamic parameter model of acoustic impedance data and real-time torque, and output the nanoparticle dispersion index and the inverse gradient mixing instruction; The reverse gradient mixing control module is linked with the dispersion determination module and the extrusion screw driver, and is configured to adjust the stirring speed and the screw advancement phase difference, and synchronously trigger the asymmetric heat conduction film layer in the spiral guide convex pattern area to establish a directional heat energy migration channel; The dynamic balance phase-locking module, including a temperature gradient sensor array and an online rheometer, is configured to monitor the temperature-viscoelastic coupling state of the paste core area and the thermal vortex ring, and generate a phase-locking control signal through a convolutional neural network and a dynamic recursive equation to terminate the parameter adjustment chain.

[0016] In summary, the present invention includes the following beneficial technical effects: 1. Through the dynamically generated nanoparticle dispersion index and inverse gradient mixing instructions, combined with the acoustic-mechanical coupling dynamic parameter model, high-precision control of the microscopic state of the material in the mixing chamber is achieved, effectively solving the problems of particle agglomeration and uneven dispersion in traditional mixing processes.

[0017] 2. Relying on the thermal vortex ring and directional thermal energy migration channel in the spiral guide convex pattern area, a dynamic balance mechanism of temperature gradient and viscoelasticity is established to synchronously coordinate mechanical action and thermodynamic response to achieve stable range control of energy consumption rate.

[0018] 3. Through the phase-locked control signal and multi-source data fusion algorithm, the temperature-viscoelastic dynamic equilibrium state of the paste core area is verified in real time, forming a closed-loop parameter adjustment chain to ensure the system's adaptability under different formulations and working conditions, reducing the need for manual intervention and process calibration errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. The drawings are used to provide a further understanding of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 A flow chart of a dynamic optimization control method for multi-component batching of an ointment making machine is disclosed.

[0021] Figure 2 It is a characteristic waveform spectrum corresponding to the molecular structure of the residue in the embodiment of the present application.

[0022] Figure 3 The structural schematic diagram of the dynamic optimization control system of the multi-component material distribution of the ointment making machine is disclosed. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0024] The following is combined with Figure 1-Figure 3 The preferred embodiments of the present invention are described in detail.

[0025] See attached Figure 1 As shown, the present invention proposes a dynamic optimization control method for multi-component batching of a paste making machine, comprising the following steps: S1, obtaining the electromagnetic fingerprint signal of the residue on the surface of the stirring component, collecting the current phase offset through the polarization probe array arranged inside the stirring shaft, and generating a characteristic waveform spectrum corresponding to the molecular structure of the residue; S2, identifying the chemical bond type of the target residue based on the characteristic waveform spectrum, matching the preset formula switching database, and dynamically generating the electric lubrication medium parameters including the electroosmotic fluid concentration and release rate; S3, controlling the microporous ceramic tube to release the polar solution according to the electric lubrication medium parameters, and triggering the low-frequency vibration module at the bottom of the material layer, so that the released solution and the vibration wave form a molecular-level synergistic exfoliation effect; S4, collecting acoustic impedance change data in the mixing cavity, calculating the nanoparticle dispersion index in combination with the real-time stirring torque value, and generating a reverse gradient mixing instruction when the index exceeds a preset threshold; S5, executing the reverse gradient mixing instruction to adjust the stirring speed and the extrusion screw advancement phase difference, synchronously activating the thermal vortex ring in the spiral guide convex pattern area, and establishing a directional thermal energy migration channel; S6. Verify the dynamic equilibrium state between the paste core area temperature and the target formula viscoelasticity. When it is detected that the energy consumption rate drops to a stable range, trigger the phase-locked control signal to terminate the parameter adjustment chain.

[0026] In one embodiment of the present invention, step S1 includes the following steps: An array of polarization probes is embedded in the working surface of the stirring shaft of the paste making machine at equal intervals; Specifically, twelve groups of microelectrode groups were welded to the stress concentration area of ​​the stirring paddle blade in the form of a ring array. Each microelectrode group consisted of a sandwich structure consisting of a gold-coated tungsten core electrode and an iridium oxide reference electrode. The spacing between adjacent electrode groups was precisely controlled at 3.5 mm. The polarization probe array is a group of microelectrodes arranged in a circular array, which uses a sandwich structure to form a polarization current loop. The stress concentration area is the area where the stirring blade is subjected to the maximum shear force when rotating, and the point arrangement is determined by finite element stress analysis.

[0027] Apply a dynamic scanning voltage to the polarization probe before switching the recipe; Specifically, a potentiostat is used to input a The step wave voltage is measured, the step rate of which is inversely related to the relaxation time of the molecules of the residue on the cleaned surface, and the compensation current value of the reference electrode is monitored at the same time; The dynamic scanning voltage is a combination of asymmetric voltage pulses with step changes. The molecular relaxation time is the basic time unit required for the residue molecular chain to desorb under the action of the electric field. The step rate mapping table corresponding to different residues is established through pre-calibration experiments.

[0028] The step rate of the step wave voltage is inversely correlated with the molecular relaxation time of the residue on the cleaned surface, satisfying the following formula:

[0029] in, The unit is seconds, indicating the voltage step holding time; The unit is seconds, which represents the average relaxation time of paraffin molecules at standard temperature; represents the dimensionless coefficient, which represents the residual viscosity correction factor; represents a dimensionless constant, which is related to the electrode surface ion diffusion coefficient through Fick's second law.

[0030] Collect current phase offset and perform differential preprocessing; Specifically, by recording the current phase difference between the main electrode and the reference electrode at each voltage step, a sliding window difference algorithm is used to eliminate power frequency interference, and the characteristic phase offset is calculated based on the residual dielectric constant change rate; The sliding window difference algorithm is a dynamic filtering method that combines the phase difference changes in adjacent time periods, and each calculation takes the weighted differential result of three consecutive point measurements. The power frequency interference is a 55HZ clutter signal from the AC power supply.

[0031] Each calculation takes the weighted differential result of three consecutive point measurements, satisfying the following formula:

[0032] in, The unit is radian, indicating the The phase change rate of each window; is a dimensionless parameter, indicating the window sampling angle ratio; The unit is radian / second, the angular velocity of the stirring shaft measured by the encoder; , , Respectively represent the ) The original phase difference of the measurement; All units are radians. is the square of the dimensionless parameter, which is converted to radians per second by multiplying it with the angular velocity of the stirring shaft. The dimension of radians per second is consistent.

[0033] Generate electromagnetic fingerprint characteristic waveform spectrum; For details, please refer to the attached Figure 2 , the pre-processed characteristic phase offset is input into the Gram angle field converter in time series, and a two-dimensional grayscale matrix corresponding to the orientation characteristics of the residue molecules is output, and the matrix is ​​encoded into a pseudo-color waveform map with an eight-bit depth; Among them, the Gram angle field converter is a data processing unit that converts one-dimensional time series signals into two-dimensional spatial geometric representations, and the molecular orientation characteristics are the distribution characteristics of the arrangement direction of the residue molecular chain under the induction of the electric field. The color gradation changes of the pseudo-color waveform spectrum reflect the energy density differences in different adsorption intensity areas.

[0034] The characteristic phase offset is input into the Gram angle field transformer in time series, and the output is a two-dimensional grayscale matrix corresponding to the orientation characteristics of the residue molecules, which satisfies the following formula:

[0035] in, is the generated two-dimensional matrix element; , Respectively represent Group, A set of standardized current phase sequences; , Respectively represent Group, The time domain signal amplitude of the kth sampling of the group probe; , Respectively represent Group, The average time domain signal amplitude sampled by a group of probes.

[0036] In one embodiment of the present invention, step S2 comprises the following steps: Perform multi-modal decomposition processing on characteristic waveform spectra; Specifically, the characteristic waveform spectrum is loaded into the bandpass filter group through the waveform decomposition module, and the sub-waveform of a specific frequency band is segmented by an adaptive sliding window; a three-dimensional feature matrix including the waveform slope, extreme point density and amplitude fluctuation rate is simultaneously constructed in the time-frequency domain space; and the spectrum recognition engine is triggered to perform pattern matching on the matrix; The waveform decomposition module is a digital processing device based on wavelet packet transform. Its core algorithm sets the decomposition layer number to 6 and the energy ratio threshold of each layer is not less than The center frequency of the bandpass filter group is automatically adjusted according to the dielectric constant range of the residue to be removed, and the single-cycle scanning bandwidth is .

[0037] Analyze the dynamic matching process of chemical bond types; Specifically, the three-dimensional feature matrix is ​​input into the pre-trained bond type recognition model, and the feature vectors of hydrogen bond, ionic bond and van der Waals force are extracted using the residual network; when it is detected that the cosine similarity between the feature vector and the standard bond type vector in the preset formula switching database exceeds 0.85, the association mapping rule in the database is activated; The preset recipe switching database is a flash memory array that stores 38 kinds of paste component chemical bond characteristics and corresponding cleaning parameters. Each entry contains a 512-dimensional feature vector and corresponding 36 sets of working condition parameters. The association mapping rule is a dynamic matching mechanism based on the particle swarm optimization algorithm, which automatically assigns matching priorities according to the weight distribution of the feature vector.

[0038] Generate a control sequence of electric lubrication medium parameters; Specifically, the microfluidic chip controller is called according to the matching chemical bond type, and the concentration gradient curve of the electroosmotic solution is calculated in combination with the real-time rheological sensor data; at the same time, the nonlinear relationship between the stirring shaft speed and the material viscosity is analyzed, and the pulse release interval of the polar solution is dynamically adjusted; a parameter instruction set including the concentration-flow rate matching coefficient is generated; Among them, the microfluidic chip controller is a miniaturized execution unit integrating a piezoelectric pump and a capillary array, and its flow control accuracy reaches .

[0039] The concentration gradient curve is a two-dimensional distribution model based on Fick's diffusion law. The horizontal axis is the time axis and the vertical axis is the continuous change trajectory of the concentration value.

[0040] The two-dimensional distribution model established based on Fick's diffusion law satisfies the following formula:

[0041] in, is the diffusion flux, in units of ; is the electroosmotic diffusion coefficient, in ; Indicates the concentration of the solution in units of ; is the spatial coordinate of the diffusion direction, in units of .

[0042] Verify the coordination constraints of parameter generation; Specifically, the parameters of the electric lubrication medium are injected into the multi-physics field coupling simulation environment to detect the matching degree between the electroosmotic effect and the mechanical vibration in the molecular bond breaking efficiency; when the product of the vibration energy transfer function and the solution diffusion rate reaches a critical value, the final parameters are locked and written into the storage register; Among them, the multi-physics field coupling simulation environment is a real-time computing platform that integrates electrostatic field, fluid mechanics and solid vibration equations. Its grid division accuracy is set to The critical value is the bond dissociation energy threshold obtained through experimental calibration, corresponding to the molecular adsorption force per unit volume decaying to the original value. the following.

[0043] Build a dynamic compensation mechanism; Specifically, the historical error database is loaded before parameter execution, and the differential exponential smoothing algorithm is used to compensate for the influence of temperature drift on electroosmotic efficiency; the reference value of solution release rate is automatically corrected so that the deviation between the actual output flow rate and the theoretical value does not exceed ; Among them, differential exponential smoothing represents a dynamic filtering method that combines sliding windows with gradient calculations. The time weight of the data point is automatically adjusted according to the sensor sampling interval. The temperature drift compensation coefficient is linearly correlated with the thermal expansion coefficient of the shaft material, and the correction amount accounts for % of the baseline value. .

[0044] The differential exponential smoothing algorithm is used to compensate for the effect of temperature drift on electroosmosis efficiency, satisfying the following formula:

[0045] in, It is the corrected flow rate reference value at the current moment, in units of ; It is the flow value actually measured at the last moment, in units of ; is the baseline flow value at the previous moment, in units of ; is the temperature drift, in K; is a dimensionless coefficient, indicating the smoothing factor of flow data; Indicates the temperature compensation coefficient in units of .

[0046] In one embodiment of the present invention, constructing the key type recognition model comprises the following steps: The Raman spectra, infrared absorption spectra, and quantum chemical calculation data of 38 paste components were collected to establish a benchmark data set. After preprocessing, a mixed feature vector containing hydrogen bond vibration frequency, ionic bond polarizability and van der Waals dispersion energy was generated. A dual-path convolutional network architecture with residual connection was designed, in which the main path was configured with 6 layers of convolution kernels to extract local features such as atomic spacing, and the bypass used a multi-head self-attention mechanism to capture long-range correlations between molecules.

[0047] In the model training phase, the distance between positive and negative samples is dynamically adjusted based on the triplet loss function, and transfer learning is used to transfer the pre-trained The general bond feature parameters of the database are fine-tuned to adapt to specific cleaning scenarios; after the final model passes the X-ray photoelectron spectroscopy verification set test, it is deployed to the bond type recognition module in the form of embedded system firmware.

[0048] The residual connection is a cross-layer jump feature fusion structure. The size of the convolution kernel of each layer in the main path is set to 3×3 and the number of channels increases by 32-64-128. The triplet loss function imposes constraints on the difference in Euclidean distance between anchor samples, similar samples and heterogeneous samples, forcing the model to expand the discrimination boundary of different bond types. The quantum chemical data of the mixed feature vector is obtained by density functional theory calculation, and the bond energy resolution is accurate to .

[0049] In one embodiment of the present invention, step S3 includes the following steps: generating a pressure difference control signal; Specifically, the solution concentration and release rate values ​​of the electric lubrication medium parameters output in step S2 are input into the embedded control system and converted into multi-level pressure gradient instructions for the microporous ceramic tube through a nonlinear mapping algorithm. The pressure adjustment range is ; Among them, the pressure difference control signal is a pressure sequence that changes dynamically according to the solution characteristics, and its generation depends on the solution viscosity-pressure compensation curve preset in the embedded control system. The compensation curve is obtained by experimentally calibrating the flow resistance of electrolytes of different concentrations in micropores.

[0050] The solution viscosity-pressure compensation curve satisfies the following formula:

[0051] in, Indicates the compensation pressure, dimension , represents the actual working pressure required by the microporous ceramic tube; Represents the dynamic viscosity of the solution, dimensionless ; Indicates the actual concentration of the electrolyte, a dimensionless percentage value; is the material permeability coefficient, dimension ; is the maximum flow rate compensation, dimension , corresponding to the limiting flow capacity of the microchannel; is the concentration response coefficient, dimensionless, and its value is the curve slope matching parameter obtained in the calibration experiment; is the concentration reference value, dimensionless, corresponding to the concentration threshold at the inflection point of the material adsorption characteristics.

[0052] Activate the solution release function of the microporous ceramic tube; Specifically, the pressure difference control signal is transmitted to the annular pressure stabilizing chamber at the end of the microporous ceramic tube, and the double-layer silicone diaphragm in the pressure stabilizing chamber is used to implement intermittent opening and closing actions. The duration of each opening is The solution permeates outward through the built-in spiral microchannels distributed on the wall of the ceramic tube. The channel spacing is And the pore size is distributed in a gradient reduction; Among them, the microporous ceramic tube is made of alumina-based composite sintered material, and its wall is provided with a multi-layer staggered microchannel network, which includes a main channel and branch channels. , branch channel diameter The gradient reduction distribution is from the inner wall to the outer wall, the pore size is The rule of decreasing returns changes.

[0053] Synchronously trigger the generation of low-frequency vibration waves; Specifically, the solution release begins after a delay The vibration module at the bottom of the material layer is started in seconds. The vibration module is composed of eight groups of piezoelectric ceramic actuator arrays distributed in a circle. The waveform synthesizer generates The composite standing wave transmits vibration energy vertically along the material layer; The vibration module is a resonance amplification structure designed based on the principle of modal superposition. The vibration energy transmission path includes an annular titanium alloy vibration plate and a polytetrafluoroethylene buffer layer. The thickness of the vibration plate is A damping groove array is provided on the surface.

[0054] Implementing the synergistic effect of solution and vibration waves; Specifically, the polar solution forms a directional infiltration flow in the microporous channel, and at the same time, the vibration wave generates an acceleration pulse perpendicular to the material layer through the vibration plate. The amplitude of the acceleration pulse controls The solution is forced into microscopic gaps during capillary penetration. Among them, the synergistic effect is manifested in that the mechanical disturbance caused by the vibration wave changes the direction of the surface tension of the solution, accelerates the spreading speed of the solution along the residue adsorption interface, and at the same time, the local cavitation effect generated by the vibration wave expands the molecular contact area.

[0055] Monitor the completion of the stripping action; Specifically, the concentration of detached particles is detected in real time by an optical turbidity sensor arranged at the edge of the material layer. When the fluctuation range of the turbidity value is lower than a preset threshold within three consecutive sampling periods, it is determined that the molecular-level stripping is completed; The stripping completion degree determination standard is based on a dynamic attenuation model of particle detachment velocity, and the preset threshold is a turbidity change reference value automatically calculated based on the formula viscosity when the system is initialized.

[0056] The turbidity fluctuation determination for three consecutive sampling periods satisfies the following formula:

[0057] in, For the Turbidity value measured in sampling period, dimension ; is the average turbidity value of three cycles, dimension ; is the turbidity reference value, dimension , corresponds to the measured turbidity of the system in the initial non-peeling state; is the dynamic threshold, dimensionless, which is calculated from the initial turbidity value according to Calculated; Indicates the initial turbidity value, dimension .

[0058] In one embodiment of the present invention, step S4 comprises the following steps: Deploy multimodal sensor arrays at specific locations in the hybrid cavity; Specifically, two sets of cross-distributed piezoelectric ceramic sensor groups are installed at an axial interval of 15 cm in the mixing cavity, each group includes three ultrasonic transmitting and receiving units distributed at 120 degrees, and an interference measurement area is formed between adjacent sensor groups; Among them, the piezoelectric ceramic sensor group, which means the use of Wafer-fabricated broadband transducer arrays operating in the frequency range Each transmitting and receiving unit is coated with a high-temperature resistant polyimide protective layer. The interference measurement area is the cross-coverage area formed by the ultrasonic beam between adjacent sensor groups, and the medium density change is measured by the beam phase difference.

[0059] Acquire the time-series acoustic wave diffraction spectrum and synchronously collect the torque pulse of the power shaft; Specifically, the material mixing process triggers the ultrasonic transmitting and receiving unit with a period of 0.5 seconds to record the propagation time difference and amplitude attenuation of the reflected wave and the transmitted wave. At the same time, the original torque fluctuation signal of the stirring shaft is continuously collected through the flange torque sensor. The propagation time difference between the reflected wave and the transmitted wave is the time delay change from the transmitting end to the receiving end of the ultrasonic wave, accurate to 0.1 microseconds. The amplitude attenuation is the percentage of energy loss after the sound wave passes through the material layer. The original signal of torque fluctuation represents the voltage timing data converted by 24-bit ADC, and the sampling rate is not less than 1kHz.

[0060] Construct a dynamic parametric model of acoustic-mechanical coupling; Specifically, the sound wave propagation time difference is input into the sound wave diffraction spectrum analysis module, and the current material dynamic viscosity coefficient is calculated based on the positive correlation between the sound velocity and density of the medium. At the same time, the torque fluctuation original signal is input into the torque fluctuation analytical formula to extract the equivalent torque mean and peak-to-valley difference; Among them, the acoustic diffraction spectrum analysis module is an analysis system that uses the time-domain finite difference algorithm to process ultrasonic echo signals, and inverts the acoustic impedance distribution at different positions in the mixing cavity through iterative calculation. The torque fluctuation analysis formula is a combined function expression of the integrated average torque and the maximum instantaneous torque difference within a set time window.

[0061] Input the original torque fluctuation signal into the torque fluctuation analytical formula, extract the equivalent torque mean and peak-to-valley difference, and satisfy the following formula:

[0062] in, is the mean value and peak-to-valley difference of equivalent torque; For the The instantaneous torque value of the sampling point, unit ; is the average torque in the current time window, in units ; is the number of sampling points in the time window, dimensionless; is the maximum instantaneous torque in the time window, in units ; is the minimum instantaneous torque in the time window, in units ; Indicates the rated torque value of the equipment, unit .

[0063] Numerator The cumulative sum of ) and the denominator (unit ) cancels out the dimensions, the second fraction (The denominator and numerator have the same unit) It is also dimensionless. The result of multiplying two dimensionless parameters is still dimensionless, which satisfies the mathematical operation rules.

[0064] Real-time calculation of nanoparticle dispersion index; Specifically, the dynamic viscosity coefficient and the equivalent torque mean are input into the pre-trained proportional differential equation to solve the instantaneous diffusion rate of the nanoparticles. At the same time, the peak-to-valley difference and the preset speed-torque change rate comparison table are combined to calculate the particle agglomeration risk level; Among them, the proportional differential equation is a nonlinear equation containing a material temperature compensation term, which is used to relate the macroscopic dynamic parameters to the microscopic dispersion characteristics.

[0065] The speed-torque change rate comparison table represents a two-dimensional parameter matrix pre-established through material rheology experiments, which stores the torque fluctuation range allowed in different speed ranges.

[0066] The proportional differential equation satisfies the following formula:

[0067] in, is the instantaneous diffusion rate, in units of , characterizing the square of the diffusion distance of nanoparticles per unit time, is obtained by comprehensive calculation of dynamic viscosity and torque parameters. is the current dynamic viscosity coefficient, in units , reflecting the changes in the rheological properties of the mixed material. is the base viscosity value, unit , a pre-calibrated reference value of material viscosity under ideal dispersion conditions. Represents the proportional gain experimental coefficient, unit , the viscosity response coefficient calibrated by orthogonal test. Represents the experimental coefficient of coupling gain, unit , the torque influence coefficient determined by rheological experiments. Indicates the torque fluctuation difference during the window period, unit , the difference between the maximum torque and the minimum torque within a set time window (e.g. 10 seconds). Indicates the rated torque value of the equipment, unit .

[0068] Generate multi-dimensional mixed quality judgment results; Specifically, when the instantaneous diffusion rate is lower than the preset process standard and the particle agglomeration risk level reaches the orange warning threshold, the reverse gradient hybrid logic controller is triggered to output an adjustment instruction set; Among them, the reverse gradient hybrid logic controller represents an embedded processor with fuzzy decision-making function. The instruction set includes stirring speed reduction ratio parameters, low-speed vibration module start and stop flags, and thermal vortex ring working mode switching codes.

[0069] In one embodiment of the present invention, step S5 comprises the following steps: Receive inverse gradient blending instructions and decouple action parameters; Specifically, a dual-channel signal separation module is set in the central processing unit to split the mixed instruction into two independent data streams: the speed control signal and the phase compensation signal. The algorithm generates the speed curve correction value, and the phase compensation signal is transmitted through the delay buffer. Cycle timing calibration; The dual-channel signal separation module is a digital filter array in the embedded system, which uses frequency domain cross-correlation analysis to eliminate signal crosstalk. The algorithm is a parameter self-tuning controller combined with the empirical membership function, which dynamically adjusts the proportional coefficient according to the historical action error. The delay buffer is a memory array with a preset multi-threaded queue to achieve millisecond-level accurate delay.

[0070] Perform dynamic reduction of stirring speed; Specifically, the stirring motor vector driver loads the torque limit function, and the speed curve correction value is calculated per The speed is gradually attenuated by reducing the speed by one level per second, and the vacuum pump pressure sensor is linked to obtain the current viscous resistance characteristic value; The torque limit function is a built-in adaptive protection program of the three-phase inverter, which calculates the maximum torque pulse allowed in real time through the flux observer. The viscous resistance characteristic value is the viscosity change gradient value calculated by inverse calculation of vacuum negative pressure fluctuation.

[0071] During the adjustment of the stirring speed, the reduced-order constraint equation is applied to satisfy the following formula:

[0072] in, for The speed adjustment at the moment is ; is the initial speed, dimension is ; is the current speed, dimension is ; represents the viscous resistance compensation coefficient, a dimensionless parameter, through Calculate and obtain; Indicates the preset maximum viscosity, dimension ; To measure viscosity in real time, dimension ; is the time correction factor, dimension , the value is equal to the vacuum pump startup time .

[0073] Configure the extrusion screw advancement phase difference; Specifically, the phase difference compensation actuator applies a timestamp offset to the stepper motor of the extrusion screw, so that the vacuum pump starts earlier than the extrusion action. seconds, and at the same time, a negative pressure pre-stretching zone is formed at the front end of the extrusion; The phase difference compensation actuator refers to an electromechanical linkage device based on time stamp synchronization control. Bus coordinated multi-axis motion controller. The negative pressure pre-stretching zone is a density decreasing transition zone formed by the material under vacuum adsorption.

[0074] Activate the thermal vortex ring in the spiral guide ridge area; Specifically, the spiral guide convex pattern surface of the outer wall of the silo is covered with a variable heat conductive film layer, and a pulsed alternating electric field is applied to make it present asymmetric heat conduction characteristics, driving the heated airflow to form a tornado-like circulation along a preset path; The variable thermal conductivity film layer is a silicon-based composite material film doped with carbon nanotubes, and its thermal conductivity changes exponentially with the increase in electric field strength. The tornado circulation is a three-dimensional rotating flow pattern generated by the airflow under the induction of the spiral guide structure.

[0075] The thermal conductivity adjustment equation of the variable thermal conductivity film layer satisfies the following formula:

[0076] in, Represents electric field Equivalent thermal conductivity under the action of ; Represents the initial thermal conductivity of the substrate, dimension ; is the applied electric field strength, dimension ; represents the asymmetry factor, a dimensionless experimental constant; represents the percolation threshold electric field of carbon nanotubes, dimension , determined by the film thickness d (dimension m): .

[0077] Establish directional heat energy migration channels; Specifically, a thermosensitive optical fiber array is set on the central axis of the thermal vortex ring to dynamically monitor the temperature difference distribution in the core area of ​​the paste, and the fuzzy clustering algorithm is used to identify the optimal heat conduction path and fix the vortex rotation speed; The thermal optical fiber array is The star topology temperature measurement network is composed of distributed optical fibers, and each node integrates a fluorescence lifetime temperature sensor element. The fuzzy clustering algorithm is a dynamic partitioning method based on density peak search, with temperature uniformity as the optimization objective function.

[0078] The objective function of the fuzzy clustering algorithm satisfies the following formula:

[0079] in, To optimize the target value, the dimension ; Temperature point Belongs to cluster The degree of membership of is dimensionless; is the fuzzy index; is the temperature value of the i-th temperature measurement point, dimension K; is the temperature value of the jth cluster center, dimension K; is the core temperature variance, dimension K²; is the adjustment coefficient.

[0080] Coordinate mechanical action and thermodynamic response; Specifically, when the extrusion screw advancement phase difference and the heat vortex ring speed reach the synchronization critical point, the PID parameter reset action is triggered, and the stirring speed reduction rate is multiplied by the heat conduction gradient value for associated control; Among them, the synchronization critical point represents the state detection point where the screw thrust and thermal expansion force form a mechanical balance, and is determined by capturing the vibration frequency characteristics of the contact interface through the stress wave sensor. Product correlation control is a joint parameter adjustment mode that establishes a nonlinear coupling equation between two variables.

[0081] In one embodiment of the present invention, step S6 comprises the following steps: Initialize the dynamic balance verification module; Specifically, during the paste extrusion stage, a temperature gradient sensor array installed on the inner wall of the double-layer silo is started. The sensor array includes twelve groups of platinum resistance probes distributed in an annular pattern. The temperature data of sixteen calibration points in the longitudinal section of the silo are collected every eight seconds. At the same time, the online detection channel of the rotational rheometer is used to obtain the rheological parameters of the paste. Among them, the temperature gradient sensor array is an array of equidistant temperature measuring components inlaid on the surface of the arc-shaped warehouse wall. Each probe group includes a main and a secondary temperature measuring head. The main temperature measuring head is embedded in the warehouse wall three millimeters deep into the paste layer, and the secondary temperature measuring head is located at the coating interface on the outer side of the warehouse wall. The heat conduction rate is calculated by the temperature difference between the main and secondary temperature measuring heads. The online detection channel is a diversion capillary led out from the side wall of the extrusion flow channel. A conical flat plate fixture is set at the end of the capillary. The interval between the upper and lower plates of the fixture is adjusted to the corresponding value of the theoretical thixotropic index of the current paste viscosity. The real-time viscoelastic parameters are obtained by measuring the slope of the resistance curve when the plates are separated.

[0082] Establishing energy dissipation monitoring model; Specifically, the longitudinal temperature data collected by the temperature gradient sensor array is input into a pre-trained convolutional neural network, which is trained based on 500 sets of historical production data, and outputs a thermodynamic state prediction diagram of the paste core area. Simultaneously, the resistance curve obtained by the rheometer is input into a dynamic recursive equation to calculate the viscoelastic energy dissipation rate. The convolutional neural network is a residual network structure with five layers of convolution kernels. The third layer is equipped with a long short-term memory unit with a forget gate. The network weight initialization adopts the thermodynamic characteristic decomposition constraint algorithm. The dynamic recursive equation is a polynomial combination of material relaxation time and shear rate. The energy loss coefficient is updated every two seconds through an iterative solver.

[0083] Trigger phase-locked control signal; Specifically, when the absolute value of the energy dissipation rate changes for three consecutive iterations does not exceed the set threshold, the system is judged to have entered a stable range. At this time, the thermodynamic state prediction diagram is topologically matched with the inclination parameters of the current spiral guide convex pattern. If the matching degree reaches a critical value, a phase locking instruction is sent to the PLC controller. Among them, the phase locking instruction represents a composite signal combining the phase synchronization pulse and the power spectral density change. Its generation mechanism includes a three-step verification cycle: the first step is to verify the standard deviation decay rate of the temperature gradient, the second step is to verify the spatial coupling coefficient between the viscoelastic parameters and the diversion angle, and the third step is to trigger the vacuum pump delayed response compensation; the topology matching is to use the Minkowski distance algorithm to calculate the similarity index between the geometric features of the predicted image and the physical parameters of the diversion convex pattern. When the similarity index exceeds 0.9, the convex pattern inclination attitude maintenance instruction is triggered.

[0084] See attached Figure 3As shown, the present invention also proposes a dynamic optimization control system for multi-component batching of a paste making machine, comprising the following modules: A polarization probe array module, based on a ring-shaped microelectrode group embedded on the surface of the stirring assembly, is configured to collect current phase offsets by dynamically scanning voltage and generate characteristic waveform spectra that characterize the molecular structure of the residue; A formula dynamic analysis module, which is in communication with the polarization probe array module and is configured to analyze the chemical bond type of the characteristic waveform spectrum and generate electric lubrication medium parameters including electroosmotic fluid concentration and release rate based on a preset formula switching database; The solution-vibration synergistic stripping module controls the microporous ceramic tube to release polar solution according to the parameters of the electric lubrication medium, triggers the low-frequency vibration module of the material layer to generate vibration waves, so that the solution and the vibration waves form a molecular-level synergistic stripping effect; The dispersion determination module is based on the piezoelectric ceramic sensor group and torque sensor deployed in the mixing cavity, and is configured to calculate the coupled dynamic parameter model of acoustic impedance data and real-time torque, and output the nanoparticle dispersion index and the inverse gradient mixing instruction; The reverse gradient mixing control module is linked with the dispersion determination module and the extrusion screw driver, and is configured to adjust the stirring speed and the screw advancement phase difference, and synchronously trigger the asymmetric heat conduction film layer in the spiral guide convex pattern area to establish a directional heat energy migration channel; The dynamic balance phase-locking module, including a temperature gradient sensor array and an online rheometer, is configured to monitor the temperature-viscoelastic coupling state of the paste core area and the thermal vortex ring, and generate a phase-locking control signal through a convolutional neural network and a dynamic recursive equation to terminate the parameter adjustment chain.

[0085] The modules described above may be implemented in whole or in part through software, hardware, or a combination thereof, supporting a processor embedded in or independent of a computer device in hardware form, and also supporting a memory stored in a computer device in software form, so that the processor can call and execute operations corresponding to the modules described above.

[0086] It should be noted that the human information (including but not limited to human device information and personal information, etc.) and data (including but not limited to data used for analysis, stored data and displayed data, etc.) involved in the present invention are all information and data authorized by the human body or fully authorized by all parties. The collection, use and processing of relevant data require relevant legal standards.

[0087] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention is described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A dynamic optimization control method for multi-component batching of a paste making machine, characterized in that: The following steps are involved: S1, obtaining the electromagnetic fingerprint signal of the residue on the surface of the stirring component, collecting the current phase offset through the polarization probe array, and generating a characteristic waveform spectrum corresponding to the molecular structure of the residue; S2, dynamically generating electric lubrication medium parameters including electroosmotic fluid concentration and release rate based on matching the characteristic waveform spectrum with a preset formula switching database; S3, controlling the microporous ceramic tube to release the polar solution according to the electric lubrication medium parameters, triggering the low-frequency vibration module of the material layer to generate vibration waves, so that the solution and the vibration waves form a molecular-level synergistic exfoliation effect; S4, collecting acoustic impedance change data and stirring torque value in the mixing chamber, calculating the nanoparticle dispersion index, and generating a reverse gradient mixing instruction when the index exceeds a threshold; S5, executing the reverse gradient mixing instruction to adjust the stirring speed and the extrusion screw advancement phase difference, activate the thermal vortex ring in the spiral guide convex pattern area and establish a directional thermal energy migration channel; S6. Verify the dynamic equilibrium state between the paste core area temperature and the target formula viscoelasticity. When the energy consumption rate reaches a stable range, trigger a phase-locked control signal to terminate the parameter adjustment chain.

2. The method for dynamic optimization control of multi-component batching of a paste making machine according to claim 1, characterized in that: Generating a characteristic waveform spectrum corresponding to the molecular structure of the residue includes the following steps: Applying a dynamic scanning voltage to the polarization probe array and collecting a current phase offset; Perform differential preprocessing on the current phase offset to eliminate interference signals; The pre-processed phase offset is input into the Gram angular field converter for time conversion, and a two-dimensional gray matrix reflecting the molecular orientation characteristics is output; The two-dimensional grayscale matrix is ​​encoded into a pseudo-color waveform spectrum as a characteristic waveform spectrum.

3. The method for dynamic optimization control of multi-component batching of a paste making machine according to claim 2, characterized in that: Dynamically generating the electrolubricating medium parameters including the electroosmotic fluid concentration and the release rate includes the following steps: Perform multimodal decomposition on the characteristic waveform spectrum and construct a three-dimensional feature matrix including waveform slope and extreme point density; The chemical bond types of the three-dimensional feature matrix are analyzed through the pre-trained bond type recognition model, and the standard bond type vectors in the formula switching database are matched; The microfluidic chip controller is called to calculate the electroosmotic fluid concentration gradient curve and pulse release interval, and generate a control sequence of the electrolubricating medium parameters.

4. The method for dynamic optimization control of multi-component batching of a paste making machine according to claim 3, characterized in that: The control sequence for generating the parameters of the electrolubrication medium includes the following steps: Inject the parameters of the electrolubricating medium into the multi-physics coupling simulation environment to verify the molecular bond breaking efficiency of electroosmosis and mechanical vibration; When the product of the vibration energy transfer function and the solution diffusion rate reaches a critical value, the final parameters are output to the microfluidic chip controller.

5. The method for dynamic optimization control of multi-component batching of a paste making machine according to claim 4, characterized in that: Triggering molecular-level synergistic exfoliation, including the following steps: The parameters of the electrolubricating medium are converted into a differential pressure control signal of the microporous ceramic tube to drive the polar solution to penetrate through the spiral microchannel. After a preset delay, the piezoelectric ceramic actuator array at the bottom of the material layer is activated to generate a vibration wave in the form of a composite standing wave; The concentration of detached particles is monitored by an optical turbidity sensor to determine the molecular level of exfoliation completion and feed back to the mixing chamber.

6. The method for dynamic optimization control of multi-component batching of a paste making machine according to claim 1, characterized in that: Generate the inverse gradient blending instruction, including the following steps: The propagation time difference and amplitude attenuation of the sound waves in the mixing cavity are collected by a piezoelectric sensor group; The acoustic-mechanical coupling dynamic parameter model was constructed in combination with the stirring torque fluctuation data to calculate the instantaneous diffusion rate and agglomeration risk level of nanoparticles; When the diffusion rate is lower than the process standard and the agglomeration risk level exceeds the threshold, the reverse gradient mixing logic controller is triggered to output the adjustment instruction set.

7. The method for dynamic optimization control of multi-component batching of a paste making machine according to claim 6, characterized in that: Establishing a directional heat energy migration channel includes the following steps: A pulse electric field is applied to the variable heat-conducting film layer on the surface of the spiral flow-guiding convex pattern to generate asymmetric heat conduction characteristics to drive a tornado-like circulation; The temperature difference distribution in the core area of ​​the paste is monitored by a thermosensitive optical fiber array, and the optimal heat conduction path is determined and the eddy current speed is fixed using a fuzzy clustering algorithm; When the extrusion screw advancement phase difference and the thermal vortex ring speed reach the synchronization critical point, the stirring speed reduction rate and the heat conduction gradient value are jointly controlled.

8. The method for dynamic optimization control of multi-component batching of a paste making machine according to claim 7, characterized in that: Verifying the dynamic equilibrium state between the paste core temperature and the target formula viscoelasticity includes the following steps: The temperature data of the longitudinal section of the paste is collected through a temperature gradient sensor array and input into a convolutional neural network to generate a thermodynamic state prediction diagram; Obtain the rheological parameters of the paste and input them into the dynamic recursive equation to calculate the viscoelastic energy dissipation rate; When the energy dissipation rate changes below the set threshold for three consecutive iterations, the thermodynamic state prediction diagram is topologically matched with the inclination parameters of the spiral guide convex pattern to trigger the phase locking signal.

9. The method for dynamic optimization control of multi-component batching of a paste making machine according to claim 8, characterized in that: Triggering a phase-locked signal includes the following steps: Verify the standard deviation decay rate of temperature gradient and the spatial coupling coefficient of viscoelastic parameters; If the matching degree exceeds the critical value, a composite phase-locked control signal including synchronization pulses and power spectrum density is sent to the PLC controller.

10. A dynamic optimization control system for multi-component batching of a paste making machine, characterized in that: Includes the following modules: A polarization probe array module, based on a ring-shaped microelectrode group embedded on the surface of the stirring assembly, is configured to collect current phase offsets by dynamically scanning voltage and generate characteristic waveform spectra that characterize the molecular structure of the residue; A formula dynamic analysis module, which is in communication with the polarization probe array module and is configured to analyze the chemical bond type of the characteristic waveform spectrum and generate electric lubrication medium parameters including electroosmotic fluid concentration and release rate based on a preset formula switching database; The solution-vibration synergistic stripping module controls the microporous ceramic tube to release polar solution according to the parameters of the electric lubrication medium, triggers the low-frequency vibration module of the material layer to generate vibration waves, so that the solution and the vibration waves form a molecular-level synergistic stripping effect; The dispersion determination module is based on the piezoelectric ceramic sensor group and torque sensor deployed in the mixing cavity, and is configured to calculate the coupled dynamic parameter model of acoustic impedance data and real-time torque, and output the nanoparticle dispersion index and the inverse gradient mixing instruction; The reverse gradient mixing control module is linked with the dispersion determination module and the extrusion screw driver, and is configured to adjust the stirring speed and the screw advancement phase difference, and synchronously trigger the asymmetric heat conduction film layer in the spiral guide convex pattern area to establish a directional heat energy migration channel; The dynamic balance phase-locking module, including a temperature gradient sensor array and an online rheometer, is configured to monitor the temperature-viscoelastic coupling state of the paste core area and the thermal vortex ring, and generate a phase-locking control signal through a convolutional neural network and a dynamic recursive equation to terminate the parameter adjustment chain.

Images