Method and system for controlling production process of heart and cerebral vessel health care products based on multi-parameter collaborative optimization

Abstract

The invention provides a cardiovascular and cerebrovascular health care product production process control method and system based on multi-parameter collaborative optimization. Wherein thermal field temperature parameters of active components in a mixing container are collected in real time, and a parameter matrix is constructed by combining degradation kinetics and phase change state parameters; establishing a correlation model of a thermal stability layer, a phase change layer and an anti-oxidation attenuation layer based on a deep learning algorithm, and generating a differential control instruction of the double-spiral stirrer by optimizing a thermodynamic analysis rule; a compound motion mode of radial sweeping stirring of the main spiral body and axial pulse overturning of the auxiliary spiral body is executed, and a semiconductor temperature control device is synchronously adopted for temperature gradient compensation, so that dynamic matching of the stirring speed and temperature control parameters is realized; mapping lipid carrier premixing and inclusion processes and sustained-release matrix crosslinking parameters to a temperature regulation and control system to form comprehensive process parameter closed-loop feedback; and the degradation kinetic parameters are reversely corrected through online detection data, and the parameter matrix is iteratively updated until a preset quality index is reached.

Description

Technical Field

[0001] The present application relates to the technical field of multi-parameter collaborative optimization, and in particular to a method and system for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization. Background Art

[0002] The production of cardiovascular health products involves the synergistic interaction of multiple components. The extraction, purification, and formulation processes of their core active ingredients (such as tanshinone and notoginseng saponins) are highly sensitive to parameters such as temperature, pH, pressure, and enzymatic hydrolysis time. Traditional single-parameter independent regulation or empirical process control methods are difficult to ensure ingredient stability and bioavailability. Dynamic, coordinated optimization of multiple parameters is required to achieve precise process control while meeting production requirements for high efficiency, low energy consumption, and high batch consistency.

[0003] The current mainstream approach uses a process optimization system based on a static response surface model (RSM). This system uses experimental design (e.g., Box-Behnken) to establish a mathematical relationship between key parameters (such as extraction temperature and solvent concentration) and target component yield. Based on this model, a fixed combination of process parameters is set. The system uses PLC-controlled actuators (such as the reactor temperature control module) to achieve automated production. Some companies also use online HPLC testing feedback to adjust key steps.

[0004] Static RSM models rely on linear assumptions under specific experimental conditions and are unable to dynamically respond to raw material fluctuations (such as batch differences in traditional Chinese medicine) or environmental disturbances (such as boiling point shifts caused by changes in atmospheric pressure), resulting in poor model extrapolation. Furthermore, multi-parameter synergistic effects are only fitted through a limited number of experimental points, making it difficult to account for complex nonlinear relationships (such as the enzymatic hydrolysis-extraction coupling stage), which can easily lead to local degradation of target ingredients or redundant energy consumption. A dynamic optimization mechanism driven by real-time data is needed to address these shortcomings. Summary of the Invention

[0005] The present application provides a method and system for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization, which is used to solve the problems in the existing technology that the production process adjustment is not flexible enough and is easily affected by raw material differences, resulting in unstable product quality and low production efficiency.

[0006] In a first aspect, the present application provides a method for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization, comprising:

[0007] Collecting thermal field temperature parameters of active ingredients in cardiovascular and cerebrovascular health products in a mixing container, and constructing a parameter matrix based on the degradation kinetic parameters and phase change state parameters corresponding to the active ingredients;

[0008] Based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer is established through a deep learning algorithm. The multi-parameter correlation model generates differential speed control instructions for the double-helix stirrer by optimizing the analytical rules of the thermodynamic properties corresponding to the active ingredients;

[0009] According to the parameter matrix and the differential control instruction, the main helix is driven to perform a radial sweep stirring mode and the auxiliary helix is driven to perform an axial pulse flipping mode. At the same time, the semiconductor temperature control device is used to perform gradient compensation for temperature fluctuations in different areas of the mixing container to form temperature control data that dynamically matches the stirring rate.

[0010] Mapping the progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking to the temperature control data, and feeding the mapped comprehensive process parameters back to the multi-parameter correlation model;

[0011] Based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, the degradation kinetic parameters in the multi-parameter association model are reversely corrected, and the parameter matrix is iteratively updated until the product meets the preset quality indicators.

[0012] Optionally, based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer is established through a deep learning algorithm. The multi-parameter correlation model generates a differential speed control instruction for the double-helix stirrer by optimizing the thermodynamic property analysis rules corresponding to the active ingredients, including:

[0013] Based on the parameter matrix, the thermal stability layer defines the thermal state maintenance capability of each spatial unit through the temperature gradient and time fluctuation range; the phase change layer defines the heat transfer path and triggering conditions through the temperature difference and change rate of adjacent units; and the antioxidant attenuation layer defines the cumulative attenuation effect of the thermal field on the active components through the temperature extreme value and duration. The output characteristics of the thermal stability layer, phase change layer, and antioxidant attenuation layer are integrated to construct a multi-parameter correlation model;

[0014] Establishing thermodynamic property analysis rules between each layer of the multi-parameter joint model. Based on the thermodynamic property analysis rules, a threshold is used to determine whether the thermal state maintenance capacity between the thermal stability layer and the phase change layer is insufficient. If the threshold is lower than the threshold, a compensating heat flow is added to the phase change layer. A critical value is used to determine whether the compensating heat flow between the phase change layer and the antioxidant attenuation layer triggers a reduction in the cumulative attenuation effect, and the hierarchical parameters are reversely iterated to meet the thermal state boundary of the active component.

[0015] The speed increment and action duration of the specified area are generated according to the spatial distribution of the compensating heat flux, the speed difference upper limit and switching frequency of adjacent areas are generated according to the attenuation reduction ratio, and the differential speed control instruction of the double-helix stirrer is generated, and its consistency with the thermal field hierarchical structure is verified.

[0016] Optionally, thermal field temperature parameters of active ingredients in cardiovascular and cerebrovascular health products are collected in a mixing container, and a parameter matrix is constructed by combining the degradation kinetic parameters and phase change state parameters corresponding to the active ingredients, including:

[0017] Dividing a plurality of independent areas marked by three-dimensional coordinates in the mixing container, and collecting thermal field temperature parameters of the independent areas in real time through a fixing device;

[0018] The decomposition rate of the active ingredient determined based on the relationship between temperature and decomposition rate, and the phase change state parameters calibrated by the container material and ingredient ratio experiment;

[0019] Finally, the thermal field temperature parameters, decomposition rate and thermal field temperature data are arranged according to spatial positions and combined into a parameter matrix reflecting the state of the active ingredient.

[0020] Optionally, according to the parameter matrix and the differential control instruction, the main helix is driven to perform a radial sweep stirring mode and the auxiliary helix is driven to perform an axial pulse flipping mode, and at the same time, a semiconductor temperature control device is used to perform gradient compensation for temperature fluctuations in different areas of the mixing container to form temperature control data that dynamically matches the stirring rate, including:

[0021] According to the differential speed control instruction, the main screw is controlled to move back and forth periodically along the radial direction of the mixing container for stirring, and the angle of the stirring blade is changed during the movement to cover the materials at different radial depths. At the same time, the auxiliary screw is controlled to intermittently push the materials to turn over along the axial direction of the container;

[0022] Through the semiconductor temperature control device, the temperature fluctuations in different areas of the mixing container are monitored in real time, and the cooling or heating power of different areas is dynamically adjusted according to the temperature gradient compensation requirements in the parameter matrix, so that the temperature changes are synchronously adapted to the stirring rates of the main helix and the auxiliary helix, forming temperature control data that can be dynamically matched in real time.

[0023] Optionally, mapping the progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking to the temperature control data, and feeding the mapped comprehensive process parameters back to the multi-parameter association model, including:

[0024] mixing the lipid carrier with the auxiliary component to form a lipid carrier premix product, and recording temperature change data during the mixing process to obtain a lipid carrier premix temperature sequence;

[0025] adding the active ingredient to the lipid carrier premix product, controlling the addition rate to form an active ingredient inclusion product, and recording temperature change data during the inclusion process to obtain an active ingredient inclusion temperature sequence;

[0026] adding a crosslinking agent to the active ingredient inclusion product, adjusting the dropwise addition rate, and recording the temperature change data during the crosslinking process to obtain a crosslinking temperature sequence for the sustained-release matrix;

[0027] The lipid carrier premixing temperature sequence, active ingredient inclusion temperature sequence, and sustained-release matrix crosslinking temperature sequence are respectively compared with the previously acquired temperature control data to establish a temperature control mapping relationship, and the process parameters of each stage are adjusted to form a comprehensive process parameter set;

[0028] The comprehensive process parameter set is input into the multi-parameter association model to adjust the execution strategy of subsequent process stages.

[0029] Optionally, based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, the degradation kinetic parameters in the multi-parameter association model are reversely corrected, and the parameter matrix is iteratively updated until the product meets the preset quality indicators, including:

[0030] The concentration and status parameters of active ingredients in different areas are obtained through real-time detection devices, and compared with the preset quality indicators item by item to generate difference results;

[0031] Based on the difference results and the comprehensive process parameters, reversely adjusting the active ingredient decomposition rate parameter degradation kinetic parameter preset in the multi-parameter association model;

[0032] The adjusted degradation kinetic parameters are substituted into the multi-parameter association model, the values of the corresponding areas in the parameter matrix are updated, and the detection, comparison, adjustment and update steps are performed cyclically until the state parameters of the active ingredient meet the preset quality indicators.

[0033] Optionally, a thermodynamic property analysis rule is established between each level of the multi-parameter joint model. Based on the thermodynamic property analysis rule, a threshold value is used to determine whether the thermal state maintenance capacity between the thermal stability layer and the phase change layer is insufficient. If the threshold value is lower than the threshold, a compensating heat flow is added to the phase change layer. A critical value is used to determine whether the compensating heat flow between the phase change layer and the antioxidant attenuation layer triggers a reduction in the cumulative attenuation effect, and the hierarchical parameters are reversely iterated and corrected to meet the thermal state boundary of the active component, including:

[0034] In the multi-parameter joint model, the threshold conditions for heat transfer from the thermal stability layer to the phase change layer and the critical conditions for heat transfer from the phase change layer to the anti-oxidation attenuation layer are defined to establish analytical rules for thermodynamic characteristics between the layers;

[0035] Based on the threshold condition, if the ability of the thermal stability layer to maintain the thermal state of the phase change layer is lower than a preset minimum heat value, then additional heat compensation is calculated and added to the phase change layer by dividing the current heat gap of the thermal stability layer by a preset compensation ratio;

[0036] After adding the additional heat compensation, based on the critical condition, if the heat transferred from the phase change layer to the anti-oxidation attenuation layer exceeds the preset cumulative attenuation safety value, the heat compensation value of the phase change layer is reduced by the excess ratio, and the reduced amount is transferred back to the thermal stability layer to update the thermal parameters of the thermal stability layer;

[0037] The parameters of each layer are modified by reverse iteration to meet the thermal boundary of the active component until the thermal parameters of the thermal stability layer, phase change layer and antioxidant attenuation layer all meet the preset range of the thermal boundary of the active component.

[0038] In a second aspect, the present application provides a cardiovascular health product production process control system based on multi-parameter collaborative optimization, comprising:

[0039] An acquisition module is used to collect thermal field temperature parameters of active ingredients in cardiovascular and cerebrovascular health products in a mixing container, and to construct a parameter matrix based on the degradation kinetic parameters and phase change state parameters corresponding to the active ingredients;

[0040] a generation module for establishing, based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer using a deep learning algorithm, wherein the multi-parameter correlation model generates a differential speed control instruction for the double-helix stirrer by optimizing the analytical rules of the thermodynamic properties corresponding to the active ingredients;

[0041] an execution module, configured to drive the main helicoid to execute a radial sweep stirring mode and the auxiliary helicoid to execute an axial pulse flipping mode according to the parameter matrix and the differential speed control instruction, and simultaneously perform gradient compensation for temperature fluctuations in different areas of the mixing container through a semiconductor temperature control device to form temperature control data that dynamically matches the stirring rate;

[0042] a mapping module for mapping progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking to the temperature control data, and feeding back the mapped comprehensive process parameters to the multi-parameter association model;

[0043] The correction module is used to reversely correct the degradation kinetic parameters in the multi-parameter association model based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, and iteratively update the parameter matrix until the product meets the preset quality indicators.

[0044] In a third aspect, an embodiment of the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a cardiovascular health product production process control method based on multi-parameter collaborative optimization as described in the first aspect above.

[0045] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements a cardiovascular health product production process control method based on multi-parameter collaborative optimization as described in the first aspect.

[0046] In the embodiment of the present application, pressure change data of the corresponding parking spaces are collected by pressure sensors deployed in parking spaces on each floor of the parking lot, and multiple vehicle size data in the vehicle identification system are simultaneously obtained. A three-dimensional parking heat map is generated based on the spatial distribution relationship between the pressure change data and the vehicle size data, which can realize real-time dynamic modeling of multi-dimensional parking demand distribution characteristics and accurately reflect the spatial occupancy status of parking spaces on each floor; by installing a displacement sensor array on the parking lot column, the displacement sensor array monitors the change of vehicle chassis height, and can capture the dynamic change characteristics of the vehicle chassis height in real time, providing basic data support for vehicle physical size matching; by constructing a vehicle size database based on the mapping relationship between the chassis height change and vehicle size data, it can dynamically associate The constraint relationship between vehicle chassis height and body size is used to generate a screening basis that adapts to the physical space conditions of different vacant parking spaces; by collaboratively processing the parking demand information and physical space constraint conditions, a matching set of vacant parking spaces is screened according to the current vehicle size data when the current vehicle enters the parking lot and the parking demand change trend predicted by the three-dimensional parking heat map, so as to achieve dynamic optimization allocation of parking space resources based on real-time data and predicted trends; by determining the correlation between historical parking path data and the three-dimensional parking heat map, and combining the spatial position of the vacant parking space set with the occupancy status of the parking lot traffic path to generate a guidance path for the current vehicle, a spatiotemporal coordinated path planning mechanism can be constructed to effectively avoid traffic congestion areas and improve parking efficiency.

[0047] Furthermore, the core subordinate steps include: based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, phase change layer and antioxidant attenuation layer is constructed through a deep learning algorithm, wherein the thermal stability layer defines the thermal state maintenance capability, the phase change layer defines the heat transfer path and trigger conditions, and the antioxidant attenuation layer defines the thermal cumulative attenuation effect; dynamic compensation and correction between layers are realized through thermodynamic characteristic analysis rules, including threshold judgment and compensating heat flow addition between the thermal stability layer and the phase change layer, critical value judgment and attenuation effect reduction between the phase change layer and the antioxidant attenuation layer; finally, the differential control instructions of the double-helix stirrer are generated according to the spatial distribution of the compensating heat flow and the attenuation reduction ratio, and their consistency with the thermal field hierarchical structure is verified; through the collaborative optimization of multi-level thermodynamic models, the thermal stability of the active ingredient is precisely controlled, the energy loss in the heat transfer process is dynamically compensated and the cumulative attenuation effect is suppressed; the differential control instructions generated based on the compensating heat flow and the attenuation reduction ratio ensure that the stirring process is highly matched with the thermal field distribution, significantly improving the uniform mixing efficiency and thermal stability of the active ingredient, and ultimately ensuring the consistency and effectiveness of product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0049] Figure 1 A flowchart of a method for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization provided by the present application is shown;

[0050] Figure 2 The present invention provides a schematic diagram of a cardiovascular and cerebrovascular health product production process control system based on multi-parameter collaborative optimization;

[0051] Figure 3 A schematic structural diagram of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0052] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0053] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0054] Researchers have found that the existing production process of cardiovascular and cerebrovascular health products has problems such as insufficient thermal field control accuracy, unclear degradation mechanism of active ingredients, and low matching degree between stirring process and thermodynamic characteristics, which leads to unstable retention rate of active ingredients in products and significant quality differences between batches. Based on this, a production process control method based on multi-parameter collaborative optimization is provided. This method can achieve precise control of the thermodynamic characteristics of active ingredients and adaptive optimization of process parameters through multi-parameter modeling of thermal stability layer-phase change layer-antioxidant attenuation layer and intelligent regulation of double-helix differential stirring. The technical solution of this application can be applied to high-value-added health product manufacturing scenarios such as mixed encapsulation of temperature-sensitive active ingredients and production of sustained-release products.

[0055] The entire research and development process embodies the technical linkage of thermodynamic property analysis and intelligent control, aiming to overcome the defects of uneven thermal field distribution, insufficient control of phase change behavior and lack of quantification of antioxidant attenuation in traditional processes. Through real-time monitoring of thermal field parameters and degradation kinetics, the limitations of traditional processes in controlling the thermal stability of active ingredients are broken through; combined with the correlation analysis of phase change state and antioxidant attenuation, a multi-parameter collaborative optimization model and an intelligent decision-making mechanism for differential stirring are constructed to achieve dynamic matching of thermodynamic properties and stirring process; based on real-time feedback of temperature control, the hysteresis problem of process parameters and product quality in traditional methods is solved. This method significantly improves the retention rate of active ingredients and the quality stability between batches of products through multi-parameter modeling and closed-loop control mechanism. The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of them. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0056] Figure 1 The present invention provides a flowchart of a method for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization. Figure 1 As shown, the method includes:

[0057] 101. Collect thermal field temperature parameters of active ingredients in the cardiovascular health product in a mixing container, and construct a parameter matrix based on degradation kinetic parameters and phase change state parameters corresponding to the active ingredients;

[0058] Thermal field temperature parameters refer to the three-dimensional spatial temperature distribution data in the mixing container obtained through distributed temperature sensors, including real-time temperature values and their temporal and spatial variation trends; degradation kinetic parameters are quantitative indicators that describe the changes in the chemical stability of active ingredients under different temperature conditions, including reaction rate constants, half-lives, etc.; phase change state parameters characterize the physical state transition characteristics of the material during temperature changes, such as melting onset temperature, crystallization enthalpy, etc.; the parameter matrix is a multidimensional data set formed by integrating the above three types of parameters according to time series and spatial coordinates, which is used to comprehensively reflect the thermodynamic state of the mixing process.

[0059] In the embodiment of the present application, a 5×5 array of PT100 temperature sensors was arranged in a 50L mixing container, and the sensor sampling accuracy reached ±0.1°C. The noise interference in the collected signal was eliminated by a sliding average filtering algorithm, and the degradation kinetic parameters of the active ingredient were retrieved from the material database, including Arrhenius equation parameters such as the rate constant k value of 1.2×10 3 and activation energy Ea of 85 kJ / mol. Phase transition parameters measured by differential scanning calorimetry, such as a melting peak temperature of 52°C and an enthalpy change of 120 J / g, were also obtained. Kriging spatial interpolation was used to expand the discrete temperature measurement points into a continuous temperature field. Tensor operations were then applied to spatially and temporally align the thermal field temperature parameters with the degradation kinetic parameters and phase transition parameters, ultimately generating a structured parameter matrix. This matrix served as the foundational dataset and provided input for the subsequent development of a multi-parameter correlation model.

[0060] During the production of Coenzyme Q10 liposomes, 36 PT1000 temperature sensors were deployed in the mixing vessel. Real-time monitoring revealed that the temperature in the northwest region rose abnormally from 75°C to 83°C within 10 minutes, exceeding the set threshold of ±2°C. The system retrieved the degradation kinetic parameters of this batch of Coenzyme Q10, such as the activation energy Ea = 82.3 kJ / mol. This, combined with the phase transition onset temperature of 81.5°C detected by DSC, was used to construct a parameter matrix. The matrix showed that the degradation rate in the northwest region reached 0.12% / min, while the average in other regions was 0.07% / min. Furthermore, the phase transition flag was triggered, marking the region as a high-risk thermal runaway area.

[0061] 102. Based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer is established through a deep learning algorithm. The multi-parameter correlation model generates a differential speed control instruction for the double-helix stirrer by optimizing the analytical rules of the thermodynamic characteristics corresponding to the active ingredients;

[0062] The thermal stability layer is a local thermal field stability assessment module constructed through a convolutional neural network, which outputs the thermal tolerance index of each spatial unit; the phase change layer is a phase transition analysis module based on a graph neural network, which identifies the material state change path driven by the temperature gradient; the antioxidant attenuation layer is an oxidation loss prediction model established using a long-short-term memory network, which quantifies the impact of thermal history on active ingredients; the thermodynamic characteristic analysis rule is a model optimization algorithm that integrates physical laws (such as the heat conduction equation) to ensure that the prediction results comply with the principle of conservation of energy; the differential control instruction includes a set of operating parameters such as the speed ratio of the main / auxiliary helices and the steering switching timing.

[0063] In the embodiment of the present application, the parameter matrix is input into a three-branch neural network architecture. The thermal stability layer uses a 5×5×3 three-dimensional convolution kernel to scan the temperature field data and output the thermal risk level of each grid point; the phase change layer converts the temperature gradient data into a graph structure and identifies the key phase change conduction path through the graph attention network; the antioxidant attenuation layer uses a bidirectional LSTM to analyze the temperature time series characteristics. The output of each layer is embedded in the non-steady-state heat conduction equation through the physical constraint layer for coupling optimization. When it is detected that the thermal risk value of a certain area exceeds the threshold of 0.8, a control instruction is generated to increase the speed of the main helix and periodically reverse the secondary helix. This process continuously adjusts the model weights through the gradient backpropagation algorithm to ensure the consistency of the prediction results with the actual thermodynamic laws.

[0064] Based on the parameter matrix from step 101, the multi-parameter correlation model detected that the thermal risk index in the northwest high-temperature zone exceeded the threshold, for example, a thermal risk index of 0.91 > the threshold of 0.8. The phase transition layer identified a 4.2°C / cm temperature gradient extending southeastward from this region, triggering abnormal melting of lipid carriers in adjacent areas. The antioxidant decay layer predicted that sustained stirring for 15 minutes would result in an 11% loss of coenzyme Q10. The model generated the following instructions: ① The main helix performed localized enhanced stirring in the northwest region, increasing the speed from 600 rpm to 850 rpm; ② The secondary helix initiated axial rotation every 0.8 seconds; ③ 50 mL of lipid carrier pre-cooled to 65°C was injected into the northwest region for thermal buffering.

[0065] 103. According to the parameter matrix and the differential speed control instruction, the main helix is driven to perform a radial sweep stirring mode and the auxiliary helix is driven to perform an axial pulse flipping mode. Simultaneously, a semiconductor temperature control device is used to perform gradient compensation for temperature fluctuations in different areas of the mixing container, thereby forming temperature control data that dynamically matches the stirring rate.

[0066] The radial sweep stirring mode is the spiral propulsion motion of the main helix along the radial direction of the container, which enhances lateral mixing through a specific trajectory design; the axial pulse flipping mode is the up and down reciprocating motion of the secondary helix according to a set timing, which is used to break the temperature stratification; the gradient compensation is to dynamically adjust the output power of the semiconductor refrigeration plate according to the thermal field prediction data, forming a partitioned temperature control strategy that matches the change in stirring rate.

[0067] In an embodiment of the present application, the main helix is driven by a servo motor to execute a preset spiral trajectory motion: it moves 1.5 cm radially for every 20° rotation, forming a progressive radial sweep stirring mode. The secondary helix is controlled by a stepper motor, and after receiving the pulse signal, it alternates forward and reverse rotation with a period of 0.8 seconds to form an axial pulse flip mode. The semiconductor temperature control system uses an adaptive PID algorithm to adjust the cooling power in different regions based on the thermal field prediction of the parameter matrix, for example, +8W in the high temperature zone and -4W in the low temperature zone. The thermal field temperature parameters are fed back to the stirring controller in real time through the CAN bus to form a closed-loop regulation to obtain temperature control data that dynamically matches the stirring rate.

[0068] After executing the differential control command in step 102, the radial sweep of the main helix tripled the material exchange rate in the northwest zone, while the high-frequency flipping of the secondary helix disrupted temperature stratification. During this process, the temperature control system detected a 3.1°C temperature fluctuation in the southeast zone due to heat conduction and immediately initiated gradient compensation: cooling power in the southeast zone increased from 25W to 38W, while regional power decreased by 12%. Ten minutes later, the parameter matrix was updated to show that the temperature in the northwest zone had returned to 78±0.5°C, and the standard deviation of the temperature across the entire container had decreased from 2.4°C to 0.9°C.

[0069] 104. Mapping the progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking to the temperature control data, and feeding the mapped comprehensive process parameters back to the multi-parameter association model;

[0070] The lipid carrier premixing parameters include process boundary conditions such as the melting temperature window and the shear rate threshold; the active ingredient inclusion parameters refer to key indicators such as the mass ratio of the carrier to the active substance and the mixing energy input; the sustained-release matrix cross-linking parameters include control points such as the timing of cross-linker addition and the gelation rate.

[0071] In the embodiment of the present application, the system dynamically adjusts the three-stage process based on real-time temperature control data: in the premixing stage, the lipid carrier is premixed, the lipid melting state is monitored, and the heating power is automatically adjusted when the temperature deviates from the set value by ±1.5°C; in the inclusion stage, the active ingredient is added to the lipid carrier premix product, the inclusion rate is monitored by an online ultraviolet spectrometer, and the stirring energy is dynamically adjusted according to the rule of increasing 3% for every 1% increase in inclusion rate; in the crosslinking stage, a crosslinker is added to the active ingredient inclusion product, and according to the viscosity-temperature relationship model, the quantitative injection of the crosslinker is triggered when the viscosity reaches 50cP. All parameters are coordinated between devices through the OPC-UA protocol.

[0072] Based on the temperature control data from step 103, the system dynamically adjusts process parameters: ① The lipid premixing stage is extended by 8 minutes to ensure complete melting (temperature stabilized at 75±0.3°C); ② During the inclusion stage, the inclusion efficiency is increased from 84% to 92%, and the stirring energy input is increased by 18%, based on infrared spectroscopy results; ③ During the crosslinking stage, when the viscosity sensor data reaches 45 cP, the crosslinker is precisely triggered to be injected in three pulses, separated by 15 seconds. These adjusted process parameters are mapped to the central controller in real time to ensure coordinated operation of the three stages.

[0073] 105. Based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, reversely correct the degradation kinetic parameters in the multi-parameter association model, and iteratively update the parameter matrix until the product meets the preset quality indicators.

[0074] Real-time detection data includes quality indicators such as active ingredient content from online chromatography analysis and particle size distribution from laser diffraction measurement; quality indicators are preset key product characteristic standards, such as the lower limit of active ingredient retention rate, dissolution range and other acceptance criteria.

[0075] In the embodiment of the present application, samples are taken at three time points for rapid testing in each production batch, wherein the HPLC analysis time is less than 3 minutes, and real-time detection data is obtained, and the real-time detection data is differentially analyzed with the model prediction value. When the deviation of the active ingredient retention rate exceeds 1.5%, the particle swarm optimization algorithm is used to adjust the degradation kinetic parameters, such as correcting the activation energy by ±2%, and recalculating the phase change trigger threshold. After the updated degradation kinetic parameters are simulated and verified by the digital twin system, they are pushed to the production line to execute the parameter matrix update until the product meets the preset quality indicators.

[0076] At the end of production, online HPLC testing showed a CoQ10 retention rate of 95.2%, with a preset target of ≥95%. However, the laser particle size analyzer found that 5% of the particles exceeded the limit, i.e., >80μm. The system traced the parameter matrix and found that the temperature in zone 3 during the crosslinking phase briefly fluctuated to 73°C, 2°C below the gel point. A reverse correction process was initiated: ① Optimizing degradation kinetic parameters, such as Ea from 82.3kJ / mol to 84.1kJ / mol; ② Adjusting the crosslinker injection logic to add a temperature-viscosity dual threshold trigger condition; ③ After re-simulation, the particle size D90 of the secondary production batch was optimized from 78μm to 52μm, and the activity retention rate stabilized at 96.1%.

[0077] In summary, steps 101 to 105 achieve full-process optimization of the production process of cardiovascular health products through a multi-step closed-loop collaborative mechanism. Take the production of coenzyme Q10 liposomes as an example: the parameter matrix is used to capture thermal field anomalies in the mixing container in real time, such as local high-temperature areas, and the multi-parameter correlation model is combined to dynamically analyze the coupling relationship between thermal stability, phase change and oxidative decay, significantly reducing the risk of thermal degradation of active ingredients. Based on the synergistic effect of differential stirring and gradient temperature control, the main helix strengthens local mixing and the secondary helix suppresses temperature stratification, thereby improving the uniformity of material thermal distribution. At the same time, the precise connection between lipid melting, inclusion and cross-linking stages is achieved through process parameter mapping. Real-time detection data is used to reversely correct model parameters such as degradation kinetics correction and phase change threshold adjustment, forming an iterative optimization chain of "abnormal warning-process control-quality verification" to ensure quality consistency between batches.

[0078] In some embodiments, based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer is established through a deep learning algorithm. The multi-parameter correlation model generates differential speed control instructions for the double-helix stirrer by optimizing the analytical rules of the thermodynamic properties corresponding to the active ingredients, including:

[0079] 201. Based on the parameter matrix, the thermal stability layer defines the thermal state maintenance capability of each spatial unit through the temperature gradient and time fluctuation range, the phase change layer defines the heat transfer path and triggering conditions through the temperature difference and change rate of adjacent units, and the antioxidant attenuation layer defines the cumulative attenuation effect of the thermal field on the active components through the temperature extreme value and duration. The output characteristics of the thermal stability layer, phase change layer and antioxidant attenuation layer are integrated to construct a multi-parameter correlation model;

[0080] The parameter matrix contains structured data such as temperature gradient, time fluctuation range, temperature difference between adjacent units, change rate, temperature extremes and duration; the thermal stability layer is defined as evaluating the ability of each unit to maintain its own temperature stability through the temperature gradient, i.e., the spatial temperature change rate, and the time fluctuation range, i.e., the temperature standard deviation in the time dimension; the phase change layer is defined as identifying the dynamic characteristics of the heat transfer path based on the temperature difference between adjacent units, i.e., the absolute value of the temperature difference between units, and the change rate, i.e., the rate of change of the temperature difference over time; the antioxidant attenuation layer is defined as calculating the degree of thermal damage to the active ingredient using the temperature extremes, i.e., the historical highest temperature, and the duration, i.e., the accumulated time of exceeding the threshold temperature.

[0081] In the embodiment of the present application, the parameter matrix is first input into the thermal stability layer constructed by the convolutional neural network, the spatial characteristics of the temperature gradients in different regions are extracted through the sliding window, and the thermal state maintenance coefficient matrix is generated in combination with the time fluctuation range; then the spatial characteristics of the temperature gradient characteristics and the temperature difference of adjacent units in the parameter matrix are input into the phase change layer constructed by the graph neural network, a heat conduction relationship diagram between units is established, the change rate weight is calculated using the attention mechanism, and the probability distribution of the heat transfer path is generated; then the heat flow distribution output by the phase change layer and the temperature extreme values in the parameter matrix are input into the anti-oxidation attenuation layer constructed by the LSTM network, and the attenuation index of the cumulative attenuation effect is obtained through time series analysis; finally, the thermal state maintenance coefficient matrix, probability distribution and attenuation index output by the three levels are tensor-spliced through the feature fusion layer, and a multi-parameter correlation model is formed after dimensionality reduction by the fully connected network. The output of the thermal stability layer is used as the spatial thermal resistance feature, the output of the phase change layer is used as the conduction path feature, and the output of the anti-oxidation attenuation layer is used as the damage feature. The three together constitute a three-dimensional dynamic representation of the thermal field.

[0082] 202. Establishing thermodynamic property analysis rules between each layer of the multi-parameter joint model. Based on the thermodynamic property analysis rules, a threshold value is used to determine whether the thermal state maintenance capacity between the thermal stability layer and the phase change layer is insufficient. If the threshold value is lower than the threshold, a compensating heat flow is added to the phase change layer. A critical value is used to determine whether the compensating heat flow between the phase change layer and the antioxidant attenuation layer triggers a reduction in the cumulative attenuation effect. The layer parameters are then iteratively modified in reverse order to meet the thermal state boundary of the active component.

[0083] The thermodynamic property analysis rules include a threshold judgment module for comparing the thermal state maintenance coefficient with a preset threshold, a compensation heat flow calculation module for generating a compensation heat value based on the difference, a critical value evaluation module for determining whether the compensation heat triggers attenuation reduction, and a reverse iterative correction module for parameter feedback optimization; the thermal state boundary is defined as the temperature range constraint condition for maintaining the effectiveness of the active ingredient.

[0084] In this embodiment, a data channel is first established between layers. The thermal state retention coefficient of each unit output by the thermal stability layer is compared with a preset threshold value, where the preset threshold value is calibrated through material heat capacity experiments. When the coefficient falls below the threshold, the compensation mechanism of the phase change layer is triggered, and a linear interpolation algorithm is used to calculate the required compensation heat flow. The compensation value is proportional to the difference between the threshold value and the actual coefficient. After receiving the compensation instruction, the phase change layer adjusts the original heat transfer path using a heat flow redistribution algorithm and transmits the adjusted heat flow distribution to the antioxidant attenuation layer. The antioxidant attenuation layer calculates a new cumulative attenuation effect based on the compensated heat flow. When the attenuation index exceeds a critical value, where the critical value is determined through active ingredient half-life experiments, a backpropagation mechanism is initiated: the attenuation excess signal is transmitted back to the thermal stability layer and the phase change layer via a residual connection. The gradient descent method is then used to iteratively adjust the weight parameters of the thermal state retention coefficient and the allocation parameters of the phase change layer's attention mechanism until the compensated heat flow satisfies the thermal state boundary and the attenuation index falls below the critical value. This process forms a closed-loop optimization process, and each iteration updates dynamic parameters such as the temperature gradient and rate of change in the parameter matrix.

[0085] 203. Generate the speed increment and action duration of the specified area based on the spatial distribution of the compensation heat flux, generate the speed difference upper limit and switching frequency of adjacent areas based on the attenuation reduction ratio, generate the differential speed control instruction of the double-helix stirrer, and verify its consistency with the thermal field hierarchical structure.

[0086] The speed increment is defined as the number of revolutions increased by the agitator blades per unit time in a specific area; the action duration corresponds to the duration of the compensated heat flow; the upper limit of the speed difference is set as the maximum allowable speed difference between adjacent blades; the switching frequency refers to the time interval for speed difference adjustment; consistency verification uses thermal-mechanical coupling simulation to compare the thermal field distribution and mechanical motion parameters.

[0087] In the embodiment of the present application, the spatial distribution matrix of the compensation heat flux is first input into the speed mapping model: the regional clustering algorithm is used to divide the high compensation demand area, and the speed increment is generated proportionally according to the compensation value. For example, the compensation value increases by 10W / m 2 The corresponding speed is increased by 50 rpm, and the duration of the action is determined by the time decay curve of the compensated heat flux; then, based on the reduction ratio of the output of the antioxidant attenuation layer, for example, a 30% reduction in cumulative attenuation, the upper limit of the speed difference between adjacent areas is calculated through a constrained optimization algorithm, such as not exceeding 200 rpm and the switching frequency, such as adjusting once every 5 seconds; the speed parameters are then input into the PID controller of the double-helix stirrer to generate a differential control instruction containing the speed sequence, phase difference and switching timing; finally, a three-dimensional thermal-mechanical field coupling model is established through the ANSYS simulation platform to verify whether the convective heat transfer effect generated by the stirrer motion trajectory matches the thermal field hierarchical structure predicted by the multi-parameter correlation model. If there is a deviation, the coefficient matrix of the speed mapping model is recalibrated.

[0088] Here's a specific example:

[0089] In the continuous synthesis process of the active ingredient of the antiviral oral liquid, a certain intelligent reactor system achieves precise temperature control through multi-level thermal field regulation. When the temperature field in the reactor is disturbed during the reactant feeding stage, the system's real-time monitoring shows that the temperature gradient of the reaction unit in area B drops to 0.6°C / cm, of which the threshold is 0.8°C / cm, and the maximum temperature difference between adjacent units due to uneven mixing of raw materials is 3.2°C. In view of the thermosensitive characteristics of the active ingredient, the system selects two groups of spare heating units in area C: by comparing the compensation heat flow demand with the warning value of the antioxidant attenuation layer, it is confirmed that the No. 3 heating unit can stabilize the cumulative thermal damage index below the safety threshold of 1.5 when the power is increased to 1500W. The phase change layer heat conduction model predicts that the phase change reaction in area D in the next 15 minutes will cause a sudden increase in heat flux density, causing the temperature fluctuation amplitude in this area to increase to 0.8°C. The system comprehensively evaluated thermal state maintenance capabilities, compensated heat flux distribution schemes, and attenuation risk levels. It generated differential speed instructions for the twin-helix agitator for the control system, increasing the speed of the upper blades in Zone B to 850 rpm for forced convection (with a baseline of 600 rpm). It also limited the speed difference between adjacent blades in Zone D to no more than 120 rpm to prevent local overheating. The resulting dynamic control scheme maintained temperature fluctuations in the reactor core within ±0.3°C, increasing the active ingredient synthesis efficiency from 120g per batch to 260g, validating the collaborative optimization capabilities of multi-parameter correlation models in complex thermodynamic environments.

[0090] This solution achieves full-dimensional thermodynamic precision control of the mixed reaction system by constructing a multi-parameter correlation model and dynamic compensation mechanism. Based on the deep learning architecture, the thermal stability layer, phase change layer, and antioxidant attenuation layer work together to form a three-dimensional thermal field dynamic analysis capability, which can detect local temperature anomalies in real time and predict the degradation risk of active ingredients. The adaptive compensation mechanism established through the thermodynamic analysis rules between layers can intelligently generate differential stirring instructions and temperature gradient compensation schemes, effectively suppressing the phase change runaway and oxidation loss of materials. The resulting closed-loop control system significantly improves the stability of the production process, ensures the chemical integrity of active ingredients in complex thermal environments, and optimizes mixing efficiency through the dynamic matching of mechanical motion and thermal field distribution, providing an intelligent solution for the large-scale production of high-value-added pharmaceutical preparations.

[0091] In some embodiments, thermal field temperature parameters of active ingredients in cardiovascular and cerebrovascular health products are collected in a mixing container, and a parameter matrix is constructed by combining the degradation kinetic parameters and phase transition state parameters corresponding to the active ingredients, including:

[0092] 301. Divide the mixing container into a plurality of independent areas marked by three-dimensional coordinates, and collect thermal field temperature parameters of the independent areas in real time using a fixing device;

[0093] The independent area refers to the cubic unit divided by the three-dimensional coordinate system (x, y, z) in the mixing container, and the volume of each unit is 0.5m 3 The fixed device refers to an annular temperature sensor array installed on the inner wall of the container, which includes an infrared temperature measurement module for monitoring the surface temperature and an optical fiber probe for monitoring the internal temperature. The thermal field temperature parameters include temperature value, spatial gradient, i.e., temperature difference between adjacent units and standard deviation of time fluctuation.

[0094] In the embodiment of the present application, the mixing container is first divided into 120 independent areas by a laser positioning system, and a three-dimensional coordinate label is set for the center point of each area, such as the A12 area coordinate (2.4, 3.0, 1.8); then the temperature measurement unit of the fixed device is started, and the temperature of each area is synchronously collected at a sampling frequency of 10 times per second. The sensor blind area data is compensated using a spatial interpolation algorithm; finally, a thermal field temperature parameter set containing the temperature value of each area and the gradient is generated for calculating the temperature extreme difference of adjacent 3×3×3 units and the temperature standard deviation within a 5-minute period. This data is stored with a timestamp and association to form the topological structure of the initial thermal field temperature parameters.

[0095] 302. The decomposition rate of the active ingredient determined based on the corresponding relationship between temperature and decomposition rate, and the phase change state parameters calibrated by the container material and ingredient ratio experiment;

[0096] Degradation kinetic parameters refer to the decomposition rate of active ingredients at a specific temperature, which is calculated by the Arrhenius equation; phase change state parameters include the critical temperature of the material phase change calibrated by DSC testing and the latent heat of phase change measured by the heat flow method; component ratio experiment refers to testing the effect of different raw material ratios on the phase change temperature under laboratory conditions.

[0097] In the embodiments of the present application, first, based on the experimental data of thermal decomposition of active ingredients, such as ginkgo biloba extract, the activation energy Ea in the Arrhenius equation, such as 52.3 kJ / mol and the pre-exponential factor A, such as 1.2×10^8s-1, are fitted to establish a temperature-decomposition rate mapping table; secondly, the phase change peak of the mixed material at a heating rate of 5°C / min is determined to be 78°C by differential scanning calorimetry (DSC), and the latent heat of phase change is calculated to be 150 J / g; finally, the thermal conductivity coefficient of the container material, 316L stainless steel, 16.3 W / m·K, is integrated with the phase change state parameters to form a degradation-phase change correlation model including the temperature sensitivity coefficient, phase change triggering conditions and thermal resistance characteristics.

[0098] 303. Finally, the thermal field temperature parameters, decomposition rate and thermal field temperature data are arranged according to spatial positions and combined into a parameter matrix reflecting the state of the active ingredient.

[0099] Spatial position arrangement refers to storing the parameters of each independent area in a multidimensional array in the order of three-dimensional coordinates; the parameter matrix is a four-dimensional tensor structure, and the dimensions include space (x, y, z), time series t and parameter type, such as temperature, decomposition rate, etc.

[0100] In an embodiment of the present application, the thermal field temperature data (including gradient and fluctuation value) obtained in step 301 is first spatially aligned with the decomposition rate calculated in step 302: for each coordinate point (x_i, y_j, z_k), its temperature T_ijk is associated and the pre-stored mapping table is queried to obtain the corresponding decomposition rate k_ijk; then the phase change state parameters (critical temperature, latent heat value) are attached to the matrix metadata as global constants; finally, the temperature field and the decomposition rate field are integrated into a four-dimensional parameter matrix (120×4×30×5) through tensor splicing technology, including 120 spatial points, 4 types of parameters such as temperature value, gradient, fluctuation, decomposition rate), 30 time slices such as one frame every 5 minutes, and 5 phase change parameter expansion dimensions.

[0101] Here's a specific example:

[0102] In the Ginkgo flavonoid ester capsule production scenario, the mixing vessel has a diameter of 3 meters and a height of 2.5 meters. The vessel is first divided into 120 independent zones, each measuring 0.5×0.5×0.5 meters. The sensor array detected zone C7 at z=1.0m in layer B. The temperature at coordinates (2.0, 3.5, 1.0) rose to 82°C due to a mixing dead zone, exceeding the phase transition threshold of 78°C. The temperature fluctuated by ±1.2°C within 5 minutes. A pre-stored model was then queried based on the temperature in this zone, calculating the decomposition rate of Ginkgo flavonoid esters at this temperature, k=0.15h⁻¹, compared to a normal value of 0.03h⁻¹ at 45°C. The generated parameter matrix showed that the decomposition rate in zone C7 was four times higher than that in adjacent zones, and the latent heat of phase transition depletion resulted in a decrease in heat transfer efficiency. This triggered an alert, adjusted the agitator speed, and initiated local cooling, bringing the temperature in this zone back down to 76°C within 3 minutes.

[0103] This solution precisely locates the thermal runaway risk area in the mixing container through the spatial coupling of three-dimensional thermal field parameters and degradation dynamics, thereby shortening the response time for abnormal temperature detection; the multidimensional structure of the parameter matrix completely characterizes the thermal stability state of the active ingredient, providing a quantitative basis for dynamic regulation, and successfully reducing the thermal decomposition loss rate in the ginkgo yellow production process from the industry average; the fusion analysis of phase change state parameters and real-time temperature effectively prevents the problem of uneven mixing caused by material phase change, and controls the batch differences in the active ingredient content of the capsules within a very small range, meeting the requirements of the pharmacopoeia standards.

[0104] In some embodiments, according to the parameter matrix and the differential control instruction, the main helicoid is driven to perform a radial sweep stirring mode and the auxiliary helicoid is driven to perform an axial pulse flipping mode. At the same time, a semiconductor temperature control device is used to perform gradient compensation for temperature fluctuations in different areas of the mixing container to form temperature control data that dynamically matches the stirring rate, including:

[0105] 401. According to the differential speed control instruction, the main screw is controlled to periodically move back and forth along the radial direction of the mixing container for stirring, and the angle of the stirring blade is changed during the movement to cover the material at different radial depths. At the same time, the auxiliary screw is controlled to intermittently push the material to turn along the axial direction of the container;

[0106] The differential control instruction refers to a set of control signals including the main / auxiliary helices' rotation speed, movement period, and angle change threshold; radial periodic back-and-forth movement refers to the reciprocating motion of the main helices along the diameter of the container in a sinusoidal wave trajectory; the change in the angle of the stirring blade refers to the dynamic deflection of 0-90° achieved by driving the blade root shaft through a servo motor; and axial intermittent push refers to the pulsed motion of the auxiliary helices in which they pause for 0.5 seconds after completing three axial pushes.

[0107] In an embodiment of the present application, the main helix receives the movement parameters in the differential control instruction, wherein the frequency is 2Hz, and after the stroke is ±15cm, the slide mechanism is driven by PID closed-loop control to perform radial sweeping. When the displacement sensor detects that it has reached the end point of the stroke, the fuzzy control algorithm is triggered to adjust the blade angle, that is, the angle increases by 2° for every 1cm of movement, so that the viscosity sensor detection value is stabilized in the range of 50±5Pa·s. At the same time, the auxiliary helix is driven by a servo motor according to the pulse width modulation signal with a duty cycle of 30%, and then enters the braking state after implementing 3 continuous push strokes. Each push of 8cm takes 0.8 seconds. The next cycle is started after the push rod is confirmed to be in place by the Hall sensor, so that the auxiliary helix can intermittently push the material to turn over along the axial direction of the container. The main helix sweep trajectory data is written into the D3 buffer area of the parameter matrix in real time for the temperature control system to call.

[0108] 402. Through the semiconductor temperature control device, the temperature fluctuations in different areas of the mixing container are monitored in real time, and the cooling or heating power of different areas is dynamically adjusted according to the temperature gradient compensation requirements in the parameter matrix, so that the temperature changes are synchronously adapted to the stirring rates of the main helix and the auxiliary helix, forming temperature control data that can be dynamically matched in real time.

[0109] The semiconductor temperature control device refers to a thermoelectric module array based on the Peltier effect, which contains 32 independent temperature control units; the temperature gradient compensation requirement refers to the target temperature-actual temperature difference matrix of each area stored in the parameter matrix; the dynamic adjustment of cooling / heating power refers to the distribution of the working current of each module according to the proportional integral algorithm based on the temperature deviation value.

[0110] In an embodiment of the present application, 64 PT1000 temperature sensors arranged on the outer wall of the container collect temperature data every 200ms, monitor the temperature fluctuations in different areas of the mixing container in real time, and generate an 8×8 partition temperature field after Kalman filtering. Comparing the preset gradient temperature curves in the parameter matrix, such as 65°C ± 0.5 in the central area and 60°C ± 1 in the edge area, the model predictive control (MPC) algorithm is used to calculate the temperature gradient compensation requirements of each area. When the main spiral speed is increased to 120rpm, the temperature control system synchronously increases the cooling power of the 3×3 modules in the central area from 150W to 300W, and at the same time switches the edge area to heating mode, with 50W increments / second. The temperature control data forms a closed loop with the stirring rate through the CAN bus to achieve a dynamic tracking accuracy of ±0.3°C.

[0111] In some embodiments, the progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking are mapped to the temperature control data, and the mapped comprehensive process parameters are fed back to the multi-parameter correlation model, including:

[0112] 501. Mixing the lipid carrier with the auxiliary component to form a lipid carrier premix product, and recording temperature change data during the mixing process to obtain a lipid carrier premix temperature sequence;

[0113] The lipid carrier premix product refers to a uniform micellar solution formed by phospholipids, cholesterol and surfactant at 60-70°C; the mixing temperature change data refers to a time series data set containing temperature fluctuation values per minute and phase change inflection point timestamps; the lipid carrier premix temperature sequence refers to the temperature-viscosity correlation data chain of the entire mixing process recorded in the ISO 14644 standard format.

[0114] In an embodiment of the present application, in a twin-screw melt mixer, the initial heating power is controlled to 2.5kW to heat the lipid carrier to a critical phase transition point of 65°C. The lipid carrier is mixed with the auxiliary components to form a lipid carrier premix product; when the infrared spectrometer detects that the intensity of the cholesterol characteristic peak wave number 2940cm-1 is stable, the high shear emulsification module is turned on at 12000rpm. The temperature data of 32 points in the mixing chamber are collected every 0.5 seconds through an embedded thermocouple array, and after removing the equipment vibration interference by using a wavelet denoising algorithm, a three-dimensional sequence matrix containing timestamps, spatial coordinates, and temperature values is generated. The three-dimensional sequence matrix is stored as a lipid carrier premix temperature sequence by the time series compression encoding LZ77 algorithm.

[0115] 502. Adding the active ingredient to the lipid carrier premix product, controlling the addition rate to form an active ingredient inclusion product, and recording temperature change data during the inclusion process to obtain an active ingredient inclusion temperature sequence;

[0116] The active ingredient addition rate refers to the drug powder addition acceleration controlled by the peristaltic pump, in mg / s; the inclusion temperature series refers to the exothermic peak data generated by hydrogen bonding when the drug molecules are embedded in the lipid micelles.

[0117] In an embodiment of the present application, when the lipid carrier premix product reaches the particle size threshold value PDI <0.3 detected by a dynamic light scattering instrument, the PID-controlled peristaltic pump is started to add ibuprofen micropowder at a gradient rate: the addition rate is linearly increased from 50 mg / s in the first 30 seconds to 300 mg / s to form an active ingredient containing product. A differential scanning calorimeter DSC is used to monitor the enthalpy change of the system in real time. When a characteristic exothermic peak is detected, wherein the peak temperature is 68.5±0.2°C, the feedback control is triggered to reduce the pump speed by 20%. The ultraviolet absorption spectrum obtained synchronously by the fiber optic spectrometer, wherein the absorbance change rate at λ=273nm, is fused with the temperature change data obtained by the temperature sensor to generate an active ingredient inclusion temperature sequence including the drug loading rate.

[0118] 503. Add a crosslinking agent to the active ingredient inclusion product, adjust the dropwise addition rate, and record the temperature change data during the crosslinking process to obtain a crosslinking temperature sequence for the sustained-release matrix;

[0119] The cross-linking agent addition acceleration refers to the injection rate of the glutaraldehyde solution dynamically adjusted based on the pH value of the reaction system; the cross-linking temperature sequence refers to the data recording the temperature mutation points caused by the condensation reaction during the covalent cross-linking process.

[0120] In an embodiment of the present application, when the Zeta potential of the inclusion product reaches -35mV, the microfluidic chip is started to control the injection of the cross-linking agent. The initial dripping acceleration is 5μL / s. When the near-infrared sensor wavelength 1450nm detects a sudden change in the turbidity of the system, it is switched to a pulse width modulation mode, wherein the pulse width modulation mode is a 0.1Hz square wave with a duty cycle of 45%. The characteristic soundprint signal of the cross-linking process is captured by a distributed acoustic wave sensor DAS, in which the frequency band is 5-8kHz. When a temperature rise rate of 0.5°C / s is detected in combination with the temperature change data, cooling compensation is automatically triggered to generate a sustained-release matrix cross-linking temperature sequence containing a cross-linking degree evaluation value.

[0121] 504. Compare the lipid carrier premixing temperature sequence, active ingredient inclusion temperature sequence, and sustained-release matrix crosslinking temperature sequence with the previously acquired temperature control data, establish a temperature control mapping relationship, and adjust the process parameters of each stage to form a comprehensive process parameter set;

[0122] The temperature control mapping relationship refers to the correspondence between the process temperature curve and the control benchmark aligned through the dynamic time warping (DTW) algorithm; the comprehensive process parameter set refers to the parameter matrix containing the optimized temperature-time gradient, mechanical energy input threshold, and stoichiometric ratio.

[0123] In the embodiment of the present application, the lipid carrier premixing temperature sequence, the active ingredient inclusion temperature sequence, and the sustained-release matrix cross-linking temperature sequence are imported into the time series analysis platform, and the dynamic time warping DTW algorithm is used to perform pattern matching with the historical optimal temperature control data, wherein the historical optimal temperature control data is stored in the golden profile table of the MySQL relational database. When it is detected that the temperature rise slope deviation in the premixing stage exceeds 15%, the particle swarm optimization algorithm PSO is automatically called to recalculate the heating power curve. In view of the exothermic delay phenomenon detected in the inclusion stage, the starting threshold of the cooling system in the subsequent process is adjusted, that is, changed from ΔT=2°C to ΔT=1.5°C. Finally, a comprehensive process parameter set containing 328 optimization parameters is generated, and virtual verification is performed through a digital twin model.

[0124] 505. Input the comprehensive process parameter set into the multi-parameter association model to adjust the execution strategy of subsequent process stages.

[0125] The multi-parameter association model refers to a process parameter-product quality prediction model built based on the deep neural network DNN; execution strategy adjustment refers to the dynamic correction of downstream parameters such as drying temperature and centrifugal speed based on the encapsulation rate prediction value output by the model.

[0126] In the embodiment of the present application, the comprehensive parameter set is input into a pre-trained 3-layer LSTM neural network, where the training data covers 100,000 historical batches. The model outputs a predicted encapsulation efficiency of 91.2±0.8% and a sudden release risk coefficient of 0.23 for sustained-release microspheres. When the risk coefficient exceeds 0.3, the trigger strategy adjustment is triggered: the spray drying inlet air temperature is reduced from 85°C to 78°C, and the centrifugal separation speed is increased from 8000rpm to 12000rpm. At the same time, the corrected parameters are written into the process route card of the MES system in real time to ensure that the parameters are updated synchronously at subsequent production stations.

[0127] In some embodiments, based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, the degradation kinetic parameters in the multi-parameter correlation model are reversely corrected, and the parameter matrix is iteratively updated until the product meets the preset quality indicators, including:

[0128] 601. Obtain the active ingredient concentration and state parameters of different areas through a real-time detection device, and compare them with the preset quality indicators item by item to generate difference results;

[0129] The real-time detection device refers to an online analysis system that integrates Raman spectroscopy, near-infrared sensors and microscopic imaging, which is used to capture the concentration gradient, crystal state and dispersion uniformity of the active ingredient; the state parameters refer to multiple dimensions including the proportion of dissolved state, crystallinity index, and proportion of chemical degradation products; the difference results refer to the deviation quantification matrix between the real-time detection value and the preset quality indicators, where the preset quality indicators include the degradation product threshold and the purity of the active ingredient.

[0130] In an embodiment of the present application, an online Raman probe collects spectral features of the mixed system every 20 seconds, calculates the concentration of the active ingredient by using the PLS partial least squares regression model, and simultaneously captures the crystalline morphology of the particles using a microscopic high-speed camera with a frame rate of 1000fps. When it is detected that the proportion of dissolved state in a certain area is lower than a preset threshold, where the target value is ≥95% and the measured value is 89%, the system generates a difference matrix of state parameters including spatial coordinates, a deviation amplitude of -6%, and a timestamp. The difference matrix identifies abnormal patterns, such as degradation caused by local overheating, through a convolutional neural network (CNN) classifier, triggering a subsequent parameter correction process.

[0131] 602. Based on the difference result and the comprehensive process parameters, reversely adjust the active ingredient decomposition rate parameter and degradation kinetic parameter preset in the multi-parameter association model;

[0132] The decomposition rate parameter refers to the coefficient of the Arrhenius equation that describes the chemical degradation rate of the active ingredient under a specific temperature / shear force; reverse adjustment refers to the reverse calculation of the optimization direction and amplitude of the kinetic model parameters based on the difference results.

[0133] In an embodiment of the present application, the difference matrix is input into an optimization engine based on Bayesian inference, and the degradation kinetics equation is solved in reverse in combination with the current comprehensive process parameters, such as temperature of 65°C and stirring power of 1.2kW: the Markov chain Monte Carlo (MCMC) method is used to generate candidate values of potential parameter combinations such as activation energy Ea; the degradation trajectory is simulated using the candidate parameters, and the KL divergence Kullback-Leibler divergence is compared with the measured difference results; the parameter combination with the smallest divergence is selected, such as Ea is corrected from 80kJ / mol to 73kJ / mol, the kinetic module in the multi-parameter association model is updated, and degradation kinetic parameters adapted to the current production conditions are generated.

[0134] 603. Substitute the adjusted degradation kinetic parameters into the multi-parameter association model, update the values of the corresponding areas in the parameter matrix, and repeat the detection, comparison, adjustment, and update steps until the state parameters of the active ingredient meet the preset quality indicators.

[0135] Parameter matrix update refers to mapping the corrected kinetic parameters to the control logic table of the corresponding process area, where the corresponding process area includes the high-temperature area and the high-shear area; cyclic execution refers to implementing physical system adjustments after pre-verifying the validity of the parameters through the digital twin model.

[0136] In an embodiment of the present application, the corrected degradation kinetic parameters are written into the D5 partition of the parameter matrix through the OPC-UA protocol, triggering the following iterations: the digital twin system loads the new parameters to simulate the degradation trend under the process conditions in the next 15 minutes; if the predicted maximum degradation rate still exceeds the threshold, such as >5%, a secondary correction instruction is automatically generated, such as reducing the stirring power by 10%; the updated parameters are sent to the PLC controller to adjust the reaction tank temperature setting value and the agitator speed; step 601 is re-executed until the dissolved proportion of active ingredients in all regions is ≥95% and the degradation product is <0.3%, that is, all meet the preset quality indicators.

[0137] In some embodiments, thermodynamic property analysis rules are established between the various levels of the multi-parameter joint model. Based on the thermodynamic property analysis rules, a threshold value is used to determine whether the thermal state maintenance capacity between the thermal stability layer and the phase change layer is insufficient. If the threshold value is lower than the threshold, a compensating heat flow is added to the phase change layer. A critical value is used to determine whether the compensating heat flow between the phase change layer and the antioxidant attenuation layer triggers a reduction in the cumulative attenuation effect, and the hierarchical parameters are reversely iterated and corrected to meet the thermal state boundary of the active component, including:

[0138] 701. In the multi-parameter joint model, define the threshold conditions for heat transfer from the thermal stability layer to the phase change layer, and the critical conditions for heat transfer from the phase change layer to the anti-oxidation attenuation layer, so as to establish analytical rules for thermodynamic characteristics between the layers;

[0139] The threshold condition refers to the minimum allowable value of heat transfer from the thermal stability layer to the phase change layer, such as temperature difference ≥ ΔT1 or heat flux density ≥ Q_min, which is derived from the thermodynamic equilibrium equation; the critical condition refers to the maximum safe threshold for heat transfer from the phase change layer to the antioxidant attenuation layer, such as temperature ≤ T_max or heat flux density ≤ Q_safe, which is used to prevent material phase change failure; thermodynamic property analysis rules: include constraint equations for heat transfer between layers and state switching logic.

[0140] In this embodiment, an embedded thermocouple array monitors the temperature gradient distribution of the thermally stable layer in real time, while a heat flux sensor collects the interface heat flux Q_current. The embedded thermocouple array has a sampling frequency of 10 Hz. A threshold Q_min = λ × ΔT / Δx is calculated using the thermodynamic equilibrium equation, where λ is the material thermal conductivity, ΔT is the allowable temperature difference, and Δx refers to the characteristic thickness at the interface between the thermally stable layer and the phase change layer, specifically the geometric dimensions of the effective heat conduction path in the thermally stable layer. A state flag Flag = 1 is generated when the interface heat flux Q_current is less than the threshold Q_min. The latent heat release rate V_pcm of the phase change layer is measured online using a differential scanning calorimeter (DSC). If the latent heat release rate V_pcm > V_max, with a preset upper limit of 180 J / s, an attenuation coefficient α is calculated as 1-(V_pcm - V_max) / V_max. A thermodynamic state matrix is generated, containing the interface heat flux Q_current, the state flag Flag, and the attenuation coefficient α, and is input into the finite state machine (FSM) to trigger the hierarchical switching logic.

[0141] 702. Based on the threshold condition, if the ability of the thermal stability layer to maintain the thermal state of the phase change layer is lower than a preset minimum heat value, then calculate and add additional heat compensation for the phase change layer by dividing the current heat gap of the thermal stability layer by a preset compensation ratio;

[0142] The heat gap refers to the difference ΔQ = Q_min - Q_current between the heat required by the phase change layer and the actual heat transferred when the threshold condition is not met; the preset compensation ratio refers to the compensation coefficient β set according to the heat capacity characteristics of the phase change material, which is used to dynamically adjust the compensation amount.

[0143] In the embodiment of the present application, when the state flag Flag = 1 in the thermodynamic state matrix, the dynamic compensation module calls the PID controller with a proportional coefficient Kp = 0.8 and an integral time Ti = 5s. The heat gap ΔQ is calculated as Q_min - Q_current, and the compensation coefficient β = 0.6 is set based on the heat capacity characteristics of the phase change material to generate a compensation heat flow Q_comp = β × ΔQ. The compensation heat flow Q_comp is added as an additional term to the heat transfer equation of the phase change layer:

[0144]

[0145] By COMSOL Solve the corrected temperature field. If the deformation of the phase change layer detected by the laser displacement sensor exceeds ±0.1 mm, reduce the compensation coefficient β to 0.4 and recalculate the compensation heat flux Q_comp.

[0146] 703. After adding the additional heat compensation, based on the critical condition, if the heat transferred from the phase change layer to the anti-oxidation attenuation layer exceeds the preset cumulative attenuation safety value, the heat compensation value of the phase change layer is reduced by the excess ratio, and the reduced amount is transferred back to the thermal stability layer to update the thermal parameters of the thermal stability layer;

[0147] The cumulative attenuation safety value refers to the upper limit of the thermal fatigue life of the anti-oxidation attenuation layer, which is calculated by the Arrhenius equation; the excess ratio refers to the percentage of the actual heat transfer exceeding the critical value γ = (Q_transfer-Q_safe) / Q_safe.

[0148] In the embodiment of the present application, an infrared thermal imager with a resolution of 0.1°C monitors the heat flux Q_transfer output from the phase change layer to the anti-oxidation attenuation layer. If the heat flux Q_transfer>Q_safe, where the safety value calculated by the Arrhenius equation is 45W, the excess ratio γ is calculated as (Q_transfer-Q_safe) / Q_safe×100%. The compensation amount is dynamically adjusted by the reduction factor η=1 / (1+γ / 10), the compensation heat flux Q_comp'=Q_comp×η is updated, and the reduction amount ΔQ_reduce=Q_comp×(1-η) is fed back to the control unit of the thermal stability layer. The thermal conductivity k_stable of the thermal stability layer is iteratively corrected using a gradient descent method with a learning rate α=0.01 until the heat flux Q_transfer≤Q_safe±2%.

[0149] 704. Reversely iterate and modify the parameters of each layer to meet the thermal boundary of the active component until the thermal parameters of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer all meet the preset range of the thermal boundary of the active component.

[0150] The thermal boundary refers to the operating temperature range of the active component [T_low, T_high]; reverse iterative correction refers to the reverse adjustment of hierarchical parameters such as material thermal conductivity and latent heat of phase change based on the thermal boundary as a constraint condition.

[0151] In the embodiment of the present application, the thermal boundary of the active ingredient [T_low = 45°C, T_high = 50°C] is constrained to construct an objective function F = 0.7×|T_active-47.5°C|+0.3×σ_anti, where σ_anti is the thermal stress of the antioxidant layer. The genetic algorithm (population size 50, mutation rate 0.1) optimizes the following parameters: thermal stability layer cooling power P_cool (range 30-80W), phase change layer latent heat L_pcm 150-200kJ / k), and antioxidant layer thickness d_anti 0.3-1.0mm. During the iterative process, if T_active = 52°C exceeds the standard, the algorithm selects the optimal solution L_pcm = 185kJ / kg, P_cool = 55W, and d_anti = 0.6mm, so that T_active converges to 48.2°C and σ_anti drops to 15MPa, meeting the termination condition of |T_active-47.5°C|≤1°C.

[0152] Figure 2 The present invention provides a schematic diagram of a control system for the production of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization. Figure 2 As shown, the system includes:

[0153] The acquisition module 21 is used to collect the thermal field temperature parameters of the active ingredients in the cardiovascular and cerebrovascular health products in the mixing container, and construct a parameter matrix based on the degradation kinetic parameters and phase change state parameters corresponding to the active ingredients;

[0154] A generation module 22 is configured to establish, based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer using a deep learning algorithm, wherein the multi-parameter correlation model generates differential speed control instructions for the double-helix stirrer by optimizing the analytical rules for the thermodynamic properties corresponding to the active ingredients;

[0155] An execution module 23 is configured to drive the main helicoid to execute a radial sweep stirring mode and the auxiliary helicoid to execute an axial pulse flipping mode according to the parameter matrix and the differential speed control instruction, and simultaneously perform gradient compensation for temperature fluctuations in different areas of the mixing container through a semiconductor temperature control device to form temperature control data that dynamically matches the stirring rate;

[0156] a mapping module 24 for mapping the progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking to the temperature control data, and feeding back the mapped comprehensive process parameters to the multi-parameter association model;

[0157] The correction module 25 is used to reversely correct the degradation kinetic parameters in the multi-parameter association model based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, and iteratively update the parameter matrix until the product meets the preset quality indicators.

[0158] Figure 2 The cardiovascular health product production process control system based on multi-parameter collaborative optimization can be executed Figure 1 The implementation principle and technical effects of the method for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization described in the illustrated embodiment will not be elaborated on here. The specific manner in which each module and unit performs operations in the above-mentioned embodiment of a control system for the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization has been described in detail in the embodiments of the method and will not be elaborated on here.

[0159] In one possible design, Figure 2 The embodiment shown is a cardiovascular health product production process control system based on multi-parameter collaborative optimization, which can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0160] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0161] The processing component 32 is used for the above Figure 1 The embodiment provides a method for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization.

[0162] The processing component 32 may include one or more processors to execute computer instructions to perform all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0163] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0164] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0165] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0166] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0167] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0168] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The embodiment shown is a method for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization.

[0169] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0170] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0171] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0172] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been 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. However, these 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 application.

Claims

1. A method for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization, characterized in that: include: Collecting thermal field temperature parameters of active ingredients in cardiovascular and cerebrovascular health products in a mixing container, and constructing a parameter matrix based on the degradation kinetic parameters and phase change state parameters corresponding to the active ingredients; Based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer is established through a deep learning algorithm. The multi-parameter correlation model generates differential speed control instructions for the double-helix stirrer by optimizing the analytical rules of the thermodynamic properties corresponding to the active ingredients; According to the parameter matrix and the differential control instruction, the main helix is driven to perform a radial sweep stirring mode and the auxiliary helix is driven to perform an axial pulse flipping mode. At the same time, the semiconductor temperature control device is used to perform gradient compensation for temperature fluctuations in different areas of the mixing container to form temperature control data that dynamically matches the stirring rate. Mapping the progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking to the temperature control data, and feeding the mapped comprehensive process parameters back to the multi-parameter correlation model; Based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, the degradation kinetic parameters in the multi-parameter association model are reversely corrected, and the parameter matrix is iteratively updated until the product meets the preset quality indicators.

2. The method according to claim 1, characterized in that Based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer is established through a deep learning algorithm. The multi-parameter correlation model generates differential speed control instructions for the double-helix stirrer by optimizing the analytical rules of the thermodynamic properties corresponding to the active ingredients, including: Based on the parameter matrix, the thermal stability layer defines the thermal state maintenance capability of each spatial unit through the temperature gradient and time fluctuation range; the phase change layer defines the heat transfer path and triggering conditions through the temperature difference and change rate of adjacent units; and the antioxidant attenuation layer defines the cumulative attenuation effect of the thermal field on the active components through the temperature extreme value and duration. The output characteristics of the thermal stability layer, phase change layer, and antioxidant attenuation layer are integrated to construct a multi-parameter correlation model; Establishing thermodynamic property analysis rules between each layer of the multi-parameter joint model. Based on the thermodynamic property analysis rules, a threshold is used to determine whether the thermal state maintenance capacity between the thermal stability layer and the phase change layer is insufficient. If the threshold is lower than the threshold, a compensating heat flow is added to the phase change layer. A critical value is used to determine whether the compensating heat flow between the phase change layer and the antioxidant attenuation layer triggers a reduction in the cumulative attenuation effect, and the hierarchical parameters are reversely iterated to meet the thermal state boundary of the active component. The speed increment and action duration of the specified area are generated according to the spatial distribution of the compensating heat flux, the speed difference upper limit and switching frequency of adjacent areas are generated according to the attenuation reduction ratio, and the differential speed control instruction of the double-helix stirrer is generated, and its consistency with the thermal field hierarchical structure is verified.

3. The method according to claim 1, characterized in that The thermal field temperature parameters of the active ingredients in the cardiovascular health product are collected in the mixing container, and the degradation kinetic parameters and phase change state parameters corresponding to the active ingredients are combined to construct a parameter matrix, including: Dividing a plurality of independent areas marked by three-dimensional coordinates in the mixing container, and collecting thermal field temperature parameters of the independent areas in real time through a fixing device; The decomposition rate of the active ingredient determined based on the relationship between temperature and decomposition rate, and the phase change state parameters calibrated by the container material and ingredient ratio experiment; Finally, the thermal field temperature parameters, decomposition rate and thermal field temperature data are arranged according to spatial positions and combined into a parameter matrix reflecting the state of the active ingredient.

4. The method according to claim 1, wherein According to the parameter matrix and the differential control instruction, the main helix is driven to perform a radial sweep stirring mode and the auxiliary helix is driven to perform an axial pulse flipping mode. At the same time, the semiconductor temperature control device is used to perform gradient compensation for temperature fluctuations in different areas of the mixing container to form temperature control data that dynamically matches the stirring rate, including: According to the differential speed control instruction, the main screw is controlled to move back and forth periodically along the radial direction of the mixing container for stirring, and the angle of the stirring blade is changed during the movement to cover the materials at different radial depths. At the same time, the auxiliary screw is controlled to intermittently push the materials to turn over along the axial direction of the container; Through the semiconductor temperature control device, the temperature fluctuations in different areas of the mixing container are monitored in real time, and the cooling or heating power of different areas is dynamically adjusted according to the temperature gradient compensation requirements in the parameter matrix, so that the temperature changes are synchronously adapted to the stirring rates of the main helix and the auxiliary helix, forming temperature control data that can be dynamically matched in real time.

5. The method according to claim 1, characterized in that: Mapping the progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking to the temperature control data, and feeding the mapped comprehensive process parameters back to the multi-parameter correlation model, including: mixing the lipid carrier with the auxiliary component to form a lipid carrier premix product, and recording temperature change data during the mixing process to obtain a lipid carrier premix temperature sequence; adding the active ingredient to the lipid carrier premix product, controlling the addition rate to form an active ingredient inclusion product, and recording temperature change data during the inclusion process to obtain an active ingredient inclusion temperature sequence; adding a crosslinking agent to the active ingredient inclusion product, adjusting the dropwise addition rate, and recording the temperature change data during the crosslinking process to obtain a crosslinking temperature sequence for the sustained-release matrix; The lipid carrier premixing temperature sequence, active ingredient inclusion temperature sequence, and sustained-release matrix crosslinking temperature sequence are respectively compared with the previously acquired temperature control data to establish a temperature control mapping relationship, and the process parameters of each stage are adjusted to form a comprehensive process parameter set; The comprehensive process parameter set is input into the multi-parameter association model to adjust the execution strategy of subsequent process stages.

6. The method according to claim 1, wherein Based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, the degradation kinetic parameters in the multi-parameter association model are reversely corrected, and the parameter matrix is iteratively updated until the product meets the preset quality indicators, including: The concentration and status parameters of active ingredients in different areas are obtained through real-time detection devices, and compared with the preset quality indicators item by item to generate difference results; Based on the difference results and the comprehensive process parameters, reversely adjusting the active ingredient decomposition rate parameter degradation kinetic parameter preset in the multi-parameter association model; The adjusted degradation kinetic parameters are substituted into the multi-parameter association model, the values of the corresponding areas in the parameter matrix are updated, and the detection, comparison, adjustment and update steps are performed cyclically until the state parameters of the active ingredient meet the preset quality indicators.

7. The method according to claim 2, characterized in that Establish thermodynamic property analysis rules between each layer of the multi-parameter joint model. Based on the thermodynamic property analysis rules, a threshold is used to determine whether the thermal state maintenance capacity between the thermal stability layer and the phase change layer is insufficient. If it is lower than the threshold, a compensating heat flow is added to the phase change layer. A critical value is used to determine whether the compensating heat flow between the phase change layer and the antioxidant attenuation layer triggers a reduction in the cumulative attenuation effect, and the hierarchical parameters are reversely iterated to meet the thermal state boundary of the active component, including: In the multi-parameter joint model, the threshold conditions for heat transfer from the thermal stability layer to the phase change layer and the critical conditions for heat transfer from the phase change layer to the anti-oxidation attenuation layer are defined to establish analytical rules for thermodynamic characteristics between the layers; Based on the threshold condition, if the ability of the thermal stability layer to maintain the thermal state of the phase change layer is lower than a preset minimum heat value, then additional heat compensation is calculated and added to the phase change layer by dividing the current heat gap of the thermal stability layer by a preset compensation ratio; After adding the additional heat compensation, based on the critical condition, if the heat transferred from the phase change layer to the anti-oxidation attenuation layer exceeds the preset cumulative attenuation safety value, the heat compensation value of the phase change layer is reduced by the excess ratio, and the reduced amount is transferred back to the thermal stability layer to update the thermal parameters of the thermal stability layer; The parameters of each layer are modified by reverse iteration to meet the thermal boundary of the active component until the thermal parameters of the thermal stability layer, phase change layer and antioxidant attenuation layer all meet the preset range of the thermal boundary of the active component.

8. A cardiovascular health product production process control system based on multi-parameter collaborative optimization, characterized in that: include: An acquisition module is used to collect thermal field temperature parameters of active ingredients in cardiovascular and cerebrovascular health products in a mixing container, and to construct a parameter matrix based on the degradation kinetic parameters and phase change state parameters corresponding to the active ingredients; a generation module for establishing, based on the parameter matrix, a multi-parameter correlation model of the thermal stability layer, the phase change layer, and the antioxidant attenuation layer using a deep learning algorithm, wherein the multi-parameter correlation model generates a differential speed control instruction for the double-helix stirrer by optimizing the analytical rules of the thermodynamic properties corresponding to the active ingredients; an execution module, configured to drive the main helicoid to execute a radial sweep stirring mode and the auxiliary helicoid to execute an axial pulse flipping mode according to the parameter matrix and the differential speed control instruction, and simultaneously perform gradient compensation for temperature fluctuations in different areas of the mixing container through a semiconductor temperature control device to form temperature control data that dynamically matches the stirring rate; a mapping module for mapping progressive process parameters of lipid carrier premixing, active ingredient inclusion, and sustained-release matrix cross-linking to the temperature control data, and feeding back the mapped comprehensive process parameters to the multi-parameter association model; The correction module is used to reversely correct the degradation kinetic parameters in the multi-parameter association model based on the real-time detection data of the active ingredient and the feedback results of the comprehensive process parameters, and iteratively update the parameter matrix until the product meets the preset quality indicators.

9. A computing device, characterized in that It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the cardiovascular and cerebrovascular health product production process control method based on multi-parameter collaborative optimization as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, the method for controlling the production process of cardiovascular and cerebrovascular health products based on multi-parameter collaborative optimization as claimed in any one of claims 1 to 7 is implemented.

Images