Medicinal liquor production process simulation method

Abstract

The present application relates to the technical field of liquor production simulation, and discloses a liquor production process simulation method. The method first acquires the physical property parameter set and the production process parameter set of the raw materials of the liquor; constructs a spatial distribution model containing solvent diffusion paths and effective component dissolution trajectories through multi-scale molecular dynamics simulation; extracts the core reaction region in the model that meets specific conditions, and screens out key simulation nodes; inputs the key simulation nodes into a pre-constructed multiphase flow coupling simulation system, the system generates optimized process parameters through iterative solution, and the convergence conditions introduce concentration field uniformity constraints and phase interface stability constraints; combines the time stage division of the optimized process parameters and the production process parameter set, establishes a dynamic production process model, simulates the spatiotemporal evolution behavior of the liquor components based on the non-equilibrium thermodynamics equation, separates the concentration evolution field and the phase distribution field from the model, multi-stage reconstructs the concentration evolution field and the original solvent ratio, and outputs the final simulation result.

Description

Technical Field

[0001] This invention relates to the field of medicinal wine production simulation technology, specifically a method for simulating the medicinal wine production process. Background Technology

[0002] As an important branch of traditional Chinese medicine, the production of medicinal wines has long faced numerous challenges in terms of both the inheritance of its craftsmanship and its modernization. Traditional production processes rely heavily on experience, lacking a systematic understanding of the patterns of component changes in each stage, from raw material formulation to extraction and aging. The dissolution of active ingredients in medicinal materials is affected by various factors, including solvent properties, temperature, time, and the physical structure of the medicinal materials themselves. Subtle fluctuations in these factors between different batches often lead to variations in product quality.

[0003] Currently, some manufacturers are attempting to introduce simple parameter control methods, adjusting processes by monitoring key indicators such as temperature and concentration. However, this approach struggles to capture the dynamic evolution of components within complex systems. Microscopic behaviors such as solvent diffusion paths within medicinal materials, interactions between active ingredients and solvents, and changes at phase interfaces in multiphase systems cannot be fully represented by conventional detection methods. Furthermore, existing simulation technologies often focus on static analysis of single stages, such as simulating concentration changes only during the extraction stage, neglecting the interrelationships between different stages of the production process. This leads to significant discrepancies between simulation results and actual production.

[0004] The flow state of multiphase flow within production equipment significantly impacts component mixing efficiency. However, traditional methods struggle to effectively couple turbulence characteristics with mass transfer processes, resulting in a lack of scientific theoretical support for optimizing process parameters. These problems become more pronounced as production scales up, leading to frequent occurrences of reduced raw material utilization and increased energy consumption, hindering the standardization and efficiency of medicinal wine production. Therefore, overcoming reliance on experience and constructing a dynamic simulation system that reflects the entire production process has become a pressing issue. Summary of the Invention

[0005] The purpose of this invention is to provide a method for simulating the production process of medicinal wine, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for simulating the production process of medicinal wine, the method comprising:

[0007] Obtain the set of physical property parameters and production process parameters of the raw materials for medicinal wine production;

[0008] Based on the physical property parameter set, the physical property dataset of the medicinal materials is analyzed, and a spatial distribution model of the medicinal wine components is constructed through multi-scale molecular dynamics simulation. The spatial distribution model includes solvent diffusion paths and dissolution trajectories of effective components.

[0009] Extract the core reaction regions in the spatial distribution model that meet the concentration gradient threshold and diffusion stability conditions, and screen key simulation nodes where the solvent diffusion path length is lower than the preset path upper limit and the dissolution rate of the effective component is higher than the dissolution rate lower limit.

[0010] The key simulation nodes are input into a pre-constructed multiphase flow coupled simulation system. The multiphase flow coupled simulation system generates optimized process parameters through iterative solution of the turbulence model and mass transfer equation. Concentration field uniformity constraints and phase interface stability constraints are introduced into the convergence conditions of the multiphase flow coupled simulation system.

[0011] By combining the optimized process parameters with the time stage division of the production process parameter set, a dynamic production process model is established, and the spatiotemporal evolution behavior of medicinal wine components is simulated based on non-equilibrium thermodynamic equations.

[0012] The concentration evolution field and phase distribution field are separated from the dynamic production process model. The concentration evolution field is reconstructed in multiple stages with the original solvent ratio to output the final simulation results of the medicinal wine production process.

[0013] Preferably, the acquisition of the set of physical property parameters and the set of production process parameters for the raw materials used in the medicinal wine production includes:

[0014] The set of physical property parameters includes the types of medicinal materials, solvent ratios, and initial concentration distributions; the set of production process parameters includes the extraction temperature sequence, stirring rate sequence, and time stage divisions.

[0015] Collect microstructure images and solvent property test data of medicinal materials, and extract porosity distribution, specific surface area and solvent surface tension coefficient of medicinal materials;

[0016] The production process parameter set is divided into time stages to generate discrete time nodes for the extraction temperature sequence and rotation speed gradients for the stirring rate sequence.

[0017] By integrating the porosity distribution, specific surface area, solvent surface tension coefficient, discrete time nodes, and rotational speed gradient, a three-dimensional production process parameter matrix is ​​constructed as the input data source.

[0018] Preferably, the construction of the spatial distribution model of the medicinal wine components includes:

[0019] Molecular conformation space clustering was performed on the initial concentration distribution of the physical property parameter set to delineate the interaction domains between solvent molecules and the effective components of the medicinal material.

[0020] The topology of the solvent diffusion path is calculated based on the energy barrier distribution of the interaction domain, and a dynamic distribution field is generated with the diffusion time step as the first dimension, the spatial coordinates of the medicinal material as the second dimension, and the concentration gradient as the third dimension.

[0021] The physical consistency of the dynamic distribution field was verified by the rate of change of curvature of the solvent diffusion path and the continuity of the dissolution trajectory of the effective components.

[0022] Preferably, the extraction of the core reaction region in the spatial distribution model that satisfies the concentration gradient threshold and diffusion stability conditions includes:

[0023] Calculate the local concentration gradient magnitude and diffusion path oscillation frequency at each simulated node in the dynamic distribution field;

[0024] The concentration gradient stability index of the statistical simulation node within a preset time window is composed of a weighted integral of the gradient magnitude variance and the oscillation frequency.

[0025] Set concentration gradient threshold and diffusion stability threshold, and select simulation nodes that simultaneously satisfy the condition that the local concentration gradient magnitude exceeds the threshold and the stability index is lower than the stability threshold as the core reaction region.

[0026] Preferably, the iterative solution of the multiphase flow coupled simulation system includes:

[0027] Map the diffusion path topology of key simulation nodes to the boundary conditions of the turbulence model;

[0028] A phase interface tracking algorithm is coupled into the mass transfer equation to generate temperature correction and stirring rate correction for optimized process parameters.

[0029] The variance decay rate of the concentration evolution field is calculated by constraining the uniformity of the concentration field, and the critical point of interface breakage of the phase distribution field is detected by constraining the stability of the phase interface.

[0030] Preferably, establishing the dynamic production process model includes:

[0031] The correction amount of the optimized process parameters is aligned with the original solvent ratio in multiple stages of time.

[0032] A set of governing equations, including concentration convection terms, diffusion terms, and chemical reaction source terms, was constructed based on non-equilibrium thermodynamic equations.

[0033] The spatiotemporal evolution behavior of the governing equations is solved using an explicit-implicit hybrid numerical scheme until the interface energy of the phase distribution field tends to a steady state.

[0034] Preferably, the multi-stage reconstruction includes:

[0035] The concentration evolution field is subjected to discrete-time inverse transformation to separate its transient concentration components from its steady-state concentration components;

[0036] The initial mixing ratio of the original solvent ratio is retained, and the transient concentration component is superimposed on the initial mixing ratio according to the weighting coefficient of the time stage;

[0037] The final simulation result is generated by integrating the steady-state concentration components and the superposition results through a multi-stage reconstruction algorithm.

[0038] Preferably, the calculation of the concentration gradient stability index includes:

[0039] Extract the concentration gradient vector sequence of the simulated node at continuous time steps, and calculate the magnitude variance of the gradient vector sequence and the rate of change of the direction cosine of adjacent vectors;

[0040] Stability indices are generated by exponentially weighted moving averages of the magnitude variance and the direction cosine rate of change.

[0041] Preferably, the construction of the governing equations includes:

[0042] Introducing the permeability tensor of the pore microstructure of medicinal materials into the non-equilibrium thermodynamic equations, the permeability tensor and the concentration convection term are tensor-merged to generate a modified convection term, and the activation energy parameter of the dissolution trajectory of the effective components is associated with the chemical reaction source term.

[0043] Preferably, the allocation of the time stage weight coefficients includes:

[0044] The oscillation amplitude and frequency spectral density of the transient concentration component within the discrete time period are statistically analyzed. The weight coefficients of each stage are assigned according to the product of the oscillation amplitude and the frequency spectral density. The weight coefficients are normalized to ensure the conservation of the multi-stage superposition.

[0045] Compared with the prior art, the beneficial effects of the present invention are:

[0046] Through multi-dimensional technological integration, a refined depiction of the production process was achieved. The physical property dataset constructed based on the set of physical property parameters provides a solid foundation for subsequent molecular dynamics simulations, enabling the dissolution trajectory and solvent diffusion path of the effective components inside the medicinal materials to be clearly presented. This multi-scale simulation method breaks through the traditional vague understanding of the component migration process, allowing the laws of change at the microscopic level to be revealed.

[0047] The extraction of the core reaction region and the selection of key simulation nodes effectively focused on the critical aspects of the production process, eliminating interference from non-critical factors and making the simulation more targeted. By setting concentration gradient thresholds and diffusion stability conditions, it was ensured that the selected nodes could accurately reflect the core changes in production, providing precise input parameters for subsequent multiphase flow coupling simulations.

[0048] The multiphase flow coupled simulation system introduces concentration field homogeneity constraints and phase interface stability constraints, enhancing its ability to simulate complex flow states and mass transfer processes. The iterative solution of the turbulence model and mass transfer equations can dynamically capture the interactions within the multiphase system, allowing process parameter optimization to no longer rely on empirical predictions but rather to be naturally generated based on the dynamic equilibrium laws within the system, making parameter adjustments more closely aligned with actual production needs.

[0049] The establishment of a dynamic production process model combines optimized process parameters with time-stage division, and utilizes non-equilibrium thermodynamic equations to fully present the evolutionary behavior of medicinal wine components in the spatiotemporal dimensions. This dynamic simulation of the entire process can reflect the impact of parameter changes at different stages on the final product, demonstrating the correlation and continuity between production links.

[0050] The separation and multi-stage reconstruction of the concentration evolution field and phase distribution field further enhance the practicality of the simulation results. By combining the simulation results with the original solvent ratio, the simulation results can more intuitively reflect the changing trends of components during the production process, providing detailed references for adjusting the production process and helping to better grasp the operational rhythm of each link in actual production. Attached Figure Description

[0051] Figure 1 This is a schematic diagram illustrating the working principle of the medicinal wine production process simulation method described in this invention.

[0052] Figure 2 A flowchart for obtaining the set of physical property parameters and the set of production process parameters;

[0053] Figure 3 A flowchart for extracting the core reaction region;

[0054] Figure 4 A flowchart for iterative solution of a multiphase flow coupled simulation system;

[0055] Figure 5 This is a flowchart for multi-stage reconstruction. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Please see Figure 1 This invention provides a method for simulating the production process of medicinal wine, the method comprising:

[0058] Obtain the set of physical property parameters of the raw materials for medicinal wine production and the set of production process parameters. The set of physical property parameters includes the types of medicinal materials, solvent ratios, and initial concentration distributions; the set of production process parameters includes the extraction temperature sequence, stirring rate sequence, and time stage divisions.

[0059] Based on a dataset of physical properties of medicinal materials analyzed using a set of physical property parameters, a spatial distribution model of the medicinal wine components was constructed through multi-scale molecular dynamics simulations. This spatial distribution model includes solvent diffusion paths and dissolution trajectories of active ingredients. The construction process involves spatial clustering of the molecular conformations of the initial concentration distribution, delineating the interaction domains between solvent molecules and the active ingredients of the medicinal materials, and calculating the topological structure of the solvent diffusion paths based on the energy barrier distribution of the interaction domains. This generates a dynamic distribution field with the diffusion time step as the first dimension, the spatial coordinates of the medicinal materials as the second dimension, and the concentration gradient as the third dimension. The physical consistency of the dynamic distribution field is verified by the rate of change of curvature of the solvent diffusion paths and the continuity of the dissolution trajectories of the active ingredients.

[0060] The core reaction regions in the spatial distribution model that satisfy the concentration gradient threshold and diffusion stability conditions are extracted. This process includes calculating the local concentration gradient magnitude and diffusion path oscillation frequency of each simulated node in the dynamic distribution field, statistically analyzing the concentration gradient stability index of the nodes within a preset time window (this index is composed of the weighted integral of the gradient magnitude variance and the oscillation frequency), and selecting simulated nodes that simultaneously satisfy the condition that the local concentration gradient magnitude exceeds the preset threshold and the stability index is lower than the preset stability threshold as core reaction regions. Simultaneously, simulated nodes whose solvent diffusion path length is lower than the preset upper limit and whose effective component dissolution rate is higher than the lower limit of the dissolution rate are selected as key simulated nodes.

[0061] Key simulation nodes are input into a pre-built multiphase flow coupled simulation system. This system maps the diffusion path topology of the key simulation nodes to the boundary conditions of the turbulence model, couples a phase interface tracking algorithm into the mass transfer equation, and generates optimized process parameters (including temperature and stirring rate corrections) through iterative solutions of the turbulence model and the mass transfer equation. During the solution process, concentration field homogeneity constraints (calculating the variance decay rate of the concentration evolution field) and phase interface stability constraints (detecting the critical point of interface breakage in the phase distribution field) are introduced into the system's convergence conditions.

[0062] A dynamic production process model is established by combining optimized process parameters with the time-stage division of the production process parameter set. This model simulates the spatiotemporal evolution of medicinal wine components based on non-equilibrium thermodynamic equations. Specifically, it includes: aligning the correction amounts of optimized process parameters with the original solvent ratios across multiple time stages; constructing a set of governing equations containing concentration convection terms, diffusion terms, and chemical reaction source terms; and solving the set of governing equations using an explicit-implicit hybrid numerical scheme until the interfacial energy of the phase distribution field tends to a steady state.

[0063] The concentration evolution field and phase distribution field are separated from the dynamic production process model. A discrete-time inverse transform is performed on the concentration evolution field to separate its transient and steady-state concentration components. The initial mixing ratio of the original solvent is retained, and the transient concentration components are superimposed with the initial mixing ratio according to the time-stage weighting coefficient (which is distributed based on the product of the oscillation amplitude and frequency spectral density of the transient concentration components in each discrete time stage and then normalized). The steady-state concentration components and the superimposed results are integrated through a multi-stage reconstruction algorithm to output the final simulation results of the medicinal wine production process.

[0064] Example 1: See Figure 2 The acquisition of physical property parameter sets and production process parameter sets constitutes the basic input for process simulation.

[0065] The physical property parameter set covers the types of medicinal materials, preset solvent mixing ratios, and initial solvent concentration distribution data inside and outside the medicinal materials. The production process parameter set includes time-varying extraction temperature records, stirring speed change sequences, and multiple time stages divided according to the process specifications. During the parameter acquisition stage, microscopic structural scanning images of the medicinal material samples and solvent laboratory physical property test reports are obtained. The microscopic images are processed using image analysis algorithms to quantify and extract the distribution probability data of the three-dimensional pore structure inside the medicinal materials and the unit mass surface area value of the medicinal material tissue in contact with the solvent. The solvent physical property report provides accurate experimental measurements of the solvent surface tension coefficient. Time discretization processing is performed on the production process parameter set. For the extraction temperature sequence, it is discretized into a discrete time node sequence containing different temperature values ​​according to the set process stage switching time points; for the stirring rate sequence, a corresponding stirring speed gradient sequence is generated according to the time interval and change amplitude of the stirring speed adjustment. Porosity distribution data describes the spatial differences in the volume ratio of micropores within the medicinal material that allow solvent penetration; specific surface area data reflects the total surface area per unit volume of medicinal material available for solvent contact and dissolution of active ingredients; solvent surface tension coefficient affects the wetting and spreading properties of the solvent on the surface of the medicinal material; discrete time nodes record key temperature switching moments in process control; and rotation speed gradient characterizes the rate of change of stirring intensity over time. These five types of data are integrated and filled into a three-dimensional data structure after spatial grid matching and time axis alignment, forming a production process parameter matrix. The first dimension of this matrix indexes the spatial grid position, the second dimension indexes the discrete time nodes, and the third dimension stores the porosity, specific surface area, surface tension coefficient, and stirring speed gradient data combination at the corresponding spatiotemporal points. This three-dimensional matrix serves as the input data source for system initialization, fully characterizing the structural features of the medicinal material-solvent system and external process driving conditions in the initial stage of extraction.

[0066] Based on the initial concentration distribution data from the aforementioned set of physical property parameters, a molecular conformational spatial clustering analysis was performed. This analysis was conducted within a molecular dynamics simulation framework, locating the relative positions and interaction strengths of solvent molecules and the main active ingredient (e.g., alkaloids, flavonoids) molecular clusters in three-dimensional space at the initial moment. By calculating the intermolecular interaction potential energy, the spatial range where significant attraction or repulsion exists between solvent molecules and active ingredients was delineated, defining multiple interaction domains. Each domain is labeled by its spatial boundary coordinates and the average interaction energy intensity within it. These domains constitute the dynamic physical boundaries of the simulated solvent diffusion process. The energy barrier distribution within and at the boundaries of the domains is a key physical factor determining the diffusion kinetic path. The energy barrier refers to the lowest energy barrier that a solvent molecule must overcome to cross adjacent domains or move within a domain. Using molecular dynamics simulation data, the typical energy barrier heights and their spatial distribution at and within the domain boundaries were calculated. This energy barrier data was used to drive the calculation of the solvent diffusion path. The topological structure of the diffusion path refers to an abstract description of the path shape, connectivity, and branching states formed when solvent molecules migrate through the networked pores of the medicinal material. The topological structure calculation constructs a path mobility map based on the energy barrier height, with low-barrier regions corresponding to high-probability diffusion paths. This topological structure maps the spatial distribution of possible solvent flows over time. The calculation process outputs a dynamic distribution field data structure. This data structure has three core dimensions. The first dimension is a discretized simulation time step sequence, quantitatively recording the dynamic evolution of the diffusion process from its initial state to a certain moment. The second dimension represents the specific spatial coordinates within the physical model of the medicinal material, typically indexed by a three-dimensional grid coordinate system, precisely marking the spatial points where solvent molecules reside. The third dimension records the concentration gradient value at each spatial coordinate point, at each time step, along the mainstream direction of solvent diffusion. The concentration gradient quantifies the concentration change per unit distance at that point, serving as a measure of the diffusion driving force. The dynamic distribution field is essentially a dataset of functional mapping relationships between time, spatial location, and concentration gradient.

[0067] The geometric characteristics of a solvent diffusion path are characterized by its spatial curvature. Path curvature is defined as the reciprocal of the degree of bending of the path curve, while the rate of change of curvature refers to the rate at which the degree of bending changes along the direction of the diffusion path. In a physical simulation scenario, an ideal diffusion path should have a smooth transition; excessively drastic curvature changes usually indicate numerical errors or non-physical assumptions. The physical plausibility of a selected diffusion path is evaluated by calculating the rate of change of curvature at continuous points along that path. The dissolution trajectory of the active ingredient describes the continuous migration process of the drug's active ingredient molecules detaching from the drug matrix surface and entering the solvent matrix. Trajectory continuity requires that the dissolution path cannot have non-physical interruptions or jumps on the time and spatial axes. Verification of the physical consistency of the dynamic distribution field is equivalent to verifying the reliability of the simulation results. The verification logic analyzes whether the rate of change of curvature of typical solvent diffusion paths generated in the dynamic distribution field falls within the pre-set reasonable physical limits, and simultaneously observes whether the positional offset of the recorded active ingredient molecule dissolution migration trajectory between adjacent simulation time steps is continuous and smooth. If the rate of curvature change of the main diffusion path remains at a low level, and the trajectories of most effective components do not exhibit spatial jumps or temporal gaps, then this dynamic distribution field is deemed to meet the physical consistency requirements and can be used for the identification of core reaction regions in subsequent stages. This verification operation confirms that the molecular dynamics simulation steps are completed under reasonable physical constraints, reducing the possibility of introducing errors into the basic data in subsequent simulation stages.

[0068] Example 2: See Figure 3 The dynamic distribution field serves as the input data source, containing a time step sequence, a set of spatial coordinates, and concentration gradient information at each spatiotemporal point. For each independent simulation node in this field, physical characteristic analysis is performed. Each node is associated with a set of spatial location and concentration gradient records that evolve over time. For the state of a node at a single time step, its local concentration gradient magnitude is calculated. This magnitude represents the geometric length of the concentration gradient vector at that node at that moment, reflecting the drastic nature of the local concentration change. Simultaneously, the diffusion path oscillation frequency of the node is calculated. The quantization of the oscillation frequency is based on the rate of change of the diffusion path direction or geometric length at the node between consecutive time points, reflecting the fluctuating characteristics of diffusion dynamics. The diffusion path originates from the molecular dynamics simulation recording of solvent molecule migration trajectories.

[0069] A continuous time window is defined, covering multiple consecutive time steps. All nodes in the dynamically distributed field are traversed, and data for all time steps experienced by each node within the selected time window are collected. For each node, the concentration gradient vector sequence corresponding to each time point within the time window is extracted. This sequence is a time-dependent vector array. Based on this vector sequence, two statistical processing methods are performed. The first statistic calculates the numerical variance of the concentration gradient magnitude at the node within the time window. The variance reflects the dispersion of the concentration gradient vector length over time. Large length variations indicate instability in the local mass transfer driving force. The second statistic focuses on the directional stability of the concentration gradient vector. The rate of change of the direction cosine of the gradient vector between adjacent time steps within the time window is calculated. The rate of change of the direction cosine quantifies the rate of change of the angle of difference between the gradient vectors at two consecutive time points; rapid angle changes indicate frequent changes in the diffusion direction. These two statistics together characterize the dynamic behavior of the nodes within a specific time window.

[0070] The magnitude variance statistic and the direction cosine rate of change statistic are incorporated into the exponentially weighted moving average processing framework. An exponential decay factor that decreases over time is applied during the processing. This factor gives higher weight to the statistical data from the most recent time point in the averaging calculation, while reducing the weight of data from more distant time points. The moving average processing outputs a comprehensive numerical result, namely the concentration gradient stability index for that node. This index integrates the dual effects of gradient magnitude fluctuations and gradient direction fluctuations; a low index value indicates that the concentration gradient state at that node remains relatively stable within the time window.

[0071] Two physical constraints were set: a concentration gradient threshold and a diffusion stability threshold. The concentration gradient threshold requires that the local concentration gradient magnitude of a node reach or exceed a certain lower limit, indicating a sufficiently strong driving force for material migration at that location. The diffusion stability threshold requires that the concentration gradient stability index of a node be below a certain upper limit, indicating sufficient temporal stability of the diffusion process at that location. All simulated nodes were scanned across the entire dynamic distribution field to identify a subset of nodes that simultaneously met both conditions. Specifically, if the local concentration gradient magnitude of a node exceeds the preset concentration gradient threshold at the time of evaluation, and the concentration gradient stability index of that node is below the preset diffusion stability threshold within the same evaluation time window, then that node is marked as a candidate node. All candidate nodes meeting the conditions constitute the initially identified core reaction region set. This set reflects the key action sites within the herb-solvent system that exhibit high driving efficiency and relatively stable states.

[0072] Based on the core reaction region set, further screening is performed to identify key simulation nodes. Screening is based on two kinetic parameters directly related to extraction efficiency. The first parameter is the solvent diffusion path length. This length refers to the theoretical shortest migration distance from the bulk solvent region through the pore network to the node location. An upper limit for the path length is set, typically determined with reference to the material properties of the medicinal material and the container size. Only nodes with path lengths not exceeding this upper limit are retained. The second parameter is the simulated dissolution rate of the active ingredient at the node. The dissolution rate quantifies the rate at which active ingredient molecules desorb from the solid phase into the solvent phase per unit time. A lower limit for the dissolution rate is set, which is associated with the commercial extraction efficiency requirements of the active ingredient in the medicinal material. Only nodes with dissolution rates not lower than this lower limit are retained. For each node in the core reaction region, its associated solvent diffusion path length is checked to ensure it is within the specified range, and its simulated active ingredient dissolution rate is verified to meet the requirements. Only when both parameters of a node meet the screening criteria is the node finally identified as a key simulation node. Nodes that do not meet either condition are removed from the key set. The set of key simulation nodes represents the most technologically valuable target locations in the spatial distribution model, and its data will be directly used to drive subsequent multiphase flow coupled simulation processes. This screening step completes the shift from physicochemical stability assessment to focusing on extraction process efficiency.

[0073] Example 3: See Figure 4 After the key simulation node set is selected from the spatial distribution model, it is imported into a pre-configured multiphase flow coupled simulation system. The core design of this system lies in integrating flow field dynamics and mass transfer processes. First, the boundary conditions of the turbulence model are set based on the diffusion path topology of the key simulation nodes. The diffusion path topology includes the network connectivity, branch characteristics, and spatial location data of the path. This data is parsed and mapped to the inlet velocity distribution, wall boundary type, and initial flow field conditions of the turbulence solver. The inlet velocity is distributed as a non-uniform flow according to the path branch density to simulate the penetration trend of solvent in the material; the wall boundary is set with no-slip conditions to correspond to the constraint effect of the solid skeleton of the medicinal material; the initial flow field is initialized as a uniform flow based on topological spatial coordinates to reduce numerical divergence. Next, a phase interface tracking algorithm is integrated into the construction of the mass transfer equation. This algorithm uses the volume fraction method to track the interface position between the solvent and the residue or bubble phase. The volume fraction is defined as the volume ratio of the solvent phase within the fluid unit. The algorithm monitors the spatiotemporal gradient change of this fraction and updates the phase interface geometry in real time. The mass transfer equation is specifically expressed including convection, diffusion, and phase transition terms. The phase interface tracking module embeds the source term of this equation to capture mass exchange at the interface front. The turbulence model uses the nonlinear eddy viscosity model in the Reynolds-averaged method to handle the stochastic fluctuations within the fluid. This model outputs transient flow field variables, including velocity, pressure, and turbulent kinetic energy distribution.

[0074] The turbulence model and the mass transfer equation coupled with phase interface tracking are solved iteratively. The iterative process is divided into multiple computation cycles. In each cycle, the turbulence equation is solved first to update the flow field data, and then the latest velocity and pressure fields are imported into the mass transfer equation to calculate the concentration distribution; concentration changes are used to correct fluid density and viscosity, and the turbulence model is updated. The number of iterations is adaptively adjusted according to the convergence criterion. During the solution process, two physical constraints are introduced to monitor the convergence of the system. The first constraint is the concentration field uniformity constraint, which is quantified by the variance decay rate of the global concentration data. The variance decay rate is defined as the negative value of the derivative of the concentration variance with time; a low value indicates a tendency towards uniformity. In specific calculations, after each iteration, the concentration variance value of all grid nodes in the simulation domain is statistically analyzed, and the rate of decrease of this value with the iteration step size is tracked. The second constraint is the phase interface stability constraint, focusing on the interface integrity within the phase distribution field. The algorithm detects the extreme points of surface tension gradient in local regions of the interface. When the gradient exceeds a preset threshold, it is marked as a critical point for interface breakage. The increase in the number of such critical points indicates the risk of interface breakage. The convergence condition is met by requiring the variance decay rate to be close to the zero threshold and the number of interface breakage critical points to show no increasing trend. Upon completion of the iteration, the system outputs optimized process parameters: temperature correction and stirring rate correction. The temperature correction is calculated based on the contribution of the heat source term in the mass transfer equation, quantifying the optimal adjustment value of the leaching temperature sequence; the stirring rate correction is derived from the mechanical energy input analysis of the turbulence model, providing the optimized increment of the stirring rate sequence.

[0075] The optimized process parameters are aligned with the original production process parameter set over time. The original parameter set contains discrete time-stage data, with each stage corresponding to an independent control time period for the extraction temperature and stirring rate sequences. The alignment process distributes the correction amounts to the corresponding time periods according to the time stage weights, forming an integrated temperature and stirring rate sequence. Based on this sequence, a dynamic production process model is constructed. The core of this model is the establishment of a multi-component non-equilibrium thermodynamic governing equation set. The governing equation set expresses the conservation relationship of the medicinal wine components and includes three key terms: the concentration convection term describes the mass migration caused by fluid motion, the diffusion term reflects the molecular-level diffusion driven by the concentration gradient, and the chemical reaction source term relates to the dissolution kinetics of the effective components. The component evolution follows the principle of mass conservation.

[0076]

[0077] in: The concentration field variable is represented by mass fraction, defined as the percentage of effective component mass per unit volume. τ represents the time variable, measured in seconds, and is distinct from the global time axis; it is used to mark independent simulation timelines within the model. This represents the velocity vector field, with dimensions in meters per second, and includes velocity components in three dimensions. Let represent the spatial gradient operator, taking its partial derivative with respect to spatial coordinates. D represents the effective diffusion coefficient, measured in square meters per second, combining molecular diffusion and turbulent diffusion effects. R s It represents the chemical reaction source term, with dimensions of kilograms per second and cubic meters, describing the rate at which the active ingredient dissolves from the solid phase of the medicinal material.

[0078] The numerical solution of this system of equations employs a hybrid explicit-implicit scheme. (Time derivative term) Discretized in an explicit format, updates are directly calculated using data from the previous time step; spatial derivative term. and An implicit scheme is used to construct an iterative solution for the sparse linear system. The explicit step size is adaptively controlled according to the Courant number stability criterion; the implicit part uses the preprocessed conjugate gradient method to solve the matrix, accelerating convergence. A special discretization scheme is applied in the region near the phase interface to avoid non-physical oscillations. After initialization of the solution process, the simulation time is progressively advanced to track the evolution of the concentration field and the phase distribution field. The phase distribution field is monitored by the interfacial energy variable, defined as the energy value per unit area of ​​the phase interface. The solution termination condition is defined as the average rate of change of the interfacial energy of the phase distribution field being lower than a minimum threshold, indicating that the interface structure approaches steady-state equilibrium. The execution output of the dynamic production process model includes a dataset of the evolution of medicinal wine components in the spatial and temporal dimensions, covering the entire extraction process. The entire implementation process precisely links the multiphase flow simulation results with the actual process operation, driving the physical quantitative analysis of the medicinal wine production process. Through this coupled simulation and dynamic modeling mechanism, the complexity of fluid-structure interaction within the medicinal wine system is addressed, and the trajectory of time-dependent variables is captured.

[0079] Example 4: See Figure 5 After the dynamic production process model is completed, concentration field data and phase distribution field data describing the evolution of the medicinal wine system are generated. This embodiment first separates these two physical field data from the model. The concentration field data contains records of the spatial distribution of effective component concentrations within the entire medicinal herb soaking container at multiple discrete time steps. The phase distribution field mainly records the position and morphological evolution of the solid-liquid interface; it is not used as the core input for reconstruction but is used to assist in stability judgment. The concentration field is selected as the main processing object, and a discrete-time stage inverse transformation operation is performed. This operation uses the signal separation principle to decompose the concentration field into two independent parts according to preset time stages: a steady-state concentration component characterizing the long-term average concentration distribution, and a transient concentration component reflecting short-term concentration fluctuations. The steady-state component in each stage represents the state after reaching macroscopic equilibrium, such as the basic dissolution distribution of a medicinal herb after extraction at a specific temperature stage. The transient component reflects rapidly changing process responses, such as a surge in local component release caused by a sudden increase in temperature. A medicinal herb production process is divided into six time stages for inverse transformation; the table below shows the length data of each stage:

[0080]

[0081] The initial mixing ratio of the original solvent is retained as the basic concentration field. This ratio is usually determined before the medicinal materials are added to the solvent, such as the volume ratio of ethanol to water, forming an initial uniform distribution field within the container. Transient concentration components need to be redistributed and integrated according to time stages. The distribution is based on the weighting coefficients of the time stages. The physical characteristics of the transient concentration within each discrete time stage are calculated: first, the peak-to-valley difference of the concentration fluctuation within the entire time window of that stage is extracted as the oscillation amplitude. For example, in stage T2, the tanshinone concentration at a certain spatial point increases from 0.8 mg / L to 1.5 mg / L, with an oscillation amplitude of 0.7 mg / L. Simultaneously, spectral analysis is performed on the time series of this point in that stage, and the energy integral of its main frequency band is taken as the frequency spectral density. The calculated spectral density value for a certain point in stage T2 is 0.32 (mg / L). 2 ·s -1 The original weighting coefficients for this stage are obtained by directly multiplying the oscillation amplitude by the frequency spectral density value.

[0082] The steady-state concentration component directly participates in the final integration. This component reflects the equilibrium state throughout the process, such as the diffusion equilibrium concentration distribution of a certain alkaloid within the entire container. Spatial field synthesis is performed through a multi-stage reconstruction algorithm: the algorithm merges the initial field superimposed with the transient component and the steady-state concentration field at the same spatial grid points. The merging uses bilinear interpolation to ensure spatial continuity and avoid boundary jumps. For example, in the middle layer region of the container, the initial + transient component is 52 mg / L, and the steady-state component is 58 mg / L, then the reconstructed value is (52×k+58×(1-k)), where k is the spatial correlation factor of the transient component. The algorithm also eliminates the time dependency term, generating a static three-dimensional concentration distribution as the final simulation result. This result is output as a container gridded data file, recording each spatial unit (e.g., 1 cm). 3 The document contains the effective component concentration values ​​(voxels). For the decision-making of medicinal wine production processes, this document can be directly mapped to a spatial dissolution state diagram of the effective components of the medicinal materials at the moment of extraction completion, displaying the three-dimensional positioning data of high concentration zones and low permeability zones. The entire implementation transforms the time-series information output by the dynamic model into a high-precision physical characterization of the process terminal state through quantitative separation, normalization allocation, and spatial integration mechanisms.

[0083] Example 5: In constructing the dynamic production process model, the development of the governing equations involves integrating physical parameters within a non-equilibrium thermodynamic framework. The microscopic pore structure of medicinal materials significantly influences the mass transfer process. To quantify this influence, permeability tensor data extracted using microscopic image analysis techniques are introduced. This tensor is a three-dimensional matrix structure, and its element values ​​are determined by calculating the flow resistance coefficients of the solvent in different spatial directions of the medicinal material. This resistance characteristic is closely related to the geometry, pore size distribution, and connectivity of the pores within the medicinal material. For example, the permeability values ​​of some root-type medicinal materials are significantly higher in the axial direction than in the radial direction, reflecting the directional permeation characteristics of their vascular structure. The introduction of the permeability tensor enables the equations to accurately express the migration behavior of the solvent in heterogeneous porous media.

[0084] In the concentration convection term of the governing equations, the original velocity vector field is first subjected to a tensor contraction operation with the permeability tensor. This operation performs mathematical operations on a spatial grid basis: the permeability tensor matrix is ​​multiplied by the velocity vector at the current grid point to generate a new, corrected velocity vector. The value of each spatial component of the new vector is modulated by the numerical attribute of the permeability tensor in that direction: the component in the high permeability direction increases, while the component in the low permeability direction remains unchanged or decreases. This correction directly reflects the directional guiding effect of the microstructure on the macroscopic flow, forming the corrected convection term. The corrected velocity field drives the concentration migration process in the equations, making the mass transfer more closely resemble the actual extraction environment of the medicinal materials.

[0085] The construction of chemical reaction source terms requires quantifying the kinetics of active ingredient transfer from the solid phase of the medicinal material to the solvent phase. This process is achieved by introducing activation energy parameters for the dissolution trajectory of the active ingredient. Activation energy parameters characterize the minimum energy barrier that a single active ingredient molecule must overcome to detach from the medicinal material matrix. The numerical values ​​of these parameters are obtained through two methods: calculating the energy curve of the interaction force between the target molecule and the matrix based on molecular dynamics simulations, or performing inversion fitting by experimentally testing dissolution rates at different temperatures. During the model initialization phase, a corresponding activation energy parameter value is assigned to each target active ingredient. For example, the dissolution activation energy values ​​of glycyrrhizic acid and tanshinone differ, reflecting the differences in energy barriers in the desorption processes of different molecules.

[0086] The specific form of the chemical reaction source term uses the activation energy parameter as the core input variable. This mathematical expression establishes a functional relationship between the dissolution rate and the current local temperature and concentration gradient: increasing temperature generally reduces the energy barrier effect, promoting the dissolution reaction; changes in the concentration gradient reflect the dynamic adjustment of diffusion resistance. The source term participates in the solution process as a dynamic variable in the governing equations, and its value is updated in real time with spatial location and simulation time. This design ensures that the dissolution behavior conforms to the energy conservation laws at the molecular scale while responding to changes in macroscopic process conditions.

[0087] The overall structure of the governing equations thus comprises three core elements: a modified convection term, which incorporates the modulation effect of the permeability tensor on the directional migration of fluids; a primary diffusion term, which maintains the concentration equalization effect caused by molecular thermal motion; and a chemical reaction source term, which couples the dynamic dissolution kinetics characterized by the activation energy parameter. This structure fully integrates the microscopic structural properties of medicinal materials, solvent flow characteristics, and chemical dissolution mechanisms, providing a physically consistent mathematical model for simulating the extraction process. During the solver's calculations, these three elements interact: permeability modulates the macroscopic flow path, affecting the solvent's contact efficiency with the medicinal materials; activation energy constrains the chemical dissolution intensity, driving the accumulation of active ingredients in the solvent; and diffusion balances local concentration differences. Through this multi-physics coupling mechanism, the model can dynamically capture the real material migration trajectories during the medicinal wine production process.

[0088] The entire implementation focuses on establishing an accurate mass transfer model within a non-equilibrium thermodynamic framework. By integrating the regulatory effect of the permeability tensor on the migration process with the constraint mechanism of the activation energy on the dissolution rate into the same set of equations, a comprehensive simulation of the microstructure of medicinal materials and macroscopic processing conditions is achieved.

[0089] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0090] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for simulating the production process of medicinal wine, characterized in that, include: Obtain the set of physical property parameters and production process parameters of the raw materials for medicinal wine production; Based on the physical property parameter set, the physical property dataset of the medicinal materials is analyzed, and a spatial distribution model of the medicinal wine components is constructed through multi-scale molecular dynamics simulation. The spatial distribution model includes solvent diffusion paths and dissolution trajectories of effective components. Extract the core reaction regions in the spatial distribution model that meet the concentration gradient threshold and diffusion stability conditions, and screen key simulation nodes where the solvent diffusion path length is lower than the preset path upper limit and the dissolution rate of the effective component is higher than the dissolution rate lower limit. The key simulation nodes are input into a pre-constructed multiphase flow coupled simulation system. The multiphase flow coupled simulation system generates optimized process parameters through iterative solution of the turbulence model and mass transfer equation. Concentration field uniformity constraints and phase interface stability constraints are introduced into the convergence conditions of the multiphase flow coupled simulation system. By combining the optimized process parameters with the time stage division of the production process parameter set, a dynamic production process model is established, and the spatiotemporal evolution behavior of medicinal wine components is simulated based on non-equilibrium thermodynamic equations. The concentration evolution field and phase distribution field are separated from the dynamic production process model. The concentration evolution field is reconstructed in multiple stages with the original solvent ratio to output the final simulation results of the medicinal wine production process. The establishment of the dynamic production process model includes: The correction amount of the optimized process parameters is aligned with the original solvent ratio in multiple stages of time. A set of governing equations, including concentration convection terms, diffusion terms, and chemical reaction source terms, was constructed based on non-equilibrium thermodynamic equations. The spatiotemporal evolution behavior of the governing equations is solved using an explicit-implicit hybrid numerical scheme until the interface energy of the phase distribution field tends to a steady state.

2. The method for simulating the production process of medicinal wine according to claim 1, characterized in that, The acquisition of the set of physical property parameters and the set of production process parameters for the raw materials used in the production of medicinal wine includes: The set of physical property parameters includes the types of medicinal materials, solvent ratios, and initial concentration distributions; the set of production process parameters includes the extraction temperature sequence, stirring rate sequence, and time stage divisions. Collect microstructure images and solvent property test data of medicinal materials, and extract porosity distribution, specific surface area and solvent surface tension coefficient of medicinal materials; The production process parameter set is divided into time stages to generate discrete time nodes for the extraction temperature sequence and rotation speed gradients for the stirring rate sequence. By integrating the porosity distribution, specific surface area, solvent surface tension coefficient, discrete time nodes, and rotational speed gradient, a three-dimensional production process parameter matrix is ​​constructed as the input data source.

3. The method for simulating the production process of medicinal wine according to claim 2, characterized in that, The construction of the spatial distribution model of medicinal wine components includes: Molecular conformation space clustering was performed on the initial concentration distribution of the physical property parameter set to delineate the interaction domains between solvent molecules and the effective components of the medicinal material. The topology of the solvent diffusion path is calculated based on the energy barrier distribution of the interaction domain, and a dynamic distribution field is generated with the diffusion time step as the first dimension, the spatial coordinates of the medicinal material as the second dimension, and the concentration gradient as the third dimension. The physical consistency of the dynamic distribution field was verified by the rate of change of curvature of the solvent diffusion path and the continuity of the dissolution trajectory of the effective components.

4. The method for simulating the production process of medicinal wine according to claim 3, characterized in that, The extraction of the core reaction regions in the spatial distribution model that satisfy the concentration gradient threshold and diffusion stability conditions includes: Calculate the local concentration gradient magnitude and diffusion path oscillation frequency at each simulated node in the dynamic distribution field; The concentration gradient stability index of the statistical simulation node within a preset time window is composed of a weighted integral of the gradient magnitude variance and the oscillation frequency. Set concentration gradient threshold and diffusion stability threshold, and select simulation nodes that simultaneously satisfy the condition that the local concentration gradient magnitude exceeds the threshold and the stability index is lower than the stability threshold as the core reaction region.

5. The method for simulating the production process of medicinal wine according to claim 4, characterized in that, The iterative solution of the multiphase flow coupled simulation system includes: Map the diffusion path topology of key simulation nodes to the boundary conditions of the turbulence model; A phase interface tracking algorithm is coupled into the mass transfer equation to generate temperature correction and stirring rate correction for optimized process parameters. The variance decay rate of the concentration evolution field is calculated by constraining the uniformity of the concentration field, and the critical point of interface breakage of the phase distribution field is detected by constraining the stability of the phase interface.

6. The method for simulating the production process of medicinal wine according to claim 5, characterized in that, The multi-stage reconstruction includes: The concentration evolution field is subjected to discrete-time inverse transformation to separate its transient concentration components from its steady-state concentration components; The initial mixing ratio of the original solvent ratio is retained, and the transient concentration component is superimposed on the initial mixing ratio according to the weighting coefficient of the time stage; The final simulation result is generated by integrating the steady-state concentration components and the superposition results through a multi-stage reconstruction algorithm.

7. The method for simulating the production process of medicinal wine according to claim 6, characterized in that, The calculation of the concentration gradient stability index includes: Extract the concentration gradient vector sequence of the simulated node at continuous time steps, and calculate the magnitude variance of the gradient vector sequence and the rate of change of the direction cosine of adjacent vectors; Stability indices are generated by exponentially weighted moving averages of the magnitude variance and the direction cosine rate of change.

8. The method for simulating the production process of medicinal wine according to claim 7, characterized in that, The construction of the governing equations includes: Introducing the permeability tensor of the pore microstructure of medicinal materials into the non-equilibrium thermodynamic equations, the permeability tensor and the concentration convection term are tensor-merged to generate a modified convection term, and the activation energy parameter of the dissolution trajectory of the effective components is associated with the chemical reaction source term.

9. The method for simulating the production process of medicinal wine according to claim 8, characterized in that, The allocation of the weighting coefficients for the time phases includes: The oscillation amplitude and frequency spectral density of the transient concentration component within the discrete time period are statistically analyzed. The weight coefficients of each stage are assigned according to the product of the oscillation amplitude and the frequency spectral density. The weight coefficients are normalized to ensure the conservation of the multi-stage superposition.

Images