Numerical simulation optimization method and system for drying process based on multi-physical field coupling

Abstract

The application discloses a drying process numerical simulation optimization method and system based on multi-physical field coupling and relates to the technical field of drying numerical simulation. The method comprises the following steps: acquiring object data, observation data, parameter constraints and process constraints; establishing a numerical model of heat and mass transfer and solid response coupling and forming an observation state; solving parameter combinations to form observation mismatch, interface closure and control body conservation; comparing positive tolerance and reference parameter combination values to form a permitted parameter set; propagating the process candidate in the domain to a common termination time according to the mark to form a termination neighborhood state envelope, cumulative heat supply and same mark over-limit duration witness; screening out feasible candidates whose propagation is completed and witness does not exceed the allowed duration; outputting no feasible process result in space-time, and determining target process data according to the common termination time, state envelope, cumulative heat supply and process candidate mark sorting. The application is beneficial to multi-field coupling drying optimization and eliminates uncoordinated parameters.

Description

Technical Field

[0001] This invention relates to the field of numerical simulation technology for industrial drying, and in particular to a method and system for optimizing numerical simulation of drying processes based on multiphysics coupling. Background Technology

[0002] Numerical simulations of industrial drying processes typically establish heat and mass transfer models based on equipment structure, material properties, and boundary conditions. Model parameters are adjusted using limited measurement results, and processes are then selected based on the simulated endpoint moisture content and energy consumption indicators. However, because flow, heat transfer, moisture migration, phase change, and solid response interact during drying, endpoint determination using a single fitting index and nominal parameters cannot fully reflect the physical consistency and local state risks of the model under different operating conditions. This can easily lead to process selection results deviating from the practically usable range.

[0003] In front-end data processing, traditional methods often interpolate observations from different locations, times, and scales and directly use them for parameter correction, using the smaller overall error as the basis for model reliability. When the coverage of observation points is limited, multiple sets of transmission parameters may compensate for each other to obtain similar observation results, and the heat and moisture exchange relationships on both sides of the exchange interface and the changes in inventory in adjacent areas may still be inconsistent. Such differences lack a unified object, time, and direction correlation, which can easily lead to parameter identification distortion and cause subsequent calculations to use parameters that are not applicable to the corresponding operating conditions. In back-end process judgment, traditional methods often compare the average endpoint state, peak state, and comprehensive target value of a single set of parameters, making it difficult to simultaneously retain the propagation results of parameter differences in the spatial unit and the time process near the endpoint. Local overheating, overhumidification, and solid response exceeding limits may be masked by spatial average quantities and instantaneous values, and it is not easy to form traceable rejection criteria for solution failure, failure to reach the termination state, and continuous constraint exceeding limits, which may lead to misjudgment of process feasibility, incomplete state output, and decreased reliability of subsequent equipment settings. Summary of the Invention

[0004] The technical problem to be solved by this invention is how to form a set of permissible parameters under the joint constraints of observation fitting, thermal mass closure of the exchange interface and conservation of adjacent control volumes, and how to ensure that candidate processes maintain the same identity propagation under all permissible parameter combinations, and determine feasible processes based on the local state of the endpoint neighborhood and the duration of over-limit.

[0005] In a first aspect, the present invention provides a numerical simulation optimization method for a drying process based on multi-physics coupling, comprising: Acquire object data, observation data, parameter constraints, and process constraints. Based on the object data, establish a numerical model coupling heat and mass transfer with solid response. Map the observation data to the exchange interface and adjacent control volumes to form the observation state. The numerical model is used to solve the combination of parameters within the parameter constraints. The observed mismatch is formed by the observed state and the solution result. The interface closure is formed by the difference between the sensible heat of the fluid and the heat conduction and phase change heat of the material, and the difference between the water exchange between the fluid and the material, according to the same interface normal and time window. The control volume conservation is formed by the difference between the energy and water storage changes of adjacent control volumes and the net exchange. The observed mismatch, interface closure, and control volume conservation quantities are compared with the corresponding positive tolerances of the parameter constraints, and the control volume conservation quantities are ensured not to exceed the reference parameter combination values, thus forming a set of permissible parameters with parameter combination identifiers and applicable boundary domains. Process candidate data with process candidate identifiers are generated according to process constraints. For process candidate data within the applicable boundary domain, each parameter combination of the permitted parameter set is propagated to the common termination time through the numerical model according to the parameter combination identifier, forming the parameter combination propagation completion status, the unit-by-unit termination neighborhood status envelope, the cumulative heat supply, and the continuous witnessing of the same identifier exceeding the limit. Select feasible process candidate data whose parameter combination propagation completion status is fully completed and whose extended duration does not exceed the allowable duration.

[0006] Secondly, the present invention provides a numerical simulation optimization system for a drying process based on multi-physics coupling, comprising: The model building module is used to acquire object data, observation data, parameter constraints and process constraints, establish a numerical model of heat and mass transfer coupled with solid response based on object data, and map observation data to the exchange interface and adjacent control volume to form the observation state. The permissible parameter formation module is used to solve the parameter combination within the parameter constraints using a numerical model. It forms the observed mismatch quantity based on the observed state and the solution result. It forms the interface closure quantity based on the difference between the sensible heat of the fluid and the heat conduction and phase change heat of the material, and the difference between the water exchange between the fluid and the material, according to the same interface normal and time window. It forms the control volume conservation quantity based on the difference between the energy and water storage changes of adjacent control volumes and the net exchange. It compares the observed mismatch quantity, interface closure quantity, and control volume conservation quantity with the corresponding positive tolerance of the parameter constraints, and ensures that the control volume conservation quantity does not exceed the reference parameter combination value, thus forming a set of permissible parameters with parameter combination identifier and applicable boundary domain. The process candidate generation module is used to generate process candidate data with process candidate identifiers according to process constraints; The state propagation module is used to propagate the process candidate data within the applicable boundary domain to the common termination time through the numerical model according to the parameter combination identifier of each parameter combination in the permitted parameter set, forming the parameter combination propagation completion status, the unit-by-unit termination neighborhood state envelope, the cumulative heat supply, and the continuous witnessing of the same identifier exceeding the limit. The process screening module is used to screen out feasible process candidate data where the parameter combination propagation completion status is fully completed and the extended duration of the witness has not exceeded the allowable duration. The results output module is used to output no feasible process results when the feasible process candidate data is empty, and to determine the target process data by sorting the data according to the common termination time, the state envelope of the termination neighborhood, the cumulative heat supply, and the process candidate identifier when the data is not empty.

[0007] In summary, the present invention has at least one of the following beneficial technical effects: This invention forms an observation state by combining object data with object identification, coordinates, time, unit, and validity information with observation data. Within the same interface normal and time window, it compares the observed mismatch, interface closure, and control volume conservation with their corresponding positive tolerance and reference parameter combinations to form a set of permissible parameters with parameter combination identifiers and applicable boundary domains. Screening solely based on observation mismatch cannot identify interface exchange differences masked by parameter compensation, while checking interface exchange and control volume conservation alone cannot guarantee that the parameter combination corresponds to the actual observation. These three types of comparison relationships jointly constrain the same parameter combination, eliminating combinations that are well-fitted by observations but have inconsistent heat and mass exchange and inventory changes. This provides intermediate data that retains physical verification basis and applicable scope for subsequent process propagation.

[0008] This invention uses the aforementioned set of permitted parameters and their combination identifiers, along with the applicable boundary domain, as continuous inputs. First, it excludes process candidate data that exceeds the applicable boundary domain. Then, it propagates each retained process candidate to a common termination time under all permitted parameter combinations, forming a spatial unit-by-spatial-unit termination neighborhood state envelope and a persistent witnessing of exceeding limits with the same identifier. Based on the propagation completion status, the duration of exceeding limits, and a fixed sorting relationship, it outputs infeasible process results or target process data. The endpoint result of a single nominal parameter is insufficient to characterize the local differences caused by all permitted parameters, and simply providing the envelope endpoint makes it difficult to trace the specific combination that caused the boundary violation. The combined effect of applicable domain limitation, full parameter propagation, identity retention, and persistent witnessing can avoid mistakenly including candidates with extrapolation errors, incomplete propagation, and local persistent boundary violations into the feasible set, providing a technical basis for backtracking by process candidate, parameter combination, spatial unit, and time. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart of a numerical simulation optimization method for a drying process based on multi-physics coupling, provided in an embodiment of the present invention.

[0011] Figure 2 This is a schematic diagram of the cross-section of the exchange interface and the adjacent control body provided in an embodiment of the present invention.

[0012] Figure 3 This is a parameter permission record and termination neighborhood witness correspondence diagram provided in the embodiments of the present invention.

[0013] Figure 4 This is a framework diagram of a numerical simulation optimization system for the drying process based on multi-physics coupling, provided in an embodiment of the present invention. Detailed Implementation

[0014] The following description, in conjunction with the implementation of the present invention, is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention, and all such modifications and additions should fall within the protection scope of the present invention.

[0015] Example 1 This invention is applicable to optimization scenarios for industrial drying equipment using hot air, vacuum, microwave-assisted hot air, or other methods that can establish coupled models of heat, moisture, and solid response. On a batch drying production line, the geometry, initial moisture content, and equipment boundaries of the same batch of materials fluctuate, and observations of boundary temperature, humidity, and flow rate can only cover a limited number of locations. If only numerical results are fitted to the observed data, the transfer parameters at different boundaries can easily compensate for each other, masking the non-conservation of the exchange interface. Furthermore, if the process is selected using the average endpoint quantity of a single nominal parameter, it may still miss the problem of individual spatial units experiencing continuous overheating, overhumidification, or solid response exceeding the limits near the endpoint.

[0016] A key concept of this invention is to assign a unique identifier, such as a parameter combination identifier or a process candidate identifier, to each parameter combination and each process candidate data throughout the entire process. When forming the set of permitted parameters, physical verification results, such as observed mismatch, interface closure, and control volume conservation quantities, are associated and stored with this identifier. During backend process propagation, it is mandatory that the same process candidate must complete propagation under all permitted parameter combinations. Based on this identifier, a termination neighborhood state envelope and an over-limit persistence witness are generated for each spatial unit, each time point, and each state, which can be traced back to the specific parameter combination.

[0017] Through the synergistic effect of the aforementioned identity preservation and full parameter traversal mechanism, this invention not only selects parameters that fit observations and are physically consistent, but also ensures that the finally selected process remains stable and reliable across all possible parameter fluctuations. This effectively avoids the technical shortcomings of traditional methods that mask local and persistent risks by using a single nominal parameter or average endpoint state.

[0018] This application discloses a numerical simulation optimization method for the drying process based on multiphysics coupling.

[0019] Reference Figure 1-3Numerical simulation optimization methods for drying processes based on multi-physics coupling include: Acquire object data, observation data, parameter constraints, and process constraints. Based on the object data, establish a numerical model coupling heat and mass transfer with solid response. Map the observation data to the exchange interface and adjacent control volumes to form the observation state. Specifically, the numerical simulation optimization computing equipment reads object data, observation data, parameter constraints, and process constraints from the current optimization task directory. Object data consists of geometric design files, mesh files, and material property files, and at least records the fluid domain, material domain, exchange interface, spatial element identifier, coordinate system, unit system, material thermophysical property relationships, moisture transfer relationships, and applicable solid response relationships. Observation data is generated by measurement components installed at the inlet, outlet, cavity boundary, or material surface of the drying equipment via acquisition interfaces. Each record includes the object identifier, coordinates, acquisition time, unit, range, resolution, calibration status, and validity flag. Parameter constraints and process constraints are formed from material data, equipment allowable range, measurement capabilities, and quality targets, and use the same version identifier as the current optimization task.

[0020] The computing device first verifies four types of inputs: object identifier, coordinate system, time base, unit, and version. If the object identifier does not match, the unit is undefined, necessary material relationships are missing, or the version is inconsistent, the current task is stopped and an invalid input record is output. After successful verification, the material domain is discretized into spatial units, and the exchange interface is discretized into interface units that correspond to adjacent control volumes on both the fluid and material sides. The numerical model couples at least fluid motion and sensible heat transfer, material heat conduction, internal moisture transfer, interfacial phase change moisture exchange, and solid response. The solid response characterizes the configured state of deformation, strain, or stress caused by temperature and humidity conditions; if the material property file does not have a corresponding relationship, this state is not generated, and the process constraints are marked as insufficient input.

[0021] The computational device exchanges boundary quantities in a fixed order: flow field, temperature field, moisture transfer field, and solid response field. This order is repeated within each simulation time step until the residuals of each field are no greater than the solution convergence tolerance and the difference in exchange quantities between adjacent fields is no greater than the coupling convergence tolerance, then proceeding to the next time step. If the maximum number of iterations is reached but the conditions are not met, or if any energy, moisture, or solid response state is not a finite value, the current solution task fails. The solution order, convergence tolerance, maximum number of iterations, and time steps are all obtained from the same model version and constraint file and are not changed during the solution process.

[0022] The computing device projects the observation location onto the exchange interface or adjacent control volume of the same object based on the object identifier and coordinates in the observation data. When multiple equidistant candidates exist, they are selected in ascending order of interface unit identifier; if the projected distance exceeds the allowable distance formed by installation positioning error and grid scale, the record is excluded. Using the simulation time as the target time, corresponding observation states are formed between adjacent valid sampling times according to the interpolation rules registered in the observation file. No extrapolation is performed if the target time exceeds the valid sampling range or the adjacent sampling interval exceeds the allowable interval. Each valid mapping result is associated with the observation record identifier, interface unit identifier, adjacent control volume identifier, and simulation time for subsequent mismatch and closure checks.

[0023] The numerical model is used to solve the combination of parameters within the parameter constraints. The observed mismatch is formed by the observed state and the solution result. The interface closure is formed by the difference between the sensible heat of the fluid and the heat conduction and phase change heat of the material, and the difference between the water exchange between the fluid and the material, according to the same interface normal and time window. The control volume conservation is formed by the difference between the energy and water storage changes of adjacent control volumes and the net exchange. Specifically, the computing device first generates a finite number of parameter combinations according to parameter names and value order, and assigns a parameter combination identifier to each combination. Then, it solves the problem using data from the same object and calibration process boundaries. For each valid observation, the observed and solved values ​​for the same object, location, time, and unit are obtained. The difference between the two values ​​is calculated, and the observation identifier, time, and parameter combination identifier are retained. Different physical quantities are not directly added. The absolute difference of the observation mismatch is saved separately for each observation, and subsequently compared with the corresponding observation positive tolerance on the same dimension. Parameter combinations without valid observations, with invalid correspondences, or with failed solutions do not form valid observation mismatches.

[0024] Within the same interface normal and the same time window, the computing device accumulates the sensible heat exchange amount entering the interface from the fluid side, the conductive heat exchange amount leaving the interface from the material side, the phase change heat corresponding to the moisture exchange with the interface, the increase in moisture on the fluid side, and the decrease in moisture on the material side. The thermal closure difference is the difference between the fluid sensible heat exchange amount and the sum of the material conductive heat exchange amount and the phase change heat, while the moisture closure difference is the difference between the increase in moisture on the fluid side and the decrease in moisture on the material side. If the original output direction of the model is opposite to the unified normal, the sign is changed first according to the normal relationship registered at the interface, and then the exchange amount within the time window is accumulated. The interface closure amount stores the absolute difference of thermal closure and the absolute difference of moisture closure according to the interface unit and the time window, respectively.

[0025] The computing device simultaneously acquires the energy and moisture reserves of adjacent control bodies on both sides of the exchange interface at the start and end of the time window, respectively generating changes in energy and moisture reserves. It then accumulates the net energy and moisture exchange amounts of the same control body within the time window, subtracting outflows from all boundary flows. The energy conservation difference is the difference between the energy reserve change and the net energy exchange amount, and the moisture conservation difference is the difference between the moisture reserve change and the net moisture exchange amount. Control body conservation quantities are stored as two types of absolute differences according to the control body identifier and the time window; one physical quantity is not used to compensate for the other. If any interface-side control body is missing, the time window is incomplete, or the reserve units do not correspond, that parameter combination will not be included in the permission screening.

[0026] The observed mismatch, interface closure, and control volume conservation quantities are compared with the corresponding positive tolerances of the parameter constraints, and the control volume conservation quantities are ensured not to exceed the reference parameter combination values, thus forming a set of permissible parameters with parameter combination identifiers and applicable boundary domains. Specifically, the observation positive tolerance is determined by the larger of the positive values ​​available from the measurement resolution, the state difference caused by the mapping distance, and the time discrete error. The interface positive tolerance is determined by the larger of the positive values ​​of the spatial projection error and the positive values ​​of the time integration error of the discrete fluxes on both sides of the exchange interface. The conservation positive tolerance is determined by the larger of the positive values ​​of the spatial discrete error and the positive values ​​of the time integration error of adjacent control volumes. Each error positive value can be formed by calculating differences based on equipment specifications, adjacent grid levels, or adjacent time steps, as well as numerical integration settings. When a formation item is missing, the remaining available positive values ​​are used; when all formation items are missing, not finite values, or not greater than 0, the corresponding comparison is invalid and the formation of the permissible parameter set is stopped.

[0027] Understandably, through the aforementioned triple screening of observed mismatch, interface closure, and control volume conservation, especially the introduction of the control volume conservation value not exceeding the reference parameter combination value as an additional constraint, a large number of physically unreasonable or distorted parameter combinations can be eliminated at an early stage. This results in a final set of permissible parameters that is much smaller than the initial set of parameter combinations, thereby significantly reducing the computational overhead required for the propagation of all subsequent parameter combinations. Furthermore, the pre-comparison of the applicable boundary domain further filters out a large number of obviously mismatched candidates before the process candidate data enters the time-consuming high-fidelity numerical simulation. The above-mentioned multi-screening mechanism is not only key to ensuring the accuracy of the results, but also a necessary guarantee for the implementation of the technical solution of this invention within reasonable computational resources.

[0028] The computing device checks each parameter combination item by item according to its identifier. A parameter combination is retained only if the difference between each observation item is no greater than the corresponding positive tolerance, the difference between each interface thermal and moisture content is no greater than the corresponding positive tolerance, the difference between each control volume energy conservation and moisture conservation is no greater than the corresponding positive tolerance, and the conservation quantity of the control volume is no greater than the reference parameter combination value obtained using the same control volume and the same time window. The reference parameter combination is pre-identified within the parameter constraints; if not pre-identified, it can be generated based on the middle position of each parameter's discrete values, and the smaller value is taken when two equidistant values ​​exist. The reference parameter combination is only used as a baseline for conservation without increase and is not represented as actual material parameters.

[0029] For each reserved parameter combination, the computing device obtains the lower and upper bounds of the components of inlet temperature, inlet humidity, flow rate, heating boundary, and other configured boundary quantities when a valid observation state is formed, thus creating an applicable boundary domain. Each permission record is associated with the parameter combination identifier, parameter values, various differences, corresponding positive tolerances, reference parameter combination values, applicable boundary domain, and input file version. These associated records constitute a permission credential, ensuring that subsequent process propagation accurately reads the same parameter combination. When the permission parameter set is empty, the output shows no permission parameters and failed items, no tolerance is allowed, and candidate process propagation is stopped.

[0030] Process candidate data with process candidate identifiers are generated according to process constraints. For process candidate data within the applicable boundary domain, each parameter combination of the permitted parameter set is propagated to the common termination time through the numerical model according to the parameter combination identifier, forming the parameter combination propagation completion status, the unit-by-unit termination neighborhood status envelope, the cumulative heat supply, and the continuous witnessing of the same identifier exceeding the limit. Specifically, the computing device reads the name, lower bound, upper bound, positive step size, and allowed switching time of each process boundary quantity from the process constraints. For each boundary quantity, values ​​are assigned starting from the lower bound with positive steps. If the last step size cannot accurately reach the upper bound, the upper bound is still used as the final value. Then, finite initial process candidate data is formed according to the boundary quantity name, time period, and ascending value order, and initial process candidate identifiers are assigned sequentially. If the number of candidates exceeds the maximum number specified in the constraint document, the result of "candidate space too large" is output and the process stops, without replacing the finite enumeration with random sampling.

[0031] The computing device writes each initial process candidate data into the initial state and boundary condition interface of the numerical model, forming a candidate boundary state. For each component of the candidate boundary state, it searches for the lower and upper bounds of the applicable boundary domains with the same name, unit, and version as that component in the set of permitted parameters. When all components are within the corresponding interval of at least one common permitted record, they are retained as process candidate data and the initial process candidate identifier is used as the process candidate identifier; when any component exceeds the applicable interval of all permitted records, the exceeding component, interval, and initial process candidate identifier are written into the rejection record, and the candidate is not propagated.

[0032] For each process candidate data, the computing device writes the process candidate identifier, parameter combination identifier, model version, and initial state into a solution task, with the process candidate data serving as the boundary conditions for that task. The permissible parameter set is traversed in ascending order according to the parameter combination identifier, the corresponding parameter values ​​are read, and a complete coupled solution is executed. If any task fails to converge, or if any of the energy stock, moisture stock, or configured solid response state is not a finite value, the corresponding process candidate identifier is marked as incomplete propagation, the failed parameter combination identifier and the last complete time step are recorded, and subsequent propagation of that process candidate is stopped. Propagation is only marked as complete when all permissible parameter combinations have been solved.

[0033] For each parameter combination after propagation, the computing device checks the humidity state of all spatial cells in ascending order of simulation time. The first moment when all spatial cells are simultaneously within the target humidity range and continuously maintain that position for the specified duration is considered the arrival time of that parameter combination. If any parameter combination does not have an arrival time within the solution time domain, the process candidate is marked as having no valid termination. If all parameter combinations have arrival times, the latest time among them is taken as the common termination time. The endpoint observation duration is taken as the larger positive value among the effective sampling interval, the minimum process switching interval, and the solid response step size. This duration is extended backward from the common termination time to form a termination neighborhood; if all formation terms are missing, the termination neighborhood is deemed invalid.

[0034] Within the termination neighborhood, the computing device reads the humidity state, temperature state, and configured solid response state according to the process candidate identifier, parameter combination identifier, spatial unit identifier, and simulation time. For each state in the same spatial unit and at the same time, the lower and upper bounds of all parameter combination results are obtained to form a termination neighborhood state envelope for each spatial unit. At the same time, the parameter combination identifiers that generate each endpoint are saved to avoid splicing different combination endpoints into a fictitious state. The cumulative heat supply is formed by accumulating the heat supply boundary quantities according to the same time base and complete time steps; if no heat supply boundary quantity is configured, the sorting level is marked as inapplicable, and no energy value is generated.

[0035] The computing device checks the termination neighborhood step by step, parameter by parameter combination, spatial unit by spatial unit, and state by state. When a state exceeds the upper allowable limit or falls below the lower allowable limit, the length of consecutive time steps is accumulated from the first time the state exceeds the limit. When a state re-enters the allowable range, the current over-limit period ends and the record is saved. Each over-limit continuous witness with the same identifier includes the process candidate identifier, parameter combination identifier, spatial unit identifier, state name, over-limit start and end times, duration, and corresponding allowable duration. If there are missing time steps in the termination neighborhood, the gap interpolation is not considered qualified, and the process candidate is marked as incomplete propagation.

[0036] The feasible process candidate data is screened out when the parameter combination propagation completion status is all completed and the extended continuous witnessing does not exceed the allowable duration; the feasible process candidate data is output as follows: when there is no feasible process result in the empty time, the target process data is determined by sorting according to the common termination time, termination neighborhood state envelope, cumulative heat supply and process candidate identifier in the non-empty time. Specifically, the computing device excludes process candidate data that has incomplete propagation, lacks valid termination, or whose duration exceeds the corresponding allowable duration for any identical identifier. Each candidate is only added to the feasible process candidate data set if the set of permitted parameters is not empty, the candidate boundary is within the applicable boundary domain throughout the entire process, all permitted parameter combinations have completed propagation, the duration of exceeding the limit for all spatial units and states does not exceed the allowable value, and all device boundary quantities are within the allowable range. The checks are performed in ascending order of process candidate identifiers, and the order of checks does not change the set results.

[0037] When the feasible process candidate data is empty, the result output module generates a result indicating no feasible process, and saves candidate rejection records, propagation failure records, no valid termination records, and excessive continuous witness records, without outputting target process data. When the feasible process candidate data is not empty, they are first sorted in ascending order by common termination time; candidates with the same value are sorted in ascending order by the maximum width of the envelope of the humid state in the termination neighborhood; if they are still the same, they are sorted in ascending order by cumulative heat supply if the cumulative heat supply ranking level is applicable; if they are still the same or the ranking level is not applicable, they are sorted in ascending order by process candidate identifier. The candidate at the top forms the target process data.

[0038] The target process data should include at least the process candidate identifier, process boundary quantities for each time period, common termination time, termination neighborhood state envelope, cumulative heat supply applicable status, list of permitted parameter combinations, and verification records with no overdue witnesses. The results can be written to a process setting file readable by the drying equipment, or they can be reviewed by technicians before being imported into the equipment. This embodiment only proposes the technical path for forming and outputting the aforementioned file and does not describe the numerical simulation results as control facts already executed by the equipment.

[0039] Furthermore, the observed mismatch is formed by combining the observed state with the solution results. The interface closure is formed by the difference between the sensible heat of the fluid and the thermal conductivity and phase change heat of the material, and the difference between the water exchange between the fluid and the material, according to the same interface normal and time window. The control volume conservation is formed by the difference between the changes in energy and water storage of adjacent control volumes and the net exchange, including: The observation data is mapped to the exchange interface and adjacent control volumes according to the object identifier, coordinates and acquisition time. The difference between the observation data at the same location, at the same time and in the same unit and the solution results corresponding to the parameter combination is calculated, and the difference of each observation item is used to form the observation mismatch. Accumulate the fluid sensible heat, material thermal conductivity, phase change heat, and fluid-material moisture exchange within the same interface normal and time window. Subtract the sum of fluid sensible heat, material thermal conductivity, and phase change heat, and subtract the fluid moisture exchange from the material moisture exchange. The interface closure value is composed of the interface thermal difference and the interface moisture difference. Within the time window, obtain the changes in energy and moisture content and the corresponding net exchange volume of adjacent control volumes on both sides of the exchange interface, calculate the difference, and use the energy conservation difference and moisture conservation difference to form the control volume conservation quantity. The parameters are associated with the observed mismatch, interface closure, and control volume conservation quantities according to the parameter combination identifier, and are then incorporated into the formation process of the permissible parameter set.

[0040] In a specific embodiment, when forming the observation mismatch quantity, a four-level correspondence is first performed on the observation data: object, coordinates, time, and unit. Then, the difference between the observed and solved values ​​at the same location and time is calculated. When forming the interface closure quantity, the interface normal and time window are first unified, and then sensible heat, thermal conductivity, phase change heat, and water exchange on both sides are accumulated separately. When forming the control volume conservation quantity, the changes in energy and water storage within the time window are compared with the net exchange quantity. All three types of results use the current parameter combination identifier, and if any upstream correspondence fails, the missing item is not recorded as 0.

[0041] Observations of mismatch, interface closure, and control volume conservation quantities retain their respective physical quantities, objects, time windows, and units, and comparisons only consider the corresponding positive tolerances. This avoids direct merging between quantities of different dimensions and allows permission records to be used to retrospectively identify observations, interface units, or adjacent control volumes that resulted in rejection.

[0042] Furthermore, the formation of the corresponding positive tolerance and reference parameter combination values ​​includes: The observation positive tolerance is determined based on the positive values ​​obtained from the measurement resolution, the state difference caused by the mapping distance, and the time discrete error, and the larger of the positive values ​​is taken. The larger of the positive values ​​of the spatial projection error and the positive values ​​of the time integration error of the discrete fluxes on both sides of the exchange interface is determined as the interface positive tolerance. Compare the positive values ​​of spatial discrete error and time integral error of adjacent control volumes, and take the larger one as the conserved positive tolerance. According to the control volume and time window used for the conservation of the control volume to be compared, read the solution results of the reference parameter combination pre-identified in the parameter constraints; Obtain the reference parameter combination values ​​from the solution results.

[0043] In one specific embodiment, the corresponding positive tolerance is frozen before the parameter combination solution is initiated. The measurement resolution is derived from the observation equipment specifications and acquisition configuration; the mapping distance state difference can be formed by the changes in the positional state of adjacent grids and the actual projected distance; the time discretization error, spatial projection error, and time integration error can be formed by the differences in results of adjacent discretization levels. For each category, only the obtainable and greater than 0 forming terms are compared and the larger one is taken. The process stops when all forming terms are invalid, and no arbitrary minimal constants are used as substitutes.

[0044] The reference parameter combination value is formed using the same control volume, time window, unit, and model version as the control volume to be compared, which are identical in terms of their conserved quantities. If the parameter constraint has a registered reference identifier, it is read directly; otherwise, it is determined by the median position of the discrete values, and the generation rule is saved. If the reference combination solution fails, the current round of permissible filtering has no valid reference values ​​and stops.

[0045] Furthermore, a set of licensed parameters with parameter combination identifiers and applicable boundary domains is formed, including: The finite values ​​within the parameter constraints are combined according to the parameter name and the order of their values, and a parameter combination identifier is assigned to each parameter combination. Ensure that the observed mismatch, interface closure, and control volume conservation are all no greater than their respective positive tolerances, and that the control volume conservation is no greater than the reference parameter combination value, and retain the parameter combination; The applicable boundary domain is formed from the lower and upper bounds of each component of the observed boundary state of the parameter combination, and the parameter combination identifier, applicable boundary domain and parameter combination association are written into the permitted parameter set. When the set of permitted parameters is empty, the result of no permitted parameters is output and the propagation of process candidate data is stopped. When it is not empty, the set of permitted parameters is output for the propagation of process candidate data.

[0046] In one specific embodiment, parameter combination generation employs deterministic Cartesian enumeration. Each parameter value is derived from a finite sequence formed by a lower bound, an upper bound, and a positive step size. The order of parameter names and the ascending order of each sequence jointly determine the parameter combination identifier. The screening process sequentially checks for observation mismatch, interface closure, control volume conservation, and relative reference conservation without incrementing the condition. If any condition fails, a failed item is recorded, but the judgment boundary for subsequent combinations remains unchanged.

[0047] When retaining a parameter combination to form an applicable boundary region, the component boundaries of different parameter combinations are not merged into a single expanded region. Each permission record separately stores its own parameter combination identifier, applicable boundary region, and parameter values. Subsequent candidates can only be propagated by that parameter combination if all boundary components of the same permission record are valid. When the permission parameter set is empty, the result with no permission parameters is output, and the current optimization task ends.

[0048] Furthermore, process candidate data with process candidate identifiers is generated according to process constraints, including: The lower bound, upper bound and positive step size of each process boundary quantity are obtained from the process constraints, and then combined into finite initial process candidate data according to the name and value order of each process boundary quantity. Assign initial process candidate identifiers to each initial process candidate data, and form candidate boundary states based on the initial state of the numerical model and process boundary quantities; Each component of the candidate boundary state is compared with the corresponding lower and upper bounds of the applicable boundary domain. The initial process candidate data in which each component is located between the corresponding lower and upper bounds is retained as the process candidate data, and its initial process candidate identifier is used as the process candidate identifier. The components that exceed the limit are written into the rejection record along with the initial process candidate identifier.

[0049] In one specific embodiment, the boundary quantities of the process candidates may include inlet temperature, inlet humidity, flow rate, heating capacity, and other adjustable boundary quantities. The lower bound, upper bound, positive step size, and allowable switching time for each boundary quantity are all derived from the same version of the process constraints. The computing device generates a finite number of combinations based on name, time period, and value order, ensuring that the same input always yields the same initial process candidate identifier.

[0050] Boundary domain checks are performed on each component of the candidate boundary state. Candidates that satisfy all component ranges of the same license record are retained; those that only satisfy partial components of different license records are rejected to avoid splicing non-existent license domains. Rejection records save the out-of-bounds component, candidate value, lower license bound, upper license bound, and initial process candidate identifier.

[0051] Furthermore, the propagation of various parameter combinations includes: Write the process candidate identifier, parameter combination identifier, and initial state of the numerical model into the corresponding solution task, and use the process candidate data as the boundary conditions of the solution task. Read parameter combinations from the set of permitted parameters according to the parameter combination identifier in the solution task, traverse all parameter combinations for each process candidate data, perform numerical model solution and complete the solution task; If a solution task fails to converge, or if the energy or water content is not finite, the corresponding process candidate will be marked as having incomplete propagation and the subsequent propagation of the process candidate data will be stopped. When all solution tasks are completed, the corresponding process candidate will be marked as propagation complete.

[0052] In one specific embodiment, each solution task is uniquely identified by both a process candidate identifier and a parameter combination identifier. The task reads parameter values ​​with the same identifier from the set of permitted parameters and uses the same type version, initial state, and process candidate boundary. For a process candidate, all permitted parameter combinations must be traversed; failed combinations cannot be deleted and the remaining results cannot be used as the complete envelope.

[0053] Non-convergence is determined by reaching the maximum number of iterations while the residual remains above tolerance; non-finite inventory is determined by numerical finiteness checks. In either case, a failure flag and the last complete time step are saved, marking the propagation as incomplete and stopping the current candidate. Propagation is marked as complete only after all tasks are finished; this flag serves as a prerequisite input for calculating the common termination time.

[0054] Furthermore, the formation of the common termination time and termination neighborhood includes: For each parameter combination that has completed propagation, check the humidity status of each spatial unit in chronological order. Record the first moment when all spatial units are continuously within the target humidity range of the process constraint for the duration of the holding time as the arrival time of the corresponding parameter combination. If any combination of parameters fails to reach a certain time in the solution time domain, the corresponding process candidate will be marked as having no valid termination. When all parameter combinations have reached their respective times, the latest time among all the times is taken as the common termination time. The effective sampling interval is taken from the observation data, the minimum switching interval is taken from the process constraints, and the solid response step size is taken from the numerical model. The larger of the three is taken as the endpoint observation time. The termination neighborhood is formed by extending the endpoint observation time forward from the common termination time.

[0055] In one specific embodiment, the time-reaching check targets all spatial cells under the same parameter combination, rather than the average humidity state. After checking in ascending order of time, only when all spatial cells are simultaneously within the target humidity range and continuously reach the holding time is the first qualified time recorded. If any combination fails to reach that time within the solution time domain, the entire process candidate is marked as having no valid termination.

[0056] The common termination time is the latest value among all permitted parameter combinations, ensuring that the termination neighborhoods of all combinations share a common time base. The termination observation duration is the larger positive value among the effective sampling interval, minimum switching interval, and solid-state response step size. After forming the termination neighborhood, only the complete time steps within that neighborhood are used to form the envelope and witness.

[0057] Furthermore, the formation of unit-by-unit termination of the neighborhood state envelope, cumulative heat supply, and continuous witnessing of exceeding the same identifier limit includes: Within the termination neighborhood, the humidity state, temperature state, and solid response state are obtained according to the process candidate identifier, parameter combination identifier, spatial unit identifier, and time. For the same spatial unit, the lower and upper bounds of all parameter combinations are taken for the humidity state, temperature state, and solid response state, respectively, to form the terminating neighborhood state envelope for each spatial unit. The humidity state, temperature state, and solid response state are compared with the corresponding allowable ranges of the process constraints. For the cumulative duration of continuous periods exceeding the allowable range, the process candidate identifier, parameter combination identifier, spatial unit identifier, state name, over-limit period, and duration are associated as a continuous over-limit witness with the same identifier.

[0058] In one specific embodiment, the termination neighborhood state envelope is indexed by spatial unit and time to obtain the humidity, temperature, and configured solid response state upper and lower bounds for all permitted parameter combinations. The endpoints store the source parameter combination identifiers, which can be used to retrieve the complete multi-physics state of the same model task, avoiding treating different combination endpoints as the same instantaneous state.

[0059] When a continuous over-limit witness is generated, the consecutive over-limit periods are identified by process candidate, parameter combination, spatial unit, and status name, and the actual time step length is accumulated. A period is considered acceptable when the value equals the allowable boundary; it enters an over-limit period when the value is less than the lower allowable limit or greater than the upper allowable limit. Each witness is associated with a corresponding allowable range and allowable duration for feasibility screening.

[0060] Furthermore, feasible process candidate data were screened out where the parameter combination propagation completion status was fully completed and the extended duration of witnessing did not exceed the allowable duration, including: Exclude process candidate data that are incompletely propagated, have no valid termination, or whose identical identifiers exceed the allowed duration, and form feasible process candidate data; When the feasible process candidate data is empty, the output will show no feasible process and the corresponding rejection record. When the feasible process candidate data is not empty, the process candidate data with the earlier common termination time, smaller envelope width of the humid state in the termination neighborhood, smaller cumulative heat supply, and smaller process candidate identifier are selected in sequence to determine the target process data.

[0061] In one specific embodiment, the screening process first performs hard exclusion, followed by hierarchical sorting. Candidates with incomplete propagation, no valid termination, or any out-of-limit duration exceeding the allowable duration are not included in the sorting. When the feasible set is empty, no feasible process result is output directly, along with a rejection record, without considering the candidate closest to the constraint as a replacement result.

[0062] When the feasible set is not empty, the termination time, the maximum width of the humid state envelope, the cumulative heat supply, and the process candidate identifier constitute the ranking levels in that order. Each level only compares candidates with the same status in the previous level. If the heat supply boundary quantity is not configured, the cumulative heat supply level is skipped, and the process candidate identifier is still used to obtain the unique first-in-command candidate.

[0063] Example 2 This application discloses a numerical simulation optimization system for the drying process based on multiphysics coupling.

[0064] Reference Figure 4 A numerical simulation optimization system for the drying process based on multi-physics coupling includes: The model building module is used to acquire object data, observation data, parameter constraints and process constraints, establish a numerical model of heat and mass transfer coupled with solid response based on object data, and map observation data to the exchange interface and adjacent control volume to form the observation state. The permissible parameter formation module is used to solve the parameter combination within the parameter constraints using a numerical model. It forms the observed mismatch quantity based on the observed state and the solution result. It forms the interface closure quantity based on the difference between the sensible heat of the fluid and the heat conduction and phase change heat of the material, and the difference between the water exchange between the fluid and the material, according to the same interface normal and time window. It forms the control volume conservation quantity based on the difference between the energy and water storage changes of adjacent control volumes and the net exchange. It compares the observed mismatch quantity, interface closure quantity, and control volume conservation quantity with the corresponding positive tolerance of the parameter constraints, and ensures that the control volume conservation quantity does not exceed the reference parameter combination value, thus forming a set of permissible parameters with parameter combination identifier and applicable boundary domain. The process candidate generation module is used to generate process candidate data with process candidate identifiers according to process constraints; The state propagation module is used to propagate the process candidate data within the applicable boundary domain to the common termination time through the numerical model according to the parameter combination identifier of each parameter combination in the permitted parameter set, forming the parameter combination propagation completion status, the unit-by-unit termination neighborhood state envelope, the cumulative heat supply, and the continuous witnessing of the same identifier exceeding the limit. The process screening module is used to screen out feasible process candidate data where the parameter combination propagation completion status is fully completed and the extended duration of the witness has not exceeded the allowable duration. The results output module is used to output no feasible process results when the feasible process candidate data is empty, and to determine the target process data by sorting the data according to the common termination time, the state envelope of the termination neighborhood, the cumulative heat supply, and the process candidate identifier when the data is not empty.

[0065] Specifically, the system is implemented by a single computing device or multiple computing devices working collaboratively through controlled data connections. The model building module reads four types of inputs and establishes observation states and multiphysics numerical models; the permissible parameter formation module calls the same numerical model to form three types of comparison quantities and a set of permissible parameters; the process candidate generation module generates finite process candidate data; the state propagation module establishes a solution task using process candidate identifiers and parameter combination identifiers and outputs the termination neighborhood state envelope, cumulative heat supply, and continuous over-limit observations with the same identifier; the process screening module generates feasible process candidate data; and the result output module outputs results for infeasible processes or target process data according to a dual-terminal branch.

[0066] Each module exchanges data through version identifiers, object identifiers, process candidate identifiers, and parameter combination identifiers in the task directory. Modules can be program units executed on the same processor or computing nodes connected through task queues, but parameter combination identities, applicable boundary domains, propagation completion statuses, or over-limit witnesses must not be deleted at module boundaries. The result output module can connect to the drying equipment process setting interface or a manual verification terminal; when the interface write fails, the target process data is retained and the write failure status is output, without describing the unwritten data as an executed process.

[0067] The material categories, equipment structures, measurement locations, material relationships, parameter ranges, distance ranges, positive tolerances, target state ranges, and permissible durations in the above implementation process should be confirmed by the inventor based on the actual drying object, equipment data, measurement capabilities, and test records, and written into the corresponding input files. The current embodiment discloses the acquisition path, formation rules, and missing data handling for the aforementioned quantities, and does not treat unconfirmed models, sample sizes, performance improvement ratios, or test results as existing facts.

[0068] Interface closure permission and endpoint neighborhood witnessing form a continuous and synergistic relationship through parameter combination identifiers and applicable boundary domains. Interface closure permission alone can only exclude parameter combinations with good observation fit but unqualified heat and mass transfer or control volume conservation; it cannot prove that the candidate process satisfies local state constraints under all retained combinations. If endpoint state propagation deviates from the permission record, it may lead to over-exclusion due to using an overly wide parameter range, or omission of local risks due to using a single nominal parameter. With the combined effect of both, each endpoint state and out-of-limit witness can be traced back to the same parameter combination that has been checked by observation, interface, and control volume.

[0069] Where there is no conflict, the data acquisition, permission screening, process propagation, and anomaly handling features can be combined according to the same object identity and version relationship. The above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection; when making equivalent substitutions for the processing entity or storage medium, the data source, same identifier propagation, judgment boundaries, and dual-terminal output as defined in the claims should still be maintained.

[0070] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention, they should all fall within the protection scope of the present invention.

[0071] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0072] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A numerical simulation optimization method for drying processes based on multiphysics coupling, characterized in that, include: Acquire object data, observation data, parameter constraints, and process constraints. Based on the object data, establish a numerical model coupling heat and mass transfer with solid response. Map the observation data to the exchange interface and adjacent control volumes to form the observation state. The numerical model is used to solve the combination of parameters within the parameter constraints. The observed mismatch is formed by the observed state and the solution result. The interface closure is formed by the difference between the sensible heat of the fluid and the heat conduction and phase change heat of the material, and the difference between the water exchange between the fluid and the material, according to the same interface normal and time window. The control volume conservation is formed by the difference between the energy and water storage changes of adjacent control volumes and the net exchange. The observed mismatch, interface closure, and control volume conservation quantities are compared with the corresponding positive tolerances of the parameter constraints, and the control volume conservation quantities are ensured not to exceed the reference parameter combination values, thus forming a set of permissible parameters with parameter combination identifiers and applicable boundary domains. Process candidate data with process candidate identifiers are generated according to process constraints. For process candidate data within the applicable boundary domain, each parameter combination of the permitted parameter set is propagated to the common termination time through the numerical model according to the parameter combination identifier, forming the parameter combination propagation completion status, the unit-by-unit termination neighborhood status envelope, the cumulative heat supply, and the continuous witnessing of the same identifier exceeding the limit. Select feasible process candidate data whose parameter combination propagation completion status is fully completed and whose extended duration does not exceed the allowable duration.

2. The numerical simulation optimization method for drying process based on multiphysics coupling according to claim 1, characterized in that, The observed mismatch is formed by combining the observed state with the solution results. The interface closure is formed by the difference between the sensible heat of the fluid and the thermal conductivity and phase change heat of the material, and the difference between the water exchange between the fluid and the material, according to the same interface normal and time window. The control volume conservation is formed by the difference between the changes in energy and water storage of adjacent control volumes and the net exchange, including: The observation data is mapped to the exchange interface and adjacent control volumes according to the object identifier, coordinates and acquisition time. The difference between the observation data at the same location, at the same time and in the same unit and the solution results corresponding to the parameter combination is calculated, and the difference of each observation item is used to form the observation mismatch. Accumulate the fluid sensible heat, material thermal conductivity, phase change heat, and fluid-material moisture exchange within the same interface normal and time window. Subtract the sum of fluid sensible heat, material thermal conductivity, and phase change heat, and subtract the fluid moisture exchange from the material moisture exchange. The interface closure value is composed of the interface thermal difference and the interface moisture difference. Within the time window, obtain the changes in energy and moisture content and the corresponding net exchange volume of adjacent control volumes on both sides of the exchange interface, calculate the difference, and use the energy conservation difference and moisture conservation difference to form the control volume conservation quantity. The parameters are associated with the observed mismatch, interface closure, and control volume conservation quantities, and are then incorporated into the formation process of the permissible parameter set.

3. The numerical simulation optimization method for the drying process based on multiphysics coupling according to claim 1, characterized in that, The formation of the corresponding positive tolerance and reference parameter combination values ​​includes: The observation positive tolerance is determined based on the positive values ​​obtained from the measurement resolution, the state difference caused by the mapping distance, and the time discrete error, and the larger of the positive values ​​is taken. The larger of the positive values ​​of the spatial projection error and the positive values ​​of the time integration error of the discrete fluxes on both sides of the exchange interface is determined as the interface positive tolerance. Compare the positive values ​​of spatial discrete error and time integral error of adjacent control volumes, and take the larger one as the conserved positive tolerance. According to the control volume and time window used for the conservation of the control volume to be compared, read the solution results of the reference parameter combination pre-identified in the parameter constraints; Obtain the reference parameter combination values ​​from the solution results.

4. The numerical simulation optimization method for drying process based on multiphysics coupling according to claim 1, characterized in that, A set of licensed parameters with parameter combination identifiers and applicable boundary domains is formed, including: The finite values ​​within the parameter constraints are combined according to the parameter name and the order of their values, and a parameter combination identifier is assigned to each parameter combination. Ensure that the observed mismatch, interface closure, and control volume conservation are all no greater than their respective positive tolerances, and that the control volume conservation is no greater than the reference parameter combination value, and retain the parameter combination; The applicable boundary domain is formed from the lower and upper bounds of each component of the observed boundary state of the parameter combination, and the parameter combination identifier, applicable boundary domain and parameter combination association are written into the permitted parameter set. When the set of permitted parameters is empty, the result of no permitted parameters is output and the propagation of process candidate data is stopped. When it is not empty, the set of permitted parameters is output for the propagation of process candidate data.

5. The numerical simulation optimization method for drying process based on multiphysics coupling according to claim 1, characterized in that, Generate process candidate data with process candidate identifiers based on process constraints, including: The lower bound, upper bound and positive step size of each process boundary quantity are obtained from the process constraints, and combined into finite initial process candidate data according to the name and value order of each process boundary quantity. Assign initial process candidate identifiers to each initial process candidate data, and form candidate boundary states based on the initial state of the numerical model and process boundary quantities; Each component of the candidate boundary state is compared with the corresponding lower and upper bounds of the applicable boundary domain. The initial process candidate data in which each component is located between the corresponding lower and upper bounds is retained as the process candidate data, and its initial process candidate identifier is used as the process candidate identifier. The components that exceed the limit are written into the rejection record along with the initial process candidate identifier.

6. The numerical simulation optimization method for drying process based on multiphysics coupling according to claim 1, characterized in that, The propagation of various parameter combinations includes: Write the process candidate identifier, parameter combination identifier, and initial state of the numerical model into the corresponding solution task, and use the process candidate data as the boundary conditions of the solution task. Read parameter combinations from the set of permitted parameters according to the parameter combination identifier in the solution task, traverse all parameter combinations for each process candidate data, perform numerical model solution and complete the solution task; If a solution task fails to converge, or if the energy or water content is not finite, the corresponding process candidate will be marked as having incomplete propagation and the subsequent propagation of the process candidate data will be stopped. When all solution tasks are completed, the corresponding process candidate will be marked as propagation complete.

7. The numerical simulation optimization method for drying process based on multiphysics coupling according to claim 1, characterized in that, The formation of the common termination time and termination neighborhood includes: For each parameter combination that has completed propagation, check the humidity status of each spatial unit in chronological order. Record the first moment when all spatial units are continuously within the target humidity range of the process constraint for the duration of the holding time as the arrival time of the corresponding parameter combination. If any combination of parameters fails to reach a certain time in the solution time domain, the corresponding process candidate will be marked as having no valid termination. When all parameter combinations have reached their respective times, the latest time among all the times is taken as the common termination time. The effective sampling interval is taken from the observation data, the minimum switching interval is taken from the process constraints, and the solid response step size is taken from the numerical model. The larger of the three is taken as the endpoint observation time. The termination neighborhood is formed by extending the endpoint observation time forward from the common termination time.

8. The numerical simulation optimization method for drying process based on multiphysics coupling according to claim 1, characterized in that, The formation of the unit-by-unit termination neighborhood state envelope, cumulative heat supply, and continuous witnessing of exceeding the limit with the same identifier includes: Within the termination neighborhood, the humidity state, temperature state, and solid response state are obtained according to the process candidate identifier, parameter combination identifier, spatial unit identifier, and time. For the same spatial unit, the lower and upper bounds of all parameter combinations are taken for the humidity state, temperature state, and solid response state, respectively, to form the terminating neighborhood state envelope for each spatial unit. The humidity state, temperature state, and solid response state are compared with the corresponding allowable ranges of the process constraints. For the cumulative duration of continuous periods exceeding the allowable range, the process candidate identifier, parameter combination identifier, spatial unit identifier, state name, over-limit period, and duration are associated as a continuous over-limit witness with the same identifier.

9. The numerical simulation optimization method for drying process based on multiphysics coupling according to claim 1, characterized in that, Feasible process candidate data for which the parameter combination propagation completion status is fully completed and the extended duration of the witness test does not exceed the allowable duration are selected, including: Exclude process candidate data that are incompletely propagated, have no valid termination, or whose identical identifiers exceed the allowed duration, and form feasible process candidate data; When the feasible process candidate data is empty, the output will show no feasible process and the corresponding rejection record. When the feasible process candidate data is not empty, the process candidate data with the earlier common termination time, smaller envelope width of the humid state in the termination neighborhood, smaller cumulative heat supply, and smaller process candidate identifier are selected in sequence to determine the target process data.

10. A numerical simulation optimization system for a drying process based on multiphysics coupling, used to implement the numerical simulation optimization method for a drying process based on multiphysics coupling as described in any one of claims 1-9, comprising: The model building module is used to acquire object data, observation data, parameter constraints and process constraints, establish a numerical model of heat and mass transfer coupled with solid response based on object data, and map observation data to the exchange interface and adjacent control volume to form the observation state. The permissible parameter formation module is used to solve the parameter combination within the parameter constraints using a numerical model. It forms the observed mismatch quantity based on the observed state and the solution result. It forms the interface closure quantity based on the difference between the sensible heat of the fluid and the heat conduction and phase change heat of the material, and the difference between the water exchange between the fluid and the material, according to the same interface normal and time window. It forms the control volume conservation quantity based on the difference between the energy and water storage changes of adjacent control volumes and the net exchange. It compares the observed mismatch quantity, interface closure quantity, and control volume conservation quantity with the corresponding positive tolerance of the parameter constraints, and ensures that the control volume conservation quantity does not exceed the reference parameter combination value, thus forming a set of permissible parameters with parameter combination identifier and applicable boundary domain. The process candidate generation module is used to generate process candidate data with process candidate identifiers according to process constraints. The state propagation module is used to propagate the process candidate data within the applicable boundary domain to the common termination time through the numerical model according to the parameter combination identifier of each parameter combination in the permitted parameter set, forming the parameter combination propagation completion status, the unit-by-unit termination neighborhood state envelope, the cumulative heat supply, and the continuous witnessing of the same identifier exceeding the limit. The process screening module is used to screen out feasible process candidate data where the parameter combination propagation completion status is fully completed and the extended duration of the witness has not exceeded the allowable duration. The results output module is used to output no feasible process results when the feasible process candidate data is empty, and to determine the target process data by sorting the data according to the common termination time, the state envelope of the termination neighborhood, the cumulative heat supply, and the process candidate identifier when the data is not empty.

Images