Digital-twin-based ultra-soft thermoplastic elastomer precision molding process parameter determination system and method

Abstract

This invention discloses a system and method for determining precision molding process parameters of ultra-soft thermoplastic elastomers based on digital twins, belonging to the field of computer-aided design technology. It includes a molding data processing module, which receives geometric data of the product to be molded, material data of the ultra-soft thermoplastic elastomer, candidate process parameter ranges, and molding quality constraints, and processes this data into a dataset that can be used for digital twin simulation. A dual-model simulation module is also included, which establishes a first molding simulation model and a second molding simulation model based on the dataset, and generates two delayed deformation prediction results for the same candidate process parameter. This invention determines a stable process parameter window based on the degree of divergence between the two models' delayed deformation prediction results by establishing the first and second molding simulation models and performing virtual perturbation simulation on the same candidate process parameter.

Description

Technical Field

[0001] This invention relates to the field of computer-aided design technology, specifically to a system and method for determining process parameters of precision molding of ultra-soft thermoplastic high elastomers based on digital twins. Background Technology

[0002] Ultra-soft thermoplastic elastomers possess low hardness, high resilience, high ductility, and thermoplastic processability, making them suitable for precision molding of flexible seals, cushioning structures, thin-walled elastic components, mesh elastic components, and irregularly shaped elastic components. The molding quality of these products is closely related to the material's rheological state, mold cavity geometry, filling process, cooling process, demolding release process, and the combination of process parameters. Digital twin technology can map product geometric data, material data, process data, and molding quality data into a computational model for simulating, predicting, and determining process parameters during molding, thus providing a data-driven design basis for the precision molding of ultra-soft thermoplastic elastomers.

[0003] In precision molding scenarios involving thin-walled, mesh-like, film-like, and parting-face adjacent structures of ultrasoft thermoplastic elastomers, existing digital twin simulations typically rely on a single molding assumption to determine the cavity filling state, pressure state, temperature state, and demolding dimensional state. Even when simulation results show sufficient filling and acceptable demolding dimensions, the product may still exhibit delayed springback, localized collapse, edge curling, and mesh misalignment during the stress release phase after demolding. This problem is not simply caused by insufficient filling, inadequate cooling, or dimensional shrinkage, but rather by a combination of near-wall shear accumulation in ultrasoft materials, the cooling and solidification sequence, and deformation release after demolding. Existing single models struggle to identify this hidden process window—where demolding is acceptable but instability occurs after release—leading to insufficient stability of the determined process parameters in actual precision molding. Summary of the Invention

[0004] The purpose of this invention is to provide a system and method for determining the process parameters of precision molding of ultra-soft thermoplastic high elastomers based on digital twins, so as to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a system for determining precision molding process parameters of ultra-soft thermoplastic high elastomers based on digital twins, comprising: a molding data processing module, which receives geometric data of the product to be molded, material data of the ultra-soft thermoplastic high elastomer, candidate process parameter ranges, and molding quality constraints, and processes the data into a dataset that can be used for digital twin simulation; a dual-model simulation module, which establishes a first molding simulation model and a second molding simulation model based on the dataset, and generates two delayed deformation prediction results for the same candidate process parameter; and a process parameter determination module, which compares the degree of divergence between the two delayed deformation prediction results under virtual perturbation, and determines a stable process parameter window from the candidate process parameters based on the degree of divergence, wherein the delayed deformation is a local deformation generated after the molded part is demolded for a preset release time.

[0006] According to the above technical solution, the molding data processing module includes a geometric unit division submodule, a material data standardization submodule, and a candidate parameter generation submodule; the geometric unit division submodule is used to divide the product to be molded into several feature units, the material data standardization submodule is used to convert the material hardness, viscosity, recovery characteristics, and shrinkage characteristics into a unified data format, and the candidate parameter generation submodule is used to generate a set of candidate process parameters within the range of candidate process parameters; The dual-model simulation module includes a first simulation submodule, a second simulation submodule, and a virtual perturbation simulation submodule. The first simulation submodule is used to predict delayed deformation under the computational assumption that the material does not slip relative to the mold cavity wall. The second simulation submodule is used to predict delayed deformation under the computational assumption that the shear history in the near-wall region is retained after cooling and released after demolding. The virtual perturbation simulation submodule is used to apply small parameter perturbations to the same candidate process parameter and call the first simulation submodule and the second simulation submodule respectively. The process parameter determination module includes a divergence calculation submodule, a risk parameter elimination submodule, and a result output submodule. The divergence calculation submodule is used to calculate the difference in late-deformation prediction between the two models. The risk parameter elimination submodule is used to eliminate candidate process parameters whose divergence exceeds the stability threshold. The result output submodule is used to output a stable process parameter window, a recommended process parameter group, and a late-deformation risk unit.

[0007] A precision molding method for ultra-soft thermoplastic elastomers based on digital twins includes the following steps: S1. Organize the molding input data, receive the three-dimensional geometric data, material data, candidate process parameter range and molding quality constraints of the product to be molded, divide the product to be molded into several feature units, and mark the thin-walled area, mesh intersection area, parting surface adjacent area and edge free area, establish the mapping relationship between each feature unit and material data and process data, and form a unified input dataset for digital twin simulation. S2. Generate candidate process parameters and virtual perturbation data. Generate multiple sets of candidate process parameters according to the range of candidate process parameters. For each set of candidate process parameters, generate a benchmark parameter set and multiple small perturbation parameter sets. The small perturbation parameter sets are used to observe whether the prediction results of the two models produce opposite trends when the process parameters change slightly. S3. Execute the first forming simulation and the second forming simulation respectively. The first forming simulation calculates the delayed deformation prediction results of each characteristic element under the assumption of wall-attached flow. The second forming simulation calculates the delayed deformation prediction results of each characteristic element under the assumption of near-wall shear cumulative release. Both forming simulations are executed based on the same candidate process parameters and its small perturbation parameter group to ensure that the divergence comparison has the same input basis. S4. Calculate the degree of divergence of delayed deformation between the two models, compare the delayed deformation prediction results obtained from the first forming simulation and the second forming simulation, determine whether the change direction of the same feature unit before and after the small disturbance is consistent, and calculate the normalized difference between the prediction results of the two models. Based on the normalized difference and the consistency of the change direction, obtain the divergence risk value of the candidate process parameters. S5. Determine the stable process parameter window, retain the candidate process parameters whose divergence risk value does not exceed the stable threshold as stable candidate parameters, generate the stable process parameter window based on the stable candidate parameters, determine the recommended process parameter group from the window, and output the stable process parameter window, the recommended process parameter group and the delayed deformation risk unit as the determination result of the precision molding process parameters of the ultra-soft thermoplastic high elastomer.

[0008] According to the above technical solution, step S1 includes the following steps: S1-1. Divide the three-dimensional geometric data of the product to be formed into a set of feature units. ,in, For the first Each feature unit The feature unit number is used as the smallest computational object for delayed deformation prediction, model divergence comparison, and risk unit output, so that different local structures of the same product can participate in subsequent digital twin simulations. S1-2. After completing the feature element division, the local thickness, free edge distance, and allowable delayed deformation upper limit of each feature element are used as the core inputs for the simulation of that feature element. This establishes a feature data vector for each feature element. Establish feature data vectors ,in, For the first Feature data vectors of each feature unit For the first The local thickness of each feature unit. For the first The distance from each feature unit to the nearest free edge, To be assigned to the first according to the molding quality constraints The formula defines the upper limit of allowed delayed deformation for each feature element; it is used to unify geometric shape and mass constraints under the same feature element, enabling subsequent simulation results to be compared element by element. The structural types of the thin-walled region, mesh intersection region, parting surface adjacent region, and edge free region are used as the corresponding feature elements. The attribute identifiers are saved and stored through local thickness. Free edge distance and the upper limit of allowed delayed deformation Participate in subsequent calculations; S1-3. Organize the material data of the ultrasoft thermoplastic elastomer into a material data vector. ,in, For material data vectors, This represents the material's hardness value. This is melt viscosity data. For stretch recovery data, This is for cooling and shrinkage data; the formula is used to take the material's flow properties, recovery properties, and shrinkage properties as common inputs for digital twin simulation; S1-4, Transfer the feature data vector With the material data vector By performing correlation, a unified input dataset for digital twin simulation is obtained. ,in, To standardize the input dataset, Indicates the first Each feature element and its corresponding feature data vector; this formula is used to ensure that each subsequent simulation uses the same geometric element, the same material data, and the same upper limit of delayed deformation as the input basis, and the feature data vector The upper limit of delayed deformation for the corresponding feature unit has been included. .

[0009] According to the above technical solution, step S2 includes the following steps: S2-1. To ensure that the first and second molding simulation models are compared under completely identical process inputs, the process conditions affecting filling, cooling, and release deformation are uniformly grouped into a candidate process parameter vector. Based on the range of candidate process parameters and the preset sampling step size for each type of process parameter, the first... Group of candidate process parameter vectors ,in, For the first Group of candidate process parameter vectors, Candidate number, The melting temperature is... For mold temperature, For injection speed, To maintain pressure, For the pressure holding time, This is the cooling time; this formula is used to fix the main forming conditions affecting delayed deformation into the same vector to avoid inconsistencies in subsequent simulation inputs. S2-2, The virtual perturbation vector is generated according to a preset perturbation step size, and the perturbation amplitude of each component is less than the allowable adjustment amplitude of the corresponding process parameter. It is used to simulate the change in prediction results when the process parameters fluctuate slightly. For each set of candidate process parameter vectors Generate virtual perturbation parameter vector ,in, For the first Group of candidate process parameters in the first The parameter vector after the virtual perturbation. This is a virtual perturbation sequence number. For the first The virtual perturbation vector; this formula is used to observe the impact of small parameter changes on simulation results without changing the actual equipment state. By applying multiple virtual perturbations to the same candidate process parameter, it is possible to determine whether there are hidden instability regions near the candidate process parameter that are highly sensitive to small parameter changes. With candidate process parameter vector Parameter vectors with the same parameter component order and obtained through virtual perturbation It remains within the range of the candidate process parameters; S2-3, Set the reference perturbation vector without applied virtual perturbation to... ,in, This is the reference perturbation vector; this formula is used to make Indicates the first The baseline input of the candidate process parameters enables subsequent steps to use the simulation results of the baseline input as the basis for calculating the changes.

[0010] According to the above technical solution, step S3 includes the following steps: S3-1, For each set of parameter vectors First, the first model is called to calculate the feature units. Based on the conventional wall-adhering flow assumption, the predicted delayed deformation after demolding is calculated using the first molding simulation model for characteristic elements. In parameter vector First delayed deformation prediction value ,in, This is the predicted value of delayed deformation output by the first molding simulation model. This is a digital twin simulation function under the assumption of near-wall flow; this formula is used to simulate the delayed deformation results when the near-wall material does not slip. The first forming simulation model uses a unified input dataset. Feature data vectors in Material data vector and parameter vector As input, output the predicted delayed deformation value of the corresponding feature unit after demolding and release; S3-2. In the second molding simulation model, first, in the same parameter vector... Lower feature extraction unit The data on near-wall shear rate versus time and cooling-to-solidification ratio versus time were collected and accumulated over the cooling time range to characterize the tendency of releaseable deformation stored in the near-wall region during molding and cooling. Feature elements were then calculated using a second molding simulation model. In parameter vector Near-wall cumulative shear ,in, This represents the cumulative near-wall shear. For simulation time Lower feature unit The near-wall shear rate, For simulation time Lower feature unit The cooling and curing ratio, parameter vector The formula is used to convert the near-wall shear history and the cooling solidification process into a calculable release risk input. The near-wall shear rate is used to characterize the shear history of the ultra-soft thermoplastic elastomer near the mold cavity wall, and the cooling solidification ratio is used to characterize the extent to which this shear history is retained until the demolding stage. Therefore, the two together determine the release risk input in the near-wall region. S3-3, Based on material data vector Stretching recovery data Cooling shrinkage data and feature units Geometric properties to determine feature units Deformation release conversion coefficient ,in This is a deformation release conversion function pre-established based on material tensile recovery data, cooling shrinkage data, and geometric properties of characteristic elements. The first value calculated by the deformation release conversion function The deformation release conversion coefficient of each feature unit is used to convert the near-wall shear accumulation into additional delayed deformation after demolding, based on the near-wall shear accumulation. Calculate the second delayed deformation prediction value ,in, The delayed deformation prediction value output by the second forming simulation model. For feature units The deformation release conversion coefficient; this formula is used to convert the cumulative near-wall shear into additional delayed deformation after demolding, so that the second molding simulation model can reflect the instability trend of the ultra-soft thermoplastic elastomer after release.

[0011] According to the above technical solution, step S4 includes the following steps: S4-1. To determine the response direction of the first modeling simulation to virtual perturbations, the difference between the predicted value of the first delayed deformation under k virtual perturbations and the predicted value of the first delayed deformation under the reference input is calculated. Based on the reference parameter vector and the virtual perturbation parameter vector, the change in the first model is calculated. ,in, For the first molded simulation model in the 1st The formula is used to determine the direction of the effect of small parameter changes on the delayed deformation under the assumption of wall-attached flow. S4-2. To determine the response direction of the second-formed simulation model to the same virtual perturbation, the first... The difference between the second delayed deformation prediction value under the virtual perturbation and the second delayed deformation prediction value under the reference input is calculated. The change in the second model is then calculated based on the reference parameter vector and the virtual perturbation parameter vector. ,in, For the second molding simulation model in the first The formula is used to determine the direction of the effect of small parameter changes on the delayed deformation under the near-wall shear cumulative release assumption. S4-3. When the product of the first model change and the second model change is less than zero, it indicates that the same virtual perturbation causes the delayed deformation prediction results of the two models to change in opposite directions. This situation is marked as a reverse response, and the reverse response label is calculated based on the first model change and the second model change. ,in, For reverse response markers, This is a conditional function; it takes the value 1 when the condition inside the parentheses is true and 0 when the condition inside the parentheses is false. This formula is used to mark the case where two models give opposite directions of change for the same virtual perturbation. S4-4. To prevent any single risk from being offset by other low-risk items, the maximum value among the first delayed deformation proportion, the second delayed deformation proportion, the difference proportion between the two model predictions, and the inverse response marker is used as the unit divergence risk value. The unit divergence risk value is calculated based on the first delayed deformation prediction value, the second delayed deformation prediction value, the inverse response marker, and the allowed upper limit of delayed deformation. ,in, For feature units In the Group of candidate process parameters and the first The unit divergence risk value under the second virtual perturbation; this formula is used to simultaneously constrain the delayed deformation of the first model, the delayed deformation of the second model, the difference in prediction between the two models, and the consistency of the response direction of the two models. Through the above maximum value determination method, as long as the candidate process parameter has a delayed deformation close to the upper limit, a large difference in model prediction, or opposite response directions of the two models under any feature unit or any virtual perturbation, it can be identified as a hidden instability risk.

[0012] According to the above technical solution, step S5 includes the following steps: S5-1. To ensure that the stability evaluation of candidate process parameters covers all characteristic elements and all virtual perturbations, the maximum value among all element divergence risk values ​​is taken as the overall risk value of the candidate process parameter. The risk value of the candidate process parameter is calculated based on the element divergence risk values ​​under all characteristic elements and all virtual perturbations. ,in, For the first The risk value of the candidate process parameters in the group of candidate process parameters; this formula is used to determine whether the candidate process parameters are stable based on the most unstable feature element and the most sensitive virtual perturbation; S5-2. Compare the risk values ​​of candidate process parameters with the stability threshold. Candidate process parameters with risk values ​​not exceeding the stability threshold are retained, while those with risk values ​​exceeding the stability threshold are eliminated. Determine the set of stable candidate parameters based on their risk values. ,in, To stabilize the candidate parameter set, For a stable threshold, and This formula is used to eliminate candidate process parameters that exhibit delayed deformation exceeding limits, model prediction discrepancies exceeding limits, or inverse responses. S5-3. If the set of stable candidate parameters is not empty, select the group of candidate process parameters with the lowest risk value from the set of stable candidate parameters as the recommended process parameter group, and determine the recommended process parameter group based on the set of stable candidate parameters. This formula is used to select the candidate process parameter with the lowest overall divergence risk value from the stable candidate parameter set; S5-4. For each type of process parameter in the stable candidate parameter set, take the minimum and maximum values ​​to form the stable value range of each process parameter, and determine the stable process parameter window based on the stable candidate parameter set. The minimum and maximum values ​​in the formula are calculated separately for each component of the process parameter vector; this formula is used to output the stable range of values ​​for melt temperature, mold temperature, injection speed, holding pressure, holding time, and cooling time. S5-5. Determine the set of delayed deformation risk units based on the unit divergence risk value. ,in, The set of delayed deformation risk units; when the set of stable candidate parameters When empty, the output shows the determination result that there is no stable process parameter window and the set of delayed deformation risk units. When the stable candidate parameter set If not empty, output the stable process parameter window. The recommended process parameter set and the set of delayed deformation risk units The results serve as the determination of process parameters for precision molding of ultra-soft thermoplastic high elastomers.

[0013] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention uses the predicted discrepancies under two reasonable molding assumptions as the basis for judging hidden instability risks. The first molding simulation model is used to simulate the molding results under conventional wall-attached flow, and the second molding simulation model is used to simulate the delayed deformation results formed by the cumulative near-wall shear after demolding. By applying small virtual perturbations to candidate process parameters, dangerous parameter areas that meet the filling and size requirements on the surface but are highly sensitive to disturbances can be identified. The resulting stable process parameter window simultaneously satisfies conventional molding quality constraints, delayed deformation constraints, and model discrepancy constraints, which can solve the problem of ultra-soft thermoplastic high-elasticity precision parts being qualified after demolding but unstable after release. Furthermore, this solution is based on computer simulation, data processing, and parameter determination, and does not rely on additional molding equipment structures. Attached Figure Description

[0014] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating the present invention; Figure 2 This is a schematic diagram of the overall modular structure of the present invention. Detailed Implementation

[0015] 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.

[0016] Please see Figure 1 and Figure 2 This invention provides a technical solution: a system for determining precision molding process parameters of ultra-soft thermoplastic high elastic body based on digital twin, comprising: a molding data processing module, which receives geometric data of the product to be molded, material data of the ultra-soft thermoplastic high elastic body, candidate process parameter ranges, and molding quality constraints, and processes the above data into a dataset that can be called by digital twin simulation; a dual-model simulation module, which establishes a first molding simulation model and a second molding simulation model based on the dataset, and generates two delayed deformation prediction results for the same candidate process parameter; and a process parameter determination module, which compares the degree of divergence between the two delayed deformation prediction results under virtual perturbation, and determines a stable process parameter window from the candidate process parameters according to the degree of divergence, wherein delayed deformation is the local deformation generated after the molded part is demolded for a preset release time; The molding data processing module includes a geometric unit division submodule, a material data standardization submodule, and a candidate parameter generation submodule. The geometric unit division submodule is used to divide the product to be molded into several feature units. The material data standardization submodule is used to convert the material hardness, viscosity, recovery characteristics, and shrinkage characteristics into a unified data format. The candidate parameter generation submodule is used to generate a set of candidate process parameters within the range of candidate process parameters. The dual-model simulation module includes a first simulation submodule, a second simulation submodule, and a virtual perturbation simulation submodule. The first simulation submodule is used to predict delayed deformation under the computational assumption that the material does not slip relative to the mold cavity wall. The second simulation submodule is used to predict delayed deformation under the computational assumption that the shear history in the near-wall region is retained after cooling and released after demolding. The virtual perturbation simulation submodule is used to apply small parameter perturbations to the same candidate process parameter and call the first simulation submodule and the second simulation submodule respectively. The process parameter determination module includes a divergence calculation submodule, a risk parameter elimination submodule, and a result output submodule. The divergence calculation submodule is used to calculate the difference in late-deformation prediction between the two models. The risk parameter elimination submodule is used to eliminate candidate process parameters whose divergence exceeds the stability threshold. The result output submodule is used to output the stable process parameter window, the recommended process parameter group, and the late-deformation risk unit. The focus of this embodiment is not simply to improve the prediction accuracy of a single molding simulation model, but to identify hidden instability risks in the precision molding of ultra-soft thermoplastic elastomers that are easily masked by a single model. For thin-walled, mesh, parting surface adjacent, and edge-free structures, the molded part may meet dimensional requirements at the moment of demolding, but the deformation accumulated in its near-wall region during filling, cooling, and demolding may not be released synchronously. Some local areas may exhibit springback, collapse, or curling only after a short recovery process following demolding. Therefore, this embodiment simulates the same candidate process parameter under two different molding assumptions and further observes whether the two simulation results diverge under slight process disturbances. This transforms delayed instability, which is difficult to observe directly at demolding, into calculable and screenable process parameter risks in advance.

[0017] A precision molding method for ultra-soft thermoplastic elastomers based on digital twins includes the following steps: S1. Organize the molding input data, receive the three-dimensional geometric data, material data, candidate process parameter range and molding quality constraints of the product to be molded, divide the product to be molded into several feature units, and mark the thin-walled area, mesh intersection area, parting surface adjacent area and edge free area, establish the mapping relationship between each feature unit and material data and process data, and form a unified input dataset for digital twin simulation. S2. Generate candidate process parameters and virtual perturbation data. Generate multiple sets of candidate process parameters according to the range of candidate process parameters. For each set of candidate process parameters, generate a baseline parameter set and multiple small perturbation parameter sets. The small perturbation parameter sets are used to observe whether the prediction results of the two models produce opposite trends when the process parameters change slightly. S3. Execute the first forming simulation and the second forming simulation respectively. The first forming simulation calculates the delayed deformation prediction results of each characteristic element under the assumption of wall-attached flow. The second forming simulation calculates the delayed deformation prediction results of each characteristic element under the assumption of near-wall shear cumulative release. Both forming simulations are executed based on the same candidate process parameters and its small perturbation parameter group to ensure that the divergence comparison has the same input basis. S4. Calculate the degree of divergence of delayed deformation between the two models, compare the delayed deformation prediction results obtained from the first forming simulation and the second forming simulation, determine whether the change direction of the same feature unit before and after the small disturbance is consistent, and calculate the normalized difference between the prediction results of the two models. Based on the normalized difference and the consistency of the change direction, obtain the divergence risk value of the candidate process parameters. S5. Determine the stable process parameter window, retain the candidate process parameters whose divergence risk value does not exceed the stable threshold as stable candidate parameters, generate the stable process parameter window based on the stable candidate parameters, determine the recommended process parameter group from the window, and output the stable process parameter window, the recommended parameter group and the delayed deformation risk unit as the determination result of the precision molding process parameters of the ultra-soft thermoplastic high elastomer. Step S1 includes the following steps: S1-1. Divide the three-dimensional geometric data of the product to be formed into a set of feature units. ,in, For the first Each feature unit The feature unit number is used as the smallest computational object for late deformation prediction, model divergence comparison, and risk unit output, so that different local structures of the same product can participate in subsequent digital twin simulations. S1-2. After completing the feature element division, the local thickness, free edge distance, and allowable delayed deformation upper limit of each feature element are used as the core inputs for the simulation of that feature element. This establishes a feature data vector for each feature element. Establish feature data vectors ,in, For the first Feature data vectors of each feature unit For the first The local thickness of each feature unit. For the first The distance from each feature unit to the nearest free edge, To be allocated to the first according to molding quality constraints The formula defines the upper limit of allowed delayed deformation for each feature element; it is used to unify geometric shape and mass constraints under the same feature element, enabling subsequent simulation results to be compared element by element. The structural types of thin-walled regions, mesh intersection regions, parting surface adjacent regions, and edge free regions are used as the corresponding feature elements. The attribute identifiers are saved and stored through local thickness. Free edge distance and the upper limit of allowed delayed deformation Participate in subsequent calculations; S1-3. Organize the material data of the ultrasoft thermoplastic elastomer into a material data vector. ,in, For material data vectors, This represents the material's hardness value. This is melt viscosity data. For stretch recovery data, This is for cooling and shrinkage data; the formula is used to take the material's flow properties, recovery properties, and shrinkage properties as common inputs for digital twin simulation; S1-4, Convert the feature data vector Material data vector By performing correlation, a unified input dataset for digital twin simulation is obtained. ,in, To standardize the input dataset, Indicates the first Each feature element and its corresponding feature data vector; this formula is used to ensure that each subsequent simulation uses the same geometric element, the same material data, and the same upper limit of delayed deformation as the input basis, and the feature data vector The upper limit of delayed deformation for the corresponding feature unit has been included. ; In this embodiment, the feature element division is not simply for discretizing the product model, but rather for locating the risk of delayed deformation to specific local structures. Abnormal deformation in ultrasoft thermoplastic elastomers often does not occur simultaneously across the entire product, but rather first appears at local locations such as abrupt thickness changes, free edges, mesh intersections, and parting line adjacencies. If molding quality is evaluated solely based on the overall dimensions of the entire product, localized release deformation can easily be masked within the overall acceptable result. This embodiment uses feature elements to store local thickness, free edge distance, and the upper limit of allowable delayed deformation, enabling subsequent comparisons between two simulation models on the same local object and ensuring that the final output risk location corresponds to a specific molding structure.

[0018] Step S2 includes the following steps: S2-1. To ensure that the first and second molding simulation models are compared under completely identical process inputs, the process conditions affecting filling, cooling, and release deformation are uniformly grouped into a candidate process parameter vector. Based on the range of candidate process parameters and the preset sampling step size for each type of process parameter, the first... Group of candidate process parameter vectors ,in, For the first Group of candidate process parameter vectors, Candidate number, The melting temperature is... For mold temperature, For injection speed, To maintain pressure, For the pressure holding time, This is the cooling time; this formula is used to fix the main forming conditions affecting delayed deformation into the same vector to avoid inconsistencies in subsequent simulation inputs. S2-2. The virtual perturbation vector is generated according to a preset perturbation step size, and the perturbation amplitude of each component is less than the allowable adjustment amplitude of the corresponding process parameter. It is used to simulate the change in prediction results when the process parameters fluctuate slightly. For each set of candidate process parameter vectors Generate virtual perturbation parameter vector ,in, For the first Group of candidate process parameters in the first The parameter vector after the virtual perturbation. This is a virtual perturbation sequence number. For the first The virtual perturbation vector; this formula is used to observe the impact of small parameter changes on simulation results without altering the actual equipment state. By applying multiple virtual perturbations to the same candidate process parameter, it is possible to determine whether there are hidden instability regions near the candidate process parameter that are highly sensitive to small parameter changes. With candidate process parameter vector Parameter vectors with the same parameter component order and obtained through virtual perturbation Still within the range of candidate process parameters; S2-3, Set the reference perturbation vector without applied virtual perturbation to... ,in, This is the reference perturbation vector; this formula is used to make Indicates the first The baseline input of the candidate process parameters enables subsequent steps to use the simulation results of the baseline input as the basis for calculating the changes. The purpose of setting up virtual perturbations in this embodiment is not to find optimal parameters in the conventional sense, but to observe whether there are regions near candidate process parameters that are abnormally sensitive to slight process changes. In actual precision molding, equipment control errors, material batch fluctuations, and local cooling differences can all cause slight deviations in the process state. Conventional parameter optimization usually only judges whether a set of parameters is qualified, while this embodiment further judges whether the local neighborhood around the set of parameters is stable. When a slight perturbation can cause two simulation models to show opposite trends in delayed deformation, it indicates that although the candidate parameter may be qualified under the benchmark simulation, there are hidden instability boundaries in its vicinity, and it is not suitable as a stable process parameter output.

[0019] Step S3 includes the following steps: S3-1, For each set of parameter vectors First, the first model is called to calculate the feature units. Based on the conventional wall-adhering flow assumption, the predicted delayed deformation after demolding is calculated using the first molding simulation model for characteristic elements. In parameter vector First delayed deformation prediction value ,in, This is the predicted value of delayed deformation output by the first molding simulation model. This is a digital twin simulation function under the assumption of near-wall flow; this formula is used to simulate the delayed deformation results when the near-wall material does not slip. The first forming simulation model uses a unified input dataset. Feature data vectors in Material data vector and parameter vector As input, output the predicted delayed deformation value of the corresponding feature unit after demolding and release; S3-2. In the second molding simulation model, first, in the same parameter vector... Lower feature extraction unit The data on near-wall shear rate versus time and cooling-to-solidification ratio versus time were collected and accumulated over the cooling time range to characterize the tendency of releaseable deformation stored in the near-wall region during molding and cooling. Feature elements were then calculated using a second molding simulation model. In parameter vector Near-wall cumulative shear ,in, This represents the cumulative near-wall shear. For simulation time Lower feature unit The near-wall shear rate, For simulation time Lower feature unit The cooling and curing ratio, parameter vector The formula is used to convert the near-wall shear history and the cooling solidification process into a calculable release risk input. The near-wall shear rate is used to characterize the shear history of the ultra-soft thermoplastic elastomer near the mold cavity wall, and the cooling solidification ratio is used to characterize the extent to which this shear history is retained until the demolding stage. Therefore, the two together determine the release risk input in the near-wall region. S3-3, Based on material data vector Stretching recovery data Cooling shrinkage data and feature units Geometric properties to determine feature units Deformation release conversion coefficient ,in This is a deformation release conversion function pre-established based on material tensile recovery data, cooling shrinkage data, and geometric properties of characteristic elements. The first value calculated by the deformation release conversion function The deformation release conversion coefficient of each feature unit is used to convert the cumulative near-wall shear into additional delayed deformation after demolding, based on the cumulative near-wall shear. Calculate the second delayed deformation prediction value ,in, The delayed deformation prediction value output by the second forming simulation model. For feature units The deformation release conversion coefficient; this formula is used to convert the cumulative near-wall shear into additional delayed deformation after demolding, so that the second molding simulation model can reflect the instability trend of the ultra-soft thermoplastic elastomer after release. In this embodiment, the first molding simulation model and the second molding simulation model are not two overlapping predictive models. The first molding simulation model represents the molding result when the material flows near the mold cavity wall in a conventional wall-attaching manner, reflecting the delayed deformation trend that can be seen in ordinary molding simulations. The second molding simulation model, on the other hand, uses the near-wall shear history and cooling retention process as sources of risk release, reflecting the additional deformation trend that only appears after demolding of ultrasoft materials. The two are not interchangeable, but rather represent two reasonable interpretations of the same molding process. This embodiment does not directly determine which model is absolutely correct, but rather uses the difference in predictions between the two models for the same input conditions to discover delayed instability risks that are not easily exposed by conventional models.

[0020] For ultra-soft thermoplastic elastomers, the shear stress in the near-wall region does not necessarily disappear immediately after filling. Especially in areas with thin walls, mesh intersections, and free edges, the internal recovery tendency of the material, locked in by the cooling process, may be delayed until after demolding. This embodiment uses both the near-wall shear history and the degree of cooling retention as sources of release risk in the second molding simulation model, enabling it to reflect the additional deformation after demolding that is difficult to directly express in conventional filling simulations. The effect of this approach is that the system does not wait until the product actually becomes unstable before correcting the parameters, but rather eliminates candidate parameters that may lead to instability after release during the parameter determination stage.

[0021] Step S4 includes the following steps: S4-1. To determine the response direction of the first modeling simulation to virtual perturbations, the difference between the predicted value of the first delayed deformation under k virtual perturbations and the predicted value of the first delayed deformation under the reference input is calculated. Based on the reference parameter vector and the virtual perturbation parameter vector, the change in the first model is calculated. ,in, For the first molded simulation model in the 1st The formula is used to determine the direction of the effect of small parameter changes on the delayed deformation under the assumption of wall-attached flow. S4-2. To determine the response direction of the second-formed simulation model to the same virtual perturbation, the first... The difference between the second delayed deformation prediction value under the virtual perturbation and the second delayed deformation prediction value under the reference input is calculated. The change in the second model is then calculated based on the reference parameter vector and the virtual perturbation parameter vector. ,in, For the second molding simulation model in the first The formula is used to determine the direction of the effect of small parameter changes on the delayed deformation under the near-wall shear cumulative release assumption. S4-3. When the product of the changes in the first model and the changes in the second model is less than zero, it indicates that the same virtual perturbation causes the delayed deformation prediction results of the two models to change in opposite directions. This situation is marked as a reverse response, and the reverse response label is calculated based on the changes in the first model and the changes in the second model. ,in, For reverse response markers, This is a conditional function; it takes the value 1 when the condition inside the parentheses is true and 0 when the condition inside the parentheses is false. This formula is used to mark the case where two models give opposite directions of change for the same virtual perturbation. S4-4. To prevent any single risk from being offset by other low-risk items, the maximum value among the proportions of the first delayed deformation, the second delayed deformation, the difference between the two model predictions, and the reverse response marker is used as the unit divergence risk value. The unit divergence risk value is calculated based on the predicted values ​​of the first and second delayed deformations, the reverse response marker, and the allowed upper limit of delayed deformation. ,in, For feature units In the Group of candidate process parameters and the first The unit divergence risk value under the second virtual perturbation; this formula is used to simultaneously constrain the delayed deformation of the first model, the delayed deformation of the second model, the difference in prediction between the two models, and the consistency of the response direction of the two models. Through the above maximum value determination method, as long as the candidate process parameter has a delayed deformation close to the upper limit, a large difference in model prediction, or opposite response directions of the two models under any feature unit or any virtual perturbation, it can be identified as a hidden instability risk. In calculating the degree of divergence, this embodiment does not use a weighted summation of multiple indicators. Instead, it treats delayed deformation exceeding limits, excessive differences in predictions between the two models, and opposite responses from the two models to the same perturbation as independent sources of risk. This design is because delayed instability in ultrasoft thermoplastic elastomers is often triggered by local critical phenomena. If a particular feature unit exhibits significant instability under a single virtual perturbation, it can lead to the subsequent failure of the entire precision-molded part. If a weighted averaging method is used, local high risks may be offset by other low-risk areas. This embodiment retains local anomalous signals through the worst-case scenario determination method, allowing hidden instability boundaries to be directly identified rather than being masked by the overall averaging result.

[0022] The reverse response marker is used to identify a special instability situation where the same slight process change manifests as a decrease in delayed deformation in the first molding simulation model, but an increase in delayed deformation in the second molding simulation model, or vice versa. This phenomenon indicates that the candidate process parameter is near the boundary where the judgments of the two molding mechanisms are inconsistent. For conventional parameter determination methods, such parameters may be retained because a model's prediction is satisfactory; however, in the precision molding of ultrasoft thermoplastic elastomers, these parameters are more sensitive to material batch fluctuations and minor process deviations. Therefore, in this embodiment, they are directly included as hidden instability risks in the screening criteria.

[0023] Step S5 includes the following steps: S5-1. To ensure that the stability evaluation of candidate process parameters covers all characteristic elements and all virtual perturbations, the maximum value among all element divergence risk values ​​is taken as the overall risk value of the candidate process parameter. The risk value of the candidate process parameter is calculated based on the element divergence risk values ​​under all characteristic elements and all virtual perturbations. ,in, For the first The risk value of the candidate process parameters in the group of candidate process parameters; this formula is used to determine whether the candidate process parameters are stable based on the most unstable feature element and the most sensitive virtual perturbation; S5-2. Compare the risk values ​​of candidate process parameters with the stability threshold. Candidate process parameters with risk values ​​not exceeding the stability threshold are retained, while those with risk values ​​exceeding the stability threshold are eliminated. Determine the set of stable candidate parameters based on their risk values. ,in, To stabilize the candidate parameter set, For a stable threshold, and This formula is used to eliminate candidate process parameters that exhibit delayed deformation exceeding limits, model prediction discrepancies exceeding limits, or inverse responses. S5-3. If the set of stable candidate parameters is not empty, select the group of candidate process parameters with the lowest risk value from the set of stable candidate parameters as the recommended process parameter group, and determine the recommended process parameter group based on the set of stable candidate parameters. This formula is used to select the candidate process parameter with the lowest overall divergence risk value from the stable candidate parameter set; S5-4. For each type of process parameter in the stable candidate parameter set, take the minimum and maximum values ​​to form the stable value range of each process parameter, and determine the stable process parameter window based on the stable candidate parameter set. The minimum and maximum values ​​in the formula are calculated separately for each component of the process parameter vector; this formula is used to output the stable range of values ​​for melt temperature, mold temperature, injection speed, holding pressure, holding time, and cooling time. S5-5. Determine the set of delayed deformation risk units based on the unit divergence risk value. ,in, For the set of late-deformation risk elements; when the set of stable candidate parameters When empty, the output shows the determination result that there is no stable process parameter window and the set of delayed deformation risk elements. When the set of stable candidate parameters When not empty, output stable process parameter window. Recommended process parameter set and the set of delayed deformation risk units The results serve as the determination of process parameters for precision molding of ultra-soft thermoplastic high elastomers.

[0024] Through the above processing, the stable process parameter window output in this embodiment is not simply a range of parameters that meet the requirements of filling, pressure, or demolding dimensions, but rather a range of parameters that has undergone benchmark simulation, virtual perturbation simulation, and dual-model divergence screening. This stable process parameter window can exclude a class of candidate parameters that superficially meet conventional molding requirements but are prone to localized delayed deformation during the demolding release stage. The recommended process parameter group is selected from the remaining stable candidate parameters, choosing the group with the lowest overall risk, ensuring that it not only meets molding quality constraints under the current simulation conditions but also maintains good delayed deformation stability under small process fluctuations. The delayed deformation risk unit is used to indicate locations in the product geometry where post-release deformation is more likely to occur, providing a clear target for subsequent product structure adjustments, process verification, and quality validation.

[0025] The beneficial effect of this embodiment is that it transforms the post-release instability problem, which is difficult to observe directly in the precision molding of ultra-soft thermoplastic elastomers, into a prediction divergence problem between two molding simulation models under the same candidate parameters and their virtual perturbation conditions. Conventional techniques typically rely on a single simulation model to determine whether parameters are qualified, easily retaining parameters that meet the dimensional requirements at demolding but exhibit localized delayed deformation after demolding. This embodiment retains conventional molding prediction results through a first molding simulation model, introduces the influence of near-wall shear history and cooling retention on release deformation through a second molding simulation model, and then eliminates unstable parameters through divergence risk assessment, resulting in a final process parameter window with stronger delayed deformation stability. This process does not require the addition of new molding equipment structures, nor does it rely on repeated trial and error based on human experience, and risk screening can be completed in advance during the digital twin simulation stage.

[0026] 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.

[0027] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A system for determining process parameters of precision molding of ultra-soft thermoplastic high elastomers based on digital twins, characterized in that: include: The molding data processing module is used to receive the geometric data of the product to be molded, the material data of the ultra-soft thermoplastic elastomer, the range of candidate process parameters, and the molding quality constraints, and to process the above data into a dataset that can be called by digital twin simulation. A dual-model simulation module is used to establish a first forming simulation model and a second forming simulation model based on the dataset, and to generate two delayed deformation prediction results for the same candidate process parameters. The process parameter determination module is used to compare the degree of divergence between the two delayed deformation prediction results under virtual perturbation, and determine a stable process parameter window from the candidate process parameters based on the degree of divergence.

2. The system for determining process parameters of ultra-soft thermoplastic high elastomer precision molding based on digital twin as described in claim 1, characterized in that: The molding data processing module includes a geometric unit division submodule, a material data standardization submodule, and a candidate parameter generation submodule. The geometric unit division submodule is used to divide the product to be molded into several feature units. The material data standardization submodule is used to convert the material hardness, viscosity, recovery characteristics, and shrinkage characteristics into a unified data format. The candidate parameter generation submodule is used to generate a set of candidate process parameters within the range of candidate process parameters. The dual-model simulation module includes a first simulation submodule, a second simulation submodule, and a virtual perturbation simulation submodule; The first simulation submodule is used to predict delayed deformation under the computational assumption that the material does not slip relative to the cavity wall. The second simulation submodule is used to predict delayed deformation under the computational assumption that the shear history in the near-wall region is retained after cooling and released after demolding. The virtual perturbation simulation submodule is used to apply small parameter perturbations to the same candidate process parameters and call the first simulation submodule and the second simulation submodule respectively. The process parameter determination module includes a divergence calculation submodule, a risk parameter elimination submodule, and a result output submodule. The divergence calculation submodule is used to calculate the difference in late-deformation predictions between the two models. The risk parameter elimination submodule is used to eliminate candidate process parameters whose divergence exceeds the stability threshold. The result output submodule is used to output a stable process parameter window, a recommended process parameter group, and a late-deformation risk unit.

3. A precision molding method for ultra-soft thermoplastic elastomers based on digital twins, characterized in that: This method is performed using the system described in any one of claims 1 to 2, and includes the following steps: S1. Organize the molding input data, receive the three-dimensional geometric data, material data, candidate process parameter range and molding quality constraints of the product to be molded, divide the product to be molded into several feature units, and mark the thin-walled area, mesh intersection area, parting surface adjacent area and edge free area, establish the mapping relationship between each feature unit and material data and process data, and form a unified input dataset for digital twin simulation. S2. Generate candidate process parameters and virtual perturbation data. Generate multiple sets of candidate process parameters according to the range of candidate process parameters. For each set of candidate process parameters, generate a benchmark parameter set and multiple small perturbation parameter sets. The small perturbation parameter sets are used to observe whether the prediction results of the two models produce opposite trends when the process parameters change slightly. S3. Execute the first forming simulation and the second forming simulation respectively. The first forming simulation calculates the delayed deformation prediction results of each characteristic element under the assumption of wall-attached flow. The second forming simulation calculates the delayed deformation prediction results of each characteristic element under the assumption of near-wall shear cumulative release. Both forming simulations are executed based on the same candidate process parameters and its small perturbation parameter group to ensure that the divergence comparison has the same input basis. S4. Calculate the degree of divergence of delayed deformation between the two models, compare the delayed deformation prediction results obtained from the first forming simulation and the second forming simulation, determine whether the change direction of the same feature unit before and after the small disturbance is consistent, and calculate the normalized difference between the prediction results of the two models. Based on the normalized difference and the consistency of the change direction, obtain the divergence risk value of the candidate process parameters. S5. Determine the stable process parameter window, retain the candidate process parameters whose divergence risk value does not exceed the stable threshold as stable candidate parameters, generate the stable process parameter window based on the stable candidate parameters, determine the recommended process parameter group from the window, and output the stable process parameter window, the recommended process parameter group and the delayed deformation risk unit as the determination result of the precision molding process parameters of the ultra-soft thermoplastic high elastomer.

4. The precision molding method for ultra-soft thermoplastic high elastic body based on digital twin according to claim 3, characterized in that: Step S1 includes the following steps: S1-1. Divide the three-dimensional geometric data of the product to be formed into a set of feature units. ,in, For the first Each feature unit The feature unit number is used as the smallest computational object for delayed deformation prediction, model divergence comparison, and risk unit output, so that different local structures of the same product can participate in subsequent digital twin simulations. S1-2, for each feature unit Establish feature data vectors ,in, For the first Feature data vectors of each feature unit For the first Local thickness of each feature unit For the first The distance from each feature unit to the nearest free edge, To be assigned to the first according to the molding quality constraints The upper limit of allowed delayed deformation for each feature unit; S1-3. Organize the material data of the ultrasoft thermoplastic elastomer into a material data vector. ,in, For material data vectors, This represents the material's hardness value. This is melt viscosity data. For stretch recovery data, This is for cooling and shrinkage data; S1-4, Transfer the feature data vector With the material data vector By performing correlation, a unified input dataset for digital twin simulation is obtained. ,in, To standardize the input dataset, Indicates the first Each feature unit and its corresponding feature data vector.

5. The precision molding method for ultra-soft thermoplastic high elastic body based on digital twin according to claim 4, characterized in that: Step S2 includes the following steps: S2-1. Based on the candidate process parameter range and the preset sampling step size for each type of process parameter, generate the first... Group of candidate process parameter vectors ,in, For the first Group of candidate process parameter vectors, Candidate number, The melting temperature is... For mold temperature, For injection speed, To maintain pressure, For the pressure holding time, Cooling time; S2-2, For each set of candidate process parameter vectors Generate virtual perturbation parameter vector ,in, For the first Group of candidate process parameters in the first The parameter vector after the virtual perturbation. This is a virtual perturbation sequence number. For the first The virtual perturbation vector; this formula is used to observe the impact of small parameter changes on simulation results without changing the actual equipment state. By applying multiple virtual perturbations to the same candidate process parameter, it is possible to determine whether there are hidden instability regions near the candidate process parameter that are highly sensitive to small parameter changes. With candidate process parameter vector Parameter vectors with the same parameter component order and obtained through virtual perturbation It remains within the range of the candidate process parameters; S2-3, Set the reference perturbation vector without applied virtual perturbation to... ,in, This is the reference perturbation vector.

6. The precision molding method for ultra-soft thermoplastic high elastic body based on digital twin according to claim 5, characterized in that: Step S3 includes the following steps: S3-1, For each set of parameter vectors First, the first model is called to calculate the feature units. Based on the conventional wall-adhering flow assumption, the predicted delayed deformation after demolding is calculated using the first molding simulation model for characteristic elements. In parameter vector First delayed deformation prediction value ,in, This is the predicted value of delayed deformation output by the first molding simulation model. For digital twin simulation functions under the assumption of wall-attached flow; S3-2, Calculate the feature units using the second molding simulation model. In parameter vector Near-wall cumulative shear ,in, This represents the cumulative near-wall shear. For simulation time Lower feature unit The near-wall shear rate, For simulation time Lower feature unit The cooling and curing ratio, parameter vector The cooling time component; S3-3, Based on material data vector Stretching recovery data Cooling shrinkage data and feature units Geometric properties to determine feature units Deformation release conversion coefficient ,in This is a deformation release conversion function pre-established based on material tensile recovery data, cooling shrinkage data, and geometric properties of characteristic elements. The first value calculated by the deformation release conversion function The deformation release conversion coefficient of each feature unit is used to convert the near-wall shear accumulation into additional delayed deformation after demolding, based on the near-wall shear accumulation. Calculate the second delayed deformation prediction value ,in, The delayed deformation prediction value output by the second forming simulation model. For feature units The deformation release conversion coefficient.

7. The precision molding method for ultra-soft thermoplastic high elastic body based on digital twin according to claim 6, characterized in that: Step S4 includes the following steps: S4-1. Calculate the changes in the first model based on the baseline parameter vector and the virtual perturbation parameter vector. ,in, For the first formed simulation model in the 1st The amount of delayed deformation relative to the reference input under sub-virtual perturbation; S4-2, the first The difference between the second delayed deformation prediction value under the virtual perturbation and the second delayed deformation prediction value under the reference input is calculated. The change in the second model is then calculated based on the reference parameter vector and the virtual perturbation parameter vector. ,in, For the second molding simulation model in the first The amount of delayed deformation relative to the reference input under sub-virtual perturbation; S4-3. When the product of the first model change and the second model change is less than zero, it indicates that the same virtual perturbation causes the delayed deformation prediction results of the two models to change in opposite directions. This situation is marked as a reverse response, and the reverse response label is calculated based on the first model change and the second model change. ,in, For reverse response markers, This is a conditional function; it takes the value 1 when the condition inside the parentheses is true, and takes the value 0 when the condition inside the parentheses is false. S4-4. Calculate the divergence risk value of the unit based on the first delayed deformation prediction value, the second delayed deformation prediction value, the reverse response flag, and the allowed upper limit of delayed deformation. ,in, For feature units In the Group of candidate process parameters and the first The unit divergence risk value under the sub-virtual perturbation; this formula is used to simultaneously constrain the delayed deformation of the first model, the delayed deformation of the second model, the difference in prediction between the two models, and the consistency of the response direction of the two models.

8. The precision molding method for ultra-soft thermoplastic high elastic body based on digital twin according to claim 7, characterized in that: Step S5 includes the following steps: S5-1. Calculate the risk value of candidate process parameters based on the element divergence risk values ​​under all characteristic elements and all virtual perturbations. ,in, For the first Risk values ​​of candidate process parameters for a group of candidate process parameters; S5-2. Compare the risk values ​​of candidate process parameters with the stability threshold. Candidate process parameters with risk values ​​not exceeding the stability threshold are retained, while those with risk values ​​exceeding the stability threshold are eliminated. Determine the set of stable candidate parameters based on their risk values. ,in, To stabilize the candidate parameter set, For a stable threshold, and ; S5-3. If the set of stable candidate parameters is not empty, select the group of candidate process parameters with the lowest risk value from the set of stable candidate parameters as the recommended process parameter group, and determine the recommended process parameter group based on the set of stable candidate parameters. ; S5-4. For each type of process parameter in the stable candidate parameter set, take the minimum and maximum values ​​to form the stable value range of each process parameter, and determine the stable process parameter window based on the stable candidate parameter set. The minimum and maximum values ​​in the formula are calculated separately for each component of the process parameter vector; S5-5. Determine the set of delayed deformation risk units based on the unit divergence risk value. ,in, The set of delayed deformation risk units; when the set of stable candidate parameters When empty, the output shows the determination result that there is no stable process parameter window and the set of delayed deformation risk units. When the stable candidate parameter set If not empty, output the stable process parameter window. The recommended process parameter set and the set of delayed deformation risk units The results serve as the determination of process parameters for precision molding of ultra-soft thermoplastic high elastomers.

Images