Injection molding process parameter optimization method and system based on hybrid algorithm and model fusion

Abstract

The invention relates to the technical field of artificial intelligence, in particular to an injection molding process parameter optimization method and system based on hybrid algorithm and model fusion, and the method comprises the steps: optimizing a parameter combination of a support vector regression model through a simulated annealing algorithm, building a weighted fusion model based on the optimized support vector regression model and a random forest, and optimizing the model; constructing a hybrid model of an adaptive selection weighted fusion model and an optimized support vector regression model; constructing a three-objective optimization model including buckling deformation, volume shrinkage and production energy consumption, and searching a Pareto optimal solution set in a process parameter space by adopting a multi-objective genetic algorithm by taking the hybrid model as a target value evaluation tool; carrying out local correction on the key process parameters by adopting a gradient descent method until the deviation falls back to be within a preset threshold value, and obtaining optimized process parameters; the defect rate of products can be reduced, and meanwhile production energy consumption is reduced.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, specifically to a method and system for optimizing injection molding process parameters based on hybrid algorithms and model fusion. Background Technology

[0002] In injection molding, product defects and energy consumption control have long been pain points in the industry. Traditional parameter optimization relies on experience or single algorithms, which suffers from weak global search capabilities and difficulty in balancing multiple objectives. With the increasing demand for intelligent manufacturing, building an efficient and precise process optimization system has become crucial. Summary of the Invention

[0003] To address the aforementioned issues, this invention provides a method and system for optimizing injection molding process parameters based on the fusion of hybrid algorithms and models. By integrating multi-dimensional algorithms and models, it optimizes the parameter combination of the support vector regression model through simulated annealing algorithm and achieves multi-objective optimization through multi-objective genetic algorithm, thereby reducing product defect rate and production energy consumption, and realizing multi-dimensional intelligent optimization of the injection molding process.

[0004] To achieve the above objectives, the present invention provides the following technical solution: On one hand, embodiments of the present invention provide a method for optimizing injection molding process parameters based on hybrid algorithm and model fusion, the method comprising the following steps: S100, Obtain a sample dataset, train the support vector regression model using the sample dataset, and optimize the parameter combination of the support vector regression model using the simulated annealing algorithm to obtain an optimized support vector regression model; the sample dataset includes multiple process parameters and corresponding target variables, and the parameter combination includes the penalty factor and kernel parameters of the support vector regression model; S200 is a hybrid model that combines an optimized support vector regression model and a random forest to construct a weighted fusion model. It also incorporates the product structure complexity and the fluctuation range of process parameters to construct an adaptive selection weighted fusion model and an optimized support vector regression model. S300 constructs a three-objective optimization model that includes warping deformation, volume shrinkage rate, and production energy consumption. A hybrid model is used as the objective value evaluation tool, and a multi-objective genetic algorithm is used to search for the Pareto optimal solution set in the process parameter space. The Pareto optimal solution set contains multiple sets of process parameters, and each set of process parameters corresponds to a set of non-dominated predicted objective values. S400 collects the actual target values ​​during the production process and compares them with the corresponding predicted target values. When the deviation exceeds the preset threshold, the gradient descent method is used to locally correct the key process parameters until the deviation falls back to within the preset threshold, thus obtaining the optimized process parameters.

[0005] Optionally, in step S100, obtaining a sample dataset, training a support vector regression model using the sample dataset, and optimizing the parameter combination of the support vector regression model using a simulated annealing algorithm to obtain an optimized support vector regression model includes: S110, Initialize the search range of the penalty factor and kernel parameter, and set the initial parameter combination; the parameter combination includes the penalty factor and kernel parameter, the penalty factor is used to control the tolerance of the support vector regression model to the mean square error, and the kernel parameter is used to map the nonlinear relationship between process parameters and target variables; S120: Obtain a sample dataset. Using the mean squared error of the support vector regression model's prediction on the sample dataset as the optimization objective function, the temperature is successively reduced at a set cooling rate. At each temperature, a new parameter combination is generated through random perturbation. If the mean squared error of the new parameter combination is lower, the new parameter combination is accepted. If the error is higher, the new parameter combination is accepted according to the Metropolis criterion with probability. The iteration stops when the termination condition is reached, and the current optimal parameter combination is output. This is then substituted into the support vector regression model to obtain the optimized support vector regression model. The termination condition is that the objective function value does not improve within a preset number of consecutive iterations, the preset number of iterations is reached, or the temperature drops to a termination threshold.

[0006] Optionally, in step S200, constructing a weighted fusion model based on an optimized support vector regression model and a random forest includes: Based on the optimized support vector regression model, the interaction between the process parameters of random forest is introduced to dynamically fuse the weight coefficients of the optimized support vector regression model and random forest, and construct a weighted fusion model. When the prediction error of the optimized support vector regression model is greater than that of the random forest, the weight ratio of the random forest is increased, and vice versa.

[0007] Optionally, in step S200, a hybrid model is constructed by combining the product structure complexity and the fluctuation range of process parameters, which includes an adaptive selection weighted fusion model and an optimized support vector regression model, comprising: When the complexity of the product structure exceeds the complexity threshold or the fluctuation range of the process parameters exceeds the fluctuation threshold, the sample dataset is clustered according to the product complexity and parameter range, and each cluster is trained separately. The support vector regression sub-model obtained by the corresponding cluster training is then called. When the complexity of the product structure does not exceed the complexity threshold and the fluctuation range of the process parameters does not exceed the fluctuation threshold, the weighted fusion model is invoked.

[0008] Optionally, in step S300, a three-objective optimization model is constructed, including warpage deformation, volume shrinkage rate, and production energy consumption. A hybrid model is used as the objective value evaluation tool, and a multi-objective genetic algorithm is employed to search for the Pareto optimal solution set in the process parameter space, including: S310 uses the objective function of minimizing warpage, volume shrinkage, and production energy consumption as a three-objective optimization model, introducing mold temperature constraints, holding pressure-injection time ratio constraints, and injection speed constraints. Among them, the warpage is obtained by fitting the relationship between process parameters and measured warpage values ​​through an optimized support vector regression model; the volume shrinkage rate is predicted based on the correlation between melt density change and process parameters, combined with the optimized support vector regression model; and the production energy consumption is calculated by combining equipment power and running time through an optimized support vector regression model. The mold temperature constraint is limited between the lower limit of material flowability and the upper limit of mold heat resistance; the holding pressure-injection time ratio constraint is set between the lower limit of avoiding insufficient holding pressure and the upper limit of avoiding excessive holding pressure leading to energy waste; and the injection speed constraint is set between the lower limit of preventing slow filling and cold material and the upper limit of preventing flash or bubbles. S320, Randomly generate an initial population of process parameters, wherein the individuals in the population of process parameters are combinations of parameters; S330 uses a multi-objective genetic algorithm to perform selection, crossover, and mutation operations on the initial process parameter population. It uses a hybrid model to calculate the target value corresponding to each individual, selects individuals with better target values ​​based on non-dominated sorting, and combines a reference point mechanism to maintain a uniform distribution of the solution set in the target space. After multiple generations of iteration, it outputs a Pareto optimal solution set containing multiple sets of non-dominated target values.

[0009] Optionally, in step S400, the gradient descent method is used to locally correct key process parameters until the deviation falls back to within a preset threshold, resulting in optimized process parameters, including: S410 uses minimizing the deviation between the actual target value and the predicted target value as the objective function to determine the key process parameters related to the production process; S420, the key process parameters are adjusted in small steps. After each round of adjustment, the correction effect is quickly evaluated by the fusion prediction model, and the objective function value corresponding to the current key process parameter is calculated. The step size range of the small step adjustment is preset according to the process parameter type. S430: When the objective function value decreases compared to the previous round, continue to adjust the key process parameters in the current direction; when the objective function value increases, adjust the key process parameters in the opposite direction or reduce the adjustment step size; until the deviation between the actual target value and the predicted target value falls back to within the preset threshold, the key process parameters output at this time are used as the optimized process parameters.

[0010] Optionally, the method further includes: Using a preset production batch as a sliding window, the actual process parameters and actual target values ​​within the window are periodically used as new training samples to retrain and optimize the support vector regression model and the weighted fusion model.

[0011] On the other hand, embodiments of the present invention provide an injection molding process parameter optimization system based on hybrid algorithm and model fusion, including: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method described above.

[0012] On the other hand, embodiments of the present invention provide a computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to perform the above-described method.

[0013] The beneficial effects of this invention are as follows: This invention discloses a method and system for optimizing injection molding process parameters based on hybrid algorithms and model fusion. This invention optimizes the parameter combination of a support vector regression model using simulated annealing, effectively improving the model's ability to fit complex nonlinear relationships; it combines random forest to construct a weighted fusion model, dynamically adapting to the prediction needs of different product structures, significantly enhancing the model's generalization performance; it employs a multi-objective genetic algorithm to achieve multi-objective collaborative optimization of process parameters, achieving a better balance between warpage, volume shrinkage rate, and production energy consumption; and it further improves the accuracy of process parameters by locally correcting key parameters using gradient descent. Experimental results show that, compared with traditional optimization methods, this method can reduce product defect rates while reducing production energy consumption, providing an effective solution for the intelligent upgrading of injection molding processes. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in 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.

[0015] Figure 1 This is a flowchart illustrating an injection molding process parameter optimization method based on hybrid algorithm and model fusion according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of an injection molding process parameter optimization system based on hybrid algorithm and model fusion in an embodiment of the present invention. Detailed Implementation

[0016] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0017] In injection molding, product defects and energy consumption control have long been pain points in the industry. Traditional parameter optimization relies on experience or a single algorithm, which suffers from weak global search capabilities and difficulty in balancing multiple objectives. With the increasing demand for intelligent manufacturing, building an efficient and precise process optimization system has become crucial. This invention breaks through the limitations of existing research by integrating multiple algorithms and models. It optimizes SVR parameters through simulated annealing and achieves multi-objective optimization through NSGA-III, forming an SA-SVR-NSGA-III hybrid architecture to realize multi-dimensional intelligent optimization of the injection molding process.

[0018] To address the shortcomings of existing technologies, this invention proposes an innovative multimodal knowledge base construction method for analyzing physical, chemical, and biological experiments. To overcome the limitations of traditional injection molding process parameter optimization methods, it proposes an intelligent optimization method for injection molding process parameters that integrates improved algorithms and multi-model collaboration. The simulated annealing algorithm is used to optimize support vector regression (SVR) parameters, constructing an SA-SVR-NSGA-III hybrid framework to achieve synergistic optimization of warpage, volume shrinkage rate, and production energy consumption. An adaptive fusion of multiple models and an online feedback mechanism are introduced to improve prediction accuracy and dynamic optimization capabilities in complex scenarios. Experimental results show that compared to traditional methods, this method significantly improves optimization efficiency, multi-objective balance, and industrial adaptability, providing a new path for intelligent optimization of injection molding processes.

[0019] refer to Figure 1 ,like Figure 1 The figure shows an embodiment of the present invention providing a method for optimizing injection molding process parameters based on hybrid algorithm and model fusion. The method includes the following steps: S100, Obtain a sample dataset, train the support vector regression model using the sample dataset, and optimize the parameter combination of the support vector regression model using the simulated annealing algorithm to obtain an optimized support vector regression model; the sample dataset includes multiple process parameters and corresponding target variables, and the parameter combination includes the penalty factor and kernel parameters of the support vector regression model; S200 is a hybrid model that combines an optimized support vector regression model and a random forest to construct a weighted fusion model. It also incorporates the product structure complexity and the fluctuation range of process parameters to construct an adaptive selection weighted fusion model and an optimized support vector regression model. S300 constructs a three-objective optimization model that includes warping deformation, volume shrinkage rate, and production energy consumption. A hybrid model is used as the objective value evaluation tool, and a multi-objective genetic algorithm is used to search for the Pareto optimal solution set in the process parameter space. The Pareto optimal solution set contains multiple sets of process parameters, and each set of process parameters corresponds to a set of non-dominated predicted objective values. S400 collects the actual target values ​​during the production process and compares them with the corresponding predicted target values. When the deviation exceeds the preset threshold, the gradient descent method is used to locally correct the key process parameters until the deviation falls back to within the preset threshold, thus obtaining the optimized process parameters.

[0020] In this embodiment, the optimization process of the support vector regression model parameters is simulated using the annealing algorithm, enabling effective exploration of initial parameter combinations. Specifically, the penalty factor and kernel parameter search range set during the initialization phase, combined with the set cooling rate and random perturbation mechanism, ensure global coverage of the parameter space. During the iteration process, the introduction of the Metropolis criterion ensures that the algorithm maintains probabilistic rationality when accepting inferior solutions, avoiding getting trapped in local optima.

[0021] During the construction of the weighted fusion model, the dynamic weight adjustment mechanism achieves the complementary advantages of optimized support vector regression and random forest. When the fluctuation range of process parameters exceeds the preset threshold, the system automatically switches to cluster training mode, generating a sub-model library for specific working conditions through cluster analysis of product complexity and parameter range.

[0022] In constructing the three-objective optimization model, warpage deformation, volume shrinkage rate, and production energy consumption are transformed into quantifiable objective functions. Combined with the parameter mapping relationship established based on the physical mechanism of injection molding, a complete optimization system is formed. The non-dominated sorting and reference point mechanism employed in the multi-objective genetic algorithm ensures the uniformity and diversity of the Pareto front solution set.

[0023] In some embodiments, in step S100, obtaining a sample dataset, training a support vector regression model using the sample dataset, and optimizing the parameter combination of the support vector regression model using a simulated annealing algorithm to obtain an optimized support vector regression model includes: S110, Initialize the search range of the penalty factor and kernel parameter, and set the initial parameter combination; the parameter combination includes the penalty factor and kernel parameter, the penalty factor is used to control the tolerance of the support vector regression model to the mean square error, and the kernel parameter is used to map the nonlinear relationship between process parameters and target variables; S120: Obtain a sample dataset. Using the mean squared error of the support vector regression model's prediction on the sample dataset as the optimization objective function, the temperature is successively reduced at a set cooling rate. At each temperature, a new parameter combination is generated through random perturbation. If the mean squared error of the new parameter combination is lower, the new parameter combination is accepted. If the error is higher, the new parameter combination is accepted according to the Metropolis criterion with probability. The iteration stops when the termination condition is reached, and the current optimal parameter combination is output. This is then substituted into the support vector regression model to obtain the optimized support vector regression model. The termination condition is that the objective function value does not improve within a preset number of consecutive iterations, the preset number of iterations is reached, or the temperature drops to a termination threshold.

[0024] In this embodiment, by refining the search for the penalty factor and kernel parameters, combined with the random perturbation characteristics of the simulated annealing algorithm, the drawback of traditional parameter optimization methods being prone to getting trapped in local optima is effectively overcome. Specifically, the parameter search range set during the initialization phase covers common numerical ranges in engineering practice, and the cooling rate parameter is verified through pre-experiments to ensure a balance between algorithm convergence and computational efficiency. During the temperature decrease process, the random perturbation mechanism uses a Gaussian distribution to generate candidate parameter combinations, ensuring both the breadth of parameter space exploration and maintaining the asymptotic stability of the algorithm through the probability acceptance mechanism of the Metropolis criterion. When the objective function value does not improve within a preset number of iterations, the system automatically triggers the termination condition to avoid unnecessary computational resource consumption; when the temperature drops to the termination threshold, the current optimal parameter combination is output. In particular, for the common nonlinear mapping problem in injection molding processes, the kernel parameters achieve feature space transformation through radial basis functions, and with the adaptive adjustment of the penalty factor, the model shows significant advantages in predicting the warpage of complex curved surface products.

[0025] In some embodiments, the step S200, which involves constructing a weighted fusion model based on an optimized support vector regression model and a random forest, includes: Based on the optimized support vector regression model, the interaction between the process parameters of random forest is introduced to dynamically fuse the weight coefficients of the optimized support vector regression model and random forest, and construct a weighted fusion model. When the prediction error of the optimized support vector regression model is greater than that of the random forest, the weight ratio of the random forest is increased, and vice versa.

[0026] In this embodiment, a dynamic weight adjustment mechanism is used to achieve complementary advantages between the optimized support vector regression model and the random forest. Specifically, the system monitors the prediction errors of the two models in real time. When the optimized support vector regression model shows prediction deviation under specific operating conditions, the weight allocation module automatically increases the contribution of the random forest; conversely, when the random forest's ability to analyze complex interactions is insufficient, the weight of the optimized support vector regression model is increased. This effectively solves the performance degradation problem of a single model under local operating conditions. In the model construction phase, the initial weight benchmark value is determined through cross-validation technology. Combined with real-time monitoring of the fluctuation range of process parameters, when the standard deviation of the parameters exceeds the threshold, the system triggers a cluster training mechanism, dividing the dataset into multiple subsets according to product complexity, and training sub-model libraries for specific operating conditions. In the prediction phase, the model selector automatically matches the optimal sub-model based on the product characteristics of the current production batch, ensuring a balance between prediction accuracy and computational efficiency. This fusion model can improve prediction accuracy compared to a single model when dealing with high-dimensional nonlinear data. In particular, in the warpage prediction of thin-walled products, the dynamic tracking of the melt flow front by random forest effectively compensates for the prediction delay of the optimized support vector regression model under time-varying conditions.

[0027] In some embodiments, in step S200, a hybrid model combining an adaptive selection weighted fusion model and an optimized support vector regression model is constructed, taking into account the product structure complexity and the fluctuation range of process parameters. This includes: When the complexity of the product structure exceeds the complexity threshold or the fluctuation range of the process parameters exceeds the fluctuation threshold, the sample dataset is clustered according to the product complexity and parameter range, and each cluster is trained separately. The support vector regression sub-model obtained by the corresponding cluster training is then called. When the complexity of the product structure does not exceed the complexity threshold and the fluctuation range of the process parameters does not exceed the fluctuation threshold, the weighted fusion model is invoked.

[0028] In this embodiment, a dual judgment mechanism based on product structure complexity and process parameter fluctuation range is introduced to achieve dynamic adaptation of the model selection strategy.

[0029] In some embodiments, in S300, a three-objective optimization model is constructed, including warpage deformation, volume shrinkage rate, and production energy consumption. A hybrid model is used as the objective value evaluation tool, and a multi-objective genetic algorithm is employed to search for a Pareto optimal solution set in the process parameter space, including: S310 uses the objective function of minimizing warpage, volume shrinkage, and production energy consumption as a three-objective optimization model, introducing mold temperature constraints, holding pressure-injection time ratio constraints, and injection speed constraints. Among them, the warpage is obtained by fitting the relationship between process parameters and measured warpage values ​​through an optimized support vector regression model; the volume shrinkage rate is predicted based on the correlation between melt density change and process parameters, combined with the optimized support vector regression model; and the production energy consumption is calculated by combining equipment power and running time through an optimized support vector regression model. The mold temperature constraint is limited between the lower limit of material flowability and the upper limit of mold heat resistance; the holding pressure-injection time ratio constraint is set between the lower limit of avoiding insufficient holding pressure and the upper limit of avoiding excessive holding pressure leading to energy waste; and the injection speed constraint is set between the lower limit of preventing slow filling and cold material and the upper limit of preventing flash or bubbles. S320, Randomly generate an initial population of process parameters, wherein the individuals in the population of process parameters are combinations of parameters; S330 uses a multi-objective genetic algorithm to perform selection, crossover, and mutation operations on the initial process parameter population. It uses a hybrid model to calculate the target value corresponding to each individual, selects individuals with better target values ​​based on non-dominated sorting, and combines a reference point mechanism to maintain a uniform distribution of the solution set in the target space. After multiple generations of iteration, it outputs a Pareto optimal solution set containing multiple sets of non-dominated target values.

[0030] In this embodiment, a complete three-objective optimization system was constructed to achieve coordinated control of warpage deformation, volume shrinkage rate, and production energy consumption. The objective function not only considered product quality indicators (warpage deformation and volume shrinkage rate) but also incorporated economic indicators of production energy consumption, forming a multi-dimensional optimization framework. The constraints were set to balance process feasibility (mold temperature range, holding pressure injection time ratio, and injection speed range) with equipment protection requirements, ensuring that the optimization results can be directly applied to actual production. The multi-objective genetic algorithm, through non-dominated sorting and reference point mechanisms, effectively addresses the shortcomings of traditional methods in terms of Pareto front solution set uniformity, especially when dealing with complex constraints, maintaining the diversity and representativeness of the solution set.

[0031] In some embodiments, in step S400, the gradient descent method is used to locally correct key process parameters until the deviation falls back to within a preset threshold, thereby obtaining optimized process parameters, including: S410 uses minimizing the deviation between the actual target value and the predicted target value as the objective function to determine the key process parameters related to the production process; S420, the key process parameters are adjusted in small steps. After each round of adjustment, the correction effect is quickly evaluated by the fusion prediction model, and the objective function value corresponding to the current key process parameter is calculated. The step size range of the small step adjustment is preset according to the process parameter type. S430: When the objective function value decreases compared to the previous round, continue to adjust the key process parameters in the current direction; when the objective function value increases, adjust the key process parameters in the opposite direction or reduce the adjustment step size; until the deviation between the actual target value and the predicted target value falls back to within the preset threshold, the key process parameters output at this time are used as the optimized process parameters.

[0032] In this embodiment, the local correction mechanism implemented by the gradient descent method effectively improves the dynamic adaptability of the production process.

[0033] In some embodiments, the method further includes: Using a preset production batch as a sliding window, the actual process parameters and actual target values ​​within the window are periodically used as new training samples to retrain and optimize the support vector regression model and the weighted fusion model.

[0034] In this embodiment, the sliding window mechanism enables the model to track the dynamic changes of production data in real time, and the incremental learning method ensures the synchronous updating of model parameters and process conditions. The prediction error of the model using the sliding window update is lower than that of the static model in continuous production, effectively avoiding performance degradation caused by data distribution offset.

[0035] The following is a specific embodiment provided by the present invention: (I) Algorithm architecture innovation: SA-SVR-NSGA-III hybrid framework Simulated Annealing (SA) optimizes the parameters of a Support Vector Regression (SVR) model. Simulated annealing is used to replace the traditional particle swarm optimization (PSO) algorithm to optimize the core parameters of the support vector regression model, namely the penalty factor C and the kernel parameters. In injection molding process parameter optimization scenarios, traditional optimization algorithms are prone to getting trapped in local optima, making it difficult to accurately find the optimal solution in a complex, high-dimensional process parameter space. Simulated annealing, inspired by the physical annealing process, introduces a probabilistic jump mechanism, allowing for the acceptance of suboptimal solutions with a certain probability during the search process, thus effectively escaping the local optimum trap.

[0036] The core principle of the simulated annealing algorithm is based on the Metropolis criterion, and its probability formula for accepting a new solution is: ; in, Let T be the difference between the objective function of the new solution and the current solution, and let T be the current temperature. As the annealing process progresses, the temperature T gradually decreases, and the probability of accepting a suboptimal solution also decreases, eventually converging to the global optimum.

[0037] In SVR parameter optimization, the penalty factor C is compared with the kernel parameter. As variables in the solution space, the objective function is set to minimize the mean squared error (MSE) of the support vector regression model: ; in, For the sample size, This is the actual value. These are the predicted values ​​from the support vector regression model. The simulated annealing algorithm continuously adjusts the penalty factor C and the kernel parameters. The Metropolis criterion is used to determine whether to accept new parameter combinations, and the performance of the support vector regression model is gradually optimized.

[0038] In the high-dimensional process parameter space of this invention, in addition to conventional parameters such as mold temperature, melt temperature, and holding pressure, key variables such as injection speed and cooling time are added, further increasing the complexity and difficulty of parameter optimization. The simulated annealing algorithm, with its unique probabilistic jump characteristics, can perform a global search in this complex space, adjusting the penalty factor C and kernel parameters of the support vector regression model. More precise tuning is then performed. The penalty factor C controls the tolerance of the support vector regression model to errors; the larger the C value, the more strictly the support vector regression model penalizes errors. Kernel parameters... This determines the complexity and generalization ability of the support vector regression model, and the appropriate kernel parameters... Values ​​enable support vector regression models to better fit the nonlinear relationship between process parameters and target variables (warpage, energy consumption, etc.).

[0039] Extensive comparative analysis of experiments revealed a significant performance improvement in the support vector regression model optimized using the simulated annealing algorithm. Specifically, its mean squared error (MSE) convergence speed was improved by 20% compared to traditional optimization methods, and the final MSE value was lower. This means that the SVector Regression model optimized by the simulated annealing algorithm can reach a stable state in a shorter time, reducing the time cost required for model training. It also provides a more stable and reliable foundation for the accurate prediction of subsequent injection molding process parameters. This SVector Regression model can quickly calculate warpage, volume shrinkage rate, and production energy consumption corresponding to any process parameter, providing efficient target value evaluation support for NSGA-III multi-objective optimization and significantly improving the efficiency of the overall optimization process.

[0040] NSGA-III Multi-Objective Collaborative Optimization: A three-objective optimization model was constructed to evaluate warpage deformation, volume shrinkage rate, and production energy consumption. A support vector regression model optimized using simulated annealing was used as the objective value evaluation tool, and a multi-objective genetic algorithm (NSGA-III) was employed to output the Pareto optimal solution set. By introducing process feasibility constraints (such as upper limits for mold temperature and the range of holding pressure-injection time ratio), the optimization results were ensured to closely align with industrial realities, achieving a deep fit between the technical solution and production conditions.

[0041] In injection molding, warpage, shrinkage, and production energy consumption are interrelated and mutually restrictive key indicators. For example, reducing energy consumption may lead to insufficient melt filling, thereby increasing warpage; while excessively increasing holding pressure can reduce shrinkage, it increases energy consumption and mold wear. To achieve synergistic optimization of these three objectives, we first establish a system that includes mold temperature based on the physical mechanism of injection molding and numerical simulation results. , melt temperature Injection pressure Holding pressure Injection time Holding time Injection speed Cooling time The mapping relationship between process parameters and target variables is efficiently calculated using a support vector regression model optimized by the simulated annealing algorithm.

[0042] The mathematical expressions for each target variable are as follows: Warpage The relationship between process parameters and measured warpage values ​​is fitted using a support vector regression model, expressed as follows: ,in This is a random error term, with a value range of less than 0.05 mm.

[0043] Volume shrinkage rate Based on the correlation between melt density variation and process parameters, combined with support vector regression model prediction, the formula is as follows: , The initial volume of the melt. An SVR sub-model for predicting the volume of the molded product.

[0044] Production energy consumption The combined power and operating time of the equipment are calculated using a support vector regression model, expressed as follows: , The fixed energy consumption of auxiliary equipment (such as cooling systems and raw material conveying systems) is determined to be constant through experiments.

[0045] The optimization objective function of NSGA-III is set to minimize three objective variables, namely: ; To avoid optimization results deviating from actual production conditions, multiple process feasibility constraints are introduced: Mold temperature constraints: in (e.g., 40℃) is the lower limit of material fluidity. (e.g., 90℃) is the upper limit of the mold's heat resistance to prevent mold deformation or material degradation.

[0046] Pressure holding time to injection time ratio constraint: (To avoid insufficient pressure holding) (Avoid excessive pressure holding time, which leads to energy waste).

[0047] Injection speed constraint: (e.g., 50mm / s) to prevent slow filling and the generation of cold material. (e.g., 300mm / s) to prevent flash or bubbles.

[0048] NSGA-III efficiently searches the process parameter space by simulating the selection, crossover, and mutation operations of biological evolution: First, an initial population of process parameters is randomly generated, and the three target values ​​for each individual are quickly calculated using a support vector regression model optimized by simulated annealing algorithm; then, individuals with better target values ​​are selected based on "non-dominated sorting", and a "reference point mechanism" is combined to ensure that the solution set is evenly distributed in the target space, avoiding the omission of key trade-offs; after multiple iterations, the Pareto optimal solution set is finally output; this Pareto optimal solution set contains multiple sets of process parameters, each set of process parameters corresponding to a set of non-dominated target values ​​(such as "warpage 3.1mm + shrinkage rate 1.7% + energy consumption 122kW·h" and "warpage 3.4mm + shrinkage rate 1.5% + energy consumption 108kW·h"), which can be flexibly selected according to production priorities (such as "energy saving priority" and "quality priority").

[0049] (II) Upgrading the prediction model based on a multi-model adaptive fusion mechanism: SVR-RF weighted fusion model: A weighted fusion model of Support Vector Regression (SVR) and Random Forest (RF) is constructed: Based on the SVR model optimized by simulated annealing algorithm, its strong fitting ability to nonlinear relationships of process parameters is retained; the interaction between analytical parameters of Random Forest (such as "the synergistic effect of melt temperature and cooling time on shrinkage") is introduced; the outputs of the two are fused through dynamic weight coefficients (based on real-time prediction error adjustment); when the prediction error of the SVR model is greater than that of the Random Forest, the weight ratio of the Random Forest is automatically increased, and vice versa, the weight of the SVR model is increased, ensuring that the weighted fusion model maintains high prediction accuracy in different parameter ranges.

[0050] Experimental results show that the prediction mean squared error of the fusion model is reduced to 0.068 (compared to 0.0827 for the traditional single support vector regression model), improving accuracy by 17.8%. Specifically, in the range of large melt temperature fluctuations (e.g., 200-260℃), the analytical ability of random forest to analyze parameter interactions effectively compensates for the bias of the support vector regression model; while in the range of stable mold temperature (e.g., 50-70℃), the nonlinear fitting advantage of the support vector regression model is more obvious, and the dynamic weighting mechanism can adaptively balance the contributions of both, avoiding the prediction failure of a single model in complex scenarios.

[0051] Scene adaptive selection strategy: Based on the distribution characteristics of the input parameters (such as melt temperature range and product complexity clustering), an automatic model switching mechanism is designed to further improve prediction robustness. The specific rules are as follows: Complexity scenarios: When the product structure is complex (e.g., the hole density of automotive mesh parts > 5 holes / cm²) or the process parameters fluctuate greatly (e.g., the injection speed variation > 100mm / s), the SVR sub-model of "clustering and grouping training" is triggered. By clustering historical samples according to "product complexity + parameter range", an SVR sub-model is trained separately for each cluster (e.g., "high-density hole + high injection speed" sub-model, "low-density hole + low injection speed" sub-model), avoiding cross-interference of samples in complex scenarios.

[0052] Typical scenarios: When the product structure is simple (such as flat parts, hole density <2 holes / cm²) and the process parameters fluctuate little (variation range of each parameter <50%), the weighted fusion model of SVR-RF is automatically called to reduce the computational cost while ensuring accuracy.

[0053] (III) Dynamic optimization closed loop: online feedback and drift compensation Real-time feedback correction module: The system integrates industrial sensors (such as laser displacement sensors to detect warpage and power sensors to detect energy consumption) to collect actual target values ​​during the production process in real time and compares them with the predicted values ​​optimized by SA-SVR-NSGA-III. When the deviation exceeds 5% (e.g., predicted warpage 3.2mm, actual warpage 3.4mm), the gradient descent method is automatically activated to locally correct key process parameters (e.g., holding pressure, cooling time). With the goal of minimizing the deviation between actual and predicted values, the key process parameters are adjusted in small steps (e.g., holding pressure ±2MPa, cooling time ±0.5s), and the correction effect is quickly evaluated through a fusion prediction model until the deviation falls back to within 5%.

[0054] In actual production testing, this mechanism improves the dynamic adaptability of process parameters by 30%. For example, when the aging of the equipment heating system causes the actual melt temperature to be 5°C lower than the set value, the feedback module can adjust the holding time from 5s to 5.3s within 2 production batches to offset the increased warpage caused by temperature deviation and effectively cope with equipment fluctuations and environmental interference (such as changes in workshop temperature and differences in raw material batches).

[0055] Parameter drift compensation algorithm: To address parameter drift caused by long-term equipment operation (such as mold wear leading to a decrease in actual clamping force, and aging of heating tubes leading to a decrease in heating efficiency), a sliding window sample update strategy is designed: using 100 production batches as a sliding window, old samples outside the window (such as equipment status data from 3 months ago) are automatically removed, and the "actual process parameters + actual target values" in the window are added as training samples. The SA-SVR model and the SVR-RF fusion model are retrained periodically (every 100 batches).

[0056] This algorithm ensures that the model always makes predictions based on the latest equipment status, avoiding optimization failures caused by equipment drift. It achieves a 15 percentage point improvement over traditional fixed-sample models (which reduce accuracy to 75%), significantly enhancing the long-term industrial adaptability of the optimization method.

[0057] refer to Figure 2 This invention also provides an injection molding process parameter optimization system based on hybrid algorithm and model fusion, comprising: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method described above.

[0058] The content of the above method embodiments is applicable to this embodiment. The specific functions implemented in this embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments. Therefore, they will not be repeated here.

[0059] Furthermore, one embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions for performing the above-described method.

[0060] It is worth noting that, since the computer-readable storage medium of the present invention is capable of executing the methods of any of the above embodiments, the specific implementation methods and technical effects of the computer-readable storage medium of the present invention can be referred to the specific implementation methods and technical effects of the methods of any of the above embodiments.

[0061] Although the description of this disclosure has been quite detailed and particularly focused on several of the described embodiments, it is not intended to limit itself to any of these details or embodiments or any particular embodiment, but should be considered as effectively covering the intended scope of this disclosure by referring to the appended claims and taking into account the broad possible interpretations of these claims provided by the prior art. Furthermore, the foregoing description of this disclosure with respect to embodiments foreseeable by the inventors is intended to provide a useful description, and non-substantial modifications to this disclosure that have not yet been foreseen may still represent equivalent modifications.

Claims

1. A method for optimizing injection molding process parameters based on hybrid algorithm and model fusion, characterized in that, The method includes the following steps: S100, Obtain a sample dataset, train the support vector regression model using the sample dataset, and optimize the parameter combination of the support vector regression model using the simulated annealing algorithm to obtain an optimized support vector regression model; the sample dataset includes multiple process parameters and corresponding target variables, and the parameter combination includes the penalty factor and kernel parameters of the support vector regression model; S200 is a hybrid model that combines an optimized support vector regression model and a random forest to construct a weighted fusion model. It also incorporates the product structure complexity and the fluctuation range of process parameters to construct an adaptive selection weighted fusion model and an optimized support vector regression model. S300 constructs a three-objective optimization model that includes warping deformation, volume shrinkage rate, and production energy consumption. A hybrid model is used as the objective value evaluation tool, and a multi-objective genetic algorithm is used to search for the Pareto optimal solution set in the process parameter space. The Pareto optimal solution set contains multiple sets of process parameters, and each set of process parameters corresponds to a set of non-dominated predicted objective values. S400 collects the actual target values ​​during the production process and compares them with the corresponding predicted target values. When the deviation exceeds the preset threshold, the gradient descent method is used to locally correct the key process parameters until the deviation falls back to within the preset threshold, thus obtaining the optimized process parameters.

2. The method according to claim 1, characterized in that, In step S100, a sample dataset is acquired, a support vector regression model is trained using the sample dataset, and the parameter combination of the support vector regression model is optimized using a simulated annealing algorithm to obtain an optimized support vector regression model, including: S110, Initialize the search range of the penalty factor and kernel parameter, and set the initial parameter combination; the parameter combination includes the penalty factor and kernel parameter, the penalty factor is used to control the tolerance of the support vector regression model to the mean square error, and the kernel parameter is used to map the nonlinear relationship between process parameters and target variables; S120: Obtain a sample dataset. Using the mean squared error of the support vector regression model's prediction on the sample dataset as the optimization objective function, the temperature is successively reduced at a set cooling rate. At each temperature, a new parameter combination is generated through random perturbation. If the mean squared error of the new parameter combination is lower, the new parameter combination is accepted. If the error is higher, the new parameter combination is accepted according to the Metropolis criterion with probability. The iteration stops when the termination condition is reached, and the current optimal parameter combination is output. This is then substituted into the support vector regression model to obtain the optimized support vector regression model. The termination condition is that the objective function value does not improve within a preset number of consecutive iterations, the preset number of iterations is reached, or the temperature drops to a termination threshold.

3. The method according to claim 1, characterized in that, In S200, a weighted fusion model is constructed based on an optimized support vector regression model and a random forest, including: Based on the optimized support vector regression model, the interaction between the process parameters of random forest is introduced to dynamically fuse the weight coefficients of the optimized support vector regression model and random forest, and construct a weighted fusion model. When the prediction error of the optimized support vector regression model is greater than that of the random forest, the weight ratio of the random forest is increased, and vice versa.

4. The method according to claim 1, characterized in that, In step S200, a hybrid model combining an adaptive selection weighted fusion model and an optimized support vector regression model is constructed, taking into account the product structure complexity and the fluctuation range of process parameters. This model includes: When the complexity of the product structure exceeds the complexity threshold or the fluctuation range of the process parameters exceeds the fluctuation threshold, the sample dataset is clustered according to the product complexity and parameter range, and each cluster is trained separately. The support vector regression sub-model obtained by the corresponding cluster training is then called. When the complexity of the product structure does not exceed the complexity threshold and the fluctuation range of the process parameters does not exceed the fluctuation threshold, the weighted fusion model is invoked.

5. The method according to claim 1, characterized in that, In S300, a three-objective optimization model is constructed, including warpage deformation, volume shrinkage rate, and production energy consumption. A hybrid model is used as the objective value evaluation tool, and a multi-objective genetic algorithm is employed to search for a Pareto optimal solution set in the process parameter space, including: S310 uses the objective function of minimizing warpage, volume shrinkage, and production energy consumption as a three-objective optimization model, introducing mold temperature constraints, holding pressure-injection time ratio constraints, and injection speed constraints. Among them, the warpage is obtained by fitting the relationship between process parameters and measured warpage values ​​through an optimized support vector regression model; the volume shrinkage rate is predicted based on the correlation between melt density change and process parameters, combined with the optimized support vector regression model; and the production energy consumption is calculated by combining equipment power and running time through an optimized support vector regression model. The mold temperature constraint is limited between the lower limit of material flowability and the upper limit of mold heat resistance; the holding pressure-injection time ratio constraint is set between the lower limit of avoiding insufficient holding pressure and the upper limit of avoiding excessive holding pressure leading to energy waste; and the injection speed constraint is set between the lower limit of preventing slow filling and cold material and the upper limit of preventing flash or bubbles. S320, Randomly generate an initial population of process parameters, wherein the individuals in the population of process parameters are combinations of parameters; S330 uses a multi-objective genetic algorithm to perform selection, crossover, and mutation operations on the initial process parameter population. It uses a hybrid model to calculate the target value corresponding to each individual, selects individuals with better target values ​​based on non-dominated sorting, and combines a reference point mechanism to maintain a uniform distribution of the solution set in the target space. After multiple generations of iteration, it outputs a Pareto optimal solution set containing multiple sets of non-dominated target values.

6. The method according to claim 1, characterized in that, In step S400, the gradient descent method is used to locally correct key process parameters until the deviation falls back to within a preset threshold, resulting in optimized process parameters, including: S410 uses minimizing the deviation between the actual target value and the predicted target value as the objective function to determine the key process parameters related to the production process; S420, the key process parameters are adjusted in small steps. After each round of adjustment, the correction effect is quickly evaluated by the fusion prediction model, and the objective function value corresponding to the current key process parameter is calculated. The step size range of the small step adjustment is preset according to the process parameter type. S430: When the objective function value decreases compared to the previous round, continue to adjust the key process parameters in the current direction; when the objective function value increases, adjust the key process parameters in the opposite direction or reduce the adjustment step size; until the deviation between the actual target value and the predicted target value falls back to within the preset threshold, the key process parameters output at this time are used as the optimized process parameters.

7. The method according to claim 1, characterized in that, The method further includes: Using a preset production batch as a sliding window, the actual process parameters and actual target values ​​within the window are periodically used as new training samples to retrain and optimize the support vector regression model and the weighted fusion model.

8. A system for optimizing injection molding process parameters based on hybrid algorithm and model fusion, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method as described in any one of claims 1 to 7.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images