Method and device for determining parameters, storage medium and program product
By optimizing die-casting parameters using a target proxy model and a parameter optimization model, the problem of low efficiency in manual experience-based adjustments was solved, achieving efficient and accurate mold temperature control, improving casting quality and reducing costs.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-27
- Publication Date
- 2026-03-27
AI Technical Summary
In existing die-casting systems, mold temperature control parameters rely on manual experience for adjustment, resulting in low efficiency, high cost, and unstable casting quality.
By employing a target surrogate model and a preset parameter optimization model, and acquiring multiple sets of die-casting parameters and defect prediction data, the mold temperature control parameters are automatically optimized. The Monte Carlo tree search and real-area Bayesian optimization algorithms are used to optimize the parameters and determine the optimal die-casting parameters.
It improves the efficiency and accuracy of setting die-casting parameters, reduces the number of experiments, significantly improves the quality of castings, and reduces costs.
Smart Images
Figure CN121744403A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of optimizing casting process parameters, and more particularly to a method, apparatus, storage medium, and program product for determining parameters. Background Technology
[0002] In the design and production of die-cast products, the process parameters of the die-casting equipment (e.g., the temperature control parameters of the mold temperature controller and mold cooling station) have a significant impact on the quality of the cast parts. Currently, in die-casting systems, mold temperature control parameters often rely on manual adjustment based on experience. For example, engineers conduct multiple experimental adjustments based on experience to refine the mold temperature control parameters. Summary of the Invention
[0003] To overcome the problems existing in related technologies, this disclosure provides a method, apparatus, storage medium, and program product for determining parameters.
[0004] According to a first aspect of the present disclosure, a method for determining parameters is provided, comprising: Obtain multiple sets of first die-casting parameters corresponding to the target die-casting equipment; For each set of the first die casting parameters, the defect prediction data of the die casting corresponding to the first die casting parameters is determined by the target surrogate model. The target surrogate model represents the mapping relationship between the die casting parameters and the casting defects. Based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, target die-casting parameters are determined through the target proxy model and the preset parameter optimization model. The target die-casting parameters are used to die-cast the castings using the target die-casting equipment. The preset parameter optimization model is used to determine the target die-casting parameters based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters.
[0005] Optionally, determining the target die-casting parameters based on the plurality of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters through the target surrogate model and the preset parameter optimization model includes: The parameter optimization steps are executed repeatedly until the preset termination condition is met. When the preset termination condition is met, the multiple sets of second die-casting parameters output by the preset parameter optimization model are used as alternative die-casting parameters. The defect prediction data corresponding to each group of candidate die casting parameters is determined by the target proxy model, and the target die casting parameter is determined from the candidate die casting parameters based on the defect prediction data corresponding to each group of candidate die casting parameters. The parameter optimization step includes: The multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters are used as input data for the preset parameter optimization model. Based on the input data, the preset parameter optimization model outputs multiple sets of second die-casting parameters, and the number of sets of second die-casting parameters is the same as the number of sets of first die-casting parameters. The second die-casting parameter is used as the updated first die-casting parameter, and the defect prediction data corresponding to the updated first die-casting parameter is determined by the target surrogate model.
[0006] Optionally, the preset parameter optimization model includes a Monte Carlo Tree Search (LAMCTS) sub-model and a True Region Bayesian Optimization (TuRBO) sub-model; the step of outputting multiple sets of second die-casting parameters based on the input data through the preset parameter optimization model includes: Based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, the target parameter range is determined by searching the parameter space through the LAMCTS sub-model. The target parameter range is the parameter value range in which the second die-casting parameter to be output is located. Based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, the multiple sets of second die-casting parameters are obtained by performing parameter search within the target parameter range through the TuRBO sub-model.
[0007] Optionally, obtaining multiple sets of first die-casting parameters corresponding to the target die-casting equipment includes: Obtain a first preset number of the plurality of first die-casting parameters from the historical die-casting parameters of the target die-casting equipment; or... The preset parameter optimization model randomly generates a first preset number of multiple sets of first die-casting parameters.
[0008] Optionally, the target agent model is pre-trained in the following manner: A second preset number of historical die casting parameters are obtained from multiple sets of historical die casting parameters of the target die casting equipment as die casting sample parameters, and the actual defect data of the die casting corresponding to each set of historical die casting parameters in the die casting sample parameters are obtained. The target surrogate model is trained using the Gaussian process regression method based on the die-casting sample parameters and the actual defect data corresponding to each group of historical die-casting parameters in the die-casting sample parameters.
[0009] Optionally, the step of training the target surrogate model using a Gaussian process regression method based on the actual defect data corresponding to each group of historical die-casting parameters in the die-casting sample parameters includes: For each set of historical die casting parameters in the die casting sample parameters, the historical die casting parameters are input into the preset proxy model to be trained, and the model output defect data corresponding to the historical die casting parameters is obtained. Obtain the preset kernel function corresponding to the Gaussian process, and construct the log marginal likelihood function based on the preset kernel function; Based on the actual defect data and the model output defect data corresponding to each group of historical die-casting parameters in the die-casting sample parameters, the target surrogate model is obtained by maximizing the logarithmic marginal likelihood function to determine the parameters of the preset kernel function.
[0010] Optionally, the method further includes: After setting the die casting parameters of the target die casting equipment to the target die casting parameters, the target die casting equipment is controlled to perform die casting to obtain the target die casting product. Obtain the defect data of the target die-casting product, and store the defect data of the target die-casting product and the target die-casting parameters in a preset database; After training the target agent model using data from the preset database, an updated target agent model is obtained.
[0011] According to a second aspect of the present disclosure, an apparatus for determining parameters is provided, comprising: The acquisition module is configured to acquire multiple sets of first die-casting parameters corresponding to the target die-casting equipment; The defect prediction module is configured to determine the defect prediction data of the die casting corresponding to each set of the first die casting parameters through a target proxy model. The target proxy model represents the mapping relationship between the die casting parameters and the casting defects. The parameter determination module is configured to determine target die casting parameters based on the plurality of first die casting parameters and the defect prediction data corresponding to each set of first die casting parameters, through the target proxy model and the preset parameter optimization model. The target die casting parameters are used to die cast parts through the target die casting equipment. The preset parameter optimization model is used to determine the target die casting parameters based on the plurality of first die casting parameters and the defect prediction data corresponding to each set of first die casting parameters.
[0012] According to a third aspect of the present disclosure, an apparatus for determining parameters is provided, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to perform the steps of the method described in the first aspect of this disclosure.
[0013] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps of the method described in the first aspect of the present disclosure.
[0014] According to a fifth aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method described in the first aspect of the present disclosure.
[0015] The technical solutions provided by the embodiments of this disclosure can include the following beneficial effects: the defects of the casting after die casting using the first die casting parameters can be predicted by the pre-trained target surrogate model, and the target die casting parameters can be determined by the parameter optimization through the target surrogate model and the preset parameter optimization model based on multiple sets of first die casting parameters and the defect prediction data corresponding to each set of first die casting parameters. This avoids the need to manually set the die casting parameters, which is highly efficient. Furthermore, the parameter setting is more accurate when the target die casting parameters are determined based on the defect prediction data, so that better die casting parameters can be found in less time (number of trials).
[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description
[0017] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.
[0018] Figure 1 This is a flowchart illustrating a method for determining parameters according to an exemplary embodiment.
[0019] Figure 2 It is based on Figure 1 The illustrated embodiment shows a flowchart of a method for determining parameters.
[0020] Figure 3 This is a flowchart illustrating a model training method according to an exemplary embodiment.
[0021] Figure 4 It is based on Figure 1 The embodiment illustrates a flowchart of a method for determining parameters.
[0022] Figure 5 This is a block diagram illustrating an apparatus for determining parameters according to an exemplary embodiment.
[0023] Figure 6 It is based on Figure 5 The illustrated embodiment shows a block diagram of an apparatus for determining parameters.
[0024] Figure 7 It is based on Figure 5 The illustrated embodiment shows a block diagram of an apparatus for determining parameters.
[0025] Figure 8 This is a block diagram illustrating an apparatus for determining parameters according to an exemplary embodiment. Detailed Implementation
[0026] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.
[0027] The embodiments described in the following examples of this disclosure are not representative of all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.
[0028] It should be noted that all actions involving the acquisition of signals, information, or data in this disclosure are carried out in compliance with the relevant data protection laws and policies of the country where the location is situated, and with authorization from the owner of the relevant device.
[0029] This disclosure is primarily applied to scenarios involving the optimization of casting process parameters, particularly the optimization of temperature control parameters for mold temperature controllers and mold cooling stations. Related technologies mainly rely on engineers' manual experience to adjust these parameters through multiple experiments, but this method is inefficient and ineffective. For example, in the automotive industry's rear floor die-casting solutions, the die-casting parameters are overly complex, human experience has significant limitations, and manual adjustments require numerous experiments, resulting in a large number of scrap parts, leading to low efficiency and high costs.
[0030] To address the aforementioned problems, this disclosure provides a method, apparatus, storage medium, and program product for determining parameters. The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings.
[0031] Figure 1 This is a flowchart illustrating a method for determining parameters according to an exemplary embodiment, such as... Figure 1 As shown, the method includes the following steps.
[0032] In step S11, multiple sets of first die-casting parameters corresponding to the target die-casting equipment are obtained.
[0033] The target die-casting equipment may include a mold temperature controller and / or a mold cooling station, or it may include a die-casting machine. The first die-casting parameter, as well as the second die-casting parameter and historical die-casting parameter mentioned later, may include mold temperature control parameters or die-casting machine parameters. The mold temperature control parameters may include the temperature, opening time, and delay time of a preset temperature regulating medium. This preset temperature regulating medium may include, for example, water or oil. Taking water as an example, the opening time refers to the water flow time of the mold temperature controller and / or mold cooling station, and the delay time refers to the delayed water flow time of the mold temperature controller and / or mold cooling station. Thus, a set of first die-casting parameters may include water temperature, water flow time, and delayed water flow time. Additionally, the die-casting machine parameters may include, for example, the vacuum value, clamping force, and injection speed of the die-casting machine.
[0034] In this step, a first preset number of multiple sets of first die casting parameters can be obtained from the historical die casting parameters of the target die casting equipment. Alternatively, the first preset number of multiple sets of first die casting parameters can be randomly generated by a preset parameter optimization model. For example, the parameter range of the multiple sets of first die casting parameters to be randomly generated can be input into the preset parameter optimization model. Then, the preset parameter optimization model can use a preset optimization algorithm to generate the multiple sets of first die casting parameters. The preset optimization algorithm may include, for example, the LAMCTS (Latent Action Monte Carlo Tree Search) algorithm and the TuRBO (True Region Bayesian Optimization) algorithm.
[0035] The first preset number can usually be set based on experience from parameter optimization experiments. For example, the first preset number can be 8, 10 or 12.
[0036] In step S12, for each group of first die casting parameters, the defect prediction data of the die casting corresponding to the first die casting parameter is determined by the target proxy model. The target proxy model represents the mapping relationship between the die casting parameters and the casting defects.
[0037] The target surrogate model can be a Gaussian Process Model (GP model) trained using the Gaussian Process Regression method. This target surrogate model can be used to map the relationship between die-casting parameters and casting defects.
[0038] In this step, each set of the first die-casting parameters can be input into the target proxy model, and the target proxy model can output the defect prediction data corresponding to each set of the first die-casting parameters.
[0039] It is understandable that if the temperature control parameters of the mold or the parameters of the die-casting machine are not set properly during die casting, the quality of the casting product will inevitably be affected. In this disclosure, improper setting of the temperature control parameters of the mold or the parameters of the die-casting machine can cause internal defects such as cracks, shrinkage cavities, and porosity in the casting, and the defect area will vary depending on the parameter settings.
[0040] For each set of first die-casting parameters, the defect prediction data refers to the defect prediction data of the casting produced after die-casting using that set of first die-casting parameters. This defect prediction data may include, for example, the defect area of the corresponding casting and the probability of occurrence corresponding to that defect area. As mentioned above, casting defects may include internal defects, such as shrinkage cavities, cracks, and porosity. In practical applications, X-ray inspection can be used to determine whether the casting has such defects and the defect area of each type of defect.
[0041] In step S13, based on multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, target die-casting parameters are determined through a target proxy model and a preset parameter optimization model. The target die-casting parameters are used to die-cast the castings using the target die-casting equipment. The preset parameter optimization model is used to determine the target die-casting parameters based on multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters.
[0042] Here, the target die-casting parameter can be understood as the optimal die-casting parameter that can be set for the target die-casting equipment at the current moment. The casting produced using this optimal die-casting parameter has the smallest defect area. This preset parameter optimization model can be used to optimize parameters based on the target proxy model. During the parameter optimization process, the optimal parameter combination can be determined as the target die-casting parameter based on multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters.
[0043] Using the above method, the defects of the casting after die casting using the first die casting parameter can be predicted by the pre-trained target surrogate model. Based on multiple sets of first die casting parameters and the defect prediction data corresponding to each set of first die casting parameters, the target die casting parameter is determined after parameter optimization by the target surrogate model and the preset parameter optimization model. This avoids manually setting the die casting parameter, which is highly efficient. Furthermore, determining the target die casting parameter based on the defect prediction data makes the parameter setting more accurate. In this way, better die casting parameters can be found in less time (number of trials).
[0044] Figure 2 It is based on Figure 1 The illustrated embodiment shows a flowchart of a method for determining parameters, as shown below. Figure 2As shown, step S13 includes the following sub-steps: In step S131, the parameter optimization step is executed cyclically until the preset termination condition is met.
[0045] In one implementation, the parameter optimization steps include: Multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters are used as input data for a preset parameter optimization model. Based on the input data, multiple sets of second die-casting parameters are output through the preset parameter optimization model, with the number of sets of second die-casting parameters being the same as the number of sets of first die-casting parameters. The second die-casting parameters are used as updated first die-casting parameters, and the defect prediction data corresponding to the updated first die-casting parameters is determined through a target proxy model.
[0046] The preset parameter optimization model includes a Monte Carlo Tree Search (LAMCTS) sub-model and a True Region Bayesian Optimization (TuRBO) sub-model. Thus, during the process of outputting multiple sets of second die-casting parameters based on the input data through the preset parameter optimization model, the target parameter range is determined by searching the parameter space using the LAMCTS sub-model based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters. The target parameter range is the range of parameter values for the second die-casting parameters to be output. Then, based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, the multiple sets of second die-casting parameters are obtained by searching the parameter space within the target parameter range using the TuRBO sub-model.
[0047] In one implementation, after inputting multiple sets of first die-casting parameters and the corresponding defect prediction data for each set of first die-casting parameters into a preset parameter optimization model, the optimal parameter setting interval (i.e., the target parameter interval) can be searched in a complex parameter space using a LAMCTS sub-model based on the defect prediction data corresponding to each set of first die-casting parameters. This process can be understood as identifying the target parameter interval containing the optimal parameters globally. The target parameter interval can be determined after searching the parameter space based on the working principle of LAMCTS provided in relevant literature (e.g., "Learning Search Space Partitioning for Black-Box Optimization Using Monte Carlo Tree Search" https: / / arxiv.org / abs / 2007.00708). Then, based on the defect prediction data corresponding to each set of first die-casting parameters, a fine-grained parameter search can be performed within the target parameter interval using a TuRBO sub-model to obtain the multiple sets of second die-casting parameters. The TuRBO sub-model can perform the fine-grained parameter search within the target parameter interval using a Bayesian optimization approach.
[0048] For example, assuming the first die-casting parameters include four sets of die-casting parameters A, B, C, and D, after inputting the defect prediction data corresponding to these four sets of die-casting parameters into the preset parameter optimization model, parameter intervals can be divided based on the LAMCTS sub-model. For instance, if the defect prediction area corresponding to the parameter in set A is 2, the defect prediction area corresponding to the parameter in set B is 4, the defect prediction area corresponding to the parameter in set C is 6, and the defect prediction area corresponding to the parameter in set D is 8, the average defect prediction areas corresponding to sets A and B can be taken, and the average defect prediction areas corresponding to sets C and D can be taken. It is found that the average defect area corresponding to sets A and B is small, while the average defect area corresponding to sets C and D is large. Thus, the parameter interval corresponding to the parameter value whose defect prediction area is smaller than the average defect area corresponding to sets A and B can be selected as the target parameter interval. After determining the target parameter interval, the multiple sets of second die-casting parameters can be determined based on the defect prediction data corresponding to each set of the first die-casting parameters using the Bayesian optimization method of the TuRBO sub-model. The above example is only illustrative and is not limited in this disclosure.
[0049] It should be noted that the above parameter optimization steps usually need to be executed multiple times. Each execution outputs a preset number of sets of second die-casting parameters. Then, each set of second die-casting parameters in the preset number of sets is input into the target proxy model to obtain the defect prediction data corresponding to that set of second die-casting parameters. Then, the defect prediction data is input into the preset parameter optimization model to output a new preset number of sets of second die-casting parameters. This process is repeated until a preset termination condition is met. At this point, the preset number of sets of second die-casting parameters output by the preset parameter optimization model are input into the target proxy model, and the target die-casting parameter is obtained based on the defect prediction data corresponding to each set of second die-casting parameters output by the model.
[0050] The preset termination condition may include, for example, the current number of iterations reaching the preset maximum number of iterations, or the convergence of the objective function value used to predict the target die-casting parameter.
[0051] In step S132, multiple sets of second die-casting parameters output by the preset parameter optimization model when the preset termination condition is met are used as candidate die-casting parameters.
[0052] It is understandable that the loop ends when the preset termination condition is met. As mentioned above, the second die-casting parameter of the preset number group output by the preset parameter optimization model when the preset termination condition is met can be used as the alternative die-casting parameter.
[0053] In step S133, the defect prediction data corresponding to each group of candidate die casting parameters is determined by the target surrogate model, and the target die casting parameter is determined from the candidate die casting parameters based on the defect prediction data corresponding to each group of candidate die casting parameters.
[0054] In this step, each set of candidate die-casting parameters can be input into the target proxy model. The target proxy model can output the defect prediction data corresponding to the set of candidate die-casting parameters. Then, the defect prediction data corresponding to each set of candidate die-casting parameters can be sorted, and the set of candidate die-casting parameters with the smallest defect prediction data (such as defect area) can be selected as the target die-casting parameter.
[0055] Figure 3 This is a flowchart illustrating a model training method according to an exemplary embodiment, such as... Figure 3 As shown, the target agent model can be pre-trained through the following steps: In step S31, a second preset number of historical die casting parameters are obtained from multiple sets of historical die casting parameters of the target die casting equipment as die casting sample parameters, and the actual defect data of the die casting corresponding to each set of historical die casting parameters in the die casting sample parameters are obtained.
[0056] The second preset quantity is also an empirical value, and the order of magnitude of the second preset quantity can be several thousand sets of data. For example, 3,000 sets of historical die-casting parameters can be selected as the die-casting sample parameters for training to obtain the target surrogate model.
[0057] By performing this step, the second preset number (referring to the number of sets of historical die-casting parameters) of the target die-casting equipment's multiple sets of historical die-casting parameters, along with the actual defect data of the die-casting parts corresponding to each set of historical die-casting parameters, can be used as die-casting sample parameters to train the target surrogate model. Specifically, for each set of historical die-casting parameters in the die-casting sample parameters, the actual defect data can include the defect area of the die-casting part obtained when die-casting using that set of historical die-casting parameters.
[0058] After obtaining the historical die-casting parameters of the second preset number of groups of the target die-casting equipment, and the actual defect data of the die-casting parts corresponding to each group of historical die-casting parameters, these data can be cleaned and standardized and used as training samples.
[0059] In step S32, the target surrogate model is trained using the Gaussian process regression method based on the actual defect data corresponding to each group of historical die-casting parameters in the die-casting sample parameters.
[0060] The target surrogate model can be a Gaussian process model (GP model).
[0061] In this step, for each set of historical die-casting parameters in the die-casting sample parameters, the historical die-casting parameters are input into the preset surrogate model to be trained, and the model output defect data corresponding to the historical die-casting parameters is obtained; the preset kernel function corresponding to the Gaussian process is obtained, and the logarithmic marginal likelihood function is constructed based on the preset kernel function; based on the actual defect data and model output defect data corresponding to each set of historical die-casting parameters in the die-casting sample parameters, the parameters of the preset kernel function are determined by maximizing the logarithmic marginal likelihood function to obtain the target surrogate model.
[0062] For example, in a Gaussian process, a kernel function needs to be defined to measure the similarity between input samples. For instance, this preset kernel function could be the RBF (Radial Basis Function) kernel function, whose formula is as follows:
[0063] in, This represents the similarity between input vectors x and y, where x represents historical die-casting parameters and y represents the defect area corresponding to those historical die-casting parameters. The square of the Euclidean distance between x and y. This represents a positive parameter that controls the width of the kernel function. As the value increases, the width of the RBF kernel decreases, corresponding to a more complex decision boundary; when... The smaller the size, the wider the RBF kernel, corresponding to a simpler decision boundary.
[0064] The training process of a Gaussian process model is essentially the process of learning the kernel function parameters. The training objective is to find a set of kernel function parameters that make the predicted defect area corresponding to historical die-casting parameters as close as possible to the true value. This can be achieved by maximizing the log-marginal likelihood function, which can be:
[0065] Where y is the defect area corresponding to each set of historical die-casting parameters, x is the historical die-casting parameter, K is the kernel matrix, and n is the number of samples (i.e., the number of sets of historical die-casting parameters). Thus, the target surrogate model can be obtained by using the gradient descent optimization algorithm to maximize the logarithmic marginal likelihood function and then determining the parameters of the preset kernel function.
[0066] The above examples are merely illustrative and are not intended to limit the scope of this disclosure.
[0067] After training the target surrogate model, for a new set of die-casting parameters (such as the second die-casting parameters output by the preset parameter optimization model), the defect area corresponding to this new set of die-casting parameters can be predicted using the following formula:
[0068] in, This indicates the new set of die-casting parameters. This indicates the defect area corresponding to the new set of die-casting parameters. This indicates the new set of die-casting parameters. The kernel matrix of the historical die-casting parameters x (i.e., the training data), This represents the inverse of the kernel matrix of the training data x.
[0069] Furthermore, this disclosure can also use the target surrogate model to calculate the variance of the prediction in order to assess the uncertainty of the prediction. This uncertainty can be characterized by the probability of occurrence of the corresponding predicted defect area, and the calculation formula is as follows:
[0070] in, The predicted defect area is... The probability of occurrence, This indicates the new set of die-casting parameters. The kernel matrix, Represents the inverse of the kernel matrix of the training data x. This indicates the new set of die-casting parameters. The kernel matrix with respect to the historical die-casting parameter x.
[0071] Figure 4 It is based on Figure 1 The embodiment illustrates a flowchart of a method for determining parameters, such as... Figure 4 As shown, the method also includes the following steps: In step S14, after setting the die casting parameters of the target die casting equipment to the target die casting parameters, die casting is performed to obtain the target die casting product.
[0072] In step S15, defect data of the target die-casting product is obtained, and the defect data and target die-casting parameters of the target die-casting product are stored in a preset database.
[0073] In step S16, the target agent model is trained using data from a preset database to obtain an updated target agent model.
[0074] By executing steps S14-S16, the collected new die-casting parameters and defect data can be fed back into the preset database, so that the next round of proxy model training and parameter optimization settings can be carried out based on the new die-casting parameters in the preset database, thereby continuously improving the quality of casting products.
[0075] It should be noted that after obtaining the target die-casting parameters, these parameters can be used as input for the die-casting experiment. Specifically, steps S14-S16 are executed during the die-casting experiment to obtain more accurate parameters. Since these target die-casting parameters are determined based on defect prediction data output by the target surrogate model, the parameter settings are more accurate. This allows for finding better die-casting parameters in less time (number of experiments), thus enabling more efficient optimization of the die-casting parameters for the target die-casting equipment.
[0076] Furthermore, this disclosure uses a target surrogate model and a preset parameter optimization model to determine the target die-casting parameters. Practical data proves that the target surrogate model and the preset parameter optimization model have good fitting effects, the experimental results are in line with expectations, the AI model prediction results are consistent with the trend of the manually set experimental group, and the defect area of the die-casting parts generated using the target die-casting parameters obtained by the AI model is significantly reduced compared with the defect area of the die-casting parts generated using the die-casting parameters set by human experience. Therefore, this disclosure uses the pre-trained target surrogate model to predict the defects of the castings after die-casting using the first die-casting parameters, and further determines the target die-casting parameters based on the defect prediction data, making the parameter setting more accurate and improving the quality of the casting products.
[0077] Figure 5 This is a block diagram illustrating an apparatus for determining parameters according to an exemplary embodiment, such as... Figure 5 As shown, the device may include: The acquisition module 501 is configured to acquire multiple sets of first die-casting parameters corresponding to the target die-casting equipment; The defect prediction module 502 is configured to determine the defect prediction data of the die casting corresponding to each set of the first die casting parameters through a target proxy model. The target proxy model represents the mapping relationship between the die casting parameters and the casting defects. The parameter determination module 503 is configured to determine target die casting parameters based on the plurality of first die casting parameters and the defect prediction data corresponding to each set of first die casting parameters, through the target proxy model and the preset parameter optimization model. The target die casting parameters are used to die cast the castings through the target die casting equipment. The preset parameter optimization model is used to determine the target die casting parameters based on the plurality of first die casting parameters and the defect prediction data corresponding to each set of first die casting parameters.
[0078] Optionally, the parameter determination module 503 is configured to repeatedly execute the parameter optimization step until a preset termination condition is met; take multiple sets of second die-casting parameters output by the preset parameter optimization model when the preset termination condition is met as candidate die-casting parameters; determine the defect prediction data corresponding to each set of candidate die-casting parameters through the target proxy model, and determine the target die-casting parameter from the candidate die-casting parameters based on the defect prediction data corresponding to each set of candidate die-casting parameters; The parameter optimization step includes: The multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters are used as input data for the preset parameter optimization model; multiple sets of second die-casting parameters are output through the preset parameter optimization model based on the input data, wherein the number of sets of second die-casting parameters is the same as the number of sets of first die-casting parameters; the second die-casting parameters are used as updated first die-casting parameters, and the defect prediction data corresponding to the updated first die-casting parameters is determined through the target proxy model.
[0079] Optionally, the preset parameter optimization model includes a Monte Carlo Tree Search (LAMCTS) sub-model and a True Region Bayesian Optimization (TuRBO) sub-model; the parameter determination module 503 is configured to determine a target parameter range by searching the parameter space through the LAMCTS sub-model based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, wherein the target parameter range is the parameter value range in which the second die-casting parameter to be output is located; and to obtain the multiple sets of second die-casting parameters by searching the parameter space within the target parameter range through the TuRBO sub-model based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters.
[0080] Optionally, the acquisition module 501 is configured to acquire a first preset number of the plurality of first die-casting parameters from the historical die-casting parameters of the target die-casting equipment; or, to randomly generate a first preset number of the plurality of first die-casting parameters through the preset parameter optimization model.
[0081] Optionally, Figure 6 It is based on Figure 5 The illustrated embodiment shows a block diagram of a device for determining parameters, as follows: Figure 6 As shown, the device also includes: Model training module 504 is configured to pre-train the target agent model in the following manner: A second preset number of historical die casting parameters are obtained from multiple sets of historical die casting parameters of the target die casting equipment as die casting sample parameters, and the actual defect data of the die casting corresponding to each set of historical die casting parameters in the die casting sample parameters are obtained. The target surrogate model is trained using the Gaussian process regression method based on the die-casting sample parameters and the actual defect data corresponding to each group of historical die-casting parameters in the die-casting sample parameters.
[0082] Optionally, the model training module 504 is configured to, for each set of historical die-casting parameters in the die-casting sample parameters, input the historical die-casting parameters into a preset surrogate model to be trained, obtain the model output defect data corresponding to the historical die-casting parameters; obtain the preset kernel function corresponding to the Gaussian process, and construct a logarithmic marginal likelihood function based on the preset kernel function; and obtain the target surrogate model by maximizing the logarithmic marginal likelihood function to determine the parameters of the preset kernel function based on the actual defect data and the model output defect data corresponding to each set of historical die-casting parameters in the die-casting sample parameters.
[0083] Optionally, Figure 7 It is based on Figure 5 The illustrated embodiment shows a block diagram of a device for determining parameters, as follows: Figure 7 As shown, the device also includes: The parameter setting module 505 is configured to set the die casting parameters of the target die casting equipment to the target die casting parameters, and then control the target die casting equipment to perform die casting to obtain the target die casting product. The model update module 506 is configured to acquire the defect data of the target die-casting product and store the defect data and the target die-casting parameters of the target die-casting product in a preset database; and to train the target proxy model using the data in the preset database to obtain the updated target proxy model.
[0084] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.
[0085] This disclosure also provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the steps of the method for determining parameters provided in this disclosure.
[0086] Figure 8 This is a block diagram illustrating an apparatus for determining parameters according to an exemplary embodiment. For example, apparatus 800 may be a mobile phone, computer, digital broadcasting terminal, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, etc.
[0087] Reference Figure 8 The device 800 may include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input / output interface 812, a sensor component 814, and a communication component 816.
[0088] Processing component 802 typically controls the overall operation of device 800, such as operations associated with display, telephone calls, data communication, camera operation, and recording. Processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the method for determining parameters described above. Furthermore, processing component 802 may include one or more modules to facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802.
[0089] Memory 804 is configured to store various types of data to support the operation of device 800. Examples of this data include instructions for any application or method operating on device 800, contact data, phonebook data, messages, pictures, videos, etc. Memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.
[0090] Power supply component 806 provides power to various components of device 800. Power supply component 806 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to device 800.
[0091] Multimedia component 808 includes a screen that provides an output interface between the device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of the touch or swipe action but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 808 includes a front-facing camera and / or a rear-facing camera. When the device 800 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.
[0092] Audio component 810 is configured to output and / or input audio signals. For example, audio component 810 includes a microphone (MIC) configured to receive external audio signals when device 800 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 804 or transmitted via communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.
[0093] Input / output interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.
[0094] Sensor assembly 814 includes one or more sensors for providing status assessments of various aspects of device 800. For example, sensor assembly 814 may detect the on / off state of device 800, the relative positioning of components such as the display and keypad of device 800, changes in the position of device 800 or a component of device 800, the presence or absence of user contact with device 800, the orientation or acceleration / deceleration of device 800, and temperature changes of device 800. Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 814 may also include an accelerometer, a gyroscope, a magnetometer, a pressure sensor, or a temperature sensor.
[0095] Communication component 816 is configured to facilitate wired or wireless communication between device 800 and other devices. Device 800 can access wireless networks based on communication standards, such as WiFi, 2G, or 3G, or combinations thereof. In one exemplary embodiment, communication component 816 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 816 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.
[0096] In an exemplary embodiment, the apparatus 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the method of determining parameters described above.
[0097] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 804 including instructions, which can be executed by a processor 820 of the device 800 to complete the method of determining parameters described above. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.
[0098] In another exemplary embodiment, a computer program product is also provided, the computer program product comprising a computer program executable by a programmable device, the computer program having a code portion for performing the method of determining the parameters described above when executed by the programmable device.
[0099] Those skilled in the art will also understand that the various illustrative logical blocks and steps listed in the embodiments of this application can be implemented by electronic hardware, computer software, or a combination of both. Whether such functionality is implemented through hardware or software depends on the specific application and the overall system design requirements. Those skilled in the art can implement the described functionality using various methods for each specific application, but such implementation should not be construed as exceeding the scope of protection of the embodiments of this application.
[0100] In the above detailed description, reference has been made to the accompanying drawings, which illustrate specific aspects of this disclosure by way of illustration. In this regard, terms indicating direction or positional relationship, such as “center,” “longitudinal,” “lateral,” “length,” “width,” “thickness,” “upper,” “lower,” “front,” “rear,” “left,” “right,” “vertical,” “horizontal,” “top,” “bottom,” “inner,” “outer,” “clockwise,” “counterclockwise,” “axial,” “radial,” and “circumferential,” are used with reference to the orientation of the described figures. Since components of the described device can be positioned in multiple different orientations, directional terms are used for illustrative purposes and not for limitation. It should be understood that other aspects can be utilized and structural or logical changes can be made without departing from the concept of this disclosure. Therefore, the following detailed description should not be considered limiting.
[0101] It should be understood that, unless otherwise specifically indicated, features of various embodiments of this disclosure described herein can be combined with each other. As used herein, the term “and / or” includes any one of the relevant listed items and any combination of any two or more; similarly, “at least one of…” includes any one of the relevant listed items and any combination of any two or more.
[0102] It should be understood that, unless otherwise expressly specified and limited, the terms "joining," "attaching," "installing," "connecting," "linking," "fixing," etc., used in the embodiments of this disclosure should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms herein based on the specific circumstances.
[0103] Furthermore, the term "above" as used herein with respect to components, elements, or material layers formed or located "above" a surface may be used to indicate that the component, element, or material layer is "indirectly" positioned (e.g., placed, formed, deposited, etc.) on the surface such that one or more additional components, elements, or layers are arranged between the surface and the component, element, or material layer. However, the term "above" as used with respect to components, elements, or material layers formed or located "above" a surface may also optionally have a specific meaning: that the component, element, or material layer is "directly" positioned (e.g., placed, formed, deposited, etc.) on the surface, for example, in direct contact with the surface.
[0104] Although terms such as “first,” “second,” and “third” may be used herein to describe various components, parts, regions, layers, or sections, these components, parts, regions, layers, or sections are not limited to these terms. Rather, these terms are used only to distinguish one component, part, region, layer, or section from another. Therefore, without departing from the teachings of the examples described herein, the first component, part, region, layer, or section mentioned in the examples may also be referred to as the second component, part, region, layer, or section. Furthermore, the terms “first” and “second” are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as “first” or “second” may explicitly or implicitly include at least one of that feature. In the description herein, “a plurality” means at least two, such as two, three, etc., unless otherwise explicitly specified.
[0105] It should be understood that spatial relative terms, such as “above,” “upper,” “below,” and “lower,” are used herein to describe the relationship between one element and another shown in the figures. In addition to the orientation depicted in the figures, these spatial relative terms are also intended to encompass different orientations of the device in use or operation. For example, if the device in the figures is flipped, an element described as “above” or “upper” relative to another element would be “below” or “lower” relative to that other element. Thus, depending on the spatial orientation of the device, the term “above” encompasses both above and below orientations. Devices may have other orientations (e.g., rotated 90 degrees or in other orientations), and the spatial relative terms used herein should be interpreted accordingly.
[0106] Furthermore, the term “exemplary” is used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as “exemplary” is not necessarily to be construed as advantageous compared to other aspects or designs. Rather, the use of the term “exemplary” is intended to present the concept in a concrete manner. As used herein, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or.” That is, unless otherwise specified or clear from the context, “X applies A or B” is intended to mean any of the natural inclusive arrangements. That is, “X applies A or B” satisfies any of the foregoing instances if X applies A; X applies B; or both X applies A and B. Additionally, unless otherwise specified or clear from the context to refer to the singular form, the articles “a” and “an” as used in this application and the appended claims are generally understood to mean “one or more.”
[0107] Similarly, although this disclosure has been shown and described with respect to one or more implementations, equivalent variations and modifications will occur to those skilled in the art upon reading and understanding this specification and the accompanying drawings. This disclosure includes all such modifications and variations and is limited only by the scope of the claims. In particular, with respect to the various functions performed by the components described above (e.g., elements, resources, etc.), unless otherwise indicated, the terminology used to describe such components is intended to correspond to any component (functionally equivalent) that performs the specific function of the described component, even if structurally not equivalent to the disclosed structure. Furthermore, although specific features of this disclosure may have been disclosed with respect to only one of several implementations, such features may be combined with one or more other features of other implementations, as may be desired and advantageous to any given or particular application. Moreover, with regard to the terms “comprising,” “owning,” “having,” “having,” or variations thereof as used in the detailed description or claims, such terms are intended to be inclusive in a manner similar to the term “including.”
[0108] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the appended claims.
[0109] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.
Claims
1. A method for determining parameters, characterized in that, include: Obtain multiple sets of first die-casting parameters corresponding to the target die-casting equipment; For each set of the first die casting parameters, the defect prediction data of the die casting corresponding to the first die casting parameters is determined by the target surrogate model. The target surrogate model represents the mapping relationship between the die casting parameters and the casting defects. Based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, target die-casting parameters are determined through the target proxy model and the preset parameter optimization model. The target die-casting parameters are used to die-cast the castings using the target die-casting equipment. The preset parameter optimization model is used to determine the target die-casting parameters based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters.
2. The method according to claim 1, characterized in that, The step of determining the target die-casting parameters based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, through the target proxy model and the preset parameter optimization model, includes: The parameter optimization steps are executed repeatedly until the preset termination condition is met. When the preset termination condition is met, the multiple sets of second die-casting parameters output by the preset parameter optimization model are used as alternative die-casting parameters. The defect prediction data corresponding to each group of candidate die casting parameters is determined by the target proxy model, and the target die casting parameter is determined from the candidate die casting parameters based on the defect prediction data corresponding to each group of candidate die casting parameters. The parameter optimization step includes: The multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters are used as input data for the preset parameter optimization model. Based on the input data, the preset parameter optimization model outputs multiple sets of second die-casting parameters, and the number of sets of second die-casting parameters is the same as the number of sets of first die-casting parameters. The second die-casting parameter is used as the updated first die-casting parameter, and the defect prediction data corresponding to the updated first die-casting parameter is determined by the target surrogate model.
3. The method according to claim 2, characterized in that, The preset parameter optimization model includes a Monte Carlo Tree Search (LAMCTS) sub-model and a True Region Bayesian Optimization (TuRBO) sub-model; the step of outputting multiple sets of second die-casting parameters based on the input data through the preset parameter optimization model includes: Based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, the target parameter range is determined by searching the parameter space through the LAMCTS sub-model. The target parameter range is the parameter value range in which the second die-casting parameter to be output is located. Based on the multiple sets of first die-casting parameters and the defect prediction data corresponding to each set of first die-casting parameters, the multiple sets of second die-casting parameters are obtained by performing parameter search within the target parameter range through the TuRBO sub-model.
4. The method according to claim 1, characterized in that, The acquisition of multiple sets of first die-casting parameters corresponding to the target die-casting equipment includes: Obtain a first preset number of the plurality of first die-casting parameters from the historical die-casting parameters of the target die-casting equipment; or... The preset parameter optimization model randomly generates a first preset number of multiple sets of first die-casting parameters.
5. The method according to any one of claims 1-4, characterized in that, The target agent model is pre-trained in the following manner: A second preset number of historical die casting parameters are obtained from multiple sets of historical die casting parameters of the target die casting equipment as die casting sample parameters, and the actual defect data of the die casting corresponding to each set of historical die casting parameters in the die casting sample parameters are obtained. The target surrogate model is trained using the Gaussian process regression method based on the die-casting sample parameters and the actual defect data corresponding to each group of historical die-casting parameters in the die-casting sample parameters.
6. The method according to claim 5, characterized in that, The step of training the target surrogate model using Gaussian process regression based on the die-casting sample parameters and the actual defect data corresponding to each group of historical die-casting parameters in the die-casting sample parameters includes: For each set of historical die casting parameters in the die casting sample parameters, the historical die casting parameters are input into the preset proxy model to be trained, and the model output defect data corresponding to the historical die casting parameters is obtained. Obtain the preset kernel function corresponding to the Gaussian process, and construct the log marginal likelihood function based on the preset kernel function; Based on the actual defect data and the model output defect data corresponding to each group of historical die-casting parameters in the die-casting sample parameters, the target surrogate model is obtained by maximizing the logarithmic marginal likelihood function to determine the parameters of the preset kernel function.
7. The method according to claim 1, characterized in that, The method further includes: After setting the die casting parameters of the target die casting equipment to the target die casting parameters, the target die casting equipment is controlled to perform die casting to obtain the target die casting product. Obtain the defect data of the target die-casting product, and store the defect data of the target die-casting product and the target die-casting parameters in a preset database; After training the target agent model using data from the preset database, an updated target agent model is obtained.
8. An apparatus for determining parameters, characterized in that, include: The acquisition module is configured to acquire multiple sets of first die-casting parameters corresponding to the target die-casting equipment; The defect prediction module is configured to determine the defect prediction data of the die casting corresponding to each set of the first die casting parameters through a target proxy model. The target proxy model represents the mapping relationship between the die casting parameters and the casting defects. The parameter determination module is configured to determine target die casting parameters based on the plurality of first die casting parameters and the defect prediction data corresponding to each set of first die casting parameters, through the target proxy model and the preset parameter optimization model. The target die casting parameters are used to die cast parts through the target die casting equipment. The preset parameter optimization model is used to determine the target die casting parameters based on the plurality of first die casting parameters and the defect prediction data corresponding to each set of first die casting parameters.
9. A device for determining parameters, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to perform the steps of the method according to any one of claims 1-7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method described in any one of claims 1-7.
11. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1-7.