A die-casting parameter tuning method and related device combined with die-casting process priority

Abstract

The present application discloses a die-casting parameter tuning method and related devices combined with the die-casting process priority, which relates to the field of die-casting optimization technology, including: taking out the die-casting parameters one by one from the sorted list in the order of the list, and constructing a decision tree node based on the historical parameter values ​​of the taken die-casting parameters in multiple die-casting production history data and the corresponding product quality grades to obtain a basic decision tree. In the case that the multiple die-casting parameters include die-casting parameters that are not in the sorted list, and the decision tree construction stop condition is not met, new nodes are incrementally constructed on the basic decision tree based on the historical parameter values ​​of the die-casting parameters that are not in the sorted list in multiple die-casting production history data and the corresponding product quality grades to obtain an updated decision tree, and the die-casting parameter values ​​to be optimized are adjusted and optimized through the basic decision tree and / or the updated decision tree. The present application constructs a decision tree that gives priority to the die-casting parameters in the sorted list, thereby improving the efficiency of die-casting parameter optimization.

Description

Technical Field

[0001] The present application relates to the field of die-casting optimization technology, and in particular to a die-casting parameter tuning method and related devices combined with die-casting process priorities. Background Art

[0002] The die-casting parameters of a die-casting machine play a decisive role in the quality of the finished product. Currently, all die-casting parameters in the die-casting process must be optimized based on manual experience before the die-casting machine can begin processing. However, adjusting all die-casting parameters throughout the die-casting process is time-consuming, resulting in inefficient tuning. Manual tuning is also prone to errors, which can be discovered only after the product is finished, wasting raw materials and reducing processing efficiency. Summary of the Invention

[0003] In view of the above problems, this application provides a die-casting parameter tuning method and related devices combined with die-casting process priority to achieve the purpose of quickly and accurately tuning the die-casting parameter values ​​to be optimized. The specific solution is as follows:

[0004] A first aspect of the present application provides a die-casting parameter tuning method combined with die-casting process priorities, comprising:

[0005] Acquire a plurality of die-casting production history data of a target die-casting machine, wherein each die-casting production history data consists of a product quality grade and a corresponding historical parameter value of a plurality of die-casting parameters;

[0006] Obtaining a sorted list, the sorted list including at least some of the die-casting parameters among the plurality of die-casting parameters, the some of the die-casting parameters being the first die-casting parameters in the die-casting process, and the some of the die-casting parameters being at the front of the sorted list;

[0007] Extracting die-casting parameters one by one from the sorted list in the order of the list, and constructing decision tree nodes based on historical parameter values ​​of the extracted die-casting parameters in the plurality of die-casting production history data and corresponding product quality grades to obtain a basic decision tree;

[0008] When the plurality of die-casting parameters include a die-casting parameter not in the sorted list and a decision tree construction stop condition is not met, incrementally constructing new nodes on the basic decision tree based on historical parameter values ​​of the die-casting parameter not in the sorted list in the plurality of die-casting production history data and corresponding product quality grades to obtain an updated decision tree;

[0009] The die-casting parameter values ​​to be optimized are regulated and optimized by means of the basic decision tree and / or the updated decision tree.

[0010] In one possible implementation, the die-casting parameters in the sorted list are arranged based on a target order, and the target order is composed of at least one of the following three orders: the order of the die-casting parameters from front to back in the die-casting process, the order of the die-casting parameters' sensitivity to product quality from high to low, and the order based on manual experience.

[0011] In a possible implementation, the adjusting and optimizing the die-casting parameter value to be optimized by using the basic decision tree and / or the updated decision tree includes:

[0012] According to each die-casting stage in the die-casting process, the die-casting parameter values ​​to be optimized in the corresponding stage are adjusted and optimized stage by stage through the basic decision tree and / or the updated decision tree.

[0013] In a possible implementation, constructing a decision tree node based on historical parameter values ​​of the extracted die-casting parameters in the plurality of die-casting production history data and corresponding product quality grades includes:

[0014] The extracted die-casting parameters are used as parameters to be processed;

[0015] Enumerate all segmentation values ​​of the parameter to be processed to form a segmentation value set corresponding to the parameter to be processed;

[0016] For each split value in the split value set, calculating a splitting criterion metric value when splitting the sample set to be split according to the split value, as the splitting criterion metric value corresponding to the split value, so as to obtain splitting criterion metric values ​​corresponding to all the split values ​​in the split value set, wherein the sample set to be split is composed of part of the die-casting production history data or all of the die-casting production history data;

[0017] Determining a first target segmentation value from the segmentation value set according to the splitting criterion metric values ​​corresponding to all segmentation values ​​in the segmentation value set;

[0018] Based on the first target segmentation value, a decision tree node corresponding to the parameter to be processed is constructed.

[0019] In a possible implementation, constructing a decision tree node corresponding to the parameter to be processed based on the first target segmentation value includes:

[0020] Segmenting the sample set to be segmented according to the first target segmentation value to obtain a segmentation result;

[0021] Determining whether the segmentation result satisfies a preset valid segmentation condition, wherein the valid segmentation condition is one or more of the following conditions: the sample set to be segmented can be segmented into only one set, and the splitting criterion metric value corresponding to the first target segmentation value is less than a preset splitting threshold;

[0022] If so, a decision tree node corresponding to the parameter to be processed is constructed based on the first target segmentation value.

[0023] In a possible implementation, the die-casting parameters not in the sorted list include at least one die-casting parameter;

[0024] The incrementally constructing a new node on the basic decision tree based on the historical parameter values ​​of the die-casting parameters not in the sorted list in the plurality of die-casting production history data and the corresponding product quality grades includes:

[0025] For each die-casting parameter of the at least one die-casting parameter, enumerate all the segmentation values ​​of the die-casting parameter to obtain all the segmentation values ​​of the at least one die-casting parameter; and form a target segmentation value set from all the segmentation values ​​of the at least one die-casting parameter;

[0026] For each split value in the target split value set, calculating a splitting criterion metric value when the sample set to be split is split according to the split value, as the splitting criterion metric value corresponding to the split value, so as to obtain splitting criterion metric values ​​corresponding to all the split values ​​in the target split value set;

[0027] Determining a second target split value from the target split value set according to the splitting criterion metrics corresponding to all the split values ​​in the target split value set;

[0028] Based on the second target cutoff value, a decision tree node corresponding to the die-casting parameter corresponding to the second target cutoff value is incrementally constructed on the basic decision tree.

[0029] A second aspect of the present application provides a die-casting parameter tuning device combined with die-casting process priority, comprising:

[0030] A historical data acquisition module is used to acquire a plurality of die-casting production historical data of a target die-casting machine, wherein each of the die-casting production historical data consists of a product quality grade and a corresponding historical parameter value of a plurality of die-casting parameters;

[0031] a sorting list acquisition module, configured to acquire a sorting list, wherein the sorting list includes at least some of the die-casting parameters among the plurality of die-casting parameters, wherein the some of the die-casting parameters are the first die-casting parameters in the die-casting process, and the some of the die-casting parameters are located at the front of the sorting list;

[0032] a decision tree construction module, configured to extract die-casting parameters one by one from the sorted list in the order of the list, construct decision tree nodes based on historical parameter values ​​of the extracted die-casting parameters in the plurality of die-casting production history data and corresponding product quality grades, so as to obtain a basic decision tree; and, if the plurality of die-casting parameters include die-casting parameters not in the sorted list and a decision tree construction stop condition is not met, incrementally construct new nodes on the basic decision tree based on historical parameter values ​​of the die-casting parameters not in the sorted list in the plurality of die-casting production history data and corresponding product quality grades, so as to obtain an updated decision tree;

[0033] The die-casting parameter tuning module is used to adjust and optimize the die-casting parameter values ​​to be optimized through the basic decision tree and / or the updated decision tree.

[0034] The third aspect of the present application provides a computer program product, including computer-readable instructions. When the computer-readable instructions are run on an electronic device, the electronic device implements the die-casting parameter tuning method combined with the die-casting process priority of the above-mentioned first aspect or any implementation method of the first aspect.

[0035] A fourth aspect of the present application provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0036] The memory is used to store computer programs;

[0037] The processor is used to execute the computer program so that the electronic device can implement the die-casting parameter tuning method combined with the die-casting process priority of the above-mentioned first aspect or any implementation method of the first aspect.

[0038] The fifth aspect of the present application provides a computer storage medium, which carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement the die-casting parameter tuning method combined with the die-casting process priority of the above-mentioned first aspect or any implementation method of the first aspect.

[0039] By means of the above technical solution, the present application provides a die-casting parameter tuning method combined with die-casting process priority. First, a plurality of die-casting production history data of a target die-casting machine is obtained, and a sorted list is obtained. Then, the die-casting parameters are taken out one by one from the sorted list in the order of the list. Based on the historical parameter values ​​of the taken die-casting parameters in the plurality of die-casting production history data and the corresponding product quality grades, a decision tree node is constructed to obtain a basic decision tree. In the case where the plurality of die-casting parameters include die-casting parameters not in the sorted list, new nodes are incrementally constructed on the basic decision tree based on the historical parameter values ​​of the die-casting parameters not in the sorted list in the plurality of die-casting production history data and the corresponding product quality grades to obtain an updated decision tree. Finally, the die-casting parameter values ​​to be optimized are adjusted and optimized through the basic decision tree and / or the updated decision tree. It can be seen from this that the present application can accurately identify whether there are erroneous values ​​in the die-casting parameter values ​​to be optimized by constructing a basic decision tree and / or updating a decision tree. When an erroneous value is found, the die-casting parameter values ​​to be optimized can be quickly and accurately adjusted based on the feedback of the basic decision tree and / or the updated decision tree, thereby reducing the probability of parameter errors caused by manual parameter adjustment and improving the efficiency and accuracy of die-casting parameter optimization.

[0040] Furthermore, the present application can give priority to the die-casting parameters that are at the forefront of the die-casting process during the decision tree construction process, so that the die-casting parameters that are at the forefront of the die-casting process can be arranged as close to the root of the tree as possible, thereby improving the stage perception ability of the basic decision tree and / or the updated decision tree. This ability enables the basic decision tree and / or the updated decision tree to identify potential defects in the early stage of product processing, and adjust the die-casting parameter values ​​to be optimized based on the identified potential defects, further improving the efficiency of die-casting parameter optimization, and the identification of potential defects can effectively avoid the subsequent processing of defective products, save raw material and equipment resources, and further improve the efficiency of product processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that the originals and elements are not necessarily drawn to scale.

[0042] Figure 1 A schematic flow chart of a die-casting parameter tuning method combined with die-casting process priority provided in this application;

[0043] Figure 2 A schematic structural diagram of a die-casting parameter tuning device combined with die-casting process priority provided by this application;

[0044] Figure 3This is a schematic diagram of the structure of an electronic device provided in this application. DETAILED DESCRIPTION

[0045] The following describes the embodiments of the present application in conjunction with the accompanying drawings. The terms used in the implementation methods of the present application are only used to explain the specific embodiments of the present application and are not intended to limit the present application.

[0046] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0047] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, and this is merely a way of distinguishing the objects of the same attributes when describing them in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0048] In order to achieve data-based, explainable and adaptable intelligent parameter adjustment optimization and solve the above-mentioned problems of clustering algorithms, the present application provides the following die-casting parameter tuning method and related devices combined with die-casting process priority.

[0049] In order to enable those skilled in the art to better understand the present application, the die-casting parameter tuning method combined with the die-casting process priority of the embodiment of the present application is described in detail below with reference to the accompanying drawings.

[0050] Reference Figure 1 , is a flow chart of a die-casting parameter tuning method combined with die-casting process priority provided by this application, such as Figure 1 As shown, the die-casting parameter tuning method combined with die-casting process priority may include:

[0051] Step S101: Acquire multiple die-casting production history data of a target die-casting machine.

[0052] Each die-casting production history data consists of a product quality grade and the historical parameter values ​​of a plurality of corresponding die-casting parameters.

[0053] Considering that even within the same production site, different die-casting machines may differ in brand, model, control algorithm, and other aspects, resulting in inconsistent operating characteristics and parameter response patterns, this exacerbates the complexity of parameter optimization. Furthermore, the cost of a single die-casting machine is substantial, potentially reaching hundreds of millions. To ensure that such a high-cost die-casting machine can produce products with the highest possible yield, this embodiment optionally allows the target die-casting machine to be the single die-casting machine for die-casting parameter optimization.

[0054] It should be understood that in the die-casting production process of the target die-casting machine, the yield rate is affected by the comprehensive influence of multiple die-casting parameter factors. This embodiment can obtain multiple die-casting production history data of the target die-casting machine, so as to study and analyze the operating characteristics and parameter response laws of the target die-casting machine through multiple die-casting production history data, obtain optimized parameters that are more suitable for the target die-casting machine, and improve the yield rate of the target die-casting machine.

[0055] Die-casting parameters refer to a series of technical parameters that must be set and controlled during the die-casting process. These parameters directly impact the quality, production efficiency, and cost of the final product. In-depth research has identified die-casting parameters that influence yield and can be optimized and adjusted, including the target die-casting machine's process parameters and parameters related to environmental climate conditions.

[0056] To this end, optionally, the multiple die-casting parameters in each die-casting production history data include: process parameters of the target die-casting machine and parameters related to environmental climate conditions. Among them, process parameters play a decisive role in the product molding process. These parameters affect the flow state, pressure distribution and heat conduction process of the molten metal in the mold cavity, thereby largely determining the density and molding quality of the final product. Improper setting of process parameters can easily lead to common defects such as short shots, pores, cold shuts, and deformations. Ambient temperature and climate conditions also have a significant impact on the die-casting process, especially in areas with large fluctuations in humidity and heat. For example, some areas have significant differences in climate throughout the year, resulting in different thermal balance response conditions for die-casting equipment in different regions. This regional environmental difference makes the parameter adjustment experience accumulated in the long-term production process in various places highly localized, making it difficult to directly reuse or migrate applications.

[0057] Optionally, the process parameters of the target die-casting machine include one or more of the following parameters: average high-speed speed, cake thickness, high-speed starting position, cycle time, pressure building time, filling distance, filling time, filling pressure, ejection pressure, reaching the highest speed position, maximum speed of the injection curve and maximum value of the filling pressure.

[0058] Optionally, parameters related to environmental climate conditions include: the temperature of the day, date (if the historical parameter values ​​of the date are not much different, this application will not consider the date parameter; if the historical parameter values ​​of the date are greatly different, this application may consider the date parameter; the same applies to the temperature of the day).

[0059] Of course, the above die-casting parameters are merely examples and are not intended to limit the present application.

[0060] Considering that in actual production, the target die-casting machine may have many die-casting parameters, it will take a lot of time to build a decision tree based on the full set of die-casting parameters. For this reason, in one possible implementation, some more important die-casting parameters (i.e., die-casting parameters that have a significant impact on product quality results) can be first screened from the full set of die-casting parameters as multiple die-casting parameters in this embodiment.

[0061] Optionally, before screening some more important die-casting parameters from the full set of die-casting parameters, the collected raw data can be cleaned and standardized to eliminate outliers, fill in or delete missing items, and ensure that the quality of the die-casting production history data for subsequent modeling is stable and reliable.

[0062] Optionally, “more important die-casting parameters” may refer to die-casting parameters that can distinguish product quality levels by adjusting parameter values, and / or die-casting parameters whose correlation with product quality levels is higher than a preset correlation threshold.

[0063] Optionally, in order to screen out "die-casting parameters that can distinguish product quality grades by adjusting parameter values", numerical statistical methods can be used for screening. Specifically, for each die-casting parameter in the full amount of die-casting parameters: obtain multiple historical parameter values ​​of the die-casting parameter and the corresponding product quality grades, and generate a data distribution graph corresponding to the die-casting parameter based on the multiple historical parameter values ​​of the die-casting parameter and the corresponding product quality grades; to obtain data distribution graphs corresponding to all die-casting parameters in the full amount of die-casting parameters; based on the data distribution graphs corresponding to all die-casting parameters, screen out die-casting parameters from all die-casting parameters that meet the requirements of being able to distinguish product quality grades by adjusting parameter values.

[0064] Optionally, in order to screen out "die-casting parameters whose correlation with product quality grade is higher than a preset correlation threshold", a correlation analysis method can be used for screening. Specifically, for each die-casting parameter in the full amount of die-casting parameters: obtain multiple historical parameter values ​​of the die-casting parameter and the corresponding product quality grade, and calculate the correlation between the die-casting parameter and the product quality grade according to the multiple historical parameter values ​​of the die-casting parameter and the corresponding product quality grade, as the correlation corresponding to the die-casting parameter; so as to obtain the correlation corresponding to all die-casting parameters in the full amount of die-casting parameters; according to the correlation corresponding to all die-casting parameters (optionally, the correlation can be a Pearson correlation), screen out die-casting parameters whose correlation with product quality grade is higher than a preset correlation threshold from all die-casting parameters.

[0065] Optionally, the product quality level includes good products and defective products. For example, good products are represented by the number 1 and defective products are represented by the number 0.

[0066] Of course, the above-mentioned product quality grades may also be other grades, which are not limited here.

[0067] Step S102: Obtain a sorted list.

[0068] It is understandable that the die-casting process of the product includes multiple die-casting stages such as mold preheating, spraying, mold closing, filling, pressure holding, cooling, and mold opening. The target die-casting machine needs to process the product step by step until the processing of the last die-casting stage is completed to obtain a formed product.

[0069] Each of the aforementioned die-casting stages corresponds to specific die-casting parameters, and these parameters can vary across different stages. As discussed in the background section, current product processing requires waiting for all die-casting parameters to be fully tuned before proceeding, significantly reducing processing efficiency. To enable simultaneous tuning and rapid optimization, the following decision tree based on die-casting stage priority was constructed.

[0070] In order to improve the decision tree's perception of each die-casting stage, this embodiment can first obtain a sorted list, which includes at least some of the die-casting parameters among the multiple die-casting parameters. The part of the die-casting parameters refers to several die-casting parameters that are at the front of the die-casting process for multiple die-casting parameters, and the part of the die-casting parameters is located at the front of the sorted list, so that the part of the die-casting parameters can be given priority when constructing the decision tree below.

[0071] Optionally, in order to further improve the stage perception ability of the decision tree, the die-casting parameters in the sorting list can be arranged based on a target order, which is composed of at least one of the following three orders: the order of die-casting parameters from front to back in the die-casting process, the order of die-casting parameters' sensitivity to product quality from high to low, and the order of manual experience.

[0072] Of course, the target order can also be other, which is not specifically limited here.

[0073] Optionally, the sorted list can be obtained by process experts or experienced engineers sorting the various die-casting parameters that need to be prioritized in the decision tree. For example, the process experts or experienced engineers sort the various die-casting parameters that need to be prioritized in the decision tree based on the above-mentioned target order.

[0074] Of course, the sorting list can also be a list of all die-casting parameters that need to be prioritized in the decision tree, given the above-mentioned target order or a list of all die-casting parameters sorted based on the target order, and is not limited here.

[0075] Step S103: extract die-casting parameters one by one from the sorted list in the order of the list, and construct decision tree nodes based on the historical parameter values ​​of the extracted die-casting parameters in multiple die-casting production history data and the corresponding product quality grades to obtain a basic decision tree.

[0076] In order to build a decision model with visual paths and logical judgment capabilities, so that the subsequent adjustment and optimization process of the die-casting parameter values ​​to be optimized has stronger transparency and explainability, this embodiment can preferentially build a decision tree starting from the root node based on the die-casting parameters in the sorted list to obtain a basic decision tree.

[0077] Since the decision tree is constructed based on the die-casting parameters in the sorted list and is constructed according to the sorting order provided in the sorted list, the die-casting parameters at the front of the die-casting process can be closer to the root of the tree. Therefore, in the early stage of product processing, potential defects of the product can be identified based on the die-casting parameters used in the early processing. If the die-casting parameters used in the early processing are wrong (here, the wrong die-casting parameters mean that the die-casting parameters are not within the range of the die-casting parameters corresponding to good products), the basic decision tree provided in this application (or the updated decision tree below) can be used to quickly identify whether the product is a good product or a defective product. When it is identified as a defective product, there is no need to continue processing, which not only saves raw materials but also improves the efficiency of product processing.

[0078] In this embodiment, during the construction of the basic decision tree, when the following decision tree construction stopping conditions are met, the division is no longer continued, a leaf node is formed, and the construction of the basic decision tree ends.

[0079] Optionally, the decision tree construction stop condition may include any of the following conditions: the number of samples of the current node (one die casting production history data corresponds to one sample) is less than the preset minimum sample threshold , the depth of the current tree reaches the preset maximum depth limit .

[0080] In practical applications, the decision tree construction stopping condition may also be other conditions, such as insufficient number of leaf node samples, etc., which is not specifically limited in this application.

[0081] Step S104: When the multiple die-casting parameters include die-casting parameters that are not in the sorted list and the decision tree construction stop condition is not met, new nodes are incrementally constructed on the basic decision tree based on the historical parameter values ​​of the die-casting parameters that are not in the sorted list in multiple die-casting production history data and the corresponding product quality grades to obtain an updated decision tree.

[0082] As mentioned above, the sorted list may include some die-casting parameters. If the multiple die-casting parameters include die-casting parameters that are not in the sorted list, if the previous basic decision tree has not reached the decision tree construction stop condition, then the historical parameter values ​​of the die-casting parameters that are not in the sorted list in multiple die-casting production history data and the corresponding product quality grades can be used to incrementally construct new nodes on the leaf nodes of the basic decision tree until the decision tree construction stop condition is reached to obtain an updated decision tree.

[0083] Step S105 : adjusting and optimizing the die-casting parameter values ​​to be optimized through the basic decision tree and / or the updated decision tree.

[0084] Specifically, in this embodiment, the die-casting parameter values ​​to be optimized may be input into the basic decision tree and / or the updated decision tree to obtain prediction results.

[0085] Among them, if the sorted list includes the above-mentioned multiple die-casting parameters, this embodiment will not obtain an updated decision tree, then the die-casting parameter values ​​to be optimized can be input into the basic decision tree to obtain a prediction result; if the sorted list includes some of the above-mentioned die-casting parameters and the basic decision tree has not reached the decision tree construction stop condition, then the die-casting parameter values ​​to be optimized can be input into the basic decision tree and / or the updated decision tree to obtain a prediction result; if the sorted list includes some of the above-mentioned die-casting parameters and the basic decision tree has reached the decision tree construction stop condition, then the die-casting parameter values ​​to be optimized can be input into the basic decision tree to obtain a prediction result.

[0086] If the prediction result indicates that the die-casting parameter value to be optimized enables the target die-casting machine to produce good products, the die-casting parameter value to be optimized is not adjusted, and the die-casting parameter value to be optimized is directly output; if the prediction result indicates that the die-casting parameter value to be optimized causes the target die-casting machine to produce defective products, the die-casting parameter value to be optimized is adjusted, and the adjusted die-casting parameter value is output.

[0087] Optionally, the process of adjusting the die-casting parameter value to be optimized may include adjusting the die-casting parameter value to be optimized based on a preset step size and / or a preset adjustment factor, and the adjustment direction may be determined based on the specific die-casting parameter, including upward adjustment and downward adjustment. For example, based on a key node that triggers defective product judgment in the path of the basic decision tree and / or the updated decision tree, the die-casting parameter corresponding to the key node is used as the parameter currently requiring adjustment, and then the parameter value of the parameter is adjusted using the preset step size and / or preset adjustment factor corresponding to the parameter.

[0088] It should also be noted that this application constructs a basic decision tree and / or updates a decision tree, and the decision tree tuning process can also serve as an auxiliary tool for parameter debugging and experience inheritance.

[0089] The present application provides a die-casting parameter tuning method in combination with die-casting process priority. First, a plurality of die-casting production history data of a target die-casting machine is obtained, and a sorted list is obtained. Then, the die-casting parameters are extracted one by one from the sorted list in the order of the list. Based on the historical parameter values ​​of the extracted die-casting parameters in the plurality of die-casting production history data and the corresponding product quality grades, a decision tree node is constructed to obtain a basic decision tree. In the case where the plurality of die-casting parameters include a die-casting parameter not in the sorted list, based on the historical parameter values ​​of the die-casting parameters not in the sorted list in the plurality of die-casting production history data and the corresponding product quality grades, a new node is incrementally constructed on the basic decision tree to obtain an updated decision tree. Finally, the die-casting parameter values ​​to be optimized are adjusted and optimized through the basic decision tree and / or the updated decision tree. It can be seen from this that the present application can accurately identify whether there are erroneous values ​​in the die-casting parameter values ​​to be optimized by constructing a basic decision tree and / or updating a decision tree. When an erroneous value is found, the die-casting parameter values ​​to be optimized can be quickly and accurately adjusted based on the feedback of the basic decision tree and / or the updated decision tree, thereby reducing the probability of parameter errors caused by manual parameter adjustment and improving the efficiency and accuracy of die-casting parameter optimization.

[0090] Furthermore, the present application can give priority to the die-casting parameters that are at the forefront of the die-casting process during the decision tree construction process, so that the die-casting parameters that are at the forefront of the die-casting process can be arranged as close to the root of the tree as possible, thereby improving the stage perception ability of the basic decision tree and / or the updated decision tree. This ability enables the basic decision tree and / or the updated decision tree to identify potential defects in the early stage of product processing, and adjust the die-casting parameter values ​​to be optimized based on the identified potential defects, further improving the efficiency of die-casting parameter optimization, and the identification of potential defects can effectively avoid the subsequent processing of defective products, save raw material and equipment resources, and further improve the efficiency of product processing.

[0091] For example, in a possible scenario, the various die-casting parameter values ​​in the die-casting process are determined by some means (such as manual experience or the method of the present application), and product processing is performed based on the determined die-casting parameter values. During the product processing process, relevant experts (such as process experts or experienced engineers) want to change a certain die-casting parameter value. The changed die-casting parameter value can be verified based on the method of the present application during the product processing process. If it is found that the changed die-casting parameter value will result in the processing of defective products, the tuning strategy in the previous step S105 can be used to search for the die-casting parameter value that can make the processed product a good product near the changed die-casting parameter value, and feedback can be given to the relevant experts. After confirmation by the relevant experts, the product can be subsequently processed according to the die-casting parameter value that can produce good products, thereby improving the efficiency and quality of product processing and improving the user experience of the relevant experts.

[0092] In some embodiments of the present application, the basic decision tree may be any one of a Classification and Regression Tree (CART), an Iterative Dichotomiser 3 (ID3), and a C4.5 Decision Tree.

[0093] Next, the process of "constructing decision tree nodes based on historical parameter values ​​of the extracted die-casting parameters in a plurality of die-casting production history data and corresponding product quality grades" in the above step S103 is introduced through the following embodiments.

[0094] For the convenience of the following introduction, the extracted die-casting parameters are defined as parameters to be processed.

[0095] In this embodiment, all split values ​​of the parameter to be processed can be enumerated to form a split value set corresponding to the parameter to be processed. Then, for each split value in the split value set, the splitting criterion metric value when the sample set to be split is split according to the split value is calculated as the splitting criterion metric value corresponding to the split value, thereby obtaining the splitting criterion metric values ​​corresponding to all split values ​​in the split value set.

[0096] Among them, if the basic decision tree is a CART tree, the splitting criterion measurement value can be a weighted Gini index, if the basic decision tree is an ID3 tree, the splitting criterion measurement value can be information gain, if the basic decision tree is a C4.5 decision tree, the splitting criterion measurement value can be an information gain rate.

[0097] Of course, the basic decision tree can also be other decision trees, and the splitting criterion metric value can also be other, which is not limited here.

[0098] The sample set to be segmented is composed of part of or all of the die-casting production history data in a plurality of die-casting production history data, and as the nodes in the decision tree continue to split, the sample set to be segmented gradually decreases.

[0099] Next, this embodiment may determine a first target segmentation value from the segmentation value set according to the splitting criterion metric values ​​corresponding to all segmentation values ​​in the segmentation value set.

[0100] Among them, if the splitting criterion measurement value is the weighted Gini index, the splitting value with the smallest corresponding splitting criterion measurement value can be determined from the splitting value set as the first target splitting value; if the splitting criterion measurement value is information gain or information gain rate, the splitting value with the largest corresponding splitting criterion measurement value can be determined from the splitting value set as the first target splitting value.

[0101] Finally, similar to the process of constructing a traditional decision tree, after determining the first target cutoff value corresponding to the parameter to be processed, a decision tree node corresponding to the parameter to be processed can be constructed based on the first target cutoff value.

[0102] Taking into account that the first target segmentation value may not be able to effectively segment the sample set to be segmented, this embodiment also sets an effective segmentation condition. Then, the process of "constructing a decision tree node corresponding to the parameter to be processed based on the first target segmentation value" may include: segmenting the sample set to be segmented according to the first target segmentation value to obtain a segmentation result; determining whether the segmentation result meets the preset effective segmentation condition, and if so, constructing a decision tree node corresponding to the parameter to be processed based on the first target segmentation value.

[0103] Optionally, the effective splitting conditions are one or more of the following conditions: the sample set to be split can only be split into one set (that is, the samples in the sample set to be split cannot be distinguished based on the first target splitting value), and the splitting criterion metric value corresponding to the first target splitting value is less than a preset splitting threshold.

[0104] In summary, this embodiment introduces the sequence of die-casting parameters in the die-casting process to construct a basic decision tree. This allows the basic decision tree to judge product quality in the early stages of the die-casting process, enabling earlier product rejection decisions. This reduces the subsequent processing of defective products, saves raw materials and equipment resources, and improves production efficiency. Furthermore, this method makes the basic decision tree structure more consistent with the logic of the die-casting process, enhances the interpretability of the prediction process, and facilitates the understanding and application of on-site engineering personnel.

[0105] The following describes the process of step S104 "incrementally constructing new nodes on the basic decision tree based on historical parameter values ​​of the die-casting parameters not in the sorted list in multiple die-casting production history data and corresponding product quality grades."

[0106] It can be understood that the die-casting parameters not in the above sorted list include at least one die-casting parameter.

[0107] Then, this embodiment can enumerate all the segmentation values ​​of each die-casting parameter in at least one die-casting parameter to obtain all the segmentation values ​​of at least one die-casting parameter; and form a target segmentation value set by all the segmentation values ​​of at least one die-casting parameter.

[0108] Next, for each segmentation value in the target segmentation value set, the splitting criterion metric value when the sample set to be segmented is segmented according to the segmentation value is calculated as the splitting criterion metric value corresponding to the segmentation value, so as to obtain the splitting criterion metric values ​​corresponding to all segmentation values ​​in the target segmentation value set.

[0109] The splitting criterion measurement value here is consistent with the introduction in the previous embodiment. Please refer to the previous introduction for details and will not be repeated here.

[0110] Furthermore, a second target segmentation value is determined from the target segmentation value set according to the splitting criterion metric values ​​corresponding to all the segmentation values ​​in the target segmentation value set.

[0111] Among them, if the splitting criterion measurement value is the weighted Gini index, the splitting value with the smallest corresponding splitting criterion measurement value can be determined from the target splitting value set as the second target splitting value; if the splitting criterion measurement value is information gain or information gain rate, the splitting value with the largest corresponding splitting criterion measurement value can be determined from the target splitting value set as the second target splitting value.

[0112] Finally, based on the second target cutoff value, a decision tree node corresponding to the die-casting parameter corresponding to the second target cutoff value is incrementally constructed on the basic decision tree.

[0113] In order to enable those skilled in the art to better understand the above process of constructing a decision tree, it is explained in detail through the following examples.

[0114] Assume that each die casting production history data has The same die-casting parameters (the historical parameter values ​​in different die-casting production history data are different) are recorded as: .in, Represents a collection of multiple die-casting parameters. Indicates the first Die casting parameters.

[0115] Assume that the die casting parameters and sorting order in the sorted list are as follows: Then, at the first few nodes of the decision tree to be constructed in this application, the die-casting parameters in the above order are preferentially tried for division. It should be noted that the inclusion of three die-casting parameters in the sorted list here is only an example and does not limit this application.

[0116] Taking CART as an example, the construction process of basic decision tree and updated decision tree is introduced.

[0117] The first layer of nodes (i.e. root nodes) are based on Perform segmentation.

[0118] enumerate All the split values ​​of The corresponding segmentation value set is recorded as Then there is The corresponding segmentation value satisfy: .

[0119] The sample set to be split is recorded as , then according to The cut value The sample set to be split can be divided into the following two sets: .

[0120] in, Represents the sample set to be split The left child node (also referred to as ), Represents the sample set to be split The right child node (also referred to as ), Indicates the current node (here The set of historical parameter values ​​of the die-casting parameters in the corresponding root node), Represents the set of product quality levels in the current node.

[0121] For the convenience of the following description, the current node is defined as the node to be split.

[0122] Then, calculate the Gini index of the current node according to the following formula (1).

[0123] Formula (1);

[0124] in, Represents the sample set to be split The corresponding Gini index of the node to be split, Indicates a category, e.g. , 0 represents defective products, 1 represents good products (It should be noted that the categories 0 and 1 here are just examples, and there can be other categories, for example, It can also be 0, 1, 2, etc.); Indicates the The sample ratio of each category is calculated as follows:

[0125] Formula (2);

[0126] in, Represents the sample set to be split The number of samples in Represents the sample set to be split The The product quality level in the samples, It means if and only if The dot in the string takes the value 1 when the condition is met, and takes the value 0 otherwise.

[0127] according to The cut value The calculation formula for the weighted Gini index when splitting the sample set to be split is as follows:

[0128] Formula (3);

[0129] in, Indicates that The cut value The weighted Gini index when splitting the sample set to be split, Indicates that The cut value The number of samples of the left child node split from the split sample set, Indicates that The cut value The number of samples of the right child node split from the split sample set, Indicates that The cut value Treat the Gini index of the left child node of the split sample set as the target. Indicates that The cut value The Gini index of the right child node of the split sample set.

[0130] Thus, you can choose to smallest ,Right now .in, Indicates smallest , which is the first target segmentation value in the root node. Then, we can follow and Build the root node and split it.

[0131] The second layer nodes (i.e. the left and right child nodes of the root node) are based on Perform segmentation.

[0132] Take the left child node as an example, that is, according to turn up , you can follow and Construct the left child node of the root node and split it, where Indicates that The cut value The weighted Gini index when splitting the sample set to be split (the sample set to be split at this time refers to the sample set corresponding to the left child node split from the first-layer node, so the sample set to be split at this time is smaller than the sample set to be split in the first-layer node).

[0133] For details not introduced here, please refer to the introduction of the first-level nodes. I will not repeat them here. The same applies to the right child node.

[0134] The third layer nodes are based on Perform segmentation.

[0135] That is, according to turn up , you can follow and Construct the left child node of the root node and split it, where Indicates that The cut value The weighted Gini index when splitting the sample set to be split.

[0136] For details not introduced here, please refer to the introduction of the first-layer nodes, which will not be repeated here.

[0137] Through the above process, a basic decision tree can be constructed.

[0138] The nodes on the fourth layer and beyond are based on to Perform segmentation.

[0139] right to For each die casting parameter in, enumerate all the segmentation values ​​of the die casting parameter to obtain the die casting parameter. For example, The set of all segmentation values ​​of , . to All the respective split values ​​compose the target split value set.

[0140] For each cut value in the target cut value set, the weighted Gini index when the cut sample set is cut according to the cut value (and the die casting parameter corresponding to the cut value) can be calculated according to the above formula (1)-formula (3): , Indicates that The cut value The weighted Gini index when splitting the sample set to be split.

[0141] Then, you can follow turn up and ,in, Indicates smallest , Indicates the selected die-casting parameters (i.e., the die-casting parameters in the nodes to be split corresponding to the sample set to be split), Indicates smallest (i.e., the second target split value in the node to be split corresponding to the sample set to be split).

[0142] In each of the above node splits, it is necessary to determine whether the split node meets the decision tree construction stop condition after the split. If so, the split node is regarded as a leaf node and no further splitting of the leaf node is performed. This results in an updated decision tree after incrementally constructing new nodes on the basic decision tree.

[0143] In a more preferred implementation, in order to ensure that the decision tree is updated more reasonably and effectively, the die-casting parameters used in the current split can be removed after each node split. For example, in the second-level node split, , eliminated in the third-level node splitting and ; The same applies to the splitting of nodes in the fourth layer and later.

[0144] It should also be noted that the construction process of the basic decision tree and the updated decision tree in ID3 and C4.5 is similar to that of CART. For details, please refer to the previous introduction. It is only necessary to replace the weighted Gini index with the information gain in ID3 or the information gain rate in C4.5.

[0145] The following only provides the calculation formulas for the information gain in ID3 and the information gain rate in C4.5, and the detailed tree construction process is not introduced.

[0146] In ID3, the information entropy of the node is first calculated, which can measure the information entropy of the sample set to be segmented. The uncertainty is as follows, and the calculation formula of information entropy is as follows:

[0147] Formula (4);

[0148] in, Represents the sample set to be split The information entropy of the corresponding node, Indicates a category, e.g. , 0 represents defective products, 1 represents good products; Indicates the The sample ratio of each category is calculated using the formula (2) above.

[0149] Then calculate the conditional entropy as follows:

[0150] Formula (5);

[0151] in, Indicates that The split value is Conditional entropy when dividing the left child node and the right child node, Indicates that The split value is The number of samples of the left child node of the partition, Indicates that The split value is The number of samples of the right child node of the partition, Indicates that The split value is The information entropy of the left child node of the partition, Indicates that The split value is The information entropy of the right child node of the partition.

[0152] Information gain represents the reduction in information entropy before and after dividing the left child node and the right child node. The formula is as follows:

[0153] Formula (6);

[0154] in, Indicates that The split value is Information gain when partitioning the left and right child nodes.

[0155] In the 4th layer and later layers, you can follow turn up and .in, Indicates The largest , Indicates the selected die-casting parameters (i.e., the die-casting parameters in the nodes to be split corresponding to the sample set to be split), Indicates The largest (i.e. the second target split value in the node to be split corresponding to the sample set to be split); in the first three layers, it has been determined by the sorted list , so the first three layers only need to determine That’s it. For the detailed process, please refer to the introduction in CART, which will not be repeated here.

[0156] In C4.5, the information gain rate further considers the amount of information in the partition itself based on the information gain. First, the intrinsic value of the partition is defined as:

[0157] Formula (7);

[0158] in, Indicates that The split value is The intrinsic value when partitioning the left and right child nodes.

[0159] The information gain ratio is defined as the ratio of information gain to the partition intrinsic value, that is:

[0160] Formula (8);

[0161] in, Indicates that The split value is The information gain rate when partitioning the left and right child nodes.

[0162] In the 4th layer and later layers, you can follow turn up and .in, Indicates The largest , Indicates the selected die-casting parameters (i.e., the die-casting parameters in the nodes to be split corresponding to the sample set to be split), Indicates The largest (i.e. the second target split value in the node to be split corresponding to the sample set to be split); in the first three layers, it has been determined by the sorted list , so the first three layers only need to determine That’s it. For the detailed process, please refer to the introduction in CART, which will not be repeated here.

[0163] In summary, when constructing a tree, the decision model in this embodiment prioritizes features at the top of the sorted list when selecting the partitioning features (here, the die-casting parameters) for each layer, arranging these previous stage features as close to the root of the tree as possible. This improves the decision model's stage perception ability, enabling it to identify and judge potential defects early in the process flow. The remaining features are automatically sorted according to traditional criteria and participate in the tree construction. The updated decision tree constructed based on this process effectively improves product processing efficiency.

[0164] In other embodiments of the present application, the process of the aforementioned step S105 of “adjusting and optimizing the die-casting parameter values ​​to be optimized through the basic decision tree and / or the updated decision tree” is introduced.

[0165] The existing manual experience tuning is a global tuning method, which cannot be used to tune each die-casting stage in the die-casting process in stages. In order to achieve stage-by-stage tuning, the following embodiment is provided.

[0166] Optionally, this embodiment may adjust and optimize the die-casting parameter values ​​to be optimized in the corresponding stages step by step through the basic decision tree and / or the updated decision tree according to each die-casting stage in the die-casting process.

[0167] That is, in this embodiment, at each die-casting stage of product processing, the die-casting parameter values ​​to be optimized in the current die-casting stage can be first determined, and then the die-casting parameter values ​​to be optimized in the current die-casting stage and the die-casting parameter values ​​that have been optimized in the previous die-casting stage can be input into the basic decision tree and / or the updated decision tree to obtain the prediction result of whether the product output by the corresponding decision tree is a good product or a defective product. If the prediction result indicates that the product is a defective product, only the die-casting parameter values ​​to be optimized in the current die-casting stage can be adjusted until the prediction result indicates that the product is a good product.

[0168] Through this stage-by-stage optimization strategy, the product can be processed based on the die-casting parameter values ​​that have been optimized in the previous die-casting stage while adjusting the parameters in the current die-casting stage, thereby improving the efficiency of product processing.

[0169] The above describes a die-casting parameter tuning method combined with die-casting process priority provided in an embodiment of the present application. The following describes an apparatus for executing the above-mentioned die-casting parameter tuning method combined with die-casting process priority.

[0170] See also Figure 2 , Figure 2 This is a schematic diagram of the structure of a die-casting parameter tuning device combined with die-casting process priority provided in an embodiment of the present application. Figure 2 As shown, the die-casting parameter tuning device combined with the die-casting process priority may include:

[0171] A historical data acquisition module 201 is used to acquire a plurality of die-casting production historical data of a target die-casting machine, wherein each die-casting production historical data is composed of a product quality grade and corresponding historical parameter values ​​of a plurality of die-casting parameters;

[0172] A sorting list acquisition module 202 is used to acquire a sorting list, where the sorting list includes at least some of the die-casting parameters among the plurality of die-casting parameters, where the some of the die-casting parameters are the first die-casting parameters in the die-casting process, and the some of the die-casting parameters are at the front of the sorting list;

[0173] A decision tree construction module 203 is configured to extract die-casting parameters one by one from the sorted list in the order of the list, and construct decision tree nodes based on historical parameter values ​​of the extracted die-casting parameters in a plurality of die-casting production history data and corresponding product quality grades to obtain a basic decision tree. If the plurality of die-casting parameters include die-casting parameters not in the sorted list and a decision tree construction stop condition is not met, new nodes are incrementally constructed on the basic decision tree based on historical parameter values ​​of the die-casting parameters not in the sorted list in a plurality of die-casting production history data and corresponding product quality grades to obtain an updated decision tree.

[0174] The die-casting parameter tuning module 204 is configured to adjust and optimize the die-casting parameter values ​​to be optimized through the basic decision tree and / or the updated decision tree.

[0175] In one possible implementation, the die-casting parameters in the above-mentioned sorting list are arranged based on a target order, and the target order is composed of at least one of the following three orders: the order of the die-casting parameters from front to back in the die-casting process, the order of the die-casting parameters' sensitivity to product quality from high to low, and the order of manual experience.

[0176] In a possible implementation, when the die-casting parameter tuning module adjusts and optimizes the die-casting parameter value to be optimized through the basic decision tree and / or the updated decision tree, it can be specifically used to:

[0177] According to each die-casting stage in the die-casting process, the die-casting parameter values ​​to be optimized in the corresponding stage are adjusted and optimized stage by stage through the basic decision tree and / or the updated decision tree.

[0178] In one possible implementation, when constructing a decision tree node based on the historical parameter values ​​of the extracted die-casting parameters in a plurality of die-casting production history data and the corresponding product quality grades, the decision tree construction module may be specifically used to:

[0179] The extracted die-casting parameters are used as parameters to be processed;

[0180] Enumerate all the segmentation values ​​of the parameters to be processed to form a segmentation value set corresponding to the parameters to be processed;

[0181] For each split value in the split value set, calculating a splitting criterion metric value when splitting the sample set to be split according to the split value, as the splitting criterion metric value corresponding to the split value, so as to obtain splitting criterion metric values ​​corresponding to all the split values ​​in the split value set, wherein the sample set to be split is composed of part of the die-casting production history data or all of the die-casting production history data in the plurality of die-casting production history data;

[0182] Determine a first target segmentation value from the segmentation value set according to the splitting criterion metric values ​​corresponding to all segmentation values ​​in the segmentation value set;

[0183] Based on the first target segmentation value, a decision tree node corresponding to the parameter to be processed is constructed.

[0184] In a possible implementation, when constructing a decision tree node corresponding to the parameter to be processed based on the first target split value, the decision tree construction module may be specifically used to:

[0185] According to the first target segmentation value, the sample set to be segmented is segmented to obtain a segmentation result;

[0186] Determine whether the segmentation result meets the preset valid segmentation conditions, where the valid segmentation conditions are one or more of the following conditions: the sample set to be segmented can only be segmented into one set, and the splitting criterion metric value corresponding to the first target segmentation value is less than a preset splitting threshold;

[0187] If so, a decision tree node corresponding to the parameter to be processed is constructed based on the first target segmentation value.

[0188] In one possible implementation, the die-casting parameters not in the sorted list include at least one die-casting parameter. Then, when incrementally constructing new nodes on the basic decision tree based on the historical parameter values ​​of the die-casting parameters not in the sorted list in multiple die-casting production history data and the corresponding product quality grades, the decision tree construction module can specifically be used to:

[0189] For each die-casting parameter of the at least one die-casting parameter, enumerate all the segmentation values ​​of the die-casting parameter to obtain all the segmentation values ​​of the at least one die-casting parameter; and form a target segmentation value set from all the segmentation values ​​of the at least one die-casting parameter;

[0190] For each split value in the target split value set, calculate the splitting criterion metric value when the sample set to be split is split according to the split value, as the splitting criterion metric value corresponding to the split value, so as to obtain the splitting criterion metric values ​​corresponding to all the split values ​​in the target split value set;

[0191] Determining a second target split value from the target split value set according to the splitting criterion metric values ​​corresponding to all split values ​​in the target split value set;

[0192] Based on the second target cutoff value, a decision tree node corresponding to the die-casting parameter corresponding to the second target cutoff value is incrementally constructed on the basic decision tree.

[0193] It should be noted that the specific implementation methods of each module in the above-mentioned die-casting parameter tuning device combined with die-casting process priority can refer to the relevant introduction in the die-casting parameter tuning method combined with die-casting process priority in the previous text, and will not be repeated here.

[0194] It should also be noted that the various modules in the aforementioned die-casting parameter optimization device incorporating die-casting process priority can be implemented in whole or in part through software, hardware, or a combination thereof. The aforementioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of the aforementioned modules.

[0195] An embodiment of the present application further provides an electronic device, which may include at least one processor and a memory connected to the processor, wherein:

[0196] Memory is used to store computer programs;

[0197] The processor is used to execute a computer program so that the electronic device can implement any one of the die-casting parameter tuning methods combined with the die-casting process priority provided in the embodiments of the present application.

[0198] refer to Figure 3 , which shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present application. The electronic device in the embodiments of the present application may include, but is not limited to, fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 3 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0199] like Figure 3 As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 602 or programs loaded from a storage device 608 into a random access memory (RAM) 603. When the electronic device is powered on, the RAM 603 also stores various programs and data required for the operation of the electronic device. The processing device 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0200] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a memory card, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Figure 3The electronic device is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead.

[0201] An embodiment of the present application also provides a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements any one of the die-casting parameter tuning methods combined with the die-casting process priority provided in the embodiment of the present application.

[0202] A computer-readable storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any one of the die-casting parameter tuning methods combined with the die-casting process priority provided in the embodiment of the present application.

[0203] It should also be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided in this application, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines.

[0204] Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course can also be implemented by special hardware including application-specific integrated circuits, special CPUs, special memories, special components, etc. In general, all functions performed by computer programs can be easily implemented with corresponding hardware, and the specific hardware structures used to implement the same function can also be diverse, such as analog circuits, digital circuits or special circuits, etc. However, for the present application, software program implementation is a better implementation method in most cases. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer's floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk or optical disk, etc., and includes a number of instructions to enable a computer device (which can be a personal computer, training equipment, or network equipment, etc.) to execute the methods described in each embodiment of the present application.

[0205] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware, or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0206] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, a computer, a training device or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website, a computer, a training device or a data center. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium can be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

Claims

1. A die casting parameter tuning method combined with die casting process priority, characterized in that: include: Acquire a plurality of die-casting production history data of a target die-casting machine, wherein each die-casting production history data consists of a product quality grade and a corresponding historical parameter value of a plurality of die-casting parameters; Obtaining a sorted list, the sorted list including at least some of the die-casting parameters among the plurality of die-casting parameters, the some of the die-casting parameters being the first die-casting parameters in the die-casting process, and the some of the die-casting parameters being at the front of the sorted list; Extracting die-casting parameters one by one from the sorted list in the order of the list, and constructing decision tree nodes based on historical parameter values ​​of the extracted die-casting parameters in the plurality of die-casting production history data and corresponding product quality grades to obtain a basic decision tree; When the plurality of die-casting parameters include a die-casting parameter not in the sorted list and a decision tree construction stop condition is not met, incrementally constructing new nodes on the basic decision tree based on historical parameter values ​​of the die-casting parameter not in the sorted list in the plurality of die-casting production history data and corresponding product quality grades to obtain an updated decision tree; The die-casting parameter values ​​to be optimized are regulated and optimized by means of the basic decision tree and / or the updated decision tree.

2. The die-casting parameter tuning method according to claim 1, characterized in that: The die-casting parameters in the sorting list are arranged based on a target order, and the target order is composed of at least one of the following three orders: the order of the die-casting parameters from front to back in the die-casting process, the order of the die-casting parameters' sensitivity to product quality from high to low, and the order of manual experience.

3. The die-casting parameter tuning method according to claim 1, characterized in that: The step of adjusting and optimizing the die-casting parameter value to be optimized by using the basic decision tree and / or the updated decision tree includes: According to each die-casting stage in the die-casting process, the die-casting parameter values ​​to be optimized in the corresponding stage are adjusted and optimized stage by stage through the basic decision tree and / or the updated decision tree.

4. The die-casting parameter tuning method according to any one of claims 1 to 3, characterized in that: The step of constructing a decision tree node based on historical parameter values ​​of the extracted die-casting parameters in the plurality of die-casting production history data and corresponding product quality grades includes: The extracted die-casting parameters are used as parameters to be processed; Enumerate all segmentation values ​​of the parameter to be processed to form a segmentation value set corresponding to the parameter to be processed; For each split value in the split value set, calculating a splitting criterion metric value when splitting the sample set to be split according to the split value, as the splitting criterion metric value corresponding to the split value, so as to obtain splitting criterion metric values ​​corresponding to all the split values ​​in the split value set, wherein the sample set to be split is composed of part of the die-casting production history data or all of the die-casting production history data; Determining a first target segmentation value from the segmentation value set according to the splitting criterion metric values ​​corresponding to all segmentation values ​​in the segmentation value set; Based on the first target segmentation value, a decision tree node corresponding to the parameter to be processed is constructed.

5. The die-casting parameter tuning method according to claim 4, characterized in that: The step of constructing a decision tree node corresponding to the parameter to be processed based on the first target segmentation value includes: Segmenting the sample set to be segmented according to the first target segmentation value to obtain a segmentation result; Determining whether the segmentation result satisfies a preset valid segmentation condition, wherein the valid segmentation condition is one or more of the following conditions: the sample set to be segmented can be segmented into only one set, and the splitting criterion metric value corresponding to the first target segmentation value is less than a preset splitting threshold; If so, a decision tree node corresponding to the parameter to be processed is constructed based on the first target segmentation value.

6. The die-casting parameter tuning method according to claim 5, characterized in that: The die-casting parameters not in the sorted list include at least one die-casting parameter; The incrementally constructing a new node on the basic decision tree based on the historical parameter values ​​of the die-casting parameters not in the sorted list in the plurality of die-casting production history data and the corresponding product quality grades includes: For each die-casting parameter of the at least one die-casting parameter, enumerate all the segmentation values ​​of the die-casting parameter to obtain all the segmentation values ​​of the at least one die-casting parameter; and form a target segmentation value set from all the segmentation values ​​of the at least one die-casting parameter; For each split value in the target split value set, calculating a splitting criterion metric value when the sample set to be split is split according to the split value, as the splitting criterion metric value corresponding to the split value, so as to obtain splitting criterion metric values ​​corresponding to all the split values ​​in the target split value set; Determining a second target split value from the target split value set according to the splitting criterion metrics corresponding to all the split values ​​in the target split value set; Based on the second target cutoff value, a decision tree node corresponding to the die-casting parameter corresponding to the second target cutoff value is incrementally constructed on the basic decision tree.

7. A die casting parameter tuning device combined with die casting process priority, characterized in that: include: A historical data acquisition module is used to acquire a plurality of die-casting production historical data of a target die-casting machine, wherein each of the die-casting production historical data consists of a product quality grade and a corresponding historical parameter value of a plurality of die-casting parameters; a sorting list acquisition module, configured to acquire a sorting list, wherein the sorting list includes at least some of the die-casting parameters among the plurality of die-casting parameters, wherein the some of the die-casting parameters are the first die-casting parameters in the die-casting process, and the some of the die-casting parameters are located at the front of the sorting list; a decision tree construction module, configured to extract die-casting parameters one by one from the sorted list in the order of the list, construct decision tree nodes based on historical parameter values ​​of the extracted die-casting parameters in the plurality of die-casting production history data and corresponding product quality grades, so as to obtain a basic decision tree; and, if the plurality of die-casting parameters include die-casting parameters not in the sorted list and a decision tree construction stop condition is not met, incrementally construct new nodes on the basic decision tree based on historical parameter values ​​of the die-casting parameters not in the sorted list in the plurality of die-casting production history data and corresponding product quality grades, so as to obtain an updated decision tree; The die-casting parameter tuning module is used to adjust and optimize the die-casting parameter values ​​to be optimized through the basic decision tree and / or the updated decision tree.

8. A computer program product, characterized in that The method comprises computer-readable instructions, which, when executed on an electronic device, enable the electronic device to implement the die-casting parameter tuning method combined with die-casting process priority as claimed in any one of claims 1 to 6.

9. An electronic device, characterized in that: comprising at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is configured to execute the computer program so that the electronic device can implement the die-casting parameter tuning method combined with die-casting process priority as described in any one of claims 1 to 6.

10. A computer storage medium, characterized in that The storage medium carries one or more computer programs, which, when executed by an electronic device, enable the electronic device to implement the die-casting parameter tuning method combined with die-casting process priority as described in any one of claims 1 to 6.

Images