Process data determination method based on sand mold additive manufacturing integration system
By using process parameter matching and optimization models in sand-type additive manufacturing, the problem of determining process parameter data is solved, and efficient and stable process parameter optimization is achieved to meet the production needs of high-quality sand-type.
Patent Information
- Application Number
- CN202411874447.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-19
- Publication Date
- 2025-04-25
AI Technical Summary
During the sand-type additive manufacturing process, it is difficult for the prior art to efficiently and stably determine process parameter data, resulting in waste of resources and increased time costs.
By obtaining process parameter data, inputting the process parameter matching model for similarity matching. If it cannot be matched, the process parameter optimization model is used to optimize the process parameter data based on the machine learning model of artificial intelligence.
It realizes rapid optimization of process parameter data, adapts to the needs of preparing high-quality sand types, and reduces resource waste and time costs.
Smart Images
Figure CN119884770A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of sand mold additive manufacturing, and in particular to an integrated circuit automatic testing system and method based on a TTL protocol. Background Art
[0002] Sand mold additive manufacturing technology is a technology that uses a method of gradually accumulating materials to manufacture solid parts. Compared with traditional material removal: cutting processing technology, it is a "top-down" manufacturing method. In recent years, additive manufacturing technology has developed rapidly. "Rapid prototyping" and "3D printing" are manufacturing systems for specific components. Various industries have repeated basic experiments for process optimization, which is easy to cause waste of resources.
[0003] With the development of modern science and technology and the advent of the information age, more and more companies are paying attention to information management and have implemented a lot of information software. The manufacturing industry is also using more automated equipment to assist production.
[0004] However, the related technology generates a large amount of data, which are stored in different databases with poor sharing. In addition, a large amount of empirical data cannot provide guidance for subsequent manufacturing processes, resulting in increased time and resource costs caused by repeated experiments. Summary of the invention
[0005] The present invention relates to a method for determining process data based on an integrated system for sand mold additive manufacturing, which can realize efficient and stable determination of process data in a sand mold additive manufacturing process. The method is applied in a computer device, and the method includes:
[0006] Acquire process parameter data, where the process parameter data is parameter data associated with the sand mold manufacturing process;
[0007] Inputting process parameter data into a process parameter matching model and outputting process parameter data matching results;
[0008] In response to the process parameter data matching result indicating that there is no close matching result, the process parameter data is input into the process parameter optimization model, and the process parameter optimization result is output.
[0009] In an optional embodiment, the process parameter data includes process parameters and process characteristic data;
[0010] The process parameters include at least one of the content of solvents such as resin / curing agent, thickness of additive manufacturing layer, width of additive manufacturing, and type and mesh size of molding sand;
[0011] The process characteristic data includes at least one of interlayer thermal coupling characteristic data, adaptive slicing and stratification characteristic data, multi-region precise sand spreading characteristic data, sand spreading vibration compaction characteristic data, and sand spreading rolling compaction characteristic data.
[0012] In an optional embodiment, the process parameter data is input into the process parameter matching model, and the process parameter data matching result is output, including:
[0013] inputting process parameter data into a process parameter matching model;
[0014] Calling a process parameter database through a process parameter matching model to perform similarity matching on process parameter data, wherein the process parameter database includes at least one set of ideal process parameters;
[0015] In response to the process parameter data and the ideal process data in the process parameter database meeting the matching rule, outputting a first matching result, the first matching result being a correction result of the process parameter data corresponding to the ideal process parameter;
[0016] In response to the process parameter data and the ideal process data in the process parameter database not meeting the matching rule, a second matching result is output, and the second matching result indicates that there is no similar matching result.
[0017] In an optional embodiment, the method further comprises:
[0018] Acquire sample process parameter data, which are process data based on experiments and physical property parameters;
[0019] A process parameter database is constructed based on sample process parameter data.
[0020] In an optional embodiment, the sample process parameter data is configured with a sample process parameter experimental value, and the sample process parameter experimental value is used to characterize the experimental result of the sample process parameter data.
[0021] In an optional embodiment, the process parameter optimization model is an artificial intelligence-based machine learning model;
[0022] In response to the process parameter data matching result indicating that there is no close matching result, the process parameter data is input into the process parameter optimization model, and the process parameter optimization result is output, including:
[0023] Inputting process parameter data and similarity matching results into a process parameter optimization model;
[0024] Determine adjustment parameter data corresponding to the process parameter data based on the similarity matching result through the process parameter optimization model;
[0025] Combined with the adjustment parameter data, the process parameter optimization results are generated through the process parameter optimization model.
[0026] In an optional embodiment, the parameter adjustment data is implemented in the form of an objective function difference.
[0027] In an optional embodiment, the process parameter optimization model is used to generate a process parameter optimization result in combination with the adjustment parameter data, including:
[0028] Determine new value acceptance criteria and iteration rules in combination with parameter adjustment data;
[0029] Based on the new value acceptance criteria and iteration rules, combined with the adjustment parameter data, the process parameter optimization model generates the process parameter optimization results.
[0030] In an optional embodiment, after the process parameter optimization result is outputted, the process includes:
[0031] Import the process parameter optimization results into the process parameter database.
[0032] In an optional embodiment, the method further comprises:
[0033] In response to receiving the discard instruction, the ideal process parameters in the process parameter database are discarded.
[0034] The technical effects of various embodiments of the present invention include at least:
[0035] After determining the process parameter data associated with the sand mold manufacturing process, the ideal process parameters are matched through the process parameter matching model. If the ideal process parameters cannot be matched, the process parameter data are optimized through the process parameter optimization model, so that the applied process parameter data can be quickly optimized and adapted to the needs of preparing high-quality sand molds. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0037] Figure 1 A schematic flow chart of a method for determining process data based on a sand mold additive manufacturing integrated system provided by an exemplary embodiment of the present application is shown.
[0038] Figure 2 A schematic flow chart of another method for determining process data based on a sand mold additive manufacturing integrated system provided by an exemplary embodiment of the present application is shown.
[0039] Figure 3 A schematic diagram of a process flow for establishing a process parameter database provided by an exemplary embodiment of the present application is shown. DETAILED DESCRIPTION
[0040] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.
[0041] Figure 1 A flow chart of a method for determining process data based on a sand mold additive manufacturing integrated system provided by an exemplary embodiment of the present application is shown. The method is described by taking the application of the method in a computer device as an example. The method includes:
[0042] Step 101, obtaining process parameter data.
[0043] In the embodiment of the present application, the process parameter data is parameter data associated with the sand mold manufacturing process. Optionally, the process parameter data is a guidance parameter in the sand mold manufacturing process, and the guidance parameter is related to the preset parameters of the manufacturing equipment in the manufacturing line. The present application does not limit the specific parameters and specific types of the process parameter data.
[0044] Step 102, input the process parameter data into the process parameter matching model, and output the process parameter data matching result.
[0045] In the embodiment of the present application, the process parameter matching model is a database-based model. The process parameter matching model can perform data matching based on the process parameter data to determine whether the process parameter data has a matching possibility with the process parameters under the ideal state. Optionally, the process parameter data matching result is embodied in the form of fields, and includes a field indicating whether the process parameter data matches the process parameters under the ideal state. If a match is possible, the computer device will simultaneously push the process parameters under the ideal state that can be matched. If a match is not possible, step 103 is executed.
[0046] Step 103 , in response to the process parameter data matching result indicating that there is no similar matching result, the process parameter data is input into a process parameter optimization model, and a process parameter optimization result is outputted.
[0047] In the embodiment of the present application, the process parameter optimization model can be implemented as an artificial intelligence model. The process parameter optimization model can adjust at least one of the process parameter data based on the parameter requirements represented by the process parameter data to obtain a process parameter optimization result. Optionally, the process parameter optimization result is consistent with the process parameter data.
[0048] To sum up, the method provided in the embodiment of the present application, after determining the process parameter data associated with the sand mold manufacturing process, matches the ideal process parameters through the process parameter matching model. If the ideal process parameters cannot be matched, the process parameter data is optimized through the process parameter optimization model, so that the process parameter data to be applied can be quickly optimized and adapted to the needs of preparing high-quality sand molds.
[0049] In this embodiment, the preset machine learning model fits all close points on a line through an algorithm in order to find the best fit. The process parameters include the content of curing agent, heating temperature, layer thickness obtained by the adaptive slicing algorithm, etc. The characteristic values are input into the machine learning model, and the output is the predetermined optimal value of the process parameter. Since the relationship between the above four characteristic indicators is not linear, it is difficult to merge them with a unified indicator. Therefore, it is necessary to establish separate sub-machine learning models for the four characteristic indicators. The output of each sub-machine learning model has only one parameter standard, that is, the machine learning model in this embodiment includes four sub-machine learning models, which output each feature one by one.
[0050] The transfer function of the machine learning model uses a curve tangent function with a variation range of [-1,1]. The average value of multiple machine learning model optimization algorithm values is used to reduce the error, so that the optimized value is closer to the model effect at the optimal convergence. The expected error is set to 0.0001-0.001, and the optimization learning algorithm uses one of the gradient fastest descent optimization algorithm, the Levenberg-Marquardt (LM) algorithm based on nonlinear least squares method, and the Bayesian regularization algorithm (Bayesianregularization, BR).
[0051] According to an embodiment of the present invention, when new experimental data containing process parameters and performance data are obtained, the current machine learning model is trained again using the experimental data; if the goodness of fit of the retrained machine learning model is greater than that of the current machine learning model, the current machine learning model is updated; otherwise, the current machine learning model remains unchanged.
[0052] Figure 2 A flow chart of another method for determining process data of an integrated system for sand mold additive manufacturing provided by an exemplary embodiment of the present application is shown. The method is described by taking the application of the method in a computer device as an example. The method includes:
[0053] Step 201, obtaining process parameter data.
[0054] In the embodiment of the present application, the process parameter data includes process parameters and process characteristic data; the process parameters include at least one of the solvent content of resin / curing agent, additive manufacturing layer thickness, additive manufacturing width, and molding sand type and mesh size; the process characteristic data includes at least one of interlayer thermal coupling characteristic data, adaptive slicing and stratification characteristic data, multi-region precise sand laying characteristic data, sand laying vibration compaction characteristic data, and sand laying rolling compaction characteristic data. Optionally, the process parameter data is data preset for production in actual production practice.
[0055] Step 202: input the process parameter data into the process parameter matching model, and call the process parameter database through the process parameter matching model to perform similarity matching on the process parameter data.
[0056] This process is to input the process parameter data into the process parameter matching model to perform similarity matching.
[0057] Optionally, in an embodiment of the present application, the process parameter matching model will call a process parameter database for data matching. The process parameter database includes at least one set of ideal process parameters. Optionally, the above-mentioned ideal process parameters are derived from process production or computer equipment reasoning. The embodiment of the present application does not limit the source of the ideal process parameters.
[0058] Optionally, the matching process is a similarity matching process combined with a machine learning model.
[0059] Step 203 , in response to the process parameter data and the ideal process data in the process parameter database meeting the matching rule, output a first matching result.
[0060] The first matching result is a correction result of the process parameter data corresponding to the ideal process parameter. In the embodiment of the present application, the first matching result is a direct adjustment of the process parameter data.
[0061] In one example, the process is to calculate the first similarity between the target process data and the sand mold process database through the optimization models corresponding to different features according to the random approximation algorithm comparison, and then calculate the second similarity between some features in the target process data and the characteristic parameters of the sand mold process database. Of course, the optimization is continuously continued, and the similarity will continue to iterate to reach the optimal value. If the second similarity is greater than the first similarity, the target process data matches the process data corresponding to the maximum second total similarity. Otherwise, there is no process plan in the currently formulated sand mold process database that matches the target sand mold process data to be verified. Continuous iteration of the similarity can obtain higher quality data.
[0062] Next, the matching rules applied in the embodiment of the present application are assisted. The matching process may include the following four steps:
[0063] S1: New data generation and preliminary similarity calculation:
[0064] 1) Through the transformation function i=1, ..., n generates new data from current values, and the transformation method affects the neighborhood structure and cooling schedule selection.
[0065] 2) Set the target data T, the initial similarity state S, and perform n iterations.
[0066] 3) Calculate the increment ΔT in each iteration. If ΔT < 0, the current data meets the requirements; otherwise, it is accepted as the optimized value with probability exp(-ΔT / T).
[0067] 4) If the similarity meets the conditions, output the optimal solution; otherwise, continue to iterate until T gradually decreases and T>0, obtain a closer similarity value, and import it into the database.
[0068] Formula verification: Correctly represents how new data is generated.
[0069] S2: Objective function difference calculation:
[0070] Calculate the objective function difference corresponding to the new value, preferably incrementally, since it only results from the transformed part.
[0071] S3: New value acceptance criteria and data tuning:
[0072] 1) Use the Metropolis criterion to determine whether the new value is accepted: if f(ω)≤0, then accept it; otherwise, accept ω′ with probability P;
[0073] 2) Through f(ω′)=￡([Z 1 Z 2 ...Z n ]W hj +β hj ), j=1, ..., m updates the feature node data, performs data optimization and iteration;
[0074] 3) When the layer thickness, width, curing agent content, interlayer thermal coupling, slicing and compaction process data reach a balanced state, relevant experiments are introduced.
[0075] Among them, the formula verification: f(ω′)=￡([Z 1 Z 2 ...Z n ]W hi +β hj ) provides the functional form of the new value acceptance criterion.
[0076] S4: New value replacement and iteration continues:
[0077] 1) If the new value is accepted, replace the current value and correct the objective function value to complete an iteration;
[0078] 2) If the new value is discarded, the next round of experiments will be continued based on the original current value.
[0079] In the embodiment of the present application, the optimization principle of the intelligent algorithm is that the indicator gradient decreases, the energy value decreases, and if the energy value decreases, then this transfer is accepted. If the energy value increases, it means that the system deviates further from the global optimal value position. At this time, the optimization algorithm will not immediately abandon this processing value, but perform probabilistic operations; therefore, the optimization value is dynamically adjusted.
[0080] Step 204 , in response to the process parameter data and the ideal process data in the process parameter database not meeting the matching rule, outputting a second matching result.
[0081] The second matching result indicates that there is no similar matching result. In this case, a subsequent process parameter optimization process is required.
[0082] Step 205 , in response to the process parameter data matching result indicating that there is no similar matching result, the process parameter data and the similarity matching result are input into the process parameter optimization model.
[0083] Step 206 , determining adjustment parameter data corresponding to the process parameter data based on the similarity matching result through the process parameter optimization model.
[0084] That is, in the embodiment of the present application, the target performance data is obtained as the goal, and the target process parameters are found and generated using a multi-objective optimization model. The optimization target of the intelligent optimization algorithm is all the features in the performance data. Based on the previously established machine learning model, the value of each feature is obtained, and then the weighted summation method is used to establish the multi-objective optimization model, and the weight factor follows the weight obtained by the hierarchical analysis method.
[0085] Step 207 , generating a process parameter optimization result through a process parameter optimization model in combination with the adjustment parameter data.
[0086] Optionally, a simulated annealing algorithm is used to correspond one-to-one to each feature in the output performance data, specifically including: optimizing the training model, accelerating the solidification of the sand mold by heating during each experiment, and preventing the step effect, matching the adaptive slicing process, and obtaining the corresponding sand mold layer thickness under different curvatures through an adaptive slicing algorithm. For the influence of heating temperature and curing agent content on the overall performance indicators of the sand mold, regular data under different states can be obtained through experiments. When new experimental data containing sand mold process parameters and performance are obtained, the current rules are automatically combined to optimize the experimental data and train the machine learning model again; the state of the current model is continuously optimized to achieve the optimal state.
[0087] In the embodiment of the present application, the process parameter optimization model is a machine learning model based on artificial intelligence
[0088] Step 208: Import the process parameter optimization results into the process parameter database.
[0089] In an embodiment of the present application, the process parameter optimization results can be restored by a computer device such as a process parameter database for further storage as data.
[0090] Optionally, in some embodiments of the present application, in response to receiving a discard instruction, the ideal process parameters in the process parameter database are discarded.
[0091] Figure 3 A schematic diagram of a process flow for establishing a process parameter database provided by an exemplary embodiment of the present application is shown. The method is described by taking the method applied in a computer device as an example. The method includes:
[0092] Step 301, obtaining sample process parameter data.
[0093] In the embodiment of the present application, the sample process parameter data is process data based on experiments and physical property parameters.
[0094] Optionally, the process data of the sand molds obtained by additive manufacturing under different process parameters are obtained experimentally, and a process parameter database is constructed using the process parameters and process data in the experimental process as cases.
[0095] In this case, the optimal process window can be preliminarily determined in the global process window through orthogonal experiments or general full-factor experiments; the optimal process parameters and corresponding process data can be determined in the optimal process window through response surface method experiments.
[0096] Step 302: construct a process parameter database based on the sample process parameter data.
[0097] In summary, the method provided in the embodiment of the present application, after determining the process parameter data associated with the sand mold manufacturing process, matches the ideal process parameters through the process parameter matching model. If the ideal process parameters cannot be matched, the process parameter data is optimized through the process parameter optimization model, so that the process parameter data to be applied can be quickly optimized and adapted to the needs of preparing high-quality sand molds.
[0098] The method provided in the embodiment of the present application can construct an initial sand mold process database through a small amount of sand mold physical property parameters and process data generated by experimental design. The constructed process database can be expanded at any time to achieve the purpose of real-time optimization of the physical property parameters of the additive manufacturing sand mold and supplementing the process data model.
[0099] (2) Generate target process parameters corresponding to target performance data through intelligent optimization algorithms, evaluate process parameters generated by intelligent optimization through machine learning models trained with experimental data, adjust the rules of intelligent optimization algorithms, and generate process parameters that meet user requirements to expand the process database, so that the model has higher performance prediction and process parameter recommendation capabilities;
[0100] (3) First, the optimal process window is preliminarily determined in the global process window, and customized experimental design is performed to determine the optimal process parameters and corresponding process data, so that the machine learning model has better fitting ability, thereby improving the accuracy, practicality and universality of the constructed database.
[0101] To summarize, the system and method provided in the embodiments of the present application are connected to a computer device through an interface module, and configure test requirements, and set a corresponding chip connection module to be tested, a test selection module, and a display module, thereby improving the flexibility of integrated circuit testing and meeting the needs of various integrated circuit tests.
[0102] The above are only optional embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for determining process data based on a sand mold additive manufacturing integrated system, characterized in that: The method is applied in a computer device, and the method comprises: Acquiring process parameter data, wherein the process parameter data is parameter data associated with a sand mold manufacturing process; Inputting the process parameter data into a process parameter matching model and outputting a process parameter data matching result; In response to the process parameter data matching result indicating that there is no similar matching result, the process parameter data is input into a process parameter optimization model, and a process parameter optimization result is output.
2. The method according to claim 1, characterized in that The process parameter data includes process parameters and process characteristic data; The process parameters include at least one of the content of solvents such as resin / curing agent, thickness of additive manufacturing layer, width of additive manufacturing, and type and mesh size of molding sand; The process characteristic data includes at least one of interlayer thermal coupling characteristic data, adaptive slicing and stratification characteristic data, multi-region precise sand spreading characteristic data, sand spreading vibration compaction characteristic data, and sand spreading rolling compaction characteristic data.
3. The method according to claim 1, characterized in that The step of inputting the process parameter data into a process parameter matching model and outputting a process parameter data matching result comprises: Inputting the process parameter data into a process parameter matching model; Calling a process parameter database through the process parameter matching model to perform similarity matching on the process parameter data, wherein the process parameter database includes at least one set of ideal process parameters; In response to the process parameter data and the ideal process data in the process parameter database meeting a matching rule, outputting a first matching result, wherein the first matching result is a correction result of the process parameter data corresponding to the ideal process parameter; In response to the process parameter data not meeting the matching rule with the ideal process data in the process parameter database, a second matching result is output, wherein the second matching result indicates that there is no similar matching result.
4. The method according to claim 3, characterized in that The method further comprises: Acquire sample process parameter data, wherein the sample process parameter data is process data based on experiments and physical property parameters; The process parameter database is constructed based on the sample process parameter data.
5. The method according to claim 4, characterized in that The sample process parameter data are configured with sample process parameter experimental values, and the sample process parameter experimental values are used to characterize experimental results of the sample process parameter data.
6. The method according to claim 3, characterized in that The process parameter optimization model is a machine learning model based on artificial intelligence; In response to the process parameter data matching result indicating that there is no similar matching result, inputting the process parameter data into a process parameter optimization model and outputting a process parameter optimization result, including: In response to the process parameter data matching result indicating that there is no similar matching result, inputting the process parameter data and the similarity matching result into the process parameter optimization model; Determining adjustment parameter data corresponding to the process parameter data based on the similarity matching result by using the process parameter optimization model; In combination with the adjustment parameter data, the process parameter optimization result is generated through the process parameter optimization model.
7. The method according to claim 6, characterized in that The parameter adjustment data is implemented in the form of objective function difference.
8. The method according to claim 7, characterized in that The step of combining the adjustment parameter data and generating the process parameter optimization result through the process parameter optimization model includes: Determining new value acceptance criteria and iteration rules in combination with the parameter adjustment data; Based on the new value acceptance criterion and the iteration rule, in combination with the adjustment parameter data, the process parameter optimization result is generated through the process parameter optimization model.
9. The method according to claim 3, characterized in that: After the process parameter optimization result is obtained, the output includes: The process parameter optimization result is imported into the process parameter database.
10. The method according to claim 9, characterized in that The method further comprises: In response to receiving the discard instruction, the ideal process parameters in the process parameter database are discarded.