A powder pressing process parameter optimization method based on digital twinning

Abstract

The application discloses a powder pressing process parameter optimization method based on digital twinning, which comprises the following steps: constructing a powder pressing process digital twinning model construction and twin data generation module, which constructs a powder pressing process digital twinning model based on filling parameters, material parameters and process parameters, and generates twin data; constructing a quality parameter and energy consumption parameter prediction model construction module, which realizes twin data expansion based on the twin data expansion network, and trains the quality parameter and energy consumption parameter prediction model based on the expanded data; and constructing a process parameter optimization module, which optimizes the process parameters through a particle swarm optimization algorithm based on the quality parameter and energy consumption parameter prediction model. The application can solve the problem that the process parameters are difficult to be accurately optimized under the condition that the powder pressing process data is insufficient, effectively improves the quality of the finished powder column while reducing the energy consumption of the pressing process.

Description

Technical Field

[0001] This invention belongs to the field of intelligent manufacturing, specifically relating to a method for optimizing process parameters in powder pressing based on digital twins. Background Technology

[0002] Powder pressing is a widely used and crucial production process in industries such as metallurgy, pharmaceuticals, and ceramics. The powder pressing process involves numerous parameters, including process parameters, material parameters, loading parameters, quality parameters, and energy consumption parameters. When producing a specific product, the material and loading parameters are largely fixed, while the process parameters can be adjusted. Quality and energy consumption parameters are determined by a combination of these parameters. Key process parameters include holding pressure, holding time, slider speed during the pressurization phase, and slider speed during the depressurization phase. While holding pressure and holding time can be set and modified, the slider speeds during the pressurization and depressurization phases are determined by the hydraulic press's performance and cannot be changed. Therefore, to produce higher-quality powder columns with lower energy consumption, optimization of the holding pressure and holding time is essential.

[0003] However, optimizing the holding pressure and holding time requires numerous powder pressing experiments, resulting in significant waste of raw materials, energy consumption, and time. Furthermore, some powder materials pose inherent risks, and hastily conducting powder pressing experiments could lead to safety accidents, such as with explosive powders or toxic powders. This makes it difficult to conduct large-scale powder pressing experiments, resulting in a lack of data on the powder pressing process. Therefore, this invention proposes a powder pressing process parameter optimization method based on digital twins. This method solves the problem of difficulty in accurately optimizing process parameters when data is lacking due to experimental difficulties, effectively improving the quality of the finished powder column while reducing production energy consumption. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method for optimizing process parameters in powder pressing based on digital twins. This method solves the problem of difficulty in accurately optimizing process parameters when data is lacking due to experimental difficulties in the powder pressing process, thereby effectively improving the quality of finished powder columns while reducing production energy consumption.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A method for optimizing process parameters in powder pressing based on digital twins includes the following steps:

[0007] Step 1: Based on the geometric-physical-behavioral-rule-dimensional digital twin model construction method, the digital twin model of the powder pressing process is constructed and twin data is generated. This includes the following steps:

[0008] (1-1) Constructing the geometric model of powder and mold: Based on the filling parameters including the inner diameter of the mold, the initial height of the powder, and the geometric positional relationship between the powder and the mold, construct a geometric model of powder and mold oriented towards the geometric structure;

[0009] (1-2) Constructing a physical model of powder deformation: Based on the material parameters of the powder, including yield strength, density, initial loose relative density, Young's modulus when the relative density is 1, Poisson's ratio when the relative density is 1, parameter γ in the Shima-Oyane model, parameter β in the Shima-Oyane model, and the friction coefficient between the powder and the inner wall of the mold, considering the interaction between the powder and the mold, the material parameters of the powder are added in the MSC Marc finite element simulation analysis software, and the contact relationship between the powder and the mold is configured, thereby constructing a physical model of powder deformation that is oriented towards the change of physical quantities in the powder deformation process;

[0010] (1-3) Constructing powder pressing behavior models: Based on the pressing stage, holding stage, depressurization stage, and springback stage of the powder pressing process, as well as process parameters including holding pressure and holding time, establish pressing behavior models, holding behavior models, depressurization behavior models, and springback behavior models. Based on the above four behavior models, construct a powder pressing behavior model that accurately reflects the behavior of each stage.

[0011] (1-4) Constructing a powder pressing rule model: Based on the variation law of holding pressure during powder pressing, construct the maximum holding pressure rule; based on the experience of setting the holding time during powder pressing, construct the maximum holding time rule; based on the variation range of the average relative density of the powder column during powder pressing, construct the maximum average relative density of the powder column rule; based on the duration of the rebound phenomenon of the powder column in the rebound stage during powder pressing, construct the rebound time setting rule; and then, based on the maximum holding pressure rule, the maximum holding time rule, the maximum average relative density of the powder column rule, and the rebound time setting rule, construct a powder pressing rule model oriented towards the constraints of the pressing process.

[0012] (1-5) Constructing a digital twin model of the powder pressing process: Based on the powder and mold geometric model, powder deformation physical model, powder pressing behavior model and powder pressing rule model in steps (1-1) to (1-4), construct a digital twin model of the powder pressing process;

[0013] (1-6) Generate powder pressing process twin data: Run the powder pressing process digital twin model constructed in step (1-5), and calculate the data obtained from the model operation to obtain the mass parameters of the finished powder column, including the average relative density, the variance of the relative density of the finished powder column, and the maximum average relative density of the powder column during the powder pressing process, as well as the energy consumption parameters including the maximum power of the upper die punch and the work done by the upper die punch. Based on the above mass parameters, energy consumption parameters and the filling parameters, material parameters and process parameters used in steps (1-1) to (1-3), generate powder pressing process twin data.

[0014] Step 2: Based on the twin data of the powder pressing process generated in steps (1-5), construct a prediction model for quality parameters and energy consumption parameters, which specifically includes the following steps:

[0015] (2-1) Training the twin data augmentation network: Using random noise and the twin data generated in step (1-5), train a generative adversarial network that includes a data generator and a data discriminator as the twin data augmentation network, where the data generator is used for twin data augmentation;

[0016] (2-2) Expanding the twin data of the powder pressing process: Based on the data generator of the twin data expansion network generated in step (2-1), the powder pressing process expansion data is generated using random noise;

[0017] (2-3) Constructing parameter prediction models: Based on twin data and augmented data of powder pressing process, train convolutional neural network as a prediction model for quality parameters and energy consumption parameters, complete the construction of quality parameter and energy consumption parameter prediction models, and realize the prediction of quality parameters and energy consumption parameters based on process parameters, material parameters and filling parameters;

[0018] Step 3: Based on the parameter prediction model constructed in Step 2, optimize the process parameters, specifically including the following steps:

[0019] (3-1) Set the boundary conditions for the optimization process: Set the process parameter boundary conditions and quality parameter boundary conditions based on the maximum holding pressure rule, the maximum holding time rule, and the maximum average relative density of the powder column rule, and use these as the boundary conditions for the optimization process.

[0020] (3-2) Optimize process parameters in multiple rounds: Based on the particle swarm optimization algorithm, with the expected value of the quality parameter as the optimization objective, and the prediction model of quality parameter and energy consumption parameter constructed in step 2 as the function to be optimized, considering the boundary conditions of the optimization process set in step (3-1), the process parameters are optimized in multiple rounds, and multiple sets of optimized values ​​of process parameters including holding pressure and holding time are obtained.

[0021] (3-3) Select the optimal process parameters: Input the optimized values ​​of multiple sets of process parameters obtained in step (3-2) and the material parameters and filling parameters used in step 1 to construct the digital powder pressing process twin model into the parameter prediction model constructed in step (2-3), calculate the optimized values ​​of process parameters that generate the minimum upper die punch work, and apply them.

[0022] Furthermore, in step (1-3), the pressurization behavior is the process in which the upper die punch of the mold moves downward at a constant speed along the longitudinal axis of symmetry of the mold from the initial pressurization position until the pressure reaches the holding pressure. The movement speed of the upper die punch is set as the average movement speed of the hydraulic press slider during the pressurization stage of the powder pressing process.

[0023] Furthermore, in steps (1-3), the pressure relief behavior is the process in which the upper die punch of the mold moves upward at a constant speed along the longitudinal axis of symmetry of the mold from the end position of pressure holding until the upper die punch reaches the initial position of pressure application, wherein the movement speed of the upper die punch is set as the average movement speed of the hydraulic press slider during the pressure relief phase of the pressing process.

[0024] Furthermore, in steps (1-3), the springback behavior is that the mold at t h The process by which the inner upper die punch remains stationary after pressure relief and the powder column springs back, t h Let t be the rebound time, and t be the rebound time. h The settings should meet the rebound time setting rules constructed in steps (1-4).

[0025] Furthermore, in steps (1-4), the maximum holding pressure rule is P. b ≤P bmax P b To maintain pressure, P bmax The maximum holding pressure is defined as t; the maximum holding time is defined as t. b ≤t bmax , t b For the pressure holding time, t bmax The maximum holding time; the maximum average relative density of the powder column is ρ. p ≤ρ pmax , ρ p ρ represents the average relative density of the maximum powder column during the powder pressing process. pmax The upper limit of the average relative density of the maximum powder column during the powder pressing process is defined by the springback time setting rule as t. h ≥t hmin , t h For the rebound time, t hmin This is the minimum set value for the rebound time.

[0026] Furthermore, in step (3-2), the expected values ​​of the quality parameters include the expected value of the average relative density of the finished powder column. The expected value σ of the relative density variance of the finished powder column e ,in Selected based on actual circumstances, σ e =0.

[0027] The advantages of this invention compared to the prior art are:

[0028] (1) Currently, the quality parameters of the powder pressing process under different process parameters, filling parameters and material parameters are mainly obtained through the powder pressing process. However, this method has low data acquisition efficiency, making it difficult to obtain sufficient data. The present invention realizes the construction of digital twin model of powder pressing process and the generation of twin data, and uses the generated twin data to realize the expansion of twin data, thus solving the problem of lack of data in powder pressing process.

[0029] (2) Traditional powder pressing processes are difficult to optimize with precise process parameters. The setting of process parameters depends on the worker's judgment of the production situation, which hinders the improvement of product quality. However, this invention optimizes process parameters based on the constructed quality parameter prediction model, taking into account the holding pressure, holding time, and the average relative density of the maximum powder column during the powder pressing process. The optimized process parameter value corresponding to the minimum upper die punch work prediction value is selected for application. In this way, the problem of difficulty in accurately optimizing process parameters is solved on the basis of comprehensively considering production safety and production energy consumption. Attached Figure Description

[0030] Figure 1 This is a flowchart of a method for optimizing process parameters in powder pressing based on digital twins according to the present invention.

[0031] Figure 2 This is a flowchart illustrating the training process of the twin data augmentation network of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0033] This invention relates to a method for optimizing process parameters in powder pressing based on digital twins. This method addresses the challenge of accurately optimizing process parameters in powder pressing processes where data is scarce due to experimental difficulties, effectively improving the quality of finished powder columns while reducing production energy consumption. For powder pressing processes of different powders, the proposed method can efficiently and accurately optimize process parameters for powder pressing.

[0034] like Figure 1 As shown, the method for optimizing process parameters of powder pressing based on digital twins according to the present invention specifically includes the following steps:

[0035] Step (1): Based on the geometric-physical-behavioral-rule dimension digital twin model construction method, the digital twin model of the powder pressing process is constructed and the twin data is generated. The specific implementation is as follows:

[0036] (1-1) Construction of the geometric model of powder and mold: Based on the filling parameters such as the inner diameter of the mold and the initial height of the powder, as well as the geometric positional relationship between the powder and the mold, a geometric model of powder and mold oriented towards the geometric structure is constructed.

[0037] (1-2) Construction of the physical model of powder deformation: Based on the material parameters such as the yield strength, density, initial loose relative density, Young's modulus when the relative density is 1, Poisson's ratio when the relative density is 1, parameter γ in the Shima-Oyane model, parameter β in the Shima-Oyane model, and friction coefficient between the powder and the inner wall of the mold, the interaction relationship between the powder and the mold is considered. The material parameters of the powder are added in the MSC Marc finite element simulation analysis software, and the contact relationship between the powder and the mold is configured, so as to construct a physical model of powder deformation that is oriented towards the change of physical quantities in the powder deformation process.

[0038] (1-3) Construction of powder pressing behavior model: Based on the pressing stage, holding stage, depressurization stage, springback stage and process parameters such as holding pressure and holding time in the powder pressing process, pressurization behavior model, holding behavior model, depressurization behavior model and springback behavior model are established. Based on the above four behavior models, a powder pressing behavior model that accurately reflects the behavior of each stage is constructed.

[0039] The pressurization behavior refers to the process in which the upper die punch moves downward at a constant speed along the longitudinal axis of symmetry of the mold from the initial pressurization position until the pressure reaches the holding pressure. The movement speed of the upper die punch is set as the average movement speed of the hydraulic press slider during the pressurization stage of the powder pressing process.

[0040] The pressure relief behavior refers to the process in which the upper die punch moves upward at a constant speed along the longitudinal axis of symmetry of the mold from the end of the pressure holding position until the upper die punch reaches the initial pressure position. The movement speed of the upper die punch is set as the average movement speed of the hydraulic press slider during the pressure relief phase of the pressing process.

[0041] The springback behavior refers to the mold's springback behavior at t h The process by which the inner upper die punch remains stationary after pressure relief and the powder column springs back, t h Let t be the rebound time, and t be the rebound time. h The settings should meet the rebound time setting rules constructed in steps (1-4).

[0042] (1-4) Construction of powder pressing rule model: Based on the change law of holding pressure during powder pressing, the maximum holding pressure rule is constructed; based on the experience of setting the holding time during powder pressing, the maximum holding time rule is constructed; based on the change range of the average relative density of the powder column during powder pressing, the maximum average relative density of the powder column rule is constructed; based on the duration of the rebound phenomenon of the powder column in the rebound stage during powder pressing, the rebound time setting rule is constructed. Then, based on the maximum holding pressure rule, the maximum holding time rule, the maximum average relative density of the powder column rule, and the rebound time setting rule, a powder pressing rule model oriented towards the constraints of the pressing process is constructed.

[0043] The maximum holding pressure rule is P b ≤P bmax P b To maintain pressure, P bmax To determine the maximum holding pressure, the maximum holding time is defined as t. b ≤t bmax , t b For the pressure holding time, t bmax The maximum holding time is determined by the rule that the average relative density of the maximum powder column is ρ. p ≤ρ pmax , ρ p ρ represents the average relative density of the maximum powder column during the powder pressing process. pmax The upper limit of the average relative density of the maximum powder column during the powder pressing process is defined by the springback time setting rule as t. h ≥t hmin , t h For the rebound time, t hmin This is the minimum set value for the rebound time.

[0044] (1-5) Construction of digital twin model of powder pressing process: Based on the powder and mold geometric model, powder deformation physical model, powder pressing behavior model and powder pressing rule model in steps (1-1) to (1-4), a digital twin model of powder pressing process is constructed.

[0045] (1-6) Generation of Twin Data for Powder Pressing Process: Run the digital twin model of powder pressing process constructed in step (15), and calculate the data obtained from the model operation to obtain quality parameters such as the average relative density of finished powder column, the variance of the relative density of finished powder column, and the maximum average relative density of powder column during powder pressing, as well as energy consumption parameters such as the maximum power of the upper die punch and the work done by the upper die punch. Based on the above quality parameters, energy consumption parameters and the filling parameters, material parameters and process parameters used in steps (1-1) to (1-3), generate twin data for powder pressing process.

[0046] Step (2): Based on the twin data of the powder pressing process generated in steps (1-5), construct a prediction model for quality parameters and energy consumption parameters, as specifically implemented below:

[0047] (2-1) Training the Siamese Data Augmentation Network: Using random noise and the Siamese data generated in step (1-5), a generative adversarial network encompassing a data generator and a data discriminator is trained as the Siamese data augmentation network. The data generator is used for Siamese data augmentation. The training process for the Siamese data augmentation network is as follows: Figure 2 As shown, the specific steps include the following:

[0048] ① Augmented data generation: Using m sets of random noise as input to the data generator, a single set of augmented data is generated, where m is any integer satisfying 1≤m≤M, and M is the number of sets of twin data;

[0049] ② Generation of discrimination results based on augmented data and twin data: Input any m sets of twin data and m sets of augmented data generated in step ① into the data discriminator to obtain 2m sets of discrimination results;

[0050] ③ Data discriminator training: Based on the discrimination results obtained in step ②, train the data discriminator according to the backpropagation algorithm;

[0051] ④ Expanded data regeneration: Using m sets of random noise different from those in step ① as input to the data generator, m sets of expanded data are generated again;

[0052] ⑤ Generation of discrimination results based on expanded data: Take the m sets of expanded data generated in step ④ as the input of the data discriminator to obtain m sets of discrimination results;

[0053] ⑥ Data generator training: Based on the discrimination results obtained in step ⑤, train the data generator according to the backpropagation algorithm;

[0054] ⑦ If the absolute error of the 2m groups of discrimination results obtained in step ② and the m groups of discrimination results obtained in step ⑤ and 0.5 is less than 0.01, then the training of the Siamese augmented network is completed; otherwise, repeat steps ① to ⑥.

[0055] (2-2) Powder pressing process twin data augmentation: Based on the data generator of the twin data augmentation network generated in step (2-1), the powder pressing process augmentation data is generated using random noise;

[0056] (2-3) Construction of parameter prediction model: Based on twin data and extended data of powder pressing process, a convolutional neural network is trained as a prediction model for quality parameters and energy consumption parameters. The construction of the prediction model for quality parameters and energy consumption parameters is completed, and the prediction of quality parameters and energy consumption parameters based on process parameters, material parameters and filling parameters is realized.

[0057] Step (3): Based on the parameter prediction model constructed in step (2), the process parameters are optimized, specifically including the following steps:

[0058] (3-1) Optimize process boundary condition setting: Based on the maximum holding pressure rule, the maximum holding time rule, and the maximum average relative density of powder column rule, set the process parameter boundary condition as holding pressure P. b ≤P bmax And the pressure holding time t b ≤t bmax The quality parameter boundary condition is set as the maximum average relative density of the powder column ρ during the powder pressing process. p ≤ρ pmax Based on production experience, let P bmax =100MPa, t bmax =10s, ρ pmax =1.5, using process parameter boundary conditions and quality parameter boundary conditions as the optimization process boundary conditions;

[0059] (3-2) Multi-round process parameter optimization: Based on the particle swarm optimization algorithm, the expected value of the average relative density of the finished powder column in the quality parameter is used. The expected value σ of the relative density variance of the finished powder column e To optimize the objective, the prediction model for quality parameters and energy consumption parameters constructed in step (2) is used as the function to be optimized. Considering the boundary conditions of the optimization process set in step (3-1), the process parameters are optimized in multiple rounds, and multiple sets of optimized process parameter values ​​are obtained. The expected values ​​of the quality parameters include the expected value of the average relative density of the finished powder column. Expected value σ of the relative density variance of the finished powder column e ,in Selected based on actual circumstances, σ e =0, each set of optimized process parameters includes optimized holding pressure and optimized holding time.

[0060] (3-3) Optimal process parameter selection: Input the multiple sets of optimized process parameter values ​​obtained in step (3-2) into the prediction models of quality parameters and energy consumption parameters constructed in step (2-3) to obtain the predicted value of upper die punch work corresponding to each set of optimized process parameter values. Select the optimized process parameter value corresponding to the minimum predicted value of upper die punch work and apply it.

[0061] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

[0062] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for optimizing process parameters in powder pressing based on digital twins, characterized in that, Includes the following steps: Step 1: Based on the geometric-physical-behavioral-rule-dimensional digital twin model construction method, the digital twin model of the powder pressing process is constructed and twin data is generated. This includes the following steps: (1-1) Constructing the geometric model of powder and mold: Based on the filling parameters including the inner diameter of the mold, the initial height of the powder, and the geometric positional relationship between the powder and the mold, construct a geometric model of powder and mold oriented towards the geometric structure; (1-2) Constructing a physical model of powder deformation: Based on the material parameters of the powder, including yield strength, density, initial loose relative density, Young's modulus when the relative density is 1, Poisson's ratio when the relative density is 1, parameter γ in the Shima-Oyane model, parameter β in the Shima-Oyane model, and the friction coefficient between the powder and the inner wall of the mold, considering the interaction between the powder and the mold, the material parameters of the powder are added in the finite element simulation analysis software, and the contact relationship between the powder and the mold is configured, thereby constructing a physical model of powder deformation that is oriented towards the change of physical quantities in the powder deformation process; (1-3) Constructing powder pressing behavior models: Based on the pressing stage, holding stage, depressurization stage, and springback stage of the powder pressing process, as well as process parameters including holding pressure and holding time, establish pressing behavior models, holding behavior models, depressurization behavior models, and springback behavior models. Based on the above four behavior models, construct a powder pressing behavior model that accurately reflects the behavior of each stage. (1-4) Constructing a powder pressing rule model: Based on the variation law of holding pressure during powder pressing, construct the maximum holding pressure rule; based on the experience of setting the holding time during powder pressing, construct the maximum holding time rule; based on the variation range of the average relative density of the powder column during powder pressing, construct the maximum average relative density of the powder column rule; based on the duration of the rebound phenomenon of the powder column in the rebound stage during powder pressing, construct the rebound time setting rule; and then, based on the maximum holding pressure rule, the maximum holding time rule, the maximum average relative density of the powder column rule, and the rebound time setting rule, construct a powder pressing rule model oriented towards the constraints of the pressing process. (1-5) Constructing a digital twin model of the powder pressing process: Based on the powder and mold geometric model, powder deformation physical model, powder pressing behavior model and powder pressing rule model in steps (1-1) to (1-4), construct a digital twin model of the powder pressing process; (1-6) Generate powder pressing process twin data: Run the powder pressing process digital twin model constructed in step (1-5), and calculate the data obtained from the model operation to obtain the mass parameters of the finished powder column, including the average relative density, the variance of the relative density of the finished powder column, and the maximum average relative density of the powder column during the powder pressing process, as well as the energy consumption parameters including the maximum power of the upper die punch and the work done by the upper die punch. Based on the above mass parameters, energy consumption parameters and the filling parameters, material parameters and process parameters used in steps (1-1) to (1-3), generate powder pressing process twin data. Step 2: Based on the twin data of the powder pressing process generated in steps (1-5), construct a prediction model for quality parameters and energy consumption parameters, which specifically includes the following steps: (2-1) Training the twin data augmentation network: Using random noise and the twin data generated in step (1-5), train a generative adversarial network that includes a data generator and a data discriminator as the twin data augmentation network, where the data generator is used for twin data augmentation; (2-2) Expanding the twin data of the powder pressing process: Based on the data generator of the twin data expansion network generated in step (2-1), the powder pressing process expansion data is generated using random noise; (2-3) Constructing parameter prediction models: Based on twin data and augmented data of powder pressing process, train convolutional neural network as a prediction model for quality parameters and energy consumption parameters, complete the construction of quality parameter and energy consumption parameter prediction models, and realize the prediction of quality parameters and energy consumption parameters based on process parameters, material parameters and filling parameters; Step 3: Based on the parameter prediction model constructed in Step 2, optimize the process parameters, specifically including the following steps: (3-1) Set the boundary conditions for the optimization process: Set the process parameter boundary conditions and quality parameter boundary conditions based on the maximum holding pressure rule, the maximum holding time rule, and the maximum average relative density of the powder column rule, and use these as the boundary conditions for the optimization process. (3-2) Optimize process parameters in multiple rounds: Based on the particle swarm optimization algorithm, with the expected value of the quality parameter as the optimization objective, and the prediction model of quality parameter and energy consumption parameter constructed in step 2 as the function to be optimized, considering the boundary conditions of the optimization process set in step (3-1), the process parameters are optimized in multiple rounds, and multiple sets of optimized values ​​of process parameters including holding pressure and holding time are obtained. (3-3) Select the optimal process parameters: Input the optimized values ​​of multiple sets of process parameters obtained in step (3-2) and the material parameters and filling parameters used in step 1 to construct the digital powder pressing process twin model into the parameter prediction model constructed in step (2-3), calculate the optimized values ​​of process parameters that generate the minimum upper die punch work, and apply them.

2. The method for optimizing process parameters of powder pressing based on digital twins according to claim 1, characterized in that: In steps (1-3), the pressurization behavior is the process in which the upper die punch of the mold moves downward at a constant speed along the longitudinal axis of symmetry of the mold from the initial pressurization position until the pressure reaches the holding pressure. The movement speed of the upper die punch is set as the average movement speed of the hydraulic press slider during the pressurization stage of the powder pressing process.

3. The method for optimizing process parameters of powder pressing based on digital twins according to claim 1, characterized in that: In steps (1-3), the pressure relief behavior is the process in which the upper die punch of the mold moves upward at a constant speed along the longitudinal axis of symmetry of the mold from the end position of pressure holding until the upper die punch reaches the initial position of pressure application. The movement speed of the upper die punch is set as the average movement speed of the hydraulic press slider during the pressure relief phase of the pressing process.

4. The method for optimizing process parameters of powder pressing based on digital twins according to claim 1, characterized in that: In steps (1-3), the springback behavior is that the mold at t h The process by which the inner upper die punch remains stationary after pressure relief and the powder column springs back, t h Let t be the rebound time, and t h The settings should meet the rebound time setting rules constructed in steps (1-4).

5. The method for optimizing process parameters of powder pressing based on digital twins according to claim 1, characterized in that: In steps (1-4), the maximum holding pressure rule is P. b ≤P bmax P b To maintain pressure, P bmax The maximum holding pressure is defined as t; the maximum holding time is defined as t. b ≤t bmax , t b For the pressure holding time, t bmax The maximum holding time; the maximum average relative density of the powder column is ρ. p ≤ρ pmax , ρ p ρ represents the average relative density of the maximum powder column during the powder pressing process. pmax The upper limit of the average relative density of the maximum powder column during the powder pressing process is defined by the springback time setting rule as t. h ≥t hmin , t h For the rebound time, t hmin This is the minimum set value for the rebound time.

6. The method for optimizing process parameters of powder pressing based on digital twins according to claim 1, characterized in that: In step (3-2), the expected values ​​of the quality parameters include the expected value of the average relative density of the finished powder column. The expected value σ of the relative density variance of the finished powder column e ,in Selected based on actual circumstances, σ e =0.

Images