Ultrasonic impact treatment process parameter optimization method and device and electronic equipment

Abstract

The invention relates to the technical field of welding processes, in particular to an ultrasonic impact treatment process parameter optimization method and device and electronic equipment.The method comprises the steps that welding simulation data of a welding part and multiple sets of process parameters of an ultrasonic impact treatment process are obtained, the process parameters comprise cooling temperature, impact frequency, impact amplitude, impact needle diameter and impact time; performing finite element simulation on each group of process parameters based on the welding simulation data to obtain residual stress and pressure stress depth data at the weld joint under different process parameters; and utilizing the trained ultrasonic peening treatment prediction model to predict the residual stress and pressure stress depth data at the weld joint, and optimizing at least one group of process parameters of the ultrasonic peening treatment process according to the residual stress and pressure stress depth data at the weld joint. Therefore, the problems that process parameters of the ultrasonic treatment process are not comprehensively optimized, and the influence of the cooling temperature on the residual stress of the welding part after ultrasonic impact treatment is neglected in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of welding processes, and particularly relates to a method and device for optimizing process parameters of ultrasonic impact treatment and an electronic device. Background Art

[0002] The vacuum chamber is a key component of a nuclear fusion reactor, which provides a safe environment for maintaining the operation of high-temperature plasma. Due to the large size of the vacuum chamber, it is usually processed in modules and then assembled during the manufacturing process. A large number of welding processes are required during the assembly process. However, during the welding process using the welding process, large residual stresses and deformations will be generated at the welding joints, which will significantly affect the fatigue strength and service life of the vacuum chamber material. As a surface strengthening technology, ultrasonic impact treatment can be applied during the manufacturing and assembly process of the vacuum chamber to reduce the residual stress at the welding joints.

[0003] Most of the equipment process parameters (impact frequency, impact amplitude, impact needle diameter) of the ultrasonic impact treatment equipment currently produced and sold on the market are factory settings and cannot be adjusted, and the process parameters during the ultrasonic impact treatment process are also less concerned.

[0004] In the related art, there are mainly prediction methods for residual stresses under different ultrasonic impact treatment process parameters based on machine learning, as well as parameter optimization methods for individual parameters. However, the consideration of process parameters is not comprehensive enough, resulting in the difficulty of fully exploring the process potential of the optimized process parameters and obtaining the best combination of processing parameters. Furthermore, the ultrasonic impact treatment process cannot achieve the optimal effect in reducing residual stresses and improving component performance. Summary of the Invention

[0005] The present invention provides a method and device for optimizing process parameters of ultrasonic impact treatment and an electronic device to solve the problems in the related art that the optimization of process parameters for the ultrasonic treatment process is not comprehensive enough and the influence of the cooling temperature on the residual stress of the welded parts after ultrasonic impact treatment is ignored.

[0006] An embodiment of the first aspect of the present invention provides a method for optimizing ultrasonic impact treatment process parameters, including the following steps: obtaining welding simulation data of a welded part and multiple groups of process parameters of the ultrasonic impact treatment process, where the process parameters include cooling temperature, impact frequency, impact amplitude, impact needle diameter, and impact time; performing finite element simulation on each group of process parameters based on the welding simulation data to obtain residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, and generating a training dataset according to the multiple groups of process parameters, welding simulation data, and residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; training a pre-constructed ultrasonic impact treatment prediction model using the training dataset, where the input of the ultrasonic impact treatment prediction model is multiple groups of process parameters and welding simulation data, and the output of the ultrasonic impact treatment prediction model is residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; predicting the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment using the trained ultrasonic impact treatment prediction model, and optimizing at least one group of process parameters of the ultrasonic impact treatment process according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment.

[0007] Optionally, the ultrasonic impact treatment prediction model includes an input layer, multiple hidden layers, an output layer, and a loss function. Among them, the input layer is used to input process parameters and welding simulation data; the multiple hidden layers are used to perform feature extraction and non-linear transformation on the process parameters and welding simulation data; the output layer is used to output the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; the loss function is used to calculate the prediction error according to the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment and the true values in the training dataset, and update the model parameters of the ultrasonic impact treatment prediction model according to the prediction error until the loss function is minimized.

[0008] Optionally, obtaining the welding simulation data of the welded part and multiple groups of process parameters of the ultrasonic impact treatment process includes: constructing a welding finite element model; performing multi-layer and multi-pass welding simulation on the welded part using the welded part finite element model to obtain the welding simulation data of the welded part.

[0009] Optionally, the welding simulation data includes: a post-welding deformation model, the post-welding temperature field and the post-welding residual stress field before ultrasonic impact treatment. Performing finite element simulation on each group of process parameters based on the welding simulation data to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters includes: constructing an ultrasonic impact treatment finite element model based on the post-welding deformation model, and loading the post-welding temperature field and the post-welding residual stress field before ultrasonic impact treatment as predefined fields into the ultrasonic impact treatment finite element model, and performing ultrasonic impact simulation on the weld under different process parameters using the ultrasonic impact treatment finite element model to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters.

[0010] Optionally, a multi-layer multi-pass welding simulation is performed on the welded part using a finite element model of the welded part to obtain welding simulation data of the welded part, including: performing failure treatment on the weld area in the welded part so that the weld area does not participate in the finite element analysis; sequentially activating each weld to participate in the finite element analysis, and performing a thermal analysis simulation on the current weld participating in the finite element analysis to obtain the temperature field distribution data of the current weld; using the temperature field distribution data of the current weld as the initial condition, performing a force analysis on the current weld to obtain the corresponding stress field distribution and deformation data until all welds are analyzed to obtain the welding simulation data of the welded part.

[0011] Optionally, an ultrasonic impact simulation is performed on the weld under different process parameters using a finite element model of ultrasonic impact treatment to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, including: cooling the welded part to the cooling temperature in the process parameter combination through the external temperature of the heat exchange condition, and setting the external temperature as the target temperature for temperature dynamics coupling simulation; setting impact pins with different diameters, applying a displacement load perpendicular to the weld surface and controlled by the amplitude to the impact pins, and setting the impact frequency and impact amplitude in the amplitude; setting the time for ultrasonic impact analysis, applying a velocity load parallel to the weld direction to the impact pins to ensure that the weld is impacted within the set time to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters.

[0012] Optionally, at least one set of process parameters of the ultrasonic impact treatment process is optimized according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment, including: setting the optimization target of the ultrasonic impact treatment process parameters, inputting the optimization target, at least one set of ultrasonic impact treatment process parameters and the corresponding residual stress and compressive stress depth data at the weld after ultrasonic impact treatment into the target algorithm; using the target algorithm for optimization and optimizing at least one set of process parameters of the ultrasonic impact treatment process according to the optimization result.

[0013] Optionally, before training a pre-constructed ultrasonic impact treatment prediction model using a training data set, it further includes: determining the perturbation amplitude of the training data in the training data set based on the prediction error; generating new training data based on the perturbation amplitude, expanding the training data set based on the new training data; training the pre-constructed ultrasonic impact treatment prediction model using the expanded training data set.

[0014] In the second aspect of the present invention, an embodiment provides an ultrasonic impact treatment process parameter optimization device, including: an acquisition module, configured to acquire welding simulation data and multiple sets of process parameters, including: acquiring welding simulation data of a welded part and multiple sets of process parameters of ultrasonic impact treatment process, where the process parameters include cooling temperature, impact frequency, impact amplitude, impact needle diameter, and impact time; a simulation module, configured to perform finite element simulation on each set of process parameters, including: performing finite element simulation on each set of process parameters based on the welding simulation data to obtain residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, and generating a training data set according to the multiple sets of process parameters, welding simulation data, and residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; a training module, configured to train an ultrasonic impact treatment prediction model, including: training a pre-constructed ultrasonic impact treatment prediction model using the training data set, where the input of the ultrasonic impact treatment prediction model is multiple sets of process parameters and welding simulation data, and the output of the ultrasonic impact treatment prediction model is residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; an optimization module, configured to perform finite element simulation on each set of process parameters, including: predicting residual stress and compressive stress depth data at the weld after ultrasonic impact treatment using the trained ultrasonic impact treatment prediction model, and optimizing at least one set of process parameters of the ultrasonic impact treatment process according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment.

[0015] Optionally, the ultrasonic impact treatment prediction model includes an input layer, multiple hidden layers, an output layer, and a loss function, where the input layer is configured to input process parameters and welding simulation data; the multiple hidden layers are configured to perform feature extraction and non-linear transformation on the process parameters and welding simulation data; the output layer is configured to output predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; the loss function is configured to calculate a prediction error according to the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment and the true values in the training data set, and update the model parameters of the ultrasonic impact treatment prediction model according to the prediction error until the loss function is minimized.

[0016] Optionally, the acquisition module is further configured to: construct a welding finite element model; perform multi-layer and multi-pass welding simulation on the welded part using the welded part finite element model to obtain welding simulation data of the welded part.

[0017] Optionally, the welding simulation data includes: a post-welding deformation model, the post-welding temperature field and the post-welding residual stress field before ultrasonic impact treatment. The simulation module is further configured to: construct a finite element model for ultrasonic impact treatment based on the post-welding deformation model, and load the post-welding temperature field and the post-welding residual stress field before ultrasonic impact treatment into the finite element model for ultrasonic impact treatment as predefined fields, and use the finite element model for ultrasonic impact treatment to perform ultrasonic impact simulation on the weld seam under different process parameters to obtain the residual stress and the depth data of the compressive stress at the weld seam after ultrasonic impact treatment under different process parameters.

[0018] Optionally, the acquisition module is further configured to: perform failure processing on the weld seam area in the welded part so that the weld seam area does not participate in the finite element analysis; sequentially activate each weld seam to participate in the finite element analysis, and perform thermal analysis simulation on the current weld seam participating in the finite element analysis to obtain the temperature field distribution data of the current weld seam; use the temperature field distribution data of the current weld seam as the initial condition, perform force analysis on the current weld seam to obtain the corresponding stress field distribution and deformation data until all weld seams are analyzed to obtain the welding simulation data of the welded part.

[0019] Optionally, the simulation module is further configured to: cool the welded part to the cooling temperature in the process parameter combination through the external temperature of the heat exchange condition, and set the external temperature as the target temperature to perform temperature dynamics coupling simulation; set impact needles with different diameters, apply a displacement load perpendicular to the weld seam surface and controlled by the amplitude to the impact needles, and set the impact frequency and the impact amplitude in the amplitude; set the time for ultrasonic impact analysis, apply a velocity load parallel to the weld seam direction to the impact needles to ensure that the weld seam is impacted within the set time to obtain the residual stress and the depth data of the compressive stress at the weld seam after ultrasonic impact treatment under different process parameters.

[0020] Optionally, the optimization module is further configured to: set the optimization objective of the ultrasonic impact treatment process parameters, input the optimization objective, at least one set of ultrasonic impact treatment process parameters and the corresponding residual stress and the depth data of the compressive stress at the weld seam after ultrasonic impact treatment into the target algorithm; use the target algorithm to perform optimization, and optimize at least one set of process parameters of the ultrasonic impact treatment process according to the optimization result.

[0021] Optionally, it further includes: a perturbation module, configured to determine the perturbation amplitude of the training data in the training data set based on the prediction error before training the pre-constructed ultrasonic impact treatment prediction model using the training data set; generate new training data based on the perturbation amplitude, expand the training data set based on the new training data; and train the pre-constructed ultrasonic impact treatment prediction model using the expanded training data set.

[0022] An embodiment of the third aspect of the present invention provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to perform the ultrasonic impact treatment process parameter optimization method as described in the above embodiment.

[0023] An embodiment of the fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program or instruction is stored, and the computer program or instruction is executed by a processor to perform the ultrasonic impact treatment process parameter optimization method as described in the above embodiment.

[0024] Therefore, the present invention has at least the following beneficial effects: The embodiment of the present invention can obtain the welding simulation data of the welded part and multiple sets of process parameters of the ultrasonic impact treatment process, and perform finite element simulation on each set of process parameters based on the welding simulation data to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters. A training data set is generated according to multiple sets of process parameters, welding simulation data, and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment. Comprehensively consider the influence of various process parameters on the residual stress and compressive stress depth of the weld after ultrasonic impact treatment, and use the training data set to train a pre-constructed ultrasonic impact treatment prediction model. Use the trained ultrasonic impact treatment prediction model to predict the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment, and optimize at least one set of process parameters of the ultrasonic impact treatment process according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment. By comprehensively considering the influence of various process parameters on the ultrasonic impact treatment process, especially the influence of the cooling temperature on the residual stress after the ultrasonic impact treatment process, the optimization of the process parameters of the ultrasonic impact treatment process is realized to minimize the residual stress, reduce the service risk of the vacuum chamber components, and can provide a reference for the parameter adjustment of the ultrasonic impact treatment equipment. Thus, the technical problems in the related art that the optimization of the process parameters of the ultrasonic treatment process is not comprehensive enough and the influence of the cooling temperature on the residual stress of the welded part after ultrasonic impact treatment is ignored are solved.

[0025] The additional aspects and advantages of the present invention will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and / or additional aspects and advantages of the present invention will become obvious and easy to understand from the following description of the embodiments in conjunction with the drawings, where: Figure 1 FIG. is a flowchart of an ultrasonic impact treatment process parameter optimization method according to an embodiment of the present invention; Figure 2Schematic diagram of the finite element model adopted for ultrasonic impact treatment simulation according to an embodiment of the present invention; Figure 3 Schematic diagram of the residual stress distribution before ultrasonic impact treatment according to an embodiment of the present invention; Figure 4 Schematic diagram of the residual stress distribution after ultrasonic impact treatment according to an embodiment of the present invention; Figure 5 Flow chart of adaptive training data augmentation according to an embodiment of the present invention; Figure 6 Flow chart of the NSGA-II algorithm according to an embodiment of the present invention; Figure 7 Execution diagram of the ultrasonic impact treatment process parameter optimization method according to a specific embodiment of the present invention; Figure 8 Example diagram of the ultrasonic impact treatment process parameter optimization device according to an embodiment of the present invention; Figure 9 Schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed implementation manners

[0027] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, in which the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.

[0028] Before describing the solution of the present invention, the research on the ultrasonic impact treatment process and its drawbacks in the related art will be introduced first.

[0029] 1. Most of the numerical simulations of ultrasonic impact treatment in the related art are carried out under ideal conditions at room temperature without considering the temperature-ultrasonic impact coupling simulation. However, the residual stress is closely related to the temperature and cooling rate of the weld, which is one of the influencing parameters that need to be studied and is likely to deviate greatly from the actual situation.

[0030] 2. The input parameters of the residual stress prediction model for the ultrasonic impact treatment process parameters in the related art are not comprehensive, and the output parameter is the single-objective optimization of the residual stress peak without considering other indicators.

[0031] 3. The residual stress prediction method in the related art is based on the training and calculation of a complex function model established, which may lead to overfitting and thus poor generalization ability.

[0032] 4. The relationship between residual stress and ultrasonic impact treatment process parameters is highly non-linear. Most of the ultrasonic impact treatment process parameter optimization algorithms in related technologies are based on traditional machine learning algorithms, with limited modeling ability for non-linear relationships and unable to capture all complex patterns in the data.

[0033] Therefore, the present invention provides an ultrasonic impact treatment process parameter optimization method. In this method, welding simulation data of a welded part and multiple groups of process parameters of the ultrasonic impact treatment process can be obtained, and finite element simulation is performed on each group of process parameters based on the welding simulation data to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters. A training data set is generated according to multiple groups of process parameters, welding simulation data, and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment. The influence of various process parameters on the residual stress and compressive stress depth of the weld after ultrasonic impact treatment is comprehensively considered, and a pre-constructed ultrasonic impact treatment prediction model is trained using the training data set. The trained ultrasonic impact treatment prediction model is used to predict the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment, and at least one group of process parameters of the ultrasonic impact treatment process is optimized according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment. By comprehensively considering the influence of various process parameters on the ultrasonic impact treatment process, especially the influence of the cooling temperature on the residual stress after the ultrasonic impact treatment process, the optimization of the process parameters of the ultrasonic impact treatment process is realized to minimize the residual stress, reduce the service risk of vacuum chamber components, and provide a reference for the parameter adjustment of ultrasonic impact treatment equipment.

[0034] Specifically, Figure 1 is a schematic flow chart of an ultrasonic impact treatment process parameter optimization method provided by an embodiment of the present invention.

[0035] As Figure 1 shown, the ultrasonic impact treatment process parameter optimization method includes the following steps: In step S101, welding simulation data of a welded part and multiple groups of process parameters of the ultrasonic impact treatment process are obtained.

[0036] Among them, the process parameters include cooling temperature, impact frequency, impact amplitude, impact needle diameter, impact time, etc.

[0037] In an embodiment of the present invention, obtaining the welding simulation data of a welded part and multiple groups of process parameters of the ultrasonic impact treatment process includes: constructing a welding finite element model; performing multi-layer and multi-pass welding simulation on the welded part using the welded part finite element model to obtain the welding simulation data of the welded part.

[0038] It can be understood that embodiments of the present invention can construct a welding finite element model, and use the welding part finite element model to perform multi-layer and multi-pass welding simulation on the welding part. The thermal-elastic-plastic finite element method can be used for multi-layer and multi-pass welding simulation to obtain the welding simulation data of the welding part.

[0039] In the embodiments of the present invention, the welding part finite element model is used to perform multi-layer and multi-pass welding simulation on the welding part to obtain the welding simulation data of the welding part, including: performing failure treatment on the weld area in the welding part so that the weld area does not participate in the finite element analysis; sequentially activating each weld to participate in the finite element analysis, and performing thermal analysis simulation on the current weld participating in the finite element analysis to obtain the temperature field distribution data of the current weld; using the temperature field distribution data of the current weld as the initial condition, performing force analysis on the current weld to obtain the corresponding stress field distribution and deformation data until all weld analyses are completed to obtain the welding simulation data of the welding part.

[0040] Specifically, after constructing the welding finite element model in the embodiments of the present invention, the thermal-elastic-plastic finite element method is used to perform sequential coupled thermo-mechanical analysis numerical simulation on the welding process. For multi-layer and multi-pass welding simulation, the specific process of sequential coupling is as follows: (1) Using the birth and death element method to perform "killing" treatment on all weld areas, that is, failure treatment, and not participating in the finite element analysis; (2) Activating the first weld to participate in the analysis, and performing thermal analysis simulation on the first weld to obtain the temperature field distribution; (3) Using the temperature field result of the first weld as the initial condition, performing force analysis on the first weld to obtain the stress field distribution and deformation; (4) Activating the second weld to participate in the analysis, considering the deformation and stress of the first weld, performing thermal analysis on the second weld to obtain the temperature field distribution of the second weld; (5) Using the temperature field result of the second weld as the condition, performing force analysis on the second weld to obtain the stress distribution and deformation; (6) And so on, sequentially completing the thermal analysis and force analysis of the subsequent welds, and saving the post-welding deformation model, post-welding temperature field and post-welding residual stress field of the welding part to the result file.

[0041] In step S102, based on the welding simulation data, finite element simulation is performed on each set of process parameters to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters. A training data set is generated according to multiple sets of process parameters, welding simulation data, and residual stress and compressive stress depth data at the weld after ultrasonic impact treatment.

[0042] Among them, the welding simulation data includes: a post-weld deformation model, the post-weld temperature field and the post-weld residual stress field before ultrasonic impact treatment, and the central point temperature of the weld temperature field, the maximum residual stress value at the weld, and the average residual stress value at the weld can be obtained based on the post-weld temperature field and the post-weld residual stress field before ultrasonic impact treatment; the residual stress data at the weld after ultrasonic impact treatment includes the maximum and average values of the residual stress at the weld.

[0043] It can be understood that the embodiments of the present invention can perform finite element simulation on each set of process parameters based on the welding simulation data to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, and generate a training dataset according to multiple sets of process parameters, welding simulation data, and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment, so as to facilitate the subsequent training of the ultrasonic impact treatment prediction model.

[0044] In the embodiments of the present invention, finite element simulation is performed on each set of process parameters based on the welding simulation data to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, including: constructing an ultrasonic impact treatment finite element model based on the post-weld deformation model, and loading the post-weld temperature field and the post-weld residual stress field before ultrasonic impact treatment as predefined fields into the ultrasonic impact treatment finite element model, and using the ultrasonic impact treatment finite element model to perform ultrasonic impact simulation on the weld under different process parameters to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters.

[0045] It can be understood that the embodiments of the present invention can construct an ultrasonic impact treatment finite element model based on the post-weld deformation model, and load the post-weld temperature field and the post-weld residual stress field before ultrasonic impact treatment as predefined fields into the ultrasonic impact treatment finite element model, and use the ultrasonic impact treatment finite element model to perform ultrasonic impact treatment simulation on the weld under different process parameters to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, so as to facilitate the subsequent training of the ultrasonic impact treatment prediction model, wherein, the schematic diagram of the ultrasonic impact treatment finite element model is as Figure 2 shown.

[0046] In an embodiment of the present invention, an ultrasonic impact treatment finite element model is used to simulate ultrasonic impact on a weld seam under different process parameters, and residual stress and compressive stress depth data at the weld seam after ultrasonic impact treatment under different process parameters are obtained, including: cooling the welded part to the cooling temperature in the process parameter combination through the external temperature under heat exchange conditions, and setting the external temperature to the target temperature for temperature dynamics coupling simulation; setting impact pins with different diameters, applying a displacement load perpendicular to the weld seam surface and controlled by amplitude to the impact pins, and setting the impact frequency and impact amplitude in the amplitude; setting the time for ultrasonic impact analysis, applying a velocity load parallel to the weld seam direction to the impact pins, and ensuring that the weld seam is impacted within the set time to obtain the residual stress and compressive stress depth data at the weld seam after ultrasonic impact treatment under different process parameters.

[0047] Specifically, in an embodiment of the present invention, an ultrasonic impact treatment finite element model can be used to simulate ultrasonic impact on a weld seam under different process parameters, and residual stress and compressive stress depth data at the weld seam after ultrasonic impact treatment under different process parameters are obtained. The simulation results of the residual stress at the weld seam before ultrasonic impact treatment are as Figure 3 shown, and the simulation results of the residual stress at the weld seam after ultrasonic impact treatment are as Figure 4 shown. During the simulation process, the control methods for each parameter are as follows: Cooling temperature: The welded part is cooled to the set cooling temperature in the process parameter combination through the external temperature under heat exchange conditions, and then the external temperature is set to room temperature for temperature - dynamics coupling simulation; Impact pin diameter: Modeling and assembling impact pins with different diameters respectively; Impact frequency, impact amplitude: Applying a displacement load perpendicular to the weld seam surface and controlled by amplitude to the impact pins, and setting the impact frequency and amplitude in the amplitude; Impact time: Setting the ultrasonic impact analysis step time, applying a velocity load parallel to the weld seam direction to the impact pins, and ensuring that the weld seam is impacted within the set time.

[0048] In step S103, a pre - constructed ultrasonic impact treatment prediction model is trained using a training data set. Among them, the input of the ultrasonic impact treatment prediction model is multiple sets of process parameters and welding simulation data, and the output of the ultrasonic impact treatment prediction model is the residual stress and compressive stress depth data at the weld seam after ultrasonic impact treatment.

[0049] It can be understood that in an embodiment of the present invention, a pre - constructed ultrasonic impact treatment prediction model can be trained using a training data set. The input of the ultrasonic impact treatment prediction model is multiple sets of process parameters and welding simulation data, and the output of the ultrasonic impact treatment prediction model is the residual stress and compressive stress depth data at the weld seam after ultrasonic impact treatment.

[0050] In an embodiment of the present invention, the ultrasonic impact treatment prediction model includes an input layer, multiple hidden layers, an output layer, and a loss function. Among them, the input layer is used to input process parameters and welding simulation data; the multiple hidden layers are used to perform feature extraction and non-linear transformation on the process parameters and welding simulation data; the output layer is used to output the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; the loss function is used to calculate the prediction error according to the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment and the true values in the training dataset, and update the model parameters of the ultrasonic impact treatment prediction model according to the prediction error until the loss function is minimized.

[0051] Among them, the loss function can be calculated by MSE (Means Squared Error), and the formula of the loss function is: ; where N is the number of training samples (process parameters and welding simulation data), is the actual value (target value) of the i-th sample, which is a vector containing three outputs, the maximum value, average value of the residual stress at the weld after ultrasonic impact treatment, and the depth of the compressive stress layer, is the model prediction value of the i-th sample, which is also a vector. ||... ||^2 represents the square of the L2 norm (Euclidean norm) of the vector, that is, the square root of the sum of the squares of each element in the vector.

[0052] In an embodiment of the present invention, MLP (Multi-Layer Perception) can be used as the ultrasonic impact treatment prediction model. Among them, the input layer is used to input process parameters and welding simulation data; the multiple hidden layers are used to perform feature extraction and non-linear transformation on the process parameters and welding simulation data; the output layer is used to output the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; the loss function is used to calculate the prediction error according to the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment and the true values in the training dataset, and update the model parameters of the ultrasonic impact treatment prediction model according to the prediction error until the loss function is minimized.

[0053] Specifically, the MLP includes an input layer, multiple hidden layers, and an output layer. Among them, the input layer includes 8 nodes, corresponding respectively to: cooling temperature, impact frequency, impact amplitude, impact needle diameter, impact time, and three features of the weld area obtained from welding simulation data (the temperature of the weld center point before ultrasonic impact treatment, the maximum residual stress value before ultrasonic impact treatment, the average residual stress value before ultrasonic impact treatment); the hidden layer contains two hidden layers, the first layer contains 64 neurons, and the second layer contains 32 neurons; the output layer includes three nodes, corresponding respectively to: the maximum residual stress at the weld after ultrasonic impact treatment, the average residual stress after ultrasonic impact treatment, and the average depth of the compressive stress layer after ultrasonic impact treatment; the connection between the input layer and the hidden layer adopts a fully connected method, the ReLU activation function is used as the activation function of the hidden layer, the Sigmoid activation function is used as the activation function of the output layer, a loss function is constructed, which consists of data loss, and the MSE mean square error is used for calculation.

[0054] In the embodiment of the present invention, before training the pre-constructed ultrasonic impact treatment prediction model using the training data set, it further includes: determining the perturbation amplitude of the training data in the training data set based on the prediction error; generating new training data based on the perturbation amplitude, expanding the training data set based on the new training data; training the pre-constructed ultrasonic impact treatment prediction model using the expanded training data set.

[0055] It can be understood that the suitable embodiments of the present invention can perform slight perturbations on the training data in the existing training data set to generate new training data, enrich the features of the training data, improve the prediction performance and generalization ability of the ultrasonic impact treatment prediction model, can determine the perturbation amplitude of the training data in the training data set based on the prediction error, generate new training data based on the perturbation amplitude, expand the training data set based on the new training data, and train the pre-constructed ultrasonic impact treatment prediction model using the expanded training data set.

[0056] Specifically, the process of adaptively enhancing the training data in the embodiment of the present invention is as Figure 5 shown. Taking the ultrasonic impact treatment prediction model as an MLP model as an example, by perturbing the existing data, new training samples are generated, and the perturbation amplitude is adaptively adjusted. Specifically: 1. Train an initial MLP model (trained using a small amount of existing finite element simulation data). Then, use the trained model to predict the existing training data, and calculate the prediction error. The prediction error is calculated using MSE.

[0057] 2. According to the magnitude of the prediction error, adaptively adjust the amplitude of data augmentation, generate new samples (i.e., new training data), and add them to the original training data set.

[0058] 3. Retrain the MLP model using the augmented dataset.

[0059] For example, assume that only the impact time t is subjected to adaptive data augmentation for now.

[0060] (1) Assume that there are initially only ten simulation points, the MLP model has one input and one output, and an initial MLP model is trained using the initial ten data points; (2) Use the trained MLP model to predict the values of the original ten data points, calculate the prediction error (i.e., MSE) of the ten points using the predicted values and the actual values, and calculate the average MSE; (3) Set a basic perturbation amplitude, for example, 20%. The meaning of the basic perturbation amplitude is that without adaptive adjustment, perturbation will be performed within the range of 20% of the original input; (4) Adjust the perturbation amplitude of each original data point according to the average MSE: the new perturbation amplitude = basic perturbation amplitude * (MSE of a certain point / average MSE); (5) Assume that for a certain point t = 10, the calculated new perturbation amplitude = 20%, and the perturbation range is from 10 * (1 - 20%) to 10 * (1 + 20%), that is, (8, 12); (6) Randomly generate a t within this interval, repeat the finite element simulation steps of ultrasonic impact treatment, calculate the output value, and obtain a new data point, that is, a new training data.

[0061] In addition, the selection of the input of the new data point is obtained by randomly perturbing the original input, and the perturbation amplitude is adaptively adjusted according to the prediction error of the MLP model. The output value of the new data point is not predicted by the model and must be obtained through simulation, so as to ensure that errors will not accumulate and guarantee data authenticity. For 5 inputs and 3 outputs, the inputs and outputs are changed to vectors (1, 2, 3, 4, 5) and (1, 2, 3). The training and prediction error calculation of the initial MLP model are the same. The difference is that the basic perturbation amplitude needs to be set separately for each input parameter, the perturbation amplitude of each parameter is calculated for each original training data respectively, and then the perturbation range of each parameter of each original training data is determined. Random numbers are generated within the range to obtain the new input, and then the output is obtained through finite element simulation to obtain the new training data.

[0062] In step S104, the trained ultrasonic impact treatment prediction model is used to predict the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment, and at least one set of process parameters of the ultrasonic impact treatment process is optimized according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment.

[0063] It can be understood that the embodiments of the present invention can use the trained ultrasonic impact treatment prediction model to predict the residual stress and the depth data of the compressive stress at the weld after the ultrasonic impact treatment process, and optimize at least one set of process parameters of the ultrasonic impact treatment process according to the residual stress and the depth data of the compressive stress at the weld after the ultrasonic impact treatment. By comprehensively considering various process parameters, especially the influence of the cooling temperature on the ultrasonic impact treatment process, the optimization of all process parameters can be realized to minimize the residual stress to the greatest extent, reduce the service risk of the vacuum chamber components, and can provide a reference for the parameter adjustment of the ultrasonic impact treatment equipment.

[0064] In the embodiments of the present invention, optimizing at least one set of process parameters of the ultrasonic impact treatment process according to the residual stress and the depth data of the compressive stress at the weld after the ultrasonic impact treatment includes: setting the optimization objective of the ultrasonic impact treatment process parameters, and inputting the optimization objective, at least one set of ultrasonic impact treatment process parameters, and the corresponding residual stress and the depth data of the compressive stress at the weld after the ultrasonic impact treatment into the target algorithm; using the target algorithm to find the optimal solution, and optimizing at least one set of process parameters of the ultrasonic impact treatment process according to the optimization result.

[0065] Among them, the target algorithm can be selected according to specific situations, such as the NSGA-II algorithm, and no specific limitation is made here; the optimization objective can be to minimize the maximum value and the average value of the residual stress and maximize the depth of the compressive stress.

[0066] It can be understood that the embodiments of the present invention can use the target algorithm, the optimization objective, and the residual stress and the depth data of the compressive stress at the weld after the ultrasonic impact treatment to optimize at least one set of process parameters of the ultrasonic impact treatment process, so as to realize the optimization of the ultrasonic impact treatment parameters of the ultrasonic impact treatment process.

[0067] For example, taking the NSGA-II algorithm to determine the optimal solution of the process parameters, specifically: 1. Individual coding.

[0068] Encode each set of process parameters of the ultrasonic impact treatment process into an individual, that is, each individual represents an ultrasonic impact treatment plan. This plan consists of a set of ultrasonic impact treatment parameters. Each individual can be represented as a vector, that is, (T, f, A, D, t) corresponding to temperature, frequency, amplitude, diameter, and time.

[0069] 2. Initialize the population.

[0070] Take N individuals as the initial parental population.

[0071] 3. Fitness evaluation.

[0072] Evaluate the fitness of each individual using the trained ultrasonic impact treatment prediction model. The purpose of fitness evaluation is to evaluate the "good or bad" of each individual, that is, to evaluate the advantages and disadvantages of each set of ultrasonic impact treatment parameters. The goal of the present invention is to minimize the maximum and average residual stresses and maximize the depth of the compressive stress layer. Therefore, the fitness is a vector composed of multiple objective function values: (f1, f2, f3) = (σ_max, σ_avg, -d). Since it is necessary to ensure that all objective functions are optimized in the same direction (minimum value), the depth of the compressive stress layer is taken as the opposite number.

[0073] 4. Non-dominated sorting.

[0074] Use the objective algorithm to sort the initial parent population, and divide the individuals in the initial parent population into multiple Pareto ranks.

[0075] 5. Selection operation.

[0076] Use the tournament selection operator for the selection operation, that is, randomly select multiple individuals from the parent population, and determine multiple parent individuals based on the ranks corresponding to the selected multiple individuals.

[0077] 6. Crossover operation.

[0078] Use the simulated binary crossover operator for the crossover operation, that is, perform the crossover operation on multiple parent individuals to obtain multiple offspring individuals.

[0079] 7. Mutation operation.

[0080] Use the polynomial mutation operator for the mutation operation, that is, perform the mutation operation on multiple offspring individuals to form an offspring population.

[0081] 8. Environmental selection.

[0082] Merge the parent population and the offspring population, and use the objective algorithm to sort the merged population. Select individuals with higher Pareto ranks to form a new population, and eliminate individuals with lower Pareto ranks.

[0083] 9. Repeat the above steps until the stopping conditions are met (using both the maximum number of iterations and the change in the Pareto front being less than the threshold as the stopping conditions). Finally, output the Pareto optimal solution set, that is, the data set of the optimal process parameter combinations.

[0084] Taking the NSGA-II algorithm as an example and the ultrasonic impact treatment prediction model as an MLP model as an example, the process of optimizing the process parameters is as Figure 6 shown. Use the trained MLP model as a surrogate model to search for the Pareto optimal solution set, including: 1. Initialize the population.

[0085] 2. Use the MLP model to evaluate the fitness of each individual. Then, perform operations such as non-dominated sorting, selection, crossover, and mutation to generate a new population.

[0086] 3. Repeat the above steps until the stopping conditions are met (using both the maximum number of iterations and the change in the Pareto front being less than a threshold as the stopping conditions), and finally output the Pareto optimal solution set.

[0087] The following describes the ultrasonic impact treatment process parameter optimization method of the embodiments of the present invention through a specific embodiment. The process is as Figure 7 shown and includes the following steps: Step 1: Based on different cooling temperatures and different ultrasonic impact treatment process parameters, including impact frequency, impact amplitude, impact needle diameter, and impact time, use the five-factor and multi-level orthogonal experiment method to obtain several groups of parameter combinations.

[0088] Step 2: Establish a finite element model of the welded part, and use the thermo-elastoplastic finite element method to perform sequential coupling thermal-mechanical analysis and numerical simulation on the welding process to obtain the post-weld deformation model, post-weld temperature field, and post-weld residual stress field of the welded part and save them to the result file.

[0089] Step 3: Based on the post-weld deformation model, establish a finite element model of ultrasonic impact treatment, and load the post-weld temperature field and post-weld residual stress field as predefined fields into the model, and perform simulation simulations for different temperature and process parameter combinations respectively.

[0090] The control methods for each parameter are as follows: Cooling temperature: Cool the welded part to the set temperature in the parameter combination through the external temperature of the heat exchange condition, and then set the external temperature to room temperature for temperature-kinetic coupling simulation; Impact needle diameter: Model and assemble impact needles with different diameters respectively; Impact frequency and impact amplitude: Apply a displacement load controlled by amplitude perpendicular to the weld surface to the impact needle, and set the impact frequency and amplitude in the amplitude; Impact time: Set the ultrasonic impact analysis step time, and apply a velocity load parallel to the weld direction to the impact needle to ensure that the weld is impacted within the set time.

[0091] Step 4: Use the cooling temperature, impact frequency, impact amplitude, impact needle diameter, and impact time as input parameters, and use the average value of the residual stress at the weld, the maximum value of the residual stress at the weld, and the average value of the depth of the compressive stress layer at the weld (the distance from the surface to the zero stress point in the depth direction) as the three output parameters.

[0092] Step 5: Construct a multi-layer perceptron (MLP) based on adaptive data augmentation.

[0093] Determine the network structure: Select a suitable multi-layer perceptron (MLP) as the prediction model. The MLP includes an input layer, multiple hidden layers, and an output layer. The input layer consists of 8 nodes, corresponding to: cooling temperature, impact frequency, impact amplitude, impact pin diameter, impact time, and three features from the welding simulation results in the weld area (weld center point temperature, maximum residual stress value, average residual stress value). The hidden layer contains two hidden layers. The first layer contains 64 neurons, and the second layer contains 32 neurons. The output layer includes three nodes, corresponding to: the maximum residual stress at the weld, the average residual stress, and the average depth of the compressive stress layer. The connection between the input layer and the hidden layer adopts a fully connected method. Determine the activation function: Use the ReLU activation function as the activation function for the hidden layer and the Sigmoid activation function as the activation function for the output layer. Construct the loss function: It consists of data loss and uses the mean square error (MSE) to calculate.

[0094] Among them, adaptive data augmentation: Generate new training samples by making small perturbations to the existing data; Train the MLP model: Use the Adam optimizer to train the MLP model, and the goal is to minimize the loss function.

[0095] Step 6: Construct the NSGA-II algorithm based on the multi-layer perceptron surrogate model.

[0096] Individual encoding: Encode the ultrasonic impact treatment parameters as individuals; Initialize the population: Randomly generate N individuals as the initial population; Fitness evaluation: Use the trained MLP model to evaluate the fitness of each individual; Non-dominated sorting: Use the NSGA-II algorithm to sort the population and divide the population into multiple Pareto ranks; Selection operation: Use the tournament selection operator for the selection operation. Crossover operation: Use the simulated binary crossover operator for the crossover operation; Mutation operation: Use the polynomial mutation operator for the mutation operation; Environmental selection: Combine the parent population and the offspring population, use the NSGA-II algorithm to sort the combined population, select individuals with higher Pareto ranks to form a new population, and eliminate individuals with lower Pareto ranks.

[0097] Step 7: Output the Pareto optimal solution set and select the best combination of process parameters according to the actual engineering requirements.

[0098] In summary, for the post-weld ultrasonic impact treatment process, the present invention proposes to perform ultrasonic impact treatment after cooling to different temperatures, taking into account the influence of the temperature factor on the elimination of residual stress; for the highly non-linear relationship between welding residual stress and process parameters, the present invention uses a multi-layer perceptron (MLP) based on adaptive data augmentation and a surrogate model evolutionary algorithm (NSGA-II) to optimize the post-weld ultrasonic impact treatment process, achieving accurate fitting and prediction; for the ultrasonic impact treatment process of large structural parts in a vacuum chamber, the multi-objective ultrasonic impact parameter optimization method proposed by the present invention can consider more output target parameters of residual stress and deformation, and select the best parameters that meet the actual engineering application scenarios.

[0099] The optimization method of the present invention focuses on considering the influence of temperature change on the regulation of residual stress in the ultrasonic impact treatment process; combines the adaptive data augmentation method with a multi-layer perceptron (MLP) to construct a high-precision prediction model for ultrasonic impact treatment parameters; considers multi-objective parameter optimization, which is more in line with the actual engineering application requirements; uses the trained MLP model as a surrogate model and combines it with the NSGA-II algorithm to accelerate the evolution process and improve the optimization efficiency.

[0100] According to the ultrasonic impact treatment process parameter optimization method proposed in the embodiment of the present invention, welding simulation data of the welded part and multiple groups of process parameters of the ultrasonic impact treatment process can be obtained, and finite element simulation is performed on each group of process parameters based on the welding simulation data to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters. A training data set is generated according to multiple groups of process parameters, welding simulation data, and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment. Comprehensively consider the influence of various process parameters on the residual stress and compressive stress depth of the weld after ultrasonic impact treatment, and use the training data set to train the pre-constructed ultrasonic impact treatment prediction model, and use the trained ultrasonic impact treatment prediction model to predict the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment, and optimize at least one group of process parameters of the ultrasonic impact treatment process according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment. By comprehensively considering the influence of various process parameters on the ultrasonic impact treatment process, especially the influence of the cooling temperature on the residual stress after the ultrasonic impact treatment process, the optimization of the process parameters of the ultrasonic impact treatment process is realized to minimize the residual stress, reduce the service risk of vacuum chamber components, and can provide a reference for the parameter adjustment of ultrasonic impact treatment equipment.

[0101] Next, refer to the drawings to describe the ultrasonic impact treatment process parameter optimization device proposed in the embodiment of the present invention.

[0102] Figure 8 It is a block diagram of the ultrasonic impact treatment process parameter optimization device according to the embodiment of the present invention.

[0103] As Figure 8 shown, the device 10 for optimizing the process parameters of ultrasonic impact treatment includes: an acquisition module 100, a simulation module 200, a training module 300, and an optimization module 400.

[0104] Among them, the acquisition module 100 is used to acquire the welding simulation data of the welded part and multiple groups of process parameters of the ultrasonic impact treatment process, specifically including: acquiring the welding simulation data of the welded part and multiple groups of process parameters of the ultrasonic impact treatment process, where the process parameters include cooling temperature, impact frequency, impact amplitude, impact needle diameter, and impact time; The simulation module 200 is used to perform finite element simulations on each group of process parameters, specifically including: performing finite element simulations on each group of process parameters based on the welding simulation data to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, and generating a training dataset according to the multiple groups of process parameters, welding simulation data, and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; The training module 300 is used to train the ultrasonic impact treatment prediction model, specifically including: using the training dataset to train the pre-constructed ultrasonic impact treatment prediction model, where the input of the ultrasonic impact treatment prediction model is multiple groups of process parameters and welding simulation data, and the output of the ultrasonic impact treatment prediction model is the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; The optimization module 400 is used to optimize at least one group of process parameters of the ultrasonic impact treatment process, specifically including: using the trained ultrasonic impact treatment prediction model to predict the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment, and optimizing at least one group of process parameters of the ultrasonic impact treatment process according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment.

[0105] In the embodiment of the present invention, the ultrasonic impact treatment prediction model includes an input layer, multiple hidden layers, an output layer, and a loss function. Among them, the input layer is used to input process parameters and welding simulation data; the multiple hidden layers are used to perform feature extraction and non-linear transformation on the process parameters and welding simulation data; the output layer is used to output the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; the loss function is used to calculate the prediction error according to the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment and the true values in the training dataset, and update the model parameters of the ultrasonic impact treatment prediction model according to the prediction error until the loss function is minimized.

[0106] In the embodiment of the present invention, the acquisition module 100 is further used to: construct a welding finite element model; perform multi-layer and multi-pass welding simulation on the welded part using the welded part finite element model to obtain the welding simulation data of the welded part.

[0107] In an embodiment of the present invention, the welding simulation data includes: a post-weld deformation model, a post-weld temperature field before ultrasonic impact treatment, and a post-weld residual stress field.

[0108] In an embodiment of the present invention, the simulation module 200 is further configured to: construct a finite element model for ultrasonic impact treatment based on the post-weld deformation model, and load the post-weld temperature field and the post-weld residual stress field before ultrasonic impact treatment into the finite element model for ultrasonic impact treatment as predefined fields, and use the finite element model for ultrasonic impact treatment to perform ultrasonic impact simulation on the weld seam under different process parameters, so as to obtain the residual stress and the depth data of the compressive stress at the weld seam after ultrasonic impact treatment under different process parameters.

[0109] In an embodiment of the present invention, the acquisition module 100 is further configured to: perform failure treatment on the weld seam area in the welded part to make the weld seam area not participate in the finite element analysis; sequentially activate each weld seam to participate in the finite element analysis, and perform thermal analysis simulation on the current weld seam participating in the finite element analysis to obtain the temperature field distribution data of the current weld seam; use the temperature field distribution data of the current weld seam as the initial condition, perform force analysis on the current weld seam to obtain the corresponding stress field distribution and deformation data, until all weld seams are analyzed, so as to obtain the welding simulation data of the welded part.

[0110] In an embodiment of the present invention, the simulation module 200 is further configured to: cool the welded part to the cooling temperature in the process parameter combination through the external temperature of the heat exchange condition, and set the external temperature as the target temperature to perform temperature dynamics coupling simulation; set impact pins with different diameters, apply a displacement load perpendicular to the weld seam surface and controlled by the amplitude to the impact pins, and set the impact frequency and the impact amplitude in the amplitude; set the time for ultrasonic impact analysis, apply a velocity load parallel to the weld seam direction to the impact pins, and ensure that the weld seam is impacted within the set time, so as to obtain the residual stress and the depth data of the compressive stress at the weld seam after ultrasonic impact treatment under different process parameters.

[0111] In an embodiment of the present invention, the optimization module 400 is further configured to: set an optimization target for the ultrasonic impact treatment process parameters, input the optimization target, at least one set of ultrasonic impact treatment process parameters, and the corresponding residual stress and the depth data of the compressive stress at the weld seam after ultrasonic impact treatment into the target algorithm; use the target algorithm to perform optimization, and optimize at least one set of process parameters of the ultrasonic impact treatment process according to the optimization result.

[0112] In an embodiment of the present invention, the ultrasonic impact treatment process parameter optimization device 10 of the embodiment of the present invention further includes: a perturbation module.

[0113] Among them, the perturbation module is used to determine the perturbation amplitude of the training data in the training data set based on the prediction error before training the pre-constructed ultrasonic impact treatment prediction model using the training data set; generate new training data based on the perturbation amplitude, augment the training data set based on the new training data; and train the pre-constructed ultrasonic impact treatment prediction model using the augmented training data set.

[0114] It should be noted that the foregoing explanation of the embodiments of the ultrasonic impact treatment process parameter optimization method also applies to the ultrasonic impact treatment process parameter optimization device of this embodiment, and will not be elaborated here.

[0115] According to the ultrasonic impact treatment process parameter optimization device provided by the embodiments of the present invention, welding simulation data of a welded part and multiple sets of process parameters of the ultrasonic impact treatment process can be obtained, and finite element simulation is performed on each set of process parameters based on the welding simulation data to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters. A training data set is generated based on multiple sets of process parameters, welding simulation data, and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment. The influence of various process parameters on the residual stress and compressive stress depth of the weld after ultrasonic impact treatment is comprehensively considered, and the pre-constructed ultrasonic impact treatment prediction model is trained using the training data set. The trained ultrasonic impact treatment prediction model is used to predict the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment, and at least one set of process parameters of the ultrasonic impact treatment process is optimized according to the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment. By comprehensively considering the influence of various process parameters on the ultrasonic impact treatment process, especially the influence of the cooling temperature on the residual stress after the ultrasonic impact treatment process, the optimization of the process parameters of the ultrasonic impact treatment process is realized in an all-round way, so as to minimize the residual stress, reduce the service risk of the vacuum chamber components, and can provide a reference for the parameter adjustment of the ultrasonic impact treatment equipment.

[0116] Figure 9 The structural schematic diagram of the electronic device provided by the embodiments of the present invention. The electronic device may include: A memory 801, a processor 802, and a computer program stored on the memory 801 and executable on the processor 802.

[0117] When the processor 802 executes the program, it implements the ultrasonic impact treatment process parameter optimization method provided in the above embodiments.

[0118] Further, the electronic device further includes: A communication interface 803 for communication between the memory 801 and the processor 802.

[0119] The memory 801 is used to store a computer program executable on the processor 802.

[0120] The memory 801 may include high-speed RAM memory and may also include non-volatile memory, such as at least one disk memory.

[0121] If the memory 801, the processor 802, and the communication interface 803 are implemented independently, the communication interface 803, the memory 801, and the processor 802 can be interconnected through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 9 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.

[0122] Optionally, in a specific implementation, if the memory 801, the processor 802, and the communication interface 803 are integrated on a single chip, the memory 801, the processor 802, and the communication interface 803 can communicate with each other through an internal interface.

[0123] The processor 802 may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0124] The embodiments of the present invention also provide a computer-readable storage medium, on which a computer program or instruction is stored. When the computer program or instruction is executed by a processor, the ultrasonic impact processing parameter optimization method as described above is implemented.

[0125] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0126] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise specifically defined.

[0127] Any process or method description in the flowchart or described in other ways herein can be understood to represent a module, segment, or part of code including one or N executable instructions for implementing a customized logical function or process, and the scope of the preferred embodiments of the present invention includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in the reverse order according to the involved functions, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0128] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by a combination of any one or more of the following technologies well known in the art: discrete logic circuits with logic gate circuits for implementing logical functions on data signals, application specific integrated circuits with suitable combinational logic gate circuits, programmable gate arrays, field programmable gate arrays, etc.

[0129] Those of ordinary skill in the technical field of the present invention can understand that all or part of the steps carried by the methods of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

Claims

1. A method for optimizing process parameters of ultrasonic impact treatment, characterized in that: The following steps are involved: Acquire welding simulation data of the welded part and multiple sets of process parameters of the ultrasonic impact treatment process, wherein the process parameters include cooling temperature, impact frequency, impact amplitude, impact needle diameter and impact time; Based on the welding simulation data, finite element simulation is performed on each set of process parameters to obtain residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, and a training data set is generated according to the multiple sets of process parameters, the welding simulation data and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; Using the training data set to train a pre-built ultrasonic impact treatment prediction model, wherein the input of the ultrasonic impact treatment prediction model is the multiple sets of process parameters and the welding simulation data, and the output of the ultrasonic impact treatment prediction model is the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; The trained ultrasonic impact treatment prediction model is used to predict the residual stress and compressive stress depth data at the weld after the ultrasonic impact treatment, and at least one set of process parameters of the ultrasonic impact treatment process is optimized according to the residual stress and compressive stress depth data at the weld after the ultrasonic impact treatment.

2. The method for optimizing process parameters of ultrasonic impact treatment according to claim 1, characterized in that: The ultrasonic impact treatment prediction model includes an input layer, multiple hidden layers, an output layer and a loss function, wherein: The input layer is used to input process parameters and welding simulation data; The multiple hidden layers are used for feature extraction and nonlinear transformation of the process parameters and welding simulation data; The output layer is used to output the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; The loss function is used to calculate the prediction error based on the predicted residual stress and compressive stress depth data at the weld after ultrasonic impact treatment and the true value in the training data set, and update the model parameters of the ultrasonic impact treatment prediction model according to the prediction error until the loss function is minimized.

3. The method for optimizing process parameters of ultrasonic impact treatment according to claim 1, characterized in that: The method of obtaining the welding simulation data of the welded part and multiple sets of process parameters of the ultrasonic impact treatment process includes: Construct welding finite element model; The finite element model of the welded part is used to perform multi-layer and multi-pass welding simulation on the welded part to obtain welding simulation data of the welded part.

4. The method for optimizing process parameters of ultrasonic impact treatment according to claim 3, characterized in that: The welding simulation data includes: a post-weld deformation model, a post-weld temperature field before ultrasonic impact treatment, and a post-weld residual stress field. The finite element simulation is performed on each set of process parameters based on the welding simulation data to obtain residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, including: An ultrasonic impact treatment finite element model is constructed based on the post-weld deformation model, and the post-weld temperature field and post-weld residual stress field before the ultrasonic impact treatment are loaded into the ultrasonic impact treatment finite element model as predefined fields. The ultrasonic impact treatment finite element model is used to perform ultrasonic impact simulation on the weld under the different process parameters, and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters are obtained.

5. The method for optimizing process parameters of ultrasonic impact treatment according to claim 3, characterized in that: The method of using the finite element model of the weldment to perform multi-layer and multi-pass welding simulation on the weldment to obtain welding simulation data of the weldment includes: Performing failure treatment on the weld area in the welded part so that the weld area does not participate in finite element analysis; Activate each weld in turn to participate in finite element analysis, and perform thermal analysis simulation on the current weld participating in the finite element analysis to obtain temperature field distribution data of the current weld; The temperature field distribution data of the current weld is used as the initial condition, and force analysis is performed on the current weld to obtain the corresponding stress field distribution and deformation data, until all weld analyses are completed to obtain the welding simulation data of the weldment.

6. The method for optimizing process parameters of ultrasonic impact treatment according to claim 4, characterized in that: The ultrasonic impact treatment finite element model is used to perform ultrasonic impact simulation on the weld under the different process parameters to obtain residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, including: Cooling the weldment to a cooling temperature in a process parameter combination by an external temperature of a heat exchange condition, and setting the external temperature as a target temperature for temperature-dynamic coupling simulation; Setting impact pins of different diameters, applying a displacement load perpendicular to the weld surface and controlled by an amplitude to the impact pins, and setting an impact frequency and an impact amplitude in the amplitude; The time for ultrasonic impact analysis is set, and a velocity load parallel to the weld direction is applied to the impact needle to ensure that the weld is impacted within the set time, so as to obtain the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters.

7. The method for optimizing process parameters of ultrasonic impact treatment according to claim 1, characterized in that: The at least one set of process parameters of the ultrasonic impact treatment process is optimized according to the residual stress and compressive stress depth data at the weld after the ultrasonic impact treatment, including: Setting an optimization target for the ultrasonic impact treatment process parameters, and inputting the optimization target, at least one set of ultrasonic impact treatment process parameters, and corresponding residual stress and compressive stress depth data at the weld after ultrasonic impact treatment into a target algorithm; The target algorithm is used to search for an optimum, and at least one set of process parameters of the ultrasonic impact treatment process is optimized according to the optimization result.

8. The method for optimizing process parameters of ultrasonic impact treatment according to claim 2, characterized in that: Before using the training data set to train the pre-built ultrasonic impact treatment prediction model, the method further includes: Determining a disturbance amplitude of training data in the training data set based on the prediction error; generating new training data based on the disturbance amplitude, and expanding the training data set based on the new training data; The pre-built ultrasonic impact treatment prediction model was trained using the expanded training dataset.

9. An ultrasonic impact treatment process parameter optimization device, characterized in that: include: An acquisition module is used to acquire welding simulation data and multiple sets of process parameters, including: acquiring welding simulation data of a welded part and multiple sets of process parameters of an ultrasonic impact treatment process, wherein the process parameters include cooling temperature, impact frequency, impact amplitude, impact needle diameter and impact time; A simulation module, used for performing finite element simulation on each set of process parameters, including: performing finite element simulation on each set of process parameters based on the welding simulation data, obtaining residual stress and compressive stress depth data at the weld after ultrasonic impact treatment under different process parameters, and generating a training data set according to the multiple sets of process parameters, the welding simulation data and the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; A training module, used for training an ultrasonic impact treatment prediction model, comprising: using the training data set to train a pre-built ultrasonic impact treatment prediction model, wherein the input of the ultrasonic impact treatment prediction model is the multiple sets of process parameters and the welding simulation data, and the output of the ultrasonic impact treatment prediction model is the residual stress and compressive stress depth data at the weld after ultrasonic impact treatment; The optimization module is used to optimize at least one set of process parameters of the ultrasonic impact treatment process, including: using the trained ultrasonic impact treatment prediction model to predict the residual stress and compressive stress depth data at the weld after the ultrasonic impact treatment, and optimizing at least one set of process parameters of the ultrasonic impact treatment process according to the residual stress and compressive stress depth data at the weld after the ultrasonic impact treatment.

10. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for optimizing process parameters of ultrasonic impact treatment as described in any one of claims 1 to 8.

Images