Hot isostatic pressure diffusion welding parameter optimization method and system based on deep learning, electronic equipment and storage medium

Abstract

The invention belongs to the technical field of heterogeneous material diffusion welding, and provides a hot isostatic pressure diffusion welding parameter optimization method and system based on deep learning, electronic equipment and a storage medium. The method comprises the steps of hot isostatic pressure diffusion physical model construction, theoretical prediction result collection, actual detection result collection, original training data construction, error compensation model construction, double-model fusion, candidate prediction result generation and scheme parameter iterative optimization. According to the method, the hot isostatic pressure diffusion physical model and the error compensation model are constructed, so that the theoretical deviation of the physical model can be accurately corrected, and the material loss and the time loss are reduced; by improving the MLP model, the utilization degree of original data is optimized, the coupling characteristics among the data are extracted, and the compensation precision of the model is improved; and through the PID controller, the optimal process parameter combination can be quickly searched by using the result prediction model, and the parameter optimization difficulty is reduced.

Description

Technical Field

[0001] This invention relates to the field of heterogeneous material diffusion welding technology, and in particular to a method, system, electronic device and storage medium for optimizing hot isostatic pressure diffusion welding parameters based on deep learning. Background Technology

[0002] Hot isostatic pressing (HIP) is a technique used in fusion reactors like EAST, where tungsten copper and chromium zirconium copper are commonly used composite structural materials due to their excellent thermal, electrical, and mechanical properties. However, traditional welding methods suffer from problems such as porous structure, high interfacial thermal resistance, and microcracks at the weld joint, severely impacting service life. HIP, with its high density and strong interfacial bonding, is considered an ideal joining process. In the field of heterogeneous material diffusion welding, existing patents on HIP diffusion welding mainly focus on methods and steps for welding different metals, with few addressing systematic optimization of the process parameters.

[0003] However, existing hot isostatic pressure diffusion welding technology has obvious limitations: the determination of hot isostatic pressure welding parameters for tungsten copper and chromium zirconium copper mostly relies on the experience of technicians or a large number of trial and error experiments. During the parameter adjustment process, only one or a few parameters such as temperature, pressure, and holding time are tested, lacking a comprehensive consideration of the interaction between multiple parameters. At the same time, the performance evaluation of the welded joint needs to be carried out through actual tests such as tensile strength testing and microstructure observation after welding. The whole process lacks effective pre-prediction and dynamic optimization methods. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a method, system, electronic device, and storage medium for optimizing hot isostatic pressure diffusion welding parameters based on deep learning, which solves the problems of traditional methods relying too much on human experience or trial-and-error experiments and lacking dynamic optimization methods.

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

[0006] A deep learning-based method for optimizing hot isostatic pressure diffusion welding parameters includes:

[0007] Construct a physical model for thermal isostatic diffusion;

[0008] Several experimental schemes are set up, and each set of experimental schemes is input into the thermo-isostatic diffusion physical model to obtain theoretical prediction results;

[0009] The hot isostatic pressure diffusion welding experiment was conducted according to the experimental scheme, and the actual test results were obtained.

[0010] Extract the error between the theoretical prediction result and the actual detection result to obtain theoretical-actual error data. Align and integrate the theoretical-actual error data, the theoretical prediction result, and the experimental scheme to obtain the original training data.

[0011] The improved MLP model is trained using the original training data to obtain an error compensation model;

[0012] By integrating the thermo-isostatic diffusion physical model and the error compensation model, a result prediction model is obtained;

[0013] Set a parameter range, randomly select a set of schemes within the parameter range and input them into the result prediction model to obtain candidate prediction results;

[0014] The schemes input to the result prediction model are iterated using a preset PID controller until the candidate prediction results meet the preset requirements, at which point the iteration stops, and the optimized parameters for hot isostatic diffusion welding are obtained.

[0015] Preferably, a thermal isostatic diffusion physical model is constructed, including:

[0016] Construct an interface atomic diffusion equation; the interface atomic diffusion equation includes: and ;in, It is the atomic diffusion flux; The diffusion coefficient is denoted as . This represents the elemental concentration value. Indicates distance; Indicates time;

[0017] Construct a temperature dependence equation; the temperature dependence equation is: ;in, It is the diffusion constant; It is the diffusion activation energy; It is the gas constant; Absolute temperature;

[0018] Construct a densification equation; the densification equation is as follows: ;in, Relative density; These are material constants; It is the hot isostatic pressure; , These are the first and second fitting indices, respectively.

[0019] Construct the free energy equation; the free energy equation is as follows: ;in, For free energy to change; For enthalpy change; It is an entropy change;

[0020] By integrating the interface atomic diffusion equation, the temperature dependence equation, the densification equation, and the free energy equation, the thermo-isostatic diffusion physical model is obtained.

[0021] Preferably, several experimental schemes are set, and each set of experimental schemes is input into the thermo-isostatic diffusion physical model to obtain theoretical prediction results, including:

[0022] Select the process parameters to be controlled; the process parameters to be controlled include: hot isostatic pressing temperature, hot isostatic pressing pressure, holding time, cooling rate, and heating rate;

[0023] Determine the material property parameters; the material property parameters include: diffusion constant, diffusion activation energy, and material constant.

[0024] The set process parameters to be controlled and the material property parameters are orthogonally combined to obtain several sets of experimental schemes;

[0025] The experimental scheme is input into the thermo-isostatic diffusion physical model to obtain the theoretical prediction results; the theoretical prediction results include: tensile strength, thermal diffusivity, porosity, and residual stress.

[0026] Preferably, a hot isostatic pressure diffusion welding experiment is performed according to the experimental scheme to obtain actual test results, including:

[0027] Select a base material that matches the parameters of the experimental scheme;

[0028] The parent material was sequentially cut, ground, ultrasonically cleaned, rinsed with deionized water, and dried to obtain the experimental sample.

[0029] The temperature, pressure, and rate of the hot isostatic pressing equipment are calibrated.

[0030] After placing the pre-set tooling and the experimental sample into the furnace cavity of the hot isostatic pressing equipment, the furnace cavity is evacuated.

[0031] The hot isostatic pressing equipment is heated according to the heating rate provided by the experimental scheme. When the temperature inside the furnace cavity meets the requirements of the experimental scheme, the experimental sample is subjected to hot isostatic pressing treatment.

[0032] During the heat preservation and pressure holding stage, the experimental sample is subjected to diffusion welding at a fixed temperature and pressure. After the heat preservation and pressure holding stage is completed, the furnace cavity is controlled to undergo cooling and depressurization treatment according to the experimental plan to obtain the welded sample.

[0033] The welded samples were subjected to mechanical property testing, thermophysical property testing, microstructure quality testing, and stress state testing to obtain the actual test results.

[0034] Preferably, the improved MLP model includes: an input layer, a weighted preprocessing layer, a hidden layer, and an output layer connected in sequence.

[0035] Preferably, a preset PID controller is used to iterate the scheme input to the result prediction model until the candidate prediction result meets the preset requirements, at which point the iteration stops, and the optimized parameters for hot isostatic pressing diffusion welding are obtained, including:

[0036] The candidate prediction results are set as the control target, and the PID input variables are set according to the control target; the expression of the PID input variables is: ;in, For the first Control deviation during the next iteration; The number of target indicators; For the first The weight of each indicator; For the first The result prediction model output during the iteration is the first... One indicator; For the first The preset target value for each indicator;

[0037] Construct the PID controller; the expression of the PID controller is:

[0038] ;

[0039] in, For parameter adjustment amount; , , These are the proportional coefficient, integral coefficient, and differential coefficient, respectively. for The deviation integral of the next iteration; The rate of change of deviation;

[0040] Determine whether the PID input variable meets the preset requirements to obtain the deviation status;

[0041] When the deviation level is not in compliance with requirements, the PID input variable is input to the PID controller to obtain the parameter adjustment amount;

[0042] The parameter adjustment amount is updated to the scheme of the result prediction model, and the process returns to the step "determine whether the PID input variable meets the preset requirements and obtain the deviation status". When the deviation status meets the requirements, the last updated scheme of the result prediction model is output to obtain the hot isostatic diffusion welding optimization parameters.

[0043] Preferably, the improved MLP model is trained using the original training data to obtain an error compensation model, including:

[0044] Construct the improved MLP model;

[0045] The original training data is received through the input layer;

[0046] The weighted preprocessing layer performs pairwise weighted calculations on the data in each group of the original training data to obtain weighted interactive data; the expression for the weighted preprocessing layer is: ;in, The first data in the same set of the original training data The data and the first The weighted interactive data of each data point; This is a learnable interaction weight matrix; , The first data point in the same group of the original training data is respectively the second data point in the original training data. The data and the first One data point;

[0047] The weighted interactive data and the original training data are synchronously input into two fully connected layers in the hidden layer for nonlinear mapping to obtain extracted features.

[0048] The extracted features are input into the output layer to perform multi-index error prediction, and the model prediction result is obtained.

[0049] Based on the model prediction results and the theoretical and actual error data, the hidden layer and the weighted preprocessing layer are backpropagated iteratively optimized to obtain the trained error compensation model.

[0050] Preferably, a deep learning-based hot isostatic pressure diffusion welding parameter optimization system includes:

[0051] The physical model building module is used to build a thermal isostatic diffusion physical model;

[0052] The theoretical data collection module is used to set up several experimental schemes, input each set of experimental schemes into the thermo-isostatic diffusion physical model, and obtain theoretical prediction results;

[0053] The actual data collection module is used to conduct hot isostatic pressure diffusion welding experiments according to the experimental scheme and obtain actual test results.

[0054] The training data construction module is used to extract the error between the theoretical prediction result and the actual detection result to obtain theoretical-actual error data, and to align and integrate the theoretical-actual error data, the theoretical prediction result, and the experimental scheme to obtain the original training data.

[0055] The compensation model construction module is used to train the improved MLP model using the original training data to obtain the error compensation model;

[0056] The model fusion module is used to fuse the thermo-isostatic diffusion physics model and the error compensation model to obtain the result prediction model.

[0057] The result prediction module is used to set a parameter range, randomly select a set of schemes within the parameter range and input them into the result prediction model to obtain candidate prediction results;

[0058] The parameter optimization module is used to iterate the scheme input to the result prediction model using a preset PID controller until the candidate prediction result meets the preset requirements, and then stop the iteration to obtain the optimized parameters for hot isostatic diffusion welding.

[0059] Preferably, an electronic device includes: at least one processor and a memory communicatively connected to the processor; wherein the memory stores instructions executable by the processor, the instructions being executed by the processor to enable the processor to perform the aforementioned deep learning-based hot isostatic diffusion welding parameter optimization method.

[0060] Preferably, a non-transitory computer-readable storage medium stores computer instructions for causing a computer to execute the aforementioned deep learning-based hot isostatic pressure diffusion welding parameter optimization method.

[0061] The present invention discloses the following technical effects:

[0062] This invention provides a method, system, electronic device, and storage medium for optimizing hot isostatic pressing (HIP) diffusion welding parameters based on deep learning. By constructing a physical model and an error compensation model for HIP diffusion, it solves the problem of large prediction errors in traditional linear fitting models and corrects the deviations of the physical model. By improving the MLP model, it solves the problem that traditional MLP models lack consideration of the interaction characteristics between data, thereby improving the accuracy of model compensation. By using a PID controller, it solves the problem that traditional methods rely too much on human experience or have high trial-and-error costs, and enables the rapid search for the optimal combination of process parameters using the result prediction model. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 A schematic diagram of the hot isostatic pressing diffusion welding parameter optimization process based on deep learning provided in an embodiment of the present invention;

[0065] Figure 2 A flowchart of the steps provided for embodiments of the present invention;

[0066] Figure 3 A flowchart of the physical model construction provided for embodiments of the present invention;

[0067] Figure 4 A flowchart for PID parameter optimization provided in an embodiment of the present invention;

[0068] Figure 5 This is a network architecture diagram provided for an embodiment of the present invention. Detailed Implementation

[0069] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0070] The purpose of this invention is to provide a method, system, electronic device, and storage medium for optimizing hot isostatic pressure diffusion welding parameters based on deep learning, which solves the problems of traditional methods relying too much on human experience or trial-and-error experiments and lacking dynamic optimization methods.

[0071] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0072] Figure 1 This is a schematic diagram of the deep learning-based hot isostatic pressing diffusion welding parameter optimization process provided in an embodiment of the present invention. Figure 2 The flowchart provided for the embodiments of the present invention is as follows: Figure 1 and Figure 2 As shown, this invention provides a method for optimizing hot isostatic pressing diffusion welding parameters based on deep learning, including:

[0073] Step 100: Construct a physical model for thermal isostatic diffusion;

[0074] Step 200: Set up several experimental schemes, input each experimental scheme into the thermo-isostatic diffusion physical model, and obtain the theoretical prediction results;

[0075] Step 300: Conduct a hot isostatic pressure diffusion welding experiment according to the experimental scheme to obtain actual test results;

[0076] Step 400: Extract the error between the theoretical prediction result and the actual detection result to obtain theoretical-actual error data. Align and integrate the theoretical-actual error data, the theoretical prediction result, and the experimental scheme to obtain the original training data.

[0077] Step 500: Train the improved MLP model using the original training data to obtain the error compensation model;

[0078] Step 600: Integrate the thermo-isostatic diffusion physical model and the error compensation model to obtain the result prediction model;

[0079] Step 700: Set the parameter range, randomly select a set of schemes within the parameter range and input them into the result prediction model to obtain candidate prediction results;

[0080] Step 800: Use a preset PID controller to iterate the scheme input to the result prediction model until the candidate prediction result meets the preset requirements, then stop the iteration to obtain the optimized parameters for hot isostatic diffusion welding.

[0081] refer to Figure 3 Construct a physical model for thermal isostatic diffusion, including:

[0082] Step 101: Construct the interface atomic diffusion equation; the interface atomic diffusion equation includes: and ;in, It is the atomic diffusion flux; The diffusion coefficient is denoted as . This represents the elemental concentration value. Indicates distance; Indicates time;

[0083] Step 102: Construct the temperature dependence equation; the temperature dependence equation is: ;in, It is the diffusion constant; It is the diffusion activation energy; It is the gas constant; Absolute temperature;

[0084] Step 103: Construct the densification equation; the densification equation is: ;in, Relative density; These are material constants; It is the hot isostatic pressure; , These are the first and second fitting indices, respectively.

[0085] Step 104: Construct the free energy equation; the free energy equation is: ;in, For free energy to change; For enthalpy change; It is an entropy change;

[0086] Step 105: Integrate the interface atomic diffusion equation, the temperature dependence equation, the densification equation, and the free energy equation to obtain the thermo-isostatic diffusion physical model.

[0087] Specifically, several experimental schemes are set up, and each set of experimental schemes is input into the thermo-isostatic diffusion physical model to obtain theoretical prediction results, including:

[0088] Select the process parameters to be controlled; the process parameters to be controlled include: hot isostatic pressing temperature, hot isostatic pressing pressure, holding time, cooling rate, and heating rate;

[0089] Determine the material property parameters; the material property parameters include: diffusion constant, diffusion activation energy, and material constant.

[0090] The set process parameters to be controlled and the material property parameters are orthogonally combined to obtain several sets of experimental schemes;

[0091] The experimental scheme is input into the thermo-isostatic diffusion physical model to obtain the theoretical prediction results; the theoretical prediction results include: tensile strength, thermal diffusivity, porosity, and residual stress.

[0092] Furthermore, a hot isostatic pressure diffusion welding experiment was conducted according to the aforementioned experimental scheme to obtain actual test results, including:

[0093] Select a base material that matches the parameters of the experimental scheme;

[0094] The parent material was sequentially cut, ground, ultrasonically cleaned, rinsed with deionized water, and dried to obtain the experimental sample.

[0095] The temperature, pressure, and rate of the hot isostatic pressing equipment are calibrated.

[0096] After placing the pre-set tooling and the experimental sample into the furnace cavity of the hot isostatic pressing equipment, the furnace cavity is evacuated.

[0097] The hot isostatic pressing equipment is heated according to the heating rate provided by the experimental scheme. When the temperature inside the furnace cavity meets the requirements of the experimental scheme, the experimental sample is subjected to hot isostatic pressing treatment.

[0098] During the heat preservation and pressure holding stage, the experimental sample is subjected to diffusion welding at a fixed temperature and pressure. After the heat preservation and pressure holding stage is completed, the furnace cavity is controlled to undergo cooling and depressurization treatment according to the experimental plan to obtain the welded sample.

[0099] The welded samples were subjected to mechanical property testing, thermophysical property testing, microstructure quality testing, and stress state testing to obtain the actual test results.

[0100] Preferably, the improved MLP model includes: an input layer, a weighted preprocessing layer, a hidden layer, and an output layer connected in sequence.

[0101] refer to Figure 4 The schemes input to the result prediction model are iterated using a preset PID controller until the candidate prediction results meet preset requirements, at which point the iteration stops, resulting in optimized parameters for hot isostatic pressing diffusion welding, including:

[0102] Step 801: Set the candidate prediction result as the control target, and set the PID input variable according to the control target; the expression of the PID input variable is: ;in, For the first Control deviation during the next iteration; The number of target indicators; For the first The weight of each indicator; For the first The result prediction model output during the iteration is the first... One indicator; For the first The preset target value for each indicator;

[0103] Step 802: Construct the PID controller; the expression of the PID controller is:

[0104] ;

[0105] in, For parameter adjustment amount; , , These are the proportional coefficient, integral coefficient, and differential coefficient, respectively. for The deviation integral of the next iteration; The rate of change of deviation;

[0106] Step 803: Determine whether the PID input variable meets the preset requirements to obtain the deviation level status;

[0107] Step 804: When the deviation level is not in compliance with requirements, the PID input variable is input into the PID controller to obtain the parameter adjustment amount;

[0108] Step 805: Update the parameter adjustment amount to the scheme of the result prediction model, and return to the step "determine whether the PID input variable meets the preset requirements and obtain the deviation status". When the deviation status meets the requirements, output the last updated scheme of the result prediction model to obtain the hot isostatic diffusion welding optimization parameters.

[0109] refer to Figure 5 The improved MLP model is trained using the original training data to obtain an error compensation model, including:

[0110] Construct the improved MLP model;

[0111] The original training data is received through the input layer;

[0112] The weighted preprocessing layer performs pairwise weighted calculations on the data in each group of the original training data to obtain weighted interactive data; the expression for the weighted preprocessing layer is: ;in, The first data in the same set of the original training data The data and the first The weighted interactive data of each data point; This is a learnable interaction weight matrix; , The first data point in the same group of the original training data is respectively the second data point in the original training data. The data and the first One data point;

[0113] The weighted interactive data and the original training data are synchronously input into two fully connected layers in the hidden layer for nonlinear mapping to obtain extracted features.

[0114] The extracted features are input into the output layer to perform multi-index error prediction, and the model prediction result is obtained.

[0115] Based on the model prediction results and the theoretical and actual error data, the hidden layer and the weighted preprocessing layer are backpropagated iteratively optimized to obtain the trained error compensation model.

[0116] Specifically, this embodiment provides a method for optimizing hot isostatic pressure diffusion welding parameters. First, a hot isostatic pressure diffusion physical model is constructed. Then, multiple sets of experiments are set up to extract the error between the physical model output and the actual measured values, thereby constructing an error compensation model. The physical model and the compensation model are then fused to obtain a result prediction model. Subsequently, a PID controller is used to iteratively optimize the parameters to find the optimal parameter combination. During model construction, not only welding process-related parameters but also the material's own property parameters must be considered.

[0117] Preferably, hot isostatic pressure diffusion welding mainly promotes atomic diffusion, porosity elimination, and joint densification at the interface of the materials to be welded through the synergistic effect of high temperature and isostatic pressure. Based on this, the physical model of this embodiment integrates diffusion kinetics, interfacial thermodynamics, and pressure-densification coupling processes, and combines process parameters and material properties to achieve theoretical prediction of the welding effect. Before simulation, it is necessary to determine the values ​​of some parameters, including diffusion activation energy, diffusion constant, welding temperature, holding time, initial interface concentration, and interface roughness. Referring to existing literature, the expressions of the HIP diffusion welding physical model constructed in this embodiment include:

[0118] Fick's Law of Diffusion:

[0119]

[0120]

[0121] In the formula, It is the atomic diffusion flux; The diffusion coefficient is denoted as . This represents the elemental concentration value. Indicates distance; Indicates time.

[0122] Arrhenius equation:

[0123]

[0124] In the formula, It is the diffusion constant; It is the diffusion activation energy; It is the gas constant; This refers to absolute temperature.

[0125] HIP compaction equation:

[0126]

[0127] In the formula, Relative density; These are material constants; It is the hot isostatic pressure; , These are the first and second fitting indices, respectively.

[0128] Gibbs free energy equation:

[0129]

[0130] In the formula, For free energy to change; For enthalpy change; This is an entropy change.

[0131] Specifically, the following description uses hot isostatic pressing (HIP) diffusion welding of tungsten copper and chromium zirconium copper as a specific example. Process parameters related to HIP diffusion welding are selected, and reasonable value ranges are set based on material properties. These parameters include HIP temperature, HIP pressure, holding time, heating rate, and cooling rate. Material property parameters of tungsten copper and chromium zirconium copper are determined, including diffusion constant, diffusion activation energy, and material constants. An experimental scheme is constructed using orthogonal experimental design, and the scheme is sequentially input into the HIP diffusion physical model for simulation. Theoretical prediction results are output, including tensile strength, thermal diffusivity, porosity, and residual stress.

[0132] Furthermore, tungsten-copper and chromium-zirconium-copper base materials matching the material property parameters in the experimental scheme were selected, and the base material size specifications were standardized, with dimensions set to 50mm to 100mm in length, 30mm to 50mm in width, and 5mm to 10mm in thickness. The base materials were cut using a wire cutting device, and the welding interface of the samples to be welded was subjected to gradient grinding, using 800-grit, 1200-grit, and 2000-grit silicon carbide sandpaper in sequence, and a mechanical grinder was used for grinding until the interface roughness was no greater than 0.8μm. The ground samples were placed in an ultrasonic cleaning tank, using anhydrous ethanol as the cleaning agent, and ultrasonic cleaning was performed after setting the ultrasonic power and cleaning time to remove residual grinding debris and oil stains at the interface. The rinsed samples were placed in a vacuum drying oven and thoroughly dried at a constant temperature.

[0133] Optionally, before conducting the experiment, the hot isostatic pressing (HIP) equipment needs to be precisely calibrated. Three to five temperature measuring points should be evenly arranged within the furnace cavity, with the calibration temperature range covering the HIP temperature set in the experimental design. A standard pressure sensor should be used to calibrate the pressure range covering the HIP pressure set in the experimental design. The heating and cooling rates should be calibrated using a temperature acquisition instrument, and the pressure increase rate should be calibrated using a pressure acquisition instrument. A graphite fixture compatible with the sample material should be selected, and the pretreated experimental sample should be fixed inside the fixture. The fixture and sample should be placed together into the HIP furnace cavity. After closing the furnace door, the vacuum system should be activated to evacuate the furnace cavity. The heating rate of the equipment was controlled according to the experimental plan. When the temperature inside the furnace reached 90% of the target hot isostatic pressure temperature, the pressure system was started, and the pressure was slowly increased to the target pressure at the set pressure increase rate. After the temperature and pressure reached the values ​​set in the experimental plan, the heat preservation and pressure holding stage was entered. During this stage, the temperature and pressure of each temperature measuring point in the furnace were monitored in real time to ensure that the parameter fluctuations were within the calibration accuracy range. The heat preservation and pressure holding time was executed according to the experimental plan. After the heat preservation and pressure holding was completed, the temperature was reduced according to the cooling rate set in the experimental plan. After the temperature dropped below 50°C and the pressure dropped to normal pressure, the furnace was opened and the welded sample was taken out.

[0134] Furthermore, the error between the theoretical prediction results and the actual detection results is extracted, and this error data, theoretical prediction results, and experimental scheme are aligned and integrated to form the original training data.

[0135] Specifically, this embodiment improves upon the original MLP model. The improved MLP model in this embodiment includes an input layer, a weighted preprocessing layer, a hidden layer, and an output layer connected in sequence. The function of each network layer is as follows:

[0136] 1) Input layer: Receives and preprocesses the original training data, which includes experimental scheme parameters, theoretical prediction results, and theoretical and actual error data. Before input, the data needs to be processed by Min-Max normalization to map the feature values ​​to the interval [0, 1].

[0137] 2) Weighted preprocessing layer: Used to extract the coupling relationships between different features in the original data. Through a learnable interaction weight matrix, weighted interaction calculations are performed on any two features of the same set of input data, as shown in the following formula:

[0138]

[0139] In the formula, The first data in the same set of original training data The data and the first Weighted interactive data of individual data; This is a learnable interaction weight matrix; , The first data point in the same set of original training data is the first... The data and the first Data points.

[0140] Iterate through all feature pairs to generate a weighted interactive dataset.

[0141] 3) Hidden layer: It consists of two fully connected layers connected in sequence. The number of neurons in the first fully connected layer is set to twice the sum of the original input features and the weighted interaction features. The number of neurons in the second fully connected layer is set to half that of the first layer.

[0142] 4) Output layer: The number of neurons is consistent with the number of indicators in the theoretical and actual error data (in this embodiment, the number of neurons in the output layer is 4, corresponding to tensile strength, thermal diffusivity, porosity and residual stress respectively), and finally outputs the multi-indicator error compensation values ​​predicted by the model.

[0143] Preferably, the multi-dimensional welding quality indicators output by the result prediction model are used as control targets, including: tensile strength, thermal diffusivity, porosity, and residual stress of the tungsten-copper-chromium-zirconium-copper welded joint; the input variables of the PID are determined as follows:

[0144]

[0145] In the formula, For the first Control deviation during the next iteration; The number of target indicators; For the first The weight of each indicator; For the first The output of the result prediction model during the nth iteration One indicator; For the first The preset target value for each indicator.

[0146] An independent PID sub-controller is designed for each parameter, and the PID output of a single parameter to be controlled is:

[0147]

[0148] In the formula, For parameter adjustment amount; , , These are the proportional coefficient, integral coefficient, and differential coefficient, respectively. for The deviation integral of the next iteration; This represents the rate of change of deviation.

[0149] Based on experience with hot isostatic pressing (HIP), initial values ​​for the proportional coefficient, integral coefficient, and derivative coefficient are set, and parameter adjustment constraints, i.e., parameter boundaries, are defined. If the overall control deviation is less than the maximum allowable deviation and the overall control deviation remains stable below the maximum allowable deviation for three consecutive iterations, it is determined that the preset requirements are met, and the iteration stops; if the overall control deviation is greater than the maximum allowable deviation, it is determined that the preset requirements are not met, and parameter adjustment is required.

[0150] As an optional implementation, this embodiment also provides a deep learning-based hot isostatic pressure diffusion welding parameter optimization system, including:

[0151] The physical model building module is used to build a thermal isostatic diffusion physical model;

[0152] The theoretical data collection module is used to set up several experimental schemes, input each set of experimental schemes into the thermo-isostatic diffusion physical model, and obtain theoretical prediction results;

[0153] The actual data collection module is used to conduct hot isostatic pressure diffusion welding experiments according to the experimental scheme and obtain actual test results.

[0154] The training data construction module is used to extract the error between the theoretical prediction result and the actual detection result to obtain theoretical-actual error data, and to align and integrate the theoretical-actual error data, the theoretical prediction result, and the experimental scheme to obtain the original training data.

[0155] The compensation model construction module is used to train the improved MLP model using the original training data to obtain the error compensation model;

[0156] The model fusion module is used to fuse the thermo-isostatic diffusion physics model and the error compensation model to obtain the result prediction model.

[0157] The result prediction module is used to set a parameter range, randomly select a set of schemes within the parameter range and input them into the result prediction model to obtain candidate prediction results;

[0158] The parameter optimization module is used to iterate the scheme input to the result prediction model using a preset PID controller until the candidate prediction result meets the preset requirements, and then stop the iteration to obtain the optimized parameters for hot isostatic diffusion welding.

[0159] As an optional implementation, this embodiment also provides an electronic device, including: at least one processor, and a memory communicatively connected to the processor; wherein the memory stores instructions executable by the processor, the instructions being executed by the processor to enable the processor to perform the aforementioned deep learning-based hot isostatic diffusion welding parameter optimization method.

[0160] As an optional implementation, this embodiment also provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the aforementioned deep learning-based hot isostatic pressure diffusion welding parameter optimization method.

[0161] The beneficial effects of this invention are as follows:

[0162] This invention, by constructing a thermal isostatic diffusion physical model and an error compensation model, can accurately correct the theoretical deviations of the physical model, reducing material and time losses; by improving the MLP model, it optimizes the utilization of the original data, extracts the coupling characteristics between data, and improves the compensation accuracy of the model; and by using a PID controller, it can quickly use the results to predict the optimal combination of process parameters, reducing the difficulty of parameter optimization.

[0163] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0164] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A deep learning-based hot isostatic pressing diffusion welding parameter optimization method, characterized in that, The application relates to a method for optimizing hot isostatic pressing diffusion welding parameters. The method comprises the following steps: constructing a hot isostatic pressing diffusion physical model; setting a plurality of experimental schemes, inputting each of the experimental schemes into the hot isostatic pressing diffusion physical model to obtain theoretical prediction results; performing hot isostatic pressing diffusion welding experiments according to the experimental schemes to obtain actual detection results; extracting errors of the theoretical prediction results and the actual detection results to obtain theoretical actual error data, and aligning and integrating the theoretical actual error data, the theoretical prediction results and the experimental schemes to obtain original training data; training an improved MLP model by using the original training data to obtain an error compensation model; fusing the hot isostatic pressing diffusion physical model and the error compensation model to obtain a result prediction model; setting a parameter range, randomly selecting a scheme in the parameter range and inputting the scheme into the result prediction model to obtain a candidate prediction result; and using a preset PID controller to iteratively input the scheme into the result prediction model until the candidate prediction result meets preset requirements, so that hot isostatic pressing diffusion welding optimization parameters are obtained. The method comprises the following steps: constructing a hot isostatic pressing diffusion physical model; fusing an interface atomic diffusion equation, a temperature dependence equation, a densification equation and a free energy equation to obtain the hot isostatic pressing diffusion physical model. The method comprises the following steps: setting a plurality of experimental schemes, inputting each of the experimental schemes into the hot isostatic pressing diffusion physical model to obtain theoretical prediction results, and the method comprises the following steps: selecting to-be-controlled process parameters; the to-be-controlled process parameters comprise hot isostatic pressing temperature, hot isostatic pressing pressure, holding time, cooling rate and heating rate; determining material attribute parameters; the material attribute parameters comprise diffusion constant, diffusion activation energy and material constant; performing orthogonal combination processing on the to-be-controlled process parameters and the material attribute parameters to obtain a plurality of experimental schemes; inputting the experimental schemes into the hot isostatic pressing diffusion physical model to obtain the theoretical prediction results; the theoretical prediction results comprise tensile strength, thermal diffusivity, porosity and residual stress. The method comprises the following steps: selecting base materials matched with the experimental scheme parameters; sequentially performing cutting, polishing, ultrasonic cleaning, deionized water flushing and drying processing on the base materials to obtain experimental samples; performing calibration processing on temperature, pressure and rate of a hot isostatic pressing device; placing a preset tooling and the experimental samples into a furnace cavity of the hot isostatic pressing device and performing vacuumizing processing on the furnace cavity; controlling the hot isostatic pressing device to heat at a heating rate provided by the experimental scheme, and performing hot isostatic pressing processing on the experimental samples when the temperature in the furnace cavity meets the requirements of the experimental scheme; performing diffusion welding on the experimental samples at a fixed temperature and pressure in a holding and pressure maintaining stage, and performing temperature and pressure reducing processing on the furnace cavity according to the experimental scheme after the holding and pressure maintaining stage is ended to obtain welded samples; and performing mechanical property detection, thermal physical property detection, microscopic quality detection and stress state detection on the welded samples to obtain actual detection results. ​ ​ ​ ​ 2. The method of claim 1, wherein the method is based on deep learning. ​ constructing an interface atomic diffusion equation; the interface atomic diffusion equation comprising: and ; wherein, is an atomic diffusion flux; is a diffusion coefficient; is an element concentration value; denotes a distance; denotes a time; constructing a temperature dependence equation; the temperature dependence equation is: ; wherein, is a diffusion constant; is a diffusion activation energy; is a gas constant; is an absolute temperature; constructing a densification equation; the densification equation is: ; wherein, is a relative density; is a material constant; is a hot isostatic pressing pressure; , is a first fitting exponent, a second fitting exponent, respectively; constructing a free energy equation; the free energy equation is: ; wherein, is a change in free energy; is a change in enthalpy; is a change in entropy; ​ 3. The deep learning-based hot isostatic pressing diffusion bonding parameter optimization method according to claim 1, characterized in that, ​ ​ ​ ​ ​ 4. The deep learning-based hot isostatic pressing diffusion bonding parameter optimization method according to claim 1, characterized in that, ​ ​ ​ ​ ​ ​ ​ ​ 5. The deep learning-based hot isostatic pressing diffusion bonding parameter optimization method according to claim 1, characterized in that, The improved MLP model comprises: an input layer, a weighted preprocessing layer, a hidden layer, and an output layer connected in sequence.

6. The deep learning-based hot isostatic pressing diffusion bonding parameter optimization method according to claim 1, characterized in that, The scheme input into the result prediction model is iterated by using a preset PID controller until the candidate prediction result meets the preset requirement, and the hot isostatic pressing diffusion welding optimization parameter is obtained, including: The candidate prediction results are set as the control target, and the PID input variables are set according to the control target; the expression of the PID input variables is: ;in, For the first Control deviation during the next iteration; The number of target indicators; For the first The weight of each indicator; For the first The result prediction model output during the iteration is the first... One indicator; For the first The preset target value for each indicator; The PID controller is constructed, and an expression of the PID controller is: ； wherein, is a parameter adjustment amount; , , are a proportional coefficient, an integral coefficient, and a differential coefficient, respectively; is is a bias integral of the nth iteration; is a bias change rate; It is judged whether the PID input variable meets the preset requirement, and a deviation degree state is obtained; When the deviation degree state does not meet the requirement, the PID input variable is input into the PID controller, and the parameter adjustment amount is obtained; The parameter adjustment amount is updated into the scheme of the result prediction model, and the step of "judging whether the PID input variable meets the preset requirement, and obtaining a deviation degree state" is returned until the deviation degree state meets the requirement, and the last updated scheme of the result prediction model is output, and the hot isostatic pressing diffusion welding optimization parameter is obtained.

7. The deep learning-based hot isostatic pressing diffusion bonding parameter optimization method according to claim 5, characterized in that, The improved MLP model is trained by using the original training data, and an error compensation model is obtained, including: The improved MLP model is constructed; The original training data is received by the input layer; The weighted preprocessing layer performs pairwise weighted calculations on the data in each group of the original training data to obtain weighted interactive data; the expression for the weighted preprocessing layer is: ;in, The first data in the same set of the original training data The data and the first The weighted interactive data of each data point; This is a learnable interaction weight matrix; , The first data point in the same group of the original training data is respectively the second data point in the original training data. The data and the first One data point; The weighted interaction data and the original training data are synchronously input into two full connection networks connected in sequence in the hidden layer for non-linear mapping, and extraction features are obtained; The extraction features are input into the output layer for multi-index error prediction, and a model prediction result is obtained; The hidden layer and the weighted preprocessing layer are iteratively optimized by back propagation according to the model prediction result and the theoretical actual error data, and the error compensation model is obtained.

8. A deep learning-based hot isostatic pressing diffusion welding parameter optimization system, characterized in that, It comprises: A physical model construction module is configured to construct a hot isostatic pressing diffusion physical model. A theoretical data collection module is configured to set a plurality of experimental schemes, input each of the experimental schemes into the hot isostatic pressing diffusion physical model, and obtain a theoretical prediction result. An actual data collection module is configured to perform a hot isostatic pressing diffusion welding experiment according to the experimental schemes, and obtain an actual detection result. A training data construction module is configured to extract errors of the theoretical prediction result and the actual detection result, obtain theoretical actual error data, and align and integrate the theoretical actual error data, the theoretical prediction result, and the experimental schemes to obtain original training data. A compensation model construction module is configured to train an improved MLP model by using the original training data, and obtain an error compensation model. A model fusion module is configured to fuse the hot isostatic pressing diffusion physical model and the error compensation model, and obtain a result prediction model. A result prediction module is configured to set a parameter range, randomly select a scheme in the parameter range, input the scheme into the result prediction model, and obtain a candidate prediction result. A parameter optimization module is configured to iterate the scheme input into the result prediction model by using a preset PID controller until the candidate prediction result meets a preset requirement, and obtain a hot isostatic pressing diffusion welding optimization parameter.

9. An electronic device, comprising: It comprises: At least one processor, and a memory connected in communication with the processor; wherein the memory stores instructions capable of being executed by the processor, and the instructions are executed by the processor to enable the processor to perform the deep learning-based hot isostatic pressing diffusion welding parameter optimization method in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to enable the computer to perform the deep learning-based hot isostatic pressing diffusion welding parameter optimization method in any one of claims 1 to 7.

Images