Method for analyzing surface differences of a metallic material

Abstract

The application provides a metal material surface difference analysis method, comprising the following steps: according to the chemical composition and initial microstructure of the material, calculating the Gibbs free energy at different temperatures by using a thermodynamic model, drawing a time-temperature-transformation curve of the material by using phase diagram analysis software, and obtaining grain growth kinetics parameters and a recrystallization temperature interval; inputting data information of strength, hardness and toughness in different application scenarios by using a multi-objective optimization model, searching for an optimal grain size distribution by using a genetic algorithm, and obtaining an optimal microstructure balancing various performance indexes; reversely designing a heat treatment process route according to the optimal microstructure, simulating the temperature field and stress field distribution in the heat treatment process by using finite element software, and determining the control parameters of heating, holding and cooling through iterative optimization.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a metal material surface difference analysis method. BACKGROUND

[0002] In the analysis of the microstructure of metal materials, grain size (i.e. granularity) is a key parameter. However, in practical applications, there is a thorny technical problem of how to maintain the strength and hardness of the material while not sacrificing its toughness too much. This problem stems from the dual influence of granularity on material performance. When the grain size is reduced through heat treatment or mechanical processing, the strength and hardness of the material will indeed increase, which is beneficial for parts that need to withstand high stress. However, at the same time, smaller grains will increase the grain boundary area, making the material more prone to cracking along the grain boundary when impacted, thereby reducing toughness. Conversely, if the grain size is increased, the toughness of the material will be improved, which is beneficial for components that need to resist fatigue failure. However, large grains will reduce the resistance of grain boundaries to dislocation movement, making the material more prone to plastic deformation and thus reducing strength and hardness. This contradictory relationship between granularity and performance is particularly pronounced in certain special applications. For example, in the aerospace field, materials need to have sufficient strength to withstand high temperature and high pressure environments, and also need to maintain a certain toughness to prevent sudden failure. How to find a balance between these two seemingly contradictory requirements has become a major challenge for materials scientists. SUMMARY

[0003] The present application provides a metal material surface difference analysis method, mainly comprising:

[0004] According to the chemical composition and initial microstructure of the material, the Gibbs free energy at different temperatures is calculated using a thermodynamic model, and the time-temperature-transformation curve of the material is plotted using phase diagram analysis software to obtain grain growth kinetics parameters and recrystallization temperature intervals;

[0005] According to the grain growth kinetics parameters and recrystallization temperature intervals, a quantitative relationship model between grain size and heat treatment process parameters is established, and the quantitative relationship model is trained through a neural network algorithm. The target grain size range is input, and the corresponding heat treatment temperature, time and cooling rate are inversely deduced;

[0006] According to the deduced heat treatment process parameters, heat treatment is performed through a small test furnace control system, the microstructure of the sample is observed using a metallographic microscope, the grain size distribution is measured using image analysis software, and statistical analysis is used to determine whether the actual grain size is within the target range;

[0007] If the actual grain size is within the target range, the mechanical properties of the heat-treated sample are tested, including tensile, impact and hardness, and the correlation between grain size and strength, hardness and toughness is analyzed using data mining algorithms, and a multi-objective optimization model of material properties and microstructure is established;

[0008] Using the established multi-objective optimization model, input the data information of strength, hardness and toughness under different application scenarios, search for the best grain size distribution through genetic algorithm, and get the optimal microstructure balancing each performance index;

[0009] According to the optimal microstructure, the heat treatment process route is designed reversely, the temperature field and stress field distribution in the heat treatment process are simulated by using finite element software, and the control parameters of heating, holding and cooling are determined through iterative optimization;

[0010] Implement the optimized heat treatment process on the industrial production line, use the online monitoring system to collect the temperature, stress and displacement of the workpiece in real time, analyze through the edge computing device, and if an abnormality is detected, adjust the process parameters of the heat treatment equipment to realize closed-loop control.

[0011] The technical scheme provided by the embodiment of the present application can include the following beneficial effects:

[0012] The metal material surface difference analysis method provided by the present application can accurately predict and control the grain growth and recrystallization behavior of the material by calculating the Gibbs free energy of the material at different temperatures and using phase diagram analysis, and ensure that the microstructure meets the performance requirements; by establishing a relationship model between grain size and heat treatment parameters, the best heat treatment conditions can be deduced reversely according to the target grain size, improving the process efficiency and consistency; after the small test furnace heat treatment, the grain size is verified by metallographic analysis and statistical method whether it meets the expectation, and the process parameters are adjusted in time to ensure the product quality; the mechanical properties of the material after heat treatment are tested, and the relationship between grain size and performance is analyzed using data mining technology, a multi-objective optimization model is established, and the comprehensive performance of the material is further improved; by generating the heat treatment process route corresponding to the optimal microstructure, the number of tests is reduced, and the research and development process is accelerated; implement the optimized heat treatment process on the production line, combine online monitoring and edge computing technology, realize real-time feedback and automatic adjustment, improve production efficiency and quality control level; through the implementation of closed-loop control strategy, the heat treatment parameters are dynamically adjusted based on real-time data, ensuring long-term stable production of high-quality products, while collecting data for continuous optimization of the model and process. BRIEF DESCRIPTION OF DRAWINGS

[0013] Figure 1 The flowchart of the metal material surface difference analysis method of the present application.

[0014] Figure 2A schematic diagram of a metal material surface difference analysis method of the present application.

[0015] Figure 3 Another schematic diagram of a metal material surface difference analysis method of the present application. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the present application will be described clearly and in detail below with reference to the drawings in the embodiments of the present application. The described embodiments are only some of the embodiments of the present application.

[0017] As Figure 1 -3, the metal material surface difference analysis method of the present embodiment can specifically include:

[0018] In step S101, according to the chemical composition and initial microstructure of the material, the Gibbs free energy at different temperatures is calculated by using a thermodynamic model, and the time-temperature-transformation curve of the material is plotted by using phase diagram analysis software to obtain the grain growth kinetics parameters and the recrystallization temperature interval.

[0019] The CALPHAD method is used for thermodynamic calculation, the thermodynamic calculation is used to calculate the Gibbs free energy at different temperatures according to the chemical composition and initial microstructure data of the material to obtain the volume fraction and chemical composition of each phase. According to the results of the thermodynamic calculation, the continuous cooling transformation diagram and isothermal transformation diagram of the material are plotted by using JMatPro software to obtain the phase transition starting temperature and phase transition completion temperature. For the continuous cooling transformation diagram and isothermal transformation diagram, the Johnson-Mehl-Avrami-Kolmogorov equation is used to analyze the time-temperature-transformation curve to calculate the grain growth index and rate constant and determine the grain growth activation energy. If the grain growth kinetics parameters are obtained, the recrystallization starting temperature and completion temperature are determined by calculating the grain growth rate at different temperatures, and the recrystallization temperature interval of the material is divided. According to the recrystallization temperature interval, the heat treatment process parameters including the heating rate and holding time are designed, and the grain size distribution under different heat treatment process conditions is calculated.

[0020] For example, according to the chemical composition and initial microstructure of the material, the CALPHAD method is used for thermodynamic calculation, the chemical composition and initial microstructure data of the material are input, the Gibbs free energy at different temperatures is calculated, and the volume fraction and chemical composition of each phase are obtained. The above calculation results are imported into the JMatPro software to draw the continuous cooling transformation diagram and isothermal transformation diagram of the material, and the phase change starting temperature and phase change completion temperature are obtained. The Johnson-Mehl-Avrami-Kolmogorov equation is used to analyze the time-temperature-transformation curve, the grain growth index and rate constant are calculated, and the grain growth activation energy is determined. According to the grain growth kinetics parameters, the grain growth rate at different temperatures is calculated to determine the recrystallization starting temperature and completion temperature, and the recrystallization temperature interval of the material is divided. According to the recrystallization temperature interval, the heat treatment process parameters such as heating rate and holding time are designed, and the grain size distribution under different heat treatment process conditions is calculated. In the CALPHAD method, the input Fe-0.2C-1.5Mn-0.5Si is the weight percentage, the chemical composition and initial ferrite+pearlite microstructure data of the alloy, and the thermodynamic database is used to calculate the Gibbs free energy every 50℃ in the temperature range of 700-1200℃. By minimizing the total Gibbs free energy, the volume fraction and chemical composition of each phase such as ferrite, austenite and cementite in the equilibrium state are obtained. These data are imported into the JMatPro software to generate a continuous cooling transformation diagram, which shows the curve of cooling rate from 0.1℃ / s to 100℃ / s. At the same time, an isothermal transformation diagram is drawn, with a temperature range of 400-900℃ and a time range of 1-10000 seconds. From the diagram, the austenitizing starting temperature Ac1 is 720℃ and the completion temperature Ac3 is 850℃. The Johnson-Mehl-Avrami-Kolmogorov equation ln[-ln(1-f)]=nlnt+lnk is used to analyze the isothermal transformation curve, where f is the transformation fraction, t is the time, n is the Avrami index, and k is the rate constant. Through logarithmic fitting, n=2.5 is obtained, indicating that the grain growth is mainly controlled by three-dimensional diffusion. The k value at different temperatures is calculated, and the relationship diagram of lnk and 1 / T is drawn, and the slope is the grain growth activation energy, which is Q=280kJ / mol. According to the grain growth rate equation v=kexp(-Q / RT), the grain growth rate every 50℃ in the temperature range of 600-1000℃ is calculated. When v reaches the preset threshold value 0.1μm / s, the recrystallization starting temperature is determined to be 750℃; when v drops to 0.01μm / s, the recrystallization completion temperature is determined to be 950℃. Based on the recrystallization temperature interval 750-950℃, the heat treatment process parameters are designed as heating rate 5℃ / s, holding temperature 900℃ and holding time 1800 seconds.The average grain size after heat treatment is calculated using the extended Burke-Turnbull equation d = (kt) ^ (1 / n), wherein d is the grain size, k is the temperature-dependent rate constant, t is the time, and n is the grain growth exponent. The average grain size after heat treatment is finally obtained as 35 μm, and the grain size distribution conforms to a lognormal distribution with a standard deviation of 0.2.

[0021] In step S102, a quantitative relationship model between the grain size and the heat treatment process parameters is established according to the grain growth kinetics parameters and the recrystallization temperature interval, the quantitative relationship model is trained by a neural network algorithm, and the corresponding heat treatment temperature, time and cooling rate are inversely deduced by inputting a target grain size range.

[0022] The grain growth kinetics parameters and the recrystallization temperature interval, including the activation energy and the frequency factor, are obtained. An Arrhenius equation is constructed according to the grain growth kinetics parameters and the recrystallization temperature interval. A Levenberg-Marquardt algorithm is used to perform nonlinear regression on the Arrhenius equation to obtain a quantitative relationship function between the grain size and the heat treatment temperature, time and cooling rate. Experimental data are collected, including grain size measurement results under different alloy compositions, heat treatment temperatures, times and cooling rates. The experimental data are proportionally divided into a training set and a validation set. A multilayer perceptron neural network is set, including an input layer, a hidden layer and an output layer. An Adam optimizer and a mean square error loss function are used to train the multilayer perceptron neural network to obtain a trained quantitative relationship model. A target grain size range is preset. If the target grain size range is a preset value, a genetic algorithm is used for inverse calculation. The fitness function of the genetic algorithm is to minimize the error between the predicted grain size and the target size. The heat treatment temperature, time and cooling rate parameter combination are obtained by the genetic algorithm. The parameter combination is used to generate grains in the target grain size range.

[0023] Illustratively, based on the grain growth kinetics parameters and the recrystallization temperature interval, including the activation energy and the frequency factor, an Arrhenius equation is constructed, a Levenberg-Marquardt algorithm is used for nonlinear regression, a quantitative relationship function between the grain size and the heat treatment temperature, time and cooling rate is fitted, and a preliminary mathematical model is obtained. 1000 sets of experimental data are collected, covering grain size measurements under different alloy compositions, heat treatment temperatures such as 600-1200°C, times such as 1-10000 seconds and cooling rates such as 0.1-100°C / s. The data set is divided into a training set and a validation set in a ratio of 8:2. A multilayer perceptron neural network is constructed, with an input layer of 4 nodes corresponding to the alloy composition, temperature, time and cooling rate, 2 hidden layers each with 20 nodes using a ReLU activation function, and an output layer of 1 node corresponding to the grain size. An Adam optimizer is used with an initial learning rate of 0.001 and a batch size of 32, and the hyperparameters are optimized through 5-fold cross-validation. After 2000 rounds of training, a mean square error of 0.05 μm2 is achieved on the validation set. Finally, a genetic algorithm is used for inverse calculation with a target grain size range of 10-50 μm. The population size is set to 100, the maximum number of iterations is 1000, the crossover probability is 0.8, and the mutation probability is 0.1.The fitness function is defined as |d predicted - d target | / d target, where d predicted is the grain size predicted by the neural network and d target is the target grain size. After 950 iterations, the optimal solution is obtained as a heat treatment temperature of 920°C, a time of 3600 seconds, a cooling rate of 5°C / s, and a predicted grain size of 32 μm, which meets the target range requirement.

[0024] In step S103, the heat treatment is performed in the small-scale furnace control system according to the derived heat treatment process parameters, the microstructure of the sample is observed by a metallographic microscope, the grain size distribution is measured by using image analysis software, and whether the actual grain size is within the target range is determined by statistical analysis.

[0025] According to the heat treatment process parameters, a PID control algorithm is used to set the heat treatment temperature, holding time and cooling rate of the small-scale furnace control system; the sample is heat treated by the small-scale furnace control system to obtain a heat-treated sample. An automatic grinding and polishing machine is used to cut, inlay, grind and polish the heat-treated sample; if the polishing precision reaches a preset threshold, a metallographic observation sample is obtained. A metallographic microscope automatic scanning system is used to scan the surface of the metallographic observation sample; the scanning results are processed by a scale-invariant feature transformation algorithm to generate a microstructure panoramic image. An image analysis software is used to perform image segmentation and grain recognition on the microstructure panoramic image; a watershed algorithm is used for image segmentation, and a roundness and area threshold method is used to recognize grains to measure the grain size distribution. The average grain size and standard deviation of the grain size distribution are calculated by statistical analysis; if the average grain size is within the target range and the standard deviation is less than a preset percentage of the target range, the metallographic observation sample is determined to be qualified.

[0026] For example, based on the derived heat treatment process parameters, the heat treatment temperature, holding time, and cooling rate were set using a PID control algorithm in a small-scale test furnace control system, with accuracies of ±5℃, ±1 minute, and ±0.1℃ / s, respectively. The sample was then heat-treated, and a temperature sensor was used to monitor the temperature change curve in real time. An automatic grinding and polishing machine was used to cut, mount, grind, and polish the heat-treated sample, ultimately polishing it to a precision of 0.05μm to obtain a sample for metallographic observation. This sample was then etched using a 3% nitric acid-alcohol solution in an electrolytic etching device for 10–15 seconds. An automatic scanning system for a metallographic microscope was used to scan the sample surface, acquiring microstructure images from multiple fields of view. A panoramic microstructure image with a resolution of 4096×3072 pixels was generated using a scale-invariant feature transform algorithm. Image analysis software was used to segment and identify grains in the panoramic image of the microstructure. The watershed algorithm was used for image segmentation, and the roundness and area thresholding methods were used to identify grains. The grain size distribution was measured, and the average grain size and standard deviation were calculated through statistical analysis. If the average grain size was within the target range and the standard deviation was less than 10% of the target range, it was judged as qualified.

[0027] In practical operation, the pilot furnace control system adopts a PID control algorithm, setting the heat treatment temperature to 920℃, the holding time to 60 minutes, and the cooling rate to 5℃ / s. The PID parameters have been optimized, with a proportional gain Kp = 10, integral time Ti = 120s, and derivative time Td = 30s, achieving a control accuracy of ±3℃.

[0028] ±30 seconds and ±0.05℃ / s. The temperature sensor sampling frequency was 1Hz, recording the entire temperature curve. After heat treatment, an automatic grinding and polishing machine used 240#, 400#, 600#, 800#, and 1200# sandpaper for rough grinding, followed by fine grinding with 3μm and 1μm diamond suspensions, and finally polishing to a mirror finish with 0.05μm alumina suspension. The electrolytic etching device used a 3% nitric acid alcohol solution, a voltage of 3V, a current of 0.2A, and etched for 12 seconds. The metallurgical microscope's automatic scanning system was set to a 50x objective lens with a scanning step size of 0.1mm, acquiring a total of 100×100 field-of-view images. The scale-invariant feature transformation algorithm was set with a feature point matching threshold of 0.7 and a minimum number of matching points of 4, generating a panoramic image of 4096×3072 pixels. The image analysis software used a watershed algorithm for segmentation, setting a roundness threshold of 0.6 and an area threshold of 50-5000 pixels to identify grains. Statistical analysis showed an average grain size of 35 μm and a standard deviation of 3 μm. The target range was 30–40 μm, and a standard deviation less than 1 μm was considered acceptable. The entire process was highly automated, requiring no manual intervention, and took approximately 120 minutes from heat treatment to result determination.

[0029] Step S104, if the actual grain size is within the target range, the mechanical properties of the heat-treated sample are tested, including tensile, impact and hardness, and a data mining algorithm is used to analyze the correlation between grain size and strength, hardness and toughness, and a multi-objective optimization model of material properties and microstructure is established.

[0030] The heat-treated sample is obtained, and tensile, impact and hardness test samples are prepared according to ASTM standards. The sample is tested to obtain mechanical property data such as yield strength, tensile strength, elongation, impact toughness and Vickers hardness. Image analysis software is used to process the metallographic structure image of the sample to extract microstructure characteristic parameters such as average grain size, grain size distribution standard deviation and grain shape factor. The correlation between grain size and strength, hardness and toughness is analyzed by random forest algorithm for the mechanical property data and microstructure characteristic parameters. If the number of trees is a preset value and the minimum leaf node sample size is a preset value, the best parameters are selected by cross-validation. According to the correlation and influence weight, a multi-objective optimization model of material properties and microstructure is constructed. The objective function is set to maximize strength and toughness while minimizing hardness deviation. Genetic algorithm is used to find the optimal combination of heat treatment process parameters to obtain the optimized microstructure parameters. According to the optimized microstructure parameters, the corresponding heat treatment temperature, time and cooling rate are obtained by reverse calculation.

[0031] Illustratively, 30 tensile specimens are prepared according to ASTM E8 standard, 30 impact specimens are prepared according to ASTM E23 standard, and 30 hardness test samples are prepared using the automated testing equipment, and tensile, impact, and hardness tests are performed on the heat treated specimens to obtain mechanical property data such as yield strength, tensile strength, elongation, impact toughness, and Vickers hardness. Image analysis software is used to process the metallographic images of the specimens to extract microstructure characteristic parameters such as average grain size, grain size distribution standard deviation, and grain shape factor such as circularity defined as 4π x area / perimeter2. A random forest algorithm is used to analyze the correlation between grain size and strength, hardness, and toughness, with 500 trees set as the number of trees, 5 set as the minimum leaf node sample size, cross-validation used to select the best parameters, root mean square error and determination coefficient used to evaluate model performance, and the influence weight of each microstructure characteristic on mechanical properties calculated. Based on the obtained correlation and influence weight, a multi-objective optimization model of material performance and microstructure is constructed, with the objective function set as maximizing strength and toughness while minimizing hardness deviation, a genetic algorithm used to find the optimal heat treatment process parameter combination, with 100 set as the population size, 0.8 set as the crossover probability, 0.1 set as the mutation probability, and 1000 set as the number of iterations. The optimized microstructure parameters are obtained through reverse calculation to obtain the corresponding heat treatment temperature, time, and cooling rate. In actual operation, 30 tensile specimens with a 10 mm gauge length are prepared according to ASTM E8 standard, 30 V-notch impact specimens with a 10 x 10 x 55 mm size are prepared according to ASTM E23 standard, and 30 hardness test samples with a 15 x 15 x 10 mm size are prepared using the automated testing equipment. A loading rate of 1 mm / min is used for the tensile test to obtain a yield strength of 480 MPa, a tensile strength of 620 MPa, and an elongation of 18%. The impact test is performed at room temperature to obtain an average impact toughness of 85 J / cm2. A 10 kg load is used for the hardness test to obtain an average Vickers hardness of 205 HV10. Image analysis software is used to process 1000 x 1000 pixel metallographic images to extract an average grain size of 35 pm, a grain size distribution standard deviation of 5 pm, and an average grain circularity of 0.85. The random forest algorithm is set with 500 decision trees and a minimum leaf node sample size of 5, and the best parameters are selected through 5-fold cross-validation. Model performance evaluation results in a root mean square error of 15 MPa for strength, 5 J / cm2 for toughness, and 8 HV for hardness, and determination coefficients R2 of 0.92, 0.88, and 0.90, respectively. Impact weight analysis shows that grain size has the greatest impact on strength, such as 60%, followed by toughness, such as 30%, and a smaller impact on hardness, such as 10%. The multi-objective optimization model sets the objective function as maximizing 0.6 x strength + 0.3 x toughness - 0.1 x | hardness - target hardness |. The genetic algorithm uses a population of 100 individuals, a crossover probability of 0.8, a mutation probability of 0.1, and 1000 iterations.The final optimization result obtains the best microstructure parameters of average grain size 30 μm, standard deviation 3 μm, and roundness 0.9. Reverse calculation obtains the corresponding heat treatment process parameters of heating temperature 900 ℃, holding time 45 minutes, and cooling rate 10 ℃ / s.

[0032] In step S105, the data information of strength, hardness and toughness in different application scenarios is input by using the established multi-objective optimization model, the genetic algorithm is used to search the best grain size distribution, and the optimal microstructure balancing various performance indicators is obtained.

[0033] According to the target values and weights set according to the application scenarios, the comprehensive performance evaluation function F is constructed, wherein F = w1 * (σ / σ_target) + w2 * (H / H_target) + w3 * (K / K_target), σ, H and K represent strength, hardness and toughness respectively, w1, w2 and w3 are corresponding weights, and target is a target value; the grain size distribution parameter population is initialized by using the genetic algorithm, including average grain size, standard deviation, roundness average value and roundness standard deviation; the fitness value of each individual is calculated by using the multi-objective optimization model, a new generation population is generated by using roulette selection, single-point crossover and Gaussian mutation operation; if the change amplitude of the optimal solution of continuous generations is less than a preset threshold or reaches a maximum iteration number, the individual with the highest fitness in the final population is selected as the optimal solution; the best grain size distribution parameters balancing various performance indicators are obtained, and the optimal microstructure is obtained; according to the multi-element regression mapping relationship between the heat treatment process parameters and the grain size distribution parameters established in advance, the heat treatment temperature, time and cooling rate are reversely calculated.

[0034] For example, the target values and weights of strength, hardness and toughness are set according to different application scenarios, and the comprehensive performance evaluation function F is constructed

[0035] F = w1 * (σ / σ_target) + w2 * (H / H_target) + w3 * (K / K_target), where σ, H, K represent strength, hardness and toughness, respectively, w1, w2, w3 are the corresponding weights, and target is the target value. The target value and weight are input into the multi-objective optimization model. The population of grain size distribution parameters is initialized using genetic algorithm, including average grain size, standard deviation, average circularity and circularity standard deviation. The population size is set to 100, the crossover probability is 0.8, and the mutation probability is 0.1. The fitness value of each individual is calculated by the multi-objective optimization model, and the roulette selection, single-point crossover and Gaussian mutation operations are performed to generate a new generation of population. The iteration process is repeated until the variation amplitude of the optimal solution is less than 1% for 50 consecutive generations or the maximum iteration number is reached, which is 1000 generations. The individual with the highest fitness value is selected from the final population as the optimal solution, and the best grain size distribution parameters balancing each performance index are obtained, and the optimal microstructure is obtained. The mapping relationship between the heat treatment process parameters and the grain size distribution parameters is established by multiple regression, and the corresponding heat treatment temperature, time and cooling rate are calculated reversely. In practical application, for the application scenario of automobile suspension spring, the strength target value is set to 1200 MPa, the hardness target value is set to 400 HV, and the toughness target value is set to 50 J / cm2. The weights are 0.5, 0.3 and 0.2, respectively. The comprehensive performance evaluation function is constructed

[0036] F = 0.5 * (σ / 1200) + 0.3 * (H / 400) + 0.2 * (K / 50). Genetic algorithm initializes a population of 100 individuals, each containing an average grain size of 10-50 μm, a standard deviation of 1-10 μm, an average circularity of 0.6-1.0, and a circularity standard deviation of 0.05-0.2. The crossover probability is set to 0.8 and the mutation probability is set to 0.1. The multi-objective optimization model uses the previously trained random forest algorithm to predict the performance indicators based on the grain parameters and calculate the fitness value. Roulette selection method is used to select individuals, and single-point crossover and Gaussian mutation are performed, such as mean 0 and standard deviation 5% of the parameter range. During the iteration process, the optimal solution gradually improves from F = 0.85 to F = 0.98 at the 782th generation, and the variation amplitude is less than 1% for 50 consecutive generations, meeting the convergence condition. The optimal grain parameters are obtained as follows: average size 28 μm, standard deviation 4 μm, average circularity 0.85, and circularity standard deviation 0.08. The multiple regression model y = β0 + β1x1 + β2x2 + β3x3 is used for reverse calculation, where y is the grain parameter and x is the heat treatment parameter. The best heat treatment process is temperature 940℃, time 40 minutes and cooling rate 15℃ / s.

[0037] Step S106, according to the optimal microstructure, reverse design heat treatment process route, using finite element software to simulate the temperature field and stress field distribution in heat treatment process, through iterative optimization to determine the heating, holding and cooling control parameters.

[0038] According to the optimal microstructure parameters, an initial heat treatment process route is obtained from the heat treatment process database, the initial heat treatment process route including heating rate, holding temperature, holding time and cooling rate; a geometric model and grid division of the heat treatment process are established using the initial heat treatment process route, and temperature field and stress field distribution are obtained; for the temperature field and the stress field distribution, a root mean square error with the target microstructure is calculated by gradient descent method; if the root mean square error is greater than a preset threshold, the heat treatment process parameters are adjusted, and finite element simulation is repeated until the root mean square error is less than the preset threshold; a heating curve, a holding curve and a cooling curve are generated according to the optimized heat treatment process parameters, and a heat treatment process control scheme is obtained; small batch tests are carried out using the heat treatment process control scheme, and an actual microstructure is obtained; it is judged whether the deviation between the actual microstructure and the target microstructure exceeds a preset threshold, and if it exceeds, the heat treatment process parameters are adjusted.

[0039] For example, according to the optimal microstructure parameters, the initial heat treatment process route including heating rate, holding temperature, holding time and cooling rate is reversely queried from the heat treatment process database constructed based on experimental data and empirical models, as the initial input parameters for finite element simulation. The geometric model and mesh partitioning of the heat treatment process are established using ANSYS software, the material physical parameters such as thermal conductivity, specific heat capacity, density, boundary conditions and initial conditions are set, and the temperature field and stress field distribution in the heating, holding and cooling stages are simulated. The difference between the simulation results and the target microstructure is compared by gradient descent method, the heat treatment process parameters are adjusted, and the finite element simulation is repeated until the root mean square error of the temperature field and stress field distribution and the target microstructure is less than 5%. According to the optimized heat treatment process parameters, the accurate heating curve, holding curve and cooling curve are generated using piecewise function to form a complete heat treatment process control scheme. Through small batch test and microstructure analysis, the actual effect of the optimized heat treatment process is verified, and the deviation between the actual obtained microstructure and the target structure is compared. If the deviation exceeds the preset threshold, the optimization step is returned for reiteration. In practical application, for a high-strength low-alloy steel, the optimal microstructure parameters are an average grain size of 30 μm and a standard deviation of 5 μm. Using a heat treatment process database containing 1000 sets of experimental data, the initial heat treatment process route is reversely queried by interpolation algorithm as a heating rate of 5°C / s, a holding temperature of 900°C, a holding time of 30 minutes and a cooling rate of 10°C / s. A geometric model of a 100x100x20mm sample is established in ANSYS, hexahedral mesh partitioning is used, and the number of elements is 50000. The material physical parameters are set as thermal conductivity 45 W / (m·K), specific heat capacity 480 J / (kg·K), and density 7850 kg / m2. The boundary conditions include convective heat transfer coefficient 20 W / (m2·K) and radiation emissivity 0.8. The transient heat analysis module is used to simulate the heat treatment process, the time step is 1s, and the total simulation time is 3600s. The simulation results show that the maximum temperature reaches 898°C and the maximum thermal stress is 250 MPa. The parameters are optimized by gradient descent method, the learning rate is set to 0.01, and after 50 iterations, the root mean square error of the temperature field and stress field distribution and the target microstructure is reduced to 4.8%. The optimized process parameters are a heating rate of 5.5°C / s, a holding temperature of 910°C, a holding time of 35 minutes and a cooling rate of 9.5°C / s. The accurate control curve of 1800 discrete points is generated using cubic spline interpolation. In small batch test, 10 samples are heat treated, and the metallographic analysis results show that the average grain size is 29.5 μm and the standard deviation is 5.2 μm, with a deviation of less than 3% from the target structure, verifying the effectiveness of the optimized process.

[0040] Step S107, the optimized heat treatment process is implemented on the industrial production line, the online monitoring system is used to collect the temperature, stress and displacement of the workpiece in real time, analysis is performed through the edge computing device, if an abnormality is detected, the process parameters of the heat treatment equipment are adjusted to realize closed-loop control.

[0041] Real-time data of the workpiece is acquired, the real-time data is collected by temperature sensors, stress sensors and displacement sensors installed on the workpiece on the industrial production line; according to the real-time data, an isolation forest algorithm is used to calculate the deviation between the current process parameters and the set value; it is judged whether the deviation exceeds a preset threshold value, if the deviation exceeds the preset threshold value, it is determined as abnormal information; for the abnormal information, a process parameter adjustment instruction is generated according to a pre-established adjustment rule library; the process parameter adjustment instruction is sent to the control system of the heat treatment equipment, after the control system of the heat treatment equipment receives the process parameter adjustment instruction, at least one process parameter of heating power, holding time or cooling rate is updated within the limits of the set adjustment range and the limit of the number of continuous adjustments; the updated process parameters of the heat treatment equipment are acquired, the control accuracy and response time are calculated; the performance of the closed-loop control system is evaluated according to the control accuracy and response time.

[0042] For example, 3 temperature sensors, 2 stress sensors and 1 displacement sensor are installed on each workpiece on the industrial production line to construct a real-time data acquisition network, the collected workpiece temperature, stress and deformation data are transmitted to the edge computing device through a high-speed industrial bus. The edge computing device uses an isolation forest algorithm to quickly analyze the collected real-time data, calculates the deviation between the current process parameters and the set value, and determines that it is an abnormal situation when the deviation exceeds a preset threshold value. If an abnormality is detected, the edge computing device automatically generates a process parameter adjustment instruction according to an adjustment rule library based on expert experience and historical data analysis, and sends the instruction to the control system of the heat treatment equipment through a control interface. After receiving the adjustment instruction, the heat treatment equipment updates the process parameters such as heating power, holding time or cooling rate in real time within the limits of the set adjustment range and the limit of the number of continuous adjustments, realizes closed-loop control of the production process, and feeds back the adjusted parameters to the monitoring system for recording and analysis. The performance of the entire closed-loop control system is evaluated by calculating the control accuracy and response time, etc., the abnormality detection algorithm and the adjustment rule library are optimized according to the evaluation results, and the control effect of the system is continuously improved. In practical application, 3 K-type thermocouples are installed on the surface of each workpiece to measure the range of -200-1300℃, the accuracy is ±1.5℃, 2 strain gauges are used to measure the stress, the range is 0-1000MPa, the accuracy is ±1%, and 1 displacement sensor is used to measure the displacement, the range is 0-1000μm, the accuracy is ±1μm.

[0043] ±3000 με, precision 0.1%, 1 laser displacement sensor like range 0-100 mm, resolution 0.01 mm. Data is transmitted to the edge computing device using EtherCAT bus like transmission rate 100 Mbps. The edge computing device uses the Isolation Forest algorithm like number of trees 100, subsample size 256, and determines an anomaly when the anomaly score exceeds 0.6. The tuning rule base contains 50 rules like "if temperature deviation > +10°C, then decrease heating power by 5%". The PID controller of the heat treatment device receives the tuning instructions and adjusts within a ±10% tuning amplitude limit, with a maximum of 3 consecutive adjustments. The heating power adjustment precision is 0.1 kW, the soak time adjustment precision is 1 second, and the cooling rate adjustment precision is 0.1 °C / s. The monitoring system records the tuning parameters every 10 seconds and calculates the average control deviation like target < ±5°C and the response time like target < 30 seconds. If the average control deviation is < ±3°C and the response time is < 20 seconds for 100 consecutive batches, then the anomaly detection threshold is increased by 0.05 and the parameter values in the tuning rule base are updated.

[0044] It will be apparent to those skilled in the art that the application is not limited to the details of the foregoing exemplary embodiments and that the present application can be implemented in other specific forms without departing from the spirit or essential characteristics thereof. The presently disclosed embodiments are therefore considered in all respects to be illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalents are intended to be embraced therein. No feature of the application is to be construed as limiting the scope of the claims to their precise configuration set forth herein.

Claims

1. A method for analyzing the surface differences of metallic materials, characterized in that, The method includes: calculating the Gibbs free energy at different temperatures using a thermodynamic model based on the material's chemical composition and initial microstructure; plotting the material's time-temperature-transformation curve using phase diagram analysis software to obtain grain growth kinetic parameters and recrystallization temperature ranges; establishing a quantitative relationship model between grain size and heat treatment process parameters based on the grain growth kinetic parameters and recrystallization temperature ranges; training the quantitative relationship model using a neural network algorithm; and inputting the target grain size range. The corresponding heat treatment temperature, time, and cooling rate are derived through reverse engineering. The grain growth kinetic parameters include activation energy and frequency factor. Based on the derived heat treatment process parameters, heat treatment is carried out through a pilot furnace control system. The microstructure of the sample was observed using a metallographic microscope. Grain size distribution was measured using image analysis software, and statistical analysis was used to determine whether the actual grain size was within the target range. The heat treatment process parameters included heat treatment temperature, time, and cooling rate. If the actual grain size was within the target range, mechanical property tests, including tensile, impact, and hardness tests, were performed on the heat-treated sample. Data mining algorithms were used to analyze the correlation between grain size and strength, hardness, and toughness, establishing a multi-objective optimization model for material properties and microstructure. Using this model, data on strength, hardness, and toughness under different application scenarios were input, and a genetic algorithm was used to search for the optimal grain size distribution, obtaining the optimal microstructure that balances various performance indicators. Based on the optimal microstructure, the heat treatment process route was designed in reverse, and finite element software was used to simulate the temperature and stress field distribution during the heat treatment process. Iterative optimization was used to determine the control parameters for heating, holding, and cooling. The optimized heat treatment process was implemented on an industrial production line. An online monitoring system was used to collect the temperature, stress, and displacement of the workpiece in real time, and the data was analyzed using edge computing devices. If an anomaly was detected, the process parameters of the heat treatment equipment were adjusted to achieve closed-loop control.

2. The method according to claim 1, wherein, The process involves calculating the Gibbs free energy at different temperatures using a thermodynamic model based on the material's chemical composition and initial microstructure, and plotting the material's time-temperature-transformation curves using phase diagram analysis software to obtain grain growth kinetic parameters and recrystallization temperature ranges. This includes: performing thermodynamic calculations using the CALPHAD method, whereby the thermodynamic calculations calculate the Gibbs free energy at different temperatures based on the material's chemical composition and initial microstructure data to obtain the volume fraction and chemical composition of each phase; and using JMatPro software to plot the material's continuous cooling transformation diagram and isothermal transformation diagram based on the thermodynamic calculation results to obtain the phase transformation initiation temperature and phase transformation time. The time-temperature-transformation curves are analyzed using the Johnson-Mehl-Avrami-Kolmogorov equation for the continuous cooling transformation diagram and isothermal transformation diagram. The grain growth index and rate constant are calculated to determine the grain growth activation energy. If the grain growth kinetic parameters are obtained, the recrystallization initiation temperature and completion temperature are determined by calculating the grain growth rate at different temperatures, and the recrystallization temperature range of the material is divided. Based on the recrystallization temperature range, heat treatment process parameters are designed, including heating rate and holding time, and the grain size distribution under different heat treatment process conditions is calculated.

3. The method according to claim 1, wherein, The process involves establishing a quantitative relationship model between grain size and heat treatment process parameters based on grain growth kinetic parameters and recrystallization temperature ranges. This model is trained using a neural network algorithm. By inputting a target grain size range, the corresponding heat treatment temperature, time, and cooling rate are derived. This includes: acquiring grain growth kinetic parameters and recrystallization temperature ranges, including activation energy and frequency factor; constructing the Arrhenius equation based on the grain growth kinetic parameters and recrystallization temperature range; performing nonlinear regression on the Arrhenius equation using the Levenberg-Marquardt algorithm to obtain the quantitative relationship function between grain size and heat treatment temperature, time, and cooling rate; and collecting experimental data, including data from various... Grain size measurement results under alloy composition, heat treatment temperature, time, and cooling rate; the experimental data are divided into training and validation sets proportionally; a multilayer perceptron neural network is set up, including an input layer, a hidden layer, and an output layer; the multilayer perceptron neural network is trained using the Adam optimizer and the mean square error loss function to obtain a trained quantitative relationship model; a target grain size range is preset; if the target grain size range is a preset value, a genetic algorithm is used for reverse calculation; the fitness function of the genetic algorithm is to minimize the error between the predicted grain size and the target size; the combination of heat treatment temperature, time, and cooling rate parameters is obtained through the genetic algorithm; the parameter combination is used to generate grains within the target grain size range.

4. The method according to claim 1, wherein, The process involves heat treatment based on derived heat treatment process parameters, conducted through a small-scale furnace control system. The microstructure of the sample is observed using a metallographic microscope. Grain size distribution is measured using image analysis software, and statistical analysis is used to determine if the actual grain size is within the target range. This includes: setting the heat treatment temperature, holding time, and cooling rate of the small-scale furnace control system using a PID control algorithm based on the heat treatment process parameters; heat-treating the sample through the small-scale furnace control system to obtain a heat-treated sample; and cutting, mounting, grinding, and polishing the heat-treated sample using an automatic grinding and polishing machine. If the polishing accuracy reaches a preset threshold, then... A metallographic observation sample is obtained; the surface of the metallographic observation sample is scanned using an automatic scanning system of a metallographic microscope; the scanning results are processed using a scale-invariant feature transformation algorithm to generate a panoramic image of the microstructure; image analysis software is used to perform image segmentation and grain identification on the panoramic image of the microstructure; a watershed algorithm is used for image segmentation, and roundness and area thresholding methods are used to identify grains and measure the grain size distribution; the average grain size and standard deviation of the grain size distribution are calculated through statistical analysis; if the average grain size is within the target range and the standard deviation is less than a preset percentage of the target range, the metallographic observation sample is deemed qualified.

5. The method according to claim 1, wherein, If the actual grain size is within the target range, mechanical property tests, including tensile, impact, and hardness tests, are performed on the heat-treated sample. Data mining algorithms are used to analyze the correlation between grain size and strength, hardness, and toughness, establishing a multi-objective optimization model for material properties and microstructure. This includes: acquiring the heat-treated sample; preparing tensile, impact, and hardness test samples according to ASTM standards; testing the sample to obtain mechanical property data such as yield strength, tensile strength, elongation, impact toughness, and Vickers hardness; and processing the metallographic images of the sample using image analysis software to extract the average grain size, standard deviation of grain size distribution, and grain shape factor microstructure. The mechanical property data and microstructure characteristics are analyzed using a random forest algorithm to determine the correlation between grain size and strength, hardness, and toughness. If the number of trees and the minimum number of leaf node samples are preset values, cross-validation is used to select the optimal parameters. Based on the correlation and influence weights, a multi-objective optimization model for material properties and microstructure is constructed. The objective function is set to maximize strength and toughness while minimizing hardness deviation. A genetic algorithm is used to find the optimal combination of heat treatment process parameters to obtain optimized microstructure parameters. For the optimized microstructure parameters, the corresponding heat treatment temperature, time, and cooling rate are calculated in reverse.

6. The method according to claim 1, wherein, The process involves using a multi-objective optimization model, inputting data on strength, hardness, and toughness under different application scenarios, and searching for the optimal grain size distribution using a genetic algorithm to obtain the optimal microstructure that balances various performance indicators. This includes: constructing a comprehensive performance evaluation function F based on the target values ​​and weights set for the application scenario, where F = w1 * (σ / σ_target) + w2 * (H / H_target) + w3 * (K / K_target), where σ, H, and K represent strength, hardness, and toughness, respectively, w1, w2, and w3 are the corresponding weights, and target is the target value; and initializing the grain size distribution parameters using a genetic algorithm. Several populations are generated, including average grain size, standard deviation, average roundness, and standard deviation of roundness. The fitness value of each individual is calculated using the multi-objective optimization model, and a new generation of population is generated using roulette wheel selection, single-point crossover, and Gaussian mutation operations. If the variation of the optimal solution is less than a preset threshold or the maximum number of iterations is reached for several consecutive generations, the individual with the highest fitness is selected as the optimal solution from the final population. The optimal grain size distribution parameters that balance various performance indicators are obtained, resulting in the optimal microstructure. Based on the pre-established multivariate regression mapping relationship between the heat treatment process parameters and the grain size distribution parameters, the heat treatment temperature, time, and cooling rate are calculated in reverse.

7. The method according to claim 1, wherein, The process involves reverse-engineering a heat treatment process route based on the optimal microstructure, simulating the temperature and stress field distributions during heat treatment using finite element method (FEM) software, and determining the control parameters for heating, holding, and cooling through iterative optimization. This includes: obtaining an initial heat treatment process route from the heat treatment process database based on the optimal microstructure parameters; the initial heat treatment process route including heating rate, holding temperature, holding time, and cooling rate; establishing a geometric model and mesh generation for the heat treatment process using the initial heat treatment process route to obtain the temperature and stress field distributions; calculating the root mean square error (RMSE) between the temperature and stress field distributions and the target microstructure using the gradient descent method; if the RMS error is greater than a preset threshold, adjusting the heat treatment process parameters and repeating the finite element simulation until the RMS error is less than the preset threshold; generating heating, holding, and cooling curves based on the optimized heat treatment process parameters to obtain a heat treatment process control scheme; conducting small-batch experiments using the heat treatment process control scheme to obtain the actual microstructure; and determining whether the deviation between the actual microstructure and the target microstructure exceeds a preset threshold. If it does, the process returns to adjust the heat treatment process parameters.

8. The method according to claim 1, wherein, The optimized heat treatment process implemented on the industrial production line employs an online monitoring system to collect real-time data on the workpiece's temperature, stress, and displacement. This data is analyzed using edge computing devices. If an anomaly is detected, the process parameters of the heat treatment equipment are adjusted to achieve closed-loop control. This includes: acquiring real-time workpiece data, collected by temperature, stress, and displacement sensors installed on the workpiece on the industrial production line; calculating the deviation between the current process parameters and set values ​​using an isolated forest algorithm based on the real-time data; determining whether the deviation exceeds a preset threshold, and identifying it as an anomaly if so; generating process parameter adjustment instructions based on a pre-established adjustment rule library for the anomaly; sending the process parameter adjustment instructions to the heat treatment equipment's control system; and updating at least one process parameter—heating power and holding time or cooling rate—within set adjustment range and consecutive adjustment count limits; acquiring the updated process parameters of the heat treatment equipment and calculating control accuracy and response time; and evaluating the performance of the closed-loop control system based on the control accuracy and response time.

Images