Titanium alloy forging process parameter optimization method based on finite element

Abstract

The present application relates to the field of titanium alloy forging technology, in particular to a large-size titanium alloy forging process parameter optimization method based on finite elements. The method acquires titanium alloy basic related data to establish a finite element simulation model; performs finite element simulation calculation according to multiple groups of preset parameters to acquire simulation data at key positions; determines a safe forming interval and an optimal grain refinement interval based on the basic data, and obtains a reference comprehensive adaptation degree corresponding to each group of parameters in combination with the forming defect risk of the key positions; trains a proxy model based on this and calculates the prediction accuracy, obtains updated forging process parameters based on the prediction accuracy, the model comprehensive adaptation degree output by the proxy model and the parameter correlation direction; and obtains optimal forging process parameters based on the updated forging process parameters, the finite element simulation model and the proxy model. The present application can effectively balance the macro forming quality and microstructure uniformity of large-size forgings, and effectively reduce the trial production scrap rate.

Description

Technical Field

[0001] This invention relates to the field of titanium alloy forging technology, and specifically to a method for optimizing forging process parameters of large-size titanium alloy forgings based on the finite element method. Background Technology

[0002] Titanium alloys, due to their high specific strength, high temperature resistance, corrosion resistance, and excellent fatigue resistance, have become the preferred material for core components such as aerospace fuselage load-bearing frames and landing gears, engine casings, ship pressure hulls, and critical valves in nuclear power plants. Among these, large-size titanium alloy forgings with a maximum profile dimension ≥1000mm are the core blanks for large integral load-bearing structures in high-end equipment, and their forming quality and microstructure directly determine the service safety and service life of the equipment. However, titanium alloys inherently possess high deformation resistance and an extremely narrow plastic forming window (e.g., the conventional two-phase forging temperature range is only 50-80℃), and their hot deformation behavior is highly sensitive to temperature and strain rate. For large-sized titanium alloy forgings, due to their large outline dimensions, large wall thickness differences, and long deformation stroke, the temperature and strain fields of the billet are extremely unevenly distributed throughout the forging process. These extreme deformation conditions can easily lead to macroscopic forming defects such as insufficient deformation, under-pressure of the cavity, surface or internal cracking, and metal flow line folding. At the same time, they are often accompanied by substandard microscopic properties such as coarse internal grains, mixed grains, and poor microstructure uniformity, which can easily cause the entire high-value forging to be scrapped directly.

[0003] Existing technologies apply finite element method (FEM) numerical simulation to titanium alloy forging process design, which can, to some extent, replace physical prototyping and trial-and-error, predicting the temperature field, strain field distribution, and macroscopic forming results of forgings under specific process parameters. However, facing the stringent conditions of narrow forging temperature windows and high deformation resistance in large-size titanium alloy forgings, traditional empirical trial-and-error process parameter design methods (even with virtual trial-and-error combined with FEM simulation, the extremely time-consuming single large-size three-dimensional thermo-mechanical coupling simulation calculation makes it impossible to conduct comprehensive and high-frequency exploration and optimization within the vast multivariable process parameter space) still cannot simultaneously consider the macroscopic forming quality and microstructure uniformity of the forgings. This easily leads to the coexistence of macroscopic forming defects and unqualified microstructure properties, resulting in process contradictions such as qualified forming but substandard performance, and qualified performance but defective forming. This process design mode, lacking multi-objective collaborative optimization methods, results in a high scrap rate for large-size titanium alloy forgings, extremely high R&D costs, and significantly extended product development and prototyping cycles. Summary of the Invention

[0004] To address the technical problems of existing finite element method (FEM)-based numerical simulation techniques in titanium alloy forging, which cannot perform high-frequency and rapid global optimization within a vast process parameter space and simultaneously fail to simultaneously ensure the macroscopic forming quality and microstructure uniformity of large-size forgings, resulting in high trial production scrap rates, high R&D costs, and long development cycles, this invention aims to provide a finite element method for optimizing forging process parameters in large-size titanium alloy forgings. The specific technical solution adopted is as follows: This invention provides a method for optimizing forging process parameters of large-size titanium alloy forgings based on the finite element method. The method includes the following steps: Acquire basic data related to titanium alloys; establish a finite element simulation model and define the mesh nodes in the finite element simulation model as key locations; Based on multiple preset forging process parameters and relevant basic data of titanium alloys, finite element simulation calculations are performed to obtain simulation data of key locations under each set of forging process parameters. Based on the relevant basic data of titanium alloys, the safe forming range and optimal grain refinement range of titanium alloy materials are determined. Based on the distribution of simulation data in the safe forming range and optimal grain refinement range, as well as the forming defect risk at key locations, the reference comprehensive adaptability of each set of forging process parameters is obtained. Using the forging process parameters of each group as input and the corresponding reference comprehensive fit degree as output, a surrogate model is trained, and the prediction accuracy of the surrogate model is calculated. Based on the model comprehensive fit degree, prediction accuracy, and correlation direction of each forging process parameter output by the surrogate model, the updated forging process parameters are obtained. The optimal forging process parameters are obtained by updating the forging process parameters, using the finite element simulation model and the surrogate model.

[0005] Furthermore, the method for obtaining the simulation data is as follows: Hot deformation parameters were extracted from basic data of titanium alloys, and hyperbolic sinusoidal Arrhenius constitutive equations were constructed. Based on the equipment slider speed and die geometry in each set of forging process parameters, the forging strain rate, forging temperature and equivalent plastic strain at each key position under each set of forging process parameters are solved by finite element thermo-mechanical coupling analysis. The corresponding rheological stress state is calculated by combining the hyperbolic sinusoidal Arrhenius constitutive equation. Forging strain rate, forging temperature, equivalent plastic strain, and rheological stress state are all used as simulation data.

[0006] Furthermore, the method for obtaining the safe forming range and the optimal grain refinement range is as follows: Based on the forging strain rate, forging temperature and preset grain evolution model in the simulation data, the dynamic recrystallized grain size at each key position under each set of forging process parameters is calculated. The range between forging strain rate and forging temperature, where no rheological instability occurs and the deformation resistance is within the preset allowable load range of the equipment, is extracted as the safe forming range; Extract the subset of dynamically recrystallized grains within the safe forming range that meet the preset qualified size and whose grain size range is less than the preset range threshold, and use it as the optimal grain refinement range.

[0007] Furthermore, the method for obtaining the dynamic recrystallized grain size is as follows: Obtain material microstructure constants, grain size strain rate exponents, dynamic recrystallization apparent activation energy, and universal gas constants from basic data related to titanium alloys; For any set of forging process parameters and any key position, obtain the forging temperature and forging strain rate of the key position at the final forging time under the set of forging process parameters. The product of the general gas constant and the forging temperature is used as the first value; The ratio of the apparent activation energy of dynamic recrystallization to the first value is used as the second value; The third value will be the value of an exponential function with the natural constant as the base and the second value as the exponent. The product of the forging strain rate and the third value is used as the Zener-Hollomon parameter value; The fourth value is the inverse power of the grain size strain rate exponent of the Zener-Hollomon parameter. The product of the material microstructure constant and the fourth value is used as the dynamic recrystallization grain size at this critical location under this set of forging process parameters.

[0008] Furthermore, the method for obtaining the reference comprehensive adaptation degree is as follows: For any set of forging process parameters, the percentage of the simulation data under that set of forging process parameters falling into the critical position of the safe forming range is taken as the safe zone matching degree corresponding to that set of forging process parameters. Based on the proportion of key positions falling into the optimal grain refinement range and the actual grain size range under the simulation data of this set of forging process parameters, the contribution of this set of forging process parameters to the microstructure uniformity is obtained; among them, the proportion of key positions falling into the optimal grain refinement range is positively correlated with the contribution of microstructure uniformity, while the actual grain size range is negatively correlated with the contribution of microstructure uniformity. The ratio of the equivalent plastic strain to the preset cracking critical strain at each defect-prone critical location under this set of forging process parameters is taken as the cracking risk value at each defect-prone critical location. The result of negatively correlated and boundary-trunculated maximum cracking risk values ​​is used as the forming risk aversion degree corresponding to this set of forging process parameters. The weighted sum of the safety zone matching degree, the contribution of the microstructure uniformity degree, and the forming risk avoidance degree is used as the reference comprehensive fit degree corresponding to the set of forging process parameters.

[0009] Furthermore, the method for obtaining the critical locations prone to defects is as follows: For any set of forging process parameters, obtain the static structural design parameters of the titanium alloy forging, as well as the stress characteristics in the simulation data corresponding to the set of forging process parameters; based on the static structural design parameters and stress characteristics, analyze the key locations where strain concentration or poor metal flow occurs during the forging process. The critical locations where strain concentration or poor metal flow occurs, and where there is a risk of cracking, folding, or insufficient filling, are identified as the critical defect-prone locations under this set of forging process parameters.

[0010] Furthermore, the method for obtaining the prediction accuracy is as follows: Input multiple sets of preset forging process parameters into the trained proxy model, and output the overall model fit degree corresponding to each set of forging process parameters; The difference between the model's overall fit degree and the reference overall fit degree corresponding to each set of forging process parameters is converted relative to the reference benchmark and used as the model's relative error corresponding to each set of forging process parameters. The result of negatively correlating the mean of the relative error of the model is used as the prediction accuracy of the surrogate model.

[0011] Furthermore, the method for obtaining the updated forging process parameters is as follows: For any forging process parameter, the difference between constant 1 and the current output model comprehensive fit is multiplied by the prediction accuracy, the preset step size adjustment coefficient and the correlation direction coefficient of the forging process parameter, and then divided by the preset sensitivity weight of the forging process parameter to obtain the adjustment increment of the forging process parameter. The adjustment increment is multiplied by the preset value range of the forging process parameter, and then added to the current base value of the forging process parameter to obtain the updated forging process parameter.

[0012] Furthermore, the method for obtaining the correlation direction coefficient is as follows: For any forging process parameter, a preset small positive increment is applied to the current benchmark value, and the current benchmark values ​​of the other forging process parameters are input into the proxy model to obtain the comprehensive adaptation degree of the pathfinding prediction. If the overall fit of the pathfinding prediction is greater than or equal to the overall fit of the currently output model, then the correlation direction coefficient of the forging process parameter is set to a first preset value; the first preset value is a positive number. If the overall fit of the pathfinding prediction is less than the overall fit of the currently output model, then the correlation direction coefficient of the forging process parameter is set to a second preset value; the second preset value is a negative number.

[0013] Furthermore, the method for obtaining the optimal forging process parameters is as follows: The updated forging process parameters are input into the proxy model to re-predict the overall model fit. If the overall fit of the model meets the preset initial screening conditions, the updated forging process parameters will be input into the finite element simulation model for simulation verification to obtain the true reference overall fit. If the degree of comprehensive adaptation of the real reference meets the preset termination condition, the iteration stops and the updated forging process parameters are used as the optimal forging process parameters. If the true reference comprehensive adaptation degree does not meet the preset termination condition, the updated forging process parameters and their true reference comprehensive adaptation degree will be added to the sample set, the training agent model will be updated, and the step of obtaining the updated forging process parameters will be returned. If the overall model fit does not meet the preset initial screening conditions, the updated forging process parameters will be used as the baseline value, and the process will return to the step of obtaining the updated forging process parameters. When the cumulative iterations reach the maximum number of iterations, the iteration stops, and the set of forging process parameters with the highest degree of comprehensive adaptation to the real reference in all iterations is output as the optimal forging process parameters.

[0014] The present invention has the following beneficial effects: This invention first uses finite element simulation calculations based on multiple pre-set forging process parameters and relevant data on titanium alloys to obtain simulation data for key locations under each set of forging process parameters. This accurately reflects the physical field distribution characteristics of the entire internal domain of the forging under each set of forging process parameters, which is beneficial for subsequent targeted local quality assessment and global parameter optimization. To define the absolute physical boundaries that ensure both macroscopic and microscopic quality of the forging meets standards, the invention further determines the safe forming range and optimal grain refinement range of the titanium alloy material based on relevant data on titanium alloys. This facilitates direct mapping and matching of the finite element simulation data under each set of forging process parameters with the aforementioned ranges, enabling quantitative evaluation of each key location. The macroscopic forming safety and microstructure uniformity at the location provide clear judgment criteria; in order to comprehensively reflect the overall performance of forging process parameters in avoiding macroscopic defects and improving microstructure properties, and further, based on the distribution of simulation data in the safe forming range and the optimal grain refinement range, as well as the forming defect risk at key locations, the reference comprehensive adaptation degree corresponding to each set of forging process parameters is obtained, accurately reflecting the degree of fit between each set of forging process parameters and the ideal perfect forging state. This is conducive to integrating the three independent and even restrictive core dimensions of safe forming breadth, microstructure uniformity, and local crack risk avoidance into a single process optimization objective function; in order to solve the large In the multi-parameter iterative optimization of dimensional titanium alloy forgings, repeated calls to finite element simulation lead to extremely high computational costs and excessive time consumption. Therefore, a surrogate model is trained using each set of forging process parameters as input and the corresponding reference comprehensive fit as output. The prediction accuracy of the surrogate model is calculated to accurately reflect the reliability of the surrogate model in replacing the time-consuming finite element simulation for result evaluation. To utilize the model's inherent credibility and guide the optimization algorithm to converge efficiently in the physically correct direction, preventing blind optimization of a single parameter from falling into local optima, the updated forging process is obtained based on the model's comprehensive fit, prediction accuracy, and the correlation direction of each forging process parameter output by the surrogate model. The parameters enable adaptive iterative correction of forging process parameters, ensuring that each adjustment effectively improves the overall forming quality of the forgings. To drive the entire optimization framework to continuously and autonomously seek optimization within the parameter space and avoid the surrogate model from falling into mathematical pseudo-peaks that lose physical meaning due to extrapolation errors, the optimal forging process parameters are obtained based on updating the forging process parameters, finite element simulation model, and surrogate model. This achieves an active learning physical closed loop of rapid initial screening of the surrogate model and high-precision verification by the finite element model, thereby obtaining a set of optimal process parameter combinations for large-size titanium alloy forgings that effectively balance macroscopic safe forming and microscopic uniformity. This effectively reduces the scrap rate of high-value forgings and significantly shortens the R&D cycle. Attached Figure Description

[0015] Figure 1A schematic flowchart illustrating a method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method, provided as an embodiment of the present invention; Figure 2 This is a structural diagram of a finite element method-based system for optimizing forging process parameters of large-size titanium alloy forgings, provided in one embodiment of the present invention. Figure 3 This is a schematic diagram of a computer device provided according to an embodiment of the present invention. Detailed Implementation

[0016] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0017] The following description, in conjunction with the accompanying drawings, details the specific scheme of the finite element-based method for optimizing forging process parameters of large-size titanium alloy forgings provided by this invention.

[0018] Example 1: This invention proposes a method for optimizing forging process parameters of large-size titanium alloy forgings based on the finite element method. Please refer to [link / reference]. Figure 1 The diagram illustrates a schematic flowchart of a method for optimizing forging process parameters of large-size titanium alloy forgings based on the finite element method, according to an embodiment of the present invention. The method includes the following steps: Step S1: Obtain relevant data on the titanium alloy basic structure; establish a finite element simulation model and define the mesh nodes in the finite element simulation model as key locations.

[0019] Specifically, large-size titanium alloy forgings are known to have large profile dimensions, large wall thickness differences, and long deformation strokes. Furthermore, titanium alloys themselves have high deformation resistance and a narrow plastic forming window. Their forming quality and microstructure are influenced by multiple core factors, including material constitutive properties, geometric boundaries, and equipment capabilities. To comprehensively and accurately characterize the physical boundaries and production constraints during the forging process, and to obtain fundamental data related to the titanium alloy, ensuring that subsequent simulation and optimization of process parameters closely reflect actual production conditions, a reliable physical and mechanical basis is required. This fundamental data includes titanium alloy base material performance data, forging geometric requirements data, and forging equipment boundary data. Specifically, titanium alloy base material performance data includes the chemical composition, thermophysical properties, hot deformation rheological characteristics, and microstructure evolution data of the titanium alloy; forging geometric requirements data includes the three-dimensional models of the forging and die, dimensional tolerances, forming defect limitation indicators, and grain size requirements; and forging equipment boundary data includes the nominal pressure of the equipment, the range of slider movement speed, the maximum allowable load threshold, and the die preheating temperature range.

[0020] To simulate the changes in the physical field across the entire forging process and accurately track the deformation behavior of titanium alloy forgings under complex stress conditions, the three-dimensional models of the forgings and dies were imported into finite element analysis software. Corresponding titanium alloy base material performance data were assigned as material properties, and forging equipment boundary data were assigned as initial and boundary conditions. A finite element simulation model was established, and the mesh nodes in the finite element simulation model were defined as key locations. This allows for the accurate extraction of microscopic state parameters from different parts of the forging's interior and surface in subsequent finite element simulation calculations. This provides discretized and high-precision data support for quantitatively evaluating the local forming quality, cracking risk, and microstructure uniformity of large-size forgings.

[0021] Step S2: Based on the preset multiple sets of forging process parameters and the relevant basic data of titanium alloy, perform finite element simulation calculations to obtain simulation data of key positions under each set of forging process parameters; determine the safe forming range and optimal grain refinement range of titanium alloy material based on the relevant basic data of titanium alloy; based on the distribution of simulation data in the safe forming range and optimal grain refinement range, as well as the forming defect risk at key positions, obtain the reference comprehensive adaptation degree corresponding to each set of forging process parameters.

[0022] Specifically, it is known that in the titanium alloy forging process, process parameters such as deformation temperature, strain rate, and deformation amount directly determine the rheological behavior and microstructure evolution of titanium alloy forgings. To comprehensively quantify the impact of process parameters on the forming quality of large-size forgings, the initial billet temperature, die preheating temperature, press speed, and deformation distribution are all included as forging process parameters. Considering the extremely complex deformation path and highly uneven temperature and strain distribution across the entire range of large-size titanium alloy forgings, to cover various possible actual production conditions and provide sufficient and representative training samples for subsequent proxy model construction, multiple sets of forging process parameters covering different process boundaries are first preset using orthogonal experimental design or Latin hypercube sampling methods. Then, based on the preset sets of forging process parameters and relevant basic data of titanium alloys, finite element simulation calculations are performed to obtain simulation data of key locations under each set of forging process parameters. This accurately reflects the physical field distribution characteristics of the entire internal domain of the forging under each set of forging process parameters, which is beneficial for subsequent targeted local quality assessment and global parameter optimization. In this embodiment, the number of preset forging process parameters is set to 200 to ensure that the sample space fully covers the process boundary and the total simulation calculation time is controllable. Implementers can set the number of preset forging process parameters according to the available computing resources and the convergence accuracy requirements of the subsequent proxy model; no limitation is imposed here. The orthogonal experimental design and Latin hypercube sampling method are well-known and will not be described further.

[0023] Considering that the hot deformation behavior of titanium alloys is highly sensitive to temperature and strain rate, and is prone to cracking or grain coarsening under specific thermodynamic conditions, in order to define the absolute physical boundary that ensures both the macroscopic and microscopic quality of forgings meet the standards, and then determine the safe forming range and optimal grain refinement range of titanium alloy materials based on basic titanium alloy data, it is beneficial to directly map and match the simulation data under each set of forging process parameters with the safe forming range and optimal grain refinement range, and provide a clear judgment benchmark for quantitatively evaluating the macroscopic forming safety and microstructure uniformity at each key location. Considering that a single indicator cannot comprehensively measure the complex quality requirements of large-size forgings, in order to comprehensively reflect the overall performance of forging process parameters in avoiding macroscopic defects and improving microstructure properties, a reference comprehensive fit degree is obtained for each group of forging process parameters based on the distribution of simulation data in the safe forming range and the optimal grain refinement range, as well as the forming defect risk at key locations. This accurately reflects the degree of fit between each group of forging process parameters and the ideal perfect forging state. Among them, the greater the reference comprehensive fit degree, the more stable the forging process under the corresponding group of forging process parameters, the lower the risk of forming defects such as cracking or folding, and the more uniform and fine the microstructure of the entire forging.

[0024] Preferably, in one feasible manner of this embodiment, the simulation data acquisition method is as follows: First, extract the hot deformation parameters from the basic data of the titanium alloy and construct a hyperbolic sinusoidal Arrhenius constitutive equation. This is beneficial for accurately describing the quantitative relationship between rheological stress, strain rate, and deformation temperature during the high-temperature thermoplastic deformation of metallic materials, providing a mechanical calculation kernel for predicting macroscopic flow laws and microstructure evolution. The hyperbolic sinusoidal Arrhenius constitutive equation is well-known and will not be elaborated further. In order to accurately solve the nonlinear thermo-mechanical coupling problem under complex geometric boundaries, the forging strain rate, forging temperature, and equivalent plastic strain at each key position under each set of forging process parameters are solved by finite element thermo-mechanical coupling analysis based on the equipment slider speed and die geometric characteristics in each set of forging process parameters. The corresponding rheological stress state is calculated by combining the hyperbolic sinusoidal Arrhenius constitutive equation. Then, the forging strain rate, forging temperature, equivalent plastic strain, and rheological stress state corresponding to each set of forging process parameters are used as the simulation data of the key positions under each set of forging process parameters.

[0025] Preferably, in one feasible manner of this embodiment, the method for obtaining the safe forming range and the optimal grain refinement range is as follows: Considering that dynamic recrystallization is the core physical mechanism for refining grains and improving mechanical properties during the hot forging process of titanium alloys, in order to quantify the specific influence of each set of forging process parameters on the microstructure, the dynamic recrystallization grain size at each key position under each set of forging process parameters is calculated based on the forging strain rate, forging temperature and preset grain evolution model in the simulation data, so as to accurately reflect the objective law of grain size evolution with thermodynamic conditions; the larger the dynamic recrystallization grain size, the more insufficient the recrystallization process, the higher the deformation temperature or the lower the strain rate, which can easily lead to a decrease in the strength properties of the forging; The method for obtaining the dynamic recrystallization grain size is as follows: Considering that the complex deformation of each node inside a large forging often accumulates along the path, and in actual engineering applications, the final microstructure is mainly determined by the severe deformation thermodynamic conditions at the end of forging (final forging stage), in order to significantly reduce the integral calculation of the entire microstructure evolution process while ensuring the reliability of the evaluation, this invention uses a final forging steady-state model for characterization, thereby obtaining the material microstructure constants, grain size strain rate exponent, dynamic recrystallization apparent activation energy, and universal gas constants in the basic data related to titanium alloys; at the same time, for any set of forging process parameters and For any key location, obtain the forging temperature and forging strain rate at the final forging moment under the forging process parameters of that key location from the output data of the previous finite element analysis module; take the product of the universal gas constant and the forging temperature as the first value, which characterizes the basic thermal energy level under the current thermodynamic state; take the ratio of the apparent activation energy of dynamic recrystallization to the first value as the second value, which characterizes the comprehensive energy barrier of the thermal activation effect during deformation; it should be noted that the first value is greater than 0, because the universal gas constant is a fixed positive constant, and in physics, the forging temperature, measured in Kelvin thermodynamic absolute temperature, is always greater than absolute zero. To accurately quantify the nonlinear promoting effect of temperature-dominated thermal activation processes on material deformation and microstructure evolution, the exponential function value with the natural constant as the base and the second value as the exponent is used as the third value. To comprehensively consider the competitive contribution mechanism of strain rate and temperature to the dynamic recrystallization process, the product of forging strain rate and the third value is further used as the Zener-Hollomon parameter value (i.e., temperature-compensated strain rate parameter). The acquisition of the Zener-Hollomon parameter value is well known and will not be elaborated further. To measure the macroscopic influence of the aforementioned comprehensive thermodynamic conditions on recrystallization nucleation and grain refinement, the grain size strain rate exponent of the Zener-Hollomon parameter is further raised to the inverse power as the fourth value, characterizing the dynamic inhibition of grain growth by the comprehensive thermodynamic effect. A smaller fourth value indicates a stronger dynamic recrystallization driving force at lower deformation temperatures or higher strain rates, resulting in a more significant inhibition of grain growth and promoting the formation of finer grains. To accurately calculate the steady-state grain size after dynamic evolution, the product of the material microstructure constant and the fourth value is used as the dynamic recrystallized grain size at this critical location under the given forging process parameters. The material microstructure constant is an empirical constant obtained through regression fitting of thermal simulation test data, possessing composite dimensions for the physical units at both ends of the equilibrium formula (e.g., ...). Taking common TC4 or TA15 titanium alloys as examples, their reference values ​​are usually within... to Between orders of magnitude), the final physical dimension of the dynamic recrystallized grain size is a unit of length (such as micrometers). It is known that exceeding the equipment load can lead to downtime or equipment damage. Furthermore, material constitutive rheological instability (such as the formation of adiabatic shear bands or localized flow concentration) can disrupt the continuity of the microstructure. To ensure dual safety in the forging process at both the instantaneous rheological characteristics and equipment capacity levels, the range of forging strain rate and forging temperature within which no rheological instability occurs and deformation resistance is within the preset allowable equipment load range is extracted as the safe forming interval. It should be noted that the risk of macroscopic cracking or ductile fracture caused by cumulative deformation (equivalent plastic strain) will be independently quantified using the critical fracture criterion in the subsequent forming risk avoidance assessment, thereby achieving physical decoupling of instantaneous rheological instability and cumulative fracture risk. In this embodiment, the preset allowable equipment load upper limit is set to 85% of the rated nominal pressure to ensure sufficient load margin in actual complex production environments. Implementers can set the preset allowable equipment load upper limit according to the tonnage, service life, and wear condition of the specific press; this is not limited here. Considering that even within the safe forming range, the microstructure may exhibit localized coarsening or unevenly distributed mixed crystal phenomena, in order to screen out the core process window that yields the best fatigue life and mechanical properties, a subset of dynamic recrystallized grain sizes within the safe forming range that meets the preset qualified size and whose grain size range (i.e., the maximum difference in dynamic recrystallized grain size between any two key locations across the entire forging region) is less than a preset range threshold is extracted as the optimal grain refinement range. In this embodiment, the preset qualified size is set as the upper limit of the physical length corresponding to the target grain size level required by the corresponding aerospace standard, ensuring that the basic mechanical properties of the forging meet the standards. Implementers can set the preset qualified size according to the specific technical protocol specifications of the forging, which is not limited here. The preset range threshold is set as the maximum physical length difference corresponding to a grain size fluctuation across the entire region that is less than or equal to one standard level, ensuring the high uniformity of the forging microstructure. Implementers can set the preset range threshold according to the severity of the forging's service environment, which is not limited here.

[0026] Preferably, in one feasible method of this embodiment, the method for obtaining the comprehensive fit degree is as follows: when a large amount of simulation data falls within the safe forming range, it indicates that the stress and deformation of most areas of the forging under the corresponding set of forging process parameters are in a safe state. In order to accurately reflect the coverage of the global safe forming corresponding to each set of forging process parameters, for any set of forging process parameters, the proportion of the simulation data under that set of forging process parameters falling into the key position of the safe forming range is taken as the safe zone matching degree corresponding to that set of forging process parameters; the larger the safe zone matching degree, the smaller the possibility of macroscopic cracking or equipment overload during the forging process under that set of forging process parameters. On the other hand, considering that in actual production, the grain size range directly determines the consistency of forging performance, in order to comprehensively evaluate that the microstructure should not only be fine but also uniform, the contribution of the microstructure uniformity corresponding to this set of forging process parameters is obtained based on the proportion of key positions falling into the optimal grain refinement range and the actual grain size range under the simulation data of this set of forging process parameters. Among them, the proportion of key positions falling into the optimal grain refinement range is positively correlated with the contribution of microstructure uniformity, while the actual grain size range is negatively correlated with the contribution of microstructure uniformity. The larger the contribution of microstructure uniformity, the finer and more uniform the forging grains are, and the lower the probability of mixed grain defects. The formula for calculating the contribution of microstructure uniformity is: In the formula, Z represents the contribution of the forging process parameters to the microstructure uniformity. N represents the number of key locations in the simulation data falling within the optimal grain refinement range under this set of forging process parameters; N is the total number of key locations. This is a function to find the maximum value. This represents the actual grain size range corresponding to this set of forging process parameters; This refers to the maximum grain size range allowed by the forging technical requirements; This is the first preset positive number. This embodiment sets... The physical length difference threshold (e.g., 10 to 30 micrometers) corresponding to 1 to 3 grain size levels ensures that the performance of each part of the forging meets the design lower limit. Implementers can set this threshold according to the specific military or civil aviation standards of the model. No restrictions are imposed here; settings The value is set to 0.001 to ensure that the denominator is not zero in extreme cases such as abnormal input data or when the upper limit of the range requirement is 0, thus preventing program calculation crashes. Implementers can set this value according to the precision requirements of floating-point operations. No specific restrictions are imposed here; among them, This is a boundary truncation function to ensure that the contribution of this term is safely truncated to 0 when the range is severely exceeded, thus avoiding the negative contribution caused by severe local non-compliance, which would interfere with the direction of overall parameter optimization. Furthermore, considering that complex local structures in large-sized forgings (such as deep cavities and extremely thin walls) are prone to metal flow interference or tearing, in order to accurately capture and amplify these local forming hazards, the ratio of the equivalent plastic strain to the preset cracking critical strain at each defect-prone key location under this set of forging process parameters is used as the cracking risk value for each defect-prone key location. This accurately reflects the degree to which the actual strain state at each defect-prone key location approaches the material's fracture limit. The larger the cracking risk value, the higher the risk of microcracks or macroscopic fracture at the corresponding key location. In this embodiment, the preset cracking critical strain is set as the strain value when the material undergoes necking instability at the corresponding temperature and strain rate (usually between 0.8 and 1.5), ensuring that it serves as a reliable mechanical constitutive boundary for assessing cracking risk. The implementer can then determine the cracking risk based on pre-conducted hot compression or hot... The tensile destructive test data sets a preset critical cracking strain, which is not limited here. In order to transform the local danger level into a positive safety evaluation index that strictly falls within the [0,1] interval, and to prevent excessive strain from causing numerical overflow and collapse, the maximum cracking risk value is negatively correlated and the result of boundary truncation is used as the forming risk aversion degree corresponding to this set of forging process parameters. This accurately reflects the ability of this set of forging process parameters to suppress the cracking potential at the weakest position. Among them, the larger the forming risk aversion degree, the more the cracking risk of the forging in the most dangerous area is controlled at an extremely low level, and the local forming quality is extremely high. In this embodiment, the difference between the constant 1 and the maximum cracking risk value is used with the constant 0 to perform a truncation formula that takes the maximum value, and the maximum cracking risk value is negatively correlated and the boundary is truncated to ensure that the output value strictly falls within the range of [0,1]. Within a unified dimension, so as to facilitate subsequent direct weighted summation; The method for obtaining the key locations prone to defects is as follows: Considering that the metal flow patterns differ under different process parameters, the defect-prone areas will dynamically shift. For any set of forging process parameters, the static structural design parameters of the titanium alloy forging (such as abrupt cross-sectional change areas and curvature extreme value areas) are extracted by reading the features of the three-dimensional model of the forging. The stress characteristics are then extracted from the simulation data corresponding to the set of forging process parameters. Based on the static structural design parameters and stress characteristics, the key locations where strain concentration or poor metal flow occurs during forging are extracted. Specifically, the strain gradient and metal velocity vector of each key location are extracted. By setting abnormal warning thresholds for strain gradient and velocity gradient, areas where the strain gradient increases abnormally or the velocity vector changes drastically are screened out. In order to narrow the monitoring range and improve the pertinence and computational efficiency of local risk assessment, the key locations where strain concentration or poor metal flow occurs, and where there is a risk of cracking, folding, or insufficient filling, are taken as the key locations prone to defects under the set of forging process parameters. Finally, in order to integrate the three independent core dimensions of safety breadth, microstructure uniformity, and local risk aversion into a single process optimization objective function, and to avoid historical label drift caused by the introduction of normalization dependent on global sample boundaries in dynamic parameter optimization iteration, thus strictly constraining... The weighted sum of the safety zone matching degree, the contribution of microstructure uniformity, and the forming risk avoidance degree within the physical dimensions is used as the reference comprehensive fit degree for this set of forging process parameters. The formula for calculating the reference comprehensive fit degree is as follows: In the formula, P represents the reference comprehensive fit degree corresponding to the set of forging process parameters; C represents the safety zone matching degree corresponding to the set of forging process parameters; Z represents the contribution degree of microstructure uniformity corresponding to the set of forging process parameters; and B represents the forming risk avoidance degree corresponding to the set of forging process parameters. As the first reference weight; As the second reference weight; As the third reference weight; this embodiment sets 0.4 0.4 and The value is set to 0.2 to ensure that while balancing overall forming safety and microstructure compliance, appropriate attention is paid to preventing local defects. Implementers can set this value based on the priority differences between macroscopic quality and microscopic properties in specific forging production. , and And satisfy No restrictions are imposed here.

[0027] Step S3: Using the forging process parameters of each group as input and the corresponding reference comprehensive fit degree as output, train the surrogate model and calculate the prediction accuracy of the surrogate model; based on the model comprehensive fit degree, prediction accuracy and correlation direction of each forging process parameter output by the surrogate model, obtain the updated forging process parameters.

[0028] Specifically, to address the issue of extremely high computational costs and excessive time consumption caused by repeated calls to finite element simulation during the multi-parameter iterative optimization of large-size titanium alloy forgings, a surrogate model is trained using each set of forging process parameters as input and the corresponding reference comprehensive fit degree as output. This model employs machine learning algorithms such as gradient boosting trees or random forests (gradient boosting trees and random forests are well-known and will not be elaborated further). This establishes a rapid mapping relationship between macroscopic process parameters and comprehensive forming quality, thereby calculating the prediction accuracy of the surrogate model. This accurately reflects the reliability level of the surrogate model in replacing time-consuming finite element simulations for result evaluation. Higher prediction accuracy indicates that the prediction results of the surrogate model are closer to the actual physical simulation results, and thus have higher credibility.

[0029] Considering that blindly seeking optimization for a single parameter can easily lead to local optima or deviate from physical laws, in order to utilize the credibility of the model itself and guide the optimization algorithm to converge efficiently in the physically correct direction, the updated forging process parameters are obtained based on the overall model fit, prediction accuracy, and correlation direction of each forging process parameter output by the surrogate model. This enables iterative correction of the forging process parameters, ensuring that each adjustment of the forging process parameters can effectively improve the overall forming quality of the forgings.

[0030] Preferably, in one feasible method of this embodiment, the prediction accuracy is obtained as follows: In order to comprehensively evaluate the error level of the trained surrogate model, multiple sets of preset forging process parameters are input into the trained surrogate model, and the overall model fit degree corresponding to each set of forging process parameters is output. When the overall model fit degree deviates more from the reference overall fit degree, it indicates that the prediction is more biased. In order to standardize the measurement of the severity of this bias relative to the true level, the absolute value of the difference between the overall model fit degree corresponding to each set of forging process parameters and the reference overall fit degree is converted relative to the reference benchmark and used as the model relative error corresponding to each set of forging process parameters. The larger the model relative error, the less accurate the prediction result of the surrogate model under the corresponding set of process parameters. In this embodiment, the absolute value of the difference between the overall model fit degree corresponding to the set of forging process parameters and the reference overall fit degree is divided by the sum of the corresponding reference overall fit degree and the second preset positive number to convert the absolute value of the difference into a relative error index, that is, by using the formula... The calculation is performed, where P is the reference comprehensive fit degree corresponding to the set of forging process parameters; The degree of model adaptation corresponding to this set of forging process parameters; The second preset positive number; The sign is absolute. In this embodiment, the size of the second preset positive number is set to 0.1, and the unit is consistent with the comprehensive adaptation degree. This ensures that when encountering samples with a reference comprehensive adaptation degree of zero or very close to zero, the denominator is not zero and the numerical explosion of relative error can be effectively suppressed, thus ensuring the numerical stability of the model prediction accuracy assessment. The implementer can set the second preset positive number according to the sensitivity requirements of the actual proxy model for the assessment of extremely low sample error and the floating-point operation accuracy. No limitation is imposed here. Considering that the final evaluation focuses on the overall predictive ability of the surrogate model rather than single-point bias, to transform the model's relative error into an intuitive positive evaluation metric, the mean of the model's relative error is negatively correlated and used as the surrogate model's prediction accuracy. This embodiment uses the difference between a constant 1 and the mean of the model's relative error, and takes the maximum value of this difference with a preset minimum threshold as the surrogate model's prediction accuracy. This serves as a protective truncation. In this embodiment, the preset minimum threshold is set to 0.01 to ensure that even in the low-confidence phase of the surrogate model's initial training, the optimization algorithm still has a basic exploration step size weight, maintaining its ability to continuously explore within the unknown parameter space. Implementers can set the preset minimum threshold based on the global exploration needs in the initial parameter optimization phase and the sensitivity of process parameters; no specific limit is imposed here. This avoids negative accuracy due to extreme deviations that could lead to deadlock in subsequent parameter optimization, while ensuring that the prediction accuracy is not zero even when the error is large in the early stages of model training. This guarantees that the multiplication factor during subsequent calculations and adjustments is not zero, driving continuous parameter updates and iterations.

[0031] Preferably, in one feasible embodiment, the method for obtaining updated forging process parameters is as follows: In order to dynamically determine the correction magnitude of each iteration by utilizing the difference between the ideal state (value 1) and the current state, while preventing the surrogate model from blindly giving excessively large step sizes in the low confidence region, and eliminating optimization oscillations caused by differences in the sensitivity of different physical dimension parameters to the objective function, for any forging process parameter, the difference between constant 1 and the current output model comprehensive fit is multiplied by the prediction accuracy, the preset step size adjustment coefficient, and the correlation direction coefficient of the forging process parameter, and then divided by the preset sensitivity weight of the forging process parameter to obtain the adjustment increment of the forging process parameter; the difference between constant 1 and the current output model comprehensive fit accurately reflects the relative change required to achieve optimal quality; the prediction accuracy, as a confidence weight, makes the adjustment more decisive when the surrogate model's prediction is more accurate; the preset sensitivity weight, as an adjustment factor, makes sensitive parameters that have a significant impact on forming quality (such as forging temperature) obtain smaller relative increments, while the impact of slow passivation is mitigated. A larger relative increment in the parameters ensures global convergence stability during multi-variable collaborative optimization. The larger the absolute value of the adjustment increment, the further the current forging process parameters are from the optimal state, requiring a larger update. In this embodiment, the preset step size adjustment coefficient is set to a decimal between 0.05 and 0.15 (e.g., 0.1) to ensure smooth convergence during the iterative process optimization of multiple forging process parameters and to avoid algorithm divergence or overshooting the optimal solution due to an excessively large preset step size adjustment coefficient. Implementers can set the preset step size adjustment coefficient according to the convergence speed and computational resources of the initial optimization iteration, without limitation here. The preset sensitivity weight is set to the absolute value of the correlation coefficient of the target parameter calculated based on the previous thermal simulation test or Latin hypercube sampling data. Implementers can set the preset sensitivity weight according to the industrial control response characteristics of the specific parameters, without limitation here. When the absolute value of the calculated target parameter correlation coefficient is 0, it is replaced with a preset minimum positive number (e.g., 0.001) as the preset sensitivity weight to prevent the program from crashing due to a denominator of 0 in subsequent calculations. The method for obtaining the correlation direction coefficient is as follows: Considering the highly nonlinear mapping between titanium alloy process parameters and forming quality, and the fact that the global proxy model (such as the tree model) is not differentiable locally, in order to dynamically sense the local gradient of the current parameter position in real time to guide the convergence direction, for any forging process parameter, a preset small positive increment is applied on the basis of the current benchmark value. This, along with the current benchmark values ​​of other forging process parameters, is input into the proxy model to obtain the comprehensive adaptation degree of the path prediction. This accurately reflects the local change trend and performance gradient direction of the forging process parameter near the current value, which is beneficial for providing real-time dynamic direction guidance for subsequent parameter adaptive iteration and avoiding missing the optimal process window. In this embodiment, the preset small positive increment is set to 0.5% to 2% (such as 1%) of the preset value range of the forging process parameter to ensure that the differential step size is small enough to accurately capture the local real gradient, while avoiding the step size being too small and being submerged by the numerical noise of the model. The implementer can set the preset small positive increment according to the industrial control accuracy of the specific parameters and the prediction sensitivity of the proxy model, which is not limited here. If the overall fit of the path prediction is greater than or equal to the overall fit of the current output model, it indicates that the current model is in a local monotonically increasing range. In order to guide the algorithm to continue to climb towards the extreme point, the correlation direction coefficient of the forging process parameter is set to the first preset value. The first preset value is a positive number. In this embodiment, the first preset value is set to a constant 1 to ensure that the optimization algorithm only provides directional guidance and does not amplify the correction range in this step. The implementer can set the first preset value according to the strength and sensitivity of the influence of each parameter on the forming quality (such as introducing a sensitivity coefficient), which is not limited here. If the overall fit of the pathfinding prediction is less than the overall fit of the current output model, it indicates that increasing the parameters in the positive direction leads to quality deterioration (e.g., the parabola vertex has been exceeded). In order to force the algorithm to turn around and find the optimal solution in the negative direction, the correlation direction coefficient of this forging process parameter is set to a second preset value. The second preset value is negative. In this embodiment, the second preset value is set to a constant -1 to ensure that the parameters are immediately corrected and reversed in the direction of deterioration. The implementer can set the second preset value according to the penalty intensity required for the parameters to deteriorate the quality, which is not limited here. Considering the significant differences in physical magnitudes and equipment allowable ranges among different forging process parameters (e.g., forging temperatures reaching thousands of degrees while slide block pressing speeds are only millimeters), directly multiplying by a uniform step size ratio would lead to excessively large single-step adjustments for large parameters (easily crossing the material phase transformation zone) and ineffective adjustments for small parameters (below the equipment control precision). To eliminate the impact of parameter magnitude differences on optimization stability, the adjustment increment is multiplied by the preset value range of the forging process parameter and then accumulated to the current baseline value of the forging process parameter to obtain the updated forging process parameter, thereby completing a closed-loop heuristic iterative optimization of the process parameter. This embodiment sets the preset value range of the forging process parameter as the difference between the maximum and minimum physical limit values ​​allowed in actual production (e.g., the range of 50°C to 80°C allowed for conventional two-phase forging temperature of titanium alloys, or the effective adjustment range of the sliding block pressing speed of a specific hydraulic press). This ensures that the absolute change in a single parameter update matches the industrial control precision of the corresponding parameter, and the iteration process is strictly limited within the material thermodynamic safety window and the rated capacity boundary of the equipment to prevent optimization overflow. The implementer can set the preset value range according to the thermal simulation measured processing diagram of the specific titanium alloy grade and the parameter control instructions of the on-site forging equipment, which is not limited here. The reference value is the actual value of the forging process parameter in the current iteration round. Specifically, it is the initial forging process parameter input during the first iteration optimization, and the updated forging process parameter output in the previous round during subsequent iterations.

[0032] Step S4: Based on the updated forging process parameters, finite element simulation model, and surrogate model, obtain the optimal forging process parameters.

[0033] Specifically, in order to drive the entire optimization framework to continuously and autonomously seek optimization within the parameter space and avoid the surrogate model from falling into mathematical pseudo-peaks that lose physical meaning due to extrapolation errors, the optimal forging process parameters are obtained based on updated forging process parameters, finite element simulation model and surrogate model. This results in a set of optimal process parameters for large-size titanium alloy forgings that effectively balances safe forming and microstructure homogenization and has been physically verified, effectively reducing the scrap rate of trial production and significantly shortening the R&D cycle.

[0034] Preferably, in one feasible method of this embodiment, the optimal forging process parameters are obtained as follows: First, the updated forging process parameters are input into the proxy model to re-predict the model's overall fit. If the model's overall fit meets the preset initial screening conditions, it indicates that the updated forging process parameters have extremely high engineering application potential under the prediction of the proxy model. In order to verify the feasibility and accuracy of the prediction result at the real physical level, the updated forging process parameters are input into the finite element simulation model for simulation verification to obtain the real reference overall fit, which accurately reflects the real forging quality of the set of process parameters under the actual thermo-coupling state. In this embodiment, the preset initial screening condition is set to the re-predicted model overall fit being greater than or equal to 0.85, ensuring that only high-potential parameters are called for time-consuming finite element simulation, greatly saving computing resources. Implementers can set the preset initial screening conditions according to the initial prediction accuracy of the proxy model and the remaining computing resources, which is not limited here. If the overall fit of the true reference meets the preset termination condition, it indicates that the updated forging process parameters have indeed enabled the forging quality to reach an extremely high engineering qualification level. In this case, the iteration stops, and the updated forging process parameters are taken as the optimal forging process parameters and directly output to guide actual production trials. If the overall fit of the true reference does not meet the preset termination condition, it indicates that the surrogate model has experienced severe prediction distortion or cognitive blind spots in this high-resolution region. To correct the errors of the surrogate model and improve its subsequent accuracy in guiding optimization, the updated forging process parameters and their overall fit of the true reference are added to the sample set to update and train the surrogate model. The model absorbs real physical feedback and returns to the step of obtaining updated forging process parameters so that it can continue to explore better process parameters under the guidance of the corrected proxy model. If the overall fit of the model does not meet the preset initial screening conditions, it means that the forming quality of the current updated forging process parameters has not yet reached the expected threshold. In order to avoid wasting expensive finite element simulation computing power on low potential parameters and to promote the parameters to continue to be rapidly iterated and optimized, the updated forging process parameters are used as the benchmark value, and the model returns to the step of obtaining updated forging process parameters. This achieves low-cost and rapid gradient climbing within the proxy model until the parameter performance is improved to meet the initial screening conditions. When the cumulative iterations reach the maximum number of iterations, iteration stops, and the set of forging process parameters with the highest comprehensive adaptation degree of the real reference in all iterations is output as the optimal forging process parameters. This embodiment sets a preset termination condition of a comprehensive adaptation degree of the real reference greater than or equal to 0.9 to ensure that the final output optimal process parameters achieve extremely high standards in both macroscopic forming safety and microscopic microstructure uniformity. Implementers can set the preset termination condition according to the specific safety requirements of the forging in the service environment of high-end equipment such as aerospace; this is not limited here. The maximum number of iterations is set to 600 to ensure that the optimization algorithm can exit in time when it gets stuck or fails to converge in extremely complex nonlinear multi-constraint conditions, preventing computational deadlock and resource exhaustion. Implementers can set the maximum number of iterations according to the time consumption of a single finite element simulation, the computing power of the computer hardware, and the time requirements for optimization and development; this is not limited here.

[0035] In summary, this embodiment acquires basic data related to titanium alloys to establish a finite element simulation model; performs finite element simulation calculations based on multiple sets of preset parameters to obtain simulation data for key locations; determines the safe forming range and optimal grain refinement range based on the basic data, and obtains the reference comprehensive fit degree corresponding to each set of parameters by combining the forming defect risk at key locations; trains a surrogate model and calculates the prediction accuracy; obtains updated forging process parameters based on the prediction accuracy, the comprehensive fit degree of the model output by the surrogate model, and the direction of parameter correlation; and obtains the optimal forging process parameters based on the updated forging process parameters, the finite element simulation model, and the surrogate model. This invention can effectively balance the macroscopic forming quality and microstructure uniformity of large-size forgings, and effectively reduce the scrap rate of trial production.

[0036] Example 2: This invention also proposes a finite element method-based optimization system for forging process parameters of large-size titanium alloy forgings. Please refer to [link / reference]. Figure 2 The diagram illustrates a structural diagram of a finite element method-based forging process parameter optimization system for large-size titanium alloy forgings, provided by an embodiment of the present invention. The system includes: a data acquisition module 10, a reference comprehensive adaptation degree acquisition module 20, a forging process parameter update module 30, and an optimal forging process parameter acquisition module 40.

[0037] Data acquisition module 10 is used to acquire basic data related to titanium alloys; establish a finite element simulation model, and define the mesh nodes in the finite element simulation model as key locations.

[0038] The reference comprehensive adaptation degree acquisition module 20 is used to perform finite element simulation calculations based on multiple preset forging process parameters and relevant basic data of titanium alloys to obtain simulation data of key positions under each set of forging process parameters; determine the safe forming range and optimal grain refinement range of titanium alloy materials based on the relevant basic data of titanium alloys; and obtain the reference comprehensive adaptation degree corresponding to each set of forging process parameters based on the distribution of simulation data in the safe forming range and optimal grain refinement range, as well as the forming defect risk at key positions.

[0039] The forging process parameter update module 30 is used to train a surrogate model with each group of forging process parameters as input and the corresponding reference comprehensive fit degree as output, and to calculate the prediction accuracy of the surrogate model; based on the model comprehensive fit degree, prediction accuracy and correlation direction of each forging process parameter output by the surrogate model, the updated forging process parameters are obtained.

[0040] The optimal forging process parameter acquisition module 40 is used to obtain the optimal forging process parameters based on the updated forging process parameters, finite element simulation model, and proxy model.

[0041] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the finite element method-based optimization system for forging process parameters of large-size titanium alloy forgings and the finite element method-based optimization method for forging process parameters of large-size titanium alloy forgings provided in the above embodiments belong to the same concept. The specific implementation process is detailed in the method embodiments and will not be repeated here.

[0042] Example 3: The present invention also proposes a computer device, see [link to relevant documentation]. Figure 3 The computer device includes a memory 401, a processor 402, and a computer program 403 stored in the memory 401 and running on the processor 402. When the processor 402 executes the computer program 403, the computer device can execute any of the aforementioned finite element-based forging process parameter optimization methods for large-size titanium alloy forgings.

[0043] Example 4: The present invention also proposes a computer-readable storage medium storing computer program code. When the computer program code is run on a computer, the computer executes the above-mentioned method steps to implement the finite element method for optimizing forging process parameters of large-size titanium alloy forgings provided in the above embodiments.

[0044] Example 5: The present invention also proposes a computer program product that, when run on a computer, causes the computer to perform the aforementioned related steps to achieve the finite element-based method for optimizing forging process parameters of large-size titanium alloy forgings provided in the above embodiments.

[0045] In this embodiment, the computer-readable storage medium, computer program product, or chip are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0046] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0047] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for optimizing forging process parameters of large-size titanium alloy forgings based on the finite element method, characterized in that, The method includes the following steps: Acquire basic data related to titanium alloys; establish a finite element simulation model and define the mesh nodes in the finite element simulation model as key locations; Based on multiple preset forging process parameters and relevant basic data of titanium alloys, finite element simulation calculations are performed to obtain simulation data of key locations under each set of forging process parameters. Based on the relevant basic data of titanium alloys, the safe forming range and optimal grain refinement range of titanium alloy materials are determined. Based on the distribution of simulation data in the safe forming range and optimal grain refinement range, as well as the forming defect risk at key locations, the reference comprehensive adaptability of each set of forging process parameters is obtained. Using the forging process parameters of each group as input and the corresponding reference comprehensive fit degree as output, a surrogate model is trained, and the prediction accuracy of the surrogate model is calculated. Based on the model comprehensive fit degree, prediction accuracy, and correlation direction of each forging process parameter output by the surrogate model, the updated forging process parameters are obtained. The optimal forging process parameters are obtained based on updated forging process parameters, finite element simulation model, and surrogate model. The method for obtaining the reference comprehensive adaptation degree is as follows: For any set of forging process parameters, the percentage of the simulation data under that set of forging process parameters falling into the critical position of the safe forming range is taken as the safe zone matching degree corresponding to that set of forging process parameters. Based on the proportion of key positions falling into the optimal grain refinement range and the actual grain size range under the simulation data of this set of forging process parameters, the contribution of this set of forging process parameters to the microstructure uniformity is obtained; among them, the proportion of key positions falling into the optimal grain refinement range is positively correlated with the contribution of microstructure uniformity, while the actual grain size range is negatively correlated with the contribution of microstructure uniformity. The ratio of the equivalent plastic strain to the preset cracking critical strain at each defect-prone critical location under this set of forging process parameters is taken as the cracking risk value at each defect-prone critical location. The result of negatively correlated and boundary-trunculated maximum cracking risk values ​​is used as the forming risk aversion degree corresponding to this set of forging process parameters. The weighted sum of the safety zone matching degree, the contribution of the microstructure uniformity degree, and the forming risk avoidance degree is used as the reference comprehensive fit degree corresponding to the set of forging process parameters.

2. The method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method as described in claim 1, characterized in that, The method for obtaining the simulation data is as follows: Hot deformation parameters were extracted from basic data of titanium alloys, and hyperbolic sinusoidal Arrhenius constitutive equations were constructed. Based on the equipment slider speed and die geometry in each set of forging process parameters, the forging strain rate, forging temperature and equivalent plastic strain at each key position are solved by finite element thermo-mechanical coupling analysis. The corresponding rheological stress state is calculated by combining the hyperbolic sinusoidal Arrhenius constitutive equation. Forging strain rate, forging temperature, equivalent plastic strain, and rheological stress state are all used as simulation data.

3. The method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method as described in claim 2, characterized in that, The method for obtaining the safe forming range and the optimal grain refinement range is as follows: Based on the forging strain rate, forging temperature and preset grain evolution model in the simulation data, the dynamic recrystallized grain size at each key position under each set of forging process parameters is calculated. The range between forging strain rate and forging temperature, where no rheological instability occurs and the deformation resistance is within the preset allowable load range of the equipment, is extracted as the safe forming range; Extract the subset of dynamically recrystallized grains within the safe forming range that meet the preset qualified size and whose grain size range is less than the preset range threshold, and use it as the optimal grain refinement range.

4. The method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method as described in claim 3, characterized in that, The method for obtaining the dynamic recrystallization grain size is as follows: Obtain material microstructure constants, grain size strain rate exponents, dynamic recrystallization apparent activation energy, and universal gas constants from basic data related to titanium alloys; For any set of forging process parameters and any key position, obtain the forging temperature and forging strain rate of the key position at the final forging time under the set of forging process parameters. The product of the general gas constant and the forging temperature is used as the first value; The ratio of the apparent activation energy of dynamic recrystallization to the first value is used as the second value; The third value will be the value of an exponential function with the natural constant as the base and the second value as the exponent. The product of the forging strain rate and the third value is used as the Zener-Hollomon parameter value; The fourth value is the inverse power of the grain size strain rate exponent of the Zener-Hollomon parameter. The product of the material microstructure constant and the fourth value is used as the dynamic recrystallization grain size at this critical location under this set of forging process parameters.

5. The method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method as described in claim 1, characterized in that, The method for obtaining the critical locations prone to defects is as follows: For any set of forging process parameters, obtain the static structural design parameters of the titanium alloy forging, as well as the stress characteristics in the simulation data corresponding to that set of forging process parameters. Based on static structural design parameters and stress characteristics, the key locations where strain concentration or poor metal flow occurs during the forging process are analyzed. The critical locations where strain concentration or poor metal flow occurs, and where there is a risk of cracking, folding, or insufficient filling, are identified as the critical defect-prone locations under this set of forging process parameters.

6. The method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method as described in claim 1, characterized in that, The method for obtaining the prediction accuracy is as follows: Input multiple sets of preset forging process parameters into the trained proxy model, and output the overall model fit degree corresponding to each set of forging process parameters; The difference between the model's overall fit degree and the reference overall fit degree corresponding to each set of forging process parameters is converted relative to the reference benchmark and used as the model's relative error corresponding to each set of forging process parameters. The result of negatively correlating the mean of the relative error of the model is used as the prediction accuracy of the surrogate model.

7. The method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method as described in claim 6, characterized in that, The method for obtaining the updated forging process parameters is as follows: For any forging process parameter, the difference between constant 1 and the current output model comprehensive fit is multiplied by the prediction accuracy, the preset step size adjustment coefficient and the correlation direction coefficient of the forging process parameter, and then divided by the preset sensitivity weight of the forging process parameter to obtain the adjustment increment of the forging process parameter. The adjustment increment is multiplied by the preset value range of the forging process parameter, and then added to the current base value of the forging process parameter to obtain the updated forging process parameter.

8. The method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method as described in claim 7, characterized in that, The method for obtaining the correlation direction coefficient is as follows: For any forging process parameter, a preset small positive increment is applied to the current benchmark value, and the current benchmark values ​​of the other forging process parameters are input into the proxy model to obtain the comprehensive adaptation degree of the pathfinding prediction. If the overall adaptation degree of the path prediction is greater than or equal to the overall adaptation degree of the currently output model, then the correlation direction coefficient of the forging process parameter is set to the first preset value. The first preset value is a positive number; If the overall fit of the pathfinding prediction is less than the overall fit of the currently output model, then the correlation direction coefficient of the forging process parameter is set to a second preset value; the second preset value is a negative number.

9. The method for optimizing forging process parameters of large-size titanium alloy forgings based on finite element method as described in claim 1, characterized in that, The method for obtaining the optimal forging process parameters is as follows: The updated forging process parameters are input into the proxy model to re-predict the overall model fit. If the overall fit of the model meets the preset initial screening conditions, the updated forging process parameters will be input into the finite element simulation model for simulation verification to obtain the true reference overall fit. If the degree of comprehensive adaptation of the real reference meets the preset termination condition, the iteration stops and the updated forging process parameters are used as the optimal forging process parameters. If the true reference comprehensive adaptation degree does not meet the preset termination condition, the updated forging process parameters and their true reference comprehensive adaptation degree will be added to the sample set, the training agent model will be updated, and the step of obtaining the updated forging process parameters will be returned. If the overall model fit does not meet the preset initial screening conditions, the updated forging process parameters will be used as the baseline value, and the process will return to the step of obtaining the updated forging process parameters. When the cumulative iterations reach the maximum number of iterations, the iteration stops, and the set of forging process parameters with the highest degree of comprehensive adaptation to the real reference in all iterations is output as the optimal forging process parameters.

Images