A temperature field control method for a large forging process

Abstract

The application discloses a large forging piece forging process temperature field control method, which is based on forging simulation experiment to establish a forging piece deformation heat generation model, a friction heat generation model and a forging piece and environment heat exchange model in a forging process, and is compiled into a subroutine to be embedded into a finite element simulation platform of the forging process; finite element simulation calculation of the forging process is carried out to obtain prior data of a forging piece temperature field, and a three-dimensional manifold equidistant mapping model of the forging piece temperature field and a time sequence prediction model of two-dimensional data after mapping are established; the surface temperature of the forging piece is measured in real time, posterior data and posterior probability of the forging piece temperature field at the current moment are calculated, the forging piece temperature field at the next moment is predicted, and the forging process parameters are adjusted accordingly. The large forging piece forging process temperature field control method provided by the application can not only effectively obtain prior data to predict the forging piece temperature field, but also correct the forging process parameters based on the posterior probability, so that the actual research and development cost and time cost of enterprises are reduced, and the accurate regulation and control of the forging piece temperature field in the forging process are realized.

Description

Technical Field

[0001] This invention belongs to the field of material forming and control, and specifically relates to a method for controlling the temperature field during the forging process of large forgings. Background Technology

[0002] Large forgings are key component materials for equipment in fields such as heavy machinery, large ships, military equipment, and aerospace.

[0003] Large forgings used in high-end equipment require high strength, high toughness, and corrosion resistance. Therefore, they are mostly made of high-performance, difficult-to-form metals such as titanium alloys, high-temperature alloys, and high-strength aluminum alloys. These materials have characteristics such as narrow forging temperature ranges and high deformation resistance. In particular, when using these materials to forge large and complex parts, forging cracking is very likely to occur, and the traditional "trial and error" method of process control is too costly.

[0004] Temperature, as a key process parameter in the forging of large forgings, not only affects the flow properties and integrity of the forging material, but also directly influences the microstructure and final product performance. At high temperatures, the material exhibits good flowability and filling properties, resulting in forged products with advantages such as high precision, high surface quality, and dimensional stability. However, in actual forging processes, due to forging deformation and thermo-mechanical coupling effects, the temperature field of the forging exhibits extremely complex distribution characteristics, which cannot be accurately described and precisely controlled using basic heat transfer principles.

[0005] Machine learning is one of the most popular branches of artificial intelligence. It can generate models based on empirical data, which can provide new judgments under new circumstances. Statistical machine learning can build probabilistic models based on empirical data and perform prediction and analysis on the data. Experts at home and abroad have repeatedly tried to use machine learning methods to predict temperature fields, but due to the difficulty in constructing three-dimensional prediction networks, the final accuracy is difficult to meet control requirements.

[0006] The patent specification with publication number CN115630565A discloses a temperature field reconstruction method based on an embedded physical knowledge neural network. However, it considers too few influencing factors, cannot achieve real-time control of the temperature field of forgings, and cannot guide the adjustment of process parameters in the forging process.

[0007] The patent specification with publication number CN115454175B discloses a real-time temperature measurement system and method for forging 690 alloy. This method only involves the surface temperature control of 690 alloy forgings and cannot regulate the internal temperature field of the forgings, thus failing to achieve accurate monitoring and control of the temperature state of the entire forging system. Summary of the Invention

[0008] To address the shortcomings of existing technologies, the present invention aims to provide a method for controlling the temperature field during the forging process of large forgings, comprising the following steps:

[0009] Step 1: Based on forging simulation experiments, establish deformation heat generation models, friction heat generation models, and heat exchange models between forgings and the environment during the forging process;

[0010] Step 2: Build a finite element simulation platform for the forging process, write subroutines to embed all the models in Step 1, carry out finite element simulation calculations of the forging process, and obtain prior data of the temperature field of the forging.

[0011] Step 3: Establish a three-dimensional manifold isometric mapping model of forging temperature and a time-series prediction model of the mapped two-dimensional data;

[0012] Step 4: Measure the surface temperature of the forging in real time during the forging process, calculate the posterior data and posterior probability of the forging temperature field at the current moment, and predict the forging temperature field at the next moment.

[0013] Step 5: Determine whether the predicted temperature field value of the forging at the next moment exceeds the preset range. If yes, continue to step 6; otherwise, continue to step 4.

[0014] Step 6: Calculate and adjust the pressing rate of the forging press and the heating temperature of the forging.

[0015] Preferably, in the temperature field control method for the forging process of large forgings as described above, step 1 includes the following steps:

[0016] Step 1.1: Based on the forging simulation experiment, measure the deformation heat absorption coefficient η of the forging, and establish a deformation heat generation model for the forging. Its specific expression is as follows: In the formula q d The heat flux density per unit volume of deformation per unit time is given by σ, where σ is the equivalent stress.

[0017] The equivalent rate of change;

[0018] Step 1.2: Based on the forging simulation experiment, measure the frictional heat absorption coefficient μ of the forging, and establish a frictional heat generation model for the forging. Its specific expression is as follows: In the formula q f Let τ be the frictional heat flux density per unit area per unit time, and τ be the shear stress on the friction surface. The relative motion rate between the forging surface and the die in the direction of shear stress;

[0019] Step 1.3: Based on forging simulation experiments, measure the heat exchange coefficient h between the forging and the environment, and the thermal radiation coefficient k between the forging and the environment, and establish a heat exchange model between the forging and the environment. The specific expression is as follows: In the formula q hT is the heat flux density exchanged between a unit area of ​​forging and the environment per unit time. E T represents the ambient temperature of the forging surface. F This refers to the surface temperature of the forging.

[0020] Preferably, in the temperature field control method for the forging process of large forgings as described above, step 3 includes the following steps:

[0021] Step 3.1: Extract the nodal coordinates corresponding to different forging times in the three-dimensional finite element simulation model of the forging. With temperature value

[0022] Step 3.2: Use the manifold isometric mapping algorithm and optimize the nearest neighbor parameter in the algorithm to obtain the node coordinates. Two-dimensional projected coordinates With projection matrix W ε ;

[0023] Step 3.3: Calculate the projected temperature

[0024] Step 3.4: Based on the loss function Training projection coordinates With projected temperature Long Short-Term Memory Recurrent Neural Network Temporal Prediction Model

[0025] Preferably, in the temperature field control method for the forging process of large forgings as described above, step 4 includes the following steps:

[0026] Step 4.1: Based on the current surface temperature of the forging The internal temperature of the forging was calculated by interpolation with the prior dataset. Then the three-dimensional temperature posterior data of the forging

[0027] Step 4.2: Calculate the posterior probability of the forging temperature field according to the following formula:

[0028]

[0029] In the formula These are the temperature field sequences of the forging at the current moment in the prior and posterior data, respectively. This is the distribution column of the three-dimensional temperature prediction data of the forging at the current moment;

[0030] Step 4.3: Store the posterior data to the set capacity and iteratively train the time series prediction model in Step 3.

[0031] Preferably, in the temperature field control method for the forging process of large forgings as described above, step 6 includes the following steps:

[0032] Step 6.1: Calculate the required heat exchange volume on the surface of the forging at the next moment. In the formula Δt ε t represents the difference between the preset temperature and the predicted temperature per unit area at the next time step. ε+Δε The preset temperature per unit area at the next moment, Ω represents all surfaces of the forging that can exchange heat;

[0033] Step 6.2: Calculate and adjust the forging heating temperature for the next moment. In the formula The average ambient temperature of the forging surface at the current moment is denoted by , and S is the surface area of ​​the forging that can be heated.

[0034] Step 6.3: Calculate the required deformation heat generation inside the forging at the next moment. In the formula, ρ is the density of the forging material, c is the specific heat capacity of the forging material, and Δτ ε The difference between the preset temperature and the predicted temperature per unit volume at the next moment is Ψ, where Ψ represents the internal structure of the forging that can deform.

[0035] Step 6.4: Calculate and adjust the pressing speed of the forging press at the next moment using the following formula:

[0036]

[0037] In the formula v ε H ε and These represent the current pressing speed of the forging press, the height of the forging, and the average equivalent stress of the forging, respectively, while V is the overall volume of the forging that can deform.

[0038] Preferably, in the temperature field control method for the forging process of large forgings as described above, the large forging refers to a three-dimensional solid model whose topological structure is simply connected, and the three side lengths x, y, z of the smallest circumscribed cube of the forging billet all satisfy the length constraint: y≥z≥0.2x≥100mm;

[0039] Compared with the prior art, the main advantages of the technical solution of the present invention are as follows:

[0040] 1. The temperature field control method for the forging process of large forgings provided by this invention can not only consider the effects of deformation heat generation, friction heat generation, and heat exchange between the forging and the environment during the forging process, but also establish a three-dimensional manifold equidistant mapping model of the forging temperature field and a time-series prediction model of the mapped two-dimensional data, so as to more accurately obtain prior data to predict the forging temperature field. At the same time, it can adjust the pressing rate of the forging press and the heating temperature of the forging in real time based on monitoring data and posterior probability, so as to achieve precise control of the temperature field of large forgings, which is conducive to the development of digital and intelligent manufacturing technology.

[0041] 2. This invention can save production costs and has wide applicability. For different types of large forgings, it can make full use of the heat generated by the deformation of the material and the heating device to quickly and accurately control the overall temperature field of the forging, reduce energy consumption, and effectively reduce material waste in large forgings. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of a method for controlling the temperature field during the forging process of large forgings in a preferred embodiment of the present invention;

[0044] Figure 2 These are photographs taken during the forging process of a large TC18 titanium alloy forging in a preferred embodiment of the present invention.

[0045] Figure 3 This is a predicted temperature field diagram of a large TC18 titanium alloy forging at the moment when the forging pressure ε = 50% during the forging process in a preferred embodiment of the present invention.

[0046] Figure 4 This is a diagram showing the overall temperature field control results of the large TC18 titanium alloy forging at the moment when the forging pressure ε = 50% during the forging process in a preferred embodiment of the present invention.

[0047] Figure 5 This is a comparison chart of the predicted and controlled temperatures during the forging process of a large TC18 titanium alloy forging in a preferred embodiment of the present invention. Detailed Implementation

[0048] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0049] like Figure 1 As shown, this invention is a method for controlling the temperature field during the forging process of large forgings, comprising the following steps:

[0050] Step 1: Based on forging simulation experiments, establish deformation heat generation models, friction heat generation models, and heat exchange models between forgings and the environment during the forging process;

[0051] Step 2: Build a finite element simulation platform for the forging process, write subroutines to embed all the models in Step 1, carry out finite element simulation calculations of the forging process, and obtain prior data of the temperature field of the forging.

[0052] Step 3: Establish a three-dimensional manifold isometric mapping model of forging temperature and a time-series prediction model of the mapped two-dimensional data;

[0053] Step 4: Measure the surface temperature of the forging in real time during the forging process, calculate the posterior data and posterior probability of the forging temperature field at the current moment, and predict the forging temperature field at the next moment.

[0054] Step 5: Determine whether the predicted temperature field value of the forging at the next moment exceeds the preset range. If yes, continue to step 6; otherwise, continue to step 4.

[0055] Step 6: Calculate and adjust the pressing rate of the forging press and the heating temperature of the forging.

[0056] Preferably, in the temperature field control method for the forging process of large forgings as described above, step 1 includes the following steps:

[0057] Step 1.1: Based on the forging simulation experiment, measure the deformation heat absorption coefficient η of the forging, and establish a deformation heat generation model for the forging. Its specific expression is as follows: In the formula q d The heat flux density per unit volume of deformation per unit time is given by σ, where σ is the equivalent stress.

[0058] The equivalent rate of change;

[0059] Step 1.2: Based on the forging simulation experiment, measure the frictional heat absorption coefficient μ of the forging, and establish a frictional heat generation model for the forging. Its specific expression is as follows: In the formula q f Let τ be the frictional heat flux density per unit area per unit time, and τ be the shear stress on the friction surface. The relative motion rate between the forging surface and the die in the direction of shear stress;

[0060] Step 1.3: Based on forging simulation experiments, measure the heat exchange coefficient h between the forging and the environment, and the thermal radiation coefficient k between the forging and the environment, and establish a heat exchange model between the forging and the environment. The specific expression is as follows: In the formula q h T is the heat flux density exchanged between a unit area of ​​forging and the environment per unit time. E T represents the ambient temperature of the forging surface. F This refers to the surface temperature of the forging.

[0061] Preferably, in the temperature field control method for the forging process of large forgings as described above, step 3 includes the following steps:

[0062] Step 3.1: Extract the nodal coordinates corresponding to different forging times in the three-dimensional finite element simulation model of the forging. With temperature value

[0063]

[0064] Step 3.2: Use the manifold isometric mapping algorithm and optimize the nearest neighbor parameter in the algorithm to obtain the node coordinates. Two-dimensional projected coordinates With projection matrix W ε ;

[0065] Step 3.3: Calculate the projected temperature

[0066] Step 3.4: Based on the loss function Training projection coordinates With projected temperature Long Short-Term Memory Recurrent Neural Network Temporal Prediction Model

[0067] Preferably, in the temperature field control method for the forging process of large forgings as described above, step 4 includes the following steps:

[0068] Step 4.1: Based on the current surface temperature of the forging The internal temperature of the forging was calculated by interpolation with the prior dataset. Then the three-dimensional temperature posterior data of the forging

[0069] Step 4.2: Calculate the posterior probability of the forging temperature field according to the following formula:

[0070]

[0071] In the formula These are the temperature field sequences of the forging at the current moment in the prior and posterior data, respectively. This is the distribution column of the three-dimensional temperature prediction data of the forging at the current moment;

[0072] Step 4.3: Store the posterior data to the set capacity and iteratively train the time series prediction model in Step 3.

[0073] Preferably, in the temperature field control method for the forging process of large forgings as described above, step 7 includes the following steps:

[0074] Step 6.1: Calculate the required heat exchange volume on the surface of the forging at the next moment. In the formula Δt ε t represents the difference between the preset temperature and the predicted temperature per unit area at the next time step. ε+Δε The preset temperature per unit area at the next moment, Ω represents all surfaces of the forging that can exchange heat;

[0075] Step 6.2: Calculate and adjust the forging heating temperature for the next moment. In the formula The average ambient temperature of the forging surface at the current moment is denoted by , and S is the surface area of ​​the forging that can be heated.

[0076] Step 6.3: Calculate the required deformation heat generation inside the forging at the next moment. In the formula, ρ is the density of the forging material, c is the specific heat capacity of the forging material, and Δτ ε The difference between the preset temperature and the predicted temperature per unit volume at the next moment is Ψ, where Ψ represents the internal structure of the forging that can deform.

[0077] Step 6.4: Calculate and adjust the pressing speed of the forging press at the next moment using the following formula:

[0078]

[0079] In the formula v ε H ε and These represent the current pressing speed of the forging press, the height of the forging, and the average equivalent stress of the forging, respectively, while V is the overall volume of the forging that can deform.

[0080] Preferred embodiment:

[0081] In this preferred embodiment, a large TC18 titanium alloy billet with a nominal size of Φ600mm×1200mm is used for temperature field control during the forging process, specifically including the following steps:

[0082] Step 1: Based on forging simulation experiments, the deformation heat absorption coefficient η of the forgings between 800℃ and 1100℃ was measured to be 0.73–0.78, the frictional heat absorption coefficient μ was 0.40–0.47, and the heat exchange coefficient with the environment was h was 0.79–0.85 W / (mm). 2 ·K), thermal emissivity k=0.0078~0.0091W / (m²) 2 ·K 4 To establish models for deformation heat generation, friction heat generation, and heat exchange between the forging and the environment during the forging process, the specific expressions are as follows:

[0083]

[0084] In the formula q d The heat flux density per unit volume of deformation per unit time is given by σ, where σ is the equivalent stress. For the equivalent rate of change, q f Let τ be the frictional heat flux density per unit area per unit time, and τ be the shear stress on the friction surface. q represents the relative velocity between the forging surface and the die in the direction of shear stress. h T is the heat flux density exchanged between a unit area of ​​forging and the environment per unit time. E T represents the ambient temperature of the forging surface. F This refers to the surface temperature of the forging.

[0085] Step 2: Build a finite element simulation platform for the forging process using Abaqus 2022 software. Write the model from Step 1 into a Dflux subroutine and embed it into the finite element model to perform finite element simulation calculations of the forging process, obtaining the calculated output frequency f. s =Prior data of the temperature field of forgings at 1Hz;

[0086] Step 3: Extract the node coordinates corresponding to different forging moments in the three-dimensional finite element simulation model of the forging. With temperature value The manifold isometric mapping algorithm was used and optimized to obtain a nearest neighbor parameter of 8, thus obtaining the node coordinates. Two-dimensional projected coordinates With projection matrix W ε ; Calculate the projected temperature According to the loss function Training projection coordinates With projected temperature Long Short-Term Memory Recurrent Neural Network Temporal Prediction Model

[0087] Step 4: As Figure 2 As shown, TC18 titanium alloy billet forging production was carried out. The surface temperature of the forging was monitored during the forging process using a thermal imager, with a sampling frequency of f. p =1Hz; based on the surface temperature of the forging at the current moment. The internal temperature of the forging was calculated by interpolation with the prior dataset. Record the three-dimensional temperature post-hoc data of the forging Substituting the values ​​into the formula, we can calculate the posterior probability of the temperature field of the forging. In the formula These are the temperature field sequences of the forging at the current moment in the prior and posterior data, respectively. The distribution of the three-dimensional temperature prediction data of the forging at the current moment is shown in the figure; the posterior data is stored, and the time-series prediction model in step 3 is trained iteratively every 5 seconds.

[0088] Step 5: Determine whether the predicted temperature field value of the forging at the next moment exceeds the preset range. If yes, continue to step 7; otherwise, continue to step 4.

[0089] Step 6: Calculate the required heat exchange volume on the surface of the forging at the next moment. In the formula Δt ε t represents the difference between the preset temperature and the predicted temperature per unit area at the next time step. ε+Δε The preset temperature per unit area for the next moment, Ω represents all surfaces of the forging capable of heat exchange; calculate and adjust the heating temperature of the forging for the next moment. In the formula Let S be the average ambient temperature of the forging surface at the current moment, and S be the surface area of ​​the forging that can be heated; calculate the required deformation heat generation inside the forging at the next moment. In the formula, ρ is the density of the forging material, c is the specific heat capacity of the forging material, and Δτ ε Let Ψ be the difference between the preset temperature and the predicted temperature per unit volume at the next moment, and Ψ be the internal integral part of the forging that can deform; calculate and adjust the pressing rate of the forging press at the next moment according to the following formula: In the formula v ε H ε and These represent the current pressing speed of the forging press, the height of the forging, and the average equivalent stress of the forging, respectively, while V is the overall volume of the forging that can deform.

[0090] The predicted temperature of the forging at the final compression ratio ε = 50% is shown in the figure below. Figure 3 As shown, the overall temperature field diagram calculated based on the surface temperature is as follows: Figure 4 As shown, from Figure 5 It can be seen that the correlation coefficient between the temperature prediction and the control result is R = 0.9946, and the root mean square error (RMSE) is about 4℃, indicating that the present invention has good practical effect.

[0091] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for controlling the temperature field during the forging process of large forgings, characterized in that... This method predicts the temperature field of forgings based on prior data and corrects the control in real time based on posterior probability. Specifically, it includes the following steps: Step 1: Based on forging simulation experiments, establish deformation heat generation models, friction heat generation models, and heat exchange models between forgings and the environment during the forging process; Step 2: Build a finite element simulation platform for the forging process, write subroutines to embed all the models in Step 1, carry out finite element simulation calculations of the forging process, and obtain prior data of the temperature field of the forging. Step 3: Establish a three-dimensional manifold isometric mapping model of forging temperature and a time-series prediction model of the mapped two-dimensional data; Step 4: Measure the surface temperature of the forging in real time during the forging process, calculate the posterior data and posterior probability of the forging temperature field at the current moment, and predict the forging temperature field at the next moment. Step 5: Determine whether the predicted temperature field value of the forging at the next moment exceeds the preset range. If yes, continue to step 6; otherwise, continue to step 4. Step 6: Calculate and adjust the forging press reduction rate and forging heating temperature, which includes the following sub-steps: Step 6.1: Calculate the required heat exchange volume on the surface of the forging at the next moment. In the formula The heat exchange coefficient between the forging and the environment, The thermal radiation coefficient of the forging and the environment. The difference between the preset temperature and the predicted temperature per unit area at the next time step. Preset the temperature per unit area at the next moment. All surfaces of the forging that can undergo heat exchange; Step 6.2: Calculate and adjust the forging heating temperature for the next moment. In the formula The average ambient temperature of the forging surface at the current moment. These are the forging temperature values ​​corresponding to different forging times. For the three-dimensional temperature post-hoc data of the forging, The surface area of ​​the forging that can be heated; Step 6.3: Calculate the required deformation heat generation inside the forging at the next moment. In the formula For the density of the forging material, For the specific heat capacity of the forging material, The difference between the preset temperature and the predicted temperature per unit volume at the next time step. Forgings are integral internal parts that can deform. Step 6.4: Calculate and adjust the pressing speed of the forging press at the next moment using the following formula: ， In the formula , and These represent the current pressing speed of the forging press, the height of the forging, and the average equivalent stress of the forging, respectively. The overall volume of the forging that can undergo deformation.

2. The method for controlling the temperature field during the forging process of large forgings as described in claim 1, characterized in that, Step 3 includes the following steps: Step 3.1: Extract the nodal coordinates corresponding to different forging times in the three-dimensional finite element simulation model of the forging. With temperature value ; Step 3.2: Use the manifold isometric mapping algorithm and optimize the nearest neighbor parameter in the algorithm to obtain the node coordinates. Two-dimensional projected coordinates With projection matrix ; Step 3.3: Calculate the projected temperature = ; Step 3.4: Based on the loss function Training projected coordinates With projected temperature Long Short-Term Memory Recurrent Neural Network Temporal Prediction Model .

3. The method for controlling the temperature field during the forging process of large forgings as described in claim 1, characterized in that, Step 4 includes the following steps: Step 4.1: Based on the current surface temperature of the forging The internal temperature of the forging was calculated by interpolation with the prior dataset. Then the three-dimensional temperature posterior data of the forging ; Step 4.2: Calculate the posterior probability of the forging temperature field according to the following formula: ， In the formula , These are the temperature field sequences of the forging at the current moment in the prior and posterior data, respectively. This is the distribution column of the three-dimensional temperature prediction data of the forging at the current moment; Step 4.3: Store the posterior data to the set capacity and iteratively train the time series prediction model in Step 3.

Images