Titanium alloy rolling process temperature monitoring and control method, system and device

Abstract

The application discloses a titanium alloy rolling process temperature monitoring and control method, system and device, and relates to the technical field. The application is specifically aimed at the problems of inaccurate core temperature monitoring, control lag and nonlinear change of titanium alloy thermal properties caused by surface temperature measurement in the prior art titanium alloy rolling process. The application constructs a prediction model of a three-dimensional internal temperature field by fusing material thermal physical properties and deformation thermal effects, collects surface temperature and calibrates the model. Based on the calibrated internal temperature distribution, a collaborative adjustment instruction of rough rolling speed, finish rolling speed and induction heating power is generated according to a temperature-speed-power mapping rule. Closed-loop control is performed, and the temperature evolution law is mined from historical data based on a time convolution network to continuously optimize the control strategy. Through the above technical scheme, the application realizes titanium alloy rolling core temperature sensing and regulation, suppresses the burning core phenomenon, significantly improves product stability, and provides a titanium alloy rolling method.

Description

Technical Field

[0001] This application relates to the field of metallurgical technology, and in particular to a method, system and equipment for monitoring and controlling the temperature during the rolling process of titanium alloys. Background Technology

[0002] Titanium alloys, due to their superior physical properties such as low specific gravity, corrosion resistance, high strength, and excellent high and low temperature performance, have become indispensable key materials in modern industry. They are increasingly widely used in high-tech fields such as aerospace, medical devices, 3C electronic products, and shipbuilding, and their market demand continues to grow. Therefore, the large-scale production and efficient processing of titanium alloys has become one of the important research directions in the field of materials science and engineering. Among various processing methods, hot rolling, with its advantages of high efficiency and low cost, is widely used in the preparation of titanium alloy wires, bars, and plates, and is of paramount importance for realizing the industrial mass production of titanium alloys.

[0003] However, with the continuous development of related technologies and the increasingly stringent requirements placed on the performance indicators of titanium alloy products by application scenarios, traditional technical solutions based primarily on surface temperature measurement cannot accurately and in real-time reflect the true temperature distribution of the core and even the entire cross-section of the rolled piece. This blind spot in the internal temperature state means that existing temperature control strategies are often only a lagging and coarse adjustment, making it difficult to achieve refined and proactive control of the overall temperature of the rolled piece, especially the core. Therefore, existing technologies urgently need improvement to address these issues. Summary of the Invention

[0004] In order to overcome at least the above-mentioned deficiencies in the prior art, the purpose of this application is to provide a method, system and equipment for monitoring and controlling the temperature in the titanium alloy rolling process to solve the above problems.

[0005] In a first aspect, this application provides a method for monitoring and controlling the temperature during the rolling process of titanium alloys, comprising the following steps: Step S1: Based on the database of thermodynamics and deformation behavior of titanium alloys, and based on the heat conduction equation, obtain the first temperature prediction model; Step S2: Based on the real-time titanium alloy rolling data and the first temperature prediction model, obtain the predicted overall temperature field of the rolled piece, which is used as the first temperature. Step S3: Obtain the measured surface temperature of the rolled piece based on real-time titanium alloy rolling data, and use it as the second temperature; Step S4: Based on the first temperature and the second temperature, obtain the overall temperature field of the rolled piece for calibration, and use it as the third temperature; Step S5: Adjust the control parameters during the rolling process according to the third temperature.

[0006] In one possible implementation, the establishment of a first temperature prediction model based on the heat conduction equation, according to a database of titanium alloy thermodynamics and deformation behavior, includes: A database of titanium alloy thermodynamic and deformation behavior is obtained based on the thermophysical properties, mechanical properties, historical batch data, and historical titanium alloy rolling data of titanium alloy material. Based on the thermophysical property parameters of the titanium alloy thermodynamic and deformation behavior database, density, specific heat capacity, and thermal conductivity are obtained by looking up tables. Based on the mechanical property parameters of the titanium alloy thermodynamics and deformation behavior database, and based on the deformation behavior during the rolling process, the plastic deformation work is converted into heat energy to obtain the rate of heat generation from plastic deformation inside the rolled piece, which is used as the internal heat source item of the rolled piece. Boundary conditions were obtained based on historical titanium alloy rolling data. Based on the density, specific heat capacity, thermal conductivity, internal heat source term of the titanium alloy, and boundary conditions, a verified three-dimensional unsteady-state partial differential equation of heat conduction is obtained based on the heat conduction equation, which serves as the first temperature prediction model.

[0007] In one possible implementation, obtaining the predicted overall temperature field of the rolled piece as the first temperature based on the first temperature prediction model, according to real-time titanium alloy rolling data, includes: Based on real-time titanium alloy rolling data and the first temperature prediction model, the predicted overall temperature distribution of the rolled piece is obtained. Based on the predicted overall temperature distribution of the rolled piece, the rolled piece is discretized into three-dimensional spatial grid elements, and the temperature of each predicted grid element is obtained by solving the problem using the finite element method. Based on the predicted temperature of each grid cell, the predicted overall temperature field of the rolled piece is obtained; the overall temperature field of the rolled piece includes the detectable surface temperature and the internal temperature field that cannot be directly measured. The overall temperature field of the rolled piece is predicted and used as the first temperature.

[0008] In one possible implementation, obtaining the measured surface temperature of the rolled piece as a second temperature based on real-time titanium alloy rolling data includes: Real-time titanium alloy rolling data is collected from monitoring points installed on the titanium alloy rolling production line. The real-time titanium alloy rolling data includes: surface temperature of the rolled piece, equipment operating parameters, and geometric parameters of the rolled piece collected at various locations. Real-time titanium alloy rolling data is collected by monitoring points set up on the titanium alloy rolling production line; Based on the real-time titanium alloy rolling data, the surface temperature of the rolled piece is obtained at the following locations: the box furnace exit, the roughing mill after rolling, the online induction furnace exit before the rolled piece enters the finishing mill, and the finished product collection device. The measured surface temperature of the rolled piece is obtained based on at least one of the surface temperatures of the rolled piece, and is used as the second temperature.

[0009] In one possible implementation, obtaining the calibrated predicted internal temperature distribution of the rolled piece as a third temperature based on the first and second temperatures includes: Based on the real-time measured surface temperature of the rolled piece obtained from the second temperature, the predicted surface temperature of the rolled piece output at the corresponding position in the first temperature distribution is obtained. The observation equation is obtained by comparing the measured temperature of the rolled surface with the predicted temperature of the rolled surface. Based on the observation equation, an extended Kalman filter algorithm is used for iterative calibration to obtain a calibrated prediction model for the rolled piece. The iterative calibration based on the observation equation using the extended Kalman filter algorithm includes: using the second temperature as the observation vector, constructing a state vector from the core temperature, surface temperature, and heat flux density in the first temperature, updating the covariance matrix of process noise and observation noise to calibrate the prediction model for the first temperature, and performing iterative calibration to achieve high-precision estimation of the internal temperature field.

[0010] Based on the calibrated prediction model for the rolled piece, the predicted internal temperature distribution of the rolled piece is obtained.

[0011] In one possible implementation, adjusting the control parameters during the rolling process based on the third temperature includes: Based on the third temperature, determine whether the third temperature of the core of the rolled piece is within the preset target processing window; If the third temperature of the core of the rolled piece is not within the target processing window, there is a risk of overheating or underheating. When there is a risk of overheating or underheating, based on the preset temperature-speed-power mapping rules, the specific adjustment amount of the rolling mill speed and heating power is obtained according to the third temperature. Based on the specific adjustment amount, smoothing is performed to obtain control parameters for controlling rolling speed and heating power.

[0012] In one possible implementation, when there is a risk of overheating or underheating, based on a preset temperature-speed-power mapping rule, and according to the third temperature, the specific adjustment amount of the rolling mill speed and heating power is obtained, including: When the core temperature of the workpiece is lower than the preset lower limit when it starts rolling, the rolling speed of the roughing mill is reduced. When the core temperature is higher than the upper limit before entering the finishing mill, the rolling speed of the finishing mill is increased, and at the same time, the power of the online induction heating furnace is reduced. One possible implementation also includes: Based on the aforementioned database of titanium alloy thermodynamics and deformation behavior, a fourth temperature prediction model is obtained using a time convolutional network. Based on the current control parameters, the second temperature, the third temperature of the current pass, the rolling speed, deformation, heating power and material code of the previous pass, the predicted core temperature of the next pass is obtained based on the fourth temperature prediction model. Based on the predicted core temperature for the next pass, the heating power of the box furnace and the preset roughing speed are determined before the start of the next roughing pass. Based on the predicted core temperature for the next pass and the predicted third temperature after the next roughing pass, the online induction furnace reheating power and the preset roughing rolling speed are obtained before finishing rolling.

[0013] Secondly, this application provides a temperature monitoring and control system for the titanium alloy rolling process, including a first temperature prediction model module, a first temperature prediction module, a second temperature acquisition module, a third temperature calibration module, and a control parameter acquisition module that are sequentially electrically connected. The first temperature prediction model module is configured to: obtain a first temperature prediction model based on the heat conduction equation according to the database of thermodynamics and deformation behavior of titanium alloys; The first temperature prediction module is configured to: obtain the predicted overall temperature field of the rolled piece based on the first temperature prediction model according to real-time titanium alloy rolling data, and use it as the first temperature; The second temperature acquisition module is configured to: obtain the measured surface temperature of the rolled piece based on real-time titanium alloy rolling data, and use it as the second temperature; The third temperature calibration module is configured to: obtain the calibrated predicted temperature distribution inside the rolled piece based on the first temperature and the second temperature, and use it as the third temperature; The control parameter acquisition module is configured to adjust the control parameters during the rolling process based on the third temperature.

[0014] The execution and feedback module issues instructions, and there is also a feedforward mechanism and a feedback mechanism.

[0015] Thirdly, this application provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of any one of the methods for monitoring and controlling the temperature in a titanium alloy rolling process.

[0016] In summary, the beneficial effects that this application can achieve are: This application proposes a method, system, and equipment for monitoring and controlling the temperature during the rolling process of titanium alloys, based on the temperature regime of titanium alloys. This achieves comprehensive, real-time, and predictive management of the hot continuous rolling process of titanium alloys. Through a physical mechanism model and command control of the equipment, adaptive and intelligent temperature monitoring and control are realized. By employing feedforward and feedback mechanisms, the balance between "core burning" defects caused by internal overheating and production efficiency is resolved. The method presented in this application demonstrates significant technological advancements and beneficial effects in improving the stability and controllability of titanium alloys during hot continuous rolling. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the method steps in an embodiment of this application; Figure 2 This is a schematic diagram of the system structure according to an embodiment of this application; Figure 3 The temperature change curves show the mechanical properties of titanium alloy TC4. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0019] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] Example 1 Please refer to the following: Figure 1The above is a schematic diagram of the steps of a method for monitoring and controlling the temperature of a titanium alloy rolling process provided in an embodiment of the present invention. Further, a method for monitoring and controlling the temperature of a titanium alloy rolling process may specifically include the contents described in steps S1-S5.

[0021] Step S1: Based on the database of thermodynamics and deformation behavior of titanium alloys, and based on the heat conduction equation, obtain the first temperature prediction model; Step S2: Based on the real-time titanium alloy rolling data and the first temperature prediction model, obtain the predicted overall temperature field of the rolled piece, which is used as the first temperature. Step S3: Obtain the measured surface temperature of the rolled piece based on real-time titanium alloy rolling data, and use it as the second temperature; Step S4: Based on the first temperature and the second temperature, obtain the overall temperature field of the rolled piece for calibration, and use it as the third temperature; Step S5: Adjust the control parameters during the rolling process according to the third temperature.

[0022] In existing technologies, titanium alloy rolling relies on surface temperature monitoring and passive feedback adjustment mechanisms. Traditional methods collect surface temperature data by placing infrared thermometers at key nodes of the rolling line and adjusting the mill speed and heating power in combination with preset process parameters. However, due to the low thermal conductivity of titanium alloys, there is a significant temperature difference between the surface and the core of the rolled piece. The surface temperature cannot accurately reflect the actual temperature state of the core. When an abnormal surface temperature is detected, irreversible damage may have already occurred in the internal area, resulting in substandard products. This lag control mode is difficult to adapt to titanium alloy processing.

[0023] In implementing the embodiments of this application, in order to solve the above-mentioned problems, the inventors found that simply relying on surface temperature feedback cannot solve the problem of dynamic changes in the internal temperature field of titanium alloys. Therefore, by analyzing the heat transfer mechanism of the rolling process and combining the heat generation and heat conduction laws of plastic deformation, a first temperature prediction model for predicting the temperature field is proposed, which can simultaneously reflect the surface and core temperature distribution as the first temperature. To solve the problem of model prediction error, real-time surface temperature measurement data is introduced to calibrate the first temperature prediction model and obtain the third temperature. Furthermore, considering the temporal correlation of rolling parameters in multiple passes, a closed-loop correction mechanism between prediction and measured data is formed.

[0024] In the implementation of this application, the method for monitoring and controlling the temperature during the titanium alloy rolling process includes: The preliminary preparation steps involve obtaining a database of the thermodynamic and deformation behavior of titanium alloys based on the thermophysical properties of the titanium alloy material, rolling process parameters, and historical batch data. Based on the material properties provided by the thermodynamic database, a three-dimensional temperature field prediction model is established to simulate the heat transfer process of the rolled piece, serving as the first temperature prediction model to predict the overall temperature distribution of the rolled piece. Finite element discretization decomposes the rolled piece into mesh elements, divides the rolled piece into three-dimensional spatial mesh elements, calculates the temperature change of each element with a set time step, and outputs the surface and core temperature distribution simultaneously using the finite element method to solve the model. The real-time measured surface temperature of the rolled piece is used as the second temperature. The real-time surface temperature data is compared with the predicted value, and the predicted temperature field is calibrated by the extended Kalman filter algorithm to obtain the third temperature; Based on the adjustment of the third temperature, the calibrated third temperature serves as the control reference, triggering the active adjustment of rolling speed and heating power to obtain rolling control parameters.

[0025] Further steps include constructing a fourth temperature prediction model based on time, employing a temporal convolutional network, analyzing historical data, performing incremental training through a sliding window strategy, introducing feedforward control and performance feedback mechanisms, adjusting and optimizing the rolling speed curve and reheating power, establishing a mapping relationship between temperature control and product performance, and forming a closed-loop optimization system.

[0026] This method overcomes the limitations of surface temperature monitoring by predicting and calibrating a three-dimensional temperature field in real time, enabling the perception of the overall temperature state of the rolled piece. By mining the temporal correlation in historical data, it enables the ability to adapt to process fluctuations. The dual mechanism of feedforward control and feedback transforms post-event adjustment into a combination of pre-event prediction and in-event control.

[0027] Through the above technical solutions, this application can obtain the core temperature status of the rolled piece in real time, avoid product defects caused by uneven temperature distribution, effectively eliminate model prediction deviations through the calibration mechanism, ensure the accuracy of temperature control commands, and significantly improve the temperature stability of the titanium alloy rolling process through the closed-loop control system, thus ensuring product quality.

[0028] Example 2 Based on Example 1, a method for monitoring and controlling the temperature during the rolling process of titanium alloys may further include the following:

[0029] Step S1: Based on the database of thermodynamics and deformation behavior of titanium alloys, and based on the heat conduction equation, obtain the first temperature prediction model; Step S2: Based on the real-time titanium alloy rolling data and the first temperature prediction model, obtain the predicted overall temperature field of the rolled piece, which is used as the first temperature. Step S3: Obtain the measured surface temperature of the rolled piece based on real-time titanium alloy rolling data, and use it as the second temperature; Step S4: Based on the first temperature and the second temperature, obtain the overall temperature field of the rolled piece for calibration, and use it as the third temperature; Step S5: Adjust the control parameters during the rolling process according to the third temperature.

[0030] In this embodiment, in the actual scenario of rolling titanium alloys, a preparation step is also included before step S1, which involves obtaining a database of titanium alloy thermodynamics and deformation behavior based on the thermophysical properties, mechanical properties, historical batch data, and historical titanium alloy rolling data of the titanium alloy material.

[0031] First, before the rolling process begins, preliminary preparations are made to construct a comprehensive database of titanium alloy thermodynamics and deformation behavior. This database is constructed based on the specific thermophysical properties of titanium alloys, their mechanical properties at different temperatures and strain rates, and rolling data from historical batches.

[0032] Thermophysical properties of the titanium alloy were precisely measured, covering different temperatures. Specific heat capacity under alloy composition conditions Thermal conductivity and density These parameters are measured at multiple points in a controlled atmosphere using specialized equipment such as differential scanning calorimeter, laser flare method, and high-precision density meter.

[0033] The database also records detailed mechanical property parameters of titanium alloys over a wide range of temperatures, strains, and strain rates. In this embodiment, the temperature is controlled between 750°C and 1100°C, the strain between 0.1 and 1.5, and the strain rate between 0.1 s⁻¹. 1 up to 100s 1 Internally, the core of mechanical performance parameters is rheological stress. Then, strain was used. strain rate Characterizing rheological stress as a function of temperature T For example, isothermal compression tests were conducted on cylindrical specimens of various titanium alloy grades at different temperatures and strain rates using a thermal simulation testing machine to obtain their rheological stress curves.

[0034] In addition, the database records historical batch data, including rolling batch numbers and production times. The rolling batch number includes the corresponding titanium material, production time, chemical composition, and traceability information on the quality of the raw materials, facilitating tracking. The database also comprehensively integrates historical titanium alloy rolling data, which covers information on at least 2,000 rolling batches. Specifically, it includes the starting rolling temperature, post-roughing rolling temperature, online induction heating power, heating time, roughing mill rolling speed, finishing mill temperature, finishing mill temperature, finishing mill rolling speed, finished product collection temperature, and online diameter measuring instrument test results for each pass. These collected data serve as monitoring results for process temperature and as a basis for process control.

[0035] Please refer to Figure 3 Taking titanium alloy TC4 as an example, the temperature change curve of the mechanical properties of titanium alloy TC4 shows that the tensile strength of TC4 decreases with increasing temperature, while the elongation reaches its maximum value at around 850℃. The rolling temperature range of 800~1000℃ is suitable for titanium alloy TC4. Table 1 shows the mechanical properties of TC4 and TC11 when rolled at 800-1100℃.

[0036] Table 1800-1100℃ Mechanical Properties

[0037] The table records the mechanical properties σ. b σ 0.2 In addition to δ, the actual mechanical properties also include σ. s σ -1 wait.

[0038] σ b It is tensile strength, measured in MPa, which refers to the maximum stress a metallic material can withstand before it breaks; σ 0.2 σ is the conditional yield strength, in MPa, which refers to the stress value when the specimen exhibits 0.2% residual plastic deformation; δ is the elongation, in %, which refers to the percentage of elongation of the gauge length after the specimen breaks compared to the original gauge length, and is calculated as δ = (L1 - L0) / L0 × 100% (where L1 is the gauge length after breakage and L0 is the original gauge length); -1 It refers to fatigue strength, also known as fatigue limit, measured in MPa. It is the maximum stress value that a material can withstand under symmetrical cyclic alternating stress for an infinite number of cycles without fatigue fracture; σ s It is the yield strength, measured in MPa, which refers to the stress at which a material begins to undergo significant plastic deformation.

[0039] Because large-scale mass production of rolled metal products is the most effective way to reduce costs and stabilize product quality, and the traceability of product quality, the summarization of quality processes, and the continuous optimization and improvement of process technology are the eternal themes of quality management, this invention is based on temperature regime, which mainly refers to process temperature control and endpoint temperature control. By recording and processing key temperature parameters, equipment operating parameters and final mechanical properties of rolled products in real time, the acquired data is analyzed to optimize the monitoring and control of temperature in the rolling process.

[0040] The thermal conductivity of TC4 titanium alloy is approximately 6.7 W / (m*K) at room temperature (25℃), decreasing to 4.5 W / (m*K) at 600℃, and approximately 2 W / (m*K) at 1000℃; TA8 titanium alloy has a thermal conductivity of approximately 15 W / (m*K) at 20℃, increasing to 17 W / (m*K) at 200℃, and approximately 20 W / (m*K) at 400℃; TA1 titanium alloy has a thermal conductivity of approximately 16 W / (m*K) at 25℃, 22 W / (m*K) at 400℃, and 1.42–3.03 W / (m*K) at 1000℃.

[0041] The thermal conductivity coefficients of different titanium alloys vary with temperature changes, but they are relatively low compared to common steels. The thermal conductivity coefficient of low alloy steel at room temperature (25°C) is 36–46 W / (m*K). As the temperature increases, the thermal conductivity first increases and then decreases, reaching 33 W / (m*K) at 1000°C.

[0042] Different titanium alloy materials require different temperature control during rolling. At the same time, it is necessary to monitor the temperature changes of each pass in a timely manner to avoid excessive temperature rise caused by rolling deformation, which could lead to core burning. For example, high core temperature can cause grain growth and reduced physical properties. Therefore, temperature monitoring and control during the titanium alloy rolling process is the key to ensuring process stability and product quality stability.

[0043] In step S1, based on the database of thermodynamics and deformation behavior of titanium alloys and the heat conduction equation, a first temperature prediction model is obtained to predict the overall temperature distribution of the rolled piece. The overall temperature distribution of the rolled piece includes the detectable surface temperature and the internal temperature field that cannot be directly measured. The heat conduction equation adopts a three-dimensional unsteady partial differential equation of heat conduction. In one possible implementation, based on a database of thermodynamic and deformation behavior of titanium alloys and the heat conduction equation, a first temperature prediction model is established, including: Density, specific heat capacity, and thermal conductivity were obtained by looking up the thermophysical property parameters from the database of titanium alloy thermodynamics and deformation behavior. Based on the mechanical property parameters of the titanium alloy thermodynamics and deformation behavior database, and based on the deformation behavior during the rolling process, the plastic deformation work is converted into heat energy to obtain the rate of heat generation from plastic deformation inside the rolled piece, which is used as the internal heat source item of the rolled piece. Boundary conditions were obtained based on historical titanium alloy rolling data. Based on the density, specific heat capacity, thermal conductivity, internal heat source term of the titanium alloy, and boundary conditions, a three-dimensional unsteady-state partial differential equation of heat conduction is obtained based on the heat conduction equation, which serves as the first temperature prediction model.

[0044] In the implementation of this application embodiment, after obtaining the database of thermodynamics and deformation behavior of titanium alloys, a first temperature prediction model is obtained based on the heat conduction equation to predict the overall temperature distribution of the rolled piece. The overall temperature distribution of the rolled piece covers the detectable surface temperature and the internal temperature field that cannot be directly measured.

[0045] In this embodiment, the heat conduction equation is characterized by a three-dimensional unsteady-state heat conduction partial differential equation.

[0046] Based on the thermophysical properties, the specific heat capacity of the rolled piece at different temperatures and compositions is first obtained by looking up tables and interpolation, or by using the previously fitted function expression. Thermal conductivity and density The values ​​of these parameters will be used as coefficients in the heat conduction equation.

[0047] Then, based on the mechanical property parameters and the deformation behavior during the rolling process, the rate of heat generation from plastic deformation inside the rolled piece is calculated. Based on the principle of converting work into heat energy during plastic deformation, the rate of heat generation during plastic deformation inside the rolled piece is obtained and used as a heat source term inside the rolled piece. This heat source term can be expressed as follows: .

[0048] in, It is the internal heat source term of the rolled piece, which quantifies the rate at which mechanical energy is converted into heat energy during the plastic deformation process of the rolled piece; The heat conversion efficiency is the Taylor-Kuni coefficient, which is usually taken between 0.9 and 1.0 to reflect that not all deformation work is converted into heat. Its value can be determined by reverse analysis of the temperature rise of the rolled workpiece after rolling. For rheological stress, its value is a complex function of temperature, strain, and strain rate. It is calculated using rolling conditions stored in the database, including rolling force, strain and strain rate derived from the geometric changes of the rolled piece, and the current temperature. It is the equivalent strain rate, derived from the change rate of roll speed and workpiece thickness.

[0049] Finally, historical titanium alloy rolling data is constructed, combined with thermophysical property parameters, rolling environment parameters, and roll parameters, to determine the boundary conditions of the heat conduction equation.

[0050] Boundary conditions should include at least: The heat transfer between the workpiece surface and the rolls is characterized by a contact thermal resistance model and roll cooling conditions; the convective heat transfer between the workpiece surface and the air is characterized by the convective heat transfer coefficient and real-time ambient temperature; the radiative heat transfer from the workpiece surface to the environment is characterized by the Stefan-Boltzmann law and workpiece emissivity; and the forced convective heat transfer under the action of the online cooling unit is characterized by a model of the flow rate, temperature, pressure, and heat transfer coefficient of the cooling medium parameters.

[0051] These parameters can all be obtained through laboratory simulation experiments, such as hot-pressing experiments or industrial field data inversion. Parameters such as heat transfer coefficient and emissivity are also stored in the database and can be adjusted according to real-time environmental parameters.

[0052] In summary, based on the specific heat capacity, thermal conductivity, density, internal heat source term, and boundary conditions mentioned above, a complete three-dimensional unsteady-state partial differential equation for heat conduction is constructed, expressed as: This mainly describes the change in internal temperature of the rolled piece over time, and the influence of heat conduction and internal heat sources. Among them, Represents material density, , where is the specific heat capacity and k is the thermal conductivity coefficient. These parameters are functions of temperature and material grade, and their values ​​are directly derived from the database established in the first step. For Hamiltonian operators, Through thermal conductivity The heat transfer term is a conduction term, where T represents temperature and t represents time. This represents the internal heat source generated by plastic deformation; This indicates that the heat introduced by external heat sources such as online induction heating furnaces can be measured in real time through their current and voltage and converted into heat flux density boundary conditions applied to the model. Its heating efficiency is usually between 60% and 80%, and its effective thermal power can reach the megawatt level.

[0053] Finally, the aforementioned three-dimensional unsteady-state heat conduction partial differential equation was calibrated and verified based on historical titanium alloy rolling data. For example, 200 representative historical batches of data were selected as input to the model, and the model's prediction results were compared with historical measured data. Historical measured data for titanium alloy rolling included, but was not limited to, historical surface temperatures of rolled pieces (recorded by an infrared thermometer), temperature probe data at key points during rolling (which could be obtained by drilling a hole in the center of the cut sample to implant a thermocouple), and the grain size and phase transition temperature of the final product. The parameters in the model, including heat transfer coefficient, surface emissivity, and the Taylor-Kuni coefficient β, were iteratively adjusted to minimize the root mean square error between the predicted and measured temperatures. For example, after calibration, under typical rolling conditions, the error in surface temperature prediction was controlled within 5°C, and the error in core temperature prediction was within 10°C. This calibrated model, i.e., the three-dimensional unsteady-state heat conduction partial differential equation with high confidence and prediction accuracy, was then used as the first temperature prediction model.

[0054] In step S2, based on real-time titanium alloy rolling data and a first temperature prediction model, the predicted overall temperature field of the rolled piece is obtained as the first temperature. In this embodiment, the finite element method is used to solve the first temperature prediction model, and the rolled piece is divided into three-dimensional spatial grid units. The temperature change of each grid unit can be predicted and calculated, and the complete three-dimensional temperature field distribution from the surface to the core is output, which is denoted as the first temperature.

[0055] In one possible implementation, based on real-time titanium alloy rolling data and a first temperature prediction model, the predicted overall temperature field of the rolled piece is obtained as the first temperature, including: Based on real-time titanium alloy rolling data and the first temperature prediction model, the predicted overall temperature distribution of the rolled piece is obtained. Based on the predicted overall temperature distribution of the rolled piece, the rolled piece is discretized into three-dimensional spatial grid elements, and the temperature of each predicted grid element is obtained by solving the problem using the finite element method. Based on the predicted temperature of each grid cell, the predicted overall temperature field of the rolled piece is obtained; the overall temperature field of the rolled piece includes the detectable surface temperature and the internal temperature field that cannot be directly measured. The overall temperature field of the rolled piece is predicted and used as the first temperature.

[0056] In the implementation of this application embodiment, during the titanium alloy rolling production process, based on the rolling data collected in real time and the first temperature prediction model, the predicted overall temperature field of the rolled piece is obtained as the first temperature. The acquisition of the first temperature is the basis for real-time temperature monitoring during the rolling process and a preliminary estimate of the internal thermal state of the rolled piece.

[0057] First, real-time data on titanium alloy rolling is collected. This data includes, but is not limited to, parameters such as mill speed (e.g., linear speed for each pass), rolling force, roll speed, roll gap, workpiece inlet temperature, current workpiece geometry, and ambient temperature. The real-time collected parameters are used as input to update and calculate the deformation heat, rolling speed, roll gap, inlet temperature, and workpiece geometry in the boundary conditions, and then input into the first temperature prediction model.

[0058] The workpiece is then discretized into three-dimensional spatial mesh elements, for example, using pre-generated and stored structured or unstructured meshes, or meshes adjusted according to the real-time geometry of the workpiece. For example, a multi-layered hexahedral mesh can be used for slab rolling, and the mesh density and element shape can be adjusted according to the reduction during rolling.

[0059] The first temperature prediction model is solved using the finite element method, with iterative calculations performed at a set small time step, such as 20 to 50 milliseconds. The selection of the time step is based on a balance between the response speed of thermal phenomena during rolling and computational resources, ensuring that transient temperature changes can be captured while meeting real-time requirements. Within each time step, the instantaneous temperature value of each mesh element is calculated and output, thereby obtaining the predicted overall temperature field of the rolled piece.

[0060] The overall temperature field is a complete three-dimensional temperature distribution data matrix containing data from the surface to the core. For example, for a 100mm thick slab, 200 nodes are set along its thickness, each with a corresponding temperature value. This transient three-dimensional temperature distribution data matrix is ​​the first temperature. The first temperature explicitly includes the directly measurable predicted surface temperature and the internal temperature information that cannot be directly measured, and its output frequency is consistent with the solution time step.

[0061] In step S3, the measured surface temperature of the rolled piece is obtained based on real-time titanium alloy rolling data, and used as a second temperature. One possible implementation includes: Real-time titanium alloy rolling data is collected from monitoring points installed on the titanium alloy rolling production line. The real-time titanium alloy rolling data includes: the surface temperature of the rolled piece, equipment operating parameters, and geometric parameters of the rolled piece at various key points on the rolling production line. Real-time titanium alloy rolling data is collected by monitoring points set up on the titanium alloy rolling production line; Based on real-time titanium alloy rolling data, the surface temperature of the rolled piece is obtained at the following locations: the box furnace exit, the roughing mill after rolling, the online induction furnace exit before the rolled piece enters the finishing mill, and the finished product collection device. The measured surface temperature of the rolled piece is obtained based on at least one of the surface temperatures of the rolled piece, and is used as the second temperature.

[0062] In the implementation of this application embodiment, the surface temperature of the rolled piece is collected based on the monitoring points installed on the titanium alloy rolling production line. The deployment position of the temperature measuring instrument is designed to cover the key thermal state monitoring points throughout the rolling process, which are defined as the second temperature. The second temperature is the key measured data for physical model calibration, and it directly reflects the true surface state of the rolled piece after heat exchange with the external environment.

[0063] Obtaining the second temperature involves the following specific operations: The titanium alloy rolling production line is configured to transport rolled parts via roller conveyors and sequentially connect multiple functionally defined process sections to form a complete production process. In this embodiment, the overall layout process of the titanium alloy rolling production line is as follows: box furnace (or induction furnace) heating → roller conveyor conveying → roughing mill rolling → flying shear head → online induction furnace reheating → finishing mill rolling → online cooling → finished product collection.

[0064] The components include: a box furnace for uniformly heating slabs to the initial rolling temperature; a roller conveyor for transporting rolled pieces; a roughing mill for initial diameter reduction; a flying shear for shearing defects at the head or tail of the rolled pieces; an online induction heating furnace for reheating the rolled pieces between passes to compensate for heat loss; a finishing mill for final diameter reduction and shaping; an online cooling section for controlling the post-rolling cooling rate to achieve the target metallographic structure, such as laminar flow cooling or atomized cooling; and a coiler or cutting device for collecting and coiling finished rolled pieces.

[0065] Multiple high-performance non-contact infrared thermometers are installed at key monitoring points in the titanium alloy rolling production line to ensure coverage of critical thermal state monitoring points throughout the rolling process and to provide high-frequency, high-precision surface temperature data. Key monitoring points include at least: The furnace outlet, located approximately a preset distance from the slab in the furnace, is used to measure the initial rolling temperature of the slab.

[0066] Before the roughing mill is fed, for example at a preset distance from the inlet of the roughing mill, the temperature of the workpiece before it enters the roughing mill is obtained.

[0067] The temperature after the roughing mill discharges, for example at a predetermined distance from the roughing mill outlet, is used to assess the deformation heat and cooling effect of the roughing mill.

[0068] Placed in front of the flying shear head, for example at a preset distance from the flying shear head, to assist in shearing strategy and temperature assessment.

[0069] Before feeding into the online induction furnace, for example at a preset distance from the furnace inlet, the temperature of the rolled piece before entering the induction heating is assessed, serving as the basis for determining the supplementary heating power.

[0070] After the material is discharged from the induction furnace, such as at a preset distance from the induction furnace outlet, it is used to evaluate the temperature rise effect after induction heating.

[0071] Before the finishing mill feeds the material, for example at a preset distance from the mill inlet, the final temperature of the workpiece before it enters the finishing mill is obtained, which is of utmost importance to the quality of the finishing mill.

[0072] After the finishing mill exits, for example at a predetermined distance from the finishing mill exit, the initial temperature is used to assess the heat of finishing deformation and the initial temperature of post-rolling cooling.

[0073] Located in the middle of the online cooling section, it is used to monitor temperature changes during the cooling process.

[0074] The infrared thermometer employs a multi-wavelength measurement principle to automatically compensate for changes in the emissivity of the rolled piece surface, thereby significantly improving measurement accuracy. Its measurement range is typically from 500℃ to 1400℃. Real-time titanium alloy rolling data also includes equipment operating parameters acquired synchronously with temperature measurements, such as mill rolling speed, rolling force, roll gap, induction furnace power, reheating time, and cooling water flow rate, as well as the geometric parameters of the rolled piece measured in real-time by the diameter gauge, such as the width and thickness of the rolled piece. All of this data is synchronized via timestamps.

[0075] Finally, the real-time measured surface temperature data of the rolled piece obtained by the non-contact infrared thermometer at the above-mentioned key monitoring points, together with its corresponding timestamp and location information, is defined as the second temperature.

[0076] In a specific embodiment, for the TC4 titanium alloy rolling production line, an infrared thermometer deployed 2 meters from the outlet of the roughing mill provides real-time feedback on the surface temperature of the rolled piece. For example, if the surface temperature is measured to be 885°C at a certain moment, the surface temperature is measured to be 930°C at 1.5 meters from the outlet of the online induction furnace. These data are synchronized with the timestamp of the rolled piece at the corresponding location so that they can be accurately compared with the first temperature later.

[0077] In step S4, the calibrated predicted temperature distribution inside the rolled piece is obtained based on the first temperature and the second temperature, and is used as the third temperature. The data of the second temperature will be used as the actual observation value for real-time calibration of the first temperature prediction model. For example, the real-time parameters before feeding into the roughing mill are obtained through the first temperature prediction model to obtain the first temperature at a preset distance from the exit of the roughing mill. The second temperature at this time is compared with the temperature at the corresponding location to correct the first temperature prediction model.

[0078] One possible implementation includes: Based on the real-time measured surface temperature of the rolled piece obtained from the second temperature, the predicted surface temperature of the rolled piece output at the corresponding position in the first temperature distribution is obtained. The observation equation is obtained by comparing the measured temperature of the rolled surface with the predicted temperature of the rolled surface. Based on the observation equation, an extended Kalman filter algorithm is used for iterative calibration to obtain a calibrated prediction model for the rolled piece. The iterative calibration based on the observation equation and the extended Kalman filter algorithm includes: using the second temperature as the observation vector, and using the core temperature, surface temperature, and heat flux density in the first temperature to form a state vector, updating the covariance matrix of process noise and observation noise to calibrate the prediction model of the first temperature, and performing iterative calibration to achieve a higher accuracy estimate of the internal temperature field.

[0079] Based on the calibrated prediction model for the rolled piece, the predicted internal temperature distribution of the rolled piece is obtained.

[0080] In the implementation of this application, the measured surface temperature is used to correct the temperature field predicted based on the physical model in real time, so as to overcome the cumulative error that may exist in the pure model prediction. The calibration process is an iterative optimization process, thereby achieving a higher accuracy estimate of the internal temperature field.

[0081] This invention observes a first temperature and a second temperature to obtain a calibrated predicted internal temperature field of the rolled piece, which is defined as a third temperature. The observation and calibration process aims to correct the prediction bias of the physical model using measured surface temperatures, thereby providing more accurate information on the internal temperature field. The specific operations for observing and obtaining the third temperature include: First, based on the real-time measured surface temperature data of the rolled piece included in the second temperature measurement, the predicted surface temperature of the rolled piece corresponding to the location and time of the temperature measuring instrument is extracted from the overall temperature distribution of the rolled piece output by the first temperature prediction model. For example, if the second temperature is collected 2 meters from the exit of the roughing mill, its timestamp is... Then the first temperature prediction model will output at that moment. The predicted temperature value of the rolled surface at this location. Then, based on the second temperature measured on the surface of the rolled piece... Predicted surface temperature of the rolled piece at the corresponding position in the first temperature range By comparing and constructing the observation equation, the difference between the predicted model's state and the actual observed values ​​was established. This difference mainly stems from factors such as model assumptions, parameter uncertainties, simplified boundary conditions, and measurement noise, and is expressed as follows:

[0082] in, It is the observation vector at time k, i.e., the second temperature of the measured surface temperature. ; The state vector to be estimated within the first temperature prediction model is composed of the core temperature of the rolled piece. Surface temperature gradient and heat flux density These parameters directly affect the accuracy of the internal temperature field of the rolled piece; The observation matrix describes the mapping relationship from the state vector to the observation vector, for example... It can be a linearized matrix that maps the core temperature, surface temperature gradient, and other states to the surface temperature; To observe noise.

[0083] Next, this invention iteratively calibrates the first temperature prediction model using an extended Kalman filter. The extended Kalman filter is a state estimation method for nonlinear systems. Its core idea is to use a linear Kalman filter to perform local linearization near the current estimation point, using the second temperature as the observation vector, and constructing the state vector from the parameters representing the thermal state of the rolled piece in the first temperature prediction model. Including core temperature of rolled piece Surface temperature gradient and heat flux density Parameters, for example, for a 50mm thick rolled piece, the state vector can include the temperature at the core center point, the temperature gradient of each layer from the surface to the center along the thickness direction, and the heat transfer coefficient between the surface and the environment.

[0084] Based on the observation equations, an iterative calibration is performed using an extended Kalman filter to obtain a calibrated prediction model for the rolled piece. The extended Kalman filter algorithm includes prediction and updating in the rolling process.

[0085] In the prediction step, the nonlinear state transition function of the first temperature prediction model is used. To predict the state vector at the next time step and its covariance matrix State transition function It describes the physical process of the evolution of the internal thermal state of the rolled piece over time, and its input This includes real-time rolling parameters such as rolling speed, reduction, and heat source power. In the update step, the second temperature is used as a real-time observation vector. , with predicted surface temperature Compare and calculate the observed residuals. .

[0086] Then, by calculating the Jacobian matrix, the state transition function and the observation function are linearized, and the process noise covariance matrix Q and the observation noise covariance matrix R are updated to correct the state vector. The update of the process noise covariance matrix Q considers the model structure error and the small fluctuations of unmodeled material parameters. Its initial value can be statistically set based on the model prediction error in historical data and adjusted during operation using maximum likelihood estimation. The update of the observation noise covariance matrix R is adjusted in real time based on the measurement accuracy, calibration status, and environmental interference of the infrared thermometer.

[0087] Through iterative calibration, the extended Kalman filter algorithm corrects the first temperature prediction model. Its key lies in utilizing surface temperature, an easily measurable physical quantity, to indirectly correct the prediction bias of the internal temperature field. This allows for real-time estimation of the internal temperature field of the rolled piece with high confidence, particularly accurate estimation of the core temperature. Finally, based on the calibrated prediction model, the calibrated predicted internal temperature field of the rolled piece is obtained. This predicted temperature field contains temperature information from all discrete grid points within the rolled piece and can estimate the temperature evolution trend over several rolling passes, such as the temperature change prediction over the next 10 to 30 seconds. By introducing a second temperature as a correction parameter, the prediction error is reduced by more than 50% compared to the uncalibrated model. This calibrated predicted internal temperature field of the rolled piece is then used as the third temperature.

[0088] In step S5, the control parameters during the rolling process are adjusted according to the third temperature. One possible implementation includes: Based on the third temperature, determine whether the third temperature of the core of the rolled piece is within the preset target processing window; If the third temperature of the core of the rolled piece is not within the target processing window, there is a risk of overheating or underheating. When there is a risk of overheating or underheating, based on the preset temperature-speed-power mapping rules, the specific adjustment amount of the mill speed and heating power is obtained according to the third temperature; Based on the specific adjustment amount, smoothing is performed to obtain control parameters for controlling rolling speed and heating power.

[0089] When implementing the embodiments of this application, it is determined whether the third temperature of the core of the rolled piece is within the target processing window. If there is a risk of overheating or underheating, the third temperature is converted into a specific adjustment amount of the mill speed and heating power according to the preset temperature-speed-power mapping rule, thereby obtaining the control command corresponding to the control parameters. This is achieved by real-time monitoring of the highest and lowest core temperatures of the rolled piece at the third temperature and comparing them with preset upper and lower temperature thresholds. When the absolute value of the core temperature deviating from the target window is greater than the preset allowable deviation, such as ±5℃, the control logic is triggered to actively adjust the parameters of the rolling production line to ensure that the temperature of the rolled piece is always maintained within the optimal plasticity processing window.

[0090] In the implementation of this application embodiment, firstly, based on the highest and lowest core temperatures of the rolled piece revealed by the third temperature, it is determined whether the temperature falls within a preset target processing window. The target processing window is set for different grades of titanium alloys. For example, for aerospace-grade TC4 titanium alloy, its optimal plastic processing temperature range has been determined through extensive experimentation to be 800℃ to 1000℃. Within this range, TC4 alloy exhibits excellent plasticity, moderate deformation resistance, and minimal grain growth, resulting in optimal mechanical properties and microstructure uniformity. Risk assessment is achieved by comparing the highest and lowest core temperatures of the rolled piece in the third temperature range with preset upper and lower temperature thresholds.

[0091] When the absolute value of the core temperature deviating from the target window exceeds a preset tolerance, such as ±5℃, the control logic is triggered. For example, if the target core temperature is 900℃ and the tolerance is ±5℃, then the control logic will be triggered when the core temperature is below 895℃ or above 905℃.

[0092] If the core temperature of the rolled piece is not within the target processing window, it is determined that there is a risk of overheating or underheating. Specifically, if the core temperature is below 800°C, there is a risk of underheating. At this time, the plasticity of the titanium alloy decreases, the deformation resistance increases sharply, and the rolling force may be too large. For example, it may exceed 80% of the mill's design load, that is, reach a force of 1800 tons for rolling, while the rated load is 2200 tons. This leads to accelerated equipment wear and may even cause the rolled piece to crack.

[0093] If the core temperature exceeds 1000℃, there is a risk of overheating, which may lead to excessive grain growth and an irreversible "core burning phenomenon," thereby severely degrading the mechanical properties of the final product. For example, the average grain size may exceed 50 micrometers, resulting in a decrease in tensile strength of more than 10%.

[0094] When there is a risk of overheating or underheating, based on a preset temperature-speed-power mapping rule, specific adjustments to the mill speed and heating power are obtained according to the real-time core temperature deviation reflected by the third temperature. The temperature-speed-power mapping rule is predetermined based on historical rolling data from a titanium alloy thermodynamics and deformation behavior database built in the preliminary preparation stage. This rule can be used to convert temperature deviations into specific control quantities using a series of piecewise linear functions, lookup table methods, or fuzzy logic-based control strategies.

[0095] In one possible implementation, when there is a risk of overheating or underheating, based on a preset temperature-speed-power mapping rule and according to a third temperature, specific adjustments to the mill speed and heating power are obtained, including: When the core temperature of the workpiece is lower than the preset lower limit when it starts rolling, the rolling speed of the roughing mill is reduced. When the core temperature is higher than the upper limit before entering the finishing mill, the rolling speed of the finishing mill is increased, and at the same time, the power of the online induction heating furnace is reduced. When implementing the embodiments of this application, if there is a risk of overheating or underheating, there are two situations: first, the temperature of the rolled piece is insufficient before entering the roughing mill when the rolling begins; second, the temperature of the rolled piece is too high before entering the finishing mill.

[0096] When the core temperature of the workpiece is lower than the set lower limit, for example, below 800°C, the rolling speed of the roughing mill is adjusted.

[0097] Based on core temperature deviation To obtain the adjustment amount ;in: Adjustment amount ,in The speed adjustment coefficient is calculated based on historical rolling data, and α is the exponential decay coefficient. The lower the core temperature, the more... For negative values, the larger the absolute value, When the value is negative, the rolling speed is reduced to increase the residence time of the rolled piece in the furnace or outside the deformation zone, thereby causing the temperature to rise.

[0098] For example, if the core temperature is 790°C, which is 10°C lower than the target lower limit of 800°C, then , At this point, the rolling speed will be reduced, allowing the workpiece more time to absorb heat.

[0099] When the core temperature of the workpiece before entering the finishing mill exceeds the set upper limit, such as above 1000°C, the finishing mill speed is increased, and the heating power of the online induction heating furnace is reduced at the same time.

[0100] Based on core temperature deviation To obtain the finishing mill speed adjustment amount and induction furnace power adjustment Among them: finishing mill speed adjustment amount ,in β is the speed regulation coefficient, and β is the exponential growth coefficient. The higher the core temperature, the more... For positive values, the larger the absolute value, the better. If the value is positive, the finishing rolling speed is increased to reduce the residence time of the rolled piece in the high-temperature zone and reduce heat accumulation.

[0101] For example, if the core temperature is 1010℃, which is 10℃ higher than the target upper limit of 1000℃, then , .

[0102] at the same time ,in γ is the power regulation coefficient, and γ is the linear regulation coefficient. When the core temperature exceeds the upper limit... If the value is negative, the induction furnace power is reduced to decrease external heat input. For example, under the above-mentioned 10°C overheating condition, the induction furnace power will be reduced by 100 × 10 = 1000 kW.

[0103] Finally, the obtained adjustment values ​​are smoothed to avoid frequent fluctuations in control commands, ensure the smooth operation of the rolling mill, and use them as control parameters. Then, the control commands corresponding to the control parameters are sent to the box furnace, roughing mill, finishing mill, and online induction heating furnace. By controlling the rolling speed and heating power, the internal temperature field of the rolled piece is actively regulated.

[0104] In this embodiment, step S6 is also included.

[0105] Step S6: Based on the control parameters obtained in step S5, generate corresponding control commands and send them to the titanium alloy rolling production line. At the same time, collect and store rolling batch number, material grade, chemical composition, second temperature at each monitoring point, rolling mill speed, induction furnace power, reheating time, and online diameter measuring instrument detection results, and store them in the titanium alloy thermodynamics and deformation behavior database. Simultaneously, obtain the fourth temperature prediction model by performing time-series analysis on historical data.

[0106] One possible implementation also includes: Based on the database of thermodynamics and deformation behavior of titanium alloys, a fourth temperature prediction model is obtained using a time convolutional network. Based on the current control parameters, the second temperature, the third temperature of the current pass, the rolling speed of the previous pass, and the heating power, the predicted core temperature for the next pass is obtained based on the fourth temperature prediction model. Based on the predicted core temperature for the next pass, the heating power of the box furnace and the preset roughing speed are determined before the start of the next roughing pass. Based on the predicted core temperature for the next pass and the predicted third temperature after the next roughing pass, the online induction furnace reheating power and the preset roughing rolling speed are obtained before finishing rolling.

[0107] In the implementation of this application embodiment, historical data is accumulated and correlation analysis is performed to extract the implicit patterns between temperature fluctuations and equipment parameters. The inputs include: the third temperature of the current pass, the rolling speed of the previous three passes, the deformation amount, the induction furnace power, and the material grade code. Using a time convolutional network, the output is the predicted value of the core temperature for the next pass. Model training uses a sliding window, and incremental training is triggered once every 100 rolling batches are completed. The updated model is used as the fourth temperature prediction model.

[0108] Preferably, it also includes feedforward control: based on the fourth temperature prediction model, the rolling speed curve is preset according to the initial heating temperature and material characteristics before the start of rough rolling; before finishing rolling, the induction furnace heating power is adjusted according to the third temperature prediction value after rough rolling to achieve active suppression of temperature fluctuations.

[0109] Preferably, it also includes a performance feedback mechanism: the mechanical property test results of the final rolled material are transmitted back to the process database to establish a mapping relationship between temperature control accuracy and product performance; when the performance of a batch of products deviates from the standard, its third temperature control trajectory is automatically traced and the parameters of the fourth temperature prediction model are targetedly corrected.

[0110] In the implementation of this application, key production data such as rolling batch number, material grade, chemical composition, second temperature obtained at each monitoring point, rolling speed of roughing mill and finishing mill, power of online induction heating furnace, reheating time, and detection results of online diameter measuring instrument are collected and stored in real time. All data are timestamped and stored in a database, which enables comprehensive quality traceability of the production history of each batch of rolled material.

[0111] At the same time, the mechanical property test results of the final rolled material, such as tensile strength, yield strength, elongation, reduction of area and impact toughness, are immediately transmitted back to the process database after the product inspection is completed, so as to establish a mapping relationship between temperature control accuracy and the final mechanical properties of the product.

[0112] Secondly, correlation analysis is performed on historical data accumulated in the structured process database to extract the implicit patterns between temperature fluctuations and equipment parameters. For example, the linear relationship between rolling speed and core temperature drop rate is identified by calculating the Pearson correlation coefficient, and the nonlinear dependence between roll coolant flow rate and surface heat flux density is analyzed by mutual information entropy analysis. Alternatively, principal component analysis is used to reduce dimensionality and identify key input features affecting core temperature changes and their nonlinear relationships. For example, it is found that the fluctuation of workpiece edge temperature is significantly correlated with roll wear during the rolling process.

[0113] A self-supervised learning model is constructed using a temporal convolutional network as the fourth temperature prediction model. The input features of the temporal convolutional network are configured, including: the third temperature obtained in the current pass, the rolling speed of the previous three rolling passes, the deformation of the rolled piece, the power of the online induction furnace, and the material grade encoding. The deformation of the rolled piece is the strain calculated by the inlet and outlet cross-sectional dimensions, and the material grade encoding is, for example, encoding TC4 as [1,0,0] and TC11 as [0,1,0], etc.

[0114] The fourth temperature prediction model outputs the predicted core temperature of the rolled piece for the next pass, with prediction accuracy superior to traditional recurrent neural network (RNN) models and a prediction lead time of up to 30 seconds. In one specific embodiment, the fourth temperature prediction model is configured as a three-layer causal dilated convolutional layer, each containing 64 filters with a filter size of 3. The dilation rates are 1, 2, and 4, respectively. The model input includes a 20-dimensional feature vector containing the current third temperature (core), the average rolling speed of the previous three passes, cumulative strain, and the average power of the online induction furnace, as well as a material type. The output is the predicted core temperature before entering the rolls for the next rolling pass.

[0115] Furthermore, based on the fourth temperature prediction model, this invention implements a feedforward control strategy. Before roughing begins, based on the initial heating temperature and material properties of the workpiece, and the prediction of the temperature evolution trend of the workpiece during roughing by the fourth temperature prediction model (e.g., predicting a core temperature drop of 100°C after roughing), a roughing speed curve is preset, thereby enabling proactive control in the early stages of rolling deformation and effectively suppressing the initial temperature fluctuations. For example, if the temperature drop during roughing is predicted to be too rapid, a lower initial roughing speed is preset to increase the deformation time and utilize deformation heat to offset some of the heat loss. Therefore, before finishing begins, based on the third temperature prediction value obtained after roughing and the prediction of temperature changes during finishing by the fourth temperature prediction model, the supplementary heating power of the online induction heating furnace is adjusted in advance to achieve proactive suppression of temperature fluctuations, rather than a passive response. For example, if the core temperature is predicted to drop to 820°C during finishing, lower than the target of 850°C, the induction furnace power is increased to 90% of the rated power in advance to ensure that the core temperature reaches 900°C when the workpiece enters the finishing mill.

[0116] Furthermore, this invention integrates a performance feedback mechanism for continuous optimization of temperature control strategies. Specifically, when the final mechanical property test results of a batch of products deviate from the preset standard, the corresponding third temperature control trajectory during the production process is traced to analyze in which rolling pass or stage the core temperature anomaly occurred. For example, if the product elongation is found to be low, tracing back reveals that in the penultimate finishing pass, the core temperature deviated from the lower limit of the target range by 15°C, resulting in insufficient recrystallization. Then, by correlating the abnormal temperature trajectory with the product performance deviation, targeted corrections to the parameters of the fourth temperature prediction model can be triggered. This improves the model's ability to predict temperature anomalies under specific materials or production conditions, ultimately optimizing the temperature-speed-power mapping rules, fundamentally improving product quality, and forming a closed-loop temperature detection and control system.

[0117] Example 3 This is the third embodiment of the present invention. Based on embodiments 1 and 2, this embodiment provides a temperature monitoring and control system for the titanium alloy rolling process. Please refer to the relevant documentation. Figure 2 The above is a schematic diagram of the system structure of an embodiment of the present invention, including a first temperature prediction model module, a first temperature prediction module, a second temperature acquisition module, a third temperature calibration module, and a control parameter acquisition module that are connected in sequence by electricity. The first temperature prediction model module is configured to obtain the first temperature prediction model based on the heat conduction equation, according to the database of titanium alloy thermodynamics and deformation behavior. The first temperature prediction module is configured to: obtain the predicted overall temperature field of the rolled piece based on the first temperature prediction model according to the real-time titanium alloy rolling data, and use it as the first temperature; The second temperature acquisition module is configured to obtain the measured surface temperature of the rolled piece based on real-time titanium alloy rolling data, and use it as the second temperature. The third temperature calibration module is configured to: obtain the calibrated predicted temperature distribution inside the rolled piece based on the first temperature and the second temperature, and use it as the third temperature; The control parameter acquisition module is configured to adjust the control parameters during the rolling process based on a third temperature.

[0118] This embodiment also includes an execution and feedback module, which is configured to issue instructions, and then there is a feedforward mechanism and a feedback mechanism.

[0119] In the implementation of this application embodiment, the execution and feedback module refers to the unit used to convert control parameters into equipment execution instructions and monitor the execution effect in real time. Specifically, it can be directly implemented by an industrial control computer and a programmable logic controller, and data interaction is performed through the rolling mill speed regulation system and the induction furnace power regulation device.

[0120] We introduced a feedforward mechanism, which is a strategy to adjust control parameters in advance based on the fourth temperature prediction model. Specifically, the fourth temperature prediction model can learn from historical data, and then make predictions based on the current temperature of the titanium alloy rolled piece. The output is then used to generate a pre-compensation signal, which is superimposed on the current control command. For example, the heating power can be adjusted in advance based on the predicted temperature before the start of the next rolling pass.

[0121] Then, the feedback mechanism refers to the strategy of dynamically correcting control parameters based on real-time monitoring data. Specifically, it can be achieved by establishing a closed-loop adjustment algorithm between temperature deviation and control quantity, such as dynamically adjusting the rolling speed based on the difference between the current pass calibration temperature and the target temperature.

[0122] In the implementation of this application embodiment, after receiving the adjustment amount output by the control parameter acquisition module, the execution and feedback module first inputs the predicted temperature data of the next pass into the pre-compensation algorithm through a feedforward mechanism to generate pre-adjustment instructions for rolling speed and heating power.

[0123] For example, when it is predicted that the core temperature of the rolled piece after the next roughing pass may be lower than the target range, the feedforward mechanism will increase the initial heating power of the box furnace in advance.

[0124] Then, the feedback mechanism adjusts the rolling speed command based on the deviation between the actual temperature calibration result of the current pass and the target temperature. For example, if the core temperature exceeds the upper limit during finishing rolling, the feedback mechanism will increase the mill speed and reduce the induction furnace power in real time.

[0125] The feedforward and feedback signals are weighted and fused to form the final control command, which is then sent to the rolling mill drive system and heating device for execution.

[0126] This application effectively solves the problem of internal temperature changes lagging behind surface temperature in rolled pieces, significantly improving the controllability of core temperature. It also introduces feedforward and feedback mechanisms. The feedforward mechanism, based on a fourth temperature prediction model learning from historical data, adjusts the process in advance to suppress temperature fluctuations caused by delayed heat conduction. The feedback mechanism dynamically corrects control values ​​through real-time calibration data, eliminating the influence of model errors and external interference. The system of this application maintains the overall temperature of the rolled piece more stably within the target processing window, avoiding core burning or low-temperature cracking caused by control lag, while reducing unnecessary rolling speed sacrifices and improving production efficiency and product quality consistency.

[0127] Example 4 The fourth embodiment of the present invention differs from the previous embodiments in that: Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0128] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices or units, or may be electrical, mechanical or other forms of connection.

[0129] The units described as separate components may or may not be physically separate. As will be apparent to those skilled in the art, the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0130] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0131] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or grid device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0132] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for monitoring and controlling the temperature during the rolling process of titanium alloys, characterized in that, include: Based on the database of thermodynamics and deformation behavior of titanium alloys, and the heat conduction equation, a first temperature prediction model is obtained. Based on real-time titanium alloy rolling data and the first temperature prediction model, the predicted overall temperature field of the rolled piece is obtained as the first temperature; based on real-time titanium alloy rolling data, the measured surface temperature of the rolled piece is obtained as the second temperature. Based on the first and second temperatures, the overall temperature field of the rolled piece for calibration is obtained and used as the third temperature. The control parameters during the rolling process are adjusted based on the third temperature.

2. The method for monitoring and controlling the temperature during the rolling process of titanium alloys according to claim 1, characterized in that, The first temperature prediction model, established based on the database of titanium alloy thermodynamics and deformation behavior and the heat conduction equation, includes: Based on the thermophysical property parameters of the titanium alloy thermodynamic and deformation behavior database, density, specific heat capacity, and thermal conductivity are obtained by looking up tables. Based on the mechanical property parameters of the titanium alloy thermodynamics and deformation behavior database, and based on the deformation behavior during the rolling process, the plastic deformation work is converted into heat energy to obtain the rate of heat generation from plastic deformation inside the rolled piece, which is used as the internal heat source item of the rolled piece. Based on the density, specific heat capacity, thermal conductivity of titanium alloys and the internal heat source term of the rolled piece, a three-dimensional unsteady-state partial differential equation for heat conduction is obtained based on the heat conduction equation. Based on historical titanium alloy rolling data, boundary conditions were set to obtain a verified three-dimensional unsteady-state heat conduction partial differential equation, which serves as the first temperature prediction model.

3. A method for monitoring and controlling the temperature during the rolling process of titanium alloys according to claim 1 or 2, characterized in that, The step of obtaining the predicted overall temperature field of the rolled piece, based on the first temperature prediction model and real-time titanium alloy rolling data, as the first temperature, includes: Based on real-time titanium alloy rolling data and the first temperature prediction model, the predicted overall temperature distribution of the rolled piece is obtained. Based on the predicted overall temperature distribution of the rolled piece, the rolled piece is discretized into three-dimensional spatial grid elements, and the temperature of each predicted grid element is obtained by solving the problem using the finite element method. Based on the predicted temperature of each grid cell, the predicted overall temperature field of the rolled piece is obtained; the overall temperature field of the rolled piece includes the detectable surface temperature and the internal temperature field that cannot be directly measured. The overall temperature field of the rolled piece is predicted and used as the first temperature.

4. The method for monitoring and controlling the temperature during the rolling process of titanium alloys according to claim 1, characterized in that, The step of obtaining the measured surface temperature of the rolled piece based on real-time titanium alloy rolling data, as the second temperature, includes: Real-time titanium alloy rolling data is collected by monitoring points set up on the titanium alloy rolling production line; Based on the real-time titanium alloy rolling data, the surface temperature of the rolled piece is obtained at the following locations: the box furnace exit, the roughing mill after rolling, the online induction furnace exit before the rolled piece enters the finishing mill, and the finished product collection device. The measured surface temperature of the rolled piece is obtained based on at least one of the surface temperatures of the rolled piece, and is used as the second temperature.

5. The method for monitoring and controlling the temperature during titanium alloy rolling process according to claim 1, characterized in that, The step of obtaining the calibrated predicted internal temperature distribution of the rolled piece, based on the first and second temperatures, as the third temperature, includes: Based on the real-time measured surface temperature of the rolled piece obtained from the second temperature, the predicted surface temperature of the rolled piece output at the corresponding position in the first temperature distribution is obtained. The observation equation is obtained by comparing the measured temperature of the rolled surface with the predicted temperature of the rolled surface. According to the observation equation, iterative calibration is performed using the extended Kalman filter algorithm to obtain a calibrated prediction model for the rolled piece. The iterative calibration using the extended Kalman filter algorithm according to the observation equation includes: using the second temperature as the observation vector, constructing a state vector from the core temperature, surface temperature, and heat flux density in the first temperature, calibrating the first temperature prediction model by updating the covariance matrix of process noise and observation noise, performing iterative calibration, obtaining a calibrated prediction model for the rolled piece, and estimating the internal temperature field. Based on the calibrated prediction model for the rolled piece, the predicted internal temperature distribution of the rolled piece is obtained. The calibrated predicted temperature field inside the rolled piece is used as the third temperature.

6. The method for monitoring and controlling the temperature during titanium alloy rolling process according to claim 1, characterized in that, The adjustment of control parameters during the rolling process based on the third temperature includes: Based on the third temperature, determine whether the third temperature of the core of the rolled piece is within the preset target processing window; If the third temperature of the core of the rolled piece is not within the target processing window, there is a risk of overheating or underheating. When there is a risk of overheating or underheating, based on the preset temperature-speed-power mapping rules, the specific adjustment amount of the rolling mill speed and heating power is obtained according to the third temperature. Based on the specific adjustment amount, smoothing is performed to obtain control parameters for controlling rolling speed and heating power.

7. The method for monitoring and controlling the temperature during titanium alloy rolling process according to claim 6, characterized in that, When there is a risk of overheating or underheating, based on a preset temperature-speed-power mapping rule and according to the third temperature, specific adjustments to the mill speed and heating power are obtained, including: When the core temperature of the workpiece is lower than the preset lower limit when it starts rolling, the rolling speed of the roughing mill is reduced. When the core temperature exceeds the upper limit before entering the finishing mill, the rolling speed of the finishing mill is increased, and at the same time, the power of the online induction heating furnace is reduced.

8. The method for monitoring and controlling the temperature during the rolling process of titanium alloys according to claim 1, characterized in that, Also includes: Based on the aforementioned database of titanium alloy thermodynamics and deformation behavior, a fourth temperature prediction model is obtained using a time convolutional network. Based on the current control parameters, the second temperature, the third temperature of the current pass, the rolling speed, deformation, heating power and material code of the previous pass, the predicted core temperature of the next pass is obtained based on the fourth temperature prediction model. Based on the predicted core temperature for the next pass, the heating power of the box furnace and the preset roughing speed are determined before the start of the next roughing pass. Based on the predicted core temperature for the next pass and the predicted third temperature after the next roughing pass, the online induction furnace reheating power and the preset roughing rolling speed are obtained before finishing rolling.

9. A temperature monitoring and control system for titanium alloy rolling process, characterized in that, It includes a first temperature prediction model module, a first temperature prediction module, a second temperature acquisition module, a third temperature calibration module, and a control parameter acquisition module, all connected in sequence. The first temperature prediction model module is configured to: obtain a first temperature prediction model based on the heat conduction equation according to the database of thermodynamics and deformation behavior of titanium alloys; The first temperature prediction module is configured to: obtain the predicted overall temperature field of the rolled piece based on the first temperature prediction model according to real-time titanium alloy rolling data, and use it as the first temperature; The second temperature acquisition module is configured to: obtain the measured surface temperature of the rolled piece based on real-time titanium alloy rolling data, and use it as the second temperature; The third temperature calibration module is configured to: obtain the calibrated predicted temperature distribution inside the rolled piece based on the first temperature and the second temperature, and use it as the third temperature; The control parameter acquisition module is configured to adjust the control parameters during the rolling process based on the third temperature.

10. A computer system device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the method according to any one of claims 1 to 8.

Images