Warm upsetting-rolling process parameter collaborative detection optimization method for high-temperature alloy bolt

Abstract

The invention relates to the technical field of bolt manufacturing, in particular to a high-temperature alloy bolt warm upsetting-rolling process parameter collaborative detection optimization method. According to the method, the energy dissipation data is obtained by calculating the difference value between energy input and heat accumulation in the warm upsetting stage, the rolling starting torque reference is determined through the dissipation rate, direct constraint of the warm upsetting heating power state on the rolling initial load is established, the problem of cross-process parameter splitting is solved, and meanwhile, the energy dissipation rate is improved. The energy density evolution trend of a rolling shaping area is deeply analyzed, forming, stabilizing and attenuation stages are divided, response intervals of errors of energy gradient, torque, linear speed and the like are constructed, load characteristics in the material deformation process are accurately mastered, and finally, energy characteristics of warm upsetting and rolling are fused and input into a support vector machine model, so that the material deformation precision is improved. According to the method, key quality indexes such as the pitch diameter, the thread angle and the hardness of a thread are predicted, rolling parameters such as driving current and preloading force are reversely corrected according to prediction deviation, and closed-loop dynamic control over the forming quality is achieved.

Description

Co-detection and optimization method for high-temperature alloy bolt upsetting-rolling process parameters Technical Field

[0001] This invention relates to the field of bolt manufacturing technology, and in particular to a method for the coordinated detection and optimization of process parameters for hot upsetting and rolling of high-temperature alloy bolts. Background Technology

[0002] The field of bolt manufacturing technology mainly involves the design, production, and processing of bolts and related components, including material selection, forming, heat treatment, and surface treatment. Bolt manufacturing typically involves multiple stages, such as cold heading, hot heading, rolling, cutting, milling, and surface treatment. Among these, the collaborative detection and optimization method for hot heading-rolling process parameters of high-temperature alloy bolts refers to the method of detecting and adjusting process parameters such as billet heating temperature, forming force, and deformation rate during the hot heading stage, and roller linear speed, rolling pressure, and feed rate during the rolling stage.

[0003] Existing technologies detect and adjust bolt warm upsetting and rolling process parameters, but they fail to establish a physical relationship between the thermodynamic state of the warm upsetting stage and the mechanical parameters of the rolling stage. The energy dissipation during the warm upsetting process is not transmitted to the subsequent rolling process, resulting in a lack of accurate preliminary state basis for the initial setting of rolling parameters. Cooperative optimization of parameters is difficult to achieve. In addition, this method only stays at the level of parameter detection and lacks in-depth analysis of energy density changes and load fluctuations during the rolling process. It cannot observe the flow and deformation resistance of the material in the die tooth surface in real time. The adjustment strategy is often based on static settings and lacks predictive feedback on the final forming quality. It is difficult to make closed-loop corrections for indicators such as thread pitch diameter or hardness distribution, resulting in limited process control accuracy and difficulty in ensuring quality consistency. Summary of the Invention

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a collaborative detection and optimization method for high-temperature alloy bolt warm upsetting-rolling process parameters, comprising the following steps: S1: collecting data on the change of heat storage at the center and surface positions of the bolt blank during the warm upsetting stage over time, and data on the change of energy input of the main drive motor of the warm upsetting press over time, to construct an energy dissipation dataset; S2: calculating the starting torque reference value of the main drive motor of the thread rolling machine based on the energy dissipation dataset, collecting the rated speed range, transmission ratio range, and current output range of the main drive motor of the thread rolling machine as constraints for optimization, and selecting the rolling parameter instruction set; S 3: Determine the operating monitoring benchmark of the thread rolling machine through the rolling parameter instruction set, and perform error analysis based on the measured data of the thread rolling machine to form a rolling load feature set; S4: Analyze the energy density change of the thread forming contact area formed by the bolt blank surface and the thread rolling die tooth surface based on the rolling load feature set, and generate an energy stratification judgment result; S5: Combine the energy stratification judgment result and the energy dissipation dataset to form a forming parameter input set for the bolt temperature upsetting and rolling stages, predict the forming quality index of the bolt, correct the rolling parameter instruction set based on the prediction result, and output the bolt forming quality control instruction.

[0005] As a further aspect of the present invention, the energy dissipation dataset includes energy input variation curves, heat storage variation curves, and energy difference variation curves; the rolling parameter instruction set includes drive current setting value, roller preload force, and roller initial linear velocity; the rolling load feature set includes torque error variation results, linear velocity error variation results, and feed displacement error variation results; the energy stratification determination results include stratification information for the forming stage, stratification information for the stable stage, and stratification information for the decay stage; and the bolt forming quality control instructions specifically include bolt thread pitch diameter error, thread angle deviation, and Vickers hardness distribution.

[0006] As a further embodiment of the present invention, step S1 specifically comprises: S101: collecting temperature change data of the center and surface of the bolt blank during the upsetting stage over time, and collecting current and voltage change data of the main drive motor during the heating and upsetting process over time, obtaining the density and specific heat parameters of the blank, converting the temperature difference between the center and the surface with the density and specific heat parameters of the blank to obtain the heat storage change data over time, and converting the current and voltage data according to the time correspondence to obtain the energy input change data over time; S102: obtaining the energy input value in the energy input change data over time and the heat storage value in the heat storage change data over time on the same time axis, comparing them point by point, calculating the difference between the two at each time point, and obtaining the energy deviation time series quantity; S103: organizing the energy deviation time series quantity on the same time axis, plotting the deviation change curve of energy input and heat storage, integrating the deviation change curve with the corresponding time axis data, and combining it with the originally collected temperature, current and voltage data to establish an energy dissipation dataset.

[0007] As a further embodiment of the present invention, step S2 specifically comprises: S201: extracting the rate of change information of the central deviation change curve of the energy dissipation data set, and calculating the starting torque reference value of the main drive motor of the thread rolling machine based on the rate of change information; S202: collecting the rated speed range, transmission ratio range and current output range of the main drive motor of the thread rolling machine, and using the rated speed range, transmission ratio range and current output range together with the starting torque reference value as constraints, inputting them into the Bayesian optimization algorithm for optimization, and obtaining the optimization result of the rolling parameters; S203: based on the optimization result of the rolling parameters, obtaining the drive current setting value according to the mapping relationship between motor torque and current, obtaining the initial linear velocity of the roller according to the correspondence between the transmission ratio, the roller pitch circle diameter and the spindle speed, and obtaining the roller preload force according to the calibration curve of hydraulic or servo displacement and loading force, combining the drive current setting value, the roller preload force and the initial linear velocity of the roller, and outputting the rolling parameter instruction set.

[0008] As a further aspect of the present invention, the process of inputting the parameters to the Bayesian optimization algorithm for optimization and obtaining the rolling parameter optimization result specifically involves: setting the rated speed range, transmission ratio range, and current output range together as the parameter search space of the Bayesian optimization algorithm; setting the starting torque reference value as the torque constraint condition for the Bayesian optimization algorithm to perform optimization within the parameter search space; the Bayesian optimization algorithm iteratively optimizing within the parameter search space and according to the torque constraint condition until the optimization process converges or reaches the preset number of iterations; and taking the parameter combination corresponding to the convergence state as the rolling parameter optimization result.

[0009] As a further aspect of the present invention, step S3 specifically comprises: S301: determining the operating monitoring benchmark of the thread rolling machine through the rolling parameter instruction set, and collecting real-time operating data of the main drive motor and roller moving parts of the thread rolling machine, including torque response change data over time, roller linear speed change data, and feed displacement change data, and integrating them into the measured data of the thread rolling machine; S302: calculating the deviations of the torque response change data over time, roller linear speed change data, and feed displacement change data in the measured data of the thread rolling machine from the starting torque benchmark value, and obtaining the relationship between the errors of torque, linear speed, and feed displacement and time; S303: based on the relationship between the errors of torque, linear speed, and feed displacement and time, extracting the fluctuation amplitude of torque error, the instantaneous change rate of linear speed error, and the trajectory deviation of feed displacement error, mapping them into characteristic indicators characterizing the intensity of rolling load fluctuation and deformation resistance, and forming a rolling load feature set.

[0010] As a further embodiment of the present invention, step S4 specifically comprises: S401: Extracting the characteristic indicators representing the rolling load fluctuation intensity and deformation resistance from the rolling load feature set, determining the energy density change of the thread forming contact area formed by the bolt blank surface and the thread rolling die tooth surface, and obtaining the energy density change trend; S402: Based on the energy density change trend, dividing the thread forming contact area into forming stage, stabilizing stage and decay stage, and calculating the energy gradient change direction of the energy density curve in each stage, integrating the stage division with the corresponding energy gradient change direction to establish a forming stage and gradient set; S403: Calling the forming stage and gradient set, obtaining the stage division and energy gradient change direction, and combining the error change results of torque, linear velocity and feed displacement from the rolling load feature set, determining the corresponding relationship interval between energy distribution and rolling pressure, feed amount and linear velocity in each stage of the thread forming contact area, and generating energy stratification judgment results.

[0011] As a further aspect of the present invention, the process of determining the corresponding relationship interval between energy distribution and rolling pressure, feed rate, and linear velocity specifically involves: extracting the error change results of torque, linear velocity, and feed displacement corresponding to the rolling load feature set in each of the forming stage, stabilization stage, and decay stage; calling the energy gradient change direction corresponding to each stage in the shaping stage and gradient set; performing correlation analysis between the energy gradient change direction and the error change result of torque to determine the response interval of energy gradient to rolling pressure; performing correlation analysis between the energy gradient change direction and the error change result of feed displacement to determine the response interval of energy gradient to feed rate; performing correlation analysis between the energy gradient change direction and the error change result of linear velocity to determine the response interval of energy gradient to linear velocity; and combining the response intervals of energy gradient to rolling pressure, energy gradient to feed rate, and energy gradient to linear velocity to establish a corresponding relationship interval.

[0012] As a further embodiment of the present invention, step S5 specifically comprises: S501: Extracting the upsetting stage feature data from the energy dissipation dataset and the rolling stage feature data from the energy stratification determination result, and establishing a set of forming parameters for the bolt upsetting and rolling stages based on the processing sequence of the same bolt blank; S502: Inputting the set of forming parameters for the bolt upsetting and rolling stages into a support vector machine model, and predicting the bolt thread pitch diameter error, thread angle deviation, and Vickers hardness distribution based on the mapping relationship between the forming parameter input set and the corresponding forming quality, thereby obtaining the bolt forming quality index; S503: Calculating the deviation between the predicted index and the target index based on the bolt forming quality index, and correcting the drive current setting, roller preload force, and roller initial linear velocity in the rolling parameter instruction set based on the deviation, and outputting the bolt forming quality control instruction.

[0013] As a further aspect of the present invention, the process of correcting the drive current setting, roller preload force, and roller initial linear velocity in the rolling parameter instruction set based on the deviation amount specifically includes: using the radial deviation value corresponding to the bolt thread pitch diameter error in the deviation amount, determining the compensation amount for the radial extrusion force based on the deformation correspondence between the roller pressing depth and the pitch diameter dimension, and using the compensation amount to adjust the roller preload force in the rolling parameter instruction set by increment or decrement; using the angle deviation value corresponding to the tooth angle deviation in the deviation amount as the basis for evaluating the meshing and slipping state between the roller and the blank, calculating and optimizing the speed adjustment coefficient for the metal flow filling effect and synchronization, and updating the roller initial linear velocity in the rolling parameter instruction set; using the hardness deviation value corresponding to the Vickers hardness distribution in the deviation amount, calculating and adjusting the motor output power based on the correlation between the degree of material work hardening and plastic deformation energy consumption, controlling the current correction value for the cold work hardening effect, and resetting the drive current setting value in the rolling parameter instruction set.

[0014] Compared with the prior art, the advantages and positive effects of this invention are as follows: In this invention, energy dissipation data is obtained by calculating the difference between energy input and heat storage during the warm upsetting stage. The dissipation rate is used to determine the rolling start torque benchmark, establishing a direct constraint of the warm upsetting thermal state on the initial rolling load, thus solving the problem of parameter fragmentation across processes. At the same time, the energy density evolution trend of the rolling forming zone is analyzed in depth, dividing the forming, stabilization and decay stages, and constructing the response range of energy gradient with errors such as torque and linear velocity, accurately grasping the load characteristics during material deformation. Finally, the energy characteristic input support vector machine model of warm upsetting and rolling is integrated to predict key quality indicators such as thread pitch diameter, tooth angle and hardness, and the rolling parameters such as drive current and preload force are corrected in reverse according to the prediction deviation, realizing closed-loop dynamic control of forming quality. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 is a schematic diagram of the steps of the present invention; Figure 2 is a detailed schematic diagram of S1 of the present invention; Figure 3 is a detailed schematic diagram of S2 of the present invention; Figure 4 is a detailed schematic diagram of S3 of the present invention; Figure 5 is a detailed schematic diagram of S4 of the present invention; Figure 6 is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0019] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0020] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0021] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0022] Please refer to Figure 1. This embodiment of the invention provides a method for collaborative detection and optimization of process parameters for warm upsetting and rolling of high-temperature alloy bolts, including the following steps: S1: Collect data on the change of heat storage at the center and surface of the bolt blank during the warm upsetting stage over time, as well as data on the change of energy input of the main drive motor of the warm upsetting press over time, and construct an energy dissipation dataset; S2: Calculate the starting torque reference value of the main drive motor of the thread rolling machine based on the energy dissipation dataset, and collect the rated speed range, transmission ratio range, and current output range of the main drive motor of the thread rolling machine as constraints for optimization, and select the rolling parameter instruction set; S3: Determine the operating monitoring benchmark of the thread rolling machine through the rolling parameter instruction set, and perform error analysis based on the measured data of the thread rolling machine to form a rolling load feature set; S4: Analyze the energy density change of the thread plastic contact area formed by the bolt blank surface and the thread rolling die tooth surface based on the rolling load feature set, and generate energy stratification judgment results; S5: Combine the energy stratification judgment results and the energy dissipation dataset to form the forming parameter input set for the bolt temperature upsetting and rolling stages, predict the forming quality index of the bolt, correct the rolling parameter instruction set based on the prediction results, and output the bolt forming quality control instruction.

[0023] The energy dissipation dataset includes energy input change curves, heat storage change curves, and energy difference change curves. The rolling parameter instruction set includes drive current setpoints, roller preload force, and roller initial linear velocity. The rolling load feature set includes torque error change results, linear velocity error change results, and feed displacement error change results. The energy stratification determination results include stratification information for the forming stage, stratification information for the stable stage, and stratification information for the decay stage. The bolt forming quality control instructions specifically include bolt thread pitch diameter error, thread angle deviation, and Vickers hardness distribution.

[0024] Please refer to Figure 2. Step S1 is as follows: S101: Collect temperature change data of the center and surface of the bolt blank during the warm upsetting stage, and collect current and voltage change data of the main drive motor during heating and upsetting. Obtain the density and specific heat parameters of the blank. Combine the temperature difference between the center and the surface with the density and specific heat parameters of the blank to obtain the heat storage change data over time. Based on the time correspondence, convert the current and voltage data to obtain the energy input change data over time. The infrared thermal imager and the pre-embedded thermocouple are started synchronously to monitor the bolt blank at the warm upsetting station. The infrared thermal imager uses a set sampling frequency. The infrared radiation intensity of the outer surface of the billet is captured at a rate (e.g., 500 frames per second), and the surface temperature data is obtained by inversion calculation using Planck's blackbody radiation law. The pre-embedded thermocouples are not implanted in real-time, but rather embedded in the center of a test billet of the same specification through prior experiments. A heat transfer correlation model between the center temperature and the surface temperature during the entire warm upsetting process is established. In actual production, the established heat transfer correlation model is called, and the real-time surface temperature data is used as input to derive and calculate the temperature change over time at the center position during the warm upsetting stage. Simultaneously, a high-frequency Hall current is installed at the power input terminal of the main drive motor of the warm upsetting machine. The voltage sensor captures the instantaneous current and voltage values ​​of the main drive motor during the heating and upsetting process in real time; it retrieves the pre-entered material property database from the memory to read the material density constant (e.g., 7.85 g / cm³ for carbon steel) and specific heat capacity constant (e.g., 0.46 kJ / kg / °C) of the bolt blanks in the current batch; it then executes the enthalpy calculation logic, specifically: taking the weighted average of the temperature data at the center and the surface (e.g., 800°C at the center, 700°C at the surface, with a weighted average of 750°C) as the overall average temperature, and using the change in the overall average temperature... (e.g., 10 degrees Celsius) × material density constant (7.85 g / cm³) × specific heat capacity constant (0.46 kJ / kg / degree Celsius) × billet volume (e.g., 20 cm³) to calculate the change in heat storage inside the bolt billet over time; simultaneously, execute the power integration logic to multiply the instantaneous current value (e.g., 50 amps) and the instantaneous voltage value (e.g., 380 volts) at the same sampling moment (50 × 380 = 19000 watts) to obtain the instantaneous input power, and perform cumulative calculation on the time axis for the instantaneous input power (i.e., integrate the power-time curve) to obtain the change in energy input over time.

[0025] S102: Obtain the energy input value from the energy input changing over time data and the heat storage value from the heat storage changing over time data on the same time axis, and compare them point by point, calculating the difference between the two at each time point to obtain the energy deviation time series quantity; map the energy input changing over time data and the heat storage changing over time data to the same millisecond-level precision universal time axis to ensure the time synchronization of data points; traverse each discrete time point on the universal time axis, extract the energy input value corresponding to the current time point (e.g., 1500 joules) as the minuend, extract the heat storage value corresponding to the current time point (e.g., 1200 joules) as the subtrahend, and perform a numerical comparison (1500-1200). =300 Joules); During the calculation process, if the time points are not completely aligned, linear interpolation is used. For example, if data for a certain intermediate time point is needed, and the data for two adjacent time points before and after it are known, the value that the intermediate time point should have is calculated by using the proportional relationship between the values ​​of these two known data points and the corresponding time difference, and the value of the time point to be aligned is calculated. After the point-by-point calculation is completed, a numerical sequence containing the difference of all discrete time points is generated, which is the energy deviation time series. The energy deviation time series reflects the dynamic fluctuation of the loss part (such as mechanical friction heat, equipment vibration energy consumption and heat dissipation to the environment) other than the electrical energy input by the motor being converted into the internal energy of the workpiece during the hot forging process.

[0026] S103: Organize the energy deviation time series on the same time axis, plot the deviation change curve of energy input and heat storage, integrate the deviation change curve with the corresponding time axis data, and combine it with the original collected temperature, current and voltage data to establish an energy dissipation dataset; construct the deviation change curve of energy input and heat storage in the data buffer area with the general time axis as the horizontal axis and the energy deviation time series as the vertical axis; concatenate the data points of the deviation change curve with the original collected surface temperature data, center temperature data, instantaneous current value and instantaneous voltage value into a multidimensional array. The specific operation is as follows: create a data structure with the timestamp as the primary key, and use the deviation value, surface temperature, center temperature, current and voltage corresponding to the timestamp as different columns, merge them into a data row, and collect the data rows of all time points to form a multidimensional array; during the concatenation process, give each group of data a unique batch ID and timestamp label to form a structured energy dissipation dataset.

[0027] Please refer to Figure 3. Step S2 is as follows: S201: Extract the rate of change information of the centralized deviation curve of the energy dissipation data, and calculate the starting torque reference value of the main drive motor of the thread rolling machine based on the rate of change information; read the numerical sequence of the centralized deviation curve of the energy dissipation data, and perform numerical differentiation operation on the numerical sequence using the five-point central difference method. For example, to calculate the rate at a certain time point, use the data points before and after the time point, and obtain the rate of change information of the deviation curve at each time point by (the value of the second data point before - 8 × the value of the first data point before + 8 × the value of the first data point after - the value of the second data point after) ÷ (12 × time step); perform absolute value processing on the rate of change information, and statistically analyze the maximum value (e.g., 150 joules per second) and the average value () of the rate of change throughout the entire process of warm upsetting. For example, 30 joules per second); the preset starting torque conversion logic is invoked. The starting torque conversion logic sets the starting torque reference value to be positively correlated with the rate of change of energy dissipation. Specifically, the maximum rate of change (150 joules per second) is multiplied by the preset inertia compensation coefficient (for example, 0.2 Nm per (joules per second). The inertia compensation coefficient is set based on the regression analysis results of the relationship between motor starting acceleration and torque in historical no-load tests, and the slope value is taken). Then, the steady-state friction resistance torque corresponding to the average rate of change (30 joules per second) (for example, 5 Nm) is added. The steady-state friction resistance torque is set based on the average torque value of the measuring equipment when running at a constant speed under no-load. A specific torque value ((150×0.2)+5=35 Nm) is calculated and defined as the starting torque reference value of the main drive motor of the thread rolling machine.

[0028] S202: Collect the rated speed range, transmission ratio range, and current output range of the main drive motor of the thread rolling machine. Use these ranges, along with the starting torque reference value, as constraints and input them into a Bayesian optimization algorithm for optimization to obtain the rolling parameter optimization results. Specifically, the process of inputting these ranges into the Bayesian optimization algorithm to obtain the rolling parameter optimization results involves: setting the rated speed range, transmission ratio range, and current output range together as the parameter search space for the Bayesian optimization algorithm; setting the starting torque reference value as the parameter search space for the Bayesian optimization algorithm. The process involves optimizing the torque constraints within the time limit; the Bayesian optimization algorithm iteratively optimizes the parameters within the search space based on the torque constraints until the optimization process converges or reaches the preset number of iterations; the parameter combination corresponding to the converged state is taken as the optimization result of the rolling parameters; by reading the electronic nameplate parameters of the thread rolling machine servo driver, the rated speed range of the main drive motor (e.g., 0 to 3000 rpm), the transmission ratio range of the mechanical transmission chain (e.g., 10:1 to 50:1), and the current output range of the driver (e.g., 0 to 50 amperes) are obtained; an instance of the Bayesian optimization algorithm is constructed. The Bayesian optimization algorithm structure includes two core components: a Gaussian process regression model (as a surrogate model) and a data acquisition function. The Gaussian process regression model is used to fit a probabilistic model based on the observed (parameter combination, objective function value) data points, predicting the mean and variance of the objective function value for any unobserved parameter combination. The acquisition function (in this example, the expected improvement strategy) uses the prediction results (mean and variance) of the Gaussian process to calculate the specific value of the "expected benefit" that can be brought by "trying" a new parameter point, and selects the point with the largest expected benefit value as the next evaluation point.

[0029] The optimization process is as follows: The algorithm randomly selects initial sample points for evaluation within the parameter search space. Then, the iteration begins: 1. The Gaussian process model updates the posterior distribution based on all evaluated points; 2. The expected lift acquisition function calculates the "expected lift" for all points in the search space; 3. The parameter combination with the largest expected lift (e.g., 2000 rpm, transmission ratio 25, current 40 amps) is selected as the next evaluation point; 4. The objective function value at this point is evaluated; 5. (New parameter point, new objective value) is added to the observation set. The iterative optimization process continues until the improvement of the objective function value in five consecutive iterations is less than the preset convergence threshold (e.g., 0.01; the convergence threshold is set based on the balance between computational efficiency and optimization accuracy, and by analyzing historical optimization data, it is determined that an improvement of less than 0.01 has less than 0.001 mm of impact on the final forming quality index (e.g., mean diameter error)). Alternatively, the number of iterations may reach the preset upper limit (e.g., 100 iterations). The optimal parameter combination corresponding to the final convergence state, namely the determined optimal speed, optimal transmission ratio, and optimal current limit value, is extracted as the optimization result of the rolling parameters.

[0030] S203: Based on the optimization results of the rolling parameters, the drive current setting value is obtained according to the mapping relationship between motor torque and current. The initial linear velocity of the roller is obtained according to the correspondence between the transmission ratio, the roller pitch circle diameter, and the spindle speed. The roller preload force is obtained according to the calibration curve of hydraulic or servo displacement and loading force. The drive current setting value, roller preload force, and roller initial linear velocity are combined to output the rolling parameter instruction set. The rolling parameter optimization results are analyzed, and the linear mapping table between motor torque constant and current is called (e.g., the motor torque constant is 0.8 Nm / Ampere). The optimal torque requirement obtained by optimization (e.g., 40 Nm) is converted into the drive current setting value (40 ÷ 0.8 = 50 Amperes). According to the mechanical transmission principle, the optimal speed (e.g., 1500 rpm) is divided by the optimal transmission ratio (e.g., 20), and then multiplied by the roller pitch circle diameter. (e.g., 0.1 meters) × pi (take 3.14), calculate the initial linear velocity of the roller ((1500÷60)×(1÷20)×0.1×3.14≈0.3925 meters per second); call the force-displacement calibration curve of the hydraulic actuator or servo electric cylinder, which is obtained by measuring the force-displacement calibration curve through a high-precision force gauge and laser displacement sensor during the calibration stage. According to the contact pressure requirement (e.g., 5000 Newtons) implied in the optimization results, find the corresponding actuator displacement (e.g., 2.5 mm) on the calibration curve and set it as the control command corresponding to the roller preload force; package the above calculated drive current setting value, roller initial linear velocity and roller preload force control command to generate a rolling parameter instruction set that conforms to the communication protocol of the thread rolling machine controller, and send it to the underlying programmable logic controller for execution.

[0031] Please refer to Figure 4. Step S3 specifically involves: S301: Determining the operating monitoring benchmark of the thread rolling machine through the rolling parameter instruction set, and collecting real-time operating data of the main drive motor and roller moving parts of the thread rolling machine, including torque response change data over time, roller linear speed change data, and feed displacement change data, integrating them into the actual measured data of the thread rolling machine; After receiving the rolling parameter instruction set, the thread rolling machine controller locks the various parameter values ​​in the instruction set as the operating monitoring benchmark for this processing cycle; After the rolling process starts, the dynamic torque sensor installed on the spindle of the thread rolling machine operates at a high frequency (e.g., 1... The actual output torque is collected at 000 Hz to form torque response data over time; the photoelectric encoder installed on the roller shaft provides real-time feedback of the rotation angle, and the roller linear velocity change data is obtained through differential calculation; the grating ruler installed on the feed slide measures the feed position in real time to generate feed displacement change data; the above three sets of real-time collected data streams are synchronized in time and formatted. Before integration, the three sets of data streams are processed by algorithms such as median filtering or low-pass filtering to remove high-frequency noise signals caused by electrical interference or minor vibrations, and then integrated into the actual measured data of the thread rolling machine.

[0032] S302: Calculate the deviations of the torque response over time, roller linear speed, and feed displacement from the measured data of the thread rolling machine compared to the starting torque reference value, to obtain the relationship between the errors of torque, linear speed, and feed displacement and time; extract the actual torque value (e.g., 41 Nm) at each moment from the measured data of the thread rolling machine, and retrieve the corresponding theoretical load curve value (e.g., the theoretical value should be 40 Nm at the same moment) from the operation monitoring reference, and compare them (41-40=1 Nm) to obtain the torque error sequence. Similarly, the linear velocity error sequence is obtained by comparing the actual roller linear velocity (e.g., 0.39 m / s) with the set initial roller linear velocity (0.3925 m / s) (0.39-0.3925=-0.0025 m / s), and the feed displacement error sequence is obtained by comparing the actual feed displacement (e.g., 1.1 mm) with the theoretical feed trajectory (1.0 mm) (1.1-1.0=0.1 mm). The above three sets of error sequences together constitute the relationship between the errors of torque, linear velocity, and feed displacement and time.

[0033] S303: Based on the time-varying relationship between torque, linear velocity, and feed displacement errors, extract the fluctuation amplitude of torque error, the instantaneous change rate of linear velocity error, and the trajectory deviation of feed displacement error, mapping them to characteristic indicators representing the intensity of rolling load fluctuation and deformation resistance, forming a rolling load feature set; perform time-domain analysis on the torque error sequence, calculate the values ​​of peaks (e.g., 1.5 Nm) and troughs (e.g., -0.5 Nm) in the error sequence (1.5 - (-0.5) = 2 Nm), and define them as the fluctuation amplitude of torque error, which reflects the degree of load oscillation during the rolling process; perform first-order difference operation on the linear velocity error sequence, that is, subtract the error value of the previous moment from the error value of the next moment, and then divide by the time step to obtain the error change rate, and screen out those whose absolute value of the change rate exceeds the preset safety threshold ( For example, the safety threshold of 0.1 meters per square second is set based on the following: (The standard deviation of the linear velocity change rate in normal processing batches is statistically analyzed, and three times the standard deviation is taken as the threshold). The frequency and intensity of abnormal points that meet the conditions are statistically analyzed and defined as the instantaneous change rate of the linear velocity error. The instantaneous change rate reflects the slippage or impact when the roller contacts the workpiece. The root mean square error of the feed displacement error sequence relative to the zero error baseline is calculated by summing the squares of all feed displacement error data points, dividing by the total number of data points, and then taking the square root. This square root is defined as the trajectory deviation of the feed displacement error. The trajectory deviation reflects the rigid deformation of the mechanical structure or the pressure fluctuation of the hydraulic system. The fluctuation amplitude, instantaneous change rate, and trajectory deviation are constructed into a feature vector as characteristic indicators representing the intensity of the rolling load fluctuation and the deformation resistance, and stored in the rolling load feature set.

[0034] Please refer to Figure 5. Step S4 is as follows: S401: Extract the characteristic indicators representing the rolling load fluctuation intensity and deformation resistance from the rolling load feature set, determine the energy density change of the thread plastic contact area formed by the bolt blank surface and the thread rolling die tooth surface, and obtain the energy density change trend; read the rolling load feature set, and combine it with the geometric parameters of the thread rolling die (such as tooth angle and pitch), and use the contact mechanics correction model (e.g., a correction model based on Hertz contact theory and considering the influence of plastic deformation) to estimate the instantaneous contact area of ​​the thread plastic contact area; convert the real-time collected torque data into tangential force, regard the preload force as radial force, and calculate the power of the resultant force; divide the calculated power by (instantaneous contact area × material rheological velocity) to calculate the energy density per unit volume; perform trend fitting on the energy density values ​​in continuous time steps, and use the least squares method for fitting, that is, find a curve that minimizes the sum of the squares of the vertical distances from all energy density data points to the curve, thereby obtaining the energy density change trend.

[0035] S402: Based on the energy density change trend, the forming stage, stable stage, and decay stage of the thread forming contact zone are divided, and the direction of energy gradient change of the energy density curve within each stage is calculated. The stage division and the corresponding energy gradient change direction are integrated to establish the forming stage and gradient set; second derivative analysis is performed on the energy density change trend curve, that is, the derivative of the derivative of the energy density change trend curve is calculated, and the G point where the second derivative value is zero or changes positive or negative is the inflection point; the first curvature abrupt change point is identified, and the time period before the first abrupt change point is defined as the forming stage (energy density rises rapidly); the second curvature abrupt change point is identified, and the time period between the two abrupt change points is defined as the stable stage (energy density remains relatively constant); the second... The time period following each mutation point is defined as the decay phase (energy density gradually decreases). Within each defined phase, the derivative of the energy density curve with respect to time, i.e., the energy gradient, is calculated, and the positive or negative direction and magnitude of the energy gradient are determined, and marked as positive enhancement gradient (e.g., gradient value greater than 10 MPa per second), zero gradient (e.g., gradient value between -10 and 10 MPa per second), or negative decay gradient (e.g., gradient value less than -10 MPa per second). The gradient threshold (10 MPa per second) is set based on: statistical analysis of the energy density data of historical stable processing phases, and taking the upper limit of its fluctuation range as the boundary between stable and unstable phases; the start and end times of the phases are packaged with the corresponding gradient attributes to establish a shaping phase and gradient set.

[0036] S403: Invoke the forming stage and gradient set to obtain the stage division and energy gradient change direction. Combined with the error change results of torque, linear velocity, and feed displacement in the rolling load feature set, determine the corresponding relationship interval between energy distribution and rolling pressure, feed amount, and linear velocity within each stage of the thread forming contact zone, and generate energy stratification judgment results. Specifically, determining the corresponding relationship interval between energy distribution and rolling pressure, feed amount, and linear velocity involves: extracting the error change results of torque, linear velocity, and feed displacement corresponding to the rolling load feature set in each of the forming stage, stabilization stage, and decay stage; and invoking the energy distribution corresponding to each stage in the forming stage and gradient set. The direction of energy gradient change is analyzed; the correlation between the direction of energy gradient change and the error change result of torque is performed to determine the response range of energy gradient to rolling pressure; the correlation between the direction of energy gradient change and the error change result of feed displacement is performed to determine the response range of energy gradient to feed amount; the correlation between the direction of energy gradient change and the error change result of linear velocity is performed to determine the response range of energy gradient to linear velocity; the response ranges of energy gradient to rolling pressure, energy gradient to feed amount, and energy gradient to linear velocity are combined to establish corresponding intervals; each stage in the forming stage and gradient set is traversed; for example, in the forming stage, the energy gradient change corresponding to the forming stage is extracted. The direction of the curve is usually positive, and the results of torque error change, feed displacement error change, and linear velocity error change during the forming stage are extracted simultaneously. Pearson correlation analysis is performed, which calculates the covariance of two sequences (e.g., energy gradient sequence and torque error sequence) and divides it by the product of the standard deviations of the two sequences to obtain a correlation coefficient value between -1 and 1. If the absolute value of the correlation coefficient (e.g., calculated to be 0.85) is greater than a preset correlation threshold (e.g., 0.8, the threshold is set based on the statistical definition of strong correlation and historical data verification, determining 0.8 as the dividing point between correlation (greater than 0.8) and no correlation (less than 0.8), then... The torque error within the current energy gradient range exhibits a clear response relationship to pressure changes, thus defining the response range of the energy gradient to rolling pressure. Similarly, the correlation between the energy gradient and feed displacement error is analyzed to define the response range to feed amount; the correlation between the energy gradient and linear velocity error is analyzed to define the response range to linear velocity. Finally, the union (i.e., including all response ranges) or intersection (e.g., if the control target is T1 level accuracy, the intersection is taken; if it is T2 level accuracy, the union is taken) of the above three response ranges is obtained to establish the correspondence range between energy distribution and various physical quantities during the forming stage, generating energy stratification judgment results, and clarifying which type of physical parameter error should be focused on under what energy state.

[0037] Please refer to Figure 6. Step S5 is as follows: S501: Extract the upsetting stage feature data from the energy dissipation dataset and the rolling stage feature data from the energy stratification judgment result. Based on the processing sequence of the same bolt blank, establish the forming parameter input set for the bolt upsetting and rolling stages. Access the database and accurately retrieve the upsetting stage feature data from the energy dissipation dataset according to the processing batch number and workpiece serial number. At the same time, extract the rolling stage feature data from the energy stratification judgment result. According to the time sequence of physical processing, use the upsetting data as the preorder vector and the rolling data as the postorder vector, and perform time alignment and feature fusion. The specific method of feature fusion is, for example, to combine the heat residual value and final temperature gradient of the upsetting stage with the initial torque and initial linear velocity of the rolling stage into a new feature vector (for example) containing 50 feature values, and establish the forming parameter input set for the bolt upsetting and rolling stages.

[0038] S502: Input the set of forming parameters from the bolt warm upsetting and rolling stages into the support vector machine model. Based on the mapping relationship between the set of forming parameters and the corresponding forming quality, predict the bolt thread pitch diameter error, thread angle deviation, and Vickers hardness distribution to obtain the bolt forming quality index; load the pre-trained support vector machine regression model. The model structure is as follows: the input layer receives the set of forming parameters from step S501; the input data is mapped to a high-dimensional feature space through a kernel function (radial basis function in this example). The radial basis function achieves nonlinear mapping by calculating the exponential decay of the Euclidean distance between the input vector and the support vector. In the high-dimensional space, the model constructs a regression hyperplane that fits the (input, quality index) data points in the training set. The fitting criterion is to minimize the sum of the total deviations of all data points from the hyperplane (within a preset tolerance, such as an error range of 0.005 mm). The model projects the input vector onto the regression hyperplane based on the position of the input vector in the high-dimensional space to obtain the predicted value, which outputs three specific predicted values: the predicted value of the bolt thread pitch diameter error (e.g., 0.02 mm), the predicted value of the thread angle deviation (e.g., 0.1 degrees), and the predicted value of the Vickers hardness distribution (e.g., 310). The above three values ​​constitute the forming quality index of the bolt.

[0039] S503: Based on the bolt forming quality indicators, calculate the deviation between the predicted and target indicators. Correct the drive current setting, roller preload force, and roller initial linear velocity in the rolling parameter command set according to the deviation, and output the bolt forming quality control command. Specifically, the process of correcting the drive current setting, roller preload force, and roller initial linear velocity in the rolling parameter command set according to the deviation is as follows: Using the radial deviation value corresponding to the bolt thread pitch diameter error in the deviation, determine the compensation amount for the radial extrusion force based on the deformation correspondence between the roller indentation depth and the pitch diameter dimension, and use the compensation amount to correct the roller preload force in the rolling parameter command set. The loading force is adjusted incrementally or decreased; the angle deviation value corresponding to the tooth profile angle deviation in the deviation quantity is used as the basis for evaluating the meshing and slippage state between the roller and the blank, calculating and optimizing the speed adjustment coefficient for metal flow filling effect and synchronization, and updating the initial linear velocity of the roller in the rolling parameter instruction set; the hardness deviation value corresponding to the Vickers hardness distribution in the deviation quantity is used to calculate and adjust the motor output power based on the correlation between the degree of material work hardening and plastic deformation energy consumption, controlling the current correction value for cold work hardening effect, and resetting the drive current setting value in the rolling parameter instruction set; the target indicators specified in the product process drawings (i.e., standard pitch diameter error of 0, standard tooth profile ... (The profile angle deviation is 0, and the standard hardness is 300). Compare the output predicted value of the bolt thread pitch diameter error (e.g., 0.02 mm) with the target index (standard pitch diameter error is 0) (0.02 - 0 = 0.02 mm). For the radial deviation value (0.02 mm) corresponding to the bolt thread pitch diameter error, consult the table of nonlinear deformation correspondence between roller indentation depth and pitch diameter, and obtain the indentation depth correction amount (e.g., 0.01 mm) required to eliminate the radial deviation. Then, based on the equipment stiffness (e.g., 10000 N / mm), calculate the incremental or decremental adjustment value of the roller preload force (0.01 × 10000 = 100 N). The pressure parameters in the rolling parameter instruction set should be directly corrected by reducing the pressure by 100 Newtons. For the angle deviation value (0.1 degrees) corresponding to the tooth profile angle deviation, it is regarded as a signal of uneven metal flow. Based on the constant volume condition in the metal plastic forming principle, the relative sliding speed required to balance the metal flow is calculated, and then the speed adjustment coefficient is obtained (e.g., 1.02, the speed adjustment coefficient is based on the geometric relationship model between tooth profile angle deviation and insufficient metal filling). The speed parameter in the instruction set is updated to 0 by multiplying the speed adjustment coefficient (1.02) by the current initial linear speed of the roller (0.3925 m / s).40035 m / s); For the hardness deviation value (10) corresponding to the Vickers hardness distribution, based on the work hardening constitutive equation of metallic materials, the plastic deformation work difference value required to generate the hardness deviation is derived in reverse, and the plastic deformation work difference value is converted into the output power adjustment amount of the motor, and then the reset value of the drive current setting value is calculated (for example, reduced from 50 amperes to 49 amperes) to change the thermal-mechanical coupling strength in the processing process; after completing all the above correction calculations, the final bolt forming quality control command is output.

[0040] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for collaborative detection and optimization of process parameters for hot upsetting-rolling of high-temperature alloy bolts, characterized in that, Includes the following steps: S1: Collect data on the change of heat storage at the center and surface of the bolt blank during the warm upsetting stage, as well as data on the change of energy input of the main drive motor of the warm upsetting press over time, and construct an energy dissipation dataset; S2: Calculate the starting torque reference value of the main drive motor of the thread rolling machine based on the energy dissipation dataset. Collect the rated speed range, transmission ratio range, and current output range of the main drive motor of the thread rolling machine as constraints for optimization, and select the rolling parameter instruction set; S3: Determine the operating monitoring benchmark of the thread rolling machine through the rolling parameter instruction set, and perform error analysis based on the measured data of the thread rolling machine to form a rolling load feature set; S4: Analyze the energy density change of the thread forming contact area formed by the bolt blank surface and the thread rolling die tooth surface based on the rolling load feature set, and generate energy stratification judgment results; S5: Combine the energy stratification judgment results and the energy dissipation dataset to form the forming parameter input set for the bolt warm upsetting and rolling stages, predict the forming quality index of the bolt, correct the rolling parameter instruction set based on the prediction results, and output the bolt forming quality control instruction.

2. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 1, characterized in that, The energy dissipation dataset includes energy input variation curves, heat storage variation curves, and energy difference variation curves. The rolling parameter instruction set includes drive current setting, roller preload force, and roller initial linear velocity. The rolling load feature set includes torque error variation results, linear velocity error variation results, and feed displacement error variation results. The energy stratification determination results include stratification information for the forming stage, stratification information for the stable stage, and stratification information for the decay stage. The bolt forming quality control instructions specifically include bolt thread pitch diameter error, thread angle deviation, and Vickers hardness distribution.

3. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 1, characterized in that, Step S1 specifically includes: S101: Collecting temperature change data over time at the center and surface of the bolt blank during the upsetting stage, and collecting current and voltage change data over time of the main drive motor during heating and upsetting. Obtaining the blank density and specific heat parameters, converting the temperature difference between the center and surface with the blank density and specific heat parameters to obtain heat storage change data over time, and converting the current and voltage data according to the time correspondence to obtain energy input change data over time; S102: Obtaining the energy input value in the energy input change data over time and the heat storage value in the heat storage change data over time on the same time axis, comparing them point by point, calculating the difference between the two at each time point, and obtaining the energy deviation time series quantity; S103: Organizing the energy deviation time series quantity on the same time axis, plotting the deviation change curve of energy input and heat storage, integrating the deviation change curve with the corresponding time axis data, and combining it with the originally collected temperature, current, and voltage data to establish an energy dissipation dataset.

4. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 1, characterized in that, Step S2 specifically includes: S201: Extracting the rate of change information of the central deviation change curve of the energy dissipation data, and calculating the starting torque reference value of the main drive motor of the thread rolling machine based on the rate of change information; S202: Collecting the rated speed range, transmission ratio range and current output range of the main drive motor of the thread rolling machine, and using the rated speed range, transmission ratio range and current output range together with the starting torque reference value as constraints, inputting them into the Bayesian optimization algorithm for optimization, and obtaining the rolling parameter optimization result; S203: Based on the rolling parameter optimization result, obtaining the drive current setting value according to the mapping relationship between motor torque and current, obtaining the initial linear velocity of the roller according to the correspondence between the transmission ratio, the roller pitch circle diameter and the spindle speed, and obtaining the roller preload force according to the calibration curve of hydraulic or servo displacement and loading force, combining the drive current setting value, the roller preload force and the roller initial linear velocity, and outputting the rolling parameter instruction set.

5. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 4, characterized in that, The process of inputting the parameters into the Bayesian optimization algorithm to obtain the optimization results of the rolling parameters is as follows: the rated speed range, transmission ratio range, and current output range are jointly set as the parameter search space of the Bayesian optimization algorithm; the starting torque reference value is set as the torque constraint condition for the Bayesian optimization algorithm to perform optimization within the parameter search space; the Bayesian optimization algorithm iteratively optimizes within the parameter search space and according to the torque constraint condition until the optimization process converges or reaches the preset number of iterations; The parameter combination corresponding to the convergence state is used as the optimization result of the rolling parameters.

6. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 1, characterized in that, Step S3 specifically comprises: S301: Determining the operating monitoring benchmark of the thread rolling machine through the rolling parameter instruction set, and collecting real-time operating data of the main drive motor and roller moving parts of the thread rolling machine, including torque response change data over time, roller linear speed change data, and feed displacement change data, and integrating them into the measured data of the thread rolling machine; S302: Calculating the deviations of the torque response change data over time, roller linear speed change data, and feed displacement change data in the measured data of the thread rolling machine from the starting torque benchmark value, and obtaining the relationship between the errors of torque, linear speed, and feed displacement and time; S303: Based on the relationship between the errors of torque, linear speed, and feed displacement and time, extracting the fluctuation amplitude of torque error, the instantaneous change rate of linear speed error, and the trajectory deviation of feed displacement error, mapping them into characteristic indicators characterizing the intensity of rolling load fluctuation and deformation resistance, and forming a rolling load feature set.

7. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 1, characterized in that, Step S4 is as follows: S401: Extract the characteristic indicators representing the rolling load fluctuation intensity and deformation resistance from the rolling load feature set, determine the energy density change of the thread forming contact area formed by the bolt blank surface and the thread rolling die tooth surface, and obtain the energy density change trend; S402: Based on the energy density change trend, divide the thread forming contact area into forming stage, stabilizing stage and decay stage, and calculate the energy gradient change direction of the energy density curve in each stage, integrate the stage division with the corresponding energy gradient change direction, and establish the forming stage and gradient set; S403: Call the forming stage and gradient set, obtain the stage division and energy gradient change direction, and combine the error change results of torque, linear velocity and feed displacement from the rolling load feature set, determine the corresponding relationship interval between energy distribution and rolling pressure, feed amount and linear velocity in each stage of the thread forming contact area, and generate energy stratification judgment results.

8. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 7, characterized in that, The process of determining the corresponding relationship range between energy distribution and rolling pressure, feed rate, and linear velocity is as follows: In each of the forming stage, stabilization stage, and decay stage, the error change results of torque, linear velocity, and feed displacement corresponding to the rolling load feature set are extracted respectively; the energy gradient change direction corresponding to each stage in the shaping stage and gradient set is called; the energy gradient change direction and the error change result of torque are correlated and analyzed to determine the response range of energy gradient to rolling pressure; By correlating the direction of energy gradient change with the error change of feed displacement, the response range of energy gradient to feed amount is determined. The correlation analysis between the direction of energy gradient change and the error change of linear velocity is performed to determine the response range of energy gradient to linear velocity; the response ranges of energy gradient to rolling pressure, energy gradient to feed rate, and energy gradient to linear velocity are combined to establish corresponding relationship ranges.

9. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 1, characterized in that, Step S5 specifically comprises: S501: Extracting the upsetting stage feature data from the energy dissipation dataset and the rolling stage feature data from the energy stratification judgment result, and establishing a set of forming parameters for the bolt upsetting and rolling stages based on the processing sequence of the same bolt blank; S502: Inputting the set of forming parameters for the bolt upsetting and rolling stages into the support vector machine model, and predicting the bolt thread pitch diameter error, thread angle deviation, and Vickers hardness distribution based on the mapping relationship between the forming parameter input set and the corresponding forming quality, thereby obtaining the bolt forming quality index; S503: Calculating the deviation between the predicted index and the target index based on the bolt forming quality index, and correcting the drive current setting, roller preload force, and roller initial linear velocity in the rolling parameter instruction set based on the deviation, and outputting the bolt forming quality control instruction.

10. The method for collaborative detection and optimization of high-temperature alloy bolt upsetting-rolling process parameters according to claim 9, characterized in that, The process of correcting the drive current setting, roller preload force, and roller initial linear velocity in the rolling parameter instruction set based on the deviation is as follows: Using the radial deviation value corresponding to the bolt thread pitch diameter error in the deviation, and based on the deformation correspondence between the roller indentation depth and the pitch diameter, the compensation amount for the radial extrusion force is determined, and the roller preload force in the rolling parameter instruction set is adjusted by increment or decrement using the compensation amount; using the angle deviation value corresponding to the tooth angle deviation in the deviation as the basis for evaluating the meshing and slippage state between the roller and the blank, the speed adjustment coefficient for optimizing the metal flow filling effect and synchronization is calculated, and the roller initial linear velocity in the rolling parameter instruction set is updated; using the hardness deviation value corresponding to the Vickers hardness distribution in the deviation, the motor output power is adjusted based on the correlation between the degree of material work hardening and plastic deformation energy consumption, the current correction value for the cold work hardening effect is controlled, and the drive current setting value in the rolling parameter instruction set is reset.

Images