Spring steel high speed wire fatigue performance improvement rolling process method

Abstract

The application discloses a kind of spring steel high-speed wire fatigue performance promotion rolling process method, it is related to rolling technical field.The application can identify abnormal temperature drop pass and dynamically correct each pass reduction distribution in real time by establishing the dynamic analysis mechanism of multiple-pass cumulative deformation and temperature drop trend, significantly improve the stability and microstructure refinement effect of rolling process by combining key deformation pass vibration characteristics and dynamic recrystallization condition judgment;And by subsection acquisition cooling data and establish the quantitative corresponding relationship of cooling rate-structure transformation, the fine control of cooling system is realized, ensure that wire obtains target phase change organization, batch performance consistency is greatly improved;Multi-parameter fatigue prediction model is constructed, which integrates finish rolling temperature, cooling rate, grain size and mechanical property, and through residual trend monitoring trigger closed-loop correction, adaptive optimization and continuous improvement of process parameters are realized.

Description

Technical Field

[0001] This invention relates to the field of rolling technology, specifically to a rolling process for improving the fatigue performance of high-speed spring steel wire rod. Background Technology

[0002] With the increasing demands of modern industry for lightweight, high-strength, and long-life components, spring steel wire rods not only need to possess excellent static mechanical properties, but also face extremely stringent requirements for fatigue life under alternating load conditions. Fatigue performance, as a core indicator for measuring the service reliability of spring steel wire rods, directly determines the safety margin and service life of spring components in actual use. Therefore, how to systematically improve the fatigue performance of high-speed spring steel wire rods through the optimization of rolling processes has become an important technical issue that urgently needs to be addressed in the current metallurgical rolling field.

[0003] Currently, the production of high-speed spring steel wire rod typically involves heating continuously cast billets in a furnace, followed by multi-pass high-speed rolling, and then controlled cooling to obtain the target microstructure. While some existing research has focused on the effects of heating regime, rolling process, and cooling regime on the microstructure and properties of spring steel, the following shortcomings remain overall:

[0004] Firstly, in the heating stage, existing heating processes primarily aim to achieve the target austenitizing temperature range for the entire continuously cast billet, lacking refined control over the temperature uniformity at different spatial locations within the same cross-section. During the heating process, due to the large billet size and uneven distribution of furnace atmosphere and heat flow, large temperature gradients are easily generated across the cross-section. This temperature inhomogeneity leads to differences in the degree of austenitization along the cross-sectional direction, resulting in inconsistent microstructure evolution during subsequent rolling deformation, ultimately creating performance fluctuations and fatigue-weak regions in the wire rod. While existing technologies provide qualitative descriptions of heating uniformity, they lack quantitative characterization methods for cross-sectional temperature uniformity and corresponding control measures for its connection with subsequent process stages.

[0005] Secondly, in multi-pass high-speed rolling, existing processes typically set the reduction amount for each pass based on experience and keep it relatively fixed during production, lacking a mechanism for dynamic correction based on actual rolling data. During actual rolling, factors such as roll wear, billet temperature fluctuations, and equipment condition changes often cause the actual deformation and temperature drop patterns of each pass to deviate from the preset values. Especially under high-speed rolling conditions, the temperature drop rate between passes may abruptly change. If the reduction amount distribution is not identified and adjusted in time, the cumulative deformation will deviate from the target value, affecting the uniformity of the microstructure and the degree of recrystallization. Furthermore, existing technologies lack a systematic analysis of the vibration characteristics and dynamic recrystallization conditions of key deformation passes, making it difficult to avoid surface quality defects and microstructure inhomogeneity caused by resonance. Simultaneously, information on the distribution of inclusions within the continuously cast billet is not fully utilized in the rolling process, resulting in areas with dense inclusions potentially failing to be sufficiently broken up due to insufficient deformation, becoming a source of fatigue crack initiation. Summary of the Invention

[0006] To solve the above-mentioned technical problems, a rolling process for improving the fatigue performance of high-speed spring steel wire rod is provided. This technical solution solves the problems mentioned in the background art.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A rolling process for improving the fatigue performance of high-speed spring steel wire rod includes:

[0009] The spring steel continuous casting billet after smelting and refining is heated, and the heating temperature data is obtained and compared and analyzed with the target austenitizing temperature range.

[0010] The heated continuous casting billet is subjected to high-speed rolling in multiple passes. During each rolling pass, rolling force, rolling temperature, cross-sectional dimensions of the rolled piece and rolling speed data are collected simultaneously. The cumulative deformation of the data from multiple passes is statistically analyzed and the temperature drop trend between passes is analyzed.

[0011] The distribution of reduction for each pass is dynamically corrected based on the analysis results of cumulative deformation and temperature drop of each pass.

[0012] The wire rod after rolling is cooled after rolling. During the cooling process, the surface temperature of the wire rod and the flow rate of the cooling medium are collected in segments, and the correspondence between cooling rate and microstructure transformation is established.

[0013] The cooled wire is subjected to microstructure testing and fatigue performance prediction and evaluation, and the evaluation results are fed back to the rolling parameter adjustment process to form a closed-loop control.

[0014] Preferably, the step of heating the refined spring steel continuous casting billet, obtaining heating temperature data, and comparing and analyzing it with the target austenitizing temperature range specifically includes:

[0015] Before the continuously cast billet is fed into the heating furnace, its surface defects are scanned and recorded, and the defect location information is associated with the deformation path of subsequent rolling passes.

[0016] The continuously cast billet is placed in the heating furnace in sections. Temperature acquisition points are set up in the preheating section, heating section and soaking section of the heating furnace respectively, and the temperature value of each acquisition point is continuously recorded at a preset time interval.

[0017] The variance of temperature values ​​at different spatial locations on the same time section is calculated. When the variance exceeds a preset uniformity threshold, the section is marked as a temperature non-uniform region and its location coordinates are recorded. The calculation method for the temperature uniformity variance of the same section is as follows:

[0018] ;

[0019] in, The variance of temperature uniformity across the same cross section. This represents the total number of temperature sampling points on this cross-section. For the first Temperature measurements at each collection point This is the arithmetic mean of the temperatures at all sampling points across this cross section;

[0020] The coordinates of the temperature non-uniformity regions at each cross section are superimposed and compared with the original segregation spectrum data of the continuously cast billet to identify the overlap range between the severely segregated regions and the temperature non-uniformity regions. The quantification method for the overlap degree is as follows:

[0021] ;

[0022] in, This represents the percentage of overlap between the segregation region and the temperature non-uniformity region. This represents the spatial overlap area between the two types of regions. The area is the union of the two types of regions;

[0023] The extension of the holding time in the heating homogenization zone is determined based on the size and distribution density of the overlapping area, and the adjusted heating regime parameters are sent to the heating furnace control terminal for execution.

[0024] Preferably, the process of performing multi-pass high-speed rolling on the heated continuously cast billet, and simultaneously collecting data on rolling force, rolling temperature, cross-sectional dimensions of the rolled piece, and rolling speed during each rolling pass, specifically includes:

[0025] Temperature sensors are installed on the inlet and outlet sides of each mill stand. The difference between the inlet and outlet temperatures of the same pass is calculated to obtain the absolute temperature drop value of that pass.

[0026] The absolute temperature drop values ​​of each pass are arranged in the rolling sequence to form a temperature drop sequence. A first-order difference operation is performed on this sequence to obtain the rate of temperature drop change between adjacent passes, where the... The calculation method for the temperature drop rate of each pass is as follows:

[0027] ;

[0028] in, For the first The order of the path relative to the first The rate of change of temperature drop per pass and The first The inlet and outlet temperatures of each pass. and The first Dao Ci and Di The rolling time of each pass;

[0029] When the ratio of the temperature drop rate of a certain pass to the temperature drop rate of the previous pass is greater than the preset mutation coefficient, the pass is determined to be an abnormal temperature drop pass, and the rolling force data corresponding to the pass is compared with the rolling force data of the normal pass to detect outliers and remove abnormal rolling force data points caused by equipment fluctuations.

[0030] The total cumulative deformation is obtained by summing the actual deformation amounts of each pass after removing outliers. The calculation methods for the area reduction rate of each pass and the total cumulative deformation are as follows:

[0031] ;

[0032] ;

[0033] in, For the first The area reduction rate of each lane, and The first Cross-sectional area of ​​the rolled piece before and after each rolling pass The total cumulative area reduction deformation. Total number of rolling passes;

[0034] The deviation between the total cumulative deformation and the target total deformation is calculated. When the deviation exceeds the allowable range, the deviation value is back-calculated and allocated according to the original reduction ratio of each pass to obtain the correction reduction amount that needs to be added or reduced for each pass. The calculation method for the correction reduction per pass is as follows:

[0035] ;

[0036] in, For the first The adjustment value for the amount of pressure that needs to be added or reduced per pass. To reduce the deformation amount to the target total area, For the first Height of the rolled piece before each rolling pass.

[0037] Preferably, the step of performing cumulative deformation statistics and inter-pass temperature drop trend analysis on multi-pass data further includes:

[0038] The area reduction rate is calculated by comparing the cross-sectional dimensions of the rolled piece after each pass with the cross-sectional dimensions before that pass.

[0039] The area reduction rate of each single pass is plotted as an area reduction rate distribution curve according to the rolling sequence. The peak value of the curve is identified, and the pass with the largest area reduction rate is identified as the critical deformation pass.

[0040] Frequency domain analysis was performed on the rolling force data of key deformation passes to extract the dominant frequency component from the rolling force signal. This dominant frequency component was then compared with the natural vibration frequency of the workpiece in that pass. The natural vibration frequency was estimated as follows:

[0041] ;

[0042] in, The natural vibration frequency of the rolled piece in this pass. This represents the elastic modulus of spring steel at this temperature. The density of spring steel, This is the effective vibration length of the rolled piece in that pass;

[0043] When the ratio of the main frequency component to the natural vibration frequency falls into the resonance range, the rolling speed of this pass is reduced by a preset ratio, and the parameters after the speed reduction are rewritten into the rolling procedure of this pass.

[0044] Simultaneously, local gradient calculations are performed on the cumulative deformation of the two passes before and after the key deformation pass. The local gradient values ​​are compared with the empirical threshold of the critical deformation amount for dynamic recrystallization to determine whether the key deformation pass meets the conditions for sufficient dynamic recrystallization. The calculation method for the local deformation gradient is as follows:

[0045] ;

[0046] in, This represents the local deformation gradient two passes before and after the critical deformation pass. and These represent the area reduction rates for the second and first two passes after the critical deformation pass, respectively. This represents the average time interval between adjacent passes.

[0047] Preferably, the dynamic correction of the reduction distribution for each pass based on the cumulative deformation and pass temperature drop analysis results specifically includes:

[0048] The target value of total cumulative deformation is distributed non-uniformly according to the rolling sequence from roughing to finishing, so that the proportion of cumulative deformation in the roughing stage is greater than that in the finishing stage.

[0049] For each pass, the theoretical rolling force corresponding to the corrected reduction is estimated. The estimated value is then compared in real time with the actual rolling force collected for that pass, and the deviation rate between the two is calculated. The calculation method for the rolling force deviation rate of each pass is as follows:

[0050] ;

[0051] in, For the first Rolling force deviation rate of each pass For the first The actual rolling force collected per pass. For the first The theoretical rolling force is estimated based on the corrected reduction amount per pass;

[0052] When the deviation rate of a certain pass exceeds three consecutive acquisition cycles and the deviation direction is consistent, the reduction correction value of that pass is locked and frozen, and no further adjustment is made. At the same time, the locking information of that pass is passed to subsequent passes as a constraint condition.

[0053] After the reduction amount of all passes is corrected, the corrected reduction amount sequence of each pass is summed and verified to ensure that the deviation between the corrected total deformation amount and the target total deformation amount is within the allowable range.

[0054] The verified reduction sequence is compared with the inclusion distribution information in the original internal quality inspection data of the continuously cast billet. If the deformation area of ​​a certain pass happens to cover the dense inclusion area, the reduction of that pass is supplemented with an additional deformation amount on the basis of the original correction value, so that the inclusions are fully broken up in that pass. The density determination of the dense inclusion area and the calculation method of the additional deformation amount are as follows:

[0055] ;

[0056] ;

[0057] in, The number density of inclusions within a certain deformation region. This refers to the number of inclusions counted in this area. The volume of this region, Thresholds are determined for areas with dense inclusions. For the first The additional reduction in pressure applied during each pass due to the coverage of areas with dense inclusions. This is the correction factor for inclusion breakage. For the first Height of the rolled piece before each rolling pass.

[0058] Preferably, the post-rolling cooling of the wire rod after rolling, wherein data on the surface temperature of the wire rod and the flow rate of the cooling medium are collected in segments during the cooling process and a correspondence between cooling rate and microstructure transformation is established, specifically includes:

[0059] The post-rolling cooling zone is divided into three temperature ranges: high temperature, medium temperature, and low temperature. A set of temperature acquisition points is set at the beginning and end of each temperature range, and the surface temperature value of the wire is collected at fixed time intervals.

[0060] The average cooling rate of a temperature range is obtained by dividing the temperature difference between the starting and ending points of the same temperature range by the spatial length of that range. The average cooling rate over the temperature range is calculated as follows:

[0061] ;

[0062] in, For the first Average cooling rate over the temperature range and The first The surface temperature of the wire at the beginning and end of the interval. For the first The spatial length of the temperature range; compare the average cooling rate of each temperature range with the critical cooling rate of the phase change corresponding to that temperature range to determine whether the wire avoids the pearlite transformation nose temperature region in each range.

[0063] Time series smoothing was performed on the cooling rate data in the medium temperature range to remove temperature fluctuation noise caused by uneven cooling medium spraying, resulting in a smoothed cooling rate curve.

[0064] The smoothed cooling rate curve is superimposed with the pre-established continuous cooling transformation diagram of the steel grade to determine the actual microstructure transformation product type and proportion of the wire in each temperature range.

[0065] The proportion of microstructure transformation products in each interval is stored in a one-to-one correspondence with the corresponding cooling rate value to form a cooling rate-microstructure transformation data table for this batch of wires.

[0066] Preferably, establishing the correspondence between cooling rate and tissue transformation further includes:

[0067] The cooling rate-microstructure transformation data of different batches of wire rod of the same steel grade under the same cooling regime were statistically summarized, and the inter-batch standard deviation of cooling rate for each temperature range was calculated. The standard deviation of the batch-to-batch cooling rate within the temperature range is calculated as follows:

[0068] ;

[0069] in, For the first Batch-to-batch standard deviation of cooling rate within temperature range To count the total number of batches, For the first Batch No. Average cooling rate over the temperature range For all batches in the first The average cooling rate over the temperature range;

[0070] When the standard deviation of the cooling rate between batches in a certain temperature range exceeds the preset stability threshold, the cooling medium injection parameters in that temperature range are traced and analyzed to find the specific injection port number that caused the cooling rate fluctuation.

[0071] The historical maintenance record data corresponding to the nozzle number is correlated with the current flow attenuation data of the nozzle to determine whether the nozzle needs to be replaced or cleaned.

[0072] After replacement or cleaning, the cooling rate data for this temperature range is re-collected, the new data is merged and updated with the previous corresponding data table, and the updated data table is checked for consistency to ensure that the deviation of the proportion of tissue transformation products between the old and new data is within the allowable range.

[0073] The finalized cooling rate-microstructure transformation relationship will be used as the benchmark for cooling in subsequent batches of this steel grade.

[0074] Preferably, the step of performing microstructure analysis and fatigue performance prediction and evaluation on the cooled wire specifically includes:

[0075] Samples were taken from the cooled wire at the specified locations, and the metallographic structure of the samples was observed to obtain data on the original austenite grain size, pearlite lamellar spacing, and carbide distribution density.

[0076] Correlation analysis was performed between the original austenite grain size data and the final rolling temperature and post-rolling cooling rate data recorded during the rolling process of this batch, establishing a three-dimensional mapping relationship between final rolling temperature, cooling rate, and grain size. The correlation coefficient was calculated as follows:

[0077] ;

[0078] in, This represents the correlation coefficient between final rolling temperature, cooling rate, and grain size. For the first The final rolling temperature of each sample, For the first Post-rolling cooling rate of each sample For the first The original austenite grain size of each sample , , Each is their respective average value. The total number of samples;

[0079] Rotational bending fatigue tests were conducted on the sampled wires, and data on the number of fracture cycles under various stress levels were collected. The data were arranged from high to low stress levels, and linear regression fitting was performed on the high-stress and low-stress regions respectively to obtain the conditional fatigue limit values ​​corresponding to the two fitted lines and their intersection points. The conditional fatigue limit was calculated as follows:

[0080] ;

[0081] in, The conditional fatigue limit, and These represent the slope and intercept of the linear regression fitted line in the high-stress region, respectively. and These are the slope and intercept of the linear regression fitted line in the low-stress area, respectively.

[0082] Multi-parameter regression analysis was performed on the conditional fatigue limit value and the tensile strength, reduction of area and hardness data of this batch of wire to establish a predictive model between the fatigue limit and mechanical performance parameters.

[0083] The residuals between the output values ​​of the prediction model and the actual fatigue test values ​​are calculated. When the absolute value of the residuals exceeds the preset accuracy requirements, the deviation information of that batch is marked and included in the model correction dataset.

[0084] Preferably, the step of feeding back the evaluation results to the rolling parameter adjustment stage to form a closed-loop control specifically includes:

[0085] The residual data of the fatigue performance prediction model for each batch of wire are arranged in chronological order to form a residual trend sequence.

[0086] A sliding window mean is calculated for the residual trend sequence, with the window size set to five consecutive batches. When the sliding mean shows a continuous upward or downward trend and the trend slope exceeds a preset drift threshold, the rolling parameters are corrected. The sliding window mean is calculated as follows:

[0087] ;

[0088] in, For the first The mean of a sliding window with batch as the endpoint. The window size is 5. For the first Batch prediction model residuals;

[0089] When a global correction is triggered, the cooling rate-microstructure transformation data table for the current batch, the reduction correction records for each pass, and the final rolling temperature data are packaged as the correction input dataset.

[0090] The sensitivity coefficients of each parameter in the corrected input dataset to the fatigue performance prediction model are ranked, and the top three parameters with the highest sensitivity coefficients are selected as priority adjustment parameters, where the first... The sensitivity coefficients of each parameter are calculated as follows:

[0091] ;

[0092] in, For the first The sensitivity coefficients of each parameter to the conditional fatigue limit For the first The current values ​​of the parameters, This represents a small perturbation to the parameter;

[0093] The current value of the priority adjustment parameter is compared with the historical best value to calculate the adjustment range and direction of each parameter and generate a parameter adjustment plan.

[0094] The parameter adjustment plan is issued to the next batch of rolling execution, and the corrected input dataset and adjustment plan for this batch are archived together for use in trend analysis of subsequent batches.

[0095] Preferably, the method further includes archiving and traceability processing of the entire process data, specifically including:

[0096] Temperature data, rolling force data, deformation data, cooling rate data, microstructure testing data, and fatigue test data collected throughout the entire process from heating the continuous casting billet to fatigue performance evaluation for each batch of wire rod are stored in association with the batch number to form a complete process data archive for that batch.

[0097] A horizontal comparison of the entire process data archives of different batches of the same steel grade was conducted to extract the batch-to-batch fluctuation range of each key parameter, and a baseline for the fluctuation of key parameters for that steel grade was established. The calculation method for the upper and lower limits of the fluctuation baseline of the key parameters is as follows:

[0098] ;

[0099] in, and The first The lower and upper limits of the fluctuation baseline for each key parameter. This is the mean of the parameter across all batches. The standard deviation of this parameter. and These are the confidence coefficients for the lower and upper limits, respectively.

[0100] When a key parameter of a batch deviates from the fluctuation baseline by more than a preset abnormal multiple, the batch is automatically marked as a batch of key concern, and its entire process data file is retrieved for item-by-item backtracking to locate the first abnormal link that caused the deviation.

[0101] Fine-grained analysis was performed on the data of the first abnormal step identified, and the distribution of the data of this step was compared with that of the normal batch of data in the same step to identify the specific characteristics of the abnormal data points.

[0102] A causal correlation analysis is performed between the description of abnormal characteristics and the recorded operational parameters of the abnormal process to form an abnormal cause determination report. The conclusions in the determination report are used as the basis for revising the subsequent operating procedures for that process.

[0103] Compared with the prior art, the present invention provides a rolling process for improving the fatigue performance of high-speed spring steel wire rod, which has the following beneficial effects:

[0104] I. By quantitatively characterizing the temperature uniformity of the heating section of the continuously cast billet and comparing the superimposed segregation regions, the heating regime was precisely controlled, effectively eliminating the difference in austenitization degree caused by temperature inhomogeneity, and ensuring the uniformity of the microstructure from the source.

[0105] Second, a multi-pass cumulative deformation statistics and temperature drop trend dynamic analysis mechanism was established, which can identify abnormal temperature drop passes in real time and dynamically correct the reduction distribution of each pass. Combined with the vibration characteristics of key deformation passes and dynamic recrystallization conditions, the stability of the rolling process and the microstructure refinement effect are significantly improved.

[0106] Third, by collecting cooling data in segments and establishing a quantitative correspondence between cooling rate and microstructure transformation, the cooling regime was precisely controlled, ensuring that the wire obtained the target phase change microstructure and significantly improving the performance consistency between batches.

[0107] Fourth, a multi-parameter fatigue prediction model integrating final rolling temperature, cooling rate, grain size and mechanical properties was constructed, and closed-loop correction was triggered by residual trend monitoring, realizing adaptive optimization and continuous improvement of process parameters.

[0108] Fifth, a full-process data association and storage and anomaly tracing mechanism has been established, which can quickly locate deviations and form a basis for revising operating procedures, promoting the systematic accumulation and iterative upgrading of process knowledge. Attached Figure Description

[0109] Figure 1 This is a schematic diagram of the method flow for S101-S105 in this invention;

[0110] Figure 2 This is a schematic diagram of the method flow for S201-S205 in this invention;

[0111] Figure 3 This is a schematic diagram of the method flow for S301-S305 in this invention;

[0112] Figure 4 This is a schematic diagram of the method flow for S401-S405 in this invention;

[0113] Figure 5 This is a schematic diagram of the method flow for S501-S505 in this invention. Detailed Implementation

[0114] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0115] Example 1

[0116] Please refer to Figure 1 As shown, a rolling process for improving the fatigue performance of high-speed spring steel wire rod includes:

[0117] S101. Heat the refined spring steel continuous casting billet, obtain the heating temperature data, and compare and analyze it with the target austenitizing temperature range.

[0118] S102. The heated continuous casting billet is subjected to multi-pass high-speed rolling. During each rolling process, rolling force, rolling temperature, cross-sectional dimensions of the rolled piece and rolling speed data are collected simultaneously. The cumulative deformation of the multi-pass data is statistically analyzed and the temperature drop trend between passes is analyzed.

[0119] S103. Dynamically correct the distribution of reduction amount for each pass based on the analysis results of cumulative deformation and temperature drop of each pass;

[0120] S104. Perform post-rolling cooling on the wire rod after rolling. During the cooling process, collect data on the surface temperature of the wire rod and the flow rate of the cooling medium in segments and establish the correspondence between cooling rate and microstructure transformation.

[0121] S105. Perform microstructure testing and fatigue performance prediction and evaluation on the cooled wire rod, and feed the evaluation results back to the rolling parameter adjustment link to form a closed-loop control.

[0122] Example 2

[0123] Please refer to Figure 2 As shown, the refined spring steel continuous casting billet is heated, and the heating temperature data is obtained and compared with the target austenitizing temperature range. Specifically, this includes:

[0124] S201. Before the continuously cast billet is fed into the heating furnace, its surface defects are scanned and recorded, and the defect location information is associated and mapped with the deformation path of the subsequent rolling passes.

[0125] S202. Place the continuously cast billet in the heating furnace in sections. Arrange temperature acquisition points in the preheating section, heating section and soaking section of the heating furnace respectively, and continuously record the temperature value of each acquisition point at a preset time interval.

[0126] S203. Calculate the variance of temperature values ​​at different spatial locations on the same time cross section. When the variance exceeds a preset uniformity threshold, mark the cross section as a temperature non-uniform region and record its location coordinates. The calculation method for the temperature uniformity variance of the same cross section is as follows:

[0127] ;

[0128] in, The variance of temperature uniformity across the same cross section. This represents the total number of temperature sampling points on this cross-section. For the first Temperature measurements at each collection point This is the arithmetic mean of the temperatures at all sampling points across this cross section;

[0129] S204. The coordinates of the temperature non-uniformity regions of each section are superimposed and compared with the original segregation spectrum data of the continuously cast billet to identify the overlap range between the severely segregated regions and the temperature non-uniformity regions. The quantitative calculation method for the overlap degree is as follows:

[0130] ;

[0131] in, This represents the percentage of overlap between the segregation region and the temperature non-uniformity region. This represents the spatial overlap area between the two types of regions. The area is the union of the two types of regions;

[0132] S205. Determine the extension of the holding time of the heating homogenization section based on the size and distribution density of the overlapping area, and send the adjusted heating system parameters to the heating furnace control terminal for execution.

[0133] Specifically, the core principle of this embodiment is to quantitatively evaluate the uniformity of the spatial temperature field, and combine it with the spatial distribution characteristics of the original segregation of the continuously cast billet to identify the weak areas where segregation and temperature inhomogeneity overlap. The homogenization and heat preservation time is extended in a directional manner for the overlapping areas to eliminate the negative impact of the superposition effect of "segregation + temperature inhomogeneity" on the austenitization uniformity from the heating source, so as to provide a homogenized initial structure for subsequent rolling deformation and reduce the dispersion of the fatigue performance of the finished wire rod.

[0134] Example 3

[0135] Please refer to Figure 3 As shown, the heated continuously cast billet undergoes multi-pass high-speed rolling. During each pass, rolling force, rolling temperature, workpiece cross-sectional dimensions, and rolling speed data are collected simultaneously. The cumulative deformation is statistically analyzed, and the temperature drop trend between passes is analyzed. Specifically, this includes:

[0136] S301. Install temperature sensors on the inlet and outlet sides of each mill stand, calculate the difference between the inlet and outlet temperatures for the same pass, and obtain the absolute temperature drop value for that pass.

[0137] S302. Arrange the absolute temperature drop values ​​of each pass according to the rolling sequence to form a temperature drop sequence. Perform a first-order difference operation on this sequence to obtain the temperature drop rate between adjacent passes, where the... The calculation method for the temperature drop rate of each pass is as follows:

[0138] ;

[0139] in, For the first The order of the path relative to the first The rate of change of temperature drop per pass and The first The inlet and outlet temperatures of each pass. and The first Dao Ci and Di The rolling time of each pass;

[0140] S303. When the ratio of the temperature drop rate of a certain pass to the temperature drop rate of the previous pass is greater than the preset mutation coefficient, the pass is determined to be an abnormal temperature drop pass, and the rolling force data corresponding to the pass is compared with the rolling force data of the normal pass to detect outliers and remove abnormal rolling force data points caused by equipment fluctuations.

[0141] S304. The actual deformation amounts of each pass after removing outliers are summed to obtain the total cumulative deformation. The calculation methods for the area reduction rate of each pass and the total cumulative deformation are as follows:

[0142] ;

[0143] ;

[0144] in, For the first The area reduction rate of each lane, and The first Cross-sectional area of ​​the rolled piece before and after each rolling pass The total cumulative area reduction deformation. Total number of rolling passes;

[0145] S305. Calculate the deviation between the total cumulative deformation and the target total deformation. When the deviation exceeds the allowable range, back-calculate and distribute the deviation value according to the original reduction ratio of each pass to obtain the correction reduction amount that needs to be added or reduced for each pass. The calculation method for the correction reduction per pass is as follows:

[0146] ;

[0147] in, For the first The adjustment value for the amount of pressure that needs to be added or reduced per pass. To reduce the deformation amount to the target total area, For the first Height of the rolled piece before each rolling pass.

[0148] Specifically, the core principle of this embodiment is to identify abnormal temperature drop passes by using the temperature drop change rate of each pass, and to eliminate equipment interference data by detecting outliers in the rolling force to obtain the true deformation amount of each pass. On this basis, the accurate total cumulative deformation amount is obtained by accumulating the area reduction rate, and the correction amount is allocated by back-calculating according to the original reduction ratio. While ensuring that the total deformation amount accurately meets the standard, the coordination of deformation of each pass is maintained, and the unevenness of the microstructure and fatigue risks caused by sudden changes in local deformation are avoided.

[0149] Example 4

[0150] Please refer to Figure 4 As shown, the cumulative deformation statistics and temperature drop trend analysis between multiple passes also include:

[0151] S401. Calculate the area reduction rate by comparing the cross-sectional dimensions of the rolled piece after each pass with the cross-sectional dimensions before that pass, and obtain the single-pass area reduction rate for each pass.

[0152] S402. Plot the area reduction rate of each single pass in the rolling sequence as an area reduction rate distribution curve, identify the peak value of the curve, and locate the pass with the largest area reduction rate as the key deformation pass.

[0153] S403. Perform frequency domain analysis on the rolling force data of key deformation passes, extract the dominant frequency component from the rolling force signal, and compare the dominant frequency component with the natural vibration frequency of the workpiece in that pass. The estimation method for the natural vibration frequency is as follows:

[0154] ;

[0155] in, The natural vibration frequency of the rolled piece in this pass. This represents the elastic modulus of spring steel at this temperature. The density of spring steel, This is the effective vibration length of the rolled piece in that pass;

[0156] S404. When the ratio of the main frequency component to the natural vibration frequency falls into the resonance range, the rolling speed of this pass is reduced by a preset ratio, and the parameters after the speed reduction are rewritten into the rolling procedure of this pass.

[0157] S405. Simultaneously, perform local gradient calculation on the cumulative deformation of the two passes before and after the key deformation pass, and compare the local gradient value with the empirical threshold of the critical deformation amount for dynamic recrystallization to determine whether the key deformation pass meets the conditions for sufficient dynamic recrystallization. The calculation method for the local deformation gradient is as follows:

[0158] ;

[0159] in, This represents the local deformation gradient two passes before and after the critical deformation pass. and These represent the area reduction rates for the second and first two passes after the critical deformation pass, respectively. This represents the average time interval between adjacent passes.

[0160] Specifically, the core principle of this embodiment is to locate the key deformation pass by the peak of the area reduction rate. On the one hand, the rolling force signal is analyzed from the frequency domain to eliminate uneven deformation and surface micro-defects by avoiding resonance of the rolled piece. On the other hand, by matching and judging the local deformation gradient and the critical deformation amount of dynamic recrystallization, the key pass is guaranteed to undergo sufficient dynamic recrystallization, thereby achieving grain refinement and deformation texture elimination, and improving the fatigue resistance of the wire rod from the source of the structure.

[0161] Example 5

[0162] Please refer to Figure 5 As shown, the distribution of reduction in each pass is dynamically corrected based on the analysis results of cumulative deformation and temperature drop in each pass. Specifically, this includes:

[0163] S501. The target value of total cumulative deformation is distributed non-uniformly according to the rolling sequence from roughing to finishing, so that the proportion of cumulative deformation in the roughing stage to the total deformation is greater than that in the finishing stage.

[0164] S502. Estimate the theoretical rolling force corresponding to the corrected reduction for each pass, compare the estimated value with the actual rolling force collected for that pass in real time, and calculate the deviation rate between the two. The calculation method for the rolling force deviation rate of each pass is as follows:

[0165] ;

[0166] in, For the first Rolling force deviation rate of each pass For the first The actual rolling force collected per pass. For the first The theoretical rolling force is estimated based on the corrected reduction amount per pass;

[0167] S503. When the deviation rate of a certain pass exceeds three consecutive acquisition cycles and the deviation direction is consistent, the reduction correction value of that pass is locked and frozen, and no further adjustment is made. At the same time, the locking information of that pass is transmitted to subsequent passes as a constraint condition.

[0168] S504. After the reduction amount of all passes is corrected, the corrected reduction amount sequence of each pass is summed and checked to ensure that the deviation between the corrected total deformation amount and the target total deformation amount is within the allowable range.

[0169] S505. Compare the verified reduction sequence with the inclusion distribution information in the original internal quality inspection data of the continuously cast billet. If the deformation area of ​​a certain pass happens to cover the dense inclusion area, then add a superimposed deformation amount to the reduction amount of that pass on the original correction value to ensure that the inclusions are fully broken up in that pass. The density determination of the dense inclusion area and the calculation method of the superimposed deformation amount are as follows:

[0170] ;

[0171] ;

[0172] in, The number density of inclusions within a certain deformation region. This refers to the number of inclusions counted in this area. The volume of this region, Thresholds are determined for areas with dense inclusions. For the first The additional reduction in pressure applied during each pass due to the coverage of areas with dense inclusions. This is the correction factor for inclusion breakage. For the first Height of the rolled piece before each rolling pass.

[0173] Specifically, the core principle of this embodiment is to adopt a non-uniform rolling reduction distribution strategy of "large deformation in rough rolling and small deformation in finish rolling". The rolling force deviation rate is continuously used to determine and lock stable passes to ensure the stability of the rolling process. At the same time, based on the spatial distribution information of inclusions, the deformation is increased in a directional manner in areas with dense inclusions to break up the inclusions, reduce the source of fatigue crack initiation, and simultaneously take into account the stability of rolling production and the improvement of fatigue performance.

[0174] Example 6

[0175] Post-rolling cooling is performed on the wire rod after rolling. During the cooling process, data on the surface temperature of the wire rod and the flow rate of the cooling medium are collected in segments, and a correlation between cooling rate and microstructure transformation is established. Specifically, this includes:

[0176] The post-rolling cooling zone is divided into three temperature ranges: high temperature, medium temperature, and low temperature. A set of temperature acquisition points is set at the beginning and end of each temperature range, and the surface temperature value of the wire is collected at fixed time intervals.

[0177] The average cooling rate of a temperature range is obtained by dividing the temperature difference between the starting and ending points of the same temperature range by the spatial length of that range. The average cooling rate over the temperature range is calculated as follows:

[0178] ;

[0179] in, For the first Average cooling rate over the temperature range and The first The surface temperature of the wire at the beginning and end of the interval. For the first The spatial length of the temperature range; compare the average cooling rate of each temperature range with the critical cooling rate of the phase change corresponding to that temperature range to determine whether the wire avoids the pearlite transformation nose temperature region in each range.

[0180] Time series smoothing was performed on the cooling rate data in the medium temperature range to remove temperature fluctuation noise caused by uneven cooling medium spraying, resulting in a smoothed cooling rate curve.

[0181] The smoothed cooling rate curve is superimposed with the pre-established continuous cooling transformation diagram of the steel grade to determine the actual microstructure transformation product type and proportion of the wire in each temperature range.

[0182] The proportion of microstructure transformation products in each interval is stored in a one-to-one correspondence with the corresponding cooling rate value to form a cooling rate-microstructure transformation data table for this batch of wires.

[0183] Specifically, the core principle of this embodiment is to accurately collect segmented cooling rates, combine them with the continuous cooling transformation diagram of the steel grade to quantitatively match the microstructure transformation products in each temperature range, and avoid the pearlite transformation nose region by controlling the cooling rate to obtain a uniform and fine sorbite microstructure; at the same time, a one-to-one correspondence between cooling parameters and microstructure results is established to provide data support for the precise control of subsequent cooling processes.

[0184] Example 7

[0185] Establishing the correlation between cooling rate and tissue transformation also includes:

[0186] The cooling rate-microstructure transformation data of different batches of wire rod of the same steel grade under the same cooling regime were statistically summarized, and the inter-batch standard deviation of cooling rate for each temperature range was calculated. The standard deviation of the batch-to-batch cooling rate within the temperature range is calculated as follows:

[0187] ;

[0188] in, For the first Batch-to-batch standard deviation of cooling rate within temperature range To count the total number of batches, For the first Batch No. Average cooling rate over the temperature range For all batches in the first The average cooling rate over the temperature range;

[0189] When the standard deviation of the cooling rate between batches in a certain temperature range exceeds the preset stability threshold, the cooling medium injection parameters in that temperature range are traced and analyzed to find the specific injection port number that caused the cooling rate fluctuation.

[0190] The historical maintenance record data corresponding to the nozzle number is correlated with the current flow attenuation data of the nozzle to determine whether the nozzle needs to be replaced or cleaned.

[0191] After replacement or cleaning, the cooling rate data for this temperature range is re-collected, the new data is merged and updated with the previous corresponding data table, and the updated data table is checked for consistency to ensure that the deviation of the proportion of tissue transformation products between the old and new data is within the allowable range.

[0192] The finalized cooling rate-microstructure transformation relationship will be used as the benchmark for cooling in subsequent batches of this steel grade.

[0193] Specifically, the core principle of this embodiment is to evaluate the operational stability of the cooling system by using the standard deviation of cooling rates between batches, to trace and locate the faulty injection port and to complete the status determination in conjunction with maintenance records; and to continuously optimize the cooling rate-microstructure transformation benchmark library through data iteration updates and consistency verification, so as to ensure the consistency of cooling microstructure of different batches of wire and reduce the batch-to-batch fluctuations in fatigue performance.

[0194] Example 8

[0195] The cooled wire was subjected to microstructure analysis and fatigue performance prediction and evaluation, specifically including:

[0196] Samples were taken from the cooled wire at the specified locations, and the metallographic structure of the samples was observed to obtain data on the original austenite grain size, pearlite lamellar spacing, and carbide distribution density.

[0197] Correlation analysis was performed between the original austenite grain size data and the final rolling temperature and post-rolling cooling rate data recorded during the rolling process of this batch, establishing a three-dimensional mapping relationship between final rolling temperature, cooling rate, and grain size. The correlation coefficient was calculated as follows:

[0198] ;

[0199] in, This represents the correlation coefficient between final rolling temperature, cooling rate, and grain size. For the first The final rolling temperature of each sample, For the first Post-rolling cooling rate of each sample For the first The original austenite grain size of each sample , , Each is their respective average value. The total number of samples;

[0200] Rotational bending fatigue tests were conducted on the sampled wires, and data on the number of fracture cycles under various stress levels were collected. The data were arranged from high to low stress levels, and linear regression fitting was performed on the high-stress and low-stress regions respectively to obtain the conditional fatigue limit values ​​corresponding to the two fitted lines and their intersection points. The conditional fatigue limit was calculated as follows:

[0201] ;

[0202] in, The conditional fatigue limit, and These represent the slope and intercept of the linear regression fitted line in the high-stress region, respectively. and These are the slope and intercept of the linear regression fitted line in the low-stress area, respectively.

[0203] Multi-parameter regression analysis was performed on the conditional fatigue limit value and the tensile strength, reduction of area and hardness data of this batch of wire to establish a predictive model between the fatigue limit and mechanical performance parameters.

[0204] The residuals between the output values ​​of the prediction model and the actual fatigue test values ​​are calculated. When the absolute value of the residuals exceeds the preset accuracy requirements, the deviation information of that batch is marked and included in the model correction dataset.

[0205] Specifically, the core principle of this embodiment is to establish a three-dimensional quantitative correlation between final rolling temperature, cooling rate, and grain size. By segmented linear fitting between high and low stress, the conditional fatigue limit is accurately determined, and a multi-parameter prediction model of conventional mechanical properties and fatigue limit is constructed. At the same time, the dataset is continuously expanded and corrected through residual judgment, and the prediction accuracy of the model is iteratively improved, so as to achieve quantitative evaluation and rapid prediction of fatigue performance.

[0206] Example 9

[0207] The evaluation results are fed back to the rolling parameter adjustment process to form a closed-loop control, specifically including:

[0208] The residual data of the fatigue performance prediction model for each batch of wire are arranged in chronological order to form a residual trend sequence.

[0209] A sliding window mean is calculated for the residual trend sequence, with the window size set to five consecutive batches. When the sliding mean shows a continuous upward or downward trend and the trend slope exceeds a preset drift threshold, the rolling parameters are corrected. The sliding window mean is calculated as follows:

[0210] ;

[0211] in, For the first The mean of a sliding window with batch as the endpoint. The window size is 5. For the first Batch prediction model residuals;

[0212] When a global correction is triggered, the cooling rate-microstructure transformation data table for the current batch, the reduction correction records for each pass, and the final rolling temperature data are packaged as the correction input dataset.

[0213] The sensitivity coefficients of each parameter in the corrected input dataset to the fatigue performance prediction model are ranked, and the top three parameters with the highest sensitivity coefficients are selected as priority adjustment parameters, where the first... The sensitivity coefficients of each parameter are calculated as follows:

[0214] ;

[0215] in, For the first The sensitivity coefficients of each parameter to the conditional fatigue limit For the first The current values ​​of the parameters, This represents a small perturbation to the parameter;

[0216] The current value of the priority adjustment parameter is compared with the historical best value to calculate the adjustment range and direction of each parameter and generate a parameter adjustment plan.

[0217] The parameter adjustment plan is issued to the next batch of rolling execution, and the corrected input dataset and adjustment plan for this batch are archived together for use in trend analysis of subsequent batches.

[0218] Specifically, the core principle of this embodiment is to identify systematic drift in fatigue performance prediction and trigger global parameter correction based on multi-batch residual sliding trend monitoring; to lock the core parameters that have the greatest impact on fatigue performance by sorting by sensitivity coefficients, and to make directional adjustments with historical best values ​​as a reference, thereby achieving closed-loop iterative optimization of parameters throughout the entire process with fatigue performance as the goal.

[0219] Example 10

[0220] The method also includes archiving and traceability of data throughout the entire process, specifically including:

[0221] Temperature data, rolling force data, deformation data, cooling rate data, microstructure testing data, and fatigue test data collected throughout the entire process from heating the continuous casting billet to fatigue performance evaluation for each batch of wire rod are stored in association with the batch number to form a complete process data archive for that batch.

[0222] A horizontal comparison of the entire process data archives of different batches of the same steel grade was conducted to extract the batch-to-batch fluctuation range of each key parameter, and a baseline for the fluctuation of key parameters for that steel grade was established. The calculation method for the upper and lower limits of the fluctuation baseline of the key parameters is as follows:

[0223] ;

[0224] in, and The first The lower and upper limits of the fluctuation baseline for each key parameter. This is the mean of the parameter across all batches. The standard deviation of this parameter. and These are the confidence coefficients for the lower and upper limits, respectively.

[0225] When a key parameter of a batch deviates from the fluctuation baseline by more than a preset abnormal multiple, the batch is automatically marked as a batch of key concern, and its entire process data file is retrieved for item-by-item backtracking to locate the first abnormal link that caused the deviation.

[0226] Fine-grained analysis was performed on the data of the first abnormal step identified, and the distribution of the data of this step was compared with that of the normal batch of data in the same step to identify the specific characteristics of the abnormal data points.

[0227] A causal correlation analysis is performed between the description of abnormal characteristics and the recorded operational parameters of the abnormal process to form an abnormal cause determination report. The conclusions in the determination report are used as the basis for revising the subsequent operating procedures for that process.

[0228] Specifically, the core principle of this embodiment is to build a batch-level traceability system by associating and archiving data throughout the entire process, and to establish a baseline for key parameter fluctuations based on historical batch big data. By automatically triggering full-process backtracking analysis through abnormal deviations, the source of the abnormality is accurately located and feedback is fed back to revise the operating procedures, thereby achieving continuous self-improvement of the process system and long-term improvement of product quality stability.

[0229] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A rolling process for improving the fatigue performance of high-speed spring steel wire rod, characterized in that, include: The spring steel continuous casting billet after smelting and refining is heated, and the heating temperature data is obtained and compared and analyzed with the target austenitizing temperature range. The heated continuous casting billet is subjected to high-speed rolling in multiple passes. During each rolling pass, rolling force, rolling temperature, cross-sectional dimensions of the rolled piece and rolling speed data are collected simultaneously. The cumulative deformation of the data from multiple passes is statistically analyzed and the temperature drop trend between passes is analyzed. The distribution of reduction for each pass is dynamically corrected based on the analysis results of cumulative deformation and temperature drop of each pass. The wire rod after rolling is cooled after rolling. During the cooling process, the surface temperature of the wire rod and the flow rate of the cooling medium are collected in segments, and the correspondence between cooling rate and microstructure transformation is established. The cooled wire is subjected to microstructure testing and fatigue performance prediction and evaluation, and the evaluation results are fed back to the rolling parameter adjustment process to form a closed-loop control.

2. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 1, characterized in that, The process of heating the refined spring steel continuous casting billet, obtaining heating temperature data, and comparing and analyzing it with the target austenitizing temperature range specifically includes: Before the continuously cast billet is fed into the heating furnace, its surface defects are scanned and recorded, and the defect location information is associated with the deformation path of subsequent rolling passes. The continuously cast billet is placed in the heating furnace in sections. Temperature acquisition points are set up in the preheating section, heating section and soaking section of the heating furnace respectively, and the temperature value of each acquisition point is continuously recorded at a preset time interval. The variance of temperature values ​​at different spatial locations on the same time section is calculated. When the variance exceeds a preset uniformity threshold, the section is marked as a temperature non-uniform region and its location coordinates are recorded. The calculation method for the temperature uniformity variance of the same section is as follows: ； in, The variance of temperature uniformity across the same cross section. This represents the total number of temperature sampling points on this cross-section. For the first Temperature measurements at each collection point This is the arithmetic mean of the temperatures at all sampling points across this cross section; The coordinates of the temperature non-uniformity regions at each cross section are superimposed and compared with the original segregation spectrum data of the continuously cast billet to identify the overlap range between the severely segregated regions and the temperature non-uniformity regions. The quantification method for the overlap degree is as follows: ； in, This represents the percentage of overlap between the segregation region and the temperature non-uniformity region. This represents the spatial overlap area between the two types of regions. The area is the union of the two types of regions; The extension of the holding time in the heating homogenization zone is determined based on the size and distribution density of the overlapping area, and the adjusted heating regime parameters are sent to the heating furnace control terminal for execution.

3. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 2, characterized in that, The process involves multi-pass high-speed rolling of the heated continuously cast billet. During each pass, rolling force, rolling temperature, workpiece cross-sectional dimensions, and rolling speed data are collected simultaneously. The cumulative deformation is statistically analyzed, and the temperature drop trend between passes is analyzed. Specifically, this includes: Temperature sensors are installed on the inlet and outlet sides of each mill stand. The difference between the inlet and outlet temperatures of the same pass is calculated to obtain the absolute temperature drop value of that pass. The absolute temperature drop values ​​of each pass are arranged in the rolling sequence to form a temperature drop sequence. A first-order difference operation is performed on this sequence to obtain the rate of temperature drop change between adjacent passes, where the... The calculation method for the temperature drop rate of each pass is as follows: ； in, For the first The order of the path relative to the first The rate of change of temperature drop per pass and The first The inlet and outlet temperatures of each pass. and The first Dao Ci and Di The rolling time of each pass; When the ratio of the temperature drop rate of a certain pass to the temperature drop rate of the previous pass is greater than the preset mutation coefficient, the pass is determined to be an abnormal temperature drop pass, and the rolling force data corresponding to the pass is compared with the rolling force data of the normal pass to detect outliers and remove abnormal rolling force data points caused by equipment fluctuations. The total cumulative deformation is obtained by summing the actual deformation amounts of each pass after removing outliers. The calculation methods for the area reduction rate of each pass and the total cumulative deformation are as follows: ； ； in, For the first The area reduction rate of each lane, and The first Cross-sectional area of ​​the rolled piece before and after each rolling pass The total cumulative area reduction deformation. Total number of rolling passes; The deviation between the total cumulative deformation and the target total deformation is calculated. When the deviation exceeds the allowable range, the deviation value is back-calculated and allocated according to the original reduction ratio of each pass to obtain the correction reduction amount that needs to be added or reduced for each pass. The calculation method for the correction reduction per pass is as follows: ； in, For the first The adjustment value for the amount of pressure that needs to be added or reduced per pass. To reduce the deformation amount to the target total area, For the first Height of the rolled piece before each rolling pass.

4. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 3, characterized in that, The method of performing cumulative deformation statistics and inter-pass temperature drop trend analysis on multi-pass data also includes: The area reduction rate is calculated by comparing the cross-sectional dimensions of the rolled piece after each pass with the cross-sectional dimensions before that pass. The area reduction rate of each single pass is plotted as an area reduction rate distribution curve according to the rolling sequence. The peak value of the curve is identified, and the pass with the largest area reduction rate is identified as the critical deformation pass. Frequency domain analysis was performed on the rolling force data of key deformation passes to extract the dominant frequency component from the rolling force signal. This dominant frequency component was then compared with the natural vibration frequency of the workpiece in that pass. The natural vibration frequency was estimated as follows: ； in, The natural vibration frequency of the rolled piece in this pass. This represents the elastic modulus of spring steel at this temperature. The density of spring steel, This is the effective vibration length of the rolled piece in that pass; When the ratio of the main frequency component to the natural vibration frequency falls into the resonance range, the rolling speed of this pass is reduced by a preset ratio, and the parameters after the speed reduction are rewritten into the rolling procedure of this pass. Simultaneously, local gradient calculations are performed on the cumulative deformation of the two passes before and after the key deformation pass. The local gradient values ​​are compared with the empirical threshold of the critical deformation amount for dynamic recrystallization to determine whether the key deformation pass meets the conditions for sufficient dynamic recrystallization. The calculation method for the local deformation gradient is as follows: ； in, This represents the local deformation gradient two passes before and after the critical deformation pass. and These represent the area reduction rates for the second and first two passes after the critical deformation pass, respectively. This represents the average time interval between adjacent passes.

5. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 4, characterized in that, The dynamic correction of the reduction distribution for each pass based on the cumulative deformation and temperature drop analysis results specifically includes: The target value of total cumulative deformation is distributed non-uniformly according to the rolling sequence from roughing to finishing, so that the proportion of cumulative deformation in the roughing stage is greater than that in the finishing stage. For each pass, the theoretical rolling force corresponding to the corrected reduction is estimated. The estimated value is then compared in real time with the actual rolling force collected for that pass, and the deviation rate between the two is calculated. The calculation method for the rolling force deviation rate of each pass is as follows: ； in, For the first Rolling force deviation rate of each pass For the first The actual rolling force collected per pass. For the first The theoretical rolling force is estimated based on the corrected reduction amount per pass; When the deviation rate of a certain pass exceeds three consecutive acquisition cycles and the deviation direction is consistent, the reduction correction value of that pass is locked and frozen, and no further adjustment is made. At the same time, the locking information of that pass is passed to subsequent passes as a constraint condition. After the reduction amount of all passes is corrected, the corrected reduction amount sequence of each pass is summed and verified to ensure that the deviation between the corrected total deformation amount and the target total deformation amount is within the allowable range. The verified reduction sequence is compared with the inclusion distribution information in the original internal quality inspection data of the continuously cast billet. If the deformation area of ​​a certain pass happens to cover the dense inclusion area, the reduction of that pass is supplemented with an additional deformation amount on the basis of the original correction value, so that the inclusions are fully broken up in that pass. The density determination of the dense inclusion area and the calculation method of the additional deformation amount are as follows: ； ； in, The number density of inclusions within a certain deformation region. This refers to the number of inclusions counted in this area. The volume of this region, Thresholds are determined for areas with dense inclusions. For the first The additional reduction in pressure applied during each pass due to the coverage of areas with dense inclusions. This is the correction factor for inclusion breakage. For the first Height of the rolled piece before each rolling pass.

6. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 5, characterized in that, The post-rolling cooling of the wire rod involves collecting data on the surface temperature of the wire rod and the flow rate of the cooling medium in segments during the cooling process, and establishing a correlation between cooling rate and microstructure transformation. Specifically, this includes: The post-rolling cooling zone is divided into three temperature ranges: high temperature, medium temperature, and low temperature. A set of temperature acquisition points is set at the beginning and end of each temperature range, and the surface temperature value of the wire is collected at fixed time intervals. The average cooling rate of a temperature range is obtained by dividing the temperature difference between the starting and ending points of the same temperature range by the spatial length of that range. The average cooling rate over the temperature range is calculated as follows: ； in, For the first Average cooling rate over the temperature range and The first The surface temperature of the wire at the beginning and end of the interval. For the first The spatial length of the temperature range; compare the average cooling rate of each temperature range with the critical cooling rate of the phase change corresponding to that temperature range to determine whether the wire avoids the pearlite transformation nose temperature region in each range. Time series smoothing was performed on the cooling rate data in the medium temperature range to remove temperature fluctuation noise caused by uneven cooling medium spraying, resulting in a smoothed cooling rate curve. The smoothed cooling rate curve is superimposed with the pre-established continuous cooling transformation diagram of the steel grade to determine the actual microstructure transformation product type and proportion of the wire in each temperature range. The proportion of microstructure transformation products in each interval is stored in a one-to-one correspondence with the corresponding cooling rate value to form a cooling rate-microstructure transformation data table for this batch of wires.

7. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 6, characterized in that, The establishment of the correspondence between cooling rate and tissue transformation also includes: The cooling rate-microstructure transformation data of different batches of wire rod of the same steel grade under the same cooling regime were statistically summarized, and the inter-batch standard deviation of cooling rate for each temperature range was calculated. The standard deviation of the batch-to-batch cooling rate within the temperature range is calculated as follows: ； in, For the first Batch-to-batch standard deviation of cooling rate within temperature range To count the total number of batches, For the first Batch No. Average cooling rate over the temperature range For all batches in the first The average cooling rate over the temperature range; When the standard deviation of the cooling rate between batches in a certain temperature range exceeds the preset stability threshold, the cooling medium injection parameters in that temperature range are traced and analyzed to find the specific injection port number that caused the cooling rate fluctuation. The historical maintenance record data corresponding to the nozzle number is correlated with the current flow attenuation data of the nozzle to determine whether the nozzle needs to be replaced or cleaned. After replacement or cleaning, the cooling rate data for this temperature range is re-collected, the new data is merged and updated with the previous corresponding data table, and the updated data table is checked for consistency to ensure that the deviation of the proportion of tissue transformation products between the old and new data is within the allowable range. The finalized cooling rate-microstructure transformation relationship will be used as the benchmark for cooling in subsequent batches of this steel grade.

8. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 7, characterized in that, The process of performing microstructure analysis and fatigue performance prediction and evaluation on the cooled wire specifically includes: Samples were taken from the cooled wire at the specified locations, and the metallographic structure of the samples was observed to obtain data on the original austenite grain size, pearlite lamellar spacing, and carbide distribution density. Correlation analysis was performed between the original austenite grain size data and the final rolling temperature and post-rolling cooling rate data recorded during the rolling process of this batch, establishing a three-dimensional mapping relationship between final rolling temperature, cooling rate, and grain size. The correlation coefficient was calculated as follows: ； in, This represents the correlation coefficient between final rolling temperature, cooling rate, and grain size. For the first The final rolling temperature of each sample, For the first Post-rolling cooling rate of each sample For the first The original austenite grain size of each sample , , Each is their respective average value. The total number of samples; Rotational bending fatigue tests were conducted on the sampled wires, and data on the number of fracture cycles under various stress levels were collected. The data were arranged from high to low stress levels, and linear regression fitting was performed on the high-stress and low-stress regions respectively to obtain the conditional fatigue limit values ​​corresponding to the two fitted lines and their intersection points. The conditional fatigue limit was calculated as follows: ； in, The conditional fatigue limit, and These represent the slope and intercept of the linear regression fitted line in the high-stress region, respectively. and These are the slope and intercept of the linear regression fitted line in the low-stress area, respectively. Multi-parameter regression analysis was performed on the conditional fatigue limit value and the tensile strength, reduction of area and hardness data of this batch of wire to establish a predictive model between the fatigue limit and mechanical performance parameters. The residuals between the output values ​​of the prediction model and the actual fatigue test values ​​are calculated. When the absolute value of the residuals exceeds the preset accuracy requirements, the deviation information of that batch is marked and included in the model correction dataset.

9. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 8, characterized in that, The process of feeding back the evaluation results to the rolling parameter adjustment stage to form a closed-loop control specifically includes: The residual data of the fatigue performance prediction model for each batch of wire are arranged in chronological order to form a residual trend sequence. A sliding window mean is calculated for the residual trend sequence, with the window size set to five consecutive batches. When the sliding mean shows a continuous upward or downward trend and the trend slope exceeds a preset drift threshold, the rolling parameters are corrected. The sliding window mean is calculated as follows: ； in, For the first The mean of a sliding window with batch as the endpoint. The window size is 5. For the first Batch prediction model residuals; When a global correction is triggered, the cooling rate-microstructure transformation data table for the current batch, the reduction correction records for each pass, and the final rolling temperature data are packaged as the correction input dataset. The sensitivity coefficients of each parameter in the corrected input dataset to the fatigue performance prediction model are ranked, and the top three parameters with the highest sensitivity coefficients are selected as priority adjustment parameters, where the first parameter is... The sensitivity coefficients of each parameter are calculated as follows: ； in, For the first The sensitivity coefficients of each parameter to the conditional fatigue limit For the first The current values ​​of the parameters, This represents a small perturbation to the parameter; The current value of the priority adjustment parameter is compared with the historical best value to calculate the adjustment range and direction of each parameter and generate a parameter adjustment plan. The parameter adjustment plan is issued to the next batch of rolling execution, and the corrected input dataset and adjustment plan for this batch are archived together for use in trend analysis of subsequent batches.

10. The rolling process for improving the fatigue performance of high-speed spring steel wire rod according to claim 9, characterized in that, The method also includes archiving and traceability processing of the entire process data, specifically including: Temperature data, rolling force data, deformation data, cooling rate data, microstructure testing data, and fatigue test data collected throughout the entire process from heating the continuous casting billet to fatigue performance evaluation for each batch of wire rod are stored in association with the batch number to form a complete process data archive for that batch. A horizontal comparison of the entire process data archives of different batches of the same steel grade was conducted to extract the batch-to-batch fluctuation range of each key parameter, and a baseline for the fluctuation of key parameters for that steel grade was established. The calculation method for the upper and lower limits of the fluctuation baseline of the key parameters is as follows: ； in, and The first The lower and upper limits of the fluctuation baseline for each key parameter. This is the mean of the parameter across all batches. The standard deviation of this parameter. and These are the confidence coefficients for the lower and upper limits, respectively. When a key parameter of a batch deviates from the fluctuation baseline by more than a preset abnormal multiple, the batch is automatically marked as a batch of key concern, and its entire process data file is retrieved for item-by-item backtracking to locate the first abnormal link that caused the deviation. Fine-grained analysis was performed on the data of the first abnormal step identified, and the distribution of the data of this step was compared with that of the normal batch of data in the same step to identify the specific characteristics of the abnormal data points. A causal correlation analysis is performed between the description of abnormal characteristics and the recorded operational parameters of the abnormal process to form an abnormal cause determination report. The conclusions in the determination report are used as the basis for revising the subsequent operating procedures for that process.

Images