A method for monitoring the heating of stainless steel seamless pipes in a heat treatment furnace

Abstract

This invention discloses a method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace, belonging to the field of metal heat treatment technology. Specifically, it includes: real-time acquisition of electromagnetic induction signals from the rotating steel tube and binding them with spatiotemporal information; kinematic inversion of the spatiotemporal information based on the tube's velocity to reconstruct the electromagnetic parameter evolution trajectory of each physical point on the surface; identification of continuously deviating abnormal points based on the evolution trajectory and clustering them into heating abnormal regions; extraction of electromagnetic features of the heating abnormal regions; obtaining temperature deviations through a preset electromagnetic-temperature mapping model and generating compensation commands; prediction of the spatiotemporal coordinates of the abnormal regions reaching the next heating zone, driving the corresponding independent heating unit to perform dynamic energy compensation within a matched time window. This invention achieves real-time, global, and precise perception and closed-loop control of the heating uniformity of moving workpieces, significantly improving the uniformity and stability of heat treatment quality and reducing energy consumption.

Description

Technical Field

[0001] This invention relates to the field of metal heat treatment technology, and more specifically to a method for monitoring the heating of seamless stainless steel tubes inside a heat treatment furnace. Background Technology

[0002] Heat treatment is a crucial process in the production of seamless stainless steel tubes. The uniformity and precision of the heating process directly determine the final microstructure, residual stress level, and dimensional accuracy of the product. Traditional heat treatment furnaces typically employ a zoned temperature control strategy and rely on several thermocouples fixed in position within the furnace for temperature monitoring. However, for a steel tube that continuously rotates and advances within the furnace, the thermal history experienced by different points on its surface varies dynamically. Fixed furnace temperature monitoring cannot accurately reflect the steel tube itself, especially the real-time temperature field distribution in three-dimensional space on its surface and near-surface areas.

[0003] Currently, some technologies have been explored for temperature monitoring of moving workpieces. For example, non-contact infrared thermal imagers are used to scan and measure the temperature of local surfaces, or fixed sensors installed at specific locations are used to perform intermittent measurements on workpieces passing through their field of view. However, firstly, methods such as infrared thermal imaging are easily affected by the high-temperature radiation background, flue gas, and viewing angle within the furnace, making it difficult to achieve stable and high-precision full-field measurements in complex furnace environments. Secondly, intermittent measurements based on fixed points cannot construct a continuous and complete temperature evolution trajectory on the steel pipe surface, failing to accurately identify localized, small, or moving areas of abnormal heating. Thirdly, most monitoring methods only remain at the "sensing" level, failing to form an efficient and precise closed-loop control link with the actuators of the heat treatment furnace, resulting in the inability to accurately compensate for identified temperature deviations in real time and with precise spatial positioning.

[0004] Therefore, in actual production, the control of heat treatment process has long been in a "black box" or "semi-blind" state, mainly relying on fixed heating curves and operating experience for rough adjustment, making it difficult to achieve dynamic and refined control based on the actual state of the workpiece, which restricts the further improvement of product quality. Summary of the Invention

[0005] The purpose of this invention is to provide a method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace, thereby solving the following technical problems:

[0006] In actual production, heat treatment process control has long been in a "black box" or "semi-blind" state, mainly relying on fixed heating curves and operational experience for rough adjustments, making it difficult to achieve dynamic and refined control based on the actual state of the workpiece.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] A method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace, comprising the following steps:

[0009] S1, the electromagnetic induction signal of the rotating steel pipe in the heat treatment furnace is collected in real time by the circumferential electromagnetic detection array, and the electromagnetic induction signal is synchronously bound with the corresponding timestamp and the coordinates of the detection array unit to obtain the original signal set;

[0010] S2, obtain the rotational angular velocity and forward linear velocity of the steel pipe, and perform kinematic inversion calculation on the original signal set to obtain the evolution trajectory of the electromagnetic parameters of any physical point on the surface of the steel pipe as a function of time.

[0011] S3, calculate the dynamic rate of change of each physical point based on its evolution trajectory, and compare the dynamic rate of change with the local background field composed of the rates of change of the physical points that are spatially adjacent to the physical point, to obtain the set of abnormal points whose rates of change continuously deviate from the local background field;

[0012] S4, perform spatial clustering on the abnormal point set to obtain the heating abnormal region, extract the electromagnetic parameters corresponding to each physical point constituting the heating abnormal region, and input the electromagnetic parameters into the preset electromagnetic-temperature mapping model to obtain the temperature compensation command.

[0013] S5. Based on the real-time rotation and forward speed of the steel pipe, predict the spatiotemporal coordinates of the heating anomaly area to reach the next heating section. Based on the spatiotemporal coordinates, determine the independent heating unit that needs to be started in the next heating section and, in conjunction with the temperature compensation command, drive the independent heating unit to perform dynamic energy compensation.

[0014] As a further aspect of the present invention: in S2, the specific process for obtaining the evolutionary trajectory is as follows:

[0015] The real-time rotational angular velocity and forward linear velocity of the steel pipe, the known radius, and the preset installation axis position inside the furnace are obtained; based on the rotational angular velocity, forward linear velocity, known radius, and preset installation axis position inside the furnace, the surface of the steel pipe is modeled as a cylindrical surface, and the physical points on the cylindrical surface are assigned labels consisting of initial circumferential angles and initial axial positions;

[0016] Based on the rotational angular velocity and the forward linear velocity, calculate the circumferential angular offset and axial position offset of the physical point from the initial mark over time, and determine its theoretical spatial coordinates relative to the fixed space of the furnace at any time.

[0017] For each sensing signal with a timestamp and spatial coordinates of the detection unit in the original signal set, the theoretical spatial coordinates are matched with the spatial coordinates of the detection unit to determine whether the theoretical spatial coordinates are located in the effective sensing area of ​​the detection unit at the time corresponding to the timestamp of the signal. When the determination is yes, the sensing signal is associated with the corresponding physical point as the state observation value of the physical point at that time.

[0018] After traversing all induced signals, a discrete state observation sequence ordered by timestamp is obtained for each physical point; the discrete state observation sequence is smoothed and curve-fitted to obtain the evolution trajectory of the electromagnetic parameters of the physical point as a function of time.

[0019] As a further aspect of the present invention: the specific process of obtaining the anomaly point set in S3 is as follows:

[0020] Calculate the instantaneous rate of change of the evolution trajectory of each physical point on the surface of the steel pipe and determine the spatially adjacent physical points of each physical point as a candidate neighbor set;

[0021] Calculate the statistical distribution of the instantaneous change rate of the candidate neighbor point set; set a preliminary judgment threshold based on the statistical distribution, and select points in the candidate neighbor point set whose instantaneous change rate exceeds the preliminary judgment threshold as candidate points, and remove the candidate points from the candidate neighbor point set to obtain a purified neighbor point set;

[0022] Calculate the statistical mean and statistical fluctuation range based on the instantaneous change rate of the purified neighbor point set; set a stability judgment threshold based on the statistical fluctuation range;

[0023] The difference between the instantaneous rate of change of each physical point and the statistical mean is calculated to obtain the purification deviation. Physical points whose purification deviation continuously exceeds the stability judgment threshold are identified as abnormal physical points, and the abnormal physical points are collected to form an abnormal point set.

[0024] As a further aspect of the present invention: in step S4, the specific construction process of the preset electromagnetic-temperature mapping model is as follows:

[0025] Multiple pipe section specimens with different characteristic temperature ranges are pre-acquired and heated in a controlled experimental thermal environment according to a preset heating curve. During the heating process, the circumferential electromagnetic detection array is used to synchronously collect the induction signals of each pipe section specimen at multiple characteristic temperature points, and a contact temperature measuring device is used to synchronously measure the actual temperature value of each pipe section specimen at the characteristic temperature point.

[0026] For each induced signal, demodulation and feature quantity calculation are performed to extract its signal amplitude change and phase shift, which are then used together as electromagnetic parameters. The electromagnetic parameters are used as independent variables, and the actual temperature value, which is strictly paired with the induced signal in time and space, is used as the dependent variable to form calibration data pairs. The least squares method is used to perform regression calculations on all the calibration data pairs to fit a continuous mapping function from the independent variable to the dependent variable. The continuous mapping function is defined as the preset electromagnetic-temperature mapping model.

[0027] As a further aspect of the present invention: the specific process of obtaining the temperature compensation command in step S4 is as follows:

[0028] Extract multiple electromagnetic characteristic parameter data corresponding to the heating abnormal area and calculate their average value to obtain the electromagnetic parameter characteristic value of the heating abnormal area; obtain multiple electromagnetic characteristic parameter data corresponding to other areas in the same axial section of the steel pipe as the heating abnormal area and calculate their average value to obtain the electromagnetic parameter reference value;

[0029] The electromagnetic parameter characteristic values ​​are input into the preset electromagnetic temperature mapping model to obtain the temperature estimate of the heating anomaly area; the electromagnetic parameter reference values ​​are input into the preset electromagnetic temperature mapping model to obtain the temperature reference values; the difference between the temperature estimate and the temperature reference values ​​is calculated to obtain the temperature deviation value.

[0030] Based on the absolute value of the temperature deviation, a heating uniformity index is defined. The heating uniformity index is compared with a preset uniformity index threshold range. If the heating uniformity index exceeds the preset uniformity index threshold range, the abnormal heating area is mapped to a specific spatial location within the furnace chamber of the heat treatment furnace and determined as the area to be controlled.

[0031] The sign of the temperature deviation value is determined. When the temperature deviation value is positive, the direction of temperature adjustment for the region to be controlled is set to increase the energy flow input. When the temperature deviation value is negative, the direction of adjustment is set to decrease the energy flow input.

[0032] According to the preset energy flow adjustment step size, a temperature compensation command is generated along the adjustment direction to adjust the output power of the independent heating unit corresponding to the region to be controlled in stages.

[0033] As a further aspect of the present invention, the specific process for defining other regions located in the same axial cross-section of the steel pipe is as follows:

[0034] Obtain the real-time rotational angular velocity and forward linear velocity of the steel pipe, the known radius, and the pre-set installation axis position inside the furnace;

[0035] Based on the preset installation axis position and forward linear velocity direction inside the furnace, the forward linear velocity direction is calibrated as the positive axial direction, and the geometric center point of the first spatially fixed circumferential electromagnetic detection array along the positive axial direction is defined as the origin of the furnace space cylindrical coordinate system; the direction perpendicular to the top of the furnace is calibrated as the circumferential angle zero point direction.

[0036] For each physical point, based on its identifier consisting of an initial circumferential angle and an initial axial position, and combined with the real-time rotational angular velocity, forward linear velocity, known radius, and the preset installation axis position within the furnace, the absolute circumferential angle and absolute axial position of the physical point at the current moment are calculated; based on the absolute circumferential angle, absolute axial position, known radius, and the preset installation axis position within the furnace, its specific spatial coordinates in the cylindrical coordinate system of the furnace space are calculated;

[0037] Obtain the specific spatial coordinates of each physical point constituting the heating anomaly region; extract the axial coordinate components from the specific spatial coordinates; determine the minimum and maximum values ​​of the axial coordinate components; define the interval defined by the minimum and maximum values ​​as the current target axial cross-sectional range;

[0038] Among all physical points on the surface of the steel pipe, physical points whose axial coordinate components are located within the current target axial section and whose circumferential coordinate components are located outside the circumferential coverage of the heating anomaly area are selected to form a set of candidate physical points; the area corresponding to the set of candidate physical points is defined as other areas within the same axial section of the steel pipe.

[0039] As a further aspect of the present invention: the specific process of performing dynamic energy compensation in S5 is as follows:

[0040] Obtain the specific spatial coordinates of each physical point constituting the heating anomaly region in the cylindrical coordinate system of the furnace space; extract the circumferential coordinate components of each physical point from the specific spatial coordinates; determine the maximum and minimum values ​​of the circumferential coordinate components, and use the difference between the maximum and minimum values ​​as the circumferential angle range of the heating anomaly region; based on the real-time rotational angular velocity of the steel pipe and the circumferential angle range, calculate the rotation time required for the heating anomaly region to rotate and sweep across the fixed circumferential action position of the independent heating unit.

[0041] Extract the axial coordinate components of each physical point from the specific spatial coordinates; determine the maximum and minimum values ​​of the axial coordinate components, and use the difference between the maximum and minimum values ​​as the axial extension length of the heating anomaly region; based on the real-time forward linear velocity of the steel pipe and the axial extension length, calculate the time required for the heating anomaly region to completely pass through the axial range of the independent heating unit.

[0042] The time required for rotation is compared with the time required for forward movement, and the smaller value is determined as the effective time of the dynamic energy compensation; the precise energy compensation window of the independent heating unit is set according to the start time in the predicted spatiotemporal coordinates and the effective time; the independent heating unit is driven to perform power adjustment according to the temperature compensation command within the precise energy compensation window.

[0043] As a further aspect of the present invention: S5 further includes adjusting the heating unit according to the temperature compensation command, then re-collecting and calculating the electromagnetic parameters and temperature deviation values ​​of the area to be controlled and other control areas, thereby updating the heating uniformity index.

[0044] Determine whether the updated heating uniformity index has returned to the preset uniformity index threshold range; if it has not returned, determine the adjustment direction again based on the sign of the current temperature deviation value, and generate further power grading adjustment instructions according to the energy flow adjustment step size, and perform iterative adjustment; repeat this process until the updated heating uniformity index returns to the normal range.

[0045] The beneficial effects of this invention are:

[0046] 1) This invention first establishes a direct sensing and full-field mapping capability based on the evolution of intrinsic material properties, which can be understood as the physical characteristic of significant changes in magnetic permeability near a specific critical temperature. Through a circumferentially arranged electromagnetic detection array, the electromagnetic signals induced by this property change are captured in real time, rather than directly measuring temperature radiation susceptible to environmental interference. Each set of original signals is strictly bound to the electromagnetic state of a tiny region on the steel pipe surface at a specific moment. Combining the precise kinematic parameters of the steel pipe, the spatiotemporally discrete observation data is reconstructed through inverse calculation into a unique continuous evolution trajectory of electromagnetic parameters for each physical point on the steel pipe surface. This trajectory is a dynamic response record of the material to the heating process. Furthermore, through a rigorously calibrated mapping relationship, this electromagnetic evolution trajectory can be converted into corresponding temperature evolution information, thereby achieving interference-resistant indirect measurement and visualization of the full-domain temperature field of a rotating workpiece based on physical principles, even in the harsh environment of an industrial furnace.

[0047] 2) This invention dynamically generates a real-time "normal response background field" applicable only to that local area by calculating the instantaneous rate of change of the electromagnetic evolution trajectory of any physical point on the surface of the steel pipe in real time, and simultaneously statistically analyzing the distribution of the rate of change of all its spatially adjacent points. This means that any real anomaly caused by microscopic inhomogeneity of the material or differences in heating will manifest as a continuous deviation of the rate of change of that point from its own local background field. This core logic based on "neighborhood comparison" effectively focuses on local relative differences, enabling the diagnostic process to concentrate on extracting real abnormal signals. Simultaneously, relying on strict spatiotemporal coordinate registration throughout the entire process, any identified anomaly pattern can be uniquely traced to its precise three-dimensional spatial location within the furnace and its specific process time point, achieving a seamless connection from anomaly perception to physical root cause localization, providing a clear spatial target for subsequent precise intervention.

[0048] 3) This invention establishes a closed-loop dynamic control system from "full-domain state perception" to "precise spatiotemporal execution." Based on electromagnetic diagnosis of abnormal areas, it generates quantitative compensation commands containing adjustment direction and energy flow intensity. According to the real-time movement of the steel pipe, it predicts in advance the precise spatiotemporal coordinates of the abnormal area reaching the downstream controllable heating unit and calculates a strictly matching energy compensation time window and duration. This ensures that the independent heating unit applies the correct dose of energy compensation at the correct time, aligned with the correct local location on the workpiece. This "prediction-matching" spatiotemporal synchronous execution mechanism ensures that energy is efficiently and directionally delivered to the moving defect, avoiding broad impacts on the fixed furnace area. This fundamentally shifts the control logic from coarse adjustment of the "static furnace" to real-time feedback and morphological repair of the "moving workpiece state," ultimately improving the ultimate uniformity of heat treatment quality, process stability, and energy utilization efficiency at its source. Attached Figure Description

[0049] The invention will now be further described with reference to the accompanying drawings.

[0050] Figure 1 This is a schematic diagram of a heating monitoring method for seamless stainless steel tubes in a heat treatment furnace according to the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Please see Figure 1As shown, the present invention is a method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace, comprising the following steps:

[0053] S1, the electromagnetic induction signals of the rotating steel pipe are collected in real time by a circumferential electromagnetic detection array arranged in the heating section of the heat treatment furnace, and each electromagnetic induction signal is bound with the precise timestamp of the collection time and the spatial coordinates of the detection array unit to generate an original signal set with spatiotemporal information.

[0054] In critical heating sections of the heat treatment furnace (such as the entrance to the soaking zone), a circumferential electromagnetic detection array is fixedly installed on the furnace wall. This array consists of at least eight independent detection units evenly distributed circumferentially along the furnace chamber to ensure complete coverage of the rotating steel tube. Each detection unit includes an excitation coil and an induction coil, operating at a low-to-mid frequency range of 1kHz-10kHz to balance penetration depth and sensitivity to changes in the permeability of stainless steel. As the seamless stainless steel tube rotates and passes through the detection array, each unit continuously collects the induced electromotive force (EMF) signal in a small spatial region in front of it at a sampling frequency of no less than 100Hz. Data acquisition binds three sets of metadata to each raw signal: a timestamp (T) accurate to milliseconds, and a unique spatial coordinate number (X, Y, Z) for the detection unit. All data is stored in the format of "timestamp-unit ID-raw voltage / current signal," generating a raw signal set with spatiotemporal registration information.

[0055] S2, obtain the rotational angular velocity and forward linear velocity of the steel pipe, and perform kinematic inversion calculation on the original signal set to obtain the evolution trajectory of the electromagnetic parameters of any physical point on the surface of the steel pipe as a function of time.

[0056] The rotational angular velocity ω (radians / second) and forward linear velocity v (meters / second) of the steel pipe are acquired in real time from the production line PLC. The steel pipe is considered an ideal cylinder with radius R, and its theoretical axis of motion within the furnace is used as a reference. For any signal sample in the original signal set, it is known that it was acquired at time t by a detection unit at spatial location Psensor. A source point on the steel pipe surface is assigned to this signal by solving the following inverse kinematic problem: find a physical point on the steel pipe surface whose spatial position at time t exactly falls within the effective sensitive region of the detection unit Psensor. The position of this physical point on the steel pipe can be uniquely identified by its initial circumferential angle θ0 and axial position z0. Through iterative calculation, (θ0, z0) satisfying the spatial position constraint is solved. Repeating this process for all signals allows for the classification of massive discrete signals. All signals originating from the same physical point (θ0, z0) are sorted by their timestamps, forming a discrete-time sequence of the electromagnetic parameters (such as the normalized amplitude of complex impedance) of that point. Smoothing and spline interpolation are performed on this sequence to generate a continuous and smooth electromagnetic parameter evolution trajectory for that physical point. This process is repeated for all physical points to complete the full-field trajectory reconstruction.

[0057] S3, calculate the dynamic rate of change of each physical point based on its evolution trajectory, and compare the dynamic rate of change with the local background field composed of the rates of change of the physical points that are spatially adjacent to the physical point, to obtain the set of abnormal points whose rates of change continuously deviate from the local background field;

[0058] The dynamic rate of change of each physical point is calculated based on its evolution trajectory. Specifically, firstly, numerical differentiation is performed on the curve of the continuous change of the electromagnetic parameters of the physical point over time to obtain a continuous rate of change signal reflecting the instantaneous rate of heating or cooling at that point. For example, when the curve rises sharply at a certain moment, the corresponding rate of change shows a positive peak. Subsequently, to construct the local background field of the point, several spatially adjacent physical points are automatically found and locked from the digital model of the surface of the steel pipe where it is located, such as the eight neighboring points directly surrounding it. The rate of change data of these neighboring points at the same time are collected in real time, and a statistical benchmark value that can represent the behavioral characteristics of the vast majority of normal points in this local area and a statistical discrete value that measures the normal fluctuation range of this area are calculated. Next, a comparison and judgment process is performed. The real-time deviation between the rate of change of the physical point and the statistical benchmark value of its local background field is continuously calculated, and this deviation is compared with a reasonable fluctuation threshold set based on statistical discrete values. If the deviation of the point consistently exceeds the threshold range for multiple consecutive sampling periods—for example, if its rate of change is consistently significantly lower than the average level of all its neighbors while traversing the entire heating section—this indicates that the point has a persistent heating lag. Finally, all spatially dispersed anomalous physical points identified through the above conditions are analyzed by a spatial clustering algorithm. This algorithm automatically merges densely located anomalous points into one or more coherent anomalous regions with clear boundaries based on the proximity relationships between points. For example, all anomalous points clustered within a 90-degree sector on the steel pipe surface are identified as the same heating anomaly region, thus completing the transformation from discrete points to structured regions.

[0059] An intrinsic diagnostic mechanism that identifies true anomalies through local self-comparison can completely eliminate global measurement interference caused by inherent non-uniformity in the overall structure, magnetic field distribution, or heating field within the heat treatment furnace. This is because the method only considers the relative behavioral differences between a single point and its immediate vicinity, and is insensitive to background fluctuations shared by all points. This allows for highly accurate identification of heating non-uniformity caused by genuine defects such as microscopic segregation of the material itself, uneven original wall thickness, or local shading, without misjudging inherent thermal field non-uniformity of the furnace itself as a problem with the steel pipe. Through a continuously deviating judgment logic, false signals caused by measurement noise or transient disturbances can be effectively filtered out, ensuring that the captured anomaly patterns have reliable process relevance. This provides data support for subsequent steps, making quantitative energy compensation for specific areas possible. Thus, quality control shifts from fuzzy adjustments to the furnace environment to precise targeted repair of defects in the steel pipe itself, laying the most critical technical foundation for achieving a substantial improvement in heating uniformity.

[0060] S4, perform spatial clustering on the abnormal point set to obtain the heating abnormal region, extract the electromagnetic parameters corresponding to each physical point constituting the heating abnormal region, and input the electromagnetic parameters into the preset electromagnetic-temperature mapping model to obtain the temperature compensation command.

[0061] First, spatial clustering analysis is performed on the identified set of anomalies to group spatially adjacent anomalies into continuous heating anomaly regions. Next, electromagnetic characteristic parameter data corresponding to all physical points constituting these anomaly regions are extracted, and the average value of these data is calculated to obtain electromagnetic parameter characteristic values ​​representing the overall electromagnetic state of the region. Simultaneously, electromagnetic characteristic parameter data from other regions located on the same axial section of the steel pipe but not spatially overlapping with the anomaly region are acquired, and their average values ​​are also calculated to obtain electromagnetic parameter reference values, representing the normal heating level of that section. Then, the electromagnetic parameter characteristic values ​​are input into a preset electromagnetic-temperature mapping model, outputting an estimated temperature value for the heating anomaly region; the electromagnetic parameter reference values ​​are input into the same model to obtain a temperature reference value. By calculating the difference between the estimated temperature value and the reference value, a precise temperature deviation value is obtained. A heating uniformity index is defined based on the absolute value of this deviation value and compared with a preset uniformity index threshold range. If the index exceeds the threshold range, the heating anomaly region is mapped to specific spatial coordinates within the heat treatment furnace chamber and identified as a region to be controlled. The temperature deviation value is further determined by its sign: if the deviation is positive, it indicates that the temperature in the area is too high, and the adjustment direction is set to increase the energy flow input; if it is negative, it indicates that the temperature is too low, and the adjustment direction is set to decrease the energy flow input. Finally, according to the preset energy flow adjustment step size, a quantitative temperature compensation command is generated along this direction to adjust the output power of the corresponding independent heating unit in the area to be controlled in stages.

[0062] By transforming the abstract "signal anomalies" detected in the preceding steps into specific, executable process control commands, a crucial link from "perceiving anomalies" to "executing compensation" is established. Calculating regional averages and reference values ​​avoids random errors based on single-point data, ensuring robustness of judgment. A mapping model converts electromagnetic parameters into temperature quantities, grounding commands in direct process objectives. Then, by calculating temperature deviations and constructing uniformity indicators for comparison with preset thresholds, automatic and objective decision-making regarding "intervention is needed" is achieved, avoiding over- or under-regulation. Only significant non-uniformities exceeding process tolerances trigger regulation, ensuring stable production rhythm and effective intervention. Then, mapping the abnormal area to the specific location in the furnace and determining the direction of control is the key operation to accurately anchor the conclusions of the "data space" to the "physical space". It ensures that the compensation energy will be delivered to the correct spatial location, and intelligently decides whether to heat or cool based on the positive or negative deviation. Finally, generating graded adjustment commands based on preset step size is a robust and easy-to-implement control strategy. It approaches the target through gradual and trial-and-error adjustments, preventing over-adjustment or under-adjustment, and leaving room for fine-tuning based on feedback. This achieves dynamic and precise correction of the heating uniformity of the steel pipe, ensuring that the heat treatment quality changes from an "open-loop" mode that depends on fixed process parameters to an "adaptive" intelligent control mode based on real-time feedback of the workpiece status.

[0063] S5. Based on the real-time rotation and forward speed of the steel pipe, predict the spatiotemporal coordinates of the heating anomaly area to reach the next heating section. Based on the spatiotemporal coordinates, determine the independent heating unit that needs to be started in the next heating section and, in conjunction with the temperature compensation command, drive the independent heating unit to perform dynamic energy compensation.

[0064] First, the precise rotational angular velocity and forward linear velocity of the steel pipe are read in real time from the production line control unit. Then, combined with the precise geometric center coordinates of the abnormal heating area on the digital model of the steel pipe surface (e.g., an elliptical area located in the middle of the steel pipe with a circumferential angle of 120 degrees), the built-in kinematic model calculates how long it would take for the geometric center of this area to move axially to the starting boundary position of the downstream heating section with independent control capabilities at the current speed. Simultaneously, based on the rotational speed, the specific circumferential orientation of the furnace corresponding to the center point of this area at the moment of arrival is calculated, thus obtaining a set of spatiotemporal coordinates containing the precise arrival time and specific circumferential angle. Next, based on these predicted spatiotemporal coordinates, the independent heating units to be activated in the next heating section are determined. The equipment layout diagram of the heating section is called, and the predicted circumferential orientation is compared with the orientation coverage of all independent heating units in the diagram. For example, the four adjacent burner units numbered 12 to 15 in this section, which are responsible for the 90-degree sector orientation, are matched, and these units are identified as the execution mechanisms for this compensation. Finally, combined with the temperature compensation command driving these independent heating units to perform dynamic energy compensation, the power adjustment direction and magnitude specified in the compensation command, such as requiring a 5% increase in power output, along with the calculated start time and duration of the compensation action, are sent to the control valves of the burner units. These units will start synchronously just before the abnormal area enters its effective range, adjust the flame intensity or power output according to the command, and return to their original state after the abnormal area has completely passed through its effective range, thereby completing a dynamic energy spraying of the moving target.

[0065] The steel pipes inside the heat treatment furnace move continuously, and the identified abnormal heating areas also move with the rotation of the pipes. Simply compensating at the point of abnormality or making general adjustments to the entire downstream heating section cannot accurately deliver energy to the moving target. By predicting spatiotemporal coordinates, "spatiotemporal synchronization" between the compensation action and the moving target is achieved, ensuring that energy is applied to the correct location within the furnace at the correct time, thus precisely targeting the abnormal area. Its core advantage is a paradigm shift from "heating a fixed area" to "tracking heating of a moving target," making energy compensation more targeted and precise in both time and space. This avoids interference with normal areas or insufficient compensation for abnormal areas, greatly improving control accuracy and efficiency. Ultimately, this mechanism ensures that the entire method forms a complete "perception-decision-precise execution" closed loop, translating all the monitoring, identification, and diagnostic analysis in the preceding steps into tangible and immediate process optimization of the workpiece itself, dynamically eliminating temperature unevenness in the steel pipe during heat treatment, thereby steadily improving the microstructure and quality consistency of the final product.

[0066] In a preferred embodiment of the present invention, the specific process of obtaining the evolutionary trajectory in step S2 is as follows:

[0067] The real-time rotational angular velocity and forward linear velocity of the steel pipe, the known radius, and the preset installation axis position inside the furnace are obtained; based on the rotational angular velocity, forward linear velocity, known radius, and preset installation axis position inside the furnace, the surface of the steel pipe is modeled as a cylindrical surface, and the physical points on the cylindrical surface are assigned labels consisting of initial circumferential angles and initial axial positions;

[0068] Based on the rotational angular velocity and the forward linear velocity, calculate the circumferential angular offset and axial position offset of the physical point from the initial mark over time, and determine its theoretical spatial coordinates relative to the fixed space of the furnace at any time.

[0069] For each sensing signal with a timestamp and spatial coordinates of the detection unit in the original signal set, the theoretical spatial coordinates are matched with the spatial coordinates of the detection unit to determine whether the theoretical spatial coordinates are located in the effective sensing area of ​​the detection unit at the time corresponding to the timestamp of the signal. When the determination is yes, the sensing signal is associated with the corresponding physical point as the state observation value of the physical point at that time.

[0070] After traversing all induced signals, a discrete state observation sequence ordered by timestamp is obtained for each physical point; the discrete state observation sequence is smoothed and curve-fitted to obtain the evolution trajectory of the electromagnetic parameters of the physical point as a function of time.

[0071] First, the rotational angular velocity and linear velocity of the steel pipe are acquired from the real-time data bus of the production line, while simultaneously accessing preset steel pipe radius parameters and the three-dimensional coordinates of the furnace installation axis. Based on these parameters, a virtual cylindrical digital model precisely corresponding to the physical steel pipe is established internally. Each addressable position on the surface of this model is assigned a unique identifier consisting of an initial circumferential angle (e.g., 0 to 360 degrees) and an initial axial position (e.g., millimeters from the furnace inlet). For example, a point can be identified as "1500 mm axially, 90 degrees circumferentially." Next, for any identified physical point, integration is performed from the initial moment based on the acquired real-time angular velocity and linear velocity to calculate the cumulative change in circumferential angle and cumulative displacement in axial position over time. This determines the precise theoretical position of the point relative to a fixed three-dimensional spatial coordinate system with the furnace body as the reference at any future or past moment.

[0072] Then, the massive amount of discrete data from the original signal set begins to be processed: for each recorded inductive signal, its data packet contains information such as "collected by detection unit No. 3 at 12:00:01.235"; the theoretical spatial coordinate set corresponding to the timestamp of the signal is retrieved, and geometric calculations are used to determine whether, at that precise moment, the theoretical coordinates of any physical point fall exactly within the preset effective inductive cube space in front of detection unit No. 3; once a match is found, for example, if a point marked "axial 1500 mm, circumferential 90 degrees" is found to be located precisely in front of detection unit No. 3 at that moment... Directly in front of the point, the core information such as the amplitude and phase of the signal is taken as a snapshot of the state of the physical point at that instant, firmly associated and stored. Thus, each physical point obtains a series of discrete observations from different detectors arranged in chronological order. Finally, filtering algorithms are applied to these seemingly discontinuous but actually connected observation sequences to remove impulse noise, and spline curve fitting technology is used to smoothly connect the discrete points, thereby forming a complete, continuous, and detailed curve that reflects how the electromagnetic parameters of the point change over time from entering the monitoring area to leaving the area, i.e., the evolution trajectory.

[0073] Through rigorous kinematic inversion, fragmented signals captured at different times and locations by a network of fixed sensors throughout the furnace are pieced together to reconstruct the complete life history of each point on the moving object. This is equivalent to equipping each inch of the steel pipe surface with a virtual, dedicated data logger that moves with it. The purpose of this method is to completely abandon the dependence on idealized uniform heating fields or absolute coordinate systems, and instead start directly from the most original and objective observation data to reconstruct a reliable individual evolutionary process through physical and geometric rules. Only by obtaining the independent, continuous, and true evolutionary trajectory of each physical point can all the precise analyses in subsequent steps, such as calculating its rate of change, making local comparisons with neighbors, and diagnosing anomalies, become possible.

[0074] In another preferred embodiment of the present invention, the specific process of obtaining the anomaly set in step S3 is as follows:

[0075] Calculate the instantaneous rate of change of the evolution trajectory of each physical point on the surface of the steel pipe and determine the spatially adjacent physical points of each physical point as a candidate neighbor set;

[0076] Calculate the statistical distribution of the instantaneous change rate of the candidate neighbor point set; set a preliminary judgment threshold based on the statistical distribution, and select points in the candidate neighbor point set whose instantaneous change rate exceeds the preliminary judgment threshold as candidate points, and remove the candidate points from the candidate neighbor point set to obtain a purified neighbor point set;

[0077] Calculate the statistical mean and statistical fluctuation range based on the instantaneous change rate of the purified neighbor point set; set a stability judgment threshold based on the statistical fluctuation range;

[0078] The difference between the instantaneous rate of change of each physical point and the statistical mean is calculated to obtain the purification deviation. Physical points whose purification deviation continuously exceeds the stability judgment threshold are identified as abnormal physical points, and the abnormal physical points are collected to form an abnormal point set.

[0079] First, numerical differentiation is performed on the electromagnetic parameter evolution trajectory of each physical point on the steel pipe surface to calculate the instantaneous rate of change (RCD) of this trajectory at each sampling moment. Next, based on the three-dimensional coordinates of each physical point on the digital steel pipe surface model, its spatially nearest neighbor points are automatically identified. For example, two points on each side of the circumference and two points in front and behind the axial direction, for a total of eight closest physical points, are defined as an initial candidate neighbor set. Then, the instantaneous rate of change of all points in this candidate neighbor set at the current sampling moment is calculated, and the statistical distribution characteristics of this data are analyzed. Based on the overall shape and dispersion of this distribution, an intelligent threshold is set for preliminary screening. Extreme points in the candidate set whose rate of change values ​​clearly fall outside this threshold are marked and temporarily removed from the candidate neighbor set, thus obtaining a purified neighbor point set that better represents the generally normal behavior of this local area. Next, based on the rate of change of all points in this cleaned neighbor set, two key statistics are calculated: one is the statistical average that characterizes the central level of the rate of change in the local area, and the other is the statistical fluctuation range that characterizes the tolerance of normal fluctuations in the area. Based on this fluctuation range, a more stringent threshold is set for the final stability determination. Then, the difference between the instantaneous rate of change of the central physical point itself and the statistical average of the cleaned neighbor set is calculated to obtain the cleaned deviation. The cleaned deviation of the central physical point is continuously tracked over a period of time. If the deviation continuously exceeds the stability determination threshold set based on the normal fluctuation range of the neighbors, the physical point is finally determined to be an anomaly. After traversing all physical points and completing the above determination, all points marked as anomalies are gathered together to form the final anomaly set.

[0080] In another preferred embodiment of the present invention, the specific construction process of the preset electromagnetic-temperature mapping model in step S4 is as follows:

[0081] Multiple pipe section specimens with different characteristic temperature ranges are pre-acquired and heated in a controlled experimental thermal environment according to a preset heating curve. During the heating process, the circumferential electromagnetic detection array is used to synchronously collect the induction signals of each pipe section specimen at multiple characteristic temperature points, and a contact temperature measuring device is used to synchronously measure the actual temperature value of each pipe section specimen at the characteristic temperature point.

[0082] For each induced signal, demodulation and feature quantity calculation are performed to extract its signal amplitude change and phase shift, which are then used together as electromagnetic parameters. The electromagnetic parameters are used as independent variables, and the actual temperature value, which is strictly paired with the induced signal in time and space, is used as the dependent variable to form calibration data pairs. The least squares method is used to perform regression calculations on all the calibration data pairs to fit a continuous mapping function from the independent variable to the dependent variable. The continuous mapping function is defined as the preset electromagnetic-temperature mapping model.

[0083] In another preferred embodiment of the present invention, the specific process of obtaining the temperature compensation command in step S4 is as follows:

[0084] Extract multiple electromagnetic characteristic parameter data corresponding to the heating abnormal area and calculate their average value to obtain the electromagnetic parameter characteristic value of the heating abnormal area; obtain multiple electromagnetic characteristic parameter data corresponding to other areas in the same axial section of the steel pipe as the heating abnormal area and calculate their average value to obtain the electromagnetic parameter reference value;

[0085] The electromagnetic parameter characteristic values ​​are input into the preset electromagnetic temperature mapping model to obtain the temperature estimate of the heating anomaly area; the electromagnetic parameter reference values ​​are input into the preset electromagnetic temperature mapping model to obtain the temperature reference values; the difference between the temperature estimate and the temperature reference values ​​is calculated to obtain the temperature deviation value.

[0086] Based on the absolute value of the temperature deviation, a heating uniformity index is defined. The heating uniformity index is compared with a preset uniformity index threshold range. If the heating uniformity index exceeds the preset uniformity index threshold range, the abnormal heating area is mapped to a specific spatial location within the furnace chamber of the heat treatment furnace and determined as the area to be controlled.

[0087] The sign of the temperature deviation value is determined. When the temperature deviation value is positive, the direction of temperature adjustment for the region to be controlled is set to increase the energy flow input. When the temperature deviation value is negative, the direction of adjustment is set to decrease the energy flow input.

[0088] According to the preset energy flow adjustment step size, a temperature compensation command is generated along the adjustment direction to adjust the output power of the independent heating unit corresponding to the region to be controlled in stages.

[0089] In another preferred embodiment of the present invention, the specific process for defining other regions located in the same axial section of the steel pipe is as follows:

[0090] Obtain the real-time rotational angular velocity and forward linear velocity of the steel pipe, the known radius, and the pre-set installation axis position inside the furnace;

[0091] Based on the preset installation axis position and forward linear velocity direction inside the furnace, the forward linear velocity direction is calibrated as the positive axial direction, and the geometric center point of the first spatially fixed circumferential electromagnetic detection array along the positive axial direction is defined as the origin of the furnace space cylindrical coordinate system; the direction perpendicular to the top of the furnace is calibrated as the circumferential angle zero point direction.

[0092] For each physical point, based on its identifier consisting of an initial circumferential angle and an initial axial position, and combined with the real-time rotational angular velocity, forward linear velocity, known radius, and the preset installation axis position within the furnace, the absolute circumferential angle and absolute axial position of the physical point at the current moment are calculated; based on the absolute circumferential angle, absolute axial position, known radius, and the preset installation axis position within the furnace, its specific spatial coordinates in the cylindrical coordinate system of the furnace space are calculated;

[0093] Obtain the specific spatial coordinates of each physical point constituting the heating anomaly region; extract the axial coordinate components from the specific spatial coordinates; determine the minimum and maximum values ​​of the axial coordinate components; define the interval defined by the minimum and maximum values ​​as the current target axial cross-sectional range;

[0094] Among all physical points on the surface of the steel pipe, physical points whose axial coordinate components are located within the current target axial section and whose circumferential coordinate components are located outside the circumferential coverage of the heating anomaly area are selected to form a set of candidate physical points; the area corresponding to the set of candidate physical points is defined as other areas within the same axial section of the steel pipe.

[0095] By strictly limiting the comparison area to the same axial section, it is ensured that the compared "abnormal area" and "reference area" are in the exact same furnace heating zone and undergo almost the same external thermal process. This eliminates the interference caused by the inherent temperature gradient due to the different axial positions of the steel pipes, allowing the calculated temperature deviation to purely reflect the circumferential heating non-uniformity. This precise spatial definition method ensures that the electromagnetic parameter reference values ​​or temperature reference values ​​calculated from the normal area are highly representative and accurate, like finding a "mirror" for the abnormal area under the exact same environment, clearly reflecting and quantifying its degree of abnormality. Understandably, only by comparing with a spatially adjacent normal area with highly similar heating conditions can the generated compensation command truly target the "non-uniformity" itself, rather than other confounding factors. This ensures that the compensation energy is used to correct real local defects, ultimately effectively guaranteeing and improving the true uniformity of heat treatment in both the circumferential and axial directions.

[0096] In another preferred embodiment of the present invention, the specific process of performing dynamic energy compensation in step S5 is as follows:

[0097] Obtain the specific spatial coordinates of each physical point constituting the heating anomaly region in the cylindrical coordinate system of the furnace space; extract the circumferential coordinate components of each physical point from the specific spatial coordinates; determine the maximum and minimum values ​​of the circumferential coordinate components, and use the difference between the maximum and minimum values ​​as the circumferential angle range of the heating anomaly region; based on the real-time rotational angular velocity of the steel pipe and the circumferential angle range, calculate the rotation time required for the heating anomaly region to rotate and sweep across the fixed circumferential action position of the independent heating unit.

[0098] Extract the axial coordinate components of each physical point from the specific spatial coordinates; determine the maximum and minimum values ​​of the axial coordinate components, and use the difference between the maximum and minimum values ​​as the axial extension length of the heating anomaly region; based on the real-time forward linear velocity of the steel pipe and the axial extension length, calculate the time required for the heating anomaly region to completely pass through the axial range of the independent heating unit.

[0099] The time required for rotation is compared with the time required for forward movement, and the smaller value is determined as the effective time of the dynamic energy compensation; the precise energy compensation window of the independent heating unit is set according to the start time in the predicted spatiotemporal coordinates and the effective time; the independent heating unit is driven to perform power adjustment according to the temperature compensation command within the precise energy compensation window.

[0100] First, the established spatial database is accessed to obtain the specific three-dimensional coordinates of all physical points constituting the heating anomaly area (e.g., these points may be distributed in an elliptical patch) in the cylindrical coordinate system of the furnace space. From these coordinate data, the circumferential coordinate components (a set of values ​​in angles) representing the specific orientation of each point in the annular direction are extracted. Through traversal and sorting algorithms, the maximum and minimum values ​​are found from these values. For example, if the maximum value is 185 degrees and the minimum value is 155 degrees, the difference of 30 degrees is determined as the coverage area of ​​the anomaly area in the circumferential direction of the steel pipe, i.e., the circumferential angle range. Next, the current rotational angular velocity of the steel pipe (e.g., 1.5 degrees per second) is read from the real-time data stream of the production line. The circumferential angle range is divided by this angular velocity to accurately calculate how many seconds it takes for the anomaly area to completely sweep across the flame coverage area of ​​an independent heating unit (e.g., a burner) fixed to the side wall of the furnace. This calculation result is the rotation time required. Simultaneously, the axial coordinate components (i.e., position values ​​along the length of the steel pipe) of each physical point are extracted from the same set of three-dimensional coordinates. The maximum and minimum values ​​are also identified, and the difference is defined as the dimension of the abnormal region in the length direction, i.e., the axial extension length. Based on the real-time forward linear velocity of the steel pipe (e.g., 10 mm per second), the axial extension length is divided by the forward linear velocity to calculate how many seconds it takes for the region to completely pass through the effective heating zone of the independent heating unit in the length direction, thus obtaining the required forward time. Then, the calculated rotation time is compared with the forward time, and the smaller value is selected (e.g., if rotation takes 20 seconds and forward takes 15 seconds, then 15 seconds is selected). This smaller value is determined as the effective action time for this dynamic energy compensation, ensuring that the heating action only occurs when the region is completely within the effective coverage area of ​​the heating unit. Finally, the previously predicted start time of arrival of the region is combined with this effective action time to set a precise energy compensation window with clearly defined start and end times. The designated independent heating unit is then driven to strictly perform the compensation operation within this time window, according to the power adjustment direction and amplitude required by the temperature compensation command.

[0101] It is understandable that for the energy of the heating unit to effectively act on the steel pipe, two spatial conditions must be met simultaneously: circumferentially, the abnormal area must rotate to a position directly facing the heating unit; axially, the abnormal area must be within the limited effective length of the heating unit. Therefore, the truly effective compensation period can only be the intersection of these two conditions. Selecting the smaller of the "rotation time" and the "advance time" is the key operation for precisely defining this intersection period—it signifies that once this shorter time is exceeded, the abnormal area has either partially rotated out of the heating range circumferentially or partially moved out of the effective range axially, and continuing to apply energy will lead to misheating of non-target areas or energy waste. The direct purpose of this is to achieve instantaneous and precise energy focusing, concentrating the limited compensation energy with the highest spatiotemporal density and without waste to the single target area being traversed, ensuring the absolute accuracy of the compensation action, avoiding interference with the heat treatment state of adjacent normal areas due to overcompensation, and maintaining the overall stability of the process; economically, it minimizes the duration of high-power output, achieving the goal of local correction with minimal energy cost, and significantly improving energy utilization efficiency.

[0102] In another preferred embodiment of the present invention, step S5 further includes, after adjusting the heating unit according to the temperature compensation command, re-collecting and calculating the electromagnetic parameters and temperature deviation values ​​of the area to be controlled and other control areas, and updating the heating uniformity index accordingly.

[0103] Determine whether the updated heating uniformity index has returned to the preset uniformity index threshold range; if it has not returned, determine the adjustment direction again based on the sign of the current temperature deviation value, and generate further power grading adjustment instructions according to the energy flow adjustment step size, and perform iterative adjustment; repeat this process until the updated heating uniformity index returns to the normal range.

[0104] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for monitoring the heating of seamless stainless steel tubes inside a heat treatment furnace, characterized in that, Includes the following steps: S1, the electromagnetic induction signal of the rotating steel pipe in the heat treatment furnace is collected in real time by the circumferential electromagnetic detection array, and the electromagnetic induction signal is synchronously bound with the corresponding timestamp and the coordinates of the detection array unit to obtain the original signal set; S2, obtain the rotational angular velocity and forward linear velocity of the steel pipe, and perform kinematic inversion calculation on the original signal set to obtain the evolution trajectory of the electromagnetic parameters of any physical point on the surface of the steel pipe as a function of time. The specific process for obtaining the evolutionary trajectory is as follows: The real-time rotational angular velocity and forward linear velocity of the steel pipe, the known radius, and the preset installation axis position inside the furnace are obtained; based on the rotational angular velocity, forward linear velocity, known radius, and preset installation axis position inside the furnace, the surface of the steel pipe is modeled as a cylindrical surface, and the physical points on the cylindrical surface are assigned labels consisting of initial circumferential angles and initial axial positions; Based on the rotational angular velocity and the forward linear velocity, calculate the circumferential angular offset and axial position offset of the physical point from the initial mark over time, and determine its theoretical spatial coordinates relative to the fixed space of the furnace at any time. For each sensing signal with a timestamp and spatial coordinates of the detection unit in the original signal set, the theoretical spatial coordinates are matched with the spatial coordinates of the detection unit to determine whether the theoretical spatial coordinates are located in the effective sensing area of ​​the detection unit at the time corresponding to the timestamp of the sensing signal. When the determination is yes, the sensing signal is associated with the corresponding physical point and used as the state observation value of the physical point at that moment. After traversing all induced signals, a discrete state observation sequence ordered by timestamp is obtained for each physical point; the discrete state observation sequence is smoothed and curve-fitted to obtain the evolution trajectory of the electromagnetic parameters of the physical point as a function of time. S3, calculate the dynamic rate of change of each physical point based on its evolution trajectory, and compare the dynamic rate of change with the local background field composed of the rates of change of the physical points that are spatially adjacent to the physical point, to obtain the set of abnormal points whose rates of change continuously deviate from the local background field; S4, perform spatial clustering on the abnormal point set to obtain the heating abnormal region, extract the electromagnetic parameters corresponding to each physical point constituting the heating abnormal region, and input the electromagnetic parameters into the preset electromagnetic-temperature mapping model to obtain the temperature compensation command. The specific process of obtaining the temperature compensation command is as follows: Extract multiple electromagnetic characteristic parameter data corresponding to the heating abnormal area and calculate their average value to obtain the electromagnetic parameter characteristic value of the heating abnormal area; obtain multiple electromagnetic characteristic parameter data corresponding to other areas in the same axial section of the steel pipe as the heating abnormal area and calculate their average value to obtain the electromagnetic parameter reference value; The electromagnetic parameter characteristic values ​​are input into the preset electromagnetic temperature mapping model to obtain the temperature estimate of the heating anomaly region; the electromagnetic parameter reference values ​​are input into the preset electromagnetic temperature mapping model to obtain the temperature reference values. The temperature deviation value is obtained by calculating the difference between the estimated temperature value and the reference temperature value. Based on the absolute value of the temperature deviation, a heating uniformity index is defined. The heating uniformity index is compared with a preset uniformity index threshold range. If the heating uniformity index exceeds the preset uniformity index threshold range, the abnormal heating area is mapped to a specific spatial location within the furnace chamber of the heat treatment furnace and determined as the area to be controlled. The sign of the temperature deviation value is determined. When the temperature deviation value is positive, the adjustment direction for the temperature of the region to be controlled is set to increase the energy flow input. When the temperature deviation value is negative, the adjustment direction is set to decrease the energy flow input. According to the preset energy flow adjustment step size, a temperature compensation command is generated along the adjustment direction to adjust the output power of the independent heating unit corresponding to the region to be controlled in stages. S5, based on the real-time rotation and forward speed of the steel pipe, predict the spatiotemporal coordinates of the abnormal heating area to the next heating section, determine the independent heating unit that needs to be started in the next heating section based on the spatiotemporal coordinates, and drive the independent heating unit to perform dynamic energy compensation in conjunction with the temperature compensation command.

2. The method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace according to claim 1, characterized in that, In S3, the specific process of obtaining the anomaly set is as follows: Calculate the instantaneous rate of change of the evolution trajectory of each physical point on the surface of the steel pipe and determine the spatially adjacent physical points of each physical point as a candidate neighbor set; Calculate the statistical distribution of the instantaneous change rate of the candidate neighbor point set; set a preliminary judgment threshold based on the statistical distribution, and select points in the candidate neighbor point set whose instantaneous change rate exceeds the preliminary judgment threshold as candidate points, and remove the candidate points from the candidate neighbor point set to obtain a purified neighbor point set; Calculate the statistical mean and statistical fluctuation range based on the instantaneous change rate of the purified neighbor point set; set a stability judgment threshold based on the statistical fluctuation range; The difference between the instantaneous rate of change of each physical point and the statistical mean is calculated to obtain the purification deviation. Physical points whose purification deviation continuously exceeds the stability judgment threshold are identified as abnormal physical points, and the abnormal physical points are collected to form an abnormal point set.

3. The method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace according to claim 1, characterized in that, In step S4, the specific construction process of the preset electromagnetic-temperature mapping model is as follows: Multiple pipe section specimens with different characteristic temperature ranges are pre-acquired and heated in a controlled experimental thermal environment according to a preset heating curve. During the heating process, the circumferential electromagnetic detection array is used to synchronously collect the induction signals of each pipe section specimen at multiple characteristic temperature points, and a contact temperature measuring device is used to synchronously measure the actual temperature value of each pipe section specimen at the characteristic temperature point. For each induced signal, demodulation and feature quantity calculation are performed to extract its signal amplitude change and phase shift, which are then used together as electromagnetic parameters. The electromagnetic parameters are used as independent variables, and the actual temperature value, which is strictly paired with the induced signal in time and space, is used as the dependent variable to form calibration data pairs. The least squares method is used to perform regression calculations on all the calibration data pairs to fit a continuous mapping function from the independent variable to the dependent variable. The continuous mapping function is defined as the preset electromagnetic-temperature mapping model.

4. The method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace according to claim 1, characterized in that, The specific process for defining other regions within the same axial cross-section of the steel pipe is as follows: Obtain the real-time rotational angular velocity and forward linear velocity of the steel pipe, the known radius, and the position of the preset installation axis inside the furnace; Based on the preset installation axis position and forward linear velocity direction inside the furnace, the forward linear velocity direction is calibrated as the positive axial direction, and the geometric center point of the first spatially fixed circumferential electromagnetic detection array along the positive axial direction is defined as the origin of the furnace space cylindrical coordinate system; the direction perpendicular to the top of the furnace is calibrated as the circumferential angle zero point direction. For each physical point, based on its identifier consisting of an initial circumferential angle and an initial axial position, and combined with the real-time rotational angular velocity, forward linear velocity, known radius, and the preset installation axis position within the furnace, the absolute circumferential angle and absolute axial position of the physical point at the current moment are calculated; based on the absolute circumferential angle, absolute axial position, known radius, and the preset installation axis position within the furnace, its specific spatial coordinates in the cylindrical coordinate system of the furnace space are calculated; Obtain the specific spatial coordinates of each physical point constituting the heating anomaly region; extract the axial coordinate components from the specific spatial coordinates; determine the minimum and maximum values ​​of the axial coordinate components; define the interval defined by the minimum and maximum values ​​as the current target axial cross-sectional range; Among all physical points on the surface of the steel pipe, physical points whose axial coordinate components are located within the current target axial section and whose circumferential coordinate components are located outside the circumferential coverage of the heating anomaly area are selected to form a set of candidate physical points; the area corresponding to the set of candidate physical points is defined as other areas within the same axial section of the steel pipe.

5. The method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace according to claim 4, characterized in that, In S5, the specific process of performing dynamic energy compensation is as follows: Obtain the specific spatial coordinates of each physical point constituting the heating anomaly region in the cylindrical coordinate system of the furnace space; extract the circumferential coordinate components of each physical point from the specific spatial coordinates; determine the maximum and minimum values ​​of the circumferential coordinate components, and use the difference between the maximum and minimum values ​​as the circumferential angle range of the heating anomaly region; based on the real-time rotational angular velocity of the steel pipe and the circumferential angle range, calculate the rotation time required for the heating anomaly region to rotate and sweep across the fixed circumferential action position of the independent heating unit. Extract the axial coordinate components of each physical point from the specific spatial coordinates; determine the maximum and minimum values ​​of the axial coordinate components, and use the difference between the maximum and minimum values ​​as the axial extension length of the heating anomaly region; based on the real-time forward linear velocity of the steel pipe and the axial extension length, calculate the time required for the heating anomaly region to completely pass through the axial range of the independent heating unit. The time required for rotation is compared with the time required for forward movement, and the smaller value is determined as the effective time of the dynamic energy compensation; the precise energy compensation window of the independent heating unit is set according to the start time in the spatiotemporal coordinates and the effective time; the independent heating unit is driven to perform power adjustment according to the temperature compensation command within the precise energy compensation window.

6. The method for monitoring the heating of seamless stainless steel tubes in a heat treatment furnace according to claim 1, characterized in that, S5 further includes adjusting the heating unit according to the temperature compensation command, then re-collecting and calculating the electromagnetic parameters and temperature deviation values ​​of the area to be controlled and other control areas, and updating the heating uniformity index accordingly. Determine whether the updated heating uniformity index has returned to the preset uniformity index threshold range; If the temperature does not return to normal, the adjustment direction is determined again based on the sign of the current temperature deviation value, and further power grading adjustment instructions are generated according to the energy flow adjustment step size for iterative adjustment. Repeat this process until the updated heating uniformity index returns to the normal range.

Images