A bearing ring heat treatment regulation method

Abstract

This invention discloses a method for controlling the heat treatment of bearing rings, applicable to a heating-holding-cooling process. The method collects temperature T, deformation D, and equivalent state S of microstructure evolution. After time synchronization, anomaly suppression, and data quality assessment, multi-source data are fused to obtain T_f, D_f, and S_f. Based on this, performance deviation e_p, deformation deviation e_d, and energy consumption deviation e_e are calculated. A constraint package is generated, including target setpoints, upper limits of heating slope / cooling intensity, deformation window, and power soft limit. Under deformation priority constraints, heating power P, holding time t_hold, and cooling rate v_cool are adjusted in a linked manner to achieve closed-loop adaptive control. Preferably, online model identification and freezing, degradation interlocking, and write-back updates are introduced to improve consistency and reduce energy consumption redundancy.

Description

Technical Field

[0001] This invention relates to the field of heat treatment process control technology in bearing manufacturing, and in particular to a multi-source information fusion adaptive process control method for the heat treatment process of bearing rings, belonging to the direction of process control and green and low-carbon manufacturing of bearing ring heat treatment. Background Technology

[0002] Heat treatment of bearing rings (including inner and / or outer rings) is a key process that determines their mechanical properties and dimensional stability. It typically includes stages such as heating, holding and cooling. By adjusting the heat input and cooling intensity, the microstructure is optimized and deformation is controlled, thereby meeting the requirements for hardness, toughness and dimensional accuracy.

[0003] Existing solutions often employ: ① Fixed-parameter heat treatment equipment, processing based on preset temperature curves and time parameters, which is difficult to adapt to differences in raw materials from different batches; ② Manual monitoring and control, relying on personnel observation and manual adjustment, resulting in slow response and large errors; ③ Single-parameter feedback control, using only temperature feedback to adjust heating power without considering deformation and microstructure evolution information, making it difficult to balance performance and dimensional accuracy. These methods often lead to difficulties in ensuring performance and dimensional consistency, resulting in insufficient batch uniformity; setting high energy consumption redundancy to ensure quality does not meet the requirements of green and low-carbon production; and the process parameters cannot be dynamically adapted, making them susceptible to defects such as insufficient hardness and excessive deformation due to raw material fluctuations, furnace loading differences, and environmental changes.

[0004] This invention aims to solve the problems in the existing heat treatment of bearing rings, such as process control relying on fixed parameters or single index feedback, lack of multi-source information fusion and linkage parameter adjustment mechanism, response lag and large error, insufficient product consistency, high energy consumption redundancy, and quality defects caused by parameter mismatch. It provides an implementable, verifiable and traceable closed-loop control method that enables performance targets, size targets and energy consumption targets to be met in a coordinated manner. Summary of the Invention

[0005] To achieve the above objectives, this invention provides a method for controlling the heat treatment of bearing rings, applicable to a heat treatment process of heating, holding, and cooling. Its key feature is that it constructs a multi-layered closed-loop control system by real-time acquisition and fusion of multi-source information such as temperature, deformation, and microstructure evolution, superimposed with online thermal-phase change model identification and confidence gating. This system consists of "data quality assessment—state estimation—risk / hotspot assessment—constraint package generation—linked parameter tuning—closed-loop execution—replayback verification and parameter write-back," simultaneously satisfying performance and dimensional targets while also optimizing energy consumption.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A method for controlling the heat treatment of bearing rings includes at least the following steps: S1. Establishing Targets and Process Baselines: Obtain information on ring type / material / batch, set performance and dimensional targets, and establish baseline process parameters; The performance targets include at least hardness and / or toughness indicators, and the size targets include at least one of critical size deviation thresholds and allowable deformation thresholds. The baseline process parameters include at least one of the following: baseline heating curve, baseline holding time, and baseline cooling rate.

[0007] Optionally, an energy consumption or peak power budget can be set, including at least one of a unit energy consumption cap, a total power cap, or a phased energy quota.

[0008] S2. Multi-source information acquisition: During the heat treatment process, a multi-source data stream is formed, including temperature data T (temperature or temperature distribution at least one key location), deformation data D (ellipticity, end face warping, radial / axial displacement, etc., at least one), and equivalent state data S (equivalent indicators used to characterize the degree of austenitization, phase transformation process, or hardening tendency).

[0009] Optionally, actuator-side electrical parameters and operating condition information such as power, current, voltage, and cooling medium flow / pressure can be collected for consistency verification and energy consumption statistics.

[0010] S3. Data Preprocessing, Synchronization and Quality Gating: Perform time alignment and synchronization on multi-source data; identify, suppress or remove abnormal data such as out-of-bounds, mutations, packet loss, and drift; establish data quality indicators and confidence levels Q_T, Q_D, and Q_S based on the stability, consistency and historical errors of each data source, and output the comprehensive data quality Q.

[0011] Optionally, when the overall data quality Q falls below a threshold, a gating strategy is triggered, entering a degradation mode or freezing some estimates.

[0012] S4. Multi-source information fusion and state estimation: Based on the confidence level, T, D, and S are correlated and fused to obtain the fused temperature state T_f, fused deformation state D_f, and tissue evolution state S_f. The fusion can be achieved by weighted fusion, constrained consistency fusion, or fusion filtering, and the fusion confidence level Q_f is output for subsequent regulation and degradation.

[0013] S5. Online thermal-phase change model identification and freezing mechanism (optional but preferred): Based on the execution side electrical parameters and temperature response, online identification of model parameters related to thermal inertia and thermal efficiency is performed to form a thermal-phase change model ThermalModel, which includes at least one of the equivalent thermal inertia parameter, thermal efficiency parameter and hysteresis parameter, and outputs the model confidence level model_conf. When model_conf is below a threshold, the model parameters are frozen to prevent model drift from misleading the control under abnormal operating conditions.

[0014] S6. Risk Index and Hotspot Assessment: Based on information such as the integrated deformation state D_f, its changing trend, and temperature gradient, a comprehensive risk index RISK for deformation risk and process risk is constructed, and hotspots are identified. Hotspot zones are used to indicate critical stages or locations that lead to increased risk, while RISK is used to drive subsequent constraint package tightening, slope limiting, and parameter tuning priority switching.

[0015] S7. Deviation Calculation and Constraint Package Generation: The fused state is compared with the preset target to obtain the performance deviation e_p (characterized by the deviation of S_f relative to the target organizational state range), deformation deviation e_d (characterized by the deviation of D_f relative to the allowable deformation threshold / window), and energy consumption deviation e_e (characterized by the deviation of unit energy consumption or stage energy relative to the target / upper limit). The comprehensive deviation e_tot can be calculated for trade-off decisions. Based on this, a constraint pack ConstraintPack is generated, which includes at least one of the following: target temperature / target tissue state setpoint, upper limit of heating / cooling slope, deformation window / threshold, upper limit of cooling intensity, and power soft limiting. Among them, deformation-related constraints have higher priority than performance compensation-related constraints, and the tightening or loosening of the constraint package is adaptively scheduled based on RISK, model_conf, and Q_f, and a smooth transition strategy is adopted to avoid sudden changes in process parameters.

[0016] S8. Linked Adaptive Control Decision: Generate control quantities based on e_p, e_d, e_e and ConstraintPack, and dynamically adjust at least one of the heating power P, holding time t_hold, and cooling rate v_cool. The regulation must at least satisfy the following rules: S81, Deformation Priority Constraint: When e_d exceeds the first threshold or RISK increases, deformation is suppressed first, limiting the heating slope, reducing the upper limit of P, adjusting t_hold and / or reducing the cooling intensity, and if necessary, applying stricter local or stage limiting to the hot spot area. S82, Performance Compensation: When e_p indicates insufficient microstructure evolution, increase the effective heat input (increase P or extend t_hold) without violating deformation constraints; when e_p indicates excessive microstructure evolution, decrease the effective heat input and / or increase v_cool. S83, Energy Budget and Peak Shaving: When e_e triggers energy consumption constraints or the total power budget is limited, the energy input during the heating / heat preservation stage is segmented and optimized under the premise of meeting performance and size constraints. Peak shaving strategies (limiting peak power, segmented pulse hold, or time-sharing hold) can be used to reduce energy redundancy and reduce peak load.

[0017] S9. Closed-loop execution, self-adaptive update and replay verification: The control quantity is sent to the actuator and executed to form a closed loop by repeating S2-S8 in the control cycle Δt; when batch raw material differences, furnace loading differences or environmental changes are detected that cause abnormal deviation trends, the fusion weights, thresholds or strategy parameters are updated automatically, and the sampling results can be used for model / parameter library write-back updates. Optionally, playback verification is performed based on the recorded process data, and a playback consistency score is calculated. When the score continues to deteriorate, parameter rollback or conservative process is triggered to ensure long-term stability and traceability.

[0018] S10, Safety Interlock and Degradation: When the confidence level of any data source is lower than the threshold or data is lost, enter the degradation mode, freeze the corresponding state estimate and switch to the conservative temperature single closed loop and preset curve. When overheating, abnormal cooling medium, or abnormal power execution occurs, an interlock is triggered to limit heating power or terminate heating and maintain a safe cooling strategy.

[0019] The equivalent state data S of tissue evolution can be obtained through at least one of the following methods: a) Mapping online signals such as magnetic properties, ultrasonic propagation properties, resistivity changes, or acoustic emission characteristics into tissue state indices; b) Establish a lookup / regression model for "temperature-time-cooling rate → tissue state" based on historical data; c) When the online organization signal is unavailable, S=F(T_f, dT / dt, cumulative heat preservation time) is used for estimation, and the model parameters are updated by sampling and writing back.

[0020] Note: The step numbers S1-10 above are only used to describe the logical order for easy understanding. In actual implementation, the steps can be merged, split, rearranged or executed in a loop according to the equipment configuration and control strategy, but should not deviate from the core technical ideas and constraints of this invention. Beneficial effects

[0021] Compared with fixed parameters, manual monitoring, or single-parameter feedback control, the present invention has at least the following beneficial effects: (1) Performance and size coordination assurance: By integrating temperature-deformation-organization evolution multi-source fusion and dual-objective (performance / size) linkage parameter tuning, and introducing risk / hot spot driven constraint package, the organization meets the standard and deformation is controlled in the same closed loop, significantly reducing size deviation and performance fluctuation.

[0022] (2) Green, low-carbon and efficiency improvement: Introduce energy consumption / power budget and peak shaving allocation strategy to reduce energy consumption redundancy and reduce peak load under the premise of meeting quality constraints, and improve the process adaptation efficiency of batch processing.

[0023] (3) Enhanced process controllability and consistency: By using data quality gating, model confidence freezing, downgrading and replay verification and write-back mechanisms, the risk of loss of control caused by abnormal data or model drift is reduced, and cross-batch stability and traceability are improved. Detailed Implementation

[0024] The invention will be more readily understood through the following detailed description of preferred embodiments and included examples. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. In case of conflict, the definitions in this specification shall prevail.

[0025] The singular form includes the plural objects of discussion unless the context clearly indicates otherwise. "Optional" or "any one" means that the matter or event described thereafter may or may not occur, and the description includes both the possibility that the event occurs and the possibility that the event does not occur.

[0026] A method for controlling the heat treatment of bearing rings, applicable to the heating-holding-cooling heat treatment process of bearing rings, characterized by comprising the following steps: S1. Establish target and process baseline: Obtain the model, material and / or batch information of the bearing ring, set performance target and size target, and establish baseline process parameters; wherein, the performance target includes at least hardness and / or toughness index, the size target includes at least critical size deviation threshold and / or allowable deformation threshold, and the baseline process parameters include at least one of baseline heating curve, baseline holding time and baseline cooling rate; S2. Multi-source information acquisition: During the heat treatment process, multi-source data streams are acquired and formed. The multi-source data streams include at least temperature data T, deformation data D, and equivalent state data of tissue evolution S. S3. Data preprocessing, synchronization and quality gating: Time alignment and synchronization of the multi-source data streams are performed, and at least one of the following anomalies, such as out-of-bounds, mutation, packet loss and drift, is identified, suppressed or eliminated. The confidence scores Q_T, Q_D and Q_S of each data source are established and the overall data quality Q is output. S4. Multi-source information fusion and state estimation: Based on the confidence level, temperature data T, deformation data D, and tissue evolution equivalent state data S are correlated and fused to obtain fused temperature state T_f, fused deformation state D_f, and tissue evolution state S_f, and output the fusion confidence level Q_f. S5. Deviation Calculation and Constraint Pack Generation: The fusion temperature state T_f, fusion deformation state D_f, and tissue evolution state S_f are compared with the performance target and size target to obtain the performance deviation e_p and deformation deviation e_d, and a constraint pack ConstraintPack is generated. The constraint pack includes at least one of the following: target setting value, upper limit of heating slope, upper limit of cooling intensity, deformation window / threshold, and power soft limit. S6. Linked Adaptive Control Decision: Based on the performance deviation e_p, deformation deviation e_d and the constraint package ConstraintPack, a control quantity is generated to dynamically adjust at least one of the heating power P, holding time t_hold and cooling rate v_cool. S7. Closed-loop execution and self-adaptive update: The control quantity is sent to the actuator and executed to repeat S2 to S6 to form a closed loop in the control cycle Δt. When abnormal deviation trends caused by batch differences, furnace loading differences and / or environmental changes are detected, the fusion weight, threshold and / or strategy parameters are updated to achieve self-adaptive update.

[0027] In step S6, the linkage adaptive control decision must at least satisfy the following rules: When the deformation deviation e_d exceeds the first threshold, deformation is suppressed first, and at least one or more of the following measures are taken: limiting the heating slope, reducing the upper limit of heating power P, adjusting the holding time t_hold, reducing the cooling intensity and / or reducing the cooling rate v_cool. When the performance deviation e_p indicates insufficient microstructure evolution, at least one or both of the following should be taken without violating the deformation-related constraints: increasing the heating power P and / or extending the holding time t_hold. When the performance deviation e_p indicates excessive microstructure evolution, at least one or more of the following should be taken: decreasing the heating power P, shortening the holding time t_hold, and / or increasing the cooling rate v_cool.

[0028] Step S1 further includes setting an energy consumption budget or power budget, which includes at least one of the following: unit energy consumption limit, total power limit, and stage energy quota; and step S5 further obtains the energy consumption deviation e_e; and step S6, when the energy consumption deviation e_e triggers energy consumption constraints or the total power budget is limited, performs segmented optimization allocation of energy input for the heating and / or heat preservation stage to implement a peak shaving strategy; the peak shaving strategy includes at least one of limiting peak power, pulse holding, and / or time-sharing holding.

[0029] The quality gating in step S3 includes: when the overall data quality Q and / or the fusion confidence Q_f is lower than the threshold, entering the degradation mode, freezing the estimation of the tissue evolution state S_f and / or the fusion deformation state D_f, and switching to temperature single closed loop and preset process curve execution.

[0030] The equivalent state data S of organizational evolution is obtained through at least one of the following methods: A) Map at least one of the following online signals—magnetic property signals, ultrasonic propagation property signals, resistivity change signals, or acoustic emission characteristic signals—to a tissue state index; B) Establish a lookup table model and / or regression model for “temperature-time-cooling rate → tissue state” based on historical data, and output the tissue evolution state S_f using real-time T_f, t_hold and / or v_cool. C) When online tissue signals are unavailable, the tissue evolution state is estimated using S=F(T_f, dT / dt, cumulative holding time), and the parameters of the function F are updated by writing back using the sampling results.

[0031] A method for controlling the heat treatment of bearing rings further includes an online thermal-phase change model identification step, which is performed after step S4 and before step S6. This step includes: based on the response relationship between at least one execution-side electrical parameter (heating power, current, voltage) and the fused temperature state T_f, identifying at least one of thermal inertia parameters, thermal efficiency parameters, and hysteresis parameters online to form a thermal-phase change model (ThermalModel), and outputting the model confidence score (model_conf); freezing the parameters of the thermal-phase change model when model_conf is below a threshold to avoid model drift affecting the linkage adaptive control decision.

[0032] A method for controlling the heat treatment of bearing rings further includes a risk index and hotspot assessment step, which is performed after step S4 and before step S7, and includes: constructing a comprehensive risk index RISK based on the fused deformation state D_f and its changing trend and / or temperature gradient information, and identifying hotspots; wherein, the tightening or loosening of the constraint pack ConstraintPack is adaptively scheduled based on the comprehensive risk index RISK, the fused confidence level Q_f and / or the model confidence level model_conf.

[0033] The constraint pack in step S5 includes at least the upper limit of the heating slope, the upper limit of the cooling intensity, the deformation window / threshold, and the soft power limit, and the deformation-related constraints have a higher priority than the performance compensation-related constraints; the constraint pack adopts a smooth transition strategy when tightening or relaxing to avoid abrupt changes in process parameters.

[0034] Step S7 also includes playback verification and parameter write-back, specifically: recording heat treatment process data, and performing playback verification based on the recorded process data to obtain a playback consistency score; when the playback consistency score continues to deteriorate, triggering parameter rollback to conservative process parameters and / or entering a conservative process mode.

[0035] A method for controlling the heat treatment of bearing rings further includes a safety interlock control, specifically triggering an interlock when an over-temperature, abnormal cooling medium, and / or abnormal power execution are detected, and at least one or two of the following strategies are adopted: limiting the heating power P or terminating heating and maintaining safe cooling; and the control period Δt is 0.1s to 2s, and the deviation calculation uses a sliding window W of 5s to 120s.

[0036] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0037] Example 1: Multi-source fusion-model identification-constraint package linkage control in batch furnace heat treatment (1) System and Object: The heat treatment system includes a multi-source sensing module, a controller, and an actuator. The multi-source sensing module is used to collect temperature data T, deformation data D, and equivalent state data S of microstructure evolution; the controller is used to complete data synchronization, quality gating, fusion state estimation, online model identification, risk / hot spot assessment, constraint package generation, and linkage parameter adjustment; the actuator is used to adjust the heating power P, the holding time t_hold, and the cooling rate v_cool, and outputs power, electrical parameters, and cooling medium conditions for consistency verification.

[0038] (2) Target and baseline: Establish a parameter library, set the target hardness range and toughness requirements, and set the critical dimension deviation threshold and allowable deformation window; and set the baseline heating curve, baseline holding time and baseline cooling rate; optionally set the upper limit of energy consumption per unit and the upper limit of total power.

[0039] (3) Acquisition and synchronization: The control period Δt can be 0.1s to 2s; the temperature sampling frequency can be 1Hz to 50Hz; the deformation sampling frequency can be 0.5Hz to 20Hz; the trend and deviation are calculated using a sliding window W (e.g., 5s to 120s); outliers are removed according to the rules of out of bounds / mutations / inconsistencies to obtain Q_T, Q_D, Q_S and comprehensive quality Q.

[0040] (4) Fusion, Model and Risk: T_f, D_f and S_f are obtained by fusion according to confidence weight; ThermalModel is identified online based on power-temperature response and model_conf is output, and frozen when the confidence is low; RISK is constructed based on D_f and its trend and hot spot areas are identified to drive the tightening of constraint package.

[0041] (5) Constraint package and linkage parameter tuning: Generate a constraint package that includes the upper limit of the heating slope, the soft limit of power, the deformation window, the upper limit of the cooling intensity, etc. When e_d increases or RISK increases, the heating slope is limited first and the upper limit of P is reduced. If necessary, the cooling intensity is reduced to suppress deformation. Then, performance compensation is achieved by extending / shortening t_hold and fine-tuning v_cool, so that e_p returns to the target range.

[0042] (6) Energy consumption budget and peak shaving: When the energy consumption of a unit is close to the upper limit or the total power budget is limited, a peak shaving strategy is adopted under the premise of satisfying the e_p and e_d constraints. For example, the peak power of the heating section is reduced, and pulse maintenance or time-sharing maintenance is adopted in the heat preservation section to reduce energy consumption redundancy and peak load.

[0043] (7) Self-adaptation and playback verification: When batch raw material differences or changes in ambient temperature cause systematic shifts in e_p and e_d, update the fusion weight, threshold or S mapping model parameters; and can perform playback verification based on recorded process data, output playback consistency score, and trigger parameter rollback or conservative process when the score deteriorates.

[0044] (8) Degradation and Interlocking: When the overall quality Q or the fusion confidence Q_f is lower than the threshold, the degradation mode is entered, the tissue / deformation estimation is frozen and the temperature single closed loop and preset curve are switched; when over-temperature, power execution abnormality or cooling medium abnormality occurs, the interlocking is triggered, P is limited or heating is terminated and safe cooling is maintained.

[0045] Example 2: Implementation of the equivalent state of organizational evolution (optional) Method A: Online signal mapping from magnetic / ultrasound / resistivity / acoustic emission and other signals to tissue state index; the mapping model is obtained by calibration of historical samples.

[0046] Method B: Establish a lookup / regression model based on historical process and sampling data, and map real-time T_f, t_hold, and v_cool to S_f.

[0047] Method C: When there is no online organization signal, S=F(T_f, dT / dt, cumulative holding time) is used for estimation, and the parameters of F are updated by random inspection at the end of the batch.

[0048] Example 3: Evaluation Indicators and Validation The following indicators can be used for verification: dimensional deviation rate, performance compliance rate, unit energy consumption, peak power, parameter response delay, batch consistency error, playback consistency score, etc. By comparing with the fixed parameter scheme, the effects of dimensional deviation reduction, energy consumption reduction, consistency improvement and long-term stability enhancement can be verified.

[0049] The examples described herein are merely illustrative, intended to explain some features of the methods described herein. The appended claims are intended to claim the broadest possible scope, and the embodiments presented herein are merely illustrative of selected implementations based on combinations of all possible embodiments. Therefore, the applicant intends that the appended claims are not limited by the selection of examples illustrating the features of the invention. Some numerical ranges used in the claims also include sub-ranges within them, and variations within these ranges should be interpreted, where possible, as covered by the appended claims.

Claims

1. A heat treatment regulation method for bearing ring, which is suitable for the heating- holding- cooling heat treatment process of bearing ring, characterized in that, Includes the following steps: S1. Establish target and process baseline: Obtain the model, material and / or batch information of the bearing ring, set performance target and size target, and establish baseline process parameters; Wherein, the performance target includes at least hardness and / or toughness indicators, the size target includes at least a critical size deviation threshold and / or an allowable deformation threshold, and the baseline process parameters include at least one of the following: baseline heating curve, baseline holding time, and baseline cooling rate; S2. Multi-source information acquisition: During the heat treatment process, multi-source data streams are acquired and formed. The multi-source data streams include at least temperature data T, deformation data D, and equivalent state data of tissue evolution S. Where T represents the collected temperature data, D represents the collected deformation data, and S represents the collected or estimated equivalent state data of tissue evolution; T_f, D_f, and S_f represent the fused temperature state, fused deformation state, and tissue evolution state after time synchronization, quality gating, and multi-source fusion, respectively. "Cooling intensity" is used to characterize the overall level of cooling effect and can be characterized by at least one of the following operating parameters: cooling rate v_cool and the flow rate, pressure, or temperature of the cooling medium. 2.S3, Data Preprocessing, Synchronization and Quality Gating: Time alignment and synchronization are performed on the multi-source data streams, and at least one of the following anomalies, such as out-of-bounds, mutation, packet loss and drift, is identified, suppressed or eliminated. The confidence scores Q_T, Q_D and Q_S of each data source are established and the overall data quality Q is output. S4. Multi-source information fusion and state estimation: Based on the confidence level, temperature data T, deformation data D, and tissue evolution equivalent state data S are correlated and fused to obtain fused temperature state T_f, fused deformation state D_f, and tissue evolution state S_f, and output the fusion confidence level Q_f. S5. Deviation Calculation and Constraint Pack Generation: The fusion temperature state T_f, fusion deformation state D_f, and tissue evolution state S_f are compared with the performance target and size target to obtain the performance deviation e_p and deformation deviation e_d, and a constraint pack ConstraintPack is generated. The constraint package includes at least one of the following: target setpoint, upper limit of heating slope, upper limit of cooling intensity, deformation window / threshold, and power soft limit; S6. Linked Adaptive Control Decision: Based on the performance deviation e_p, deformation deviation e_d and the constraint package ConstraintPack, a control quantity is generated to dynamically adjust at least one of the heating power P, holding time t_hold and cooling rate v_cool. S7. Closed-loop execution and self-adaptive update: The control quantity is sent to the actuator and executed to repeat S2 to S6 to form a closed loop in the control cycle Δt. When abnormal deviation trends caused by batch differences, furnace loading differences and / or environmental changes are detected, the fusion weight, threshold and / or strategy parameters are updated to achieve self-adaptive update.

3. The method of claim 1, wherein: In step S6, the linkage adaptive control decision satisfies at least the following rules: when the deformation deviation e_d exceeds the first threshold, deformation is suppressed first, and at least one or more of the following measures are taken: limiting the heating slope, reducing the upper limit of the heating power P, adjusting the holding time t_hold, reducing the cooling intensity and / or reducing the cooling rate v_cool. When the performance deviation e_p indicates insufficient tissue evolution, at least one or both of the following should be taken without violating the deformation-related constraints: increasing the heating power P and / or extending the holding time t_hold. When the performance deviation e_p indicates excessive tissue evolution, at least one or more of the following should be taken: reducing the heating power P, shortening the holding time t_hold, and / or increasing the cooling rate v_cool.

4. The method of claim 1, wherein: Step S1 also includes setting an energy consumption budget or power budget, wherein the energy consumption budget or power budget includes at least one of the following: unit energy consumption limit, total power limit, and stage energy quota; Furthermore, step S5 further obtains the energy consumption deviation e_e, and step S6, when the energy consumption deviation e_e triggers energy consumption constraints or the total power budget is limited, performs segmented optimization allocation of energy input in the heating and / or heat preservation stage to implement peak shaving strategy. The peak clipping strategy includes at least one of peak power limiting, pulse hold, and / or time-division hold.

5. The method of claim 1, wherein: The quality gating in step S3 includes: when the overall data quality Q and / or the fusion confidence Q_f is lower than the threshold, entering the degradation mode, freezing the estimation of the tissue evolution state S_f and / or the fusion deformation state D_f, and switching to temperature single closed loop and preset process curve execution.

6. The method of claim 1, wherein: The equivalent state data S of organizational evolution is obtained through at least one of the following methods: A) Map at least one of the following online signals—magnetic property signals, ultrasonic propagation property signals, resistivity change signals, or acoustic emission characteristic signals—to a tissue state index; B) Establish a lookup table model and / or regression model for "temperature-time-cooling rate → tissue state" based on historical data, and output the tissue evolution state S_f using real-time T_f, t_hold and / or v_cool. C) When online tissue signals are unavailable, the tissue evolution state is estimated using S=F(T_f, dT / dt, cumulative holding time), and the parameters of the function F are updated by writing back using the sampling results.

7. The method of claim 1, wherein: It also includes an online thermal-phase change model identification step, which is performed after step S4 and before step S6. The online thermal-phase change model identification step includes: based on the response relationship between at least one execution-side electrical parameter of heating power, current, and voltage and the fusion temperature state T_f, identifying at least one of thermal inertia parameter, thermal efficiency parameter, and hysteresis parameter online, forming a thermal-phase change model ThermalModel, and outputting the model confidence level model_conf; When model_conf is below a threshold, the parameters of the thermal-phase change model are frozen to prevent model drift from affecting the linkage adaptive control decision.

8. The method of claim 1, wherein: It also includes a risk index and hotspot assessment step, which is performed after step S4 and before step S7, including: constructing a comprehensive risk index RISK based on the fused deformation state D_f and its changing trend and / or temperature gradient information, and identifying hotspots; The tightening or loosening of the constraint pack ConstraintPack is adaptively scheduled based on the comprehensive risk index RISK, the fusion confidence level Q_f, and / or the model confidence level model_conf.

9. The method of claim 1, wherein: The constraint pack in step S5 includes at least the upper limit of the heating slope, the upper limit of the cooling intensity, the deformation window / threshold, and the power soft limit, and the deformation-related constraints have a higher priority than the performance compensation-related constraints. The constraint package employs a smooth transition strategy when tightening or loosening to avoid abrupt changes in process parameters.

10. The method of claim 1, wherein: Step S7 also includes playback verification and parameter write-back, specifically: recording heat treatment process data and performing playback verification based on the recorded process data to obtain a playback consistency score; When the playback consistency score continues to deteriorate, the parameters are triggered to roll back to conservative process parameters and / or enter conservative process mode.

11. The method of claim 1, wherein: The control method also includes safety interlock control, specifically, when an over-temperature, abnormal cooling medium and / or abnormal power execution is detected, an interlock is triggered, and at least one or two of the following strategies are adopted: limiting heating power P or terminating heating and maintaining safe cooling. Furthermore, the control period Δt is 0.1s to 2s, and the deviation calculation uses a sliding window W of 5s to 120s.