A magnesium-aluminum alloy heat treatment process monitoring and adaptive control system

Abstract

The application relates to the technical field of intelligent heat treatment production lines, and discloses a magnesium-aluminum alloy heat treatment process monitoring and self-adaptive regulation system, which comprises a multi-modal sensing and preprocessing module, a digital twin parameter identification module, a predictive control decision module and a self-adaptive execution module. The system collects workpiece resistivity and acoustic emission signals in real time, dynamically corrects the effective activation energy by using the deviation between the measured resistivity and the theoretical value, predicts the future phase change trend in combination with a JMAK equation, and generates temperature regulation instructions. At the same time, an acoustic emission fuse judgment mechanism is executed: in the subcritical stress accumulation area, an adaptive gain damping strategy is executed; and in the crack initiation risk area, rapid cooling protection is forcibly triggered. The application can dynamically adjust the temperature regulation instructions according to the material microstate in real time, solves the problems of insufficient solid solution or coarse grains caused by traditional open-loop control, guarantees the quality of magnesium-aluminum alloy heat treatment workpieces, and meets the production requirements of high-quality and precise magnesium-aluminum alloy components.

Description

Technical Field

[0001] This invention relates to the field of intelligent heat treatment production line technology, specifically to a monitoring and adaptive control system for the heat treatment process of magnesium-aluminum alloys. Background Technology

[0002] Magnesium-aluminum alloys, due to their low density, high specific strength, and excellent casting properties, are widely used in aerospace and automotive lightweighting applications. In the production of magnesium-aluminum alloys, heat treatment is a crucial process that determines the final mechanical properties of the material. The purpose of this process is to remove the non-equilibrium eutectic phase (mainly...) from the alloy through heating. Mutually, Fully dissolve and enter In the matrix, casting stress is eliminated, significantly improving the plasticity and toughness of the material.

[0003] Existing heat treatment processes for magnesium-aluminum alloys typically rely on standardized temperature-time process curves for control. In actual production, operators set the heating rate, holding temperature, and holding time according to the work instructions, and the equipment adjusts the heating power through a PID controller to track the set temperature. This control mode is essentially an open-loop control based on the microscopic state of the material, assuming that all workpieces to be treated will produce completely identical physical and metallurgical reactions under the same process parameters, and assuming that the furnace environment and workpiece response perfectly conform to an ideal model.

[0004] However, the aforementioned existing technologies have some shortcomings in practical applications: Firstly, due to differences in impurity content, initial grain size, and casting cooling history among different batches of raw materials, the activation energy required for solid solution varies. Using a fixed process curve can easily lead to insufficient solid solution due to insufficient reaction kinetics, or grain coarsening or even overheating due to excessive heating. Secondly, during heat treatment, if microscopic damage or grain boundary slip occurs inside the workpiece due to thermal stress release, traditional temperature controllers cannot detect such abnormal evolution of microstructures and often continue to execute heating commands, leading to crack propagation or even workpiece scrap, failing to meet the production requirements of high-quality precision magnesium-aluminum alloy components. Therefore, a magnesium-aluminum alloy heat treatment process monitoring and adaptive control system is urgently needed to solve these problems. Summary of the Invention

[0005] To address the problems in related technologies, this invention provides a monitoring and adaptive control system for the heat treatment process of magnesium-aluminum alloys, thereby overcoming the aforementioned technical problems in existing related technologies.

[0006] To solve the aforementioned technical problem, the present invention is achieved through the following technical solution: In a first aspect, embodiments of the present invention provide a monitoring and adaptive control system for the heat treatment process of magnesium-aluminum alloys, specifically comprising: a multimodal sensing and preprocessing module for acquiring the dynamic parameters, instantaneous resistivity, and acoustic emission signals of the workpiece; measuring the reference resistivity of the workpiece at room temperature and eliminating power frequency noise using a moving average filtering algorithm; a digital twin parameter identification module for calculating the current theoretical resistivity based on the activation energy estimated at the previous moment, and calculating the effective activation energy estimated at the current moment by combining the deviation between the measured instantaneous resistivity and the current theoretical resistivity; a predictive control decision module for predicting the evolution trend of the phase transition fraction in a future preset time domain based on the effective activation energy estimated at the current moment, calculating the deviation ratio between the predicted rate and the ideal rate, and generating a temperature adjustment command by combining the thermistor; and an adaptive execution module for executing an acoustic emission melting judgment mechanism, determining the micro-damage risk level based on the intensity of the acoustic emission signal, executing the corresponding strategy, and finally issuing the final execution command.

[0007] As a preferred embodiment of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system of the present invention, the specific steps of the moving average filtering algorithm include: establishing a window with a length of... A first-in-first-out queue is used, and a recursive algorithm is employed to update the accumulated sum of the queue data. And based on the updated cumulative sum Calculate the smoothed resistivity at the current moment. As valid input data.

[0008] As a preferred embodiment of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system of the present invention, the current theoretical resistivity is calculated based on the JMAK equation, and its expression is: ; In the formula, For the current moment Theoretical resistivity, The reference resistivity of the workpiece at room temperature. is the temperature coefficient of resistance of magnesium alloy. The solid solution scattering coefficient is... Based on room temperature, This represents the total phase transition fraction at the current moment.

[0009] As a preferred embodiment of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system of the present invention, the effective activation energy estimated at the current moment is calculated using a dynamic activation energy adaptive correction formula, the expression of which is: ; In the formula, for Effective activation energy estimated at any time The initial theoretical activation energy, This is the gain coefficient. Noise tolerance factor, Adjust the integral weight for energy. The measured instantaneous resistivity Compared with the current theoretical resistivity The deviation.

[0010] In a preferred embodiment of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system of the present invention, the calculation formula for the temperature adjustment command is as follows: ; In the formula, For the current moment Temperature adjustment instructions For the target phase transition rate, The actual phase transition rate is based on real-time resistivity feedback. This is the gain coefficient. The gas constant is... The current temperature of the workpiece. To prevent the minimum value where the denominator is zero.

[0011] As a preferred embodiment of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system of the present invention, the determination of the micro-damage risk level includes: setting a first-level threshold. and secondary threshold ; If acoustic emission signal The system determines that the current microstructure is safe and directly executes the temperature regulation command. ;like The system determines that the workpiece faces the risk of overheating and cracking, and directly blocks the temperature adjustment command. The rapid cooling protection program is forcibly triggered, and an overburned fuse alarm is sent to the host computer; if The system determines that the workpiece is in the subcritical stress accumulation zone and executes an adaptive gain damping strategy based on risk gradient.

[0012] As a preferred embodiment of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system of the present invention, the adaptive gain damping strategy based on risk gradient specifically includes: constructing a penalty function to calculate the discount factor of the heating command. Using the discount factor Temperature adjustment command Make corrections to obtain the final execution instructions. :like This is determined to be a heating request, and will be executed. ;like If the request is determined to be a cooling or insulation request, it will be executed. .

[0013] As a preferred embodiment of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system of the present invention, the penalty function is based on the current acoustic emission signal. At the first threshold With secondary threshold The relative positions within the constructed risk interval are defined by the following expression: .

[0014] As a preferred embodiment of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system of the present invention, wherein the first-level threshold The determination method includes: continuously collecting data within the initial metallurgical quiescent zone during the heating phase. A set of background noise samples is formed by several acoustic emission root mean square voltage samples. The mean of the background noise sample set is calculated. and standard deviation According to the preset heat treatment time and control cycle frequency Calculate the false alarm probability ; Calculate the statistical significance factor corresponding to the false alarm probability using the inverse function of the standard normal distribution. The first-level threshold is calculated. The secondary threshold The first-level threshold Three times that.

[0015] Secondly, embodiments of the present invention provide an intelligent precision magnesium-aluminum alloy heat treatment equipment, specifically including: a vacuum heat treatment furnace body, a multi-channel resistivity acquisition instrument, a wideband acoustic emission detector, a PID power regulator, a main control computer, and a magnesium-aluminum alloy heat treatment process monitoring and adaptive control system. The main control computer is electrically connected to the multi-channel resistivity acquisition instrument, the wideband acoustic emission detector, and the PID power regulator. The multi-channel resistivity acquisition instrument is connected to the workpiece using a four-probe method, the wideband acoustic emission detector is coupled to the surface of the workpiece, and the PID power regulator is used to control the furnace body heating power according to the temperature adjustment command issued by the main control computer.

[0016] The present invention has the following beneficial effects: 1. This invention acquires the resistivity change of the workpiece in real time, and combines the deviation between the measured resistivity and the theoretical value to identify and correct the current effective activation energy of the material online. Unlike the rigid temperature-time fixed curve in the prior art, this invention can accurately calculate the actual solid solution progress (phase transformation fraction) inside the material based on the corrected micro parameters, and use model prediction control to dynamically adjust the heating command. This not only avoids insufficient solid solution caused by impurities or differences in grain size of raw materials, but also prevents grain coarsening caused by blind heating, which helps to ensure that the quality of each batch of workpieces meets the standards.

[0017] 2. This invention introduces broadband acoustic emission monitoring technology to capture high-frequency stress wave signals generated during grain boundary slip or microcrack initiation. By setting multi-level early warning thresholds, it can automatically intervene in the early stages of micro-damage: for slight stress accumulation, it adaptively reduces the heating rate to give the material time to recover; for emergency situations with the risk of overheating or cracking, it directly blocks the heating command and triggers rapid cooling protection, thereby solving the problem that existing temperature control systems cannot detect and respond to internal physical damage to materials, and ensuring the safety and quality of the heat treatment process in the magnesium-aluminum alloy heat treatment production line.

[0018] 3. This invention employs a moving average filtering algorithm to process resistivity signals, eliminating the influence of power frequency noise. Simultaneously, it resolves the issue of inconsistent sampling frequencies among multiple sensors. Through timestamp alignment and feature merging, it ensures the consistency of the input state vector in physical space and time, thereby improving the accuracy of parameter identification and preventing control misjudgments caused by signal noise or timing misalignment.

[0019] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the module of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system provided by the present invention.

[0022] Figure 2 The flowchart shows the method for monitoring and adaptive control of the heat treatment process of magnesium-aluminum alloys provided by the present invention.

[0023] Figure 3 This is a schematic diagram of model predictive control provided by the present invention.

[0024] Figure 4 A schematic diagram of the architecture of the intelligent precision magnesium-aluminum alloy heat treatment equipment provided by the present invention. Detailed Implementation

[0025] 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. Example

[0026] Existing magnesium-aluminum alloy heat treatment technologies typically employ fixed temperature-time process curves, ignoring the actual impact of batch-to-batch variations in raw materials on phase transformation kinetics. This open-loop control approach can easily lead to insufficient solid solution or overheating, resulting in coarse grains, and lacks a real-time protection mechanism against sudden microcracks.

[0027] To solve the above technical problems, such as Figure 1 As shown, Embodiment 1 of the present invention provides a monitoring and adaptive control system for the heat treatment process of magnesium-aluminum alloys, specifically including: a multimodal sensing and preprocessing module, a digital twin parameter identification module, a predictive control decision module, and an adaptive execution module. The multimodal sensing and preprocessing module is used to acquire the workpiece's dynamic parameters, instantaneous resistivity, and acoustic emission signals; measure the workpiece's reference resistivity at room temperature, and eliminate power frequency noise using a moving average filtering algorithm. The module is used to calculate the current theoretical resistivity based on the activation energy estimated at the previous moment, and calculate the effective activation energy estimated at the current moment by combining the deviation between the measured instantaneous resistivity and the current theoretical resistivity. The predictive control decision module is used to predict the evolution trend of the phase transition fraction within a preset time domain based on the effective activation energy estimated at the current moment, calculate the deviation ratio between the predicted rate and the ideal rate, and generate temperature adjustment commands by combining the thermistor factor. The adaptive execution module is used to execute the acoustic emission melting judgment mechanism, determine the micro-damage risk level based on the intensity of the acoustic emission signal, execute the corresponding strategy, and finally issue the final execution command.

[0028] Specific embodiment 1 involves performing T4 solution treatment on a batch of magnesium-aluminum alloy die-cast parts with grade AZ91D. The system is equipped with high-temperature platinum electrodes for resistivity monitoring and a broadband acoustic emission sensor is deployed to monitor grain boundary slip signals, providing the data basis for embodiment 1 of this invention.

[0029] In the specific implementation of Example 1, firstly, initialization and benchmark parameter calibration are performed. The benchmark for subsequent calculations is established by reading the standard dynamic parameters of the workpiece and measuring the benchmark resistivity at room temperature. Secondly, during the heating stage, dynamic data acquisition and preprocessing are performed. A moving average filtering algorithm is used to eliminate power frequency noise, and a global clock is used to timestamp and align the high-frequency acoustic emission characteristics with the low-frequency resistivity data, solving the data heterogeneity problem caused by asynchronous sampling from multiple sensors and ensuring the spatiotemporal consistency of the input data. Then, the current theoretical resistivity is calculated based on the JMAK equation. Combining the deviation between the measured resistivity and the theoretical value, the current effective activation energy is inverted in real time using the dynamic activation energy adaptive correction formula. This transforms the invisible microstructure evolution into a quantifiable real-time state variable, solving the problem that traditional techniques cannot adapt to the batch heterogeneity of materials. Afterwards, model predictive control (MPC) is executed based on the corrected activation energy parameters to predict the future phase transition evolution trend in the time domain. By calculating the deviation ratio between the predicted rate and the ideal rate and combining it with the thermistor, a temperature adjustment command is generated, achieving pre-emptive intervention and avoiding the risks of coarse grains or insufficient solid solution. Next, safety constraint verification and adaptive adjustment are performed. The system determines the level of micro-damage risk by monitoring the intensity of acoustic emission signals, implements an adaptive gain damping strategy to gently apply the brakes on subcritical stress accumulation, and implements rapid cooling protection to force melting and prevent crack initiation, thus endowing the system with the ability to sense material physical damage and proactively avoid risks. Finally, when the real-time calculated phase transformation fraction reaches the preset target, the system automatically terminates the heat treatment process and enters the quenching stage, archiving the effective activation energy evolution curve of the entire process as a micro-quality fingerprint, providing a third dimension of data support for product quality traceability in addition to macroscopic mechanical properties.

[0030] Furthermore, to better illustrate the technical solution of Embodiment 1 of the present invention, such as... Figure 2 As shown, a method for monitoring and adaptively controlling the heat treatment process of magnesium-aluminum alloys is proposed, and the heat treatment control system for magnesium-aluminum alloys is described in detail, including the following: S1. Initial stage: The system performs initialization and baseline parameter calibration, which includes the following sub-steps: S11. The magnesium-aluminum alloy workpiece is placed in a heat treatment furnace, and electrodes are spot-welded to non-critical parts of the workpiece using a four-probe method to eliminate the influence of contact resistance. Furthermore, to overcome contact instability caused by high-temperature thermal cycling, Pt-Rh alloy wire is used for the electrode leads. After spot-welding to the workpiece surface, an alumina-based high-temperature ceramic sealant is applied to prevent oxidation of the weld joint. Simultaneously, an S-shaped stress-relief bend is pre-formed at the electrode lead-out point to compensate for displacement caused by thermal expansion of the workpiece and prevent the weld joint from detaching due to shear force.

[0031] Simultaneously, the standard kinetic parameters of this grade of magnesium alloy were read, including the initial theoretical activation energy. Pre-exponential factors and Avrami index .

[0032] S12. Measure the reference resistivity of the workpiece at room temperature. And set a target phase transition fraction. ,For example The phase solubility reaches 95%.

[0033] S2, Heating Phase: Perform dynamic data acquisition and preprocessing, including the following sub-steps: S21. Start the heating program. The system collects the instantaneous resistivity of the workpiece in real time with a period of 100ms. Furthermore, considering environmental electromagnetic noise such as that from induction heating, a moving average filtering algorithm is used to smooth the original signal. The specific implementation steps are as follows: S21-1, The system allocates a window in memory with a length of... First-In-First-Out (FIFO) circular buffer queue Used to store historical resistivity data. Window length. Based on the operating frequency and sampling frequency of the induction heating power supply Determined. For example, when the sampling period is 100ms ( When the main noise is power frequency interference, set This is to cover the entire noise fluctuation cycle and achieve the effect of wave peaks and troughs canceling each other out.

[0034] S21-2, at the initial stage of the heating program startup During each sampling period, the system is in fill mode. No filtering is performed at this time; instead, the raw resistivity values ​​are displayed. Fill the queue in sequence At the same time, calculate this The sum of data This is used as the initial accumulator value.

[0035] S21-3, When the queue is full (i.e.) When a new raw sample value is obtained, the steady-state filtering phase begins. Upon arrival, perform the following operations: Slide out the old value: Retrieve the earliest entered data from the head of the queue. ; Slide into new value: The latest Store it at the tail of the queue and update the queue pointer; Update accumulator: Update the accumulated sum using a recursive algorithm. To avoid The formula for summing repeated data is: ; This recursive algorithm reduces computational complexity from Reduced to This ensures that even with limited embedded controller computing resources, it can still handle high-frequency sampling data within a large window. Furthermore, to eliminate accumulated floating-point errors, all data in the queue is summed again at fixed intervals (e.g., every 1000 samples), resetting the process. .

[0036] S21-4, Based on the updated cumulative sum Calculate the smoothed resistivity at the current moment. : ; Ultimately, this will Used as valid data to be transmitted to step S3.

[0037] S22. Synchronously acquire acoustic emission signals, calculate root mean square voltage (RMS) and cumulative ring count, and align the timestamps with resistivity data to generate a state vector. The specific implementation steps are as follows: S22-1. The system continuously acquires internal stress wave signals from the workpiece using a wideband acoustic emission sensor, with an operating frequency range of 100kHz-1MHz. The stress wave signals are amplified by a preamplifier (gain, e.g., 40dB) and then sent to a high-speed analog-to-digital converter (ADC) at a sampling rate of... (For example, 2 MSPS, or 2 million samples per second) converts analog signals into digital sequences. .

[0038] Set a hard float threshold Used to filter out ambient background noise ,in The system only processes amplitudes exceeding Valid waveform data.

[0039] S22-2. To match the resistivity acquisition period, the system does not directly store massive amounts of raw waveforms, but instead uses... (like () represents a time window, and the content within that window... Statistical characteristics are calculated for each sampling point. Points: Calculate the root mean square voltage (RMS): This reflects the average energy intensity of the acoustic emission signal within the given time window, characterizing the overall activity level of microscopic processes. The calculation formula is as follows: ; In the formula, For the first RMS value for each control cycle This represents the voltage values ​​of discrete sampling points within the window.

[0040] Calculate the cumulative ring count This is used to reflect the number of bursts of microcrack propagation or grain boundary slip: the system counts the number of signal waveforms exceeding the threshold within this 100ms window. The sum of the number of oscillations is denoted as . .

[0041] S22-3. The system establishes a global master clock.

[0042] Low-frequency data labeling: when resistivity and temperature In the When a data point is collected, the system timestamp is recorded. .

[0043] High-frequency feature merging: The system will merge the high-frequency features within the time interval. Acoustic emission characteristics calculated internally and Forced merging to timestamp superior.

[0044] This method solves the problem of data asynchrony caused by inconsistent sampling rates of multiple sensors, and ensures the instantaneous correspondence of the subsequently constructed state vectors in a physical sense.

[0045] S22-4. Assemble the aligned data of each dimension into the first... The enhanced state vector at each time step : ; Store the vector in a first-in-first-out (FIFO) queue for use in subsequent step S3.

[0046] In this embodiment 1, the high-frequency sampled acoustic emission characteristics are mapped to the low-frequency sampled resistivity / temperature time axis by using the global clock timestamp forced alignment mechanism. This solves the problem of data asynchrony caused by inconsistent sampling frequencies between heterogeneous sensors. It not only filters out power frequency noise interference in the induction heating environment, but also ensures the consistency of the state vector input to subsequent modules in physical space and time, preventing causal inversion or control misjudgment caused by signal timing misalignment.

[0047] S3. Calculate the current theoretical resistivity based on the JMAK equation, and update the estimated effective activation energy in real time according to the deviation between the measured resistivity and the theoretical value using the dynamic activation energy adaptive correction formula. This includes the following sub-steps: S31, Activation energy estimated based on the previous moment. The theoretical resistivity at the current moment is calculated using the JMAK equation. The specific implementation steps are as follows: S31-1, The system reads the cumulative phase transition fraction stored at the end of the previous control cycle from memory. (i.e., the percentage of solid solution already achieved) and the activation energy identified at the previous moment. Simultaneously, obtain the current measured furnace temperature / workpiece temperature. .

[0048] S31-2. Using the Arrhenius equation and the activation energy estimated at the previous moment, calculate the instantaneous reaction rate constant at the current temperature. : ; In the formula, The pre-exponential factor is the preset factor. is the gas constant.

[0049] S31-3. Considering that heat treatment is a variable-temperature process, a correction factor is introduced. The current control period is calculated using the differential form of the JMAK equation. Phase transition fraction increment within : ; ; In the formula, Indicates the rate of temperature increase; For standard reference heating rate (e.g., 10°C / min) This is the thermal hysteresis coefficient; Then update the total phase transition fraction at the current moment: ; In the formula, This is the Avrami index. And considering... The value may be less than 1, in order to avoid the formula Item in When a mathematical singularity occurs, the system sets a phase transition initiation threshold. When the cumulative phase transition fraction of the previous moment At that time, the system will force a recognition. Substitute the values ​​into the formula to ensure the numerical stability of the initial reaction rate calculation.

[0050] For example, in this embodiment 1, the thermal hysteresis coefficient The method was obtained through experimental fitting, and the steps are as follows: Circular samples with a diameter of 4 mm and a thickness of 1 mm were cut from the same batch of as-cast AZ91D magnesium-aluminum alloy. The sample surface was polished to a bright finish, oxide scale was removed, and quality was controlled within specified limits. Within [a certain range]. The DSC instrument was calibrated for temperature and heat flux using high-purity indium (In) and zinc (Zn) standards. The heating rate was set to [a certain value]. (In engineering, this rate is often used to approximate the equilibrium state). Perform a DSC scan at room temperature. Record the heat flow curve and integrate it to obtain the baseline phase transition fraction curve. and reaction rate curve Set up at least three sets of heating rates covering the actual production range for the experiment: Experimental group A: Experimental Group B: Experimental group C: (Or set according to equipment limits). Record heat flux data for each group separately and calculate the corresponding values. and Because the temperature at which the same degree of solid solution (phase transition fraction) is reached differs under different heating rates, it is necessary to use the phase transition fraction as the determining factor. Align the anchor point, rather than using temperature as the anchor point. Select a characteristic phase transition fraction point; a specific value is recommended. At this point (the reaction is most intense, and the data signal-to-noise ratio is highest). In the baseline group ( Find in the data Corresponding physical temperature And record the reaction rate at this time. In experimental groups A, B, and C, respectively, [the following were found] The corresponding physical temperature and the corresponding instantaneous reaction rate Then to Fitting: To calculate the correction factor observed in the experiment For each heating rate ( This requires calculating the degree of attenuation relative to the reference rate, and theoretically, if there is no hysteresis, the rate should be... If they increase proportionally, then the normalized rate ratio is: ,like This indicates no lag; usually This indicates the existence of thermal hysteresis. Transforming the formula into a linear form... : ; make: , The data representation is constructed as follows: Using the least squares method By performing a linear fit, the slope of the resulting straight line is... .

[0051] S31-4. Using the physical mapping model, the calculated microscopic phase transition fractions are... Transformed into macroscopic physical quantities : ; In the formula, Theoretical resistivity; The room temperature reference resistivity is measured in step S1. For magnesium alloys, the temperature coefficient of resistance (TCR) is used. The solid solution scattering coefficient; Based on room temperature; The total phase transition fraction at the current moment is calculated in step S31-3.

[0052] Final output This refers to the resistivity value that should theoretically be observed at the current temperature and the current degree of solid solution.

[0053] For example: In this embodiment 1, the room temperature reference... The determination method is as follows: Before the heating power output, the system continuously collects the furnace temperature. and workpiece surface temperature The sampling period is 100ms. Calculate the rate of temperature change over the last 60 seconds. and the temperature difference between the workpiece and the environment Only if the following conditions are met: and At this point, the system determines that the workpiece has reached thermal equilibrium. The arithmetic mean of the workpiece temperature over the most recent 10 seconds is then taken as the equilibrium temperature. And simultaneously record the resistivity at this time as .

[0054] In this Example 1, the temperature coefficient of resistance of the magnesium alloy This characterizes the linear contribution of lattice thermal vibrations caused by temperature increase to resistivity, and its determination method is as follows: within the pure thermal effect range of the heating process (i.e., to At this point, the atomic diffusion coefficient is extremely low and can be considered as solid solubility. Constant and unchanging, that is Or the initial state. The system collects discrete data point pairs within this interval in real time. The least squares method was used to perform linear regression analysis on the data, and the equation was constructed as follows: The calculated slope This refers to the rate of change of resistivity due to pure thermal effects. At this point, the temperature coefficient of resistance... The calculation formula is: .

[0055] In this embodiment 1, the solid solution scattering coefficient It represents The change in resistivity caused by the increased degree of lattice distortion due to phase dissolution can be determined as follows: Select a standard sample of the same grade (e.g., AZ91D) and prepare two extreme states: State A (fully precipitated state): After sufficient annealing, assuming the solid solubility is... Its resistivity was measured at room temperature. ; State B (fully dissolved): After prolonged high-temperature solution treatment and water quenching, assuming that at this point... The phase completely dissolves, and the solid solubility reaches its maximum value. (Determined by chemical composition analysis of the alloy, e.g., Al content 9%). Its resistivity was measured at room temperature. State B was prepared under a protective atmosphere, and the surface was polished to remove the oxide layer. Differential calculation: Using Matthiassen's rule, background values ​​of phonon scattering and grain boundary scattering are eliminated to calculate the solid solution scattering coefficient. : ; In the formula, This represents the total atomic percentage of aluminum in the alloy. This represents the limiting solubility of the matrix at room temperature (which is usually negligible or can be taken as the theoretical value from the phase diagram). Calculated The inherent material constants of this alloy grade are stored in the system database. During actual heat treatment, the system reads the workpiece grade (e.g., AZ91D) and automatically loads the corresponding material constants. value.

[0056] S32. Calculate the measured instantaneous resistivity. With theoretical resistivity deviation : .

[0057] S33. Calculate the estimated effective activation energy at the current moment using the dynamic activation energy adaptive correction formula. The formula is: ; In the formula, for Effective activation energy estimated at specific times; This is the initial theoretical activation energy; is the gain coefficient (dimensionless); Noise tolerance factor; Adjust the integral weight for energy.

[0058] For example: In this embodiment 1, the noise tolerance factor The calibration method is as follows: After system initialization and before heating power is turned on, the system maintains a 10-second silent sampling period to obtain... A static resistivity sample Calculate the standard deviation of this sample set. : ; according to The criterion (99.7% confidence level) sets the criteria for determining a valid signal. Defined as dimensionless normalized noise level: .

[0059] In this embodiment 1, the gain coefficient Determines activation energy The maximum correction range is determined as follows: the system database stores the activation energy fluctuation range of this grade of magnesium alloy (such as AZ91D) observed in existing literature or experiments. Let the standard value be... (e.g., 135 kJ / mol), the physically permissible minimum activation energy is (Energy barrier at complete grain boundary liquefaction, e.g., 115 kJ / mol), maximum activation energy is (Energy barrier when impurities are highly segregated, such as 155 kJ / mol). Calculate the maximum drift rate: Set the maximum drift rate .

[0060] In this embodiment 1, the energy correction integral weight This characterizes the sensitivity of rate deviation to energy correction, and its determination method is as follows: According to the Arrhenius equation Taking the logarithm of both sides of the equation and differentiating, we obtain the relationship between the rate of change of reaction rate and the change of activation energy: ; The result of the conversion is: ; Integral terms in the dynamic activation energy adaptive correction formula Equivalent in dimensional terms to the cumulative relative rate deviation Therefore, in order to achieve dimensional balance, the dimensionless deviation is converted into the energy unit J / mol. It should be equal to the system's thermal energy reference: ; In the formula, It is the gas constant (8.314 J / (mol·K)). The target solution temperature set for the current heat treatment process (e.g., 693K) is then J / mol = 5.760 kJ / mol.

[0061] Specifically, for example: assuming that for AZ91D magnesium-aluminum alloy, the system presets an initial activation energy. Gain coefficient Noise tolerance factor and reference resistivity During the heat treatment process... At that time, it is assumed that the integral term of the relative error between the measured rate of change of resistivity and the model prediction over a past period is... The system detected the measured resistivity. The resistivity deviation was calculated to be lower than the theoretical value. The system is calculated using the dynamic activation energy adaptive correction formula: , Next, calculate the hyperbolic tangent: Last update: Current effective activation energy .

[0062] In this embodiment 1, the measured resistivity Compared with theoretical value deviation As a feedback signal, the effective activation energy at the current moment is updated in reverse through a nonlinear adaptive algorithm. This transforms the invisible microstructure evolution process in traditional heat treatment into an observable and quantifiable real-time variable, enabling the control system to accurately perceive the actual reaction barrier changes inside the material. This provides high-fidelity initial values ​​for subsequent model predictive control (MPC), solving the problem that traditional open-loop control cannot cope with material heterogeneity.

[0063] S4. Model Predictive Control (MPC) based on modified parameters predicts future evolution trends, specifically including the following steps: S41. Estimated effective activation energy Substituting into the JMAK equation, we predict that, assuming the current furnace temperature remains constant, the future... Phase transition fraction after minutes The specific implementation steps are as follows: S41-1, Based on the current time Measured furnace temperature and estimated effective activation energy The reaction rate constant is calculated using the Arrhenius equation if the current temperature remains constant. : .

[0064] S41-2. Since the previous heat treatment process was at varying temperatures, the system needs to calculate the equivalent time. Its physical meaning is: assuming that from the beginning at the current temperature... Under constant temperature treatment, the current accumulated phase change fraction is reached. The required time. Solving using the inverse operation of the JMAK equation. : ; In the formula, The total phase transition fraction updated in real time in step S31-3; This is the Avrami exponent. And if... If it is close to 0, then let ;when At that time, it will no longer be updated. The value from the previous time step is used to enhance robustness.

[0065] S41-3, in equivalent time Based on this, the prediction time domain is overlaid. (For example, predicting the next 5 minutes, i.e.) Substituting into the standard JMAK equation, calculate the phase transition fraction at future time points. : ; S41-4. To provide quantitative basis for subsequent temperature control, the system further calculates the deviation ratio between the predicted rate and the ideal rate: ; like This indicates that, based on the current temperature, the future solution treatment progress will lag behind the ideal process, requiring temperature compensation.

[0066] S42. Combine the thermosensitive factor to generate a temperature regulation command. : ; In the formula, For the target phase transition rate, The actual phase transition rate is based on real-time resistivity feedback. This is the gain coefficient. The gas constant is... The current temperature of the workpiece. To prevent the minimum value where the denominator is zero; It is a thermosensitive factor that characterizes the ability of temperature changes to affect the reaction rate.

[0067] Furthermore, considering the background noise present in sensors during actual engineering, in order to prevent the target phase transition rate from being affected... When the value approaches zero, the denominator in the formula becomes too small, causing even a weak noise signal to be amplified infinitely, thus triggering violent oscillations in the control command and affecting the minimum value in the formula. and calculation results Implement a stability constraint mechanism. Specifically, Not only does it serve as a mathematical correction term to prevent the denominator from being zero, but it is also defined as the system noise tolerance threshold (its value is not lower than...). (that is, slightly higher than the background noise floor) is used to forcibly limit the control gain in the low-rate range; at the same time, the system controls the output temperature adjustment command. Execute amplitude saturation limit: when the calculated temperature regulation command When the temperature exceeds the positive safety step size (e.g., +0.5℃), the system forcibly outputs that positive safety step size as a command; when the calculated temperature adjustment command... When the temperature is less than the negative safety step size (e.g., -0.5℃), the system will forcibly output the negative safety step size as an instruction to prevent temperature overshoot or oscillation caused by excessively large calculation results, avoid system divergence caused by mathematical singularities, and ensure the robustness of control throughout the heat treatment process.

[0068] For example: In this embodiment 1, the target phase transition rate This represents the solid solution rate that should be achieved at the current moment under perfect, ideal conditions. The method for determining this rate is as follows: The system reads the standard process temperature profile of this grade of magnesium alloy from the database. (i.e., the ideal heating path specified in the process card) and standard kinetic parameters ( The reaction rate constant at standard temperature was calculated using the Arrhenius equation. : ; The instantaneous phase transition rate under standard conditions is calculated based on the differential form of the JMAK equation. : ; In the formula, The phase transition fraction that should accumulate at standard temperatures, calculated in parallel by the system.

[0069] In this embodiment 1, the actual phase transition rate The actual solid solution rate occurring within the material is represented and determined as follows: Using the physical mapping model in step S31-4, the measured resistivity is... The lattice thermal vibration components in the crystal were stripped to extract the net resistivity caused solely by the phase transition. : ; In the formula, This is the current measured temperature. The calibrated temperature coefficient of resistance; Using the calibrated solid solution scattering coefficient The net resistivity is converted into the current real-time phase transition fraction. : ; Logical constraints: If the calculation result (Due to noise fluctuations), force set to 0; if Forced to set to 1; Real-time phase transition fraction Differentiation is performed to obtain the rate. Furthermore, to suppress the amplification effect of high-frequency noise on the differentiation operation, a five-point Savitzky-Gore smoothing differential filter is used. ; In real-time systems, the above formula can be easily applied using backward differential combined low-pass filtering: .

[0070] In this embodiment 1, the gain coefficient The method for determining it is as follows: Real-time calculations are performed using data from the heating phase (non-phase transition region, such as 150°C-200°C). According to a variant of Newton's law of cooling, the thermal response hysteresis of the workpiece can be characterized by the difference between furnace temperature and surface temperature. ; This represents the characteristic time required for the workpiece temperature to respond to changes in furnace temperature under the current furnace heat capacity and workpiece load. The system selects data from 10 consecutive seconds during the heating phase, calculates the average value of the above formula, and locks it as the inherent thermal inertia constant for the current furnace cycle. Read the set control sampling period (i.e., 100ms = 0.1s). According to the stability criterion of discrete control systems, [the following is a calculation method]... Defined as a single-period response factor: ; when Time (Inherent thermal inertia constant of industrial furnace) Typically on the order of 100 seconds, that is... The above formula can be Taylorically expanded into a more intuitive linear form: .

[0071] Specifically, for example: suppose a certain batch of furnaces was measured , Workpiece heating rate Calculate the thermal time constant: Known control period Calculate the gain coefficient. Meanwhile, in this embodiment 1, the current time is set. The workpiece temperature is at the target solid solution plateau. ( The effective activation energy estimated in real time is: The system detected a hysteresis in the solid solution reaction rate: the target phase transition rate The actual rate based on resistivity feedback Based on the gain coefficient calculated above... and gas constant Substitute into the formula to calculate , Finally, the temperature adjustment command is generated: This result indicates that the system will add temperature to the original set temperature within the current control cycle. The positive incremental value.

[0072] S43. Logical Judgment: If This indicates that the current solution rate is too slow, and the system issues a heating command; if This indicates that the reaction is too fast and there is a risk of coarse grains. The system then issues a heat preservation or cooling command.

[0073] In this embodiment 1, as Figure 3 The model predictive control diagram shown illustrates how phase transition fractions are calculated in advance based on current trends to predict future times. and ideal target score The evolutionary deviation between them is mapped to the current temperature regulation command through the thermosensitive factor. Overcoming the inherent challenges of thermal inertia hysteresis and nonlinear material response in industrial heating furnaces, temperature control is transformed from traditional post-error correction to pre-trend intervention, thereby avoiding process oscillations or overshoots (such as grain coarsening) caused by slow control response and ensuring that the solid solution path of the workpiece can conform to the optimal kinetic curve.

[0074] S5. Perform security constraint verification and adaptive adjustment, which includes the following steps: S51, during execution Before issuing the command, check the acoustic emission signal. ,in The corresponding current control cycle calculated in step S22-2 discrete values .

[0075] Set a first-level threshold and secondary threshold Execute the acoustic emission fuse triggering mechanism: like The system determines that the current microstructure is safe and allows the execution of the heating / holding command calculated in step S4. .

[0076] like The system determines that the workpiece faces the risk of overheating or cracking. At this point, regardless of the prediction result of step S4, even if the model requires a temperature increase, the main controller will directly block the temperature increase command and forcibly trigger the rapid cooling protection program, such as opening the nitrogen rapid cooling valve, and at the same time sending an overheating and melting alarm to the host computer.

[0077] like The system determines that the workpiece is in the subcritical stress accumulation zone. At this time, the material is undergoing drastic stress adjustment through lattice distortion and dislocation movement. Irreversible failure has not yet occurred, but it is extremely sensitive to thermal shock. An adaptive gain damping strategy based on risk gradient is executed, with the following steps: The system constructs a penalty function to calculate the discount factor for the heating command based on the relative position of the current acoustic emission value within the danger zone. ; This formula characterizes: when Just over hour, This allows the system to retain most of its heating capacity. Approaching hour, The system completely suppresses the heating capacity, forcing the temperature rise to stop; Then, the damping coefficient is used to adjust the original temperature control command output in step S4. Make corrections to obtain the final execution instructions. : If the original instruction is to increase the temperature ( ),but: This method reduces the heating rate while maintaining the heat treatment process by gently applying the brakes, giving the material enough time to release internal stress through creep and recovery mechanisms. If the original instruction was to cool down or keep warm ( Since cooling or insulation helps reduce thermal stress, no restrictions are imposed, and full permission is granted: .

[0078] For example, in this embodiment 1, the secondary threshold Characterizing grain boundary liquefaction or crack initiation, set as Level 1 threshold The method for characterizing the initiation of minute slip at grain boundaries is as follows: In the initial range of the heating phase ( to The system marks this area as the metallurgical silent zone. Within this zone, the system continuously collects data at a period of 100ms. Each acoustic emission root mean square voltage sample, for example, collected for 5 minutes, corresponds to These points constitute the background noise sample set. ; Assuming the background noise follows a normal distribution, the system calculates the mean of the sample set online. and standard deviation : ; ; Set false alarm probability This is used to identify abnormal signals. It requires a typical heat treatment process (duration...) The expected number of false alarms due to random noise is less than 1 per hour (e.g., 10 hours). Calculation : ; In the formula, To control the periodic frequency (e.g., 10Hz). For example: if If the sampling rate is 10Hz, then the total number of sampling points is Therefore ; Using the inverse function of the standard normal distribution (i.e., the quantile function) ), calculate the statistical significance factor corresponding to the false alarm rate. : ; The first-level threshold is determined as follows: .

[0079] Furthermore, to address the non-stationary nature of background noise drifting over time during long-cycle heat treatment operations, a threshold adaptive dynamic adjustment mechanism is introduced: the system maintains a sliding time window (e.g., 300s) in memory, adjusting the threshold only when the acoustic emission signal... Background noise samples are collected and updated during the metallurgical quiet period when the noise level is below the current first-level threshold. The statistical mean of the background noise is periodically updated using the exponentially weighted moving average (EWMA) algorithm. with standard deviation Based on this, the time-varying first-level threshold is calculated in real time. This mechanism ensures that when the furnace enters a stable operating period and the ambient noise level decreases, the system automatically lowers the threshold to improve the sensitivity to capture weak damage signals. When environmental interference temporarily increases, the system automatically raises the threshold to suppress false alarms, thereby achieving robust adaptation to non-steady-state industrial noise environments.

[0080] S52, The controller will ultimately execute the instructions. The signal is sent to the power regulator, which adjusts the power of the heating rod to intervene in the flow field inside the furnace.

[0081] In this embodiment 1, before sending temperature control commands to the underlying hardware, a sound emission melting judgment mechanism based on the material health status is set up to compare in real time. With grading threshold Heating request from MPC module Nonlinear modulation is performed: for micro-stress accumulation, a penalty function is used. Flexible damping is applied to allow for stress release time by gently applying the brakes; for risks of crack initiation, heating commands are forcibly blocked and rapid cooling is triggered. This method endows the control system with the ability to sense internal physical damage to the material, solving the problem of crack propagation or even workpiece scrapping caused by the inability of traditional temperature control systems to detect grain boundary slip or signs of overheating, thus maximizing process efficiency while ensuring heat treatment safety.

[0082] S6. Perform cooling and data archiving, which includes the following steps: S61, When the system calculates the real-time phase transition fraction Achieve the preset target phase transition fraction If the heat treatment process is interrupted, the process should immediately end and the quenching stage should begin.

[0083] S62. The system automatically saves the entire process. The evolution curve serves as a microscopic quality fingerprint for this batch of materials. If the curve fluctuates wildly, it indicates poor uniformity of the raw materials, providing a third dimension of data support for subsequent quality traceability, in addition to hardness and strength. Example

[0084] As a second embodiment of the present invention, based on embodiment 1, an intelligent precision magnesium-aluminum alloy heat treatment equipment is also disclosed, such as... Figure 4 As shown, the system specifically includes: a vacuum heat treatment furnace body, a multi-channel resistivity acquisition instrument, a broadband acoustic emission detector, a PID power regulator, a main control computer, and the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system described in Embodiment 1 above. The main control computer runs the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system and is electrically connected to the multi-channel resistivity acquisition instrument, the broadband acoustic emission detector, and the PID power regulator, respectively.

[0085] The vacuum heat treatment furnace body is used to accommodate magnesium-aluminum alloy workpieces and provide a controlled temperature field; the multi-channel resistivity acquisition instrument acquires the voltage and current signals of the workpiece in real time using the four-probe method to obtain the resistivity; the broadband acoustic emission detector is coupled to the workpiece surface or furnace wall waveguide rod to monitor micro-stress waves; the PID power regulator is used to adjust the output power of the heating element according to the instructions of the main control computer.

[0086] In the specific implementation of the above embodiment 2, the main control computer starts and executes the following process: During the process preparation and initialization phase, the magnesium-aluminum alloy workpiece is loaded and fixed, and the system automatically performs an electrode contact impedance self-test. Once the furnace temperature stabilizes at the reference temperature... At that time, the system records the initial resistivity. And load the standard kinetic parameters of the alloy grade from the database. and solid solution scattering coefficient Complete system initialization.

[0087] During the heat treatment heating and holding stages, the system enters real-time monitoring mode. The multi-channel resistivity acquisition instrument and the broadband acoustic emission detector work synchronously, and the main control computer uses a global clock to collect resistivity data. Acoustic emission characteristics Timestamp alignment is performed. The system is based on the deviation between measured resistivity and theoretical model values. The current effective activation energy is inverted in real time using the dynamic activation energy adaptive correction formula. Based on this, the JMAK model is updated to predict future phase transition paths.

[0088] Depending on the different microscopic evolution states inside the workpiece, the system automatically switches between the following scenario logics: Scenario 1, Normal Solution Treatment Period: When the root mean square voltage of acoustic emission... Below the first-level threshold At that time, it is determined that there is no risk of internal damage to the workpiece. The system executes model predictive control (MPC) decision-making. The main control computer calculates the current actual phase transition rate. With the target phase transition rate The deviation, combined with the predicted horizon The evolution trend within the body, calculating temperature regulation commands. .

[0089] If a delay in the solid solution process is detected, then the effective activation energy is... Too high a temperature hinders atomic diffusion, prompting the system to generate a positive temperature regulation command. The system instructs the PID power regulator to slightly increase the temperature based on the current set temperature, providing additional thermal driving force. If the solution treatment process is found to be ahead of schedule (risk of grain coarsening), the system generates a negative command to reduce the heating power. Through this dynamic fine-tuning, the solution treatment path of the workpiece is ensured to conform to the ideal process curve.

[0090] Scenario 2, Damage Risk Period: When abnormal fluctuations in acoustic emission signals are detected, the system intervenes through a multi-level early warning and circuit breaker mechanism: Subcritical state: if This indicates that grain boundary slip or stress concentration has occurred. The system calculates the penalty function. The temperature adjustment commands originally generated by the MPC module Applying damping forces a reduction in the heating rate, giving the material time to release stress through creep, thus achieving flexible braking.

[0091] Critical danger state: if This indicates a risk of overheating liquefaction or crack propagation. Regardless of the current solution treatment progress, the main control computer immediately blocks all heating commands and triggers a rapid cooling protection program (such as opening the nitrogen rapid cooling valve), while simultaneously alarming the host computer to prevent workpiece scrap.

[0092] Furthermore, to verify the stability and repeatability of the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system described in this invention in actual production, this embodiment designed a set of parallel comparative experiments: Experimental subjects: 500 batches of AZ91D magnesium-aluminum alloy die-cast parts were selected, with 10 samples randomly chosen from each batch, totaling 5000 parts. To simulate raw material fluctuations in actual production, the initial grain size of these 500 batches of raw materials was... Randomly distributed among impurity elements (Fe, Cu, Ni), the total content of these elements is within [a certain range]. Fluctuations between.

[0093] Control group: Utilized traditional time-temperature PID open-loop control. Strictly adhered to standard process: temperature increased to... Keep warm for 12 hours, with the heating rate set to [value missing]. .

[0094] Experimental group: Employing the adaptive control system described in this invention. Target phase transition fraction set. This allows the system to calculate the effective activation energy in real time. Automatically adjusts the heating path and holding time.

[0095] After the evaluation index experiment, all samples underwent physicochemical testing, and two key quality indicators were selected for statistical analysis: Solid solubility (%): Characterizes whether the heat treatment was sufficient. Target value is... The higher the better, but overheating should be avoided.

[0096] Average grain size ( ): Characterizes whether grain coarsening has occurred. Target value is The smaller and more uniform the better.

[0097] The experimental data were statistically analyzed by averaging the test results of 500 batches, resulting in the following statistical data table: The results show that the standard deviation data indicates that the standard deviation of solid solubility in the control group is [value missing]. This indicates that traditional processes are easily affected by batch fluctuations in raw materials, leading to inconsistent product quality; while the experimental group, by identifying effective activation energy online and compensating in real time, reduced the standard deviation to [missing information]. This indicates that the system can effectively eliminate disturbances caused by differences in raw materials, demonstrating extremely strong robustness. Meanwhile, the process capability index of the control group was only 0.28, implying significant incomplete solid solution treatment. The risk of non-conformity in production is inherent to the present invention, as the production process is uncontrollable; however, the process capability index of the system of this invention reaches 2.86 (typically...). (This is considered excellent), indicating that the system consistently operated within a high-quality range across 500 consecutive batches of production, validating the technology's extremely high reliability and repeatability in large-scale continuous industrial production. Furthermore, the control group, unable to detect overheating risks, experienced grain coarsening in some batches. The above; however, the experimental group utilized an acoustic emission melting mechanism to limit heat input in the early stages of grain boundary slip, thereby controlling the final average grain size within... And its distribution is extremely narrow ( This enables precision-level heat treatment.

[0098] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0099] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A monitoring and adaptive control system for the heat treatment process of magnesium-aluminum alloys, characterized in that, include: The multimodal sensing and preprocessing module is used to acquire the workpiece's dynamic parameters, instantaneous resistivity, and acoustic emission signals. The reference resistivity of the workpiece at room temperature is measured, and the power frequency noise is eliminated using a moving average filtering algorithm. The digital twin parameter identification module is used to calculate the current theoretical resistivity based on the activation energy estimated at the previous moment, and to calculate the effective activation energy estimated at the current moment by combining the deviation between the measured instantaneous resistivity and the current theoretical resistivity. The predictive control decision module is used to predict the evolution trend of phase transition fraction in a preset time domain based on the effective activation energy estimated at the current moment, calculate the deviation ratio between the predicted rate and the ideal rate, and generate temperature regulation commands in combination with the thermosensitive factor. The adaptive execution module is used to execute the acoustic emission fuse judgment mechanism. It determines the microscopic damage risk level based on the intensity of the acoustic emission signal, executes the corresponding strategy, and finally issues the final execution command.

2. The magnesium-aluminum alloy heat treatment process monitoring and adaptive control system according to claim 1, characterized in that, The specific steps of the moving average filtering algorithm include: establishing a window with a length of [missing information] in memory. A first-in-first-out queue is used, and a recursive algorithm is employed to update the accumulated sum of the queue data. And based on the updated cumulative sum Calculate the smoothed resistivity at the current moment. As valid input data.

3. The magnesium-aluminum alloy heat treatment process monitoring and adaptive control system according to claim 1, characterized in that, The current theoretical resistivity is calculated based on the JMAK equation, and its expression is: ； In the formula, For the current moment Theoretical resistivity, The reference resistivity of the workpiece at room temperature. is the temperature coefficient of resistance of magnesium alloy. The solid solution scattering coefficient is... Based on room temperature, This represents the total phase transition fraction at the current moment.

4. The magnesium-aluminum alloy heat treatment process monitoring and adaptive control system according to claim 3, characterized in that, The effective activation energy estimated at the current moment is calculated using the dynamic activation energy adaptive correction formula, the expression of which is: ； In the formula, for Effective activation energy estimated at any time The initial theoretical activation energy, This is the gain coefficient. Noise tolerance factor, Adjust the integral weight for energy. The measured instantaneous resistivity Compared with the current theoretical resistivity The deviation.

5. The magnesium-aluminum alloy heat treatment process monitoring and adaptive control system according to claim 4, characterized in that, The calculation formula for the temperature adjustment command is as follows: ； In the formula, For the current moment Temperature adjustment instructions For the target phase transition rate, The actual phase transition rate is based on real-time resistivity feedback. This is the gain coefficient. The gas constant is... The current temperature of the workpiece. To prevent the minimum value where the denominator is zero.

6. The magnesium-aluminum alloy heat treatment process monitoring and adaptive control system according to claim 1, characterized in that, The determination of the risk level of microscopic damage includes: setting a first-level threshold. and secondary threshold ; If acoustic emission signal The system determines that the current microstructure is safe and directly executes the temperature regulation command. ;like The system determines that the workpiece faces the risk of overheating and cracking, and directly blocks the temperature adjustment command. The rapid cooling protection program is forcibly triggered, and an overburned fuse alarm is sent to the host computer; if The system determines that the workpiece is in the subcritical stress accumulation zone and executes an adaptive gain damping strategy based on risk gradient.

7. The magnesium-aluminum alloy heat treatment process monitoring and adaptive control system according to claim 6, characterized in that, The adaptive gain damping strategy based on risk gradient specifically includes: constructing a penalty function and calculating the discount factor of the heating command. Using the discount factor Temperature adjustment command Make corrections to obtain the final execution instructions. :like This is determined to be a heating request, and will be executed. ;like If the request is determined to be a cooling or insulation request, it will be executed. .

8. The magnesium-aluminum alloy heat treatment process monitoring and adaptive control system according to claim 7, characterized in that, The penalty function is based on the current acoustic emission signal. At the first threshold With secondary threshold The relative positions within the constructed risk interval are defined by the following expression: 。 9. The magnesium-aluminum alloy heat treatment process monitoring and adaptive control system according to claim 6, characterized in that, The first-level threshold The determination method includes: continuously collecting data within the initial metallurgical quiescent zone during the heating phase. A set of background noise samples is formed by several acoustic emission root mean square voltage samples. The mean of the background noise sample set is calculated. and standard deviation According to the preset heat treatment time and control cycle frequency Calculate the false alarm probability ; Calculate the statistical significance factor corresponding to the false alarm probability using the inverse function of the standard normal distribution. The first-level threshold is calculated. The secondary threshold The first-level threshold Three times that.

10. An intelligent precision magnesium-aluminum alloy heat treatment equipment, characterized in that, Specifically, it includes: Vacuum heat treatment furnace body, multi-channel resistivity acquisition instrument, wideband acoustic emission detector, PID power regulator, main control computer, and the magnesium-aluminum alloy heat treatment process monitoring and adaptive control system as described in any one of claims 1-9; The main control computer is electrically connected to the multi-channel resistivity acquisition instrument, the broadband acoustic emission detector, and the PID power regulator, respectively. The multi-channel resistivity acquisition instrument is connected to the workpiece using a four-probe method, the broadband acoustic emission detector is coupled to the surface of the workpiece, and the PID power regulator is used to control the furnace heating power according to the temperature adjustment command issued by the main control computer.

Images