A composite magnetic core material and its preparation method

Abstract

This invention relates to the field of material preparation process control technology, and particularly to a composite magnetic core material and its preparation method. The method includes collecting the target performance indicators and basic material properties of the composite magnetic core to be processed, and inversely solving the process trajectory matrix using a predictive model. During annealing, the system collects the actual temperature and input electrical power in real time, calculates the theoretical heating temperature rise, and extracts the thermal residual time series. If the thermal residual time derivative exceeds a preset threshold, a decoupling control command is generated. A power attenuation command is used to reduce the input electrical energy, and a pulse modulation command is used to switch the protective atmosphere to a pulsed airflow to dissipate latent heat. Finally, the holding time is dynamically compensated based on the accumulated heat release, and annealing is terminated when the temperature field stabilizes. This invention effectively solves the problem of coordinated control of annealing parameters, improving the consistency of magnetic core performance and yield.

Description

Technical Field

[0001] This invention relates to the field of material preparation process control technology, and in particular to a composite magnetic core material and its preparation method. Background Technology

[0002] Composite magnetic core materials are functional materials created by combining two or more soft magnetic alloys with complementary properties, such as high-permeability nanocrystalline alloys and high-temperature-stability permalloy, through specific processes to meet specific electromagnetic performance requirements. The core of composite magnetic core material preparation lies in annealing, which involves heating the material to a specific temperature and holding it for a certain time in a protective atmosphere such as hydrogen or nitrogen. This aims to eliminate internal stress and optimize the microstructure, thereby enabling the material to acquire the desired magnetic properties. The protective atmosphere prevents oxidation or compositional changes at high temperatures. Ideally, the annealing process can endow the final material with the advantages of both constituent materials. However, the preparation of composite magnetic core materials involves the coordinated control of several key process parameters.

[0003] Existing methods for preparing composite magnetic core materials suffer from the following technical challenges: In the fabrication of voltage transient disturbance suppression devices based on magnetic saturation, the annealing process of the composite magnetic core is a complex process involving strong coupling of multiple variables. Key parameters such as annealing temperature, holding time, and the type and flow rate of the protective atmosphere do not act independently but rather have profound interactive effects. For example, while a higher annealing temperature is beneficial for eliminating internal stress in the material, improper control of the holding time can easily lead to abnormal grain growth, damaging the nanocrystalline structure and causing a decrease in magnetic permeability. Simultaneously, the combination of temperature and time directly determines the diffusion uniformity of alloying elements, thus affecting the temperature stability of the material. Furthermore, if the purity and flow field distribution of the protective atmosphere do not match the set temperature profile, it may lead to oxidation or carburization on the magnetic core surface, introducing additional impurities and stress, thus increasing hysteresis loss. Due to the lack of precise modeling and coordinated control methods for the above-mentioned multi-parameter coupling relationship, relying solely on fixed process windows or experience adjustments in actual production makes it difficult to ensure that each batch of magnetic cores simultaneously achieves the optimal matching state of high permeability, low loss, and wide temperature stability. This ultimately results in large batch-to-batch dispersion of the overall performance of the magnetic cores, making it difficult to improve the yield. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a composite magnetic core material and its preparation method. This invention solves the technical problem of unstable overall magnetic core performance and low yield caused by the difficulty in accurately and synergistically controlling multiple process parameters such as annealing temperature, time, and protective atmosphere during the preparation of nanocrystalline and permalloy composite magnetic cores.

[0005] To solve the above-mentioned technical problems, the specific contents of the present invention are as follows: This invention provides a composite magnetic core material and its preparation method, comprising: Step 1: Collect the target performance index data and basic material property data of the composite magnetic core to be processed, extract the key feature variables from the target performance index data and the basic material property data, and structurally splice the key feature variables to generate an initial feature vector characterizing the initial state of the composite magnetic core to be processed. Step 2: Input the target performance index data and the material basic property data into the prediction model for inverse solving, and output the process trajectory matrix. The process trajectory matrix includes annealing temperature setting data, protective atmosphere flow rate setting data, and preset holding time data. Step 3: Input the annealing heat according to the annealing temperature setting data, and input the protective atmosphere according to the protective atmosphere flow setting data to perform annealing on the composite magnetic core to be processed, and collect the actual temperature data and actual input power data consumed during the annealing process. Step 4: Calculate the theoretical heating temperature rise data by combining the actual input electrical power data and the preset heat capacity mapping rule; perform a difference operation between the actual temperature data and the theoretical heating temperature rise data to obtain the thermal residual time series. Step 5: Perform time derivative on the thermal residual time series to obtain the exothermic derivative. When the exothermic derivative is greater than a preset derivative threshold, generate a decoupled control command including a power attenuation command and a pulse modulation command. Use the power attenuation command to attenuate the input power corresponding to the actual input power data, and use the pulse modulation command to modulate the constant airflow corresponding to the protective atmosphere flow rate setting data into a pulsed airflow.

[0006] Furthermore, the composite magnetic core material preparation method of the present invention inputs the target performance index data and the basic material property data into a prediction model for inverse solving, and outputs a process trajectory matrix, including: Extract the set crystallization volume fraction and set grain size values ​​from the target performance index data, and screen out the crystallization activation energy data of the nanocrystalline phase and the grain growth activation energy data of the permalloy phase from the material basic physical property data. After using the maximum-minimum normalization algorithm to uniformly and linearly map the crystallization activation energy data, the grain growth activation energy data, and the target performance index data to a dimensionless numerical range, the data are input into the prediction model including the Avramie phase transformation kinetic equation for forward mapping operation. This establishes a first numerical correlation mapping between the annealing temperature setting data and the volume fraction of the nanocrystalline phase, and a second numerical correlation mapping between the annealing temperature setting data and the grain size of the permalloy phase. The optimization algorithm is used to update the algorithm through multiple rounds of iteration. The iteration is terminated when the number of iterations reaches a preset maximum iteration threshold or the continuous change of the global optimal fitness value is less than a preset convergence accuracy. Then, the intersection data points that simultaneously satisfy the set crystal volume fraction and the set grain size value are extracted from the first numerical correlation map and the second numerical correlation map. The parameters corresponding to the intersection data points are summarized into the process trajectory matrix.

[0007] Furthermore, the composite magnetic core material preparation method of the present invention, wherein the annealing heat is input according to the annealing temperature setting data, and the protective atmosphere is input according to the protective atmosphere flow rate setting data, annealing is performed on the composite magnetic core to be processed, and the actual temperature data and actual input power consumed during the annealing process are collected, including: Obtain the geometric dimensions of the annealing space in which the annealing process is performed, as well as the kinetic viscosity data of the protective atmosphere; The spatial thermal inertia delay time value is derived by performing correlation calculations on the geometric dimension data and the kinetic viscosity data. The preset sampling frequency is set according to the reciprocal of the spatial thermal inertia delay time value; The annealing heat is input using the annealing temperature setting data as the heat input condition, and the protective atmosphere is delivered to the annealing processing space in a constant state using the protective atmosphere flow rate setting data as the gas transmission condition. The actual temperature data inside the annealing processing space is continuously collected according to the preset sampling frequency, and the actual input electrical power data of the annealing heat consumption is recorded simultaneously.

[0008] Furthermore, the composite magnetic core material preparation method of the present invention involves calculating theoretical heating temperature rise data by using the actual input electrical power data and a preset heat capacity mapping rule, and performing a difference operation between the actual temperature data and the theoretical heating temperature rise data to obtain a thermal residual time series, including: Extract the heat capacity values ​​of the nanocrystalline phase, the heat capacity values ​​of the permalloy phase, the mass data of the nanocrystalline phase, the mass data of the permalloy phase, and the background heat capacity values ​​of the heating hardware used to input the annealing heat of the composite magnetic core to be processed. The total input heat value is obtained by performing time integration on the actual input electrical power data in the same time dimension. The heat capacity of the nanocrystalline phase is obtained by multiplying the heat capacity value of the nanocrystalline phase by the mass data of the nanocrystalline phase. The heat capacity value of the permalloy phase is obtained by multiplying the heat capacity value of the permalloy phase by the mass data of the permalloy phase. The total heat capacity value of the system is obtained by summing the heat capacity values ​​of the nanocrystalline phase, the permalloy phase, and the background heat capacity value. Dividing the total input heat value by the total heat capacity of the system, the theoretical heating temperature rise data under the condition of no material phase change heat release is calculated; Under the same time dimension, the temperature difference between the actual temperature data and the theoretical heating temperature rise data is calculated, and the temperature difference values ​​obtained at consecutive time points are arranged to generate the thermal residual time series.

[0009] Furthermore, in the composite magnetic core material preparation method of the present invention, the exothermic derivative is obtained by performing time derivative on the thermal residual time series. When the exothermic derivative is greater than a preset derivative threshold, a decoupling control command including a power attenuation command and a pulse modulation command is generated, including: Obtain the critical temperature gradient value for the secondary crystallization of the permalloy phase; The critical temperature gradient value is calculated by ratioing the latent heat of phase transformation per unit mass of the nanocrystalline phase to obtain the maximum safe heat release rate value that ensures no secondary crystallization occurs. The maximum safe heat release rate value is set as the preset derivative threshold value. The exothermic derivative is obtained by performing time derivative operation on the thermal residual time series; When the exothermic derivative is greater than the preset derivative threshold, it is determined that the exothermic rate of the nanocrystal primary crystal exceeds the system's thermal conductivity, and the decoupling control command is generated to simultaneously limit the input electrical energy and increase convective heat transfer.

[0010] Furthermore, the composite magnetic core material preparation method of the present invention utilizes the power attenuation command to attenuate the input electrical energy corresponding to the actual input electrical power data, including: Extract the maximum permissible derating rate value of the power supply hardware; Calculate the absolute value of the difference between the exothermic derivative and the preset derivative threshold, and calculate the exothermic excess ratio by comparing the absolute value of the difference with the preset derivative threshold. The target decay slope value is obtained by multiplying the exothermic excess ratio value by the maximum allowable derating rate value. The input electrical energy is dated and limited by the target attenuation slope value to restrict the heat source input.

[0011] Furthermore, the method for preparing composite magnetic core material according to the present invention utilizes the pulse modulation command to modulate the constant airflow corresponding to the protective atmosphere flow rate setting data into a pulsed airflow, comprising: Extract the surface area value of the composite magnetic core to be processed and the volume value of the annealing processing space; The target pulse frequency value is derived by multiplying the surface area value and the heat release excess ratio value, and the target duty cycle value to ensure the turbulent heat transfer depth is derived based on the volume value. Adjust the opening and closing frequency and opening duration of the gas delivery channel for delivering the protective atmosphere according to the target pulse frequency value and the target duty cycle value. The constant airflow is divided into pulsed airflows with peak velocity characteristics and trough velocity characteristics, and the latent heat accumulated on the surface of the composite magnetic core to be treated is dissipated by the dynamic pressure difference of the pulsed airflows.

[0012] Furthermore, the method for preparing composite magnetic core material according to the present invention, after modulating the constant airflow corresponding to the protective atmosphere flow rate setting data into a pulsed airflow using the pulse modulation command, includes: The zero-crossing time data of the exothermic derivative from positive to negative is extracted as the end point of the exothermic interval, and the time period from the start time of the annealing process to the end point of the exothermic interval is defined as the effective exothermic interval. Perform time integration on the thermal residual time series within the effective heat release interval to obtain the cumulative heat release value; Extract the total theoretical latent heat of complete crystallization, divide the cumulative heat release value by the total theoretical latent heat of complete crystallization value, and obtain the phase transformation completion value characterizing the crystallization ratio.

[0013] Furthermore, the method for preparing the composite magnetic core material according to the present invention, after obtaining the phase transition completion value, includes: Extract the real-time crystallization rate value corresponding to the current time node from the actual temperature data; Subtract the phase transformation completion value from 1 to obtain the non-crystallization ratio value; Multiply the uncrystallized ratio by the theoretical total mass of the composite magnetic core to be processed to obtain the remaining mass to be crystallized. Divide the remaining mass of crystallizers by the real-time crystallization rate to calculate the delay time that needs to be compensated. The delay duration value is added to the preset insulation time data to generate the target insulation time data.

[0014] Furthermore, the method for preparing the composite magnetic core material according to the present invention, after generating the target heat preservation time data, includes: The composite magnetic core to be processed is continuously annealed based on the pulsed airflow, the input electrical energy after derating and limiting, and the target heat preservation time data. Extract the temperature fluctuation derivative at the current time point; When the annealing process reaches the target holding time and the temperature fluctuation derivative is less than the steady-state fluctuation threshold during the continuous observation period, the input of the annealing heat is terminated, and the preparation of the composite magnetic core material is completed.

[0015] Beneficial effects of this invention: This invention provides a composite magnetic core material and its preparation method. By constructing a phase space prediction model with dynamic compensation for latent heat of crystallization, it achieves precise mapping of the heat treatment window for both nanocrystalline and permalloy phase biphase materials, fundamentally solving the problem of initial setting deviation caused by the coupling of multiple process parameters. During the annealing stage, the thermodynamic observer established by the system can accurately extract the thermal residual time series reflecting the microscopic phase transformation behavior by comparing actual temperature data with theoretical heating temperature rise data in real time, thereby overcoming the monitoring lag of traditional temperature control methods for the exothermic phenomenon of primary crystallization of nanocrystals. When the exothermic derivative exceeds the preset derivative threshold, the decoupling control command is triggered. Through the power attenuation command, the external heat energy supply is actively cut off, and the high-frequency pulse airflow generated by the pulse modulation command enhances convective heat transfer, rapidly dissipating the transient accumulated latent heat on the surface of the composite magnetic core to be treated, effectively avoiding the risk of abnormal grain growth and secondary crystallization of the permalloy phase caused by local overheating. Furthermore, based on the phase transition completion value and target holding time data dynamically generated from the cumulative heat release value, a deep closure between the holding time and the material's solidification process is achieved, eliminating the magnetic property dispersion caused by material differences or environmental fluctuations. This enables the final composite magnetic core to exhibit high consistency, high permeability, and excellent wide-temperature stability in the application scenario of voltage transient disturbance suppression device based on the magnetic saturation principle, significantly improving the overall yield of product preparation. Attached Figure Description

[0016] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the accompanying drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a composite magnetic core material and its preparation method according to the present invention. Detailed Implementation

[0018] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. 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. The present invention provided by various embodiments will be described in detail below with reference to the accompanying drawings. To better understand the purpose of the present invention, the present invention will be described in further detail below.

[0019] Please see Figure 1 The present invention provides a composite magnetic core material and its preparation method, comprising: Step 1: Collect the target performance index data and basic material property data of the composite magnetic core to be processed, extract the key feature variables from the target performance index data and the basic material property data, and structurally splice the key feature variables to generate an initial feature vector characterizing the initial state of the composite magnetic core to be processed. Step 2: Input the target performance index data and the material basic property data into the prediction model for inverse solving, and output the process trajectory matrix. The process trajectory matrix includes annealing temperature setting data, protective atmosphere flow rate setting data, and preset holding time data. Step 3: Input the annealing heat according to the annealing temperature setting data, and input the protective atmosphere according to the protective atmosphere flow setting data to perform annealing on the composite magnetic core to be processed, and collect the actual temperature data and actual input power data consumed during the annealing process. Step 4: Calculate the theoretical heating temperature rise data by combining the actual input electrical power data and the preset heat capacity mapping rule; perform a difference operation between the actual temperature data and the theoretical heating temperature rise data to obtain the thermal residual time series. Step 5: Perform time derivative on the thermal residual time series to obtain the exothermic derivative. When the exothermic derivative is greater than a preset derivative threshold, generate a decoupled control command including a power attenuation command and a pulse modulation command. Use the power attenuation command to attenuate the input power corresponding to the actual input power data, and use the pulse modulation command to modulate the constant airflow corresponding to the protective atmosphere flow rate setting data into a pulsed airflow.

[0020] In the fabrication of voltage transient disturbance suppression devices based on magnetic saturation, the composite magnetic core annealing process faces the challenge of controlling multiple strongly coupled variables. The system collects target performance index data and basic material property data of the composite magnetic core to be processed. The system extracts key feature variables from these data and structurally concatenates them to generate a uniformly formatted initial feature vector. Specifically, the target performance index data represents permeability (characterizing magnetization) and hysteresis loss (characterizing energy loss). The basic material property data includes the crystallization activation energy data for the nanocrystalline phase, the grain growth activation energy data for the permalloy phase, and the initial mass ratio data for the two-phase material. These data form the foundational information for establishing the phase transformation kinetics. The system converts the collected data into a digitally quantized feature vector that can be recognized by the algorithm.

[0021] The predictive model receives target performance index data and basic material property data and then performs inverse solving. Embedded within the predictive model is Avramie phase transition kinetic logic, which establishes a nonlinear mapping relationship between input parameters and final magnetic properties through forward mapping operations. The system uses an optimization algorithm to find intersection data points in the nonlinear mapping relationship that satisfy the set crystallization volume fraction. The output process trajectory matrix is ​​not a single numerical value, but a multi-dimensional process control array that varies over time. Specifically, the process trajectory matrix includes annealing temperature setting data to guide the heating element's operation, protective atmosphere flow rate setting data to adjust the gas delivery valve opening, and preset holding time data to limit the duration of the isothermal stage. The annealing temperature setting data specifies the target heat value to be achieved in different heating stages. The protective atmosphere flow rate setting data determines the baseline value of the gas flow rate injected into the furnace cavity to prevent material oxidation.

[0022] After receiving the annealing temperature setting data, the heating hardware inputs annealing heat into the annealing processing space. A synchronously operating gas delivery channel inputs a protective atmosphere into the annealing processing space according to the protective atmosphere flow rate setting data. The composite magnetic core to be processed is enveloped in a constant atmosphere within the annealing processing space and continuously absorbs annealing heat. During the annealing process, temperature-measuring thermocouples continuously collect the actual temperature data inside the annealing processing space according to a preset sampling frequency. The power transmitter collects the actual input electrical power data consumed by the heating hardware in real time. The actual temperature data reflects the macroscopic thermal field state, including the heat released by the composite magnetic core to be processed and the heat from external heating. The actual input electrical power data characterizes the pure external heat source energy injected into the annealing processing space by the external control system.

[0023] The data processing engine inputs the actual electrical power data into a preset heat capacity mapping rule for integral calculation. This preset rule integrates the specific heat capacity of the composite magnetic core and the inherent heat capacity of the heating hardware. Based on the actual electrical power data, the engine calculates theoretical heating temperature rise data under conditions of no material phase change heat release. This theoretical temperature rise data represents the theoretical temperature change caused purely by external electrical energy conversion. The system aligns the actual temperature data and the theoretical heating temperature rise data onto the same time axis and performs differential calculations. The system arranges the values ​​obtained from the differential calculations in chronological order to generate a thermal residual time series. This thermal residual time series effectively eliminates interference from external heating, directly mapping the dynamic evolution trajectory of the latent heat released during the nanocrystalline phase change process.

[0024] The system performs time derivative calculations on the thermal residual time series to obtain the exothermic derivative, which characterizes the instantaneous exothermic rate. The exothermic derivative reflects the intensity of microcrystalline formation in the material. The control system internally sets a preset derivative threshold to characterize the safe phase transition boundary. When the exothermic derivative exceeds the preset threshold, the system determines that the primary crystallization exothermic rate of the nanocrystals exceeds the upper limit of physical thermal conductivity, reaching a critical point where secondary crystallization is imminent. The control system then activates an adaptive intervention mechanism and generates decoupled control commands, including power attenuation and pulse modulation commands. The system uses the power attenuation command to derating and limiting the input electrical energy of the heating hardware, directly attenuating the input electrical energy corresponding to the actual input power data to cut off the external heat source supply. Simultaneously, the control system uses the pulse modulation command to adjust the opening and closing frequency of the gas delivery valve. The pulse modulation command modulates the stable and constant airflow corresponding to the protective atmosphere flow rate setting data into a high-frequency fluctuating pulsed airflow. The pulsed airflow creates a strong convective heat transfer effect and rapidly dissipates the microscopic transient latent heat accumulated on the surface of the composite magnetic core to be treated. The aforementioned dynamic intervention process breaks through the limitations of conventional open-loop control and achieves deep cross-coupling of thermodynamic and fluid dynamic fields.

[0025] In the process of fabricating a voltage transient disturbance suppression device, the prediction model needs to handle complex input parameters across multiple dimensions. The system extracts the set crystal volume fraction representing the target crystallization degree and the set grain size representing the desired grain size from the target performance index data. Simultaneously, it filters the crystallization activation energy data of the nanocrystalline phase (characterizing the ease of crystallization) and the grain growth activation energy data of the permalloy phase (characterizing the grain growth trend) from the basic material property data. The system synchronously inputs the crystallization activation energy data, grain growth activation energy data, and target performance index data into the prediction model, which includes the Avramyi phase transformation kinetic equation, for forward mapping calculation. The forward mapping calculation establishes a first numerical correlation mapping between the annealing temperature setting data and the nanocrystalline phase crystal volume fraction in a multidimensional space, and simultaneously establishes a second numerical correlation mapping between the annealing temperature setting data and the permalloy phase grain size value. The system calls the optimization algorithm to extract the intersection data points that simultaneously satisfy the set crystallization volume fraction and set grain size values ​​in the first numerical correlation map and the second numerical correlation map. The optimization algorithm merges and recombines the process parameters corresponding to the intersection data points and summarizes them into the output process trajectory matrix.

[0026] The control of annealing hardware inherently suffers from physical delay. To overcome measurement errors caused by this delay, the system pre-acquires the geometric dimensions of the annealing space and the kinetic viscosity of the protective atmosphere. The control system inputs these data into a preset thermofluid equation for correlation calculations, deriving the spatial thermal inertia delay time, which reflects the temperature conduction lag. The system sets the preset sampling frequency of the temperature sensing hardware based on the reciprocal of this spatial thermal inertia delay time. Subsequently, the heating hardware inputs annealing heat using the set annealing temperature, while the gas supply hardware delivers a constant-state protective atmosphere to the annealing space using the set protective atmosphere flow rate. The temperature sensor continuously collects the actual temperature data inside the annealing space according to the preset sampling frequency, and the power recording hardware synchronously records the actual input electrical power consumed by the annealing heat.

[0027] Extracting heat changes purely caused by material phase transitions requires eliminating interference from external electrical heating. The system extracts the specific heat capacity values ​​of the nanocrystalline phase, the specific heat capacity value of the permalloy phase, the mass data of the nanocrystalline phase, the mass data of the permalloy phase, and the background heat capacity value of the heating hardware used to input annealing heat from the composite magnetic core to be processed. The system performs time integration on the actual input electrical power data over the same time dimension to obtain the total input heat value provided by the external hardware. The data processing unit multiplies the specific heat capacity value of the nanocrystalline phase by the mass data of the nanocrystalline phase to obtain the heat capacity value of the nanocrystalline phase, and multiplies the specific heat capacity value of the permalloy phase by the mass data of the permalloy phase to obtain the heat capacity value of the permalloy phase. The system sums the heat capacity values ​​of the nanocrystalline phase, the permalloy phase, and the background heat capacity value to obtain the total system heat capacity value representing the global heat absorption capacity. The system divides the total input heat value by the total system heat capacity value to calculate the theoretical heating temperature rise data under conditions without material phase transition heat release. Under the same time dimension, the system calculates the temperature difference by subtracting the actual temperature data from the theoretical heating temperature rise data. The system then arranges multiple temperature difference values ​​obtained at consecutive time points in chronological order to generate a thermal residual time series that reflects the dynamic process of latent heat release inside the material.

[0028] While the thermal residual time series can reflect the exothermic trend, it cannot directly determine the critical point of crystallization danger. The control system pre-acquires the critical temperature gradient value representing the critical state of secondary crystallization of the permalloy phase. The system calculates the maximum safe exothermic rate value to ensure that the permalloy phase does not undergo secondary crystallization by comparing the critical temperature gradient value with the latent heat of phase transformation per unit mass of the nanocrystalline phase. The control system then sets the maximum safe exothermic rate value as a preset derivative threshold. The algorithm module performs a derivative operation on the thermal residual time series with respect to the time variable to obtain the exothermic derivative reflecting the intensity of instantaneous exothermic activity. When the calculation unit compares the values ​​and finds that the exothermic derivative is greater than the preset derivative threshold, the system immediately determines that the current exothermic rate of the initial crystallization of the nanocrystalline phase completely exceeds the system's thermal conductivity. The control system then generates decoupling control commands to simultaneously limit the input electrical energy and increase convective heat transfer in response to the heat over-limit state.

[0029] Executing power intervention actions requires consideration of the safe operating range of the power supply hardware. The system retrieves the maximum permissible derating rate value of the power supply hardware. The algorithm engine performs subtraction calculations to obtain the absolute value of the difference between the heat release derivative and the preset derivative threshold. The arithmetic unit calculates the ratio between the absolute value of the difference and the preset derivative threshold to obtain a heat release over-limit ratio value reflecting the severity of heat release. The system multiplies the heat release over-limit ratio value with the maximum permissible derating rate value to derive the target attenuation slope value. The power supply control system uses the target attenuation slope value as the adjustment benchmark to perform derating and limiting actions on the input electrical energy to directly limit the input of external heat sources.

[0030] In addition to cutting off external heat sources, it is also necessary to utilize fluid dynamics to quickly remove the heat accumulated on the surface of the material. The system extracts the surface area value of the composite magnetic core to be processed and the volume value of the annealing processing space. The data processing unit performs a product operation on the surface area value and the heat release excess ratio value derived from the previous step to derive the target pulse frequency value guiding the gas valve action. The system combines the volume value of the annealing processing space with the fluid simulation algorithm to derive the target duty cycle value to ensure the turbulent heat transfer depth. The gas path control unit dynamically adjusts the opening and closing frequency of the gas delivery channel for delivering the protective atmosphere and the duration of each opening according to the target pulse frequency value and the target duty cycle value. The mechanical action of the gas delivery channel divides the original constant airflow into pulsed airflow with peak and trough velocity characteristics. The dynamic pressure difference formed by the pulsed airflow on the surface of the composite magnetic core directly dissipates the microscopic transient latent heat accumulated on the surface of the composite magnetic core to be processed.

[0031] After intervention with strong convective heat transfer, the crystallization progress of the material needs to be reassessed. The monitoring module continuously tracks the numerical change trajectory of the exothermic derivative and extracts the zero-crossing time data when the exothermic derivative changes from a positive to a negative value as the endpoint of the exothermic interval marking the end of the intense exothermic phase. The system defines the time period from the start time of annealing to the endpoint of the exothermic interval as the effective exothermic interval representing the core crystallization process. The system calls an integral algorithm to perform time integration on the thermal residual time series within the effective exothermic interval to obtain the actual cumulative heat release value. The control system extracts the total theoretical latent heat of complete crystallization for the corresponding material ratio from the material property database. The system uses division logic to divide the cumulative heat release value by the total theoretical latent heat of complete crystallization to calculate the phase transformation completion value representing the current microcrystalline completion ratio.

[0032] Dynamically adjusted holding time is a key control variable for ensuring stable crystal growth. The system calls the temperature measurement interface to extract the real-time crystallization rate value corresponding to the current time node from the actual temperature data. The calculation unit subtracts the phase transition completion value from the numerical value to calculate the proportion of uncrystallized material that has not yet undergone phase transition. The system retrieves the theoretically complete total mass of the composite magnetic core to be processed and multiplies it by the uncrystallized proportion to deduce the remaining mass to be crystallized. The algorithm module uses division to divide the remaining mass to be crystallized by the real-time crystallization rate value to calculate the delay time required to compensate for the uncrystallized portion. The time control system adds the delay time value to the preset holding time data to generate updated target holding time data.

[0033] The final stage of annealing requires a comprehensive evaluation of both the time dimension and the thermal field stability dimension. The process execution unit continuously performs annealing on the composite magnetic core under treatment, based on high-frequency switching pulsed airflow, derating and limiting input electrical energy, and the recalculated target holding time. A real-time monitoring network operates synchronously, continuously extracting the temperature fluctuation derivative at the current time point to reflect subtle disturbances in the thermal field. When the annealing duration recorded by the timer reaches the target holding time and the temperature fluctuation derivative remains within a relatively flat state below the steady-state fluctuation threshold during the continuous observation period, the control system issues a physical shutdown command and terminates the input of annealing heat.

[0034] In the fabrication of a voltage transient disturbance suppression device based on magnetic saturation, the annealing process of the composite magnetic core is in an extremely sensitive multivariate coupled state. The system collects target performance index data and basic material property data of the composite magnetic core to be processed, extracts key feature variables from these data, and structurally concatenates them to generate an initial feature vector characterizing the initial state of the composite magnetic core. Since different application scenarios have different standards for magnetization and energy loss, the target performance index data specifically represents preset permeability-related parameters and hysteresis loss data. Simultaneously, the system filters out crystallization activation energy data for the nanocrystalline phase and grain growth activation energy data for the permalloy phase from the basic material property data. The system inputs the extracted set crystallization volume fraction, set grain size values, along with the crystallization activation energy data and grain growth activation energy data, into a prediction model including Avramyan phase transformation kinetics logic for forward mapping calculation. The prediction model establishes a first numerical correlation mapping between the annealing temperature setting data and the volume fraction of the nanocrystalline phase in a complex multidimensional space, and simultaneously establishes a second numerical correlation mapping between the annealing temperature setting data and the grain size of the permalloy phase. The optimization algorithm extracts the intersection data points that simultaneously satisfy the set volume fraction and set grain size values ​​in the first and second numerical correlation mappings, and then summarizes the parameters corresponding to the intersection data points into a process trajectory matrix including the annealing temperature setting data, the protective atmosphere flow rate setting data, and the preset holding time data.

[0035] Industrial-grade annealing equipment commonly suffers from physical hysteresis in temperature conduction. The system acquires the geometric dimensions of the annealing space and the kinetic viscosity data of the protective atmosphere. The control unit performs correlation calculations on the geometric dimensions and kinetic viscosity data to derive the spatial thermal inertia delay time, reflecting the resistance to heat transfer in the physical space. The system sets the preset sampling frequency of the temperature sensing element based on the reciprocal of this spatial thermal inertia delay time. The heating module inputs annealing heat using the set annealing temperature as the heat input condition, and the gas supply hardware coordinates by delivering a constant protective atmosphere to the annealing space using the set protective atmosphere flow rate as the gas transmission condition. Thermocouples continuously collect the actual temperature data inside the annealing space according to the preset sampling frequency, and the power recording instrument synchronously records the actual input electrical power consumed by the input annealing heat.

[0036] Eliminating interference from external heating sources is a core prerequisite for identifying the internal phase transition state of materials. The system extracts the specific heat capacity values ​​of the nanocrystalline phase, the specific heat capacity value of the permalloy phase, the mass data of the nanocrystalline phase, the mass data of the permalloy phase, and the background heat capacity value of the heating hardware used to input annealing heat from the composite magnetic core to be processed. The system performs time integration on the actual input electrical power data over the same time dimension to obtain the total externally injected heat value. The data processing engine performs a weighted summation of the specific heat capacity values ​​of the nanocrystalline phase, the specific heat capacity value of the permalloy phase, and the background heat capacity value to obtain the total system heat capacity value, representing the overall heat absorption capacity of the system. The system divides the total input heat value by the total system heat capacity value to calculate the theoretical heating temperature rise data under conditions of no material phase transition heat release. Under the same time dimension, the system calculates the temperature difference between the actual temperature data and the theoretical heating temperature rise data. The data flow module arranges the temperature difference values ​​acquired at continuous time points to generate a thermal residual time series reflecting the trajectory of the material's latent heat of crystallization release.

[0037] The intense primary crystallization exothermic reaction of the nanocrystalline phase can easily induce abnormal grain growth in the permalloy phase. The control unit acquires the critical temperature gradient value for secondary crystallization of the permalloy phase and calculates the maximum safe exothermic rate to prevent secondary crystallization by comparing the critical temperature gradient value with the latent heat of phase transformation per unit mass of the nanocrystalline phase. The system sets the maximum safe exothermic rate value as a preset derivative threshold. The algorithm module performs time derivative calculation on the thermal residual time series to obtain the exothermic derivative, which characterizes the intensity of instantaneous exothermic reaction. When the exothermic derivative value exceeds the preset derivative threshold, the control center determines that the primary crystallization exothermic rate of the nanocrystalline phase exceeds the system's thermal conductivity. The control center immediately generates a decoupling control command to simultaneously limit the input electrical energy and increase convective heat transfer.

[0038] The decoupling control commands include power attenuation commands and pulse modulation commands to guide hardware actions. The system extracts the maximum permissible derating rate value of the power supply hardware. The algorithm module calculates the absolute value of the difference between the exothermic derivative and the preset derivative threshold, and calculates the exothermic over-limit ratio by comparing the absolute value of the difference with the preset derivative threshold. The system multiplies the exothermic over-limit ratio value with the maximum permissible derating rate value to obtain the target attenuation slope value. The power supply control system drates and limits the input electrical energy based on the target attenuation slope value to directly limit the input of external heat sources. The system simultaneously extracts the surface area value of the composite magnetic core to be processed and the volume value of the annealing processing space. The data node multiplies the surface area value with the exothermic over-limit ratio value to derive the target pulse frequency value, and derives the target duty cycle value to ensure the turbulent heat transfer depth based on the volume value. The gas valve adjusts the opening and closing frequency and opening duration of the gas delivery channel for delivering the protective atmosphere according to the target pulse frequency value and the target duty cycle value. The gas delivery channel divides the original constant airflow into pulsed airflows with peak and trough velocity characteristics. The high-frequency dynamic pressure difference enhances convective heat transfer to dissipate the microscopic latent heat accumulated on the surface of the composite magnetic core to be treated.

[0039] Dynamic flow field intervention alters the original expected crystallization progress. The system extracts the zero-crossing time data where the exothermic derivative changes from positive to negative as the end point of the exothermic interval, and defines the time period from the start of annealing to the end of the exothermic interval as the effective exothermic interval. The system performs time integration on the thermal residual time series within the effective exothermic interval to obtain the cumulative heat release value. The computational core extracts the total theoretical latent heat of complete crystallization and divides the cumulative heat release value by the total theoretical latent heat of complete crystallization to obtain the phase transition completion value, which characterizes the crystallization ratio. The control system extracts the real-time crystallization rate value corresponding to the current time node from the actual temperature data and subtracts the phase transition completion value from the number to obtain the uncrystallized ratio value. The system multiplies the uncrystallized ratio value by the theoretical total mass of the composite magnetic core to be processed to obtain the remaining mass to be crystallized value. The algorithm node divides the remaining mass to be crystallized value by the real-time crystallization rate value to calculate the delay time value that needs to be compensated, and the timing module adds the delay time value to the preset holding time data to generate the updated target holding time data.

[0040] The final stage of the annealing process requires comprehensive consideration of both time factors and thermodynamic stability. The process control bus continuously executes the annealing process on the composite magnetic core under treatment based on high-frequency fluctuating pulsed airflow, derating and limiting input electrical energy, and target holding time data. The temperature measurement network continuously extracts the temperature fluctuation derivative at the current time point. When the annealing time reaches the target holding time and the temperature fluctuation derivative is less than the steady-state fluctuation threshold within the continuous observation period, the control system determines that the material microstructure has reached thermodynamic equilibrium and physically shuts off the heating circuit to terminate the input of annealing heat.

[0041] In the engineering practice of fabricating voltage transient disturbance suppression devices, target performance index data serves as the logical starting point for process design. This data encompasses a set of key parameters characterizing the electromagnetic performance of the composite magnetic core, specifically including the target permeability, maximum allowable hysteresis loss, and set saturation magnetic induction intensity. These data directly define the technical specifications of the final product in magnetically saturated application scenarios. To ensure the prediction model obtains complete material background information, the system simultaneously collects basic material property data, including not only the mass ratio between the nanocrystalline phase and the permalloy phase, but also in-depth microscopic kinetic parameters, such as the crystallization activation energy data characterizing the difficulty of atomic diffusion and the activation energy value determining the grain growth rate.

[0042] The process trajectory matrix is ​​a multi-dimensional control instruction set output by the predictive model after inverse solving, presented as a dynamic sequence of process parameters distributed along the time axis. The matrix highly integrates annealing temperature setpoint data, protective atmosphere flow rate setpoint data, and preset holding time data. The annealing temperature setpoint data specifies the target heat thresholds for each stage of heating, isothermal control, and cooling, while the protective atmosphere flow rate setpoint data specifies the real-time charging rate of nitrogen or hydrogen in the furnace cavity. These process parameters are not discrete but are strongly correlated along the time dimension through a matrix structure.

[0043] The actual temperature data is a real-time feedback of the thermal field state from temperature sensors arranged inside the annealing processing space, reflecting the macroscopic temperature rise trajectory of the composite magnetic core under treatment at a specific time point after being affected by both external heating and its own heat dissipation. Correspondingly, the actual input electrical power data records the electrical energy consumed by the heating power supply hardware at the same sampling moment. By converting the actual input electrical power data into heat output, the system can accurately dissipate the contribution of external energy input, thereby providing a clean data benchmark for subsequent detection of the phase transition behavior inside the material.

[0044] The theoretical heating temperature rise data represents the expected physical temperature rise that can be induced solely by external electrical energy injection under ideal conditions without phase change and heat release. The calculation process relies heavily on a preset heat capacity mapping rule, which pre-calibrates the background heat capacity values ​​of the hardware, including the heating furnace body, as well as the specific heat capacity variation of the composite magnetic core under test in different temperature ranges. By substituting the actual input electrical power data into the preset heat capacity mapping rule, the system derives the theoretical temperature rise curve in real time, providing a comparative benchmark for capturing the microscopic latent heat release signal hidden in the macroscopic thermal field.

[0045] The thermal residual time series is a chain of characteristic signals extracted by performing a difference operation on actual temperature data and theoretical heating temperature rise data. It is specifically used to characterize the evolution of transient latent heat released during the crystallization process of nanocrystals. The exothermic derivative is the rate of change obtained by performing a derivative operation on the thermal residual time series over time, which directly reflects the intensity of the microstructural transformation within the material. When the exothermic derivative exceeds a preset derivative threshold, the system determines that the exothermic intensity exceeds the environmental heat dissipation capacity, which can easily induce the risk of abnormal grain growth.

[0046] The decoupling control command is a closed-loop intervention strategy triggered after the system detects the risk of crystallization. It consists of interrelated power attenuation commands and pulse modulation commands. The power attenuation command directly acts on the power regulation hardware, reducing the external heat energy supply by forcibly lowering the percentage of input electrical energy. The pulse modulation command, on the other hand, adjusts the opening and closing duty cycle of the gas flow valve at high frequency, transforming the stable protective atmosphere flow into a pulsed airflow with strong turbulence characteristics. It utilizes the pressure fluctuation effect of fluid mechanics to forcibly dissipate the microscopic heat energy accumulated on the surface of the composite magnetic core, achieving synchronous decoupling control of the temperature field and the flow field.

[0047] The phase transition completion rate is a quantitative indicator for evaluating the progress of the microstructure evolution of the composite magnetic core. It is calculated by dividing the cumulative heat energy released within the effective exothermic zone by the theoretical total heat of complete crystallization. The cumulative heat release rate is the physical result of performing area integration on the thermal residual time series within the exothermic zone. The system uses the phase transition completion rate to derive the proportion of uncrystallized material and dynamically calculates the required compensation delay time based on the real-time crystallization rate. This data flow ensures that the holding time can be adaptively adjusted according to the actual phase transition process of the material, ultimately generating target holding time data that guarantees consistent magnetic properties.

[0048] In the reverse engineering phase, where target performance data and basic material properties are input into the prediction model for inverse solving, a rigorous mathematical derivation path is constructed within the prediction model. When performing the forward mapping operation, the prediction model uses the Avramie phase transition kinetic equation to establish a numerical correlation mapping. The specific formula for the first numerical correlation mapping is as follows:

[0049] In the operational formula of the first numerical correlation mapping, Represents the volume fraction of nanocrystalline phase crystals. This represents the preset nanocrystalline phase frequency factor parameter. Represents the base of the natural logarithm. Data on crystallization activation energy representing nanocrystalline phases. Represents the ideal gas constant. This represents the annealing temperature setting data. This represents the preset heat preservation time data. The exponential parameter represents the nanocrystalline phase. The specific calculation formula for the second numerical correlation mapping is:

[0050] In the operational formula of the second numerical correlation mapping, This represents the grain size of the permalloy phase. This represents the preset permalloy phase frequency factor parameter. Data on the activation energy of grain growth representing the permalloy phase. Represents the ideal gas constant. This represents the annealing temperature setting data. This represents the preset heat preservation time data. The growth index parameter represents the permalloy phase. The particle swarm optimization algorithm iteratively searches for annealing temperature and preset holding time data in a multi-dimensional space. The algorithm sets the set crystal volume fraction and set grain size values ​​as the target fitness function. By updating the particle velocity and position variables, the algorithm extracts the intersection data points that simultaneously satisfy the set crystal volume fraction and set grain size values ​​and summarizes and outputs the process trajectory matrix.

[0051] The annealing process requires eliminating the physical hysteresis error caused by temperature conduction. The control system derived the spatial thermal inertia delay time value using data. After acquiring the geometric dimensions of the annealing space and the kinetic viscosity data of the protective atmosphere, the data processing module performs correlation calculations using thermodynamic principles. The specific derivation formula is as follows:

[0052] In the derivation of the formula, Represents the spatial thermal inertial delay time. Geometric dimensional data representing the annealing processing space. The kinetic viscosity data representing the protective atmosphere This represents the preset equivalent thermal conductivity parameter. The control system sets the preset sampling frequency based on the reciprocal of the spatial thermal inertia delay time. The formula for setting the preset sampling frequency is:

[0053] In the formula for setting the preset sampling frequency, Represents the preset sampling frequency. This represents the spatial thermal inertia delay time. The process of generating theoretical heating temperature rise data by extrapolating from actual input electrical power data and preset heat capacity mapping rules relies on the law of conservation of energy for data conversion. The computing engine performs time integration on the actual input electrical power data over the same time dimension to obtain the total input heat value. The calculation formula is:

[0054] In the formula for calculating the total input heat value, This represents the total input calorie value. Represents the variable during integration time. The actual input power data at each moment. This represents the current time point. The data node calculates the total system heat capacity by weighted summing of the relative heat capacity values ​​of nanocrystalline materials, permalloy, and the baseline heat capacity. The weighted summation formula is:

[0055] In the weighted summation formula, This represents the total heat capacity of the system. This represents the relative heat capacity of nanocrystals. Quality data representing nanocrystalline phases, This represents the relative heat capacity of permalloy. Mass data representing the permalloy phase. This represents the baseline heat capacity of the heating hardware. The formula for calculating the theoretical heating temperature rise is: In the formula for calculating theoretical heating temperature rise data Represents the current time point The theoretical heating temperature rise data below, This represents the total input calorie value. This represents the total heat capacity of the system. The logic unit performs a difference operation between the actual temperature data and the theoretical heating temperature rise data to derive the formula for the thermal residual time series:

[0056] In the difference operation formula, Represents the current time point Discrete numerical values ​​of the thermal residual time series. Represents the current time point The actual temperature data below, Represents the current time point The theoretical heating temperature rise data is as follows.

[0057] The logic for generating decoupling control commands is based on dynamic calculus operations on the thermal residual time series. The division formula for calculating the maximum safe heat release rate using the computational network is:

[0058] In the division formula for the maximum safe heat release rate, This represents the maximum safe heat release rate. The critical temperature gradient representing the secondary crystallization of the permalloy phase. The latent heat of phase transition per unit mass represents the nanocrystalline phase. The control system sets the maximum safe heat release rate as a preset derivative threshold. The algorithm calculates the heat release derivative by differentiating the thermal residual time series with respect to the time variable. The derivative calculation formula is as follows:

[0059] In the formula for differentiation, Represents the current time point The exothermic derivative of the following, Represents the current time point Discrete numerical values ​​of the thermal residual time series. This represents the time differential variable. The input energy stage, which utilizes power attenuation commands to reduce the actual input power data, includes specific slope derivation formulas. The core logic for deriving the heat release over-limit ratio is as follows:

[0060] In the logical formula for the exothermic excess ratio, This represents the percentage of heat release exceeding the limit. Represents the current time point The exothermic derivative of the following, This represents the preset derivative threshold. The formula for calculating the target attenuation slope is: In the formula for calculating the target attenuation slope, This represents the target attenuation slope value. This represents the percentage of heat release exceeding the limit. This represents the maximum permissible derating rate of the power supply hardware. The control system uses pulse modulation commands to modulate the constant airflow corresponding to the protective atmosphere flow rate setting data into a pulsed airflow. The formula for deriving the target pulse frequency value is:

[0061] In the formula for calculating the target pulse frequency, Represents the target pulse frequency value. This represents the surface area value of the composite magnetic core to be processed. This represents the percentage of heat release exceeding the limit. This represents the preset frequency conversion coefficient. The formula for deriving the target duty cycle value is:

[0062] In the formula for the target duty cycle value, This represents the target duty cycle value. The numerical value representing the volume of the annealing processing space. This represents the preset volume conversion benchmark parameter. In the process of evaluating the progress of microcrystallization after convective heat transfer intervention, the data engine performs a time-integration calculation on the thermal residual time series within the effective heat release range to obtain the cumulative heat release value. The formula for performing the integration calculation is:

[0063] In the formula for performing integration, This represents the cumulative heat release value. This represents the start time of the annealing process. Data representing the zero-crossing time when the exothermic derivative changes from a positive to a negative value. Represents the variable during integration time. Discrete numerical values ​​of the time series of thermal residuals at each moment.

[0064] The formula for calculating the phase transition completion value, which characterizes the crystallization ratio, is:

[0065] In the formula for the numerical value of phase transition completion, This represents the degree of phase transition completion. This represents the cumulative heat release value. This represents the total latent heat of complete crystallization. The formula for calculating the proportion of immiscible material is:

[0066] In the formula for the percentage of uncrystallized material, This represents the percentage of uncrystallized material. This represents the degree of phase transformation completion. The formula for calculating the remaining mass to be crystallized is: In the formula for calculating the remaining mass to be crystallized, This represents the remaining mass of material to be crystallized. This represents the percentage of uncrystallized material. This represents the theoretically complete crystallized total mass of the composite magnetic core to be processed. The formula for calculating the required compensation delay is as follows:

[0067] In the formula for calculating the delay duration, This represents the numerical value indicating the delay period. This represents the remaining mass of material to be crystallized. This represents the real-time crystallization rate value corresponding to the current time point in the actual temperature data. The timing control system accumulates the delay duration value to the preset holding time data to generate the target holding time data. The accumulation formula is:

[0068] In the cumulative expression, This represents the target insulation time data. This represents the preset heat preservation time data. This represents the numerical value of the delay period.

[0069] During the annealing process, as the duration approaches the target holding time, the monitoring module needs to extract the temperature fluctuation derivative at the current time point. The formula for calculating the temperature fluctuation derivative is:

[0070] In the formula for the derivative of temperature fluctuation, The derivative of temperature fluctuation at the current time point. Represents the current time point The actual temperature data below, This represents the time-dependent differential variable. The computational core compares the absolute value of the temperature fluctuation derivative with the steady-state fluctuation threshold to determine if the temperature field is approaching a stable state and outputs a physical shutdown command.

[0071] The construction of the prediction model includes establishing a library of fundamental equations for Avramian phase transformation kinetics and encapsulating a particle swarm optimization algorithm module. Thermodynamic fundamental data of nanocrystalline and permalloy phases under different stoichiometric ratios were collected using differential scanning calorimetry (DSC). Regression analysis was used to fit the thermodynamic fundamental data to determine the preset nanocrystalline and permalloy phase frequency factor parameters in the Avramian phase transformation kinetics formula. A multidimensional data mapping network was constructed internally for the prediction model. The input nodes of the multidimensional data mapping network are bound to target performance index data and fundamental material property data. Before performing the forward mapping operation, the multidimensional data mapping network uses a max-min normalization algorithm to linearly map the target performance index data and fundamental material property data with different physical dimensions to a dimensionless numerical range of 0 to 1, eliminating the magnitude bias caused by cross-dimensional calculations. The intermediate nodes of the multidimensional data mapping network deploy the Avramian phase transformation kinetics formula for forward mapping operations, thereby establishing a nonlinear relationship between the annealing temperature setting data and the microstructure state. The output nodes of the multidimensional data mapping network are connected to the particle swarm optimization algorithm. The particle swarm optimization (PSO) algorithm sets the crystal volume fraction and grain size as the convergence targets of the fitness function. Through multiple iterations, the PSO algorithm updates the position and velocity parameters of the particles. It terminates the iteration when the number of iterations reaches a preset maximum iteration threshold or when the continuous change in the global optimal fitness value is less than a preset convergence accuracy, thus locking in the optimal combination of process parameters in the solution space. The data processing path of the prediction model is as follows: target performance index data and basic material property data are input into a multidimensional data mapping network. The multidimensional data mapping network performs Avramie phase transition kinetic mapping operations on the target performance index data and basic material property data to generate multiple sets of candidate process parameters. The PSO algorithm performs fitness evaluation and iterative optimization on these candidate process parameters. Finally, the PSO algorithm packages the candidate process parameters that meet the convergence conditions into a process trajectory matrix and sends it to the external control bus via the output interface.

[0072] To further clarify the safety boundaries and judgment criteria of the system control, this invention specifically defines the preset derivative threshold and the steady-state fluctuation threshold. The preset derivative threshold represents the maximum allowable latent heat release rate of phase transition within the composite magnetic core material. The system extracts the unit mass latent heat of phase transition of the nanocrystalline phase, which typically falls within the range of 45 joules to 60 joules per gram. The research team also determined the critical temperature gradient for secondary crystallization of the permalloy phase to be between 8 and 12 degrees Celsius per second. The system's computational core divides the critical temperature gradient value by the unit mass latent heat of phase transition to obtain the maximum safe heat release rate. The computational core sets this maximum safe heat release rate as the preset derivative threshold. In specific industrial applications, the specific value range of the preset derivative threshold set by the computational core is 0.15 to 0.25.

[0073] The steady-state fluctuation threshold is used to assess whether the temperature field has reached absolute thermodynamic equilibrium at the end of the annealing process. The steady-state fluctuation threshold characterizes the tolerance to minute fluctuations in the actual temperature data within the annealing space over time. The system, combining the system background noise of high-precision thermocouples and the weak thermal response characteristics at the end of material crystallization, precisely sets the steady-state fluctuation threshold to 0.5 degrees Celsius per minute. When the absolute value of the temperature fluctuation derivative is consistently less than 0.5 degrees Celsius per minute over a continuous 10-minute observation period, the system control logic determines that the microcrystalline structure has completely stabilized and outputs a physical shutdown command.

[0074] Embodiment 1 of this invention: In the fabrication process of a voltage transient disturbance suppression device based on the principle of magnetic saturation, addressing the issue of low overlap in the annealing windows of the nanocrystalline alloy and permalloy composite magnetic core, the system executes an initial fabrication process. Researchers pre-acquire target performance index data of the composite magnetic core to be processed using acquisition equipment, specifically including target permeability values ​​and maximum permissible hysteresis loss data, while also combining the nanocrystalline phase crystallization activation energy and biphase mass ratio from the material's basic physical properties data. After receiving the above parameters, the prediction model performs inverse solving using Avramie phase transformation kinetics logic, outputting a process trajectory matrix including annealing temperature setting data, protective atmosphere flow rate setting data, and preset holding time data. During the actual processing stage, temperature-measuring thermocouples record actual temperature data at a preset sampling frequency based on the reciprocal of the spatial thermal inertia delay time, while the power transmitter records the actual input electrical power data of the input annealing heat consumption. The data processing engine inputs the actual input electrical power data into a preset heat capacity mapping rule, calculates the theoretical heating temperature rise data, and performs differential calculations with the actual temperature data to obtain the thermal residual time series. When the calculated exothermic derivative exceeds the preset derivative threshold based on the critical temperature gradient for secondary crystallization of the permalloy phase, the adaptive controller immediately issues a decoupling control command. The power control unit uses a power attenuation command to derating and limiting the input electrical energy, while simultaneously using a pulse modulation command to switch the protective atmosphere to a high-frequency pulsed airflow, dissipating the accumulated latent heat through the pressure difference generated by the flow velocity at the peak and trough. After the intense exothermic phase ends, the system calculates the phase change completion value by integrating the accumulated heat release and dynamically calculates the compensated target holding time data.

[0075] Embodiment 2 of this invention: To address the fabrication requirements of power disturbance suppression components with high magnetic permeability, the system refined the generation logic of the process trajectory matrix. Researchers precisely selected the crystallization activation energy data of the nanocrystalline phase and the grain growth activation energy data of the permalloy phase from the basic material property data, and input them into the prediction model for forward mapping calculation. The model establishes a first numerical correlation mapping between the annealing temperature setting data and the crystal volume fraction of the nanocrystalline phase, and a second numerical correlation mapping between the annealing temperature setting data and the grain size value of the permalloy phase. The optimization algorithm searches for the intersection data points in the multi-dimensional correlation space that simultaneously satisfy the set crystal volume fraction and the set grain size value, thereby locking in the optimal parameter combination. Upon entering the annealing process, the system acquires the geometric dimensions of the annealing processing space and the kinetic viscosity data of the protective atmosphere, and derives the spatial thermal inertia delay time value representing the physical delay. The heating hardware outputs heat according to the annealing temperature setting data in the process trajectory matrix, while the gas transmission channel delivers a constant-state protective atmosphere to the annealing processing space. In the real-time monitoring cycle, the system extracts the specific heat capacity values ​​of nanocrystalline materials, permalloy, and the background heat capacity values ​​of the heating hardware, and then performs a weighted summation to construct the total system heat capacity value, which reflects the overall heat absorption capacity. This value is used to convert the time integral of the actual input electrical power data into accurate theoretical heating temperature rise data, thereby ensuring that the thermal residual time series can truly reflect the phase transition trajectory of the microstructure.

[0076] Embodiment 3 of this invention: In the mass production process of the voltage transient disturbance suppression device, the system solves the technical pain point of difficulty in accurately coordinating multiple process parameters through a closed-loop feedback mechanism. In the mid-to-late stages of annealing, the monitoring module extracts the zero-crossing time data of the exothermic derivative changing from a positive to a negative value, and defines the effective exothermic interval from the start time of annealing to the end of the exothermic interval. The system performs area integration on the thermal residual time series within the effective exothermic interval to obtain the cumulative heat release value, and performs a division operation with the theoretical latent heat of complete crystallization to produce a phase transition completion value characterizing the crystallization ratio. The calculation unit derives the uncrystallized ratio value based on the phase transition completion value, and calculates the delay time value used to compensate for the phase transition lag by combining it with the real-time crystallization rate value. The system accumulates this delay time value to the preset holding time data, dynamically updating it to the final target holding time data. The process execution unit continuously processes the composite magnetic core to be processed according to the pulsed airflow, the input electrical energy after derating and limiting, and the updated target holding time data. When the annealing process reaches the target holding time and the temperature fluctuation derivative remains within a relatively flat range below the steady-state fluctuation threshold during continuous observation, the control system determines that the material has reached the ideal thermodynamic equilibrium point and terminates the input of annealing heat. This control mode, based on the active driving of the bulk phase change process, fundamentally eliminates the performance dispersion caused by empirical adjustments.

[0077] The preparation method of the composite magnetic core material of this invention includes a comparative verification experiment. The experimental group prepared the composite magnetic core material using the preparation method of this invention, while the control group prepared the composite magnetic core material using a conventional, known method with a fixed temperature and duration. 100 composite magnetic core test samples were prepared in both the experimental and control groups. High-precision measuring equipment measured the average initial permeability of the composite magnetic cores prepared in the experimental group to be 125,000, while the high-precision measuring equipment simultaneously measured the average initial permeability of the composite magnetic cores prepared in the control group to be only 98,000. Dedicated testing instruments detected a hysteresis loss of 12 watts per kilogram for the composite magnetic cores prepared in the experimental group, while the hysteresis loss of the composite magnetic cores prepared in the control group was as high as 18 watts per kilogram. Quality control personnel calculated the yield rate of the composite magnetic cores prepared in the experimental group to be 98%, while the yield rate of the composite magnetic cores prepared in the control group was 82%. The quantitative test data output by various measuring devices and testing instruments objectively prove that the method provided in the application document does indeed significantly improve the permeability and overall yield of the composite magnetic core material. The quantitative test data output by various measuring devices and testing instruments also prove that the method provided in the application document effectively reduces the high-frequency hysteresis loss of the composite magnetic core material.

Claims

1. A composite magnetic core material and its preparation method, characterized in that, include: Step 1: Collect the target performance index data and basic material property data of the composite magnetic core to be processed, extract the key feature variables from the target performance index data and the basic material property data, and structurally splice the key feature variables to generate an initial feature vector characterizing the initial state of the composite magnetic core to be processed. Step 2: Input the target performance index data and the material basic property data into the prediction model for inverse solving, and output the process trajectory matrix. The process trajectory matrix includes annealing temperature setting data, protective atmosphere flow rate setting data, and preset holding time data. Step 3: Input the annealing heat according to the annealing temperature setting data, and input the protective atmosphere according to the protective atmosphere flow setting data to perform annealing on the composite magnetic core to be processed, and collect the actual temperature data and actual input power data consumed during the annealing process. Step 4: Calculate the theoretical heating temperature rise data by combining the actual input electrical power data and the preset heat capacity mapping rule; perform a difference operation between the actual temperature data and the theoretical heating temperature rise data to obtain the thermal residual time series. Step 5: Perform time derivative on the thermal residual time series to obtain the exothermic derivative. When the exothermic derivative is greater than a preset derivative threshold, generate a decoupled control command including a power attenuation command and a pulse modulation command. Use the power attenuation command to attenuate the input power corresponding to the actual input power data, and use the pulse modulation command to modulate the constant airflow corresponding to the protective atmosphere flow rate setting data into a pulsed airflow.

2. The method for preparing composite magnetic core material according to claim 1, characterized in that, The target performance index data and the basic material property data are input into the prediction model for inverse solving, and the process trajectory matrix is ​​output, including: Extract the set crystallization volume fraction and set grain size values ​​from the target performance index data, and screen out the crystallization activation energy data of the nanocrystalline phase and the grain growth activation energy data of the permalloy phase from the material basic physical property data. After using the maximum-minimum normalization algorithm to uniformly and linearly map the crystallization activation energy data, the grain growth activation energy data, and the target performance index data to a dimensionless numerical range, the data are input into the prediction model including the Avramie phase transformation kinetic equation for forward mapping operation. This establishes a first numerical correlation mapping between the annealing temperature setting data and the volume fraction of the nanocrystalline phase, and a second numerical correlation mapping between the annealing temperature setting data and the grain size of the permalloy phase. The optimization algorithm is used to update the algorithm through multiple rounds of iteration. The iteration is terminated when the number of iterations reaches a preset maximum iteration threshold or the continuous change of the global optimal fitness value is less than a preset convergence accuracy. Then, the intersection data points that simultaneously satisfy the set crystal volume fraction and the set grain size value are extracted from the first numerical correlation map and the second numerical correlation map. The parameters corresponding to the intersection data points are summarized into the process trajectory matrix.

3. The method for preparing composite magnetic core material according to claim 2, characterized in that, The annealing heat is input according to the annealing temperature setting data, and the protective atmosphere is input according to the protective atmosphere flow rate setting data. Annealing is then performed on the composite magnetic core to be processed. Actual temperature data and actual power consumption data during the annealing process are collected, including: Obtain the geometric dimensions of the annealing space in which the annealing process is performed, as well as the kinetic viscosity data of the protective atmosphere; The spatial thermal inertia delay time value is derived by performing correlation calculations on the geometric dimension data and the kinetic viscosity data. The preset sampling frequency is set according to the reciprocal of the spatial thermal inertia delay time value; The annealing heat is input using the annealing temperature setting data as the heat input condition, and the protective atmosphere is delivered to the annealing processing space in a constant state using the protective atmosphere flow rate setting data as the gas transmission condition. The actual temperature data inside the annealing processing space is continuously collected according to the preset sampling frequency, and the actual input electrical power data of the annealing heat consumption is recorded simultaneously.

4. The method for preparing composite magnetic core material according to claim 3, characterized in that, The actual input electrical power data and the preset heat capacity mapping rule are used to calculate the theoretical heating temperature rise data. The actual temperature data and the theoretical heating temperature rise data are then compared to obtain the thermal residual time series, including: Extract the heat capacity values ​​of the nanocrystalline phase, the heat capacity values ​​of the permalloy phase, the mass data of the nanocrystalline phase, the mass data of the permalloy phase, and the background heat capacity values ​​of the heating hardware used to input the annealing heat of the composite magnetic core to be processed. The total input heat value is obtained by performing time integration on the actual input electrical power data in the same time dimension. The heat capacity of the nanocrystalline phase is obtained by multiplying the heat capacity value of the nanocrystalline phase by the mass data of the nanocrystalline phase. The heat capacity value of the permalloy phase is obtained by multiplying the heat capacity value of the permalloy phase by the mass data of the permalloy phase. The total heat capacity value of the system is obtained by summing the heat capacity values ​​of the nanocrystalline phase, the permalloy phase, and the background heat capacity value. Dividing the total input heat value by the total heat capacity of the system, the theoretical heating temperature rise data under the condition of no material phase change heat release is calculated; Under the same time dimension, the temperature difference between the actual temperature data and the theoretical heating temperature rise data is calculated, and the temperature difference values ​​obtained at consecutive time points are arranged to generate the thermal residual time series.

5. The method for preparing composite magnetic core material according to claim 4, characterized in that, The thermal residual time series is time-derived to obtain the exothermic derivative. When the exothermic derivative is greater than a preset derivative threshold, a decoupling control command including a power attenuation command and a pulse modulation command is generated, including: Obtain the critical temperature gradient value for the secondary crystallization of the permalloy phase; The critical temperature gradient value is calculated by ratioing the latent heat of phase transformation per unit mass of the nanocrystalline phase to obtain the maximum safe heat release rate value that ensures no secondary crystallization occurs. The maximum safe heat release rate value is set as the preset derivative threshold value. The exothermic derivative is obtained by performing time derivative operation on the thermal residual time series; When the exothermic derivative is greater than the preset derivative threshold, it is determined that the exothermic rate of the nanocrystal primary crystal exceeds the system's thermal conductivity, and the decoupling control command is generated to simultaneously limit the input electrical energy and increase convective heat transfer.

6. The method for preparing composite magnetic core material according to claim 5, characterized in that, The power attenuation command is used to attenuate the input electrical energy corresponding to the actual input electrical power data, including: Extract the maximum permissible derating rate value of the power supply hardware; Calculate the absolute value of the difference between the exothermic derivative and the preset derivative threshold, and calculate the exothermic excess ratio by comparing the absolute value of the difference with the preset derivative threshold. The target decay slope value is obtained by multiplying the exothermic excess ratio value by the maximum allowable derating rate value. The input electrical energy is dated and limited by the target attenuation slope value to restrict the heat source input.

7. The method for preparing composite magnetic core material according to claim 6, characterized in that, Modulating the constant airflow corresponding to the protective atmosphere flow rate setting data into a pulsed airflow using the pulse modulation command includes: Extract the surface area value of the composite magnetic core to be processed and the volume value of the annealing processing space; The target pulse frequency value is derived by multiplying the surface area value and the heat release excess ratio value, and the target duty cycle value to ensure the turbulent heat transfer depth is derived based on the volume value. Adjust the opening and closing frequency and opening duration of the gas delivery channel for delivering the protective atmosphere according to the target pulse frequency value and the target duty cycle value. The constant airflow is divided into pulsed airflows with peak velocity characteristics and trough velocity characteristics, and the latent heat accumulated on the surface of the composite magnetic core to be treated is dissipated by the dynamic pressure difference of the pulsed airflows.

8. The method for preparing composite magnetic core material according to claim 7, characterized in that, After modulating the constant airflow corresponding to the protective atmosphere flow rate setting data into a pulsed airflow using the pulse modulation command, the process includes: The zero-crossing time data of the exothermic derivative from positive to negative is extracted as the end point of the exothermic interval, and the time period from the start time of the annealing process to the end point of the exothermic interval is defined as the effective exothermic interval. Perform time integration on the thermal residual time series within the effective heat release interval to obtain the cumulative heat release value; Extract the total theoretical latent heat of complete crystallization, divide the cumulative heat release value by the total theoretical latent heat of complete crystallization value, and obtain the phase transformation completion value characterizing the crystallization ratio.

9. The method for preparing composite magnetic core material according to claim 8, characterized in that, After obtaining the phase transition completion value, the following is included: Extract the real-time crystallization rate value corresponding to the current time node from the actual temperature data; Subtract the phase transformation completion value from 1 to obtain the non-crystallization ratio value; Multiply the uncrystallized ratio by the theoretical total mass of the composite magnetic core to be processed to obtain the remaining mass to be crystallized. Divide the remaining mass of crystallizers by the real-time crystallization rate to calculate the delay time that needs to be compensated. The delay duration value is added to the preset insulation time data to generate the target insulation time data.

10. The method for preparing composite magnetic core material according to claim 9, characterized in that, After generating the target insulation time data, the following is included: The composite magnetic core to be processed is continuously annealed based on the pulsed airflow, the input electrical energy after derating and limiting, and the target heat preservation time data. Extract the temperature fluctuation derivative at the current time point; When the annealing process reaches the target holding time and the temperature fluctuation derivative is less than the steady-state fluctuation threshold during the continuous observation period, the input of the annealing heat is terminated, and the preparation of the composite magnetic core material is completed.

Images