Industrial energy-saving electromagnetic induction heating control method

Abstract

The application discloses an industrial energy-saving electromagnetic induction heating control method, and relates to the technical field of industrial energy-saving heating equipment. The method is characterized in that the sliding window discrete Fourier transform is performed on the resonant tank voltage and current signals, the equivalent impedance envelope and the equivalent compactness index of bulk charging materials are obtained on line, the feedforward frequency correction amount is constructed in combination with the electrical parameter mutation rate, the feedback frequency correction amount based on the phase error is superposed, the switching frequency command is obtained, the partition power limitation command is generated according to the compactness distribution of the charging materials, and the dead time is adjusted according to the phase error and the impedance change rate. The method does not need to increase physical sensors in the furnace, can maintain the zero-voltage open state under the conditions of the severe fluctuation of the charging material impedance and the avalanche mutation of the curie point, and can inhibit the local current concentration. The method is used for the energy-saving control of the charging material preheating process in the metallurgical and foundry industries.

Description

Technical Field

[0001] This invention relates to the field of industrial energy-saving heating equipment technology, specifically to an industrial energy-saving electromagnetic induction heating control method. Background Technology

[0002] In modern metallurgical and casting industries, in order to improve smelting efficiency, electromagnetic induction heating devices are usually used to preheat the furnace charge before it enters the main smelting furnace.

[0003] The object of heating is the bulk furnace charge before entering the main melting furnace, such as scrap steel and metal ore. This type of furnace charge is in a natural accumulation state in the preheating cavity and has random porosity distribution characteristics. The electromagnetic induction heating device usually includes a rectifier module, an inverter module, a resonant tank circuit composed of an induction coil and a resonant capacitor, and a control module. The control module is used to generate drive pulses for the inverter bridge switching tube and control the switching frequency and duty cycle. Compared with the traditional gas preheating method, electromagnetic induction preheating has the advantages of fast heating speed and no exhaust gas emission.

[0004] Existing industrial electromagnetic induction heating control systems generally employ a frequency tracking control method based on phase-locked loops (PLLs). This method collects voltage and current signals from both ends of the resonant tank circuit, detects their phase difference, and adjusts the switching frequency according to the phase difference. This gradually brings the inverter bridge's switching frequency closer to the natural resonant frequency of the resonant tank circuit, maintaining a zero-voltage turn-on state. This is then combined with constant current or constant power outer loop control to achieve preheating control of the furnace charge. However, this scheme treats the load as a quasi-constant impedance model, failing to consider the dynamic changes in the bulk charge's stacking state on the equivalent impedance, or the impact of the sudden drop in permeability near the Curie point of the ferromagnetic charge on the resonant frequency. When the load impedance fluctuates drastically, frequency tracking lag or even loss of lock-on can easily occur, causing the inverter to leave the zero-voltage turn-on state and resulting in significant switching losses. Furthermore, existing schemes cannot detect local differences in the density of the charge stacking, which can easily lead to concentrated current at local overlap points and premature melting of the charge, affecting the uniformity of the preheating process and the energy-saving effect. Therefore, we propose an industrial energy-saving electromagnetic induction heating control method. Summary of the Invention

[0005] The technical problem this invention aims to solve is as follows: Existing industrial electromagnetic induction heating control systems employ a phase-locked loop (PLL)-based frequency tracking control method, treating the bulk furnace charge load as a quasi-constant impedance model. When the bulk furnace charge experiences non-stationary and nonlinear fluctuations in equivalent load impedance due to random accumulation, thermal movement, and local softening and collapse, and when the permeability of the ferromagnetic charge drops sharply near the Curie point, causing an avalanche-like decrease in equivalent inductance, the progressive frequency adjustment of the PLL exhibits a response lag, easily leading to frequency lockout. This causes the inverter to leave the zero-voltage turn-on state, resulting in significant switching losses. Furthermore, existing solutions cannot detect local differences in the density of the furnace charge, easily causing current concentration at local overlap points and premature melting of the charge, reducing the uniformity of the preheating process and overall energy utilization.

[0006] To solve the above-mentioned technical problems, the present invention provides an industrial energy-saving electromagnetic induction heating control method for electromagnetic induction preheating of bulk furnace material placed in the cavity of an induction coil. The method includes: acquiring voltage signals at both ends of the resonant channel and current signals in the induction coil.

[0007] The voltage signal and the current signal are subjected to sliding window discrete Fourier transform respectively to extract the amplitude and phase of specific harmonic components. The equivalent impedance complex value is calculated accordingly, and the equivalent impedance envelope is obtained according to its historical sequence. Then, the equivalent density index of the furnace charge is calculated.

[0008] The electrical parameter mutation rate is calculated based on the current and voltage values ​​at adjacent sampling times. The equivalent inductance value is obtained by converting the equivalent impedance complex value and the natural resonant frequency of the resonant tank circuit is calculated. The feedforward frequency correction is calculated based on the variation trend of the natural resonant frequency and the electrical parameter mutation rate.

[0009] The phase error between the voltage and current in the resonant tank circuit is detected, and the feedback frequency correction is obtained through proportional-integral-differential calculations.

[0010] Add the feedforward frequency correction to the feedback frequency correction to obtain the switching frequency command for the next switching cycle.

[0011] Based on the equivalent density index of the furnace charge, the induction coil is divided into multiple heating sections along the axial direction. The local density deviation of each heating section is calculated, and power limiting commands for each heating section are generated accordingly.

[0012] The phase error between the zero crossing moment of the induction coil current and the theoretical turn-on moment of the switch is detected, and the dead time command for the next switching cycle is calculated by combining the rate of change of the equivalent impedance amplitude.

[0013] Based on the switching frequency command, power limit command, and dead time command, a pulse width modulation drive pulse is generated to drive the inverter bridge to supply power to the induction coil, thereby achieving preheating of the bulk furnace material.

[0014] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows:

[0015] 1. This invention can obtain the dynamic change envelope of the equivalent impedance of the furnace charge and the density index online without adding physical sensors in the furnace, breaking through the control blind spot of the traditional constant load model that is not sensitive to the random stacking state of the furnace charge.

[0016] 2. This invention uses feedforward-feedback joint predictive control to perform advance adjustment of the switching frequency before impedance avalanche mutation occurs, avoiding the frequency lockout problem caused by response lag in traditional phase-locked loop control. It can maintain zero-voltage turn-on state under conditions of severe load impedance fluctuations and avalanche mutation near the Curie point, thereby reducing the switching losses of the inverter.

[0017] 3. This invention, through the coordinated modulation of spatial adaptive power envelope and dynamic dead time, can proactively reduce the power limit value for sections with large local density deviations in the furnace charge, suppress the phenomenon of premature melting at local overlap points due to current concentration, improve the uniformity of the preheating process, thereby improving the overall energy utilization rate of the industrial-grade induction preheating process and achieving energy-saving control effects. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the overall closed-loop control process of the system in this embodiment of the invention.

[0019] Figure 2 This is a flowchart illustrating the online identification process of the equivalent impedance envelope of bulk furnace charge in an embodiment of the present invention.

[0020] Figure 3 This is a block diagram illustrating the principle of the feedforward-feedback joint predictive control method in an embodiment of the present invention. Detailed Implementation

[0021] The following is in conjunction with the appendix Figure 1-3 The specific embodiments of the present invention will be further described below. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0022] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be noted that, without conflict, the various embodiments and technical features in the embodiments of this application can be combined with each other.

[0023] I. Overall Architecture of the Control System

[0024] The industrial energy-saving electromagnetic induction heating control method provided in this embodiment of the invention relies on an electromagnetic induction heating control system. The system includes a three-phase AC power supply, a rectifier module, a DC bus, an inverter module and a resonant tank circuit in sequence along the power transmission direction, and is equipped with an induction heating furnace chamber for holding bulk furnace materials. The rectifier module is used to rectify the industrial frequency AC power output from the three-phase AC power supply into DC power and output it to the DC bus.

[0025] The inverter module is a full-bridge or half-bridge inverter circuit used to invert the DC power on the DC bus into high-frequency AC power.

[0026] The resonant circuit consists of an induction coil and a resonant capacitor forming an LC resonant circuit. The induction coil is placed on the periphery of the induction heating furnace cavity to electromagnetically induction heat the bulk furnace charge placed in the furnace cavity. The bulk furnace charge refers to metal scrap, such as scrap steel and metal ore, which are placed in the furnace cavity in a natural pile state before entering the main smelting furnace without being arranged in a regular manner. Its shape, particle size and pile gap are random. The proportion of void volume in the pile to the total volume of the pile is called porosity. Porosity directly affects the magnetic field coupling effect and equivalent load impedance of the induction coil.

[0027] The system also includes a voltage and current sampling unit connected to the resonant tank circuit to collect the voltage signal at both ends of the resonant tank circuit and the current signal in the induction coil. The digital controller is connected to the voltage and current sampling unit to receive the voltage signal and the current signal. The digital controller is equipped with an impedance online identification module, a feedforward-feedback joint predictive control module, and an adaptive power envelope and dead-time co-modulation module. The outputs of the three modules are fed into the PWM drive pulse generation unit, which generates drive pulses and outputs them to the inverter module to control the switching frequency, duty cycle, and dead time of the switching transistors in the inverter module. The digital controller can be implemented using a combination of a digital signal processor and a field-programmable gate array to meet the computational speed requirements of real-time high-frequency signal processing.

[0028] II. Online Identification of Equivalent Impedance Envelope of Bulk Furnace Charge

[0029] The online impedance identification module performs the following process for identifying the equivalent impedance envelope of bulk furnace charge: using a fixed sampling period... Voltage signal at both ends of the resonant tank circuit and the current signal in the induction coil Perform synchronous data acquisition and store data of length [length missing]. A sliding cache window;

[0030] Perform sliding-window discrete Fourier transform on the voltage and current sequences within the buffer window to extract the fundamental component corresponding to the inverter bridge switching frequency and the pre-defined specific harmonic components, obtaining the complex form of amplitude and phase information. The formula for sliding-window discrete Fourier transform is:

[0031] ;

[0032] In the formula, Indicates the first Each sampling time, with window length The calculated first The complex components of the second harmonic;

[0033] For the first in the window Voltage or current sample value at any given time;

[0034] The sliding window length is the number of sampling points contained within the window.

[0035] For harmonic order;

[0036] The imaginary unit;

[0037] This is the sampling sequence number variable within the window.

[0038] Based on the complex components of the same harmonic of voltage and current, the complex value of the equivalent impedance at the current moment is calculated according to the following formula:

[0039] ;

[0040] In the formula, For the first The complex value of the equivalent impedance at time t;

[0041] , The voltage signal and the current signal are respectively in the first... Results of sliding window discrete Fourier transform under subharmonics;

[0042] , The phase angles of the voltage and current harmonic components are respectively. The equivalent impedance amplitude and phase obtained at each sampling time are sequentially stored into the historical sequence, thus forming the equivalent impedance envelope. This envelope is used to characterize the equivalent load impedance characteristics that dynamically fluctuate with the change of the furnace charge stacking state.

[0043] The equivalent impedance envelope is differentially processed, and the rate of change of impedance amplitude at adjacent sampling times is calculated, i.e., the local slope of the envelope, which is used to characterize the severity and trend direction of the change in the state of the furnace charge.

[0044] Combined with the historical limit of impedance amplitude recorded during the current batch heating process Compared with historical lower limit The current impedance amplitude is normalized to obtain the equivalent density index of the furnace charge:

[0045] ;

[0046] In the formula, For the first The equivalent density index of the furnace charge obtained by time estimation has a value range of 0 to 1. The larger the value, the denser the furnace charge and the lower the porosity.

[0047] , These are the historical lower limit and historical upper limit values ​​of the equivalent impedance amplitude during the heating process of this batch, respectively, used for normalization processing. The equivalent impedance complex value, impedance change rate and density index are output to the feedforward-feedback joint predictive control module and the adaptive power envelope and dead zone cooperative modulation module as the input basis for subsequent control decisions, and are periodically updated to form an online identification closed loop.

[0048] The above identification process relies solely on the existing voltage and current sampling points at both ends of the resonant channel to obtain signals. No physical sensors need to be installed inside the high-temperature furnace throughout the process, and a dynamic equivalent load model reflecting the changes in the porosity and packing state of the furnace charge can be obtained.

[0049] III. Feedforward-Feedback Joint Predictive Control for Addressing Curie Point Impedance Avalanche Transitions

[0050] The feedforward-feedback joint predictive control module adopts a dual-closed-loop predictive control structure, including a feedforward prediction channel and a feedback correction channel. The feedforward prediction channel reads the equivalent inductance estimate output by the online impedance identification module. The natural resonant frequency of the resonant tank circuit at the current moment is calculated using the following formula:

[0051] ;

[0052] In the formula, For the first The natural resonant frequency of the resonant tank circuit at any given moment;

[0053] The equivalent inductance value is obtained by the impedance online identification module;

[0054] The capacitance value of the resonant capacitor in the resonant tank circuit is a fixed parameter.

[0055] Calculate the electrical parameter mutation rate based on the current and voltage sampling values ​​of adjacent sampling periods:

[0056] ;

[0057] In the formula, , The first The rate of change between sampling points of the induction coil current and the resonant tank voltage at all times;

[0058] , These are the current values ​​for the current and previous sampling periods, respectively;

[0059] , These are the voltage values ​​for the current and previous sampling periods, respectively;

[0060] The sampling period.

[0061] The electrical parameter mutation rate is compared with a pre-set avalanche mutation detection threshold. When the mutation rate exceeds the threshold, it is determined that the furnace charge is in the stage of impedance drastic change near the Curie point, and the feedforward channel gain coefficient is switched to the pre-tuned enhancement mode gain.

[0062] When the mutation rate does not exceed the threshold, the gain of the normal mode is maintained. Based on the trend of resonant frequency change and current mutation rate, the feedforward frequency correction is calculated according to the following formula:

[0063] ;

[0064] In the formula, This is the frequency correction amount for the next switching cycle obtained through the feedforward channel prediction.

[0065] The feedforward gain coefficient represents the trend of resonant frequency variation.

[0066] The feedforward gain coefficient is the current mutation rate. Both are obtained through offline tuning and automatically switch to enhanced gain mode when an avalanche mutation is detected.

[0067] The feedback correction channel detects the phase error between the voltage and current in the resonant tank circuit. This phase error is the deviation relative to the target zero-voltage turn-on phase margin. It is input into a proportional-integral-derivative (PID) controller, and the feedback frequency correction is calculated using the following formula:

[0068] ;

[0069] In the formula, The frequency correction amount obtained from feedback channel calibration;

[0070] , , The proportional, integral, and derivative coefficients are respectively used. The feedforward frequency correction and the feedback frequency correction are superimposed at the summation node to obtain the switching frequency command for the next switching cycle.

[0071] ;

[0072] The switching frequency command is combined with the current power demand to synchronously calculate the corresponding duty cycle command, which is then output to the PWM drive pulse generation unit. Since the feedforward channel calculates the frequency correction amount in advance based on the resonant frequency change trend and current mutation rate at the initial stage of impedance avalanche mutation, compared with the traditional phase-locked loop control that relies on phase error gradual adjustment, it can avoid frequency lockout caused by response lag, so that the inverter module can continuously maintain the zero voltage turn-on state.

[0073] IV. Multidimensional Adaptive Power Envelope and Dead Time Co-modulation

[0074] The adaptive power envelope and dead-zone co-modulation module reads the overall density index output by the online impedance identification module. The induction coil is divided into multiple heating sections along the axial direction, and the local density deviation of each heating section is estimated separately. This deviation characterizes the position. The degree to which the density of the furnace charge deviates from the overall average density is used to calculate the power limit command section by section according to the following formula:

[0075] ;

[0076] In the formula, The axial position of the induction coil Place, No. Power limiting command at any given time;

[0077] This is the base target power for this heating stage;

[0078] This is the local power suppression coefficient;

[0079] For position The deviation of the furnace charge density from the overall average density is measured. The larger the deviation, the more likely it is that current concentration will form at the overlap point. For heating sections where the local density deviation is greater than the preset deviation threshold, the power limit command value of the section is automatically reduced, thereby suppressing the phenomenon of local current concentration.

[0080] Simultaneously, this module detects the zero-crossing moment of the induction coil current and compares it with the current theoretical switching moment of the switch, calculating the phase error. ;

[0081] Based on the real-time rate of change of the equivalent impedance magnitude, the dead time command for the next switching cycle is calculated according to the following formula:

[0082] ;

[0083] In the formula, Dead time instruction for the next switching cycle;

[0084] The preset base dead time;

[0085] , These are the adjustment coefficients for the phase error term and the impedance magnitude change rate term, respectively.

[0086] This is the phase error between the zero crossing moment of the coil current and the theoretical turn-on moment of the switch.

[0087] The rate of change of the equivalent impedance amplitude, the power limiting command and dead time command of each section are summarized and combined with the switching frequency command and duty cycle command output by the feedforward-feedback joint predictive control module as the final driving parameters, and sent to the PWM drive pulse generation unit.

[0088] V. Overall Closed-Loop Control Process

[0089] In the overall control method of this invention, after the system is started, the voltage and current sampling unit first performs high-speed sampling of the voltage and current signals at both ends of the resonant tank circuit;

[0090] Subsequently, the online impedance identification module performs online identification of the equivalent impedance envelope and outputs the equivalent impedance complex value, impedance change rate and equivalent density index of the furnace charge.

[0091] Next, determine whether the current electrical parameter mutation rate exceeds the avalanche mutation detection threshold. If it does, switch the feedforward gain coefficient to enhanced mode; otherwise, maintain the normal mode.

[0092] The switching frequency command for the next switching cycle is then calculated by the feedforward-feedback joint predictive control module.

[0093] Simultaneously, the adaptive power envelope and dead-time co-modulation module calculates the power limiting command and dead-time command for each heating section;

[0094] Finally, after all instructions are summarized, a PWM drive pulse is generated and sent to the inverter module to drive the induction coil to heat the bulk furnace material, and then the process returns to the sampling step, forming a closed-loop control that is executed periodically. Specific Implementation

[0095] Taking an industrial scrap steel electromagnetic induction preheating production line as an example, the implementation process of this invention is explained. In this production line, the induction heating furnace is filled with scrap steel material in a naturally stacked state. The inverter module adopts a full-bridge topology, and the resonant capacitor value of the resonant tank circuit is fixed. After the system starts, the voltage and current sampling unit synchronously collects the voltage and current of the resonant tank circuit at a fixed sampling period. The impedance online identification module extracts the fundamental wave and the set harmonic components periodically based on the sliding window discrete Fourier transform, calculates the equivalent impedance complex value and forms the impedance envelope. In the initial stage of heating, the density distribution of the stacked material is uneven, and the density index shows regional differences. The adaptive power envelope and dead zone cooperative modulation module issues a lower power limit command to the section with a large density deviation to suppress the current concentration in the section. As the heating process continues, the scrap steel temperature gradually rises and approaches the Curie point, the equivalent inductance decreases rapidly, and the electrical parameter mutation rate increases accordingly.

[0096] When the mutation rate exceeds the preset avalanche mutation detection threshold, the feedforward-feedback joint predictive control module automatically switches the feedforward gain coefficient to enhancement mode, calculates in advance and issues the switching frequency command for the next switching cycle, so that the switching frequency of the inverter module can follow the jump of the natural resonant frequency of the resonant tank circuit, and continuously maintain the zero-voltage turn-on state, avoiding the loss of lock-up and hard switching phenomena commonly found in traditional phase-locked loop control schemes. During the entire preheating process, each functional module executes in a rolling manner according to the above closed-loop process until the furnace charge completes the preheating process.

[0097] VII. Other Implementation Methods

[0098] In another embodiment of the present invention, the feedforward prediction channel may not use linear prediction calculation based on voltage and current mutation rate, but instead use a lightweight recurrent neural network model trained based on historical impedance envelope data. The equivalent density index of the furnace charge, impedance amplitude and phase history sequence are used as inputs to directly output the frequency correction and duty cycle prediction value of the next switching cycle. The same technical effect as the aforementioned feedforward-feedback joint prediction control is achieved by combining offline training and online inference. This embodiment is particularly suitable for application scenarios with more complex and variable furnace charge types and stacking characteristics.

[0099] In another embodiment of the present invention, the impedance online identification module may not use the sliding window discrete Fourier transform, but instead use a state estimation method based on Kalman filtering. The equivalent resistance and equivalent inductance of the resonant tank circuit are used as system state variables, and the sampled voltage and current are used as observations. The dynamic change trajectory of the equivalent impedance is estimated in real time through recursive filtering. Similarly, a dynamic load model reflecting the change of furnace material porosity can be obtained without adding physical sensors, as an alternative implementation of the aforementioned core identification method.

[0100] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims

1. An industrial energy-saving electromagnetic induction heating control method for electromagnetic induction preheating of bulk furnace charge placed inside an induction coil cavity, the method comprising: Collect the voltage signal at both ends of the resonant tank circuit and the current signal in the induction coil; The voltage signal and the current signal are subjected to sliding window discrete Fourier transform respectively to extract the amplitude and phase of specific harmonic components in the voltage signal and the current signal. The equivalent impedance complex value is calculated based on the amplitude and phase. The equivalent impedance envelope is obtained based on the historical sequence of the equivalent impedance complex value. The equivalent density index of the furnace charge is calculated based on the equivalent impedance envelope. The electrical parameter mutation rate is calculated based on the current and voltage values ​​at adjacent sampling times. The equivalent inductance value is obtained by converting the equivalent impedance complex value. The natural resonant frequency of the resonant tank circuit is calculated based on the equivalent inductance value. The feedforward frequency correction is calculated based on the changing trend of the natural resonant frequency and the electrical parameter mutation rate. When the electrical parameter mutation rate exceeds the preset avalanche mutation detection threshold, the gain coefficient used to calculate the feedforward frequency correction is switched from the conventional mode gain coefficient to the enhanced mode gain coefficient to achieve an advanced response to Curie point avalanche impedance mutation. The electrical parameter mutation rate is compared with a preset avalanche mutation detection threshold. When the electrical parameter mutation rate exceeds the avalanche mutation detection threshold, the gain coefficient used to calculate the feedforward frequency correction is switched to the enhanced mode gain coefficient. When it does not exceed the threshold, the normal mode gain coefficient is maintained. The phase error between the voltage and current of the resonant tank circuit is detected, and the feedback frequency correction is obtained by proportional-integral-differential calculation based on the phase error. The feedforward frequency correction amount is added to the feedback frequency correction amount to obtain the switching frequency command for the next switching cycle. Based on the equivalent density index of the furnace charge, the induction coil is divided into multiple heating sections along the axial direction. The local density deviation of each heating section is calculated, and a power limiting command for each heating section is generated based on the local density deviation. The zero-crossing moment of the induction coil current is detected, the phase error between the zero-crossing moment and the theoretical turn-on moment of the switch is calculated, and the dead time command for the next switching cycle is calculated by combining the rate of change of the complex value of the equivalent impedance. A pulse width modulation (PWM) drive pulse is generated according to the switching frequency command, the power limiting command, and the dead time command, and the PWM drive pulse is output to the inverter bridge to drive the inverter bridge to supply power to the induction coil for preheating the bulk furnace charge.

2. The method according to claim 1, characterized in that, The calculation of the equivalent density index of the furnace charge based on the equivalent impedance envelope includes: recording the historical upper limit and lower limit of the equivalent impedance amplitude within the current heating batch, and performing normalization calculation based on the current equivalent impedance amplitude, the historical upper limit and the historical lower limit to obtain the equivalent density index of the furnace charge. The equivalent density index of the furnace charge is used to characterize the density of the furnace charge accumulation.

3. The method according to claim 1, characterized in that, The method further includes: The electrical parameter mutation rate is compared with a preset avalanche mutation detection threshold; When the electrical parameter mutation rate exceeds the avalanche mutation detection threshold, the gain coefficient used to calculate the feedforward frequency correction will be switched to the enhancement mode gain coefficient. When the electrical parameter mutation rate does not exceed the avalanche mutation detection threshold, the gain coefficient used to calculate the feedforward frequency correction is maintained as the normal mode gain coefficient.

4. The method according to claim 1, characterized in that, The step of generating power limiting instructions for each heating segment based on the local density deviation includes: for heating segments where the local density deviation is greater than a preset deviation threshold, reducing the power limiting instruction value corresponding to the heating segment to suppress current concentration in the heating segment.

5. The method according to claim 1, characterized in that, The specific subharmonic components include the fundamental component corresponding to the inverter bridge switching frequency.

6. The method according to claim 1, characterized in that, The acquisition of voltage signals at both ends of the resonant tank and current signals in the induction coil includes: synchronously acquiring voltage signals at both ends of the resonant tank and current signals in the induction coil at a fixed sampling period, and storing the acquired voltage signals and current signals into a sliding buffer window of a preset length for use by the sliding window discrete Fourier transform.

7. The method according to claim 1, characterized in that, After obtaining the equivalent impedance envelope based on the historical sequence of the equivalent impedance complex values, the method further includes: performing differential processing on the equivalent impedance envelope, calculating the impedance amplitude change rate at adjacent sampling times, and outputting the impedance amplitude change rate as a parameter characterizing the severity of changes in the furnace charge state to the calculation process of the feedforward frequency correction amount and the calculation process of the dead time command.

8. The method according to claim 1, characterized in that, The bulk furnace charge includes ferromagnetic furnace charge. When the temperature of the ferromagnetic furnace charge is within ±20°C of the preset Curie temperature point, the feedforward frequency correction amount compensates for the resonant frequency shift caused by the decrease in the magnetic permeability of the ferromagnetic furnace charge.

9. The method according to claim 1, characterized in that, The electrical parameter mutation rate includes the rate of change of current in the induction coil and the rate of change of voltage in the resonant tank circuit.

10. An industrial energy-saving electromagnetic induction heating control system, used to execute the method according to any one of claims 1 to 9, characterized in that, It includes a rectifier module, an inverter module, a resonant tank circuit, a voltage and current sampling unit, and a digital controller; The resonant groove is composed of an induction coil and a resonant capacitor, and the induction coil is arranged around the periphery of the induction heating furnace cavity used to accommodate bulk furnace materials. The voltage and current sampling unit is connected to the resonant tank circuit and is used to collect the voltage signal at both ends of the resonant tank circuit and the current signal in the induction coil. The digital controller includes an online impedance identification module, a feedforward-feedback joint predictive control module, and an adaptive power envelope and dead-zone co-modulation module. The impedance online identification module is connected to the voltage and current sampling unit and is used to perform the sliding window discrete Fourier transform and output the equivalent impedance complex value and the equivalent density index of the furnace charge. The feedforward-feedback joint predictive control module is connected to the impedance online identification module and is used to output the switching frequency command; The adaptive power envelope and dead-time co-modulation module is connected to the online impedance identification module and is used to output the power limiting command and the dead-time command. The digital controller generates pulse width modulation drive pulses according to the switching frequency command, the power limiting command, and the dead time command, and outputs the pulse width modulation drive pulses to the inverter module.

Images