Wide aluminum alloy plate continuous annealing system

Abstract

The invention relates to the technical field of aluminum processing equipment, and discloses a wide aluminum alloy plate continuous annealing system which comprises an air cushion furnace, an air supply cooling mechanism, a flow field sensing array and a central control device. The sensing array collects the pressure in the furnace, and the signal collecting module decomposes the pressure into a time-average component and a pulsation component; the dynamic baseline reconstruction module calculates a normalized pressure factor by using a flow resistance network model in combination with a fan rotating speed and a valve opening, and eliminates equipment action interference; the plate shape analysis module adjusts the gate array according to the normalization factor so as to control the plate shape; the flutter monitoring module adjusts the rotating speed of the fan according to the pulsation component, and changes the pneumatic rigidity to suppress flutter; the heat balance module corrects the plate linear speed according to the heat exchange capacity change. According to the invention, a plate shape detection signal and an equipment operation state are decoupled, multi-physical field cooperative control is realized, and plate shape flatness, operation stability and process temperature constancy are ensured.

Description

Technical Field

[0001] This invention relates to the field of aluminum processing equipment technology, specifically to a wide-width continuous annealing system for aluminum alloy sheets. Background Technology

[0002] With the development of lightweighting in the automotive and aerospace industries, the demand for high-surface-quality wide aluminum alloy sheets is increasing. Air-cushion continuous annealing furnaces utilize high-speed airflow from upper and lower nozzles to form an air cushion, achieving non-contact suspension support and heat treatment of the sheet material, thus avoiding surface scratches caused by roller hearth furnaces.

[0003] However, existing systems suffer from signal coupling issues in plate shape evaluation. Common techniques rely on the static pressure distribution of the air cushion to infer plate shape, but the static pressure value is simultaneously affected by the plate suspension gap, the circulating fan speed, and the opening of the regulating valve. When adjusting the air volume to adapt to the process, the pressure baseline drifts, and the control system struggles to distinguish whether the pressure change originates from plate defects or equipment adjustments, resulting in insufficient accuracy in plate shape identification.

[0004] Furthermore, wide-format sheets are prone to flutter under airflow support. Existing air cushion supports have fixed stiffness and lack active suppression mechanisms, making them susceptible to instability, slapping, or even scraping under high-speed conditions. Simultaneously, adjusting fans or valves alters the convective heat transfer coefficient. Current control strategies often overlook the impact of this flow field change on sheet temperature, leading to annealing temperature fluctuations during sheet shape correction or vibration suppression operations, thus affecting product performance consistency. Summary of the Invention

[0005] This invention provides a wide-width continuous annealing system for aluminum alloy sheets, which can eliminate the coupling interference between sheet shape correction, aerodynamic stability control and process temperature maintenance, and achieve high-precision sheet shape flatness and high-speed stable operation.

[0006] This invention provides a continuous annealing system for wide-width aluminum alloy sheets, comprising an air-cushion annealing furnace body, an air supply and cooling actuator, a multi-dimensional flow field sensing array, and a central control device. The air-cushion annealing furnace body has a furnace chamber channel along the sheet's running direction, and upper and lower static pressure boxes for generating air cushion support are arranged on the upper and lower sides of the furnace chamber channel.

[0007] The air supply and cooling actuator includes a circulating fan connected to the upper and lower static pressure boxes, and an array of airflow regulating valves controlling the lateral airflow distribution in the upper and lower static pressure boxes. The multi-dimensional flow field sensing array includes pressure sensors arranged inside the upper and lower static pressure boxes.

[0008] The central control device includes a signal acquisition and splitting module, which is used to acquire the signal from the pressure sensor and decompose the original pressure signal into a time-averaged pressure component characterizing hydrostatic characteristics and a pulsating pressure component characterizing hydrodynamic characteristics. The dynamic baseline reconstruction module is used to calculate the theoretical reference pressure based on the circulating fan speed and the opening of the air volume regulating valve array, and to generate a normalized pressure factor by combining the time-averaged pressure component, thereby eliminating the interference of the air supply and cooling actuators on the measurement data. The plate shape feature analysis module is used to adjust the plate shape of the air volume regulating valve array based on the normalized pressure factor. The flutter monitoring and variable stiffness module is used to adjust the speed of the circulating fan according to the pulsating pressure component to suppress the flutter of the plate.

[0009] Preferably, the multidimensional flow field sensing array adopts a layout of horizontal arraying and vertical alignment, that is, pressure tapping points are arranged at equal intervals along the horizontal width direction inside the upper static pressure box and the lower static pressure box, and the corresponding pressure tapping points in the upper static pressure box and the lower static pressure box are aligned in the vertical direction.

[0010] The pressure sensor is connected to the upper and lower static pressure tanks through a high-temperature pressure-sensing tube. The pressure sensor performs high-frequency synchronous acquisition based on the same clock source to ensure that the fluid force state at the same micro-area can be obtained.

[0011] The signal acquisition and splitting module performs dual-channel digital filtering: it extracts the time-averaged pressure component from the original pressure signal using a low-pass filtering algorithm, and extracts the pulsating pressure component using a high-pass or band-pass filtering algorithm. During processing, the system performs time-series alignment compensation on the filtered signal to ensure that the time-averaged pressure component and the pulsating pressure component output at the same time correspond to the same physical sampling time.

[0012] The dynamic baseline reconstruction module stores a flow resistance network model, defines the circulating fan as a pressure potential source, and defines the airflow regulating valve array as a variable fluid resistance. Based on a pre-calibrated equipment fluid characteristic spectrum, the system determines the theoretical reference pressure that the upper and lower static pressure boxes should possess under standard no-load conditions, according to the current circulating fan speed and the opening of the airflow regulating valve array. The normalized pressure factor is generated by calculating the ratio of the measured average pressure component to the theoretical reference pressure. The normalized pressure factor characterizes the deviation of the actual suspension gap of the plate material from the nominal design gap, and its value is independent of the speed adjustment of the circulating fan and the opening adjustment of the airflow regulating valve array.

[0013] The plate shape feature analysis module calculates the plate shape feature index based on the difference ratio of the normalized pressure factors corresponding to the upper and lower static pressure boxes on the same transverse cross section, and reconstructs the transverse plate shape profile of the plate. The plate shape feature analysis module further calculates the transverse temperature difference distribution required to eliminate geometric deformation in the transverse plate shape profile based on the principle of thermoelasticity, and maps the transverse temperature difference distribution to the target opening command of the airflow regulating valve array in the cooling section, generating corrective thermal stress by establishing a non-uniform temperature field across the transverse width of the plate.

[0014] The flutter monitoring and variable stiffness module performs spectral analysis on the pulsating pressure components to obtain the power spectral density and calculates the flutter intensity index. When the flutter intensity index indicates that the system is in a flutter latent state or a flutter divergent state, the system generates a speed and frequency adjustment command for the circulating fan. The adjustment command is used to change the aerodynamic vertical stiffness of the air cushion support system, so that the natural frequency of the air cushion support system avoids the current dominant flutter frequency of the plate.

[0015] The central control device also includes a heat balance collaborative compensation module. The heat balance collaborative compensation module calculates the rate of change of the comprehensive heat exchange capacity factor based on the real-time speed frequency of the circulating fan and the flow correction coefficient of the air volume regulating valve array for the adjustment of the opening of the air volume regulating valve array and the adjustment of the speed of the circulating fan. Based on the law of conservation of energy, and according to the rate of change of the calculated comprehensive heat transfer capacity factor, a correction command is generated for the linear velocity of the plate, and the linear velocity of the plate is dynamically compensated to maintain the target metal temperature of the plate constant during the annealing process.

[0016] The fluid characteristic profile of the equipment is generated through a full-condition traversal scanning calibration. The calibration process includes: Under standard no-load conditions, the circulating fan is controlled to perform step-by-step adjustment within the rated operating range, while the air volume regulating valve array is controlled to perform step-by-step sweeping within the full stroke range. The time-averaged pressure measurement values ​​of the upper and lower static pressure boxes corresponding to each operating point are recorded, and the function relationship between the rotational speed, frequency, and opening position is obtained by fitting using the least squares method.

[0017] When executing the target opening command, the air supply and cooling actuator uses a preset valve flow characteristic inverse model to calculate the opening. Based on the target volumetric flow rate and the pressure difference information across the air volume regulating valve array, it calculates the target opening command after eliminating the influence of the valve's nonlinear flow characteristics, and performs dead-zone verification and amplitude limiting on the output command.

[0018] This invention provides a continuous annealing system for wide-width aluminum alloy sheets. It offers the following advantages: 1. This invention constructs a flow resistance network model and introduces a normalized pressure factor to compare the measured time-averaged pressure component with the theoretical reference pressure calculated based on the fan speed and valve opening. This eliminates background pressure fluctuations caused by air volume adjustment or fan speed change, allowing the control system to respond only to flow field changes caused by plate geometric deformation. This solves the technical problem of traditional systems being unable to distinguish between plate shape changes and equipment adjustment interference under variable air volume conditions.

[0019] 2. This invention utilizes the spectral analysis results of the pulsating pressure component to change the vertical stiffness of the air cushion support system by adjusting the speed of the circulating fan. When flutter signs are detected, the system actively changes the air cushion stiffness so that the system's natural frequency avoids the current dominant flutter frequency of the plate. No additional mechanical damping device is required. This can suppress aerodynamic instability during the high-speed operation of wide plates and prevent the plates from scraping against the furnace body.

[0020] 3. The present invention adopts a thermal balance synergistic compensation strategy, which maps the changes in heat transfer capacity caused by plate shape control and flutter suppression into correction commands for plate linear velocity in real time. By dynamically adjusting the residence time of the plate in the furnace according to the law of conservation of energy, the system offsets the influence of airflow parameter fluctuations on plate temperature, ensuring that the final annealing temperature of the aluminum alloy plate is always kept within the target range of the annealing process while plate shape correction or vibration suppression operations are performed. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall structure of the present invention.

[0022] Figure 2 This is the overall control logic flowchart of the present invention.

[0023] Among them, 10. Air cushion annealing furnace body; 11. Upper static pressure box; 12. Lower static pressure box; 20. Air supply and cooling actuator; 21. Circulating fan; 22. Air volume regulating valve array; 30. Multi-dimensional flow field sensing array; 31. Pressure sensor; 40. Central control device; 41. Signal acquisition and diversion module; 42. Dynamic baseline reconstruction module; 43. Plate shape feature analysis module; 44. Flutter monitoring and variable stiffness module; 45. Thermal balance collaborative compensation module; 2. Plate material. Detailed Implementation

[0024] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] See attached document Figure 1The present invention provides a wide-width continuous annealing system for aluminum alloy plates, the system comprising an air cushion annealing furnace body 10, an air supply and cooling actuator 20, a multi-dimensional flow field sensing array 30, and a central control device 40.

[0026] The main body 10 of the air cushion annealing furnace is provided with a furnace passage for the plate 2 to pass through. An upper static pressure box 11 and a lower static pressure box 12 are arranged on the upper and lower sides of the passage, respectively. The surfaces of the upper static pressure box 11 and the lower static pressure box 12 are provided with nozzles to generate an air cushion that supports the suspension of the plate 2. The air supply and cooling actuator 20 includes a circulating fan 21 connected to the upper static pressure box 11 and the lower static pressure box 12 and an air volume regulating valve array 22. The circulating fan 21 changes the total pressure and aerodynamic stiffness of the output airflow through frequency conversion speed regulation. The air volume regulating valve array 22 consists of multiple independent air valves arranged in parallel, which are used to regulate the air intake distribution of the upper static pressure box 11 and the lower static pressure box 12 along the width direction of the plate 2, thereby establishing a non-uniform flow field and temperature field.

[0027] The multi-dimensional flow field sensing array 30 includes pressure sensors 31 arranged inside the upper static pressure box 11 and the lower static pressure box 12, used for high-frequency acquisition of fluid pressure data in the upper static pressure box 11 and the lower static pressure box 12. The central control device 40 is connected to the circulating fan 21, the air volume regulating valve array 22 and the pressure sensor 31, and is internally equipped with a signal acquisition and splitting module 41, a dynamic baseline reconstruction module 42, a plate shape feature analysis module 43, a flutter monitoring and variable stiffness module 44 and a thermal balance collaborative compensation module 45.

[0028] During operation, the signal acquisition and splitting module 41 decomposes the raw signal acquired by the pressure sensor 31 into a time-averaged pressure component characterizing hydrostatics and a pulsating pressure component characterizing dynamics. The dynamic baseline reconstruction module 42 calculates the theoretical reference pressure based on a pre-stored flow resistance network model, according to the current rotational speed of the circulating fan 21 and the opening degree of the airflow regulating valve array 22, and compares it with the measured time-averaged pressure component to generate a normalized pressure factor, thereby eliminating the interference of equipment operation on the measurement data.

[0029] The plate shape feature analysis module 43 identifies the waviness characteristics of the plate 2 based on the normalized pressure factor and calculates the required transverse temperature difference distribution. Thermal stress correction is achieved by adjusting the opening of the airflow regulating valve array 22. The flutter monitoring and variable stiffness module 44 performs spectral analysis on the pulsating pressure components. When flutter characteristics are detected, a command is generated to adjust the speed of the circulating fan 21, thereby changing the aerodynamic stiffness of the air cushion system to avoid the resonant frequency. The thermal balance collaborative compensation module 45 calculates the change in the comprehensive heat transfer coefficient based on the adjustment actions of the fan and valves, and corrects the operating linear velocity of the plate 2 accordingly to maintain a constant annealing process temperature.

[0030] See attached document Figure 2The central control unit 40 is based on a multi-task parallel processing architecture and performs tasks such as... Figure 2 The closed-loop control method shown is as follows. This control method achieves decoupled control of plate geometry regulation and fluid dynamic stability through a series of logical steps.

[0031] Step S100: Synchronous Acquisition of Full-State Data. The central control device 40 synchronously acquires multi-dimensional state data of the system at a preset sampling period. This multi-dimensional state data includes: the original pressure signals of the upper and lower static pressure boxes in each control zone inside the furnace, acquired by the multi-dimensional flow field sensing array 30. The real-time opening position of each segment air volume regulating valve array 22 is fed back by the air supply and cooling actuator 20. and the real-time speed and frequency of the circulating fan 21 The current operating linear speed of plate 2 All the above data are marked with a uniform timestamp to eliminate timing errors caused by communication delays between hardware modules.

[0032] Step S200: Time-frequency domain separation of the pressure signal. Signal acquisition and splitting module 41 performs time-frequency domain separation on the original pressure signal. Dual-channel digital filtering is performed. Channel one uses a low-pass filtering algorithm to extract the time-averaged pressure component, which reflects the hydrostatic characteristics. The low-pass cutoff frequency is set to a value that can filter out high-frequency airflow disturbances. Channel 2 uses a high-pass filtering algorithm to extract the pulsating pressure component that reflects the fluid dynamics characteristics. The frequency band is preserved to cover the characteristic frequency range of aerodynamic flutter in aluminum alloy sheets. For the specific implementation of the digital filtering algorithm, those skilled in the art can employ conventional techniques such as Infinite Impulse Response (IIR) filtering or Finite Impulse Response (FIR) filtering, which will not be elaborated upon here.

[0033] Step S300: Dynamic baseline reconstruction based on the flow resistance model. The dynamic baseline reconstruction module 42, based on the fluid network impedance principle, eliminates the interference of actuator actions on the measurement data. This is based on the real-time speed and frequency of the circulating fan 21. Real-time opening position of air volume regulating valve array 22 The theoretical reference pressure of the calculation device under standard no-load conditions. This calculation is based on a pre-calibrated equipment fluid characteristic spectrum, and its functional relationship is expressed as follows: ,in, The flow channel coefficient is related to the geometry of the furnace body and bellows. For the air volume regulating valve at its real-time opening position The dimensionless flow resistance characteristic function.

[0034] Subsequently, the time-averaged pressure component was calculated. relative to theoretical reference pressure Normalized stress factor : Among them, the normalized stress factor This is used to characterize the change in flow resistance caused by the actual position of plate 2. When the speed adjustment of the circulating fan 21 causes a change in the absolute pressure of the static pressure box, due to the theoretical reference pressure... The normalized stress factor is updated accordingly. The values ​​remain stable, thereby decoupling the measurement signal from the execution action.

[0035] Step S400: Plate topology reconstruction and inverse thermal stress calculation. The plate shape feature analysis module 43 calculates the plate shape feature index based on the normalized pressure factors corresponding to the upper and lower static pressure boxes on the same transverse section. : ,in, and The horizontal coordinates are respectively Normalized pressure factors for the lower and upper static pressure chambers. Plate shape characteristic index. Directly map the normalized displacement distribution of plate 2 in the vertical direction. If the plate shape characteristic index... If the distribution shows a pattern of high in the middle and low on both sides, it is judged as a mid-wave characteristic; if it shows a pattern of low in the middle and high on both sides, it is judged as a side-wave characteristic.

[0036] Based on the calculated plate shape characteristic index, the system uses the principle of thermoelasticity to calculate the lateral temperature difference distribution required to eliminate the geometric deformation. The temperature difference distribution is then mapped to a target opening command for the cooling section airflow regulating valve array 22. For example, for edge wave characteristics, the control system increases the opening of the airflow regulating valve array 22 corresponding to the edge region of the plate 2, thereby increasing local cooling to generate contraction thermal stress and tightening the edge of the plate 2.

[0037] Step S500: Pneumatic flutter monitoring and variable stiffness execution. The flutter monitoring and variable stiffness module 44 processes the pulsating pressure components acquired in step S200. A Fast Fourier Transform (FFT) is performed to obtain the real-time power spectral density. The system monitors the maximum energy peak in the power spectrum. When this peak exceeds a preset safety threshold, it is determined that aeroelastic flutter has occurred in plate 2.

[0038] Upon detecting flutter, the system generates a speed and frequency adjustment command for the circulating fan 21. This is achieved by changing the real-time speed and frequency. Adjusting the nozzle outlet flow rate changes the aerodynamic vertical stiffness of the air cushion support system, allowing the natural frequency of the air cushion system to avoid the current excitation frequency range of the plate 2.

[0039] Step S600: Decoupling of heat transfer coefficient and linear velocity compensation. The heat balance collaborative compensation module 45 calculates the comprehensive heat transfer capacity factor based on the adjustment of the opening degree of the airflow regulating valve array 22 in step S400 and the adjustment of the rotational speed and frequency of the circulating fan 21 in step S500. The rate of change.

[0040] Specifically, the comprehensive heat transfer capacity factor Calculate according to the following formula: ; in, This refers to the real-time speed and frequency of the circulating fan 21. It is the Reynolds number exponent (ranging from 0.6 to 0.8). For the air volume regulating valve at its real-time opening position The flow correction factor below.

[0041] To maintain a constant target metal temperature (PMT) during the annealing process, the system adjusts the linear velocity of the plate according to the law of conservation of energy. The linear velocity correction logic satisfies the following relationship: ,in, and These are the combined heat transfer capacity factors before and after adjustment, respectively. and These are the linear velocity commands before and after adjustment, respectively. Through the S600 procedure, the system compensates for the resulting thermodynamic parameter drift while implementing plate shape correction and chatter suppression, ensuring the stability of the annealing process.

[0042] The multidimensional flow field sensing array 30 consists of several pressure detection units distributed in various control temperature zones of the annealing furnace. For each independent control temperature zone, the pressure detection units are arranged in a horizontal array and vertically aligned layout to construct a flow field mapping over the entire width of the plate.

[0043] Specifically, inside the upper static pressure chamber 11 and the lower static pressure chamber 12, along the transverse width direction (i.e., the Y-axis direction) perpendicular to the running direction of the plate 2, there are equally spaced... Group pressure tapping point. The value of depends on the maximum width of the board 2 and the required board shape control resolution. It is typically set to at least three groups, corresponding to the left edge, center area, and right edge of the board 2 respectively. Preferably, The value ranges from 5 to 9 to cover more precise quarter-wave detection requirements. The coordinates of each pressure tap within the upper static pressure chamber 11 are... Coordinates of the corresponding pressure taps inside the lower static pressure chamber 12 Precise alignment is maintained in the vertical direction to form corresponding measurement groups. This geometric alignment ensures that the subsequent calculation of the upper and lower differential pressures can accurately reflect the fluid force state at the same micro-element area, eliminating calculation errors introduced by positional deviations.

[0044] Considering that the internal temperature of an air-cushion annealing furnace is typically between 500°C and 600°C, directly placing electronic sensors inside the furnace chamber presents engineering challenges. Therefore, this embodiment employs a remote pressure tapping structure. Each pressure tapping point extends through the furnace insulation layer to the outside of the furnace wall via a high-temperature resistant stainless steel pressure tapping tube, connecting to a high-frequency pressure transmitter installed in a low-temperature environment. The structural design of the pressure tapping tube must meet frequency response requirements. Preferably, the inner diameter of the pressure tapping tube is designed to be no less than 8 mm, and the length no more than 1.5 m, to ensure that dynamic changes in the furnace pressure are transmitted to the sensing diaphragm of the pressure transmitter with extremely low attenuation and phase delay. Specifically, the aerodynamic response bandwidth of the pressure tapping tube system must be greater than the highest frequency of sheet flutter of interest in this system.

[0045] Regarding the sampling strategy, in order to simultaneously meet the requirements of static plate shape detection (low-frequency / DC component) and dynamic flutter monitoring (high-frequency / AC component), the central control unit 40 performs high-frequency synchronous acquisition of all pressure sensors. Sampling frequency The sampling frequency is set strictly according to the Nyquist sampling theorem, which states that the sampling frequency must be greater than twice the highest frequency component of the measured signal. Given that the aeroelastic flutter frequency of aluminum alloy sheets in an air cushion furnace is typically distributed in the range of 10 Hz to 150 Hz, this embodiment sets the sampling frequency accordingly. satisfy: ,in, The estimated maximum flutter frequency of the board material. For safety margin. In specific engineering implementations, the sampling frequency... The frequency is set to no less than 500 Hz, preferably 1000 Hz.

[0046] Furthermore, to construct a real-time pressure snapshot of the entire field, all pressure sensors distributed in different lateral positions and different longitudinal temperature zones share the same high-precision clock source. The central control unit 40 triggers all sensors to latch data at the same microsecond level via a real-time Ethernet communication bus. This system-wide time synchronization ensures that subsequent differential calculations and modal analyses are based on the flow field state at the same physical moment, avoiding spurious phase differences caused by sampling timing misalignments, thereby guaranteeing the accuracy of reconstructing the instantaneous spatial morphology of the plate. For the selection of pressure transmitters, piezoresistive or capacitive differential pressure transmitters with a range of 0 to 5000 Pascals and an accuracy class of not less than 0.1 are used to capture small-amplitude airflow pulsation signals.

[0047] The signal acquisition and splitting module 41 receives a discrete sequence of raw pressure signals from the multidimensional flow field sensing array 30 and performs the following digital signal processing steps to decouple the hydrostatic and hydrodynamic characteristics.

[0048] Step S210: Preprocessing and outlier removal of the original signal. Before performing frequency domain separation, the system first preprocesses the original signal at each sampling time. The original pressure signal A validity check is performed. This step uses an amplitude limiting filter algorithm or a rate limiting algorithm to eliminate non-physical glitches caused by electromagnetic interference or momentary sensor malfunctions. Specifically, a threshold for the rate of change of physical pressure is set. Calculate the difference between the current sampled value and the valid value at the previous moment. If the absolute value of this difference is greater than the physical pressure change rate threshold... Then, linear interpolation or the effective value of the previous time step is used to replace the current sampled value, and the corrected original pressure signal sequence is output.

[0049] Step S220, Construction of the Static Baseline Channel (DC Channel). This step extracts the time-averaged pressure component from the broadband signal, reflecting the overall inflation level of the plenum chamber and the average position of the plates. The system uses a digital low-pass filter to process the corrected original pressure signal sequence. In order to suppress high-frequency noise while ensuring the tracking speed of plate shape change trend, this embodiment uses a first-order exponentially weighted moving average (EWMA) filtering algorithm.

[0050] Average pressure component The iterative calculation formula is as follows: ,in, For the current moment The time-averaged pressure component, For the previous moment The time-averaged pressure component, This represents the original pressure signal at the current moment. This is the smoothing coefficient.

[0051] Smoothing coefficient The value of is related to the set low-pass cutoff frequency. and system sampling frequency The correlation is as follows: ; In this embodiment, the low-pass cutoff frequency The value is set to be lower than the lowest natural frequency of the aluminum alloy sheet, preferably within the range of 1 Hz to 5 Hz. The data in this frequency band mainly characterize the quasi-static pressure changes caused by sheet geometry deformation (such as waviness) or fan airflow adjustments.

[0052] Step S230: Construction of the dynamic pulsation channel (AC channel). This step extracts the pulsating pressure component containing aerodynamic flutter feature information. To ensure the completeness of signal component decomposition, this step is achieved by subtracting the time-averaged pressure component calculated in step S220 from the original pressure signal. ; Alternatively, it can be obtained by configuring a separate digital bandpass filter. If a bandpass filter is used, its passband frequency range is set to cover the expected flutter frequency range of the material, for example, setting the lower cutoff frequency to 10 Hz and the upper cutoff frequency to 200 Hz. This pulsating pressure component The DC bias and extremely low frequency drift were filtered out, and only the high-frequency pressure fluctuation information caused by turbulent pulsation, acoustic resonance and fluid-structure interaction vibration of the plate was retained.

[0053] Step S240, signal synchronization alignment. Because the digital filtering process introduces a phase delay into the output signal, it causes the time-averaged pressure component... There is a lag compared to the original pressure signal. To ensure that the time-averaged pressure component is subsequently... With pulsating pressure component In conjunction with timing alignment during analysis, the system maintains a first-in-first-out (FIFO) alignment buffer in memory. The system calculates the number of delayed sampling points based on the group delay characteristics of the filter and performs corresponding clock cycle compensation for channels with smaller delays to ensure that the average pressure component output at the same time is equal. and pulsating pressure component Corresponding to the same physical sampling time.

[0054] The dynamic baseline reconstruction module 42 has a pre-built flow resistance network model that can describe the aerodynamic characteristics of the equipment. This model is based on the principle of fluid network impedance and abstracts the complex turbulent physical field into a lumped parameter network composed of pressure potential sources, fluid impedance, and nodal pressures.

[0055] In this flow resistance network model, the circulating fan 21 is defined as a controllable pressure potential source, and its total output pressure head is... Directly controlled by real-time rotational speed and frequency The delivery pipeline from the outlet of the circulating fan 21 to the interior of the plenum chamber is defined as a series branch, in which the airflow regulating valve array 22 constitutes the variable fluid resistance of this branch. The plenum chamber itself is defined as a capacitive node, and its internal pressure corresponds to the time-averaged pressure component measured by the pressure sensor 31. The flow path from the static pressure chamber through the nozzle to the surface of the plate is defined as the load branch, which includes a fixed flow resistance determined by the nozzle geometry and a variable gap flow resistance determined by the distance between the plate 2 and the nozzle.

[0056] Based on Bernoulli's equation in fluid mechanics and the law of conservation of energy in pipeline networks, the pressure balance state at the nodes of the plenum is determined by the pressure input on the supply side and the flow resistance characteristics on the exhaust side. For any control unit, the following steady-state pressure balance equation is established: ,in, The total pressure head generated by the circulating fan 21 The pressure drop caused by airflow passing through the air volume regulating valve. The volumetric flow rate through this branch is... This refers to the real-time opening position of the air volume regulating valve. This represents the time-averaged pressure component within the static pressure chamber.

[0057] To achieve plate-shaped decoupling, the model further analyzes the flow resistance characteristics of each component.

[0058] The total pressure output characteristics of the circulating fan 21 follow the law of fluid mechanical similarity: ,in, This is the characteristic coefficient of the wind turbine. air density, This refers to the real-time rotational speed and frequency of the circulating fan 21.

[0059] Pressure drop generated by air volume regulating valve It is directly proportional to the square of the flow rate and inversely proportional to the valve's flow coefficient. ,in, For the air volume regulating valve at its real-time opening position The flow coefficient function is obtained in advance through offline experimental calibration or computational fluid dynamics (CFD) simulation; This represents the cross-sectional area of ​​the pipe.

[0060] By combining the above physical relationships, an analytical expression for the theoretical pressure of the static pressure tank is constructed. Under standard no-load conditions, since plate 2 is not in or at its nominal position, the load impedance of the system is constant, and the volumetric flow rate is... It's about real-time rotational speed and frequency. and real-time opening position The dependent variable. By measuring the flow rate Substituting the dependency relationship into the pressure balance equation, the theoretical reference pressure inside the static pressure chamber is... Determined to be only related to real-time rotational speed and frequency and real-time opening position The relevant functions. This model quantifies the change in static pressure in the pressure chamber caused solely by equipment operation in the absence of plate-shaped disturbances, providing a benchmark for calculating the normalized pressure factor.

[0061] In order to obtain the theoretical reference pressure in real time during the production process The system pre-constructs a theoretical reference pressure map during the initialization or maintenance phase. This map describes the deterministic relationship between the system's input control variables and the static pressure chamber response pressure under standard conditions. The map construction process includes the following calibration steps.

[0062] Step S310: Establish a standard calibration environment. Calibration is performed with the annealing furnace empty or with a straight traction belt under standard tension inside the furnace. This state is defined as the zero plate shape interference state, where pressure fluctuations in the static pressure box are caused only by the operation of the circulating fan and air volume regulating valve, eliminating the interference of plate waviness on the flow field.

[0063] Step S320: Full-condition traversal scan. The central control unit controls the circulating fan to perform stepped adjustments within its rated operating range, while simultaneously controlling the airflow regulating valve array to perform stepped sweeps across its entire stroke range. Specifically, the real-time speed and frequency of the circulating fan are set. From minimum speed Up to maximum speed The speed increases in preset increments (e.g., 5 Hz). At each fixed rotational speed point, the real-time opening position of the airflow regulating valve is controlled. The system operates in preset steps (e.g., 5%) from 0.96 to 100%. It monitors the static pressure tank pressure in real time, and records each operating point when the fluctuation variance of the pressure signal is less than a preset stability threshold. Corresponding static pressure chamber time-averaged pressure measurement value .

[0064] Step S330: Characteristic function fitting and spectrum generation. Based on the discrete dataset acquired in step S320, the parameters of the fluid dynamics model are identified using the least squares method. According to the fluid similarity law, pressure is proportional to the square of the rotational speed. Therefore, the discrete data is fitted into an associated model of the following form to construct a continuous theoretical reference pressure model: ,in, This is the theoretical reference pressure under any working condition; The calibrated system flow coefficient characterizes the basic flow resistance characteristics of the pipe network; Real-time rotational speed and frequency; For information about the real-time opening position The dimensionless flow resistance characteristic function. For the dimensionless flow resistance characteristic function... A high-order polynomial is used for fitting and description: ,in, , , , The coefficients are the polynomial coefficients determined through regression analysis. This fitting process smooths out measurement noise and provides a continuous description of discrete data.

[0065] Step S340, Map Storage and Indexing. The calculated model parameters or discretized 3D data point matrix are stored in the storage unit of the central control device. In actual production operation, the dynamic baseline reconstruction module uses the currently sampled real-time rotational speed frequency... and real-time opening position The current theoretical reference pressure can be obtained in real time by directly substituting into the above formula or by looking up a table using a bilinear interpolation algorithm. .Should The value is used as a dynamic baseline for subsequent calculation of the normalized pressure factor.

[0066] The dynamic baseline reconstruction module 42 performs the following calculation steps in each operation cycle to generate dimensionless characteristic parameters that can characterize the true floating state of the plate, thereby achieving real-time elimination of interference with equipment operation.

[0067] Step S350: Timing synchronization and acquisition of state variables. The dynamic baseline reconstruction module reads the current time-averaged pressure component of the static pressure chamber from the signal acquisition and splitting module 41. Simultaneously, the real-time speed and frequency of the circulating fan are read from the air supply and cooling actuator 20 at the current moment. and the real-time opening position of the air volume regulating valve To ensure the timing synchronization of the calculated data, the system performs hysteresis compensation on the read actuator feedback signals. Specifically, considering the mechanical delays in fan speed response and valve action, the system uses the hysteresis time determined based on actuator step response testing to adjust the real-time speed frequency. and real-time opening position Perform time axis alignment to ensure that the state variables involved in the calculation correspond to the flow field state at the same physical moment.

[0068] Step S360: Real-time reconstruction of dynamic baseline pressure. Based on the real-time rotational speed frequency obtained in step S350. and real-time opening position The module calls the theoretical reference pressure map or correlation model pre-stored in the storage unit. Through bilinear interpolation or polynomial calculation, it obtains the theoretical reference pressure that the static pressure tank should possess under standard no-load conditions. The theoretical benchmark pressure This represents the baseline state of the flow field after eliminating the interference of sheet geometry deformation.

[0069] Step S370: Calculation of the normalized pressure factor. The system utilizes the time-averaged pressure component. Compared with theoretical benchmark pressure The ratio relationship is used to calculate the normalized pressure factor. The calculation formula is as follows: ,in, The normalized stress factor is a dimensionless scalar. The measured and filtered time-averaged pressure component; This is the theoretical baseline pressure reconstructed based on the current equipment status. During the calculation process, the system is configured with denominator protection logic; if the theoretical baseline pressure... If the pressure is less than the preset minimum maintenance pressure threshold, the normalized pressure factor will be applied. Set it to 1 or keep the value from the previous time step to prevent numerical calculations from diverging.

[0070] Step S380: Physical meaning mapping and state decoupling. The normalized pressure factor obtained through the above calculations... Directly reflects the actual suspension gap of the board. Relative to the nominal design clearance The degree of deviation. According to the principles of air cushion mechanics, the static pressure chamber pressure and the suspension gap exhibit a monotonically inverse relationship.

[0071] when When the measured pressure is equal to the theoretical reference pressure, plate 2 is at the nominal suspension height and there is no obvious deformation; when When the measured pressure is higher than the theoretical reference pressure, the corresponding outflow resistance increases, indicating a localized suspension gap in the plate. Less than the nominal design gap Plate 2 is close to the nozzle; when When the measured pressure is lower than the theoretical reference pressure, the corresponding outflow resistance decreases, meaning there is a localized suspension gap in the plate. Greater than the nominal design gap Keep plate 2 away from the nozzle.

[0072] By introducing a normalized pressure factor This embodiment achieves decoupling of the measurement signal. That is, regardless of fluctuations in the circulating fan speed or adjustments to the valve opening, as long as the geometric position of plate 2 relative to the nozzle remains unchanged, The value will always remain constant. This allows subsequent plate shape control algorithms to be based on... The system independently identifies plate shape and wave characteristics without requiring recalibration of the pressure threshold for each airflow adjustment, thus improving the robustness and response bandwidth of the control system.

[0073] The plate shape feature analysis module 43 reconstructs the instantaneous geometry of plate 2 in the annealing furnace based on the spatial distribution of the normalized pressure factor in the transverse and longitudinal directions. This process transforms the discrete pressure signal into a continuous plate shape feature curve, providing quantitative geometric feedback for subsequent thermal stress control.

[0074] Step S410: Calculate the transverse plate shape characteristic index distribution. The plate shape characteristic analysis module 43, taking each independent control temperature zone as a unit, synchronously reads the normalized pressure factor corresponding to all paired pressure detection units on the same transverse cross section. For the transverse coordinate... The first For each detection point, the module calculates the plate shape characteristic index according to the following formula. : ,in, Located on the horizontal axis Normalized pressure factor of the lower static pressure chamber at the location, Located on the horizontal axis Normalized pressure factor of the upper static pressure box.

[0075] The plate shape characteristic index The influence of the system common-mode pressure is eliminated by differential ratio calculation, and its value range is usually between -1 and 1. The symbol represents the direction in which plate 2 deviates from the horizontal plane of the furnace center (positive value represents downward displacement, negative value represents upward displacement), and its absolute value represents the relative magnitude of the vertical displacement.

[0076] Step S420: Construct the full-width plate shape profile curve. Since the sensors are physically discretely arranged, to obtain a continuous plate shape description along the plate width direction, the module uses a cubic spline interpolation algorithm to interpolate the discrete plate shape feature index point set. Perform a smooth fit to generate a continuous transverse plate-shaped distribution function. Based on this, the pre-calibrated position sensitivity coefficient is used. The dimensionless distribution function is mapped to the vertical displacement curve of plate 2 relative to the nominal center plane. : ,in, The meaning is the nominal design clearance. This is the linearized gain coefficient related to the nozzle geometry. The vertical displacement curve is determined by the reciprocal of the tangent slope at the nominal operating point of the pressure-height characteristic curve obtained from offline static suspension experiments. The wave pattern of plate 2 on the current cross section is quantitatively described.

[0077] Step S430: Parametric decomposition of plate shape defect modes. To identify specific plate shape defect types (such as center waves, edge waves, or quarter waves), an orthogonal polynomial decomposition method is used to decompose the vertical displacement curves. Feature extraction is performed. This embodiment uses Legendre polynomials as basis functions. Decomposed into components of different orders: ,in, The horizontal coordinate variable is normalized to the interval [-1, 1]. for Legendre polynomials; , , , These are the corresponding amplitude coefficients.

[0078] Each order coefficient has a clear physical meaning: The coefficient characterizes the overall vertical floating deviation of plate 2; The coefficient characterizes the lateral tilt of plate 2; The coefficients characterize the components of the quadratic parabola, i.e., the characteristics of the intermediate and lateral waves. According to the properties of the Legendre polynomial, when... When the symbol corresponds to a shape that is high in the middle and low on both sides, it is judged as a mid-wave characteristic; when When the symbol corresponds to a shape that is low in the middle and high on both sides, it is judged as a side wave feature; The coefficients characterize the fourth-order higher-order components and are used to identify complex wave patterns or quarter-wave features.

[0079] Step S440, 3D topology synthesis. The system calculates the vertical displacement curves of multiple control temperature zones arranged along the running direction (X-axis) of plate 2. Vertical splicing is performed. Using a spatiotemporal mapping algorithm, the real-time linear velocity of board 2 is monitored. By performing time integration, the longitudinal position of the current sampling section in the plate coordinate system is calculated. The cross-sectional profiles collected at different times are mapped onto the same physical coordinate system to construct a three-dimensional topological map reflecting the surface flatness of the entire roll of plate 2. This three-dimensional map can characterize the longitudinal evolution of plate defects, such as identifying periodic warping or continuous longitudinal warping, thus providing full-dimensional geometric boundary conditions for the accurate calculation of the subsequent inverse thermal stress field.

[0080] Based on the three-dimensional morphology data of the plate output by the plate shape feature analysis module 43, the internal residual stress distribution that causes plate shape defects is deduced in reverse according to the principle of thermoelasticity, and the transverse temperature gradient compensation required to eliminate the residual stress is calculated.

[0081] Step S450: Calculation of the longitudinal elongation difference of the sheet material. The geometric cause of the sheet shape defect is the inconsistent longitudinal length of the fiber strips in the width direction of sheet 2. First, based on the vertical displacement curve obtained in step S420... Calculate the longitudinal relative elongation distribution of plate 2 along the width direction. Based on the small deflection plate and shell theory, and assuming that plate 2 exhibits a sinusoidal buckling mode between adjacent support points, the longitudinal relative elongation distribution is... With vertical displacement curve The first derivative exhibits the following geometrically nonlinear relationship: ,in, The span distance between two adjacent support rollers or air cushion units; For board 2, the horizontal coordinate The amplitude of the vertical displacement curve at that location; Characterized by the horizontal coordinate The extra elongation of the longitudinal fibers relative to an ideal straight state.

[0082] Step S460: Inversion of the equivalent residual stress field. The system, based on Hooke's law, calculates the longitudinal relative elongation distribution. Converted to the equivalent residual stress distribution inside plate 2 This stress characterizes the internal stress difference that causes buckling deformation in plate 2: ,in, For aluminum alloy materials at the current annealing temperature The elastic modulus is provided by the system's preset material property database; This represents the average elongation over the entire width of the plate. In the notation system defined in this embodiment, when... When this occurs, it indicates that the fiber at that location is relatively too long, resulting in compressive stress accumulation, corresponding to the peak or trough of the plate shape; when When the fiber is too short at that position, there is tensile stress, corresponding to a straight or taut position.

[0083] Step S470: Solving for the target transverse temperature difference distribution. To eliminate plate shape defects, the system constructs an equivalent residual stress distribution. Conversely, a compensating thermal stress field is generated. Based on the principle of linear expansion, the target transverse temperature difference distribution required to offset this elongation difference is calculated. : ,in, For aluminum alloy materials at the current annealing temperature The linear expansion coefficient is as follows; the negative sign in the formula indicates that the compensation logic is: for excessive elongation ( Apply cold contraction (negative temperature difference) to areas with large elongation, and apply thermal expansion (positive temperature difference) to areas with small elongation.

[0084] Step S480: Discretized temperature control command generation. Calculated... It is a theoretical curve that varies continuously along the width of the plate. To accommodate physically discrete heating or cooling actuators, the system... Perform region discretization mapping. Assume the annealing furnace is divided horizontally into... A separate temperature control zone, for the first Each temperature control zone (covering the horizontal range) to Its local temperature correction setting value The result was obtained through weighted integration: ,in, The weighting function, which is a pre-defined Gaussian or trapezoidal distribution function, is used to characterize the thermal influence range distribution of the nozzle corresponding to the zone across the lateral width. The system ultimately calculates the local temperature correction setpoints for each zone. The final control command is generated and sent to the multi-zone heater or segmented cooling spray beam by superimposing the basic process temperature setpoint, thereby constructing the required non-uniform temperature field on the physical width of the plate and realizing the thermal stress correction of plate shape defects.

[0085] In order to adjust the discretized local temperature correction setpoint calculated in step S480 This is transformed into specific mechanical actions, and the air volume regulation controller in the air supply and cooling actuator 20 executes the opening degree calculation process based on the inverse model of fluid characteristics.

[0086] Step S490: Calculation of the target flow rate of the cooling medium. The system first establishes a mapping relationship from temperature correction requirements to cooling medium flow rate requirements. Based on the principle of convection heat transfer, the heat transfer coefficient of the plate surface is non-linearly positively correlated with the cooling airflow velocity (i.e., volumetric flow rate). Given the current baseline flow rate set by the basic process... Under the premise that, for the first Each temperature-controlled zone calculates the temperature correction value. Required target volume flow rate The calculation employs a linearized incremental model in the following form: ,in, For the first Target volumetric flow rate for each temperature-controlled zone; The global base flow rate is preset based on the sheet thickness and production speed; The target temperature for the process; The heat flux sensitivity coefficient, determined through offline thermal calibration experiments, characterizes the gain of the effect of the rate of change of flow rate on the rate of change of heat transfer efficiency under the nozzle structure. When the value is negative (requiring stronger cooling), the target flow rate Increase; conversely, decrease.

[0087] Step S492: Inverse mapping calculation of valve opening. Since industrial airflow regulating valves typically have nonlinear flow characteristics, directly using a linear control algorithm will lead to inconsistent response sensitivity at different opening degrees. Therefore, the controller has a pre-built inverse model of the valve flow characteristics. The inverse model is the flow coefficient function of the air volume regulating valve in the flow resistance network model. The mathematical inverse operation. The system is based on the target volumetric flow rate. Calculate the corresponding target opening command : ,in, It is the inverse function of the flow coefficient function, which can be realized by looking up a table or by inverse transformation of a higher-order polynomial; This refers to the real-time pressure difference across the airflow regulating valve. This value is obtained in real-time by pressure sensors installed before and after the valve, or based on a flow resistance network model using real-time rotational speed and frequency. The estimation was obtained. By introducing this inverse model, the system achieves linearization of flow control, eliminating the influence of valve nonlinear characteristics on the control loop gain.

[0088] Step S494: Dead zone and amplitude limiting processing of the actuator. The calculated target opening command is then processed. Before being sent to the servo motor, the signal needs to be conditioned to avoid mechanical vibration and excessive wear. The system compares the current opening position with the real-time opening position acquired at the current moment. With target opening command The difference. Only when the absolute value of this difference exceeds the preset mechanical dead zone threshold. The driver instructions are only updated at that time.

[0089] At the same time, the system performs saturation limiting processing on the output commands: ;in, and These are the physical minimum and maximum opening limits of the airflow regulating valve, respectively. After processing... The signal is converted into a standard 4-20mA current signal or fieldbus command to drive the electric actuator of the corresponding zone, thereby changing the local cooling airflow sprayed onto the plate 2 and realizing the lateral temperature difference distribution planned in step S470.

[0090] To prevent self-excited vibration of plate 2 within the air cushion furnace, the flutter monitoring and variable stiffness module 44 monitors the dynamic stability of the flow field in real time. This module not only focuses on the static mean pressure but also emphasizes the analysis of high-frequency dynamic components in the pressure signal to identify precursory features of aerodynamic instability.

[0091] Step S510, Extraction of dynamic pressure components. The system uses a high-frequency pressure sensor to continuously sample the pressure in the static pressure tank at a frequency that satisfies the sampling theorem, obtaining the original pressure sequence. To separate the dynamic components containing vibration information, the flutter monitoring and variable stiffness module 44 uses a high-pass filtering algorithm or a moving average removal method to process the original signal. Specifically, the dynamic pressure component... The calculation formula is: ,in, The width of the sliding time window should be greater than three times the period of the lowest natural frequency of the plate material to ensure complete filtration of the average pressure component. And low-frequency drift, only retaining AC signals that reflect flow field turbulence and plate vibration.

[0092] Step S520, Power Spectral Density Analysis. To extract periodic vibration features from the time-domain cluttered signal, the module analyzes the dynamic pressure component. Perform a Fast Fourier Transform. Before the transform, apply a Hanning window to the data to reduce spectral leakage. Then, calculate the one-sided power spectral density function. To characterize the distribution of vibrational energy in the frequency domain: ,in, For frequency variables; This represents the number of sampling points; The sampling frequency; It is a discretized dynamic pressure sequence; For frequency variables The corresponding discrete frequency domain index value satisfies Relationship; The unit is the imaginary unit. The specific implementation algorithm for the Fast Fourier Transform and its power spectrum calculation is well-known in this field and will not be elaborated upon here.

[0093] Step S530: Determination of the dominant flutter frequency. This is done after obtaining the power spectral density. Then, the system monitors the preset frequency band. The system searches for spectral peaks within a specified range. This monitoring frequency band is typically set to the first three natural frequencies of plate 2. The system identifies the frequency point corresponding to the maximum spectral density and defines it as the dominant flutter frequency. And record the peak energy density at that frequency. .

[0094] Step S540: Calculation and determination of flutter intensity index. To distinguish between random broadband turbulent disturbances and destructive narrowband self-excited flutter, the system calculates the dimensionless flutter intensity index. This index is defined as the percentage of peak energy in total vibrational energy, or in the form of signal-to-noise ratio: , where the denominator represents the total vibration energy within the monitoring frequency band.

[0095] Based on this calculation result, the system executes the following state determination logic. The first preset threshold is described below. Second preset threshold These are empirical values ​​determined based on offline flutter boundary testing or fluid-structure interaction simulation. like Less than the first preset threshold The state is determined to be a stable suspension state, at which point the vibration is mainly caused by broadband random noise. like At the first preset threshold With the second preset threshold Between these values, it is determined to be a flutter latent state, indicating that frequency energy has accumulated in the flow field. Although it has not yet dissipated, it is necessary to enter the early warning mode. like Greater than the second preset threshold The condition is determined to be a flutter divergence state, indicating that plate 2 has undergone self-excited vibration in a single mode, and a variable stiffness vibration suppression mechanism must be triggered immediately. Through this identification method based on frequency domain energy concentration, this embodiment can identify the aerodynamic instability trend before visible mechanical resonance occurs in plate 2.

[0096] When the flutter monitoring module determines that the system is in a flutter latent state or a flutter divergent state, the flutter monitoring and variable stiffness module 44 activates the variable stiffness adjustment strategy. By changing the stiffness of the pneumatic spring of the air cushion system, it actively adjusts the natural frequency of the plate air cushion coupling system, thereby changing the resonance conditions of fluid-structure interaction.

[0097] Step S550: Real-time estimation of current aerodynamic stiffness. Aerodynamic stiffness Defined as the sensitivity of the air cushion support force to changes in the air film thickness. Since direct measurement of stiffness is difficult, the module is based on a quasi-static assumption, using the current static pressure chamber pressure and the nominal design clearance of the air cushion system for analytical calculation. For the... Each control unit has aerodynamic stiffness Estimate based on the following formula: ,in, This is the current measured pressure value of the static pressure chamber; The effective bearing area of ​​the nozzle; The nominal design clearance for the air cushion system; The nozzle geometry factor, a dimensionless parameter, is determined by the nozzle flow coefficient and the nozzle distribution density per unit area, and is calibrated through offline bench tests. This formula shows that the air cushion stiffness is directly proportional to the static pressure chamber pressure and inversely proportional to the suspension height.

[0098] Step S560, Target Stiffness and Frequency Avoidance Path Planning. This is to suppress the dominant flutter frequency determined in step S530. The system needs to move the natural frequency of the air cushion out of the resonant region. The system calculations demonstrate how to make the natural frequency avoid... Target stiffness required for at least ±15% safety margin : ,in, This is the frequency shift coefficient, used to set the safety margin for frequency avoidance; The system's natural frequency is currently estimated and is calculated using the following formula: ,in, The equivalent vibration mass of the plate area supported by a single control unit is calculated from the density and thickness of plate 2 and the coverage area of ​​the control unit. If Approaching the current inherent frequency, the system preferentially selects a positive frequency shift ( ), that is, increase stiffness; if the system is already in a high-voltage limiting state, then choose a negative frequency shift ( .

[0099] Step S570, dual-modal adjustment command calculation. To achieve the target stiffness... The system offers two execution modes: pressure regulation and height regulation, which can be selected based on response speed requirements and actuator stroke limitations.

[0100] Step S571, rapid pressure regulation mode. When the flutter intensity index When in a flutter divergence state, the system prioritizes adjusting the frequency of the faster-responding air supply fan. The nominal design clearance of the air cushion system is maintained. Keeping constant, calculate the required target static pressure chamber pressure. : The system will This is converted into speed control commands for the fan inverter. The response time for this method is typically in the second range, making it suitable for rapid suppression.

[0101] Step S572, wide-range height adjustment mode. When the pressure adjustment is calculated... When the system's allowable safe pressure range is exceeded, or when the system is in a flutter latent state and requires continuous adjustment, the system activates the nozzle lifting mechanism to maintain the static pressure chamber pressure. The stiffness is adjusted by changing the physical air gap, keeping it constant. The required target suspension height is then calculated. : The system drives servo screw mechanisms installed on both sides of the static pressure box to synchronously adjust the vertical position of the upper and lower static pressure boxes, thus adjusting them to the desired position. .

[0102] Step S580: Adjust dead zone and hysteresis reset. To prevent the system from repeatedly oscillating and adjusting near the critical frequency, hysteresis logic is introduced into the controller. When the chatter intensity index... Reduce to the first preset threshold The following will continue for the set time. Then, the system gradually reverts the stiffness parameters back to the initial settings determined by the thermal process. If during the revert process... If the stiffness increases again, the current parameters are immediately locked or step S560 is re-executed. This variable stiffness strategy, without changing the sheet tension, blocks the transmission of vibration energy in the fluid-structure interaction system through aerodynamic means.

[0103] To achieve high-precision temperature control, the thermal balance collaborative compensation module 45 needs to accurately determine the actual heat transfer capacity under the current furnace conditions. Since the heat transfer coefficient is affected by various time-varying factors such as fan speed, nozzle blockage, and surface emissivity of plate 2, the system cannot rely solely on theoretical design values ​​but needs to establish a real-time observation model based on energy balance.

[0104] Step S610: Real-time calculation of the heat absorption rate of the plate. The system collects the inlet temperature of plate 2 using high-precision infrared thermometers or contact thermocouples installed at the inlet and outlet of the annealing furnace. and outlet temperature Combined with the current production line speed Based on the specifications and parameters of the sheet material, calculate the actual heat absorption power of sheet material 2 when it passes through the heating section. : ,in, This refers to the density of the aluminum alloy sheet. The width of the sheet material; The thickness of the sheet material; This represents the current production line speed. This is the specific heat capacity of the material at the average temperature, a value obtained by querying the system's pre-set material property database. This formula characterizes the total sensible heat carried away by plate 2 as it moves along the production line per unit time.

[0105] Step S620: Analysis of the effective convective heat transfer temperature difference. In the air-cushion continuous annealing furnace, the heat transfer process is mainly dominated by forced convection between the hot air ejected from the nozzles and the moving plate 2. The system collects the temperature of the hot air inside the static pressure chamber. The effective heat transfer driving temperature difference of the heating section was calculated based on the logarithmic mean temperature difference (LMTD) theory. : ; like and If the difference is less than a preset small amount (e.g., 1°C), to avoid instability in numerical calculations, the system automatically switches to arithmetic mean temperature difference calculation, i.e. .

[0106] Step S630: Observation of the heat transfer coefficient under current operating conditions. According to Newton's law of cooling, the system utilizes the calculated heat absorption power... and effective temperature difference Calculate the observed heat transfer coefficient at the current moment. : ,in, This represents the effective heating length of the temperature control zone; coefficient 2 indicates that the upper and lower surfaces of plate 2 participate in heat exchange simultaneously. This is a radiation correction factor used to eliminate the influence of radiative heat transfer from the furnace wall and extract the heat transfer characteristics of the pure convection portion. This factor is usually close to 1.0, and the specific value is calibrated by the offline thermal model based on the furnace emissivity and temperature level.

[0107] Step S640, decoupled calculation of heat transfer performance factor. Observe the heat transfer coefficient. This is a comprehensive quantity that couples the inherent performance of the equipment with the current operating conditions. To predict the heat transfer capacity under different wind speeds, the system needs to decouple the equipment performance factor from the dynamic state. Based on the power-law relationship between the Nusselt number (Nu) and the Reynolds number (Re), and assuming that the wind speed is proportional to the square root of the plenum pressure, the following decoupling model is established: ,in, This is the current measured pressure value of the static pressure chamber; The heat exchange pressure correlation index typically ranges from 0.3 to 0.4, depending on the specific geometry of the nozzle and determined through offline thermal experiments. Defined as the comprehensive heat transfer performance factor.

[0108] Through the above calculations, the system obtains a performance index independent of the current wind turbine speed in real time. This indicator characterizes the heat exchange efficiency of the furnace (e.g., whether there is nozzle blockage, flow channel ash accumulation, etc.), and its value drifts over time at a much lower rate than the rate of pressure change. In subsequent velocity feedforward compensation, the system utilizes the updated... By combining the predicted pressure setpoint, the temperature response after a change in production line speed can be predicted without waiting for the temperature deviation to occur.

[0109] To address the significant temperature lag issue that occurs during production speed adjustment in air-cushion continuous annealing furnaces, the heat balance collaborative compensation module 45 utilizes the comprehensive heat transfer performance factor updated in real-time during step S640. A velocity-pressure linkage control mechanism based on a reverse thermodynamic model was constructed.

[0110] Step S650, Energy demand prediction for the speed change process. When the main control system issues a new production line speed command... At that time, the system predicts the steady-state heat load required to maintain the current process target temperature based on this command value. Assume the inlet temperature of the plate... and process target temperature Keeping the speed constant, the system calculates the expected heat absorption power after the speed change. : ; in, The target production line speed to be implemented; The target temperature for the process; The inlet temperature is denoted as . The procedure establishes the heat flow required to ensure that plate 2 reaches the target temperature under the new steady-state velocity.

[0111] Step S660: Inverse solution of the feedforward pressure command. Based on the decoupled model established in step S640, the system transforms the thermal demand into specific actuator actions. The system first calculates the... Required theoretical heat transfer coefficient Then, combined with the latest comprehensive heat transfer performance factors The corresponding feedforward pressure setpoint is solved in reverse. : ; in, This is the effective heating length of the temperature control section; To determine the effective heat exchange temperature difference under the operating conditions, this value is based on the target air temperature set by the process. Calculated based on the average temperature of plate 2; This is the heat exchange pressure correlation index, and its value is completely consistent with the value determined in step S640. This formula utilizes a real-time calibrated comprehensive heat exchange performance factor. Directly calculate when the velocity becomes At that time, the static pressure box pressure needs to be adjusted to what level to maintain thermal balance, thereby compensating for model deviations caused by equipment characteristic drift.

[0112] Step S670: Synthesis and synchronization of control signals. To balance dynamic response and steady-state accuracy, the system will synthesize the calculated feedforward pressure setpoint. Feedback output value of a conventional temperature PID controller The commands are superimposed to generate the final static pressure control command for the pressure chamber. : ; in, The feedforward weighting coefficient ranges from 0.8 to 1.0. A dynamic synchronization mechanism is introduced during the superposition process. Because the response characteristics of the wind turbine inverter differ from the mechanical response characteristics of the production line speed changes, the system... A discretized first-order hysteresis filter is applied to the signal to ensure that pressure changes and velocity changes remain synchronized on the time axis. The filter formula is as follows: ,in, These are the filter coefficients. To control the cycle, The time constant is used to match the acceleration time of the wind turbine inverter.

[0113] Step S680: Safety limiting and anti-saturation processing. Before outputting the final instruction, the system performs verification. Is it within the safe range allowed by the process? If the calculated... Exceeded the system's maximum allowable pressure The system will output the pressure value after limiting, and simultaneously calculate the maximum allowable production line speed based on the pressure gap. Send a speed limit request to the main control PLC to prevent insufficient annealing temperature due to forced speed increase. The calculation formula is: ; Through this feedforward compensation mechanism, the system can reduce the temperature fluctuation of the board 2 caused by large fluctuations in production line speed to less than 20% of that in the pure feedback control mode.

Claims

1. A continuous annealing system for wide-width aluminum alloy sheets, characterized in that, It includes the air cushion annealing furnace body (10), air supply and cooling actuator (20), multi-dimensional flow field sensing array (30) and central control device (40); The air cushion annealing furnace body (10) is provided with a furnace passage along the running direction of the plate (2), and an upper static pressure box (11) and a lower static pressure box (12) are provided on the upper and lower sides of the furnace passage. The air supply and cooling actuator (20) includes a circulating fan (21) and an air volume regulating valve array (22). The multidimensional flow field sensing array (30) includes a pressure sensor (31); The central control device (40) is used for: The signal acquisition and splitting module (41) decomposes the raw pressure signal acquired by the pressure sensor (31) into a time-averaged pressure component and a pulsating pressure component. The dynamic baseline reconstruction module (42) calculates the theoretical reference pressure based on the rotational speed of the circulating fan (21) and the opening of the air volume regulating valve array (22), and generates a normalized pressure factor in combination with the time-averaged pressure component; The plate shape feature analysis module (43) adjusts the air volume regulating valve array (22) according to the normalized pressure factor to control the plate shape; The flutter monitoring and variable stiffness module (44) adjusts the rotation speed of the circulating fan (21) according to the pulsating pressure component to suppress the flutter of the plate (2).

2. The wide-width aluminum alloy sheet continuous annealing system according to claim 1, characterized in that, The multidimensional flow field sensing array (30) follows a layout of horizontal arraying and vertical alignment; The circulating fan (21) is connected to the upper static pressure box (11) and the lower static pressure box (12), and the air volume regulating valve array (22) is used to control the lateral air intake of the upper static pressure box (11) and the lower static pressure box (12); The upper static pressure box (11) and the lower static pressure box (12) are equipped with pressure taps at equal intervals along the transverse width direction. The pressure taps in the upper static pressure box (11) are aligned with the corresponding pressure taps in the lower static pressure box (12) in the vertical direction. The pressure sensor (31) is connected to the upper static pressure box (11) and the lower static pressure box (12) through a high-temperature pressure-sensing tube. All pressure sensors (31) perform high-frequency synchronous acquisition based on the same clock source.

3. The wide-width aluminum alloy sheet continuous annealing system according to claim 1, characterized in that, The signal acquisition and splitting module (41) performs dual-channel digital filtering: The time-averaged pressure component reflecting the hydrostatic characteristics is extracted from the original pressure signal using a low-pass filtering algorithm. The pulsating pressure component reflecting the fluid dynamics characteristics is extracted from the original pressure signal using a high-pass filtering algorithm or a band-pass filtering algorithm. During processing, the system performs time alignment compensation on the filtered signal to ensure that the average pressure component and the pulsating pressure component output at the same time correspond to the same physical sampling time.

4. The wide-width aluminum alloy sheet continuous annealing system according to claim 1, characterized in that, The dynamic baseline reconstruction module (42) stores a flow resistance network model, which is used to eliminate the interference of the air supply and cooling actuator (20) on the measurement data in the plate shape decoupling calculation; The flow resistance network model defines the circulating fan (21) as a pressure potential source and the air volume regulating valve array (22) as a variable fluid resistance. The theoretical reference pressure is calculated based on a pre-calibrated equipment fluid characteristic spectrum. The calculation is based on the current rotational speed of the circulating fan (21) and the opening of the air volume regulating valve array (22) to determine the pressure values ​​that the upper static pressure box (11) and the lower static pressure box (12) should have under standard no-load conditions.

5. A continuous annealing system for wide-width aluminum alloy sheets according to claim 4, characterized in that, The dynamic baseline reconstruction module (42) generates the normalized pressure factor by calculating the ratio of the measured time-averaged pressure component to the calculated theoretical reference pressure. The normalized pressure factor is used to characterize the degree of deviation of the actual suspension gap of the plate (2) from the nominal design gap. The value of the normalized pressure factor is independent of the speed adjustment action of the circulating fan (21) and the opening adjustment action of the air volume regulating valve array (22).

6. The wide-width aluminum alloy sheet continuous annealing system according to claim 5, characterized in that, The plate shape feature parsing module (43) is used for: The plate shape characteristic index is calculated based on the difference ratio of the normalized pressure factors corresponding to the upper static pressure box (11) and the lower static pressure box (12) on the same transverse section, and the transverse plate shape profile of the plate (2) is reconstructed based on the plate shape characteristic index. Based on the principle of thermoelasticity, the lateral temperature difference distribution required to eliminate geometric deformation in the lateral plate profile is calculated, and the lateral temperature difference distribution is mapped to the target opening command of the air volume regulating valve array (22) in the cooling section. Corrective thermal stress is generated by establishing a non-uniform temperature field on the lateral width of the plate (2).

7. The wide-width aluminum alloy sheet continuous annealing system according to claim 1, characterized in that, The flutter monitoring and variable stiffness module (44) is used for: The power spectral density is obtained by performing spectral analysis on the pulsating pressure component, and the flutter intensity index is calculated. When the flutter intensity index indicates that the system is in a flutter latent state or a flutter divergent state, a speed and frequency adjustment command is generated for the circulating fan (21); The speed and frequency adjustment command is used to change the aerodynamic vertical stiffness of the air cushion support system so that the natural frequency of the air cushion support system avoids the current dominant flutter frequency of the plate (2).

8. The wide-width aluminum alloy sheet continuous annealing system according to claim 1, characterized in that, The central control device (40) further includes a thermal balance coordination compensation module (45), which is used for: For the adjustment of the opening of the air volume regulating valve array (22) and the adjustment of the speed of the circulating fan (21), the rate of change of the comprehensive heat exchange capacity factor is calculated based on the real-time speed frequency of the circulating fan (21) and the flow correction coefficient of the air volume regulating valve array (22). Based on the law of conservation of energy, a correction command is generated for the running linear velocity of the plate (2) according to the rate of change of the comprehensive heat transfer capacity factor, so as to maintain the target metal temperature of the plate (2) constant during the annealing process.

9. A continuous annealing system for wide-width aluminum alloy sheets according to claim 4, characterized in that, The fluid characteristic profile of the equipment was generated through a full-condition traversal scanning calibration. The calibration process includes: Under standard no-load conditions, the circulating fan (21) is controlled to perform step-by-step adjustment within the rated working range, while the air volume regulating valve array (22) is controlled to perform step-by-step sweeping within the full stroke range. The average pressure measurement values ​​of the upper static pressure box (11) and the lower static pressure box (12) corresponding to each working point are recorded, and the function relationship between the rotational speed frequency and the opening position is obtained by fitting using the least squares method.

10. A continuous annealing system for wide-width aluminum alloy sheets according to claim 6, characterized in that, When the air supply and cooling actuator (20) executes the target opening command, it uses a preset valve flow characteristic inverse model to calculate the opening. The valve flow characteristic inverse model calculates the target opening command after eliminating the influence of the valve's nonlinear flow characteristics based on the target volume flow rate and the real-time or estimated pressure difference at both ends of the air volume regulating valve array (22), and performs dead zone verification and amplitude limiting processing on the output command.

Images