Intelligent control system for short fiber reinforced fiberboard hot press forming process

Abstract

The application discloses a kind of short fiber reinforced fiberboard hot-press forming process intelligent control system, it is related to intelligent control technical field, including: under the constraint of initial forming reference parameter, stress wave pulse is captured by monitoring window that is dynamically opened and closed with pressure load, and acoustic emission signal is projected to time-frequency collaborative domain, and fiber damage variable is extracted by the intensity of energy distribution stripping;Disturbance module, fiber damage variable is mapped as the structure disturbance parameter of short fiber reinforced plate blank internal structure evolution state, and coupling analysis is carried out with the pressure change information detected in forming process, and structure disturbance estimation parameter is formed;Adjusting module, structure disturbance estimation parameter is mapped as synchronous out-of-phase compensation component, and transient driving correction instruction that is offset with disturbance energy is injected in pressure control loop, and output forming state data stream;Reduce finished product delamination burst and warping risk and improve short fiber reinforced fiberboard comprehensive mechanical property.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, and in particular to an intelligent control system for the hot pressing process of short fiber reinforced fiberboard. Background Technology

[0002] In the field of engineered wood products and composite material processing, the hot pressing process of short fiber reinforced fiberboard is a complex physicochemical process involving rheology, heat transfer, and material curing kinetics. Traditional control technologies mainly rely on preset pressure-temperature-time curves for open-loop or semi-closed-loop control, using pressure and displacement sensors to monitor changes in macroscopic physical parameters and thus adjust the power output of the hydraulic drive system. With the increasing demand for precision manufacturing, some advanced solutions have begun to introduce feedback control logic, adjusting the linear gain of the actuator based on the displacement creep rate or pressure feedback deviation, in order to improve the density uniformity and mechanical property consistency of the slab.

[0003] However, existing hot-pressing control technologies still have limitations in responding to the real-time dynamic changes in the internal structure of the slab. Due to the complex fiber reorganization, matrix flow, and microstructural damage that occur during hot pressing of short-fiber slabs, the macroscopic pressure signal often contains a large amount of mechanical vibration noise and system hysteresis characteristics. This makes it difficult for existing feedback mechanisms to accurately isolate and identify the real disturbances caused by microscopic fiber skeleton damage or evolution from the chaotic macroscopic pressure fluctuations. This information gap between macroscopic monitoring and microscopic evolution prevents the system from performing timely phase offset compensation for the transient energy release of the internal structure. Consequently, the transient stress generated by local structural damage inside the slab cannot be effectively relaxed, easily forming microcracks or stress concentration zones inside the finished product. In severe cases, this can induce quality defects such as delamination and bursting of the fiberboard, excessive thickness expansion rate, and macroscopic warping deformation. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an intelligent control system for the hot pressing process of short fiber reinforced fiberboard to solve the problem of difficulty in accurately identifying and compensating for pressure disturbances caused by the evolution of the internal microstructure of the slab, which in turn leads to internal stress defects in the slab.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides an intelligent control system for the hot pressing process of short fiber reinforced fiberboard, comprising: The acquisition module collects multidimensional physical property parameters of the short fiber reinforced slab to be processed, and performs environmental noise suppression on the acoustic emission acquisition channel to obtain the initial reference parameters for molding. The damage sensing module, under the constraint of the initial reference parameters of molding, captures stress wave pulses through a monitoring window that dynamically opens and closes with pressure load, and projects the acoustic emission signal into the time-frequency co-domain. Through intensity stripping of energy distribution, it extracts fiber damage variables. The perturbation module maps fiber damage variables to structural perturbation parameters of the internal structural evolution state of short fiber reinforced slabs, and couples them with pressure change information detected during the forming process to form structural perturbation estimation parameters. The adjustment module maps the structural disturbance estimation parameters into synchronous out-of-phase compensation components, injects transient drive correction commands that cancel out the disturbance energy into the pressure control loop, and outputs the forming state data stream. The pressure relief module determines the curing completion status based on the molding state data stream and drives the hot pressing actuator to perform nonlinear stress relaxation pressure reduction control based on the structural disturbance estimation parameters.

[0007] Preferably, the method for obtaining the initial molding reference parameters includes: Information on fiber content, thickness, initial density, and moisture content of the short fiber reinforced slab to be processed was collected and the parameters were normalized to obtain multidimensional physical property parameters of the slab. The acoustic emission acquisition channel is sampled under no-load conditions to obtain environmental noise characteristic parameters. Dynamic zero-point calibration is performed on the acoustic emission acquisition channel using environmental noise characteristic parameters, and multi-dimensional feature correlation mapping is performed in combination with the multi-dimensional physical property characteristic parameters of the slab to construct the initial reference parameters for forming.

[0008] Preferably, the method for obtaining the fiber damage variable includes: Based on the upper limit of the ambient background signal amplitude in the initial molding reference parameters as the opening threshold, the execution range of the monitoring window that dynamically opens and closes with the pressure load is determined by combining the real-time pressure load partition. Within the execution range of the monitoring window, stress wave pulses are captured and bandpass filtered to obtain acoustic emission signals; short-time Fourier transform is performed on the acoustic emission signals, and the acoustic emission signals are projected into the time-frequency co-domain to generate a three-dimensional distribution map; Energy centroid localization is performed on the three-dimensional distribution map, and energy threshold values ​​are extracted based on the deviation of the background energy plane from the energy amplitude of the centroid region. Using the energy threshold values ​​as a criterion, characteristic frequency components are identified by stripping the intensity of the energy distribution. A pulse feature set is obtained by amplitude sampling and counting accumulation of characteristic frequency components. The contribution of acoustic emission energy release rate and pulse accumulation count in the pulse feature set is coupled to extract fiber damage variables.

[0009] Preferably, the method for mapping fiber damage variables to structural perturbation parameters of the internal structural evolution state of short fiber reinforced slabs includes: The fiber damage variable is smoothed to obtain the rate of change of the fiber damage variable; the rate of change of the fiber damage variable is fused with the initial molding reference parameters by feature gain to obtain the transient energy consumption characteristic value of the energy fluctuation at the current moment; the transient energy consumption characteristic value is stacked by the feature evolution of the time domain trajectory to obtain the micro-energy consumption increment inside the short fiber reinforced slab. By using the initial molding reference parameters, a correlation gain mapping based on stiffness sensitivity is performed on the micro energy consumption increment inside the short fiber reinforced slab to determine the structural degradation index characterizing the degree of stiffness decay of the fiber skeleton. The structural degradation index is converted into a feedback compensation value consistent with the pressure unit to obtain the structural disturbance parameters of the internal structural evolution state of the short fiber reinforced slab.

[0010] Preferably, the method for forming structural perturbation estimation parameters includes: The real-time pressure feedback value of the hot press actuator is obtained by using a pressure sensor, and the pressure change information detected during the molding process is obtained by extracting the fluctuation displacement characteristics of adjacent sampling positions. Based on the timestamp difference between the structural disturbance parameters and the pressure change information, the structural disturbance parameters are shifted along the time axis to achieve phase alignment between the structural disturbance parameters and the pressure change information detected during the molding process. A continuous fixed-length sequence of structural disturbance parameters and pressure change information detected during the forming process is extracted before the current sampling time. Temporal correlation matching and fusion between the sequences are performed, and waveform similarity is evaluated based on the changing trend of fused feature quantities over time. Pressure fluctuation components caused by internal structural evolution are extracted. By utilizing the pressure fluctuation component, gain correlation fusion is performed on the structural disturbance parameters to filter out mechanical error interference and form estimated structural disturbance parameters.

[0011] Preferably, the method for mapping structural disturbance estimation parameters to synchronous out-of-phase compensation components includes: The transient phase is identified by the waveform slope and amplitude polarity of the structural disturbance estimation parameters; the complementary phase values ​​are selected in the phase coordinate system based on the transient phase to determine the reverse compensation phase offset. Using the pressure neutral axis as a reference, the polarity of the amplitude of the structural disturbance estimation parameters is reversed to obtain a mirror amplitude signal. The mirror amplitude signal is then loaded onto the reverse compensation phase offset, and the amplitude and phase signals are recombined by locking the sampling time reference. The recombined reverse offset waveform is used as a synchronous out-of-phase compensation component.

[0012] Preferably, the method for outputting the forming state data stream includes: The servo valve opening degree and hydraulic pump displacement of the actuator in the pressure control loop are extracted in real time, and the current control gain of the pressure control loop is obtained by analysis; the synchronous out-of-phase compensation component and the current control gain of the pressure control loop are scaled and converted to determine the electrical signal amplitude of the transient drive correction command. The amplitude of the electrical signal is superimposed on the original given signal of the pressure control loop to complete the injection of transient drive correction command into the pressure control loop to cancel out the disturbance energy; The system collects feedback pressure, displacement change rate, and slab surface temperature of the hydraulic actuator under transient drive correction commands in real time, encapsulates them into a continuous data sequence with timestamps, and outputs a forming status data stream.

[0013] Preferably, the method for driving the thermo-pressurizing actuator to perform nonlinear stress relaxation pressure reduction control includes: Based on the feedback pressure and displacement change rate marked with timestamps in the forming state data stream, the slab compression rate is analyzed; when the feedback pressure remains constant and the slab compression rate is continuously lower than the preset displacement creep threshold, it is determined that the solidification is complete. The initial decompression slope is determined by using the amplitude of the structural disturbance estimation parameters, and the decompression pressure guidance benchmark that varies with time is established by combining the pressure decay evolution criterion, thus constructing a nonlinear stress relaxation decompression control trajectory. The opening degree of the servo valve is adjusted according to the nonlinear stress relaxation pressure reduction control trajectory, so that the actual pressure of the hot pressure actuator is reduced along the trajectory.

[0014] Preferably, the method for identifying characteristic frequency components by intensity stripping of energy distribution includes: Mark the background energy plane determined by environmental noise characteristic parameters in the three-dimensional distribution map; locate the energy concentration center in the non-background area by sliding the search window, and perform local energy distribution location analysis based on centroid logic; Based on the centroid coordinates, the energy threshold value is determined according to the degree of deviation between the background energy plane and the energy amplitude of the centroid region. Using the energy threshold value as the criterion, discrete energy components in the three-dimensional distribution spectrum that are below the energy threshold value and are unrelated to the centroid coordinates are set to zero and removed, thus extracting high-energy clusters that characterize fiber damage. The high-energy clusters are mapped and projected onto the frequency axis to extract the center frequency and bandwidth corresponding to the high-energy clusters in order to identify the characteristic frequency components.

[0015] Preferably, the method for extracting the pressure fluctuation component caused by the evolution of the internal structure includes: The structural disturbance parameters and the pressure change information detected during the molding process are compared and fused point by point within a sliding window to generate a cross-correlation coefficient sequence that slides with the sampling points. Slope analysis was performed on the cross-correlation coefficient series, and the correlation strength between macroscopic pressure fluctuations and microstructural damage was determined based on the slope change trend. By using the cross-correlation coefficient as the proportional gain coefficient, the pressure change information detected during the molding process is projected onto the time characteristic axis of the structural disturbance parameters, and the pressure fluctuation component caused by the evolution of the internal structure is obtained.

[0016] The beneficial effects of this invention are as follows: By performing energy centroid positioning and strength stripping of the three-dimensional distribution spectrum, and combining the cross-correlation projection analysis of macroscopic pressure fluctuations and microscopic damage characteristics, the precise extraction of microscopic evolution characteristics and deep decoupling of macroscopic and microscopic disturbance components during the hot pressing of short fiber slabs are achieved. This is mainly used to sense the stiffness decay state of the fiber skeleton in real time and generate synchronous out-of-phase compensation commands to cancel it out. By injecting transient correction commands into the pressure loop and performing nonlinear stress relaxation and pressure relief control, the invention effectively suppresses stress concentration inside the slab, reduces the risk of delamination, cracking, and warping of the finished product, and improves the comprehensive mechanical properties of short fiber reinforced fiberboard. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of an intelligent control system for the hot pressing process of short fiber reinforced fiberboard in this invention.

[0019] Figure 2 This is a flowchart illustrating the process of generating structural disturbance estimation parameters in this invention.

[0020] Figure 3 This is a flowchart of the output molding state data stream in this invention.

[0021] Figure 4 This is a flowchart of the nonlinear voltage reduction process in this invention. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figure 1 , Figure 2 , Figure 3 and Figure 4 This is one embodiment of the present invention, which provides an intelligent control system for the hot pressing process of short fiber reinforced fiberboard, including the following steps: Methods for obtaining initial reference parameters for molding include: Information on fiber content, thickness, initial density, and moisture content of the short fiber reinforced slab to be processed was collected and the parameters were normalized to obtain multidimensional physical property parameters of the slab. It should be noted that the fiber content, thickness, density, and moisture content of the slab are collected and standardized to construct multi-dimensional physical property parameters. Simultaneously, environmental background signals are sampled under no-load conditions, and the amplitude and dominant frequency range are extracted as environmental noise characteristic parameters. These parameters are used to perform dynamic zero-point calibration on the acquisition channels, and the physical property parameters are correlated with timestamps to construct initial molding reference parameters.

[0026] The acoustic emission acquisition channel is sampled under no-load conditions to obtain environmental noise characteristic parameters. Dynamic zero-point calibration is performed on the acoustic emission acquisition channel using environmental noise characteristic parameters, and multi-dimensional feature correlation mapping is performed in combination with the multi-dimensional physical property characteristic parameters of the slab to construct the initial reference parameters for forming.

[0027] Specifically, the pressure plate is kept suspended and stationary without the hot pressing equipment contacting the slab. The acoustic emission acquisition channel is turned on to continuously sample and record the original acoustic emission signal within a preset time period, such as 20 to 60 seconds. Amplitude statistics and spectrum analysis are performed on the acquired signal to extract the average amplitude level, peak amplitude range and main frequency distribution interval corresponding to the environmental noise, and obtain the characteristic parameters of the environmental noise. After obtaining the environmental noise characteristic parameters, the environmental noise characteristic parameters are used as a reference to perform dynamic zero-point calibration on the acoustic emission acquisition channel, so that the acquired signal can stably return to the background reference range under no-load conditions. The environmental noise characteristic parameters and the multi-dimensional physical property characteristic parameters of the slab are combined and associated in the order of timestamps to form the initial reference parameters for molding.

[0028] In the hot pressing process of existing short fiber reinforced fiberboard, acoustic emission monitoring is often affected by strong noise from hydraulic mechanical vibration, electromagnetic interference, and pressure load fluctuations. Traditional methods often use fixed thresholds, which are difficult to adapt to the complex working conditions where the noise reference constantly drifts during hot pressing. This results in the signal being easily submerged by noise, making it impossible to accurately distinguish micro-fiber damage. In addition, existing analysis methods are too simplistic and fail to combine time and frequency characteristics for in-depth analysis, making it difficult to extract the true characteristic information representing structural degradation from the complex pressure field. This limits the accuracy of the control system in sensing the internal state of the slab. Therefore, this invention designs a fiber damage variable acquisition method that can adapt to environmental noise and achieve in-depth multi-dimensional feature extraction. The specific details are as follows: Methods for obtaining fiber damage variables include: Based on the upper limit of the ambient background signal amplitude in the initial molding reference parameters as the opening threshold, the execution range of the monitoring window that dynamically opens and closes with the pressure load is determined by combining the real-time pressure load partition. Specifically, the upper limit of the ambient background signal amplitude in the initial reference parameters of molding is read and set as the threshold for opening the acoustic emission signal; real-time pressure load data during the hot pressing process is collected simultaneously, and the real-time pressure load is divided into three continuous pressure load zones: low-load contact stage, medium-load compaction stage and high-load curing stage according to the pressure load change amplitude. The upper limit of the ambient background signal amplitude is jointly determined with the pressure load zone. When the real-time pressure load is in any pressure load zone and the instantaneous amplitude of the acoustic emission signal exceeds the upper limit of the ambient background signal amplitude, the monitoring window is triggered to open. When the instantaneous amplitude of the acoustic emission signal falls back to below the upper limit of the ambient background signal amplitude, the monitoring window is closed, thereby determining the execution range of the monitoring window that is dynamically opened and closed with the pressure load.

[0029] Within the execution range of the monitoring window, stress wave pulses are captured and bandpass filtered to obtain acoustic emission signals; short-time Fourier transform is performed on the acoustic emission signals, and the acoustic emission signals are projected into the time-frequency co-domain to generate a three-dimensional distribution map; Specifically, within the execution range of the monitoring window, the raw signal output by the acoustic emission acquisition channel is continuously sampled to capture stress wave pulses generated by changes in the internal structure of the slab; based on the main frequency distribution range corresponding to the environmental noise characteristic parameters in the initial forming reference parameters, which is the continuous frequency segment with the highest energy proportion of the environmental background signal in the spectrum analysis results, the effective frequency range of the acoustic emission signal is determined, and the effective frequency range is used as the bandpass filtering range to perform bandpass filtering on the captured stress wave pulses to filter out mechanical vibration interference components below the effective frequency range and electromagnetic noise components above the effective frequency range, thereby obtaining the acoustic emission signal; After obtaining the acoustic emission signal, the acoustic emission signal is continuously segmented according to a preset fixed time length (e.g., 2 milliseconds to 10 milliseconds), and a short-time Fourier transform is performed on the acoustic emission signal in each time period to convert the acoustic emission signal in the corresponding time period into frequency distribution data. The frequency distribution data is matched with the corresponding timestamps and combined with the energy amplitude of each frequency component to form a correspondence between the acoustic emission signals in the three dimensions of time axis, frequency axis and energy axis. This allows the acoustic emission signals to be projected into the time-frequency co-domain to generate a three-dimensional distribution map.

[0030] Energy centroid localization is performed on the three-dimensional distribution map, and energy threshold values ​​are extracted based on the deviation of the background energy plane from the energy amplitude of the centroid region. Using the energy threshold values ​​as a criterion, characteristic frequency components are identified by stripping the intensity of the energy distribution. It should be noted that the frequency energy distribution corresponding to each time slice is scanned segment by segment within the time-frequency co-domain, and the amplitude distribution of each frequency component on the energy axis is statistically analyzed; based on the gain weight relationship between energy amplitude and frequency position, the energy centroid position corresponding to the overall energy concentration area in the three-dimensional distribution map is determined. After obtaining the location of the energy centroid, the low-energy distribution area far from the energy centroid region in the three-dimensional distribution map is used as the background energy plane, and the average energy amplitude corresponding to the background energy plane is calculated. The difference between the energy amplitude in the energy centroid region and the energy amplitude in the background energy plane is calculated to obtain the energy threshold value that distinguishes the effective acoustic emission signal from the background noise. Using the energy threshold value as the criterion, intensity stripping is performed on the energy distribution in the three-dimensional distribution map, retaining the frequency components with energy amplitudes higher than the energy threshold value, and the retained frequency components are determined as characteristic frequency components.

[0031] A pulse feature set is obtained by amplitude sampling and counting accumulation of characteristic frequency components. The contribution of acoustic emission energy release rate and pulse accumulation count in the pulse feature set is coupled to extract fiber damage variables.

[0032] Specifically, the instantaneous amplitude of the characteristic frequency components in the time series is collected in real time, and the number of pulses exceeding the energy threshold value is recorded during the monitoring window. The collected amplitude sequence and the total number of pulses obtained by statistics are logically encapsulated to obtain the pulse feature set. The acoustic emission energy release rate is extracted from the pulse feature set by using the fused feature quantity of the squared amplitude per unit time as the acoustic emission energy release rate. Simultaneously, the pulse cumulative count from the pulse feature set is retrieved, and a high weighting coefficient is assigned to the acoustic emission energy release rate. The weighting ratio and contribution of the acoustic emission energy release rate and pulse cumulative count are fused to output a scalar value indicating the degree of damage to the short fiber reinforced slab, thus completing the extraction of fiber damage variables. For example, the weight of the acoustic emission energy release rate is set to 0.7, and the weight of the pulse cumulative count is set to 0.3. This dynamic window and energy intensity separation design achieves high-purity extraction of slab damage characteristics. The dynamic opening and closing mechanism allows the monitoring window to be adjusted in real time according to pressure load zones, effectively avoiding environmental noise peaks. Through energy centroid positioning and intensity separation in the three-dimensional distribution spectrum, the system can act like a "filter" to remove background impurities and accurately lock the high-energy characteristic frequency of fiber fracture. By coupling the weights of the energy release rate and pulse count, complex physical phenomena are transformed into precise scalar values ​​that can quantify the degree of slab damage. This not only greatly improves the signal-to-noise ratio of damage identification.

[0033] Methods for mapping fiber damage variables to structural perturbation parameters of the internal structural evolution state of short fiber reinforced slabs include: The fiber damage variable is smoothed to obtain the rate of change of the fiber damage variable; the rate of change of the fiber damage variable is fused with the initial molding reference parameters by characteristic gain to obtain the transient energy consumption characteristic value of the energy fluctuation at the current moment; the transient energy consumption characteristic value is stacked by the characteristic evolution of the time domain trajectory to obtain the microscopic energy consumption increment inside the short fiber reinforced slab.

[0034] Specifically, the fiber damage variable is read and a time series is constructed. The fiber damage variable is then denoised and smoothed using a moving average filtering algorithm. A first-order difference operation is performed on the fiber damage variable to obtain the numerical displacement of the fiber damage variable within a unit sampling period and to obtain the rate of change of the fiber damage variable. Using the rate of change of fiber damage variables as a dynamic input, feature gain fusion is performed in conjunction with the multidimensional physical property parameters of the slab. By quantifying the energy contribution of multiple dimensions, the instantaneous energy dissipation intensity inside the slab is analyzed, thereby obtaining the transient energy consumption characteristic value. Numerical integration is performed on the hot-pressing time axis for the transient energy consumption characteristic value at each moment. The time-domain trajectory of the transient energy consumption characteristic value is accumulated to obtain the total energy fluctuation dissipation value from the start of hot pressing to the current moment, and the micro-energy consumption increment inside the short fiber reinforced slab is obtained.

[0035] By using the initial molding reference parameters, a correlation gain mapping based on stiffness sensitivity is performed on the micro-energy consumption increment inside the short fiber reinforced slab to determine the structural degradation index characterizing the degree of stiffness decay of the fiber skeleton.

[0036] Specifically, characteristic constants related to slab thickness and initial density information are retrieved from the initial molding reference parameters. The product of these two parameters is defined as an evaluation benchmark reflecting the initial compressive strength of the fiber skeleton. The incremental micro-energy consumption within the short-fiber reinforced slab is divided by the evaluation benchmark to distribute this energy consumption increment across slabs with different initial physical properties. A correlation gain mapping based on stiffness sensitivity is then performed. The ratio of the energy consumption increment to the initial evaluation benchmark is calculated to quantify the reduction in the slab's initial load-bearing capacity, thus determining the structural deterioration index characterizing the degree of stiffness reduction in the fiber skeleton.

[0037] The structural degradation index is converted into a feedback compensation value consistent with the pressure unit to obtain the structural disturbance parameters of the internal structural evolution state of the short fiber reinforced slab.

[0038] Specifically, by using a preset pressure mapping scaling factor, the dimensionless structural degradation index is mapped onto the pressure amplitude axis, converting the structural degradation index into a feedback compensation value consistent with the pressure unit. This quantifies the degree of microscopic damage into the macroscopic pressure disturbance amplitude, yielding structural disturbance parameters of the internal structural evolution state of the short fiber reinforced slab. For example, if the pressure mapping scaling factor is set to 100 and the structural degradation index is 0.05, the converted feedback compensation value is 5 MPa.

[0039] Existing technologies for handling pressure fluctuations in hot presses often rely solely on closed-loop regulation based on pressure feedback. This "lagging" regulation method is ill-suited to address transient shocks caused by abrupt changes in the slab's internal structure. More importantly, existing pressure control systems cannot distinguish whether pressure fluctuations originate from mechanical vibration, hydraulic pulsation, or deterioration of the slab's internal framework. This leads to frequently "misguided" control commands and even system oscillations due to phase delays. Without decoupling structural evolution from mechanical disturbances and injecting targeted counterbalancing commands, the impact of internal evolution on molding quality cannot be fundamentally eliminated. Therefore, this invention establishes a structural disturbance estimation and out-of-phase mapping mechanism to achieve precise decoupling and reverse counterbalancing of internal evolution components, as detailed below: Methods for generating structural perturbation estimation parameters include: The real-time pressure feedback value of the hot press actuator is obtained by using a pressure sensor, and the pressure change information detected during the molding process is obtained by extracting the fluctuation displacement characteristics of adjacent sampling positions. Specifically, pressure sensors are used to acquire real-time pressure feedback values ​​of the hot press actuator during the molding process. By calculating the fluctuation deviation characteristics of the real-time pressure feedback values ​​at the current sampling time and the adjacent previous sampling time, a value representing the macroscopic load fluctuation is obtained as pressure change information.

[0040] Based on the timestamp difference between the structural disturbance parameters and the pressure change information, the structural disturbance parameters are shifted along the time axis to achieve phase alignment between the structural disturbance parameters and the pressure change information detected during the molding process. Specifically, the starting timestamps of the structural disturbance parameters and the pressure change information detected during the molding process are read separately to obtain the time delay between them. Using this time delay as the translation step size, the structural disturbance parameters are translated ahead or behind on the time axis to eliminate the phase deviation caused by the difference in sensor sampling frequency and signal processing delay, thereby achieving phase alignment between the structural disturbance parameters and the pressure change information detected during the molding process.

[0041] A continuous fixed-length sequence of structural disturbance parameters and pressure change information detected during the forming process is extracted before the current sampling time. Temporal correlation matching and fusion between the sequences are performed, and waveform similarity is evaluated based on the changing trend of fused feature quantities over time. Pressure fluctuation components caused by internal structural evolution are extracted.

[0042] Specifically, taking the current sampling time as the base point, the structural disturbance parameter sequence and the pressure change information sequence detected during the forming process after phase alignment are extracted backward, and the extracted sequence is a preset continuous fixed length sequence (exemplarily 200 sampling points). The elements at corresponding positions in the two sequences are subjected to temporal correlation matching and fusion between the sequences to calculate the cross-correlation coefficient, and the trend of the cross-correlation coefficient with the sampling time is analyzed to evaluate the waveform similarity. When the waveform similarity exceeds the exemplary 0.85, it is determined that the macroscopic pressure fluctuation is caused by internal structural damage, and the part that highly overlaps with the waveform characteristics of the structural disturbance parameter is extracted from the pressure change information detected during the forming process. That is, within the preset continuous fixed length sequence, the peak occurrence time deviation of the two sets of waveforms is less than the preset time window (exemplarily 5 milliseconds), and the slope change direction of the waveforms is completely consistent. The part that meets the consistency characteristics is extracted as the pressure fluctuation component caused by the evolution of the internal structure.

[0043] By utilizing the pressure fluctuation component, gain correlation fusion is performed on the structural disturbance parameters to filter out mechanical error interference and form estimated structural disturbance parameters.

[0044] Specifically, the pressure fluctuation component caused by the internal structural evolution is used as a correction gain to perform gain correlation fusion processing on the structural disturbance parameters. During the superposition process, by comparing the amplitude envelope of the structural disturbance parameters and the pressure fluctuation component caused by the internal structural evolution, components with amplitude fluctuation frequencies exceeding the upper limit of fiber damage evolution frequency are identified as mechanical vibration noise, and components with periodic amplitude cycles consistent with the pumping frequency of the hydraulic pump are identified as hydraulic pulsation interference. The identified mechanical vibration noise and hydraulic pulsation interference are regarded as interference components that do not conform to the internal evolution logic. During the superposition process, the difference cancellation operation is performed with the structural disturbance parameters to filter out mechanical error interference and form the structural disturbance estimation parameters. For example, the upper limit of the fiber damage evolution frequency is 200 Hz. When an amplitude fluctuation frequency of 500 Hz is detected, it is determined to be mechanical vibration noise. The pumping frequency of the hydraulic pump is 50 Hz. When a periodic signal of 50 Hz is detected, it is determined to be hydraulic pulsation interference.

[0045] Methods for mapping structural disturbance estimation parameters to synchronous out-of-phase compensation components include: The transient phase is identified by the waveform slope and amplitude polarity of the structural disturbance estimation parameters; the complementary phase values ​​are selected in the phase coordinate system based on the transient phase to determine the reverse compensation phase offset. Specifically, the first-order derivative of the structural disturbance estimation parameters is performed to obtain the waveform slope, and the amplitude polarity of the waveform relative to zero level is detected in real time. By analyzing the positive and negative signs of the waveform slope and the positive and negative combinations of amplitude polarity, the transient phase of the fluctuation at the current moment is identified in the phase coordinate system. Based on the fluctuation cancellation principle, a phase value complementary to the transient phase is selected in the phase coordinate system, that is, a phase value with a difference of 180 degrees from the transient phase is selected to obtain the reverse compensation phase offset.

[0046] Using the pressure neutral axis as a reference, the polarity of the amplitude of the structural disturbance estimation parameters is reversed to obtain a mirror amplitude signal. The mirror amplitude signal is then loaded onto the reverse compensation phase offset, and the amplitude and phase signals are recombined by locking the sampling time reference. The recombined reverse offset waveform is used as a synchronous out-of-phase compensation component.

[0047] Specifically, a pressure neutral axis is established as a reference line within the pressure control coordinate system. The amplitude of the structural disturbance estimation parameters at each sampling moment is obtained. A polarity reversal mapping of the amplitude of the structural disturbance estimation parameters is performed with the pressure neutral axis as the center of symmetry. That is, the absolute value of the amplitude remains unchanged while the sign of the value is inverted, generating a mirror amplitude signal that is completely opposite to the direction of the disturbance in amplitude space. Example: If the amplitude of the structural disturbance estimation parameter at a certain moment is +2 MPa, after performing the mirror operation with the pressure neutral axis as the reference, the resulting mirror amplitude signal is -2 MPa.

[0048] The mirror amplitude signal is loaded into the reverse compensation phase offset. By locking the unified sampling time reference of the pressure feedback loop, the transformed amplitude data and phase data are recombined to synthesize a dynamic compensation curve with the same frequency and waveform characteristics as the structural disturbance estimation parameters, but with completely opposite phase and mirror amplitude. This recombined reverse offset waveform is used as the synchronous out-of-phase compensation component.

[0049] Methods for outputting the forming state data stream include: The servo valve opening degree and hydraulic pump displacement of the actuator in the pressure control loop are extracted in real time, and the current control gain of the pressure control loop is obtained by analysis; the synchronous out-of-phase compensation component and the current control gain of the pressure control loop are scaled and converted to determine the electrical signal amplitude of the transient drive correction command. Specifically, the servo valve opening degree and hydraulic pump displacement of the actuator in the pressure control loop are extracted in real time by the sensors of the hydraulic actuator. The current control gain of the pressure control loop is obtained by multiplying the servo valve opening degree and the hydraulic pump displacement, which is used to characterize the transfer function relationship between the electrical signal input and the pressure output. The synchronous out-of-phase compensation component is divided by the current control gain of the pressure control loop to perform scale coordination conversion, and the compensation requirement in the pressure domain is converted into the control variable in the voltage or current domain to determine the electrical signal amplitude of the transient drive correction command.

[0050] The amplitude of the electrical signal is superimposed on the original given signal of the pressure control loop to complete the injection of transient drive correction command into the pressure control loop to cancel out the disturbance energy; Specifically, the amplitude of the electrical signal is superimposed and injected into the original given signal of the pressure control loop. By changing the instantaneous control voltage of the servo valve, the hydraulic actuator is driven to generate a reverse pressure with the opposite direction and equal amplitude to the energy of the internal structural evolution disturbance, thus completing the injection of a transient drive correction command that cancels out the disturbance energy into the pressure control loop.

[0051] The system collects feedback pressure, displacement change rate, and slab surface temperature of the hydraulic actuator under transient drive correction commands in real time, encapsulates them into a continuous data sequence with timestamps, and outputs a forming status data stream.

[0052] Specifically, during the transient drive correction command applied to the hydraulic actuator, the feedback pressure, displacement change rate, and slab surface temperature of the hydraulic actuator are collected in real time through pressure sensors, displacement sensors, and temperature sensors. The collected feedback pressure, displacement change rate, and slab surface temperature are logically encapsulated according to a unified sampling frequency to construct a continuous data sequence with timestamps. The continuous data sequence is then output in real time as a forming state data stream reflecting the dynamic characteristics of the entire hot pressing process. Through the aforementioned structural disturbance estimation and synchronous out-of-phase compensation strategy, a leap from passive feedback to active counterbalancing is achieved. Phase alignment technology eliminates sampling delays between sensors, ensuring zero deviation in the compensation command over time. Furthermore, the difference cancellation operation, performed using waveform characteristics, isolates hydraulic pump pulsations and mechanical vibrations from the complex pressure background, correcting only disturbances generated by internal structural evolution. This generated mirrored amplitude signal, through scaling with the control loop gain, can drive the servo valve to produce instantaneous reverse pressure, physically offsetting the impact energy from damage evolution.

[0053] Methods for driving a thermo-pressure actuator to perform nonlinear stress relaxation pressure reduction control include: Based on the feedback pressure and displacement change rate marked with timestamps in the forming state data stream, the slab compression rate is analyzed; when the feedback pressure remains constant and the slab compression rate is continuously lower than the preset displacement creep threshold, it is determined that the solidification is complete. Specifically, the feedback pressure and displacement change rate with timestamps are extracted in real time from the forming state data stream. The slab compression rate is analyzed by integrating the displacement change rate over time or by directly reading the displacement difference output by the displacement sensor. The numerical fluctuation of the feedback pressure is monitored in real time. When the feedback pressure remains constant and the slab compression rate is continuously lower than the preset displacement creep threshold, it indicates that the internal physical structure of the slab has become stable and no longer undergoes significant volume compression. Thus, it is determined that the curing is complete. Example: The preset displacement creep threshold is 0.01 mm / s. When the feedback pressure is maintained at 15 MPa and the slab compression rate is lower than 0.01 mm / s for 5 consecutive seconds, it is determined that the curing is complete.

[0054] The initial decompression slope is determined by using the amplitude of the structural disturbance estimation parameters, and the decompression pressure guidance benchmark that varies with time is established by combining the pressure decay evolution criterion, thus constructing a nonlinear stress relaxation decompression control trajectory. Specifically, the structural disturbance estimation parameters are retrieved and their terminal amplitudes are extracted. The terminal amplitudes are then converted into a pressure reference at the decompression start point. The initial decompression slope is determined using the pressure reference. A logarithmic decay logic function is introduced, and the initial decompression slope is used as the evolution boundary to calculate a sequence of decompression pressure given values ​​that decreases nonlinearly with decompression time. By arranging and combining the decompression pressure given value sequence along the time axis, a nonlinear stress relaxation decompression control trajectory that simulates the physical properties of material stress relaxation is constructed. For example: The final amplitude of the structural disturbance estimation parameters is 5 MPa. Based on this, the initial depressurization slope is set to 0.5 MPa per second, combined with the logarithmic decay formula: Calculate the given pressure for each moment, where, During the depressurization process The corresponding pressure relief setpoint at any given time. It is the initial pressure value at the moment of depressurization. This is the stress relaxation amplitude coefficient, determined by weighting the final amplitude of the structural disturbance estimation parameters with the initial density information in the multidimensional physical property parameters of the slab. For example, it is obtained by multiplying and summing the final amplitude of the structural disturbance estimation parameters with the normalized values ​​of the initial density information. An exemplary value of 1.25 determines the total intensity of the pressure drop. It is the symbol for the natural logarithm. It is a time evolution term. It is the pressure relief rate attenuation factor, which is determined by the slab surface temperature in the forming state data stream. For example, by establishing a linear mapping relationship between the slab surface temperature and the pressure relief rate, the attenuation intensity can be determined in real time. Since the higher the temperature, the stronger the plasticity of the fibers inside the slab and the faster the stress is released, the real-time slab surface temperature is multiplied by a preset temperature sensing coefficient, and the corresponding pressure relief rate attenuation factor is directly calculated, thereby realizing the real-time adjustment of the slope of the pressure reduction trajectory. Example: Setting the temperature sensing coefficient to 0.00045, when the real-time collected slab surface temperature is 180 degrees Celsius, the pressure relief rate attenuation factor calculated through proportional mapping is, for example, 0.081; if the temperature increases, the calculated... As the value increases, the driving pressure decreases at a faster rate.

[0055] The opening degree of the servo valve is adjusted according to the nonlinear stress relaxation pressure reduction control trajectory, so that the actual pressure of the hot pressure actuator is reduced along the trajectory.

[0056] Specifically, the nonlinear stress relaxation pressure reduction control trajectory is used as the dynamic command target of the pressure control loop. By comparing the actual feedback pressure of the hot-pressing actuator with the corresponding pressure relief setpoint in the trajectory, the opening degree of the servo valve is adjusted in real time to change the return oil flow of the hydraulic chamber. This forces the actual pressure of the hot-pressing actuator to decrease nonlinearly along the nonlinear stress relaxation pressure reduction control trajectory, thereby achieving a smooth release of residual stress inside the short fiber reinforced slab and executing nonlinear stress relaxation pressure reduction control.

[0057] Methods for identifying characteristic frequency components by intensity stripping of energy distribution include: Mark the background energy plane determined by environmental noise characteristic parameters in the three-dimensional distribution map; locate the energy concentration center in the non-background area by sliding the search window, and perform local energy distribution location analysis based on centroid logic; Specifically, in the generated 3D distribution map, the background energy plane is marked based on the average amplitude level extracted from the environmental noise characteristic parameters; a sliding search window of a preset size is used to perform a full-domain scan in the time-frequency plane of the 3D distribution map, and the energy concentration center in the non-background area is located by capturing areas with abrupt changes in energy amplitude; the centroid algorithm is called on the located area, and a weighted average operation is performed with frequency and time as coordinate axes and energy amplitude as a weighting factor to calculate the centroid coordinates of the local energy distribution; for example, the size of the sliding search window is set to an exemplary 50×50 time-frequency cell grid, and the centroid algorithm is used to calculate the centroid coordinates of the area reflecting the most energy-dense region.

[0058] Based on the centroid coordinates, the energy threshold value is determined according to the degree of deviation between the background energy plane and the energy amplitude of the centroid region. Using the energy threshold value as the criterion, discrete energy components in the three-dimensional distribution spectrum that are below the energy threshold value and are unrelated to the centroid coordinates are set to zero and removed, thus extracting high-energy clusters that characterize fiber damage.

[0059] Specifically, using the centroid coordinates as a reference, the energy amplitude in the neighborhood of the coordinate point is extracted, the difference between the energy amplitude and the energy amplitude of the background energy plane is analyzed, and the difference or the proportional coefficient of the difference is determined as the energy threshold value. Then, using the energy threshold value as a hard threshold judgment condition, logical judgment is performed on each pixel unit in the three-dimensional distribution map, and discrete energy components with energy amplitudes lower than the energy threshold value and not connected to the centroid coordinates in the topological structure are zeroed out. In the cluttered signal background, high-energy clusters with continuous energy and high intensity, representing fiber damage characteristics, are extracted.

[0060] The high-energy clusters are mapped and projected onto the frequency axis to extract the center frequency and bandwidth corresponding to the high-energy clusters in order to identify the characteristic frequency components.

[0061] Specifically, the high-energy clusters representing fiber damage characteristics are geometrically projected onto the frequency axis, and the starting and ending frequencies covered on the frequency axis after projection are statistically analyzed to extract the center frequency and bandwidth corresponding to the high-energy clusters representing fiber damage characteristics. The frequency range falling within the bandwidth is determined as the characteristic frequency components that can reflect fiber breakage or matrix cracking.

[0062] Methods for extracting pressure fluctuation components caused by internal structural evolution include: The structural disturbance parameters and the pressure change information detected during the molding process are compared and fused point by point within a sliding window to generate a cross-correlation coefficient sequence that slides with the sampling points. Specifically, the generated sequence of structural perturbation parameters and the sequence of pressure changes detected during the molding process after phase alignment are used as dynamic inputs. Point-to-point correlation comparison and energy fusion convolution operations are performed within a set sliding window (an example of 50 sampling points and an example of a sliding step size of 1 sampling point). At each moving step, the values ​​of structural perturbation parameters within the area covered by the sliding window are multiplied and accumulated one by one with the values ​​of pressure changes detected during the molding process at the corresponding positions. As the sliding window continues to shift on the time axis, the accumulated calculation results corresponding to each sampling point position are recorded, generating two sets of cross-correlation coefficient sequences of signal waveform similarity that change dynamically with time as the sampling points slide.

[0063] Slope analysis was performed on the cross-correlation coefficient series, and the correlation strength between macroscopic pressure fluctuations and microstructural damage was determined based on the slope change trend. Specifically, a first-order difference operation is performed on the cross-correlation coefficient sequence that slides with the sampling points to perform slope analysis, and the growth rate and peak characteristics of the cross-correlation coefficient are observed. The correlation between macroscopic pressure fluctuations and microstructural damage is judged based on the slope change trend. If the slope change rate is greater than the example 0.15 and the cross-correlation coefficient value enters the preset correlation range of the example 0.8 to 1.0, it is determined that the macroscopic pressure fluctuation is caused by damage to the internal fiber skeleton, and the current cross-correlation value is extracted as the quantified correlation strength.

[0064] By using the cross-correlation coefficient as the proportional gain coefficient, the pressure change information detected during the molding process is projected onto the time characteristic axis of the structural disturbance parameters, and the pressure fluctuation component caused by the evolution of the internal structure is obtained.

[0065] Specifically, by using the correlation strength as a proportional gain coefficient, the pressure change information detected during the molding process is mapped and projected onto the time characteristic axis of the structural disturbance parameters. By performing gain adjustment and fusion of the pressure change information detected during the molding process and the correlation strength at the corresponding moment, the pressure component that is consistent with the micro-damage characteristics is extracted from the complex total pressure fluctuation, and irrelevant mechanical interference is filtered out to obtain the pressure fluctuation component caused by the evolution of the internal structure.

[0066] In summary, this invention achieves precise extraction of microscopic evolution characteristics and deep decoupling of macroscopic and microscopic disturbance components during the hot pressing of short fiber slabs by performing energy centroid positioning and strength stripping of three-dimensional distribution maps, combined with cross-correlation projection analysis of macroscopic pressure fluctuations and microscopic damage characteristics. It is mainly used to sense the stiffness decay state of the fiber skeleton in real time and generate synchronous out-of-phase compensation commands to cancel it out. By injecting transient correction commands into the pressure loop and performing nonlinear stress relaxation and pressure relief control, it effectively suppresses stress concentration inside the slab, reduces the risk of delamination, cracking, and warping of the finished product, and improves the comprehensive mechanical properties of short fiber reinforced fiberboard.

[0067] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An intelligent control system for the hot pressing process of short fiber reinforced fiberboard, characterized in that, include: The acquisition module collects multidimensional physical property parameters of the short fiber reinforced slab to be processed, and performs environmental noise suppression on the acoustic emission acquisition channel to obtain the initial reference parameters for molding. The damage sensing module, under the constraint of the initial reference parameters of molding, captures stress wave pulses through a monitoring window that dynamically opens and closes with pressure load, and projects the acoustic emission signal into the time-frequency co-domain. Through intensity stripping of energy distribution, it extracts fiber damage variables. The perturbation module maps fiber damage variables to structural perturbation parameters of the internal structural evolution state of short fiber reinforced slabs, and couples them with pressure change information detected during the forming process to form structural perturbation estimation parameters. The adjustment module maps the structural disturbance estimation parameters into synchronous out-of-phase compensation components, injects transient drive correction commands that cancel out the disturbance energy into the pressure control loop, and outputs the forming state data stream. The pressure relief module determines the curing completion status based on the molding state data stream and drives the hot pressing actuator to perform nonlinear stress relaxation pressure reduction control based on the structural disturbance estimation parameters.

2. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 1, characterized in that, The method for obtaining the initial reference parameters for molding includes: Information on fiber content, thickness, initial density, and moisture content of the short fiber reinforced slab to be processed was collected and the parameters were normalized to obtain multidimensional physical property parameters of the slab. The acoustic emission acquisition channel is sampled under no-load conditions to obtain environmental noise characteristic parameters. Dynamic zero-point calibration is performed on the acoustic emission acquisition channel using environmental noise characteristic parameters, and multi-dimensional feature correlation mapping is performed in combination with the multi-dimensional physical property characteristic parameters of the slab to construct the initial reference parameters for forming.

3. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 2, characterized in that, The methods for obtaining the fiber damage variables include: Based on the upper limit of the ambient background signal amplitude in the initial molding reference parameters as the opening threshold, the execution range of the monitoring window that dynamically opens and closes with the pressure load is determined by combining the real-time pressure load partition. Within the execution range of the monitoring window, stress wave pulses are captured and bandpass filtered to obtain acoustic emission signals; short-time Fourier transform is performed on the acoustic emission signals, and the acoustic emission signals are projected into the time-frequency co-domain to generate a three-dimensional distribution map; Energy centroid localization is performed on the three-dimensional distribution map, and energy threshold values ​​are extracted based on the deviation of the background energy plane from the energy amplitude of the centroid region. Using the energy threshold values ​​as a criterion, characteristic frequency components are identified by stripping the intensity of the energy distribution. A pulse feature set is obtained by amplitude sampling and counting accumulation of characteristic frequency components. The contribution of acoustic emission energy release rate and pulse accumulation count in the pulse feature set is coupled to extract fiber damage variables.

4. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 3, characterized in that, The method for mapping fiber damage variables to structural perturbation parameters of the internal structural evolution state of short fiber reinforced slabs includes: The fiber damage variable is smoothed to obtain the rate of change of the fiber damage variable; the rate of change of the fiber damage variable is fused with the initial molding reference parameters by feature gain to obtain the transient energy consumption characteristic value of the energy fluctuation at the current moment; the transient energy consumption characteristic value is stacked by the feature evolution of the time domain trajectory to obtain the micro-energy consumption increment inside the short fiber reinforced slab. By using the initial molding reference parameters, a correlation gain mapping based on stiffness sensitivity is performed on the micro energy consumption increment inside the short fiber reinforced slab to determine the structural degradation index characterizing the degree of stiffness decay of the fiber skeleton. The structural degradation index is converted into a feedback compensation value consistent with the pressure unit to obtain the structural disturbance parameters of the internal structural evolution state of the short fiber reinforced slab.

5. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 4, characterized in that, The method for forming structural perturbation estimation parameters includes: The real-time pressure feedback value of the hot press actuator is obtained by using a pressure sensor, and the pressure change information detected during the molding process is obtained by extracting the fluctuation displacement characteristics of adjacent sampling positions. Based on the timestamp difference between the structural disturbance parameters and the pressure change information, the structural disturbance parameters are shifted along the time axis to achieve phase alignment between the structural disturbance parameters and the pressure change information detected during the molding process. A continuous fixed-length sequence of structural disturbance parameters and pressure change information detected during the forming process is extracted before the current sampling time. Temporal correlation matching and fusion between the sequences are performed, and waveform similarity is evaluated based on the changing trend of fused feature quantities over time. Pressure fluctuation components caused by internal structural evolution are extracted. By utilizing the pressure fluctuation component, gain correlation fusion is performed on the structural disturbance parameters to filter out mechanical error interference and form estimated structural disturbance parameters.

6. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 5, characterized in that, The method for mapping structural disturbance estimation parameters to synchronous out-of-phase compensation components includes: The transient phase is identified by the waveform slope and amplitude polarity of the structural disturbance estimation parameters; the complementary phase values ​​are selected in the phase coordinate system based on the transient phase to determine the reverse compensation phase offset. Using the pressure neutral axis as a reference, the polarity of the amplitude of the structural disturbance estimation parameters is reversed to obtain a mirror amplitude signal. The mirror amplitude signal is then loaded onto the reverse compensation phase offset, and the amplitude and phase signals are recombined by locking the sampling time reference. The recombined reverse offset waveform is used as a synchronous out-of-phase compensation component.

7. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 6, characterized in that, The method for outputting the forming state data stream includes: The servo valve opening degree and hydraulic pump displacement of the actuator in the pressure control loop are extracted in real time, and the current control gain of the pressure control loop is obtained by analysis; the synchronous out-of-phase compensation component and the current control gain of the pressure control loop are scaled and converted to determine the electrical signal amplitude of the transient drive correction command. The amplitude of the electrical signal is superimposed on the original given signal of the pressure control loop to complete the injection of transient drive correction command into the pressure control loop to cancel out the disturbance energy; The system collects feedback pressure, displacement change rate, and slab surface temperature of the hydraulic actuator under transient drive correction commands in real time, encapsulates them into a continuous data sequence with timestamps, and outputs a forming status data stream.

8. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 7, characterized in that, The method for driving the thermo-pressure actuator to perform nonlinear stress relaxation and pressure reduction control includes: Based on the feedback pressure and displacement change rate marked with timestamps in the forming state data stream, the slab compression rate is analyzed; when the feedback pressure remains constant and the slab compression rate is continuously lower than the preset displacement creep threshold, it is determined that the solidification is complete. The initial decompression slope is determined by using the amplitude of the structural disturbance estimation parameters, and the decompression pressure guidance benchmark that varies with time is established by combining the pressure decay evolution criterion, thus constructing a nonlinear stress relaxation decompression control trajectory. The opening degree of the servo valve is adjusted according to the nonlinear stress relaxation pressure reduction control trajectory, so that the actual pressure of the hot pressure actuator is reduced along the trajectory.

9. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 3, characterized in that, The method for identifying characteristic frequency components by intensity stripping of energy distribution includes: Mark the background energy plane determined by environmental noise characteristic parameters in the three-dimensional distribution map; locate the energy concentration center in the non-background area by sliding the search window, and perform local energy distribution location analysis based on centroid logic; Based on the centroid coordinates, the energy threshold value is determined according to the degree of deviation between the background energy plane and the energy amplitude of the centroid region. Using the energy threshold value as the criterion, discrete energy components in the three-dimensional distribution spectrum that are below the energy threshold value and are unrelated to the centroid coordinates are set to zero and removed, thus extracting high-energy clusters that characterize fiber damage. The high-energy clusters are mapped and projected onto the frequency axis to extract the center frequency and bandwidth corresponding to the high-energy clusters in order to identify the characteristic frequency components.

10. The intelligent control system for the hot pressing process of short fiber reinforced fiberboard as described in claim 5, characterized in that, The method for extracting pressure fluctuation components caused by internal structural evolution includes: The structural disturbance parameters and the pressure change information detected during the molding process are compared and fused point by point within a sliding window to generate a cross-correlation coefficient sequence that slides with the sampling points. Slope analysis was performed on the cross-correlation coefficient series, and the correlation strength between macroscopic pressure fluctuations and microstructural damage was determined based on the slope change trend. By using the cross-correlation coefficient as the proportional gain coefficient, the pressure change information detected during the molding process is projected onto the time characteristic axis of the structural disturbance parameters, and the pressure fluctuation component caused by the evolution of the internal structure is obtained.

Images