A four-corner leveling control system for a composite press

Abstract

The present application relates to the technical field of press control, in particular to a kind of composite material press four-corner leveling control system, comprising: data perception module real-time acquisition press four-corner displacement load synchronous data, worktable global temperature field, blank forming thickness deviation data;Four-corner load decoupling analysis module is based on hyperstatic mechanics model and solves four-corner load coupling interference, exports each corner point independent leveling control correction quantity;Multi-cylinder collaborative leveling execution module uses improved deep reinforcement learning leveling algorithm to synchronously control four-cylinder output pressure and displacement feed;Thermal creep adaptive compensation module combines temperature field, worktable thermal coupling deformation law and material thermal creep shrinkage characteristics, generates each corner point leveling displacement compensation quantity;Process linkage calibration unit combines blank thickness deviation and real-time process parameters, and periodically micro-calibrates four-corner leveling reference in pressure maintaining stage.

Description

Technical Field

[0001] This invention relates to the field of press control technology, specifically to a four-corner leveling control system for a composite material press. Background Technology

[0002] Composite material presses are core equipment in composite material compression molding (such as lightweight automotive parts, aerospace composite structural components, and fiberglass products). The levelness of the worktable directly determines the molding accuracy, thickness uniformity, and product quality stability of composite material products. During long-term service, factors such as frequent impacts from off-center loading conditions, uneven temperature gradient distribution on the worktable surface, and thermal creep shrinkage during composite material curing can easily lead to problems such as unbalanced load distribution at the four corners of the worktable, accumulated thermal deformation, and drift of the leveling reference. This results in increased thickness deviation and internal defects in composite material products, and may even cause serious consequences such as mold damage and batch scrapping. As composite material products develop towards larger, more complex, and higher precision, higher requirements are placed on presses in terms of dynamic decoupling of four-corner loads, multi-cylinder collaborative leveling control, adaptive thermal creep compensation, and process linkage calibration. There is an urgent need for a four-corner leveling control system that can deeply integrate load perception and decoupling analysis, intelligent decision-making and collaborative execution, and thermal deformation compensation and process linkage to ensure the high-precision and high-stability operation of composite material presses in complex molding scenarios.

[0003] However, traditional four-corner leveling methods for composite material presses have significant shortcomings: severe load coupling, with strong elastic and pressure coupling interference between the four corner leveling cylinders; traditional control methods fail to effectively decouple the corner loads, leading to mutual interference between leveling commands at each corner and limited control accuracy; lagging digital decision-making, relying heavily on traditional control algorithms such as PID, making parameter tuning difficult and exhibiting weak anti-interference capabilities under complex off-center load conditions, hindering efficient multi-cylinder coordination and rapid response; lack of thermal deformation compensation, failing to incorporate thermal deformation caused by the table temperature gradient and composite material curing shrinkage into the leveling compensation system, resulting in significant leveling accuracy drift due to thermal creep effects; and insufficient process linkage, with leveling control disconnected from composite material molding process parameters (temperature, holding pressure, curing stage, etc.), making dynamic micro-calibration of the leveling benchmark based on billet thickness deviation impossible. Overall, the technology faces multiple challenges, including insufficient load decoupling capability, low intelligent decision-making level, lack of thermal compensation mechanism, and poor process coordination. Summary of the Invention

[0004] This application provides a four-corner leveling control system for a composite material press to solve problems such as severe load coupling and lack of thermal deformation compensation in the prior art.

[0005] The first aspect of this application provides a four-corner leveling control system for a composite material press, comprising: a data sensing module, a four-corner load decoupling analysis module, a multi-cylinder collaborative leveling execution module, a thermal creep adaptive compensation module, and a process linkage calibration unit; wherein: the data sensing module is used to collect real-time displacement load synchronous data of the four corner points of the press, temperature field distribution data of the entire worktable surface, and forming thickness deviation data of the composite material billet; the four-corner load decoupling analysis module is used to perform coupling interference calculation on the four-corner loads based on a statically indeterminate mechanical model, and output independent leveling control correction amounts for each corner point; the multi-cylinder collaborative leveling execution module is used to calculate the decoupling loads based on the decoupled loads... The leveling control correction amount, combined with the improved deep reinforcement learning leveling algorithm, synchronously regulates the output pressure and displacement feed of the four leveling cylinders; the thermal creep adaptive compensation module is used to dynamically generate the leveling displacement compensation amount of each corner point based on the temperature field distribution data of the entire worktable surface, combined with the thermo-mechanical coupling deformation law of the worktable structure and the thermal creep curing shrinkage characteristics of the composite material, to offset the interference of the thermal deformation of the worktable caused by the temperature gradient on the leveling accuracy; the process linkage calibration unit is used to periodically perform micro-calibration of the four corner leveling benchmarks during the pressure holding stage based on the molding thickness deviation data of the composite material blank and the current molding process parameters.

[0006] Preferably, the data sensing module includes a corner displacement load synchronous acquisition unit, a workbench global temperature field monitoring unit, and a billet forming thickness online detection unit. The corner displacement load synchronous acquisition unit is used to synchronously acquire real-time displacement and load data of four leveling corner points using high-precision displacement and pressure sensors. The workbench global temperature field monitoring unit is used to acquire the temperature of multiple points on the workbench surface through an array of infrared temperature measurement nodes to obtain global temperature field distribution data. The billet forming thickness online detection unit is used to acquire the thickness of the composite material billet in multiple regions through a laser thickness measurement array and calculate the forming thickness deviation data.

[0007] Preferably, the four-corner load decoupling analysis module includes a statically indeterminate mechanical modeling unit and a coupling interference calculation unit. The statically indeterminate mechanical modeling unit is used to construct a four-corner support statically indeterminate mechanical analysis model based on the press frame structural parameters and the cylinder arrangement. The coupling interference calculation unit is used to calculate the elastic coupling interference of the four-corner loads based on the statically indeterminate mechanical model, separate the coupling effect caused by the elastic deformation of the frame and the coaxiality error of the cylinder installation, and output the independent leveling control correction amount of each corner point.

[0008] Preferably, the multi-cylinder collaborative leveling execution module includes a deep reinforcement learning decision unit and a multi-cylinder servo drive control unit. The deep reinforcement learning decision unit is used to take the decoupled leveling control correction amount as input and optimize the output control command of each cylinder through an improved deep reinforcement learning leveling algorithm. The multi-cylinder servo drive control unit is used to synchronously regulate the output pressure and displacement feed of the four leveling cylinders according to the control command, and regulate the levelness of the worktable in real time.

[0009] Preferably, the adaptive thermal creep compensation module includes a workbench thermal deformation calculation unit, a billet creep prediction unit, and a compensation amount dynamic generation unit. The workbench thermal deformation calculation unit is used to calculate the thermal deformation at each corner point caused by the temperature gradient by combining global temperature field distribution data with the thermo-mechanical coupling deformation law of the workbench structure. The billet creep prediction unit is used to predict the thickness creep shrinkage of the billet during the pressure holding process based on the thermal creep curing shrinkage characteristics of the composite material. The compensation amount dynamic generation unit is used to fuse the workbench thermal deformation and billet creep to dynamically generate the leveling displacement compensation amount at each corner point, offsetting the interference of thermal deformation on the leveling accuracy.

[0010] Preferably, the process linkage calibration unit includes a process parameter matching unit, a thickness deviation analysis unit, and a leveling benchmark micro-calibration unit. The process parameter matching unit retrieves the molding temperature, holding pressure, and curing stage process parameters of the current composite material blank, and matches the corresponding leveling benchmark calibration strategy and allowable correction range. The thickness deviation analysis unit analyzes the thickness uniformity deviation of the corresponding area at each corner point based on the blank molding thickness deviation data. The leveling benchmark micro-calibration unit performs micro-calibration of the four corner leveling benchmarks during the holding pressure stage, combining process parameters and thickness deviation.

[0011] A second aspect of this application provides a method for four-corner leveling control of a composite material press, comprising: acquiring synchronous displacement-load data of the four corner points of the press, global temperature field distribution data of the worktable surface, and forming thickness deviation data of the composite material billet; based on the synchronous displacement-load data, performing decoupling calculations on the displacement-load coupling relationship of the four corners using a statically indeterminate mechanical model, and outputting leveling control correction amounts for decoupling control at each corner point; and, based on the decoupled leveling control correction amounts, combining an improved deep reinforcement learning leveling algorithm, synchronously controlling the movement of the four leveling cylinders with displacement feed as the core. The system operates in a state where overload limiting constraints are applied to the output pressure. Based on the global temperature field distribution data of the worktable surface and the thermo-mechanical coupling deformation law of the worktable structure, the leveling displacement compensation amount for each corner point is dynamically generated. At the same time, the thermal creep curing shrinkage characteristics of the composite material are incorporated into the process parameter dimension for adaptation and correction. Based on the molding thickness deviation data of the composite material blank, combined with the current molding process parameters and the leveling displacement compensation amount, the interpretable deviation components corresponding to thermal deformation compensation and material thermal creep curing shrinkage are stripped, the residual error of the leveling control link is extracted, and the four corner leveling benchmarks are iteratively micro-calibrated.

[0012] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the program to implement a four-corner leveling control method for a composite material press as described in the above embodiments.

[0013] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a four-corner leveling control method for a composite material press as described in the above embodiments.

[0014] The fifth aspect of this application provides a computer program product, including a computer program or instructions, for implementing a four-corner leveling control method for a composite material press as described in the above embodiments.

[0015] Therefore, this application has the following beneficial effects: This application embodiment comprehensively collects synchronous data of the four corner displacement loads of the press, the temperature field distribution data of the entire worktable, and the billet forming thickness deviation data through a data sensing module, providing multi-dimensional basic data for leveling control. The four corner load decoupling analysis module calculates the coupling interference of the four corner loads based on a statically indeterminate mechanical model, effectively separating the coupling effect caused by the elastic deformation of the frame and the installation error of the cylinder, breaking through the limitation of mutual interference of four corner commands in traditional control. The multi-cylinder collaborative leveling execution module uses an improved deep reinforcement learning leveling algorithm to achieve intelligent collaboration and rapid response of multiple cylinders, overcoming the shortcomings of traditional PID in parameter tuning difficulties and weak anti-interference ability under complex off-center load conditions. The thermal creep adaptive compensation module integrates the thermo-mechanical coupling deformation law of the worktable and the thermal creep curing shrinkage characteristics of composite materials to dynamically generate leveling displacement compensation, effectively offsetting the interference of thermal deformation caused by temperature gradient on leveling accuracy. The process linkage calibration unit performs periodic micro-calibration of the four corner leveling benchmarks during the holding pressure stage based on the billet forming thickness deviation and current process parameters, realizing deep collaboration between leveling control and forming process. This solves the problems of severe load coupling and lack of thermal deformation compensation in existing technologies.

[0016] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0017] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a schematic diagram of the structure of a four-corner leveling control system for a composite material press according to an embodiment of this application; Figure 2 This is a flowchart illustrating a four-corner leveling control method for a composite material press according to an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] The following description, with reference to the accompanying drawings, illustrates a four-corner leveling control system for a composite material press according to an embodiment of this application. Addressing the severe load coupling problem mentioned in the background art, this application provides a four-corner leveling control system for a composite material press. In this system, a data sensing module comprehensively collects synchronous data of the press's four-corner displacement loads, temperature field distribution data across the entire worktable surface, and billet forming thickness deviation data, providing multi-dimensional foundational data for leveling control. A four-corner load decoupling analysis module calculates the coupling interference of the four-corner loads based on a statically indeterminate mechanical model, effectively separating the coupling effects caused by the elastic deformation of the frame and the cylinder installation error, overcoming the limitations of mutual interference between four-corner commands in traditional control. A multi-cylinder collaborative leveling execution module, utilizing an improved deep reinforcement learning leveling algorithm, achieves intelligent collaboration and rapid response among multiple cylinders, overcoming the shortcomings of traditional PID control, such as difficulty in parameter tuning and weak anti-interference capability under complex off-center load conditions. A thermal creep adaptive compensation module integrates the thermo-mechanical coupling deformation law of the worktable with the thermal creep curing shrinkage characteristics of the composite material, dynamically generating leveling displacement compensation, effectively offsetting the interference of thermal deformation caused by temperature gradients on leveling accuracy. The process linkage calibration unit performs periodic micro-calibration of the four corner leveling benchmarks during the pressure holding stage based on the billet forming thickness deviation and current process parameters, achieving deep synergy between leveling control and forming process. This solves problems such as severe load coupling and lack of thermal deformation compensation in existing technologies.

[0020] Figure 1 This is a schematic diagram of the structure of a four-corner leveling control system for a composite material press, provided in an embodiment of this application.

[0021] This application provides a four-corner leveling control system for a composite material press, the system 10 including: Data sensing module 100, four-corner load decoupling analysis module 200, multi-cylinder collaborative leveling execution module 300, thermal creep adaptive compensation module 400, and process linkage calibration unit 500.

[0022] The data sensing module 100 is used to collect real-time synchronous data of displacement loads at the four corners of the press, data of temperature field distribution across the entire worktable surface, and data of molding thickness deviation of the composite material blank. The four-corner load decoupling analysis module 200 is used to perform coupling interference calculation on the four-corner loads based on a statically indeterminate mechanical model, and output independent leveling control corrections for each corner. The multi-cylinder collaborative leveling execution module 300 is used to synchronously control the output pressure and displacement feed of the four leveling cylinders based on the decoupled leveling control corrections and an improved deep reinforcement learning leveling algorithm. The thermal creep adaptive compensation module 400 is used to dynamically generate leveling displacement compensation for each corner based on the temperature field distribution data across the entire worktable surface, combined with the thermo-mechanical coupling deformation law of the worktable structure and the thermal creep curing shrinkage characteristics of the composite material, to offset the interference of thermal deformation of the worktable caused by temperature gradient on the leveling accuracy. The process linkage calibration unit 500 is used to periodically perform micro-calibration of the four-corner leveling benchmark during the pressure holding stage based on the molding thickness deviation data of the composite material blank and the current molding process parameters.

[0023] It is understood that in this embodiment, the data sensing module comprehensively collects synchronous data of the four corner displacement loads of the press, the temperature field distribution data of the entire worktable, and the billet forming thickness deviation data, providing multi-dimensional basic data for leveling control. The four corner load decoupling analysis module calculates the coupling interference of the four corner loads based on the statically indeterminate mechanical model, effectively separating the coupling effect caused by the elastic deformation of the frame and the installation error of the cylinder, breaking through the limitation of mutual interference of four corner commands in traditional control. The multi-cylinder collaborative leveling execution module uses an improved deep reinforcement learning leveling algorithm to achieve intelligent collaboration and rapid response of multiple cylinders, overcoming the defects of traditional PID in difficult parameter tuning and weak anti-interference ability under complex off-center load conditions. The thermal creep adaptive compensation module integrates the thermo-mechanical coupling deformation law of the worktable and the thermal creep curing shrinkage characteristics of composite materials to dynamically generate the leveling displacement compensation amount, effectively offsetting the interference of thermal deformation caused by temperature gradient on the leveling accuracy. The process linkage calibration unit performs periodic micro-calibration of the four corner leveling benchmark during the holding pressure stage according to the billet forming thickness deviation and the current process parameters, realizing deep collaboration between leveling control and forming process. This solves the problems of severe load coupling and lack of thermal deformation compensation in existing technologies.

[0024] In this embodiment, the data sensing module 100 includes: a corner displacement load synchronous acquisition unit, a workbench global temperature field monitoring unit, and a billet forming thickness online detection unit.

[0025] Among them, the corner displacement load synchronous acquisition unit is used to synchronously acquire real-time displacement and load data of four leveling corner points through high-precision displacement sensors and pressure sensors; the workbench global temperature field monitoring unit is used to acquire the temperature of multiple points on the workbench surface through an array of infrared temperature measurement nodes to obtain global temperature field distribution data; the billet forming thickness online detection unit is used to acquire the thickness of composite material billets in multiple regions through a laser thickness measurement array to calculate the forming thickness deviation data.

[0026] It is understood that, in this embodiment, high-precision displacement and pressure sensors are used to simultaneously collect real-time displacement and load data of four leveling corner points, accurately control the force-position coordination state during the leveling process, and provide real-time feedback for the closed-loop control of the four corners leveling; the workbench global temperature field monitoring unit collects the temperature of multiple points on the workbench and reconstructs the global temperature field distribution data, comprehensively grasps the thermal deformation law of the workbench, and provides data support for thermal creep compensation and temperature homogenization control; the billet forming thickness online detection unit collects the thickness of composite material billets in multiple regions and calculates the forming thickness deviation, monitors the billet forming uniformity in real time, reduces process lag and monitoring errors caused by offline sampling inspection, and provides comprehensive raw sensing data for leveling control, thermal field regulation and forming process optimization, improving the globality and accuracy of forming process monitoring.

[0027] In this embodiment of the application, the four-corner load decoupling analysis module 200 includes: a statically indeterminate mechanical modeling unit and a coupling interference solution unit.

[0028] Among them, the statically indeterminate mechanical modeling unit is used to construct a four-corner support statically indeterminate mechanical analysis model based on the press frame structural parameters and cylinder arrangement; the coupling interference solution unit is used to solve the elastic coupling interference of the four-corner load based on the statically indeterminate mechanical model, separate the coupling effect caused by the elastic deformation of the frame and the coaxiality error of the cylinder installation, and output the independent leveling control correction amount of each corner point.

[0029] It is understood that, in this embodiment of the application, a statically indeterminate mechanical modeling unit is used to construct a four-corner support statically indeterminate mechanical analysis model based on the press frame structural parameters and cylinder arrangement. This model accurately depicts the force flow transmission and deformation coupling characteristics of the four-corner support system, providing a reliable mechanical calculation basis for load coupling interference solution. The coupling interference solution unit calculates the elastic coupling interference of the four-corner loads based on the statically indeterminate mechanical model, separating the coupling effect caused by the elastic deformation of the frame and the coaxiality error of the cylinder installation. It outputs independent leveling control correction values ​​for each corner point, eliminating control deviations caused by inter-corner linkage, reducing leveling distortion caused by coupling interference and installation errors, providing accurate correction basis for high-precision four-corner leveling closed-loop control, and improving the decoupling capability and control accuracy of the leveling system.

[0030] It should be noted that a statically indeterminate mechanical analysis model with four corner supports is constructed based on the structural parameters of the press frame and the arrangement of the hydraulic cylinders. First, the structural design parameters of the press frame are used as a basis to extract the geometric characteristic parameters (such as support span, cross-sectional dimensions, and beam structure morphology), material mechanical properties (such as elastic modulus, Poisson's ratio, and shear modulus), and nominal load-bearing capacity parameters (such as rated forming load and allowable elastic deformation), forming the initial theoretical calculation framework for the statically indeterminate mechanical model. Simultaneously, considering the actual arrangement of the four corner hydraulic cylinders, the spatial coordinates of each cylinder, the stiffness of individual supports, installation coaxiality deviations, and assembly constraints are analyzed to represent the global elastic deformation of the frame as equivalent to the supports at each corner point. The stiffness components are transformed into numerically solvable stiffness matrix elements after mechanical equivalence transformation, parameter mapping calibration, and boundary constraint definition. Subsequently, the equivalent support stiffness parameters are registered, combined, and integrated with the rigid body displacement equilibrium equations to correct calculation errors caused by ideal symmetry assumptions, manufacturing and assembly deviations, and elastic deformation coupling of the frame. Key information such as force-displacement constitutive relations, moment balance constraints, and load transfer paths are integrated to finally construct a statically indeterminate mechanical analysis model that can accurately map the force flow transmission law, elastic coupling characteristics, and deformation linkage effect of the four-corner support system. This provides a reliable mechanical calculation basis for subsequent four-corner load coupling interference calculation, leveling control correction amount derivation, and high-precision closed-loop leveling control. In this embodiment, the multi-cylinder collaborative leveling execution module 300 includes: a deep reinforcement learning decision unit and a multi-cylinder servo drive control unit.

[0031] Among them, the deep reinforcement learning decision unit is used to take the decoupled leveling control correction amount as input, and optimize the output control command of each cylinder through the improved deep reinforcement learning leveling algorithm; the multi-cylinder servo drive control unit is used to synchronously regulate the output pressure and displacement feed of the four leveling cylinders according to the control command, and regulate the levelness of the worktable in real time.

[0032] It is understood that the deep reinforcement learning decision unit in this application takes the decoupled leveling control correction amount as input, and optimizes the output control commands of each cylinder through the improved deep reinforcement learning leveling algorithm. It adaptively adapts to the leveling requirements of different forming loads and off-center load conditions, avoiding the limitations of traditional control algorithm parameters being fixed and dynamic response lagging. It provides accurate and optimal decision command basis for multi-cylinder collaborative drive. The multi-cylinder servo drive control unit synchronously adjusts the output pressure and displacement feed of the four leveling cylinders according to the control commands, and adjusts the level of the worktable in real time to ensure the synchronization accuracy and output consistency of each cylinder's action. It quickly offsets the table tilt caused by forming off-center load and coupling disturbance, reduces overshoot oscillation and steady-state error in the leveling process, and improves the dynamic response speed and overall control accuracy of the leveling system.

[0033] A deep reinforcement learning decision unit is constructed using the decoupled leveling control correction as the core input. First, the decoupled correction and real-time collected multi-source working condition data, including corner displacement, load, and levelness error, are standardized and preprocessed. Each parameter is mapped and integrated into a multi-dimensional state vector recognizable by the algorithm, forming the initial input system for reinforcement learning decision-making. Simultaneously, a customized improvement is made to the deep deterministic policy gradient algorithm framework. The Actor-Critic dual-network structure is optimized for continuous control scenarios of corner leveling. A multi-objective reward function integrating levelness accuracy, load balance, motion smoothness, and safe travel constraints is designed. A soft update mechanism for the target network and an experience playback mode are introduced to improve training stability and decision generalization ability. Offline iterative training using historical leveling working condition data is then conducted. The parameter weights of the strategy network are further optimized. During the real-time operation phase, the current state vector is input into the trained strategy network, and the optimal control command increment for each leveling cylinder is obtained quickly through forward inference. This adaptively adapts to the leveling requirements under different forming loads, off-center loading conditions, and thermal deformation disturbances, correcting the inherent defects of traditional fixed parameter control algorithms such as response lag, overshoot oscillation, and poor adaptability to coupled conditions. Subsequently, the output commands are subjected to amplitude limiting and smoothing processing based on the servo system response characteristics, cylinder output boundaries, and process accuracy requirements. Finally, the pressure and displacement control commands of each corner cylinder are output to be independent and coordinated, providing accurate and optimal decision-making basis for the multi-cylinder servo drive unit, effectively improving the dynamic response speed, steady-state control accuracy, and adaptive capability of the leveling system under complex conditions.

[0034] In this embodiment, the thermal creep adaptive compensation module 400 includes a workbench thermal deformation calculation unit, a billet creep prediction unit, and a compensation amount dynamic generation unit.

[0035] The workbench thermal deformation calculation unit is used to calculate the thermal deformation at each corner point caused by the temperature gradient by combining the global temperature field distribution data with the thermo-mechanical coupling deformation law of the workbench structure; the billet creep prediction unit is used to predict the thickness creep shrinkage of the billet during the holding pressure process based on the thermal creep curing shrinkage characteristics of composite materials; the compensation amount dynamic generation unit is used to integrate the workbench thermal deformation amount and the billet creep amount to dynamically generate the leveling displacement compensation amount at each corner point to offset the interference of thermal deformation on the leveling accuracy.

[0036] It is understood that, in this embodiment of the application, the thermal deformation calculation unit of the workbench, combined with the global temperature field distribution data and the thermo-mechanical coupling deformation law of the workbench structure, calculates the amount of thermal deformation at each corner point caused by the temperature gradient, accurately quantifies the table warping deviation caused by the uneven temperature field, and provides a reliable quantitative basis for directional compensation of thermal errors; the billet creep prediction unit, based on the thermal creep curing shrinkage characteristics of composite materials, predicts the amount of thickness creep shrinkage of the billet during the holding pressure process, predicts the dimensional evolution trend of the billet in advance during the molding cycle, and compensates for the thickness uniformity deviation caused by creep shrinkage during the holding pressure stage; the compensation amount dynamic generation unit integrates the thermal deformation amount of the workbench and the curing amount of the billet, dynamically generates the leveling displacement compensation amount of each corner point, offsets the dual interference of thermal deformation and creep shrinkage on the leveling accuracy, reduces the accuracy decay and molding error under long-term high-temperature holding pressure conditions, and improves the thermal adaptability and molding dimensional stability of the leveling system.

[0037] It should be noted that, based on the structural design parameters and material thermophysical properties of the workbench, the geometric dimensions, cross-sectional shape, support constraint method, and thermo-mechanical properties of the material such as the coefficient of thermal expansion, elastic modulus, and Poisson's ratio are extracted. A theoretical calculation framework for the thermo-mechanical coupled deformation of the workbench is then constructed by combining this with the thermoelastic deformation mechanism. Simultaneously, global temperature field distribution data reconstructed from array infrared temperature measurement nodes is used to extract real-time temperature values ​​for each region of the workbench surface, calculate the transverse and longitudinal temperature gradient distribution, and locate the temperature difference amplitude and spatial distribution characteristics of high and low temperature regions. Subsequently, the global temperature field data is used as... Thermal load boundary conditions are applied to the coupled deformation model. Based on the basic equations of thermoelasticity and the equivalent finite element method, the superposition effects of frame support constraints and forming loads are considered simultaneously. The overall warping deformation of the workbench caused by the temperature gradient and the displacement deviation of each local area are solved step by step. The calculation results are then corrected and calibrated by combining historically calibrated measured thermal deformation data to eliminate parameter deviations between ideal model assumptions and actual working conditions. Finally, the vertical thermal deformation caused by temperature field inhomogeneity at the four leveling corners is accurately calculated, providing reliable quantitative thermal error data support for subsequent adaptive thermal creep compensation and dynamic correction of leveling accuracy.

[0038] This study predicts the thickness creep shrinkage of the billet during the holding pressure process based on the thermal creep curing shrinkage characteristics of composite materials. First, using the physical properties of the resin matrix and reinforcing fibers of the composite material as a foundation, the glass transition temperature, curing kinetic parameters, and viscoelastic creep constitutive coefficient of the matrix are extracted, along with material property parameters such as the elastic modulus, thermal expansion coefficient, and layup orientation of the reinforcing phase. Combining this with the coupling mechanism between curing shrinkage and viscoelastic creep under hot pressing conditions, a predictive theoretical model for thickness creep shrinkage of the billet is constructed. Simultaneously, real-time molding process data during the holding pressure stage is incorporated, including the billet heat distribution corresponding to the full-domain temperature field of the worktable, the molding holding pressure value, the holding pressure duration, and initial working condition parameters such as the initial thickness, layup structure, and pre-curing degree of the billet. This analysis elucidates the relationship between temperature, pressure, and curing properties. The dynamic evolution of the degree of curing with holding time was investigated. Subsequently, the real-time process and initial parameters were substituted into the creep-curing coupled constitutive equation, and the solution was obtained step by step based on the linear viscoelastic mechanics and curing kinetics theory. The constraint effect of fiber layup anisotropy on creep deformation was considered simultaneously, and the progressive deformation in the thickness direction of the billet during the entire holding period was calculated. Then, the prediction results were iteratively calibrated by combining the measured thickness data of the billet forming batches of the same process in history, correcting the prediction deviation caused by the dispersion of material parameters and process fluctuations. Finally, the thickness creep shrinkage of each region of the composite billet during the holding process was accurately predicted, providing a preliminary deformation prediction basis for the dynamic generation of subsequent leveling displacement compensation, and making up for the thickness uniformity deviation and leveling accuracy reduction caused by the progressive shrinkage of the billet during the holding stage.

[0039] In this embodiment, the process linkage calibration unit 500 includes a process parameter matching unit, a thickness deviation analysis unit, and a leveling reference micro-calibration unit.

[0040] The process parameter matching unit is used to retrieve the molding temperature, holding pressure, and curing stage process parameters of the current composite material blank, and match the corresponding leveling benchmark calibration strategy and allowable correction range; the thickness deviation analysis unit is used to analyze the thickness uniformity deviation of the corresponding area of ​​each corner point based on the blank molding thickness deviation data; the leveling benchmark micro-calibration unit is used to perform micro-calibration of the four corner leveling benchmarks in the holding pressure stage, combining process parameters and thickness deviation.

[0041] It is understood that, through the process parameter matching unit, the current composite material blank's molding temperature, holding pressure, and curing stage process parameters are retrieved, and the corresponding leveling benchmark calibration strategy and allowable correction range are matched to accurately adapt to the calibration requirements of different molding process stages. The correction range is reasonably limited to avoid over-adjustment and loss of control, providing a rule basis for process adaptation for the dynamic calibration of the leveling benchmark. The thickness deviation analysis unit analyzes the thickness uniformity deviation of the corresponding area of ​​each corner point based on the blank molding thickness deviation data, accurately locates the spatial distribution and deviation amplitude of the thickness unevenness, establishes a direct mapping relationship between molding quality and leveling action, and provides quantitative quality feedback support for benchmark micro-calibration. The leveling benchmark micro-calibration unit combines process parameters and thickness deviation during the holding pressure stage to perform micro-calibration on the four corner leveling benchmarks, dynamically corrects the benchmark offset caused by process evolution during curing, and makes up for the inherent limitation that fixed benchmarks are difficult to adapt to changes in the working conditions of the entire molding cycle. This provides a precise benchmark that dynamically adapts to the process for the four corner leveling control system, effectively reducing molding thickness deviation caused by process parameter fluctuations and curing shrinkage, and improving the process adaptability of leveling calibration and the uniformity of finished product thickness.

[0042] It should be noted that, based on historical molding process data and leveling calibration experience for different material systems and blanks of different specifications, the corresponding mapping relationship between molding temperature range, holding pressure level, curing reaction stage and leveling control logic is sorted out. A leveling benchmark calibration strategy library and correction range threshold table covering multiple working conditions are constructed, forming an initial matching system between process parameters and calibration rules. Simultaneously, process parameters such as blank molding temperature field data, system holding pressure value, and real-time monitoring results of curing degree are retrieved in real time during the current molding process to determine the current curing reaction stage (heating pre-impregnation, gel crosslinking, isothermal curing, cooling shaping) and process fluctuation status. Then, the current process parameters are matched with the working condition tags in the strategy library using feature matching and rule retrieval to match and adapt to the current material. The system employs a dedicated leveling benchmark calibration strategy for specific operating conditions and stages, clearly defining calibration triggering conditions, calibration frequency, and calibration adjustment direction. It also matches corresponding allowable correction ranges, defining the upper limit of single-step correction amplitude and the boundary of total correction. For thin-walled and high-precision parts, the range is narrowed to avoid over-adjustment and loss of control, while for thick-walled parts and large deformation conditions, the range is widened to ensure compensation effectiveness. Furthermore, the matching results are further verified and corrected by considering the hardware performance limits of the press equipment, cylinder stroke constraints, and safety control rules. The final output is a leveling benchmark calibration execution strategy and allowable correction range highly adapted to the current process conditions. This provides process-adaptive rules and boundary constraints for subsequent thickness deviation analysis and leveling benchmark micro-calibration, ensuring that the calibration process meets both molding accuracy requirements and process safety and equipment operation specifications.

[0043] Based on the analysis of the thickness uniformity deviation of the blank forming thickness deviation data, the spatial mapping rules between the four leveling corner points and the design nominal thickness, ply layout, and spatial correspondence of the four corner leveling corner points are established. The thickness detection zones and weight allocations covered by the control range of each corner point are clarified, forming a spatial correspondence analysis framework for thickness deviation analysis. Simultaneously, the raw thickness deviation data of the blank forming obtained from the laser thickness gauge array are input, and the difference between the real-time measured thickness value and the design reference thickness at each detection point is extracted. The raw data is then subjected to noise reduction, sliding filtering, and outlier removal to eliminate measurement errors caused by detection jitter, surface impurities, and local resin enrichment. Subsequently, the regions corresponding to the four corners are divided... The rules are as follows: The thickness deviation distribution characteristics within the corresponding area of ​​each corner point are statistically analyzed, and the average thickness deviation, extreme deviation, standard deviation, and gradient of each area are calculated. The spatial distribution pattern of uneven thickness is analyzed, the location and amplitude of areas with concentrated deviations are identified, and the different forms of overall thickness offset and local uniformity deviation are distinguished. The correlation weight between the deviation of each area and the leveling action at the corresponding corner point is simultaneously quantified. Then, combined with the current forming and curing stage, process parameter status, and the deviation evolution pattern of the same batch of blanks in history, the causes of deviations are initially identified and distinguished. Interference from uncontrollable leveling factors (such as layup overlap error and raw material dispersion) is eliminated, and invalid deviation signals are filtered out. Finally, the quantitative results and dynamic trends of the thickness uniformity deviation in the corresponding area of ​​each leveling corner point are accurately output.

[0044] This application proposes a four-corner leveling control system for a composite material press. Through a data sensing module, it comprehensively collects synchronous data on the displacement loads at the four corners of the press, the temperature field distribution data across the entire worktable, and the billet forming thickness deviation data, providing multi-dimensional foundational data for leveling control. A four-corner load decoupling analysis module, based on a statically indeterminate mechanical model, calculates the coupling interference of the four-corner loads, effectively separating the coupling effects caused by the elastic deformation of the frame and the installation errors of the hydraulic cylinders, overcoming the limitations of mutual interference between four-corner commands in traditional control. A multi-cylinder collaborative leveling execution module, utilizing an improved deep reinforcement learning leveling algorithm, achieves intelligent collaboration and rapid response among multiple hydraulic cylinders, overcoming the shortcomings of traditional PID control, such as difficulty in parameter tuning and weak anti-interference capability under complex off-center load conditions. A thermal creep adaptive compensation module integrates the thermo-mechanical coupling deformation law of the worktable with the thermal creep curing shrinkage characteristics of the composite material, dynamically generating leveling displacement compensation, effectively offsetting the interference of thermal deformation caused by temperature gradients on leveling accuracy. The process linkage calibration unit performs periodic micro-calibration of the four corner leveling benchmarks during the pressure holding stage based on the billet forming thickness deviation and current process parameters, achieving deep synergy between leveling control and forming process. This solves problems such as severe load coupling and lack of thermal deformation compensation in existing technologies.

[0045] The following will illustrate a four-corner leveling control system for a composite material press through a specific embodiment, including: In the hot press molding workshop of an aerospace composite materials manufacturing company, a 2,000-ton four-column hot press undertakes the batch molding of carbon fiber reinforced epoxy resin fuselage panel components for civil aircraft. With the increase in component size and the rise in aerospace-grade precision requirements, the traditional fixed-parameter open-loop leveling mode suffers from multiple factors: elastic coupling of corner loads, thermal deformation due to the temperature gradient of the worktable, and creep shrinkage during curing of the blank. This results in large deviations in the uniformity of the molding thickness, making it difficult to meet the leveling accuracy standards. To address this, a four-corner leveling control system was introduced to achieve high-precision closed-loop intelligent leveling control throughout the entire molding cycle.

[0046] The data sensing module constructs a multi-dimensional synchronous acquisition network. The corner displacement load acquisition unit deploys rope displacement sensors (accuracy 0.005mm) and spoke-type pressure sensors (accuracy 0.1 grade) at the four leveling cylinder support positions, with a sampling frequency of 50Hz and a timestamp error <10ms. The temperature field monitoring unit arranges 32 infrared temperature measurement nodes (0~300℃, error ±0.5℃) at 0.5m intervals on the lower surface of the workbench, with a sampling frequency of 1Hz, reconstructing the full-domain temperature field distribution through Kriging interpolation. The online thickness detection unit deploys a 12-set laser displacement sensor thickness measurement array (accuracy 0.002mm, sampling 2Hz) on the discharge side, comparing the deviation data with the nominal thickness. All data is transmitted to the edge control terminal via industrial Ethernet, filtered for noise reduction and anomaly removal, and then stored in the local database, providing a high-quality data foundation for subsequent calculations and decisions.

[0047] The four-corner load decoupling analysis module works collaboratively with statically indeterminate mechanical modeling and coupling interference resolution. Using the rigid body displacement method, the worktable is equivalent to a rigid body, and the elastic deformation of the frame and cylinders is equivalent to corner support springs. A stiffness balance equation with three degrees of freedom—vertical displacement and rotation around the X / Y axes—is established. The cylinder coaxiality error parameter is added to ensure model matching. Real-time four-corner loads are substituted into the solution, and matrix inversion separates the coupling effects. After eliminating coupling components, independent correction values ​​for each corner are output. The solution response is ≤50ms, the decoupling accuracy is ≥95%, and corner linkage interference is effectively eliminated.

[0048] The multi-cylinder collaborative leveling execution module consists of a deep reinforcement learning decision unit and a multi-cylinder servo drive control unit. The decision unit takes the decoupling correction amount and real-time displacement, load, and levelness error at the four corners as inputs. It employs an improved DDPG algorithm, with three fully connected hidden layers (128 neurons per layer) in both the policy network and the value network. It integrates a multi-objective reward function that considers levelness accuracy, load balance, and motion smoothness. After offline training on 5000 sets of historical data, the decision response is ≤100ms, representing a more than 30% improvement in dynamic response and a 40% reduction in overshoot compared to traditional PID. The servo drive unit adopts a position-pressure dual closed-loop structure, achieving a displacement accuracy of 0.01mm, a pressure accuracy of 0.5Bar, a four-cylinder synchronization error of <0.02mm, and stable control of the platform levelness within 0.02mm / m.

[0049] The thermal creep adaptive compensation and process linkage calibration modules operate in tandem. The thermal deformation calculation unit calculates the vertical thermal deformation at each corner point by combining the temperature field and thermo-mechanical coupling law (error ≤ 5%); the creep prediction unit predicts the thickness creep shrinkage during the holding pressure stage based on the curing kinetics and viscoelastic constitutive model (accuracy ≥ 90%); the compensation amount generation unit integrates the two and outputs the leveling displacement compensation amount through a weighted algorithm, which is then superimposed on the control command in real time. The process linkage calibration unit matches parameters such as molding temperature, holding pressure, and curing stage in real time, retrieves the corresponding calibration strategy from the strategy library, and triggers a micron-level micro-calibration of the four corner references every 10 minutes during the holding pressure stage. Combined with thickness deviation analysis, it distinguishes between controllable and inherent deviations, filters invalid signals, and ensures that the leveling reference is adapted to the process state throughout the entire cycle.

[0050] In summary, this application's embodiments utilize a data sensing module to build a multi-dimensional synchronous acquisition network, providing high-precision, highly synchronous raw data across all operating conditions, laying a reliable data foundation for subsequent calculations and decisions. The four-corner load decoupling analysis module, leveraging a statically indeterminate mechanical model, accurately separates elastic coupling interference, eliminating leveling distortion caused by corner linkage and outputting independent and controllable correction criteria. The multi-cylinder collaborative leveling execution module optimizes decision-making efficiency through an improved deep reinforcement learning algorithm, and, in conjunction with a dual-closed-loop servo drive, achieves high-precision synchronous control of the four cylinders, significantly improving the dynamic response speed of leveling, reducing control overshoot, and stabilizing the table level. The thermal creep adaptive compensation module integrates thermal deformation calculation and creep prediction results to dynamically generate compensation amounts, offsetting the dual interference of temperature gradient and curing shrinkage on leveling accuracy and compensating for accuracy decay under long-cycle pressure holding conditions. The process linkage calibration unit combines molding process parameters with a phased matching calibration strategy to periodically perform benchmark micro-calibration, adapting to the changes in working conditions throughout the entire molding cycle and effectively improving the uniformity of component thickness. At the same time, it avoids the limitations of traditional open-loop leveling parameters being fixed and having poor working condition adaptability, reducing the workload of repeated manual adjustments and subjective errors, improving the finished product qualification rate, stabilizing molding quality, and reducing overall production costs.

[0051] Next, referring to the accompanying drawings, a method for controlling the leveling of the four corners of a composite material press according to an embodiment of this application is described.

[0052] like Figure 2 As shown, the four-corner leveling control method for a composite material press includes the following steps: In step S101, the displacement-load synchronization data of the four corner points of the press, the temperature field distribution data of the entire worktable surface, and the molding thickness deviation data of the composite material blank are obtained.

[0053] Among them, displacement-load synchronous data refers to a set of monitoring data obtained through synchronous acquisition technology, in which the displacement at the leveling corner point corresponds one-to-one with the bearing pressure value at the same moment. It is the basic input data for carrying out load coupling calculation and leveling control.

[0054] It is understood that by acquiring the displacement-load synchronization data of the four corner points of the press, the embodiments of this application can accurately grasp the force-position linkage state and load distribution law of each leveling corner point, capture the elastic coupling disturbance of the four corner loads, the dynamic change of displacement deviation and the signs of off-center load imbalance in real time, and provide accurate and synchronized input basis for subsequent statically indeterminate mechanical model solving, load coupling interference calculation, leveling control correction calculation and multi-cylinder collaborative drive control, and force-position collaborative control and error correction in the leveling process in advance, reduce leveling distortion, table tilt and forming thickness deviation caused by coupling interference, and improve leveling control accuracy and system response speed.

[0055] In step S102, based on the displacement-load synchronous data, the displacement-load coupling relationship at the four corners is decoupled and calculated using a statically indeterminate mechanical model, and the leveling control correction amount for decoupling the control coupling at each corner point is output.

[0056] Among them, the statically indeterminate mechanical model is constructed for the four-corner support system of the press. Since the number of unknown constraint forces is greater than the number of static equilibrium equations, it is necessary to introduce deformation compatibility conditions and material elastic constitutive relations to solve the load distribution and deformation law simultaneously, so as to support the mechanical analysis model for solving load coupling interference.

[0057] It is understood that the embodiments of this application use a statically indeterminate mechanical model to decouple the displacement-load coupling relationship at the four corners, characterize the force flow transmission law and elastic linkage characteristics of the four corner support system, separate the coupling interference components caused by the elastic deformation of the frame and the coaxiality error of the cylinder installation in real time, clarify the linkage effect of the load displacement change of each corner point on the other corner points, provide a reliable mechanical calculation basis for the subsequent independent leveling control correction output of each corner point and the precise control of multi-cylinder collaborative leveling, quantitatively strip and precisely correct the coupling interference of the four corner loads, reduce the leveling distortion, control overshoot and table tilt deviation caused by corner linkage, and improve the accuracy of load calculation and leveling control precision.

[0058] In step S103, based on the decoupled leveling control correction amount and combined with the improved deep reinforcement learning leveling algorithm, the operating status of the four leveling cylinders is synchronously controlled with displacement feed as the core, and the output pressure is subjected to overload limiting constraint. Based on the temperature field distribution data of the entire worktable surface and combined with the thermo-mechanical coupling deformation law of the worktable structure, the leveling displacement compensation amount of each corner point is dynamically generated. At the same time, the thermal creep curing shrinkage characteristics of composite materials are incorporated into the process parameter dimension for adaptation correction.

[0059] Among them, the improved deep reinforcement learning leveling algorithm is based on the deep deterministic policy gradient algorithm. It is designed for the continuous control scenario of leveling the four corners of a composite material press. It optimizes the multi-objective reward function and introduces a target network soft update mechanism. It is an intelligent leveling control algorithm that uses the decoupled leveling control correction as input to find the optimal control command for each cylinder.

[0060] It is understood that the embodiments of this application utilize an improved deep reinforcement learning leveling algorithm to process the decoupled leveling control correction amount. With displacement feed as the core, it adaptively adapts to the leveling requirements of different molding loads, off-center loading conditions, and thermal deformation disturbances. It optimizes the output of the optimal control commands for each leveling cylinder in real time, synchronously regulates the operating status of the four cylinders, and imposes overload limiting constraints on the output pressure. At the same time, it can adapt to the compensation correction dimensions of thermo-coupling deformation and thermal creep curing shrinkage characteristics of composite materials. It effectively avoids the inherent limitations of traditional fixed parameter control algorithms, such as response lag, overshoot oscillation, and poor adaptability to complex working conditions. It provides intelligent decision support for the synchronous and precise control and dynamic compensation of the four corner cylinders, and reduces the risks of control overshoot, response lag, and overload by performing multi-objective dynamic optimization and full-condition collaborative management of the leveling process. It improves the dynamic response speed, steady-state control accuracy, and adaptive capability of the leveling system under complex working conditions, and ensures the leveling stability and molding quality consistency of the composite material throughout the entire molding cycle.

[0061] In step S104, based on the molding thickness deviation data of the composite material blank and the current molding process parameters and leveling displacement compensation, the interpretable deviation components corresponding to thermal deformation compensation and material thermal creep curing shrinkage are stripped, the residual error of the leveling control link is extracted, and the four corner leveling benchmarks are iteratively micro-calibrated.

[0062] Among them, the leveling displacement compensation amount is a corner displacement correction value that is dynamically generated and superimposed on the basic leveling control amount in order to offset the interference of factors such as thermal deformation due to temperature gradient of the workbench and creep shrinkage of composite material blanks during curing on the leveling accuracy.

[0063] It is understood that in this embodiment, the leveling displacement compensation is superimposed on the basic leveling control amount to participate in the hydraulic cylinder regulation, so as to offset the warping thermal deformation caused by the temperature gradient of the worktable and the dimensional deviation caused by the thermal creep curing shrinkage during the pressure holding stage of the composite material blank in advance. The interpretable deviation components corresponding to the two types of deformation are separated, which makes it easier to extract the residual error in the leveling control link from the molding thickness deviation. This provides an accurate error distinction basis for the iterative micro-calibration of the four corner leveling benchmark, avoids the continuous offset of the leveling benchmark caused by thermal deformation and creep deformation, reduces the problem of continuous decay of molding thickness uniformity accuracy under long-cycle high temperature pressure holding conditions, reduces the accumulation of residual error in the leveling system, and improves the stability of the leveling benchmark and the molding dimensional accuracy of the component throughout the curing cycle.

[0064] According to the embodiments of this application, a four-corner leveling control method for a composite material press is proposed. A data sensing module comprehensively collects synchronous data of the four-corner displacement loads, temperature field distribution data across the entire worktable surface, and billet forming thickness deviation data, providing multi-dimensional basic data for leveling control. A four-corner load decoupling analysis module calculates the coupling interference of the four-corner loads based on a statically indeterminate mechanical model, effectively separating the coupling effects caused by the elastic deformation of the frame and the installation error of the hydraulic cylinders, overcoming the limitations of mutual interference between four-corner commands in traditional control. A multi-cylinder collaborative leveling execution module, utilizing an improved deep reinforcement learning leveling algorithm, achieves intelligent collaboration and rapid response among multiple hydraulic cylinders, overcoming the shortcomings of traditional PID control, such as difficulty in parameter tuning and weak anti-interference capability under complex off-center load conditions. A thermal creep adaptive compensation module integrates the thermo-mechanical coupling deformation law of the worktable with the thermal creep curing shrinkage characteristics of the composite material, dynamically generating leveling displacement compensation, effectively offsetting the interference of thermal deformation caused by temperature gradients on leveling accuracy. The process linkage calibration unit performs periodic micro-calibration of the four corner leveling benchmarks during the pressure holding stage based on the billet forming thickness deviation and current process parameters, achieving deep synergy between leveling control and forming process. This solves problems such as severe load coupling and lack of thermal deformation compensation in existing technologies.

[0065] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: The memory 301, the processor 302, and the computer program stored on the memory 301 and capable of running on the processor 302.

[0066] When the processor 302 executes the program, it implements a four-corner leveling control method for a composite material press provided in the above embodiments.

[0067] Furthermore, electronic devices also include: Communication interface 303 is used for communication between memory 301 and processor 302.

[0068] The memory 301 is used to store computer programs that can run on the processor 302.

[0069] The memory 301 may include high-speed RAM (Random Access Memory) and may also include non-volatile memory, such as at least one disk storage.

[0070] If the memory 301, processor 302, and communication interface 303 are implemented independently, then the communication interface 303, memory 301, and processor 302 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0071] Optionally, in a specific implementation, if the memory 301, processor 302, and communication interface 303 are integrated on a single chip, then the memory 301, processor 302, and communication interface 303 can communicate with each other through an internal interface.

[0072] Processor 302 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of this application.

[0073] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for leveling the four corners of a composite material press.

[0074] Furthermore, this application also provides a computer program product, including a computer program or instructions, which, when executed, implement the aforementioned method for leveling the four corners of a composite material press.

[0075] In the description of this specification, the references to "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0076] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0077] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0078] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0079] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0080] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A four-corner leveling control system for a composite material press, characterized in that, include: The system comprises a data sensing module, a four-corner load decoupling analysis module, a multi-cylinder collaborative leveling execution module, a thermal creep adaptive compensation module, and a process linkage calibration unit; among which: The data sensing module is used to collect real-time displacement load synchronous data of the four corner points of the press, temperature field distribution data of the entire worktable surface, and molding thickness deviation data of the composite material blank. The four-corner load decoupling analysis module is used to perform coupling interference calculation on the four-corner loads based on the statically indeterminate mechanical model, and outputs the leveling control correction amount that is independent of each corner point. The multi-cylinder collaborative leveling execution module is used to simultaneously regulate the output pressure and displacement feed of the four leveling cylinders based on the decoupled leveling control correction amount and the improved deep reinforcement learning leveling algorithm. The thermal creep adaptive compensation module is used to dynamically generate the leveling displacement compensation amount of each corner point based on the temperature field distribution data of the entire worktable surface, combined with the thermo-mechanical coupling deformation law of the worktable structure and the thermal creep curing shrinkage characteristics of the composite material, so as to offset the interference of the thermal deformation of the worktable caused by the temperature gradient on the leveling accuracy. The process linkage calibration unit is used to periodically perform micro-calibration on the four corner leveling references during the pressure holding stage, based on the molding thickness deviation data of the composite material blank and the current molding process parameters.

2. The four-corner leveling control system for a composite material press according to claim 1, characterized in that, The data sensing module includes a corner displacement load synchronous acquisition unit, a workbench global temperature field monitoring unit, and a billet forming thickness online detection unit. The corner displacement load synchronous acquisition unit uses high-precision displacement and pressure sensors to synchronously acquire real-time displacement and load data at four leveling corner points. The workbench global temperature field monitoring unit uses an array of infrared temperature measurement nodes to acquire multi-point temperatures on the workbench surface, obtaining global temperature field distribution data. The billet forming thickness online detection unit uses a laser thickness measurement array to acquire the thickness of the composite material billet in multiple regions and calculates the forming thickness deviation data.

3. The four-corner leveling control system for a composite material press according to claim 1, characterized in that, The four-corner load decoupling analysis module includes a statically indeterminate mechanical modeling unit and a coupling interference calculation unit. The statically indeterminate mechanical modeling unit is used to construct a four-corner support statically indeterminate mechanical analysis model based on the press frame structural parameters and the cylinder arrangement. The coupling interference calculation unit is used to calculate the elastic coupling interference of the four-corner load based on the statically indeterminate mechanical model, separate the coupling effect caused by the elastic deformation of the frame and the coaxiality error of the cylinder installation, and output the independent leveling control correction amount of each corner point.

4. The four-corner leveling control system for a composite material press according to claim 1, characterized in that, The multi-cylinder collaborative leveling execution module includes a deep reinforcement learning decision unit and a multi-cylinder servo drive control unit. The deep reinforcement learning decision unit is used to take the decoupled leveling control correction amount as input and optimize the output control command of each cylinder through an improved deep reinforcement learning leveling algorithm. The multi-cylinder servo drive control unit is used to synchronously adjust the output pressure and displacement feed of the four leveling cylinders according to the control command, and adjust the levelness of the worktable in real time.

5. The four-corner leveling control system for a composite material press according to claim 1, characterized in that, The adaptive thermal creep compensation module includes a workbench thermal deformation calculation unit, a billet creep prediction unit, and a compensation amount dynamic generation unit. The workbench thermal deformation calculation unit combines global temperature field distribution data with the thermo-mechanical coupling deformation law of the workbench structure to calculate the thermal deformation at each corner point caused by the temperature gradient. The billet creep prediction unit predicts the thickness creep shrinkage of the billet during the pressure holding process based on the thermal creep curing shrinkage characteristics of the composite material. The compensation amount dynamic generation unit integrates the workbench thermal deformation and billet creep to dynamically generate the leveling displacement compensation amount at each corner point, offsetting the interference of thermal deformation on leveling accuracy.

6. The four-corner leveling control system for a composite material press according to claim 1, characterized in that, The process linkage calibration unit includes a process parameter matching unit, a thickness deviation analysis unit, and a leveling benchmark micro-calibration unit. The process parameter matching unit is used to retrieve the molding temperature, holding pressure, and curing stage process parameters of the current composite material blank, and match the corresponding leveling benchmark calibration strategy and allowable correction range. The thickness deviation analysis unit is used to analyze the thickness uniformity deviation of the corresponding area of ​​each corner point based on the blank molding thickness deviation data. The leveling benchmark micro-calibration unit is used to perform micro-calibration of the four corner leveling benchmarks during the holding pressure stage, combining process parameters and thickness deviation.

7. A method for applying a four-corner leveling control system for a composite material press according to any one of claims 1-6, characterized in that, The method includes: Acquire displacement-load synchronization data at the four corners of the press, temperature field distribution data across the entire worktable, and molding thickness deviation data of the composite material blank; Based on the displacement-load synchronization data, the displacement-load coupling relationship at the four corners is decoupled using a statically indeterminate mechanical model, and the leveling control correction amount for decoupling the control coupling at each corner point is output. Based on the decoupled leveling control correction amount, combined with the improved deep reinforcement learning leveling algorithm, with displacement feed as the core, the operating status of the four leveling cylinders is synchronously controlled and the output pressure is subjected to overload limiting constraint. Based on the temperature field distribution data of the entire worktable surface, combined with the thermo-mechanical coupling deformation law of the worktable structure, the leveling displacement compensation amount of each corner point is dynamically generated. At the same time, the thermal creep curing shrinkage characteristics of composite materials are incorporated into the process parameter dimension for adaptation and correction. Based on the molding thickness deviation data of the composite material blank, combined with the current molding process parameters and the leveling displacement compensation amount, the interpretable deviation components corresponding to thermal deformation compensation and material thermal creep curing shrinkage are stripped off, the residual error of the leveling control link is extracted, and the four corner leveling benchmarks are iteratively micro-calibrated.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the program to implement the four-corner leveling control method for a composite material press as claimed in claim 7.

9. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When a computer program or instruction is executed, it implements the four-corner leveling control method for a composite material press as claimed in claim 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When a computer program or instruction is executed, it implements the four-corner leveling control method for a composite material press as claimed in claim 7.

Images