A method for predicting temperature field of hot-pressing core layer of plywood

Abstract

The application provides a plywood hot pressing core layer temperature field prediction method, and relates to the technical field of data processing. The method performs time series analysis on the heating plate temperature, the plate blank surface layer temperature, the hot pressing pressure and the initial physical property parameters of the plate blank at different time points in the hot pressing process, and combines a heat and moisture coupled heat transfer equation and a recursive parameter correction algorithm to perform dynamic modeling, and then judges the real-time state and water-heat evolution trend of the core layer cross section temperature field. In this way, the accuracy and timeliness of the core layer temperature prediction can be improved without damaging the structure of the plate blank, which helps to early identify the risk of insufficient solidification or overheating degradation caused by physical property batch fluctuations, so as to timely adjust the heating curve and process parameters, reduce the product rejection rate and improve the batch quality consistency.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a method for predicting the temperature field of the hot-pressed core layer of plywood. Background Technology

[0002] As one of the leading products in the engineered wood products industry, the hot-pressing process in plywood manufacturing directly determines the final product's bonding strength, dimensional stability, and internal quality uniformity. The hot-pressing process is essentially a complex physicochemical process involving the coupling of heat transfer, mass transfer, and chemical reactions. The evolution of the core layer temperature directly affects the adhesive curing process, the internal stress distribution of the board, and energy efficiency. With increasingly scarce timber resources and ever-increasing product quality requirements, precise control of the hot-pressing process has become a core issue for technological upgrading in the plywood industry. Traditional hot-pressing processes mainly rely on empirical "temperature-time-pressure" curves, indirectly controlling the internal state of the board by adjusting the temperature and pressure of the heating plate. However, this open-loop control mode lacks direct perception of the true temperature state of the core layer. In recent years, with the development of computational fluid dynamics, reverse heat transfer theory, and data assimilation technology, intelligent monitoring methods for the hot-pressing process based on the fusion of numerical simulation and online measurement have gradually become a research hotspot, providing a new technical path for real-time prediction and dynamic control of the core layer temperature.

[0003] However, existing techniques for predicting the temperature field of plywood hot pressing still have significant limitations. First, traditional finite element or finite difference simulation methods heavily rely on the accurate input of the slab's physical properties. Wood, as a natural heterogeneous material, exhibits significant batch-to-batch fluctuations in its thermal properties, such as thermal conductivity and specific heat capacity, influenced by factors like tree species, moisture content, density, and fiber orientation. Fixed-parameter models struggle to adapt to the raw material variability in actual production, leading to a cumulative and amplified error in core layer temperature prediction. Second, the contact thermal resistance between the heating plate and the slab surface changes dynamically with pressure. Existing methods often employ simplified boundary conditions (such as directly assigning the heating plate temperature to the slab surface temperature), neglecting the time-varying characteristics of interfacial heat transfer resistance, causing the initial temperature field calculation to deviate from the actual state. Furthermore, while some studies have attempted to introduce surface temperature sensors as feedback signals, they mostly employ static parameter correction or offline optimization strategies, failing to achieve dynamic recursive correction of heat transfer parameters during hot pressing, thus failing to meet the real-time and robustness requirements of industrial settings. More importantly, existing prediction methods generally output core temperature as a purely positive calculation result, lacking a closed-loop correction mechanism based on observation data. When there are deviations in model parameters or boundary conditions, the prediction accuracy degrades rapidly, failing to provide a reliable basis for the dynamic optimization of hot-pressing curves. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method for predicting the temperature field of the core layer in hot-pressed plywood. This method, to a certain extent, solves the problem that existing hot-pressing processes suffer from unmeasurable core layer temperatures and large batch-to-batch fluctuations in veneer properties, resulting in yield fluctuations exceeding 15% due to the use of fixed heating curves.

[0005] According to one aspect of the present invention, a method for predicting the temperature field of a hot-pressed core layer of plywood is provided, comprising:

[0006] Acquire heating plate temperature data, slab surface temperature data, hot pressing pressure data, and initial physical property parameters of the slab during the hot pressing process, and align them according to the same hot pressing time axis to construct a multi-source input matrix;

[0007] Based on the hot pressing pressure data and heating plate temperature data, the interfacial heat flow boundary conditions of the slab surface are determined. Several spatial grid nodes are divided along the slab thickness direction. The preliminary temperature response of each grid node is calculated based on the heat transfer equation, and a preliminary temperature field distribution matrix is ​​generated.

[0008] The observation residual is formed by the measured surface temperature of the slab at the current moment and the calculated temperature of the corresponding surface node in the preliminary temperature field distribution matrix. The heat transfer parameters are corrected by the parameter recursion correction algorithm, and the temperature field is updated based on the corrected heat transfer parameters to obtain the corrected temperature field distribution matrix.

[0009] The temperature values ​​of the corresponding grid nodes in the core region are extracted from the corrected temperature field distribution matrix to generate the core temperature gradient curve or the core cross-sectional temperature field, and the core temperature field prediction result is output.

[0010] Furthermore, the interfacial heat flux boundary conditions include the contact heat conduction flux determined by the pressure contact state and the phase change-assisted heat transfer flux determined by the outward migration and condensation of water vapor. The interfacial heat flux boundary conditions are determined according to the following relationship:

[0011]

[0012] in, The heat flux density entering the slab surface, For contact heat transfer coefficient, The surface temperature of the heating plate. The temperature is either the temperature corrected at the previous time step or the temperature iterated at the current time step. The latent heat of condensation distribution coefficient, For latent heat of vaporization, This refers to the mass flux of water vapor migrating from the interior of the slab to the surface interface.

[0013] Furthermore, the interfacial heat flow boundary conditions of the slab surface are determined by determining the contact heat transfer coefficient based on the hot-pressing pressure data to quantify the interfacial thermal resistance. The contact heat transfer flux is calculated by multiplying the contact heat transfer coefficient by the difference between the heating plate temperature and the slab surface temperature, and the phase change-assisted heat transfer flux formed by the outward migration and condensation of water vapor is superimposed to jointly determine the interfacial heat flow boundary conditions of the slab surface.

[0014] Furthermore, the thermal conductivity in the thickness direction corresponding to the spatial grid node is determined based on the texture direction, moisture content, and temperature of the single-layer plate;

[0015] When the material calibration data includes longitudinal thermal conductivity, radial thermal conductivity, and chordal thermal conductivity, the anisotropic thermal conductivity tensor is projected onto the thickness direction of the slab to obtain the basic thermal conductivity of each slab layer in the thickness direction. The interlayer thermal resistance between adjacent grid nodes is then determined based on the adhesive layer thickness, adhesive thermal conductivity, and degree of curing.

[0016] Furthermore, the heat transfer equation includes moisture migration flux, local moisture evaporation rate, moisture content time-progression term, and temperature time-progression term, wherein the moisture content time-progression term is determined according to the following relationship:

[0017]

[0018] in, This represents the initial water content at the next time point. This represents the water content after correction at the previous time point. and This represents the water mass flux at the adjacent interface. The density of absolutely dry wood. To control the thickness of the body, This represents the localized rate of water evaporation.

[0019] Furthermore, the heat transfer equation is solved using a sequential operator splitting method, which includes:

[0020] First, solve the water content transport equation based on the temperature field of the previous moment or the current iteration, and then update the water content field and water vapor outflow flux.

[0021] Then, based on the updated moisture content field, the thermal conductivity, moisture evaporation rate, interlayer thermal resistance, and interfacial heat flow boundary are calculated; finally, the temperature field is implicitly solved.

[0022] Furthermore, the heat transfer parameters are corrected based on an ensemble Kalman filter framework. The direction and magnitude of the correction are determined by the ensemble covariance between the surface temperature observation residual and the heat transfer parameters. An adaptive observation noise matrix is ​​constructed by subtracting the observation increment residual after the predicted heating trend to suppress the initial heating interference of hot pressing. At the same time, the corrected parameters are constrained within the physical operating boundary of the material calibration to realize the online recursive update of the heat transfer parameters.

[0023] According to another aspect of the present invention, a system for predicting the temperature field of a hot-pressed core layer of plywood is provided, comprising:

[0024] The data acquisition module is used to acquire heating plate temperature data, slab surface temperature data, hot pressing pressure data, and initial physical property parameters of the slab during the hot pressing process, and align them according to the same hot pressing time axis to construct a multi-source input matrix;

[0025] The forward thermal-humidity coupling solution module is used to determine the interfacial heat flow boundary conditions on the slab surface based on the hot-pressing pressure data and the heating plate temperature data, divide a number of spatial grid nodes along the slab thickness direction, calculate the preliminary temperature response of each grid node based on the heat transfer equation, and generate a preliminary temperature field distribution matrix.

[0026] The recursive correction module is used to form an observation residual based on the measured surface temperature of the slab at the current moment and the calculated temperature of the corresponding surface node in the preliminary temperature field distribution matrix. The module then uses a parameter recursive correction algorithm to correct the heat transfer parameters and updates the temperature field based on the corrected heat transfer parameters to obtain the corrected temperature field distribution matrix.

[0027] The core temperature extraction and output module is used to extract the temperature values ​​of the corresponding grid nodes in the core region from the corrected temperature field distribution matrix, generate the core temperature gradient curve or the core cross-sectional temperature field, and output the core temperature field prediction result.

[0028] Compared with existing technologies, the plywood hot-pressing core layer temperature field prediction method provided by this invention performs time-series analysis on the heating plate temperature, slab surface temperature, hot-pressing pressure, and initial physical property parameters of the slab at different time points during the hot-pressing process. It then combines this with dynamic modeling using the thermo-humidity coupling heat transfer equation and a recursive parameter correction algorithm to determine the real-time state and hydrothermal evolution trend of the core layer cross-section temperature field. This improves the accuracy and timeliness of core layer temperature prediction without damaging the slab structure, helping to identify early risks of insufficient curing or overheating degradation caused by batch fluctuations in physical properties. This allows for timely adjustment of heating curves and process parameters, reducing product scrap rates and improving batch quality consistency. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0030] Figure 1 This is a flowchart of a method for predicting the temperature field of the hot-pressed core layer of plywood according to an embodiment of the present invention.

[0031] Figure 2 This diagram illustrates an application scenario of the plywood hot-pressed core layer temperature field prediction method according to an embodiment of the present invention. Detailed Implementation

[0032] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0033] As mentioned in the background section, existing hot-pressing processes use fixed heating durations based on surface or heating plate temperatures, making it impossible to monitor the real-time temperature evolution of the core layer. Since the core layer temperature cannot be directly measured (embedding thermocouples would damage the slab structure and only provide single-point data), and the physical property fluctuations between different batches of slabs (density deviation ±8%, moisture content deviation ±3%) lead to significant differences in the core layer temperature field distribution, using a uniform heating curve results in insufficient curing in some batches due to slow core layer heating, and degradation in others due to overheating, with yield fluctuations exceeding 15%. The core challenge lies in how to accurately predict the real-time temperature distribution of the entire core layer cross-section without damaging the slab structure or relying on single-point temperature measurements, based on measurable surface temperatures and process parameters, thereby guiding the dynamic adjustment of the heating curve.

[0034] Figure 1 This is a flowchart of a method for predicting the temperature field of a hot-pressed core layer of plywood according to an embodiment of the present invention. Figure 1 and Figure 2 As shown, the method for predicting the temperature field of the hot-pressed core layer of plywood includes:

[0035] S1. The hot-pressing temperature field prediction device first acquires the external working condition data, surface thermal response data, and initial physical property data of the slab that can be measured non-destructively during the hot-pressing process. Then, it aligns the heating plate surface temperature sequence, the measured surface temperature sequence of the slab, the hot-pressing pressure sequence, the initial moisture content of the slab, the density of the slab, the texture direction of the slab, and the structural parameters of the slab layers according to the same hot-pressing time axis to form a multi-source input matrix for interface heat flow boundary conversion, heat and moisture coupling heat transfer solution, and online correction of thermal conductivity parameters.

[0036] Specifically, the system reads the start and end signals of the hot pressing cycle from the hot press controller, taking the mold closing signal, the pressure rise start signal, or the hot pressing program start signal as the start point of the hot pressing time axis, and the pressure release signal, the mold opening signal, or the hot pressing program end signal as the end point of the hot pressing time axis. Subsequently, it reads the surface temperature sequences of the upper and lower heating plates. These temperature sequences are obtained by the temperature sensors built into the heating plates, the temperature control sampling port of the hot press, or a temperature acquisition unit attached to the surface of the heating plates. Each sampled value is stored corresponding to a sampling timestamp. If the hot press only collects the temperature of one side of the heating plate, the system records the surface temperature sequence of that side of the heating plate and processes it in subsequent boundary value determinations using either unilateral heating or symmetrical heating methods according to the actual equipment structure.

[0037] The system then reads the measured temperature sequence of the slab surface. This measured temperature sequence is obtained by an infrared temperature measurement unit, a thin-film attached temperature sensor, or a non-embedded temperature measurement unit located near the pressure plate. This unit only acquires the temperature of the slab's outer surface, does not insert into the core layer, and does not penetrate the veneer layer or adhesive layer. Therefore, it does not change the interlayer contact state within the slab, nor does it form a localized foreign heat conduction channel in the core layer. If temperature measurement channels are provided on both the upper and lower surfaces, the system generates measured temperature sequences for the upper and lower surfaces respectively. If only one side has a surface temperature measurement channel, the system uses the surface temperature of that side as the effective observation, and only uses the observation residual from that side when correcting subsequent thermal conductivity parameters.

[0038] Furthermore, the system reads the hot-pressing pressure sequence from the pressure acquisition unit. This hot-pressing pressure sequence is preferably the actual feedback pressure of the hot press, rather than the target pressure in the process formula. If the hot press outputs hydraulic cylinder pressure or the total press pressure, the system converts it into unit area pressure acting on the slab surface based on the effective pressure-bearing area of ​​the pressure plate, and then uniformly converts it to Pa. For the pressure-increasing, pressure-holding, and pressure-reducing sections, the system retains the complete process of pressure change over time, ensuring that subsequent interface thermal resistance calculations can reflect the actual state of the contact between the heating plate and the slab surface as pressure is applied, held, and released.

[0039] The system then reads the initial moisture content, density, grain direction, and adhesive layer structure parameters of the veneer. The initial moisture content can be obtained from a pre-laying moisture content testing device, a pre-pressing sampling testing device, or batch testing records from the production line. If the test results are expressed as a percentage, the system converts them into a dimensionless form of mass moisture content. For multi-layer plywood, the system stores the initial moisture content, density, and grain direction of each veneer layer along with its layer number in the board thickness direction. If only the average moisture content of the same batch or the same board can be obtained at the production site, the system assigns this average moisture content as the initial physical property value for each veneer layer and records the dispersion of the sampled moisture content and density of the same batch as a source of uncertainty for subsequent online correction of thermal conductivity parameters.

[0040] Once the temperature, pressure, moisture content, and slab structure data are prepared, the system establishes a unified hot pressing time node. For data with a sampling frequency higher than a unified time node, the system extracts representative values ​​corrected by the sensor calibration curve within the corresponding time interval. For data with a sampling frequency lower than a unified time node, the system interpolates based on the time positions of adjacent sampling points, preserving the physical order of the continuous evolution of temperature, pressure, and water content over time. The sensor calibration curve is determined by the sensor's factory calibration data, on-site comparison calibration records, or periodic calibration records of the hot-pressing equipment. If there is a short-term sampling gap at a certain time point, the system fills in the gap based on the data evolution trend of adjacent time points and generates a data quality identifier for that data point, enabling it to participate in calculations according to its reliability in subsequent observation corrections.

[0041] The system then writes the data corresponding to each hot-pressing time node into a multi-source input matrix. For double-sided hot pressing where temperature can be measured on both sides, each row of the multi-source input matrix includes at least the current surface temperature of the upper heating plate, the surface temperature of the lower heating plate, the measured temperature of the upper surface layer, the measured temperature of the lower surface layer, the hot-pressing pressure value, the initial moisture content of each veneer layer, the density of each veneer layer, and the texture direction identifier of each layer. For single-sided surface temperature measurement, the multi-source input matrix retains the surface temperature of the effective measurement side and marks the other side as unobserved. Since the initial moisture content and density of the veneer are usually not directly detected in real time within a hot-pressing cycle, they can be repeatedly written at each time node, or stored separately as a layer parameter matrix bound to that hot-pressing cycle, as long as subsequent calculations can read the initial physical property data corresponding to the thickness direction position at any time node.

[0042] On the other hand, the system also associates and stores the total thickness of the slab, the number of veneer layers, the thickness of each veneer layer, the location and thickness of the adhesive layer, the type of adhesive, the initial temperature of the slab before pressing, the calibration data of the wood's cross-grain and parallel-grain thermal conductivity, the calibration data of moisture content-temperature-diffusion coefficient, and the calibration data of adhesive layer thermal conductivity-curing degree with the multi-source input matrix. If the slab is not preheated before pressing, the system can use the surface temperature before pressing as the initial temperature of each node in the thickness direction; if the slab has been preheated, the system forms an initial temperature distribution based on the preheating equipment records, preheating time, and the surface temperature before pressing. If the hot press arranges multiple surface temperature measurement points along the width or length of the slab, the system further establishes a data structure including time dimension, variable dimension, and measurement area dimension, so that each measurement area has corresponding heating plate temperature, surface temperature, pressure input, and layer property parameters. At this point, the system completes the construction of the multi-source input matrix and sends it to S2.

[0043] S2. The hot-pressing temperature field prediction device calculates the pressure-water content-phase change coupling interface heat flow boundary between the slab and the heating plate based on the hot-pressing pressure sequence, the heating plate surface temperature sequence, the slab surface node temperature, the surface water content state, and the water vapor migration state. Subsequently, the system divides the slab thickness direction into several spatial grid nodes and calculates the transient temperature response of each grid node based on the thermo-humidity coupling discrete heat transfer equation that includes interlayer thermal resistance, water evaporation heat absorption, water migration conservation, and water vapor migration heat carrying terms, generating a preliminary temperature field distribution matrix.

[0044] Specifically, the system first reads the current hot-pressing pressure value, upper heating plate surface temperature, lower heating plate surface temperature, surface slab moisture content, slab surface node temperature, material calibration parameters, and water vapor migration flux obtained from the previous time step or the current iteration, all output by S1. For the upper and lower heating surfaces, the system calculates the contact heat transfer flux and the phase change-assisted heat transfer flux, respectively. The contact heat transfer flux reflects the heat conduction channel formed between the heating plate and the slab surface by roughness, air gap closure, pressure compression, surface compression layer, and isolation film; the phase change-assisted heat transfer flux reflects the additional heat transfer channel formed by the migration of moisture from the slab's interior to the surface after vaporization and the release of latent heat through condensation on the surface or near the interface.

[0045] In one implementation, the system uses the following relationship to determine the interfacial heat flux boundary:

[0046]

[0047]

[0048]

[0049] in, Indicates the upper heating surface or the lower heating surface. Indicates the hot pressing time point; The pressure contact saturation factor is dimensionless. This is the current hot-pressing pressure, in Pa. The contact calibration pressure corresponding to the heating surface, in Pa, is obtained from the contact heat transfer calibration test; The surface water content is dimensionless and is used for the current boundary calculation. It is either the corrected water content of the previous time step or the water content of the previous iteration in the current time step. The moisture content benchmark for stable production batches of the same material species; To stabilize the batch moisture content distribution scale; The contact heat transfer coefficient is expressed in W / (m²·K). The equivalent thermal resistance of the air gap when not fully compressed is expressed in m²·K / W. , , All are dimensionless calibration coefficients; The equivalent thickness of the surface compression contact layer, which decreases monotonically with increasing pressure and remains positive, is expressed in meters (m). The thickness-direction thermal conductivity of the surface veneer under the current moisture content and surface node temperature is expressed in W / (m·K). Thermal resistance per unit area formed by the coating, insulating film or adhesive layer on the surface of the heating plate; The surface temperature of the heating plate; The surface node temperature of the slab used for the current boundary calculation is taken as the correction temperature of the previous time step or the temperature of the previous iteration in the current time step; The heat flux density entering the slab surface is expressed in W / m². The distribution coefficient for the latent heat of interfacial condensation entering the slab side is determined by interfacial condensation test or on-site calibration of the hot press. The latent heat of vaporization of water is expressed in J / kg. It represents the non-negative water vapor outward migration mass flux from the interior of the slab to the corresponding surface interface, expressed in kg / (m²·s), and is calculated from moisture migration along the thickness direction.

[0050] The above , and The calculations do not use direct circular references of unknowns. Specifically, in the explicit update method, the system uses the corrected surface node temperature and water content output from the previous time node S3 to calculate the interfacial heat flux boundary at the current time node; in the iterative update method, the system first uses the corrected value from the previous time node as the initial iteration value for the current time node, calculates the interfacial heat flux boundary and solves the temperature field, and then updates the thermal conductivity, diffusion coefficient, and interfacial heat flux with the solved surface node temperature and water content, repeating the iteration until the surface node temperature difference and water content difference between two adjacent iterations satisfy the convergence boundary determined by the sensor resolution, water content detection resolution, and the allowable error of the control system. Thus, the temperature and water content variables in the thermal conductivity and interfacial thermal resistance both have clear sources.

[0051] After the heat flux density boundary calculations for the upper and lower surfaces are completed, the system establishes spatial coordinates along the thickness direction of the slab, with the upper surface as the starting point and the lower surface as the ending point. Mesh nodes are then created based on the total slab thickness, the number of veneer layers, the thickness of each veneer layer, and the location of the adhesive layer. Each mesh node has thickness coordinates, its corresponding veneer layer or adhesive layer identifier, initial temperature, current moisture content, density, specific heat capacity, basic thermal conductivity, moisture diffusion coefficient, and interlayer thermal resistance parameters. For the core layer and its adjacent adhesive layer regions, a denser mesh is used to allow for the analysis of the core layer temperature plateau and the temperature gradient across the core layer.

[0052] Furthermore, the system determines the thermal conductivity in the thickness direction based on the texture direction of each layer of the veneer. For the first... For each veneer layer, the system reads the longitudinal, radial, and chordal thermal conductivity from the material calibration data. Based on the orientation of the layer's texture coordinate system relative to the slab thickness direction, it projects the anisotropic thermal conductivity tensor onto the thickness direction to obtain the basic thermal conductivity in the thickness direction for that layer. If only transverse grain calibration data is provided during production, the transverse grain thermal conductivity is used as the basic thermal conductivity in the thickness direction. When adjacent veneers have intersecting textures, the basic thermal conductivity of each layer is calculated separately, and a uniform constant for the entire slab is not used.

[0053] The system then employs a sequential operator splitting strategy to solve the thermo-humidity coupling equations. Specifically, within a time step, the system first uses the temperature field from the previous time step or the current iteration to explicitly or semi-implicitly solve the moisture content transport equations, updating the moisture content field and water vapor outward flux. Then, based on the updated moisture content field, it calculates the thermal conductivity, moisture evaporation rate, interlayer thermal resistance, and interfacial heat flux boundary. Finally, it implicitly solves the temperature field. This sequential operator splitting strategy avoids constructing a global nonlinear Jacobian matrix in each time step, making it suitable for real-time execution by embedded controllers or industrial computers. For operating conditions where the sampling period does not exceed the calibration period of the hot-pressing control system, a single splitting update between adjacent time steps is sufficient to proceed to the next time step. If the sampling period is long or the slab moisture content is high, the system can perform a small number of local iterations within the same time step to maintain consistency in the thermo-humidity state.

[0054] In one implementation, the system calculates the thickness direction state using the following thermal-humid coupling discrete relationship:

[0055]

[0056]

[0057]

[0058]

[0059]

[0060] in, This represents the interfacial heat conduction channel coefficient between adjacent nodes, expressed in W / (m²·K). The interlayer thermal resistance per unit area is determined by the adhesive layer thickness, thermal conductivity of the adhesive, and degree of curing when the corresponding interface is an adhesive layer; it can be zero when the corresponding interface is located inside the same veneer. The mass flux of water vapor or moving moisture in the thickness direction is expressed in kg / (m²·s), with the positive direction pointing from the top surface to the bottom surface. The equivalent moisture diffusion coefficient driven by the moisture content gradient is expressed in m² / s. The density of oven-dry wood is expressed in kg / m³. The temperature gradient-driven thermal and moisture migration coefficient is expressed in kg / (m·s·K). The local evaporation rate is expressed in kg / (m³·s). It is the phase change mass transfer coefficient, with units of 1 / s, which is obtained by thermogravimetric analysis, drying test or hot pressing test of the same material. This is the saturation temperature corresponding to the current pore water vapor pressure; The phase transformation development temperature range of the material is determined by the temperature plateau range in the heating test of moisture-containing wood. This represents the initial water content at the next time point; This represents the water content state after correction at the previous time point; This is the volumetric heat capacity, expressed in J / (m³·K). This is the exothermic term of the adhesive curing reaction, expressed in W / m³. This is the heat endothermic term for water evaporation, with units of W / m³. This represents the sensible enthalpy flux carried by water vapor migration, expressed in W / m².

[0061] It should be noted that in this embodiment... Defined as the sensible enthalpy of water vapor, excluding the latent heat of vaporization; for example, we can take... ,in The specific heat capacity of water vapor at constant pressure. To standardize the reference temperature, the latent heat of vaporization is only considered in the evaporation endothermic phase. In the intermediate processing, the latent heat of interfacial condensation releases only through the boundary terms. Entering the corresponding surface heat flux boundary. Through the above definition, the system avoids repeatedly subtracting or adding latent heat in the water vapor migration term, maintaining the consistency of energy conservation.

[0062] In the moisture content advancement equation, the fluxes at the upper and lower surface boundaries are consistent with the outward migration fluxes at the interface. Specifically, when moisture migrates from the interior of the slab to the upper surface, the outward migration mass flux at the upper surface boundary is denoted as... When moisture migrates from the interior of the slab to the lower surface, the outward migration mass flux at the lower surface boundary is denoted as . The system applies this outward migration flux to both the boundary terms of the moisture content conservation equation and the interface latent heat of condensation boundary terms, creating a closed loop between the reduction of moisture from internal nodes, its migration to the surface, and the release of latent heat near the interface. If the calculated moisture content of a node is lower than the material's physical lower limit, the system projects it onto the moisture operating boundary determined by the absolute dry state and the fiber-bound water state, and simultaneously adjusts the evaporation rate at that time step to prevent local moisture from being calculated as negative.

[0063] Adhesive layer thermal resistance The system updates according to temperature and curing state. Specifically, when the interface corresponds to an adhesive layer, the system calculates the thermal resistance of the adhesive layer based on the adhesive layer thickness and the thermal conductivity of the adhesive, which is obtained from the adhesive temperature and curing degree calibration curve. In one embodiment, the adhesive curing degree... The Nth-order reaction kinetics can be updated as follows: ,in Pre-exponential factor, As the apparent activation energy, The gas constant is The absolute temperature of the adhesive layer. The reaction order is given by the parameters obtained from differential scanning calorimetry (DSC) experiments on the adhesive; the curing exothermic term can be written as... ,in The density of the adhesive layer, This refers to the heat of curing reaction per unit mass. Optionally, within the target adhesive curing temperature window, if calibration confirms that the total heat release from the curing reaction is more than an order of magnitude lower than the heat absorption from water evaporation, and the impact on the core layer temperature prediction error is within the allowable error range of the control system, the system can incorporate this into the main real-time calculation. Set it to zero, and process the thermal resistance of the adhesive layer according to the temperature calibration curve or constant calibration value; when the user provides DSC calibration data, then enable the curing kinetic exothermic term and the degree of curing-related thermal resistance of the adhesive layer.

[0064] Once the equations for the current time node are solved, the system writes the preliminary temperature values ​​and water content states of each thickness grid node into the preliminary temperature vector and water content state vector in thickness coordinate order. After multiple time nodes have completed calculations, the system combines the preliminary temperature vectors of consecutive time nodes into a preliminary temperature field distribution matrix. The calculated temperatures of the surface nodes in this matrix are entered into S3, forming an observation residual with the measured temperature of the slab surface. The temperatures of nodes near the core layer in this matrix already include the temperature plateau effect caused by moisture evaporation heat absorption, the moisture conservation propagation effect, and the sensible heat effect of water vapor migration.

[0065] S3. The hot-pressing temperature field prediction device extracts the current surface temperature observation value from the measured temperature sequence of the slab surface and calculates the temperature of the corresponding surface node from the preliminary temperature field distribution matrix to form the surface observation residual. Then, the system uses an ensemble Kalman filter framework to recursively correct the correction factors of the thickness direction thermal conductivity parameters, interface heat transfer parameters or thermal-humid coupling parameters, and re-inputs the corrected parameters into the heat transfer equation to obtain the corrected temperature field distribution matrix.

[0066] Specifically, in the main implementation, the system expresses the uncertainty of the thermal conductivity in the thickness direction as a dimensionless thermal conductivity correction factor. and with As the first The corrected thermal conductivity of the current batch for each grid node. The effects of layer position, texture orientation, moisture content, and temperature on the basic thermal conductivity have been included. This method is used to correct the overall offset of the current slab batch relative to the material calibration sample. This single-parameter correction method is suitable for working conditions where the material sources of each layer of the same slab are similar and the degree of dispersion of moisture content along the thickness direction can be covered by the material calibration range.

[0067] In one optional implementation, when the surface temperature data from multiple measurement zones, the moisture content detection data from different layers, or the physical property data of the paving layer are sufficient to support the identification of higher-dimensional parameters, the system will... The correction vector can be expanded into layered or partitioned correction vectors. For example, thermal conductivity correction factors can be set separately for the surface layer, core layer, and lower surface layer, or thermal conductivity correction factors can be set separately for different measurement areas. The dimension of the layered or partitioned correction vector does not exceed the dimension of the effective surface temperature observation at the current moment. When the number of observations is insufficient to support multidimensional parameter identification, the system automatically switches back to single-parameter correction mode and records multidimensional correction requests as low-confidence information, which is not included in the core filter calculation. The physical continuity between components can be maintained through adjacent layer smoothing constraints or material calibration boundaries to avoid parameters becoming unidentifiable due to insufficient observations.

[0068] Furthermore, the system generates multiple sets of correction factor members. The initial set is centered on the material calibration state, and its dispersion is determined by the dispersion of moisture content of the same batch of single plates, the dispersion of density sampling, the material thermal conductivity calibration error, the diffusion coefficient calibration error, and the surface temperature prediction residual under historical stable operating conditions. Each set member corresponds to a set of candidate heat transfer parameters. The system substitutes each candidate parameter into the thermal-humid coupling heat transfer equation of S2 to obtain the corresponding predicted surface node temperature and predicted internal node temperature.

[0069] The system employs a trend-compensated observation uncertainty update method for ensemble Kalman filtering correction.

[0070]

[0071]

[0072]

[0073]

[0074] in, This represents the parameter to be corrected, which in the main implementation is a single thermal conductivity correction factor. In an optional implementation, this is a hierarchical or partitioned correction vector; This represents the measured surface temperature vector of the slab at the current moment. The predicted average surface temperature for the set; This is the average surface temperature after correction at the previous moment; To deduct the observation increment residual after the predicted temperature rise trend, in order to avoid the observation uncertainty caused by the rapid rise of surface temperature in the early stage of hot pressing being misjudged as temperature measurement noise; The variance of the observation increment residuals updated exponentially weighted, with dimensions K²; The dimensionless update weights are determined by the sampling period and the sensor response time. This represents the current observation residual; The current residual is predicted based on the recent residual evolution trend; The observation uncertainty reference matrix is ​​formed by the sensor calibration error; This represents the current observation uncertainty matrix; The sample covariance between the parameter set and the surface predicted temperature set; The sample covariance matrix of the surface predicted temperature set; This is the filter gain; To and The corresponding zero-mean observation perturbation; This means that the updated parameters will be kept within the parameter space determined by the material calibration samples, the equipment operating boundary, and the thermal-humid coupling physical constraints.

[0075] In the above recursive relationship, the system does not directly superimpose the surface temperature residual onto the core temperature, but rather determines the direction and magnitude of parameter correction based on the ensemble covariance between the surface residual and the heat transfer parameters. Because Using the observation increment residuals after trend compensation, instead of directly using the original temperature variance within the sliding time window, ensures that the rapid rise in surface temperature during the initial stage of hot pressing cannot be simply interpreted as measurement noise. When sensor fluctuations, localized obstructions, or short-term anomalies do occur in the surface temperature measurement, the observation increment residuals and residual trend deviation terms will... This increases the value of the thermal conductivity parameter, thereby reducing the disturbance to the parameters caused by a single observation. This processing maintains the stability of the filter and distinguishes the thermal conductivity parameter correction from the actual temperature rise trend of the hot-pressing process.

[0076] After the correction factor set is updated, the system checks whether the thermal conductivity, interfacial heat transfer parameter, or diffusion parameter corresponding to each set member falls within the physical operating space supported by the material calibration sample. If a set member results in a negative thermal conductivity, abnormal moisture diffusion coefficient, negative moisture content, or a parameter exceeding the coverage range of the calibration sample under the same material type, moisture content, and temperature range, the system projects that set member to the adjacent physical operating boundary, or regenerates a replacement set member based on the calibration sample distribution. Subsequently, the system calculates representative parameter values ​​based on the updated set members. These representative values ​​can be the set mean or the robust center value after physical projection.

[0077] Furthermore, once the representative values ​​of the parameters at the current moment are determined, the system applies them to the base thermal conductivity, interfacial heat transfer coefficient, or moisture diffusion parameters of each layer, and resolves the thickness-direction heat transfer equation for the current time step using the pressure-water-phase change coupled heat flux boundary of S2 and the corrected thermo-humidity coupled parameters, obtaining the corrected temperature vector and corrected water-bearing state vector for the current moment. If improved temporal continuity is required, the system can recalculate short time windows prior to the current moment to ensure that the evolution of the temperature and water-bearing fields within those time windows is consistent with the latest parameters. If surface temperature observations are missing at the current moment, the system does not perform observation updates but instead uses the parameter set from the previous moment and its evolution uncertainty for prediction and continues to output the temperature field; once observations are recovered, the parameters are updated using the new surface residuals.

[0078] After each hot-pressing time node completes recursive correction, the system writes the corrected temperature vector of each time node into the corrected temperature field distribution matrix in chronological order, and writes the corrected water-bearing state vector of each time node into the water-bearing state matrix. The corrected temperature field distribution matrix includes three dimensions: thickness grid nodes, hot-pressing time nodes, and selectable measurement area locations. It also simultaneously records the thermal conductivity correction factor, corrected thermal conductivity, water-bearing state, surface observation residuals, observation uncertainties, and parameter physical boundary projection records. This matrix, after being processed by S4, is used to extract core region temperature values ​​and generate core section temperature field prediction results.

[0079] S4. The hot-pressing temperature field prediction device extracts the grid node temperature corresponding to the core layer region from the corrected temperature field distribution matrix, and generates the core layer temperature gradient curve or core layer cross-sectional temperature field according to the slab thickness coordinates, core layer position and selectable test area position, and outputs the core layer temperature prediction result corresponding to the hot-pressing time axis.

[0080] Specifically, the system first reads the slab plywood arrangement information to determine the position of the core layer region along the slab thickness direction. For odd-layer plywood, the system identifies the veneer containing the geometric mid-surface and its adjacent inner glue lines as the core layer region; for even-layer plywood, the system identifies the adjacent veneer layers and glue layers on both sides of the geometric mid-surface as the core layer region; for composite plywood with a clearly defined functional core or load-bearing core, the system determines the core layer region based on the core layer position specified in the process formula. Subsequently, the system matches the thickness coordinate range of the core layer region with the grid node coordinates in the corrected temperature field distribution matrix output by S3 to identify the grid nodes falling within the core layer region.

[0081] When the core layer region boundary coincides with the grid node position, the system directly extracts the corrected temperature value of the corresponding node. When the core layer region boundary is located between two adjacent grid nodes, the system interpolates based on the temperature and thickness coordinates of the adjacent nodes to obtain the temperature values ​​of the core layer boundary, core layer center surface, or core layer adhesive line position. If adjacent nodes are located within the same veneer material, the system uses thickness coordinate interpolation. If adjacent nodes cross adhesive layers or interlayer interfaces, the system performs interpolation on both sides of the interface, preserving the temperature change characteristics caused by interlayer thermal resistance. The system combines the directly extracted core layer node temperatures and the interpolated core layer position temperatures to form a core layer temperature data set.

[0082] The system then sorts the core layer temperature data according to the slab thickness coordinates and generates a core layer temperature gradient curve based on the temperature difference and thickness distance between adjacent core layer temperature points. The temperature gradient is expressed in K / m or ℃ / m and characterizes the degree of temperature change along the thickness direction near the core layer. Since the corrected temperature field output by S3 already includes the effects of water vapor evaporation endothermic effect, water conservation propagation, water vapor migration sensible heat, interfacial condensation latent heat, and online correction of thermal conductivity parameters, the core layer temperature gradient curve can reflect the core layer temperature plateau period during hot pressing, the asymmetric heating on both sides of the core layer, and the local temperature lag caused by interlayer thermal resistance.

[0083] Furthermore, the system synchronously extracts the water content of the core region from the water content matrix output by S3 to identify the water phase transition stage corresponding to the predicted core temperature. If the core temperature is in a temperature range where water evaporation and heat absorption are significant, and the water content is still higher than the low water content operating boundary determined by material calibration, the system records in the output data that the core temperature point is in the temperature plateau influence stage; if the water content is close to the low water content operating boundary, the system records that the influence of the evaporation and heat absorption term in this region on subsequent temperature rise is weakened. This record is not used as a manual judgment rule, but is directly generated from the thermo-humidity coupling state quantities of S2 and S3.

[0084] Furthermore, if the hot press is equipped with multiple surface temperature measurement channels along the width or length of the slab, the system performs core layer node extraction and core layer temperature gradient generation for each measurement area. For areas where there is no direct surface temperature measurement location between adjacent measurement areas, the system interpolates based on the spatial coordinates of the measurement area and the spatial continuity of the core layer temperature in adjacent measurement areas to form a core layer cross-sectional temperature field. The core layer cross-sectional temperature field can be represented as a two-dimensional matrix, where one dimension represents the measurement area location on the slab surface and the other dimension represents the thickness coordinates of the core layer region, and the matrix elements are the correction temperature values ​​at the corresponding hot pressing time; it can also be represented as a set of core layer temperature gradient curves updated over time.

[0085] For example, a 18 mm plywood has 9 grid nodes along its thickness, with the geometric mid-surface located at 9 mm, and the core layer covering 7 mm to 11 mm. If the corrected temperature field output by S3 at a certain hot-pressing moment shows temperatures of 63.5℃, 61.8℃, and 64.1℃ at positions 7 mm, 9 mm, and 11 mm, respectively, the system will use the 9 mm position as the core layer center temperature point and generate temperature gradients on both sides of the core layer based on the temperature changes from 7 mm to 9 mm and from 9 mm to 11 mm. If there are three measurement zones along the width of the same slab, with core layer center temperatures of 61.8℃, 64.0℃, and 62.7℃, respectively, the system will further arrange the core layer temperatures of the three zones according to their coordinates to form the core layer cross-sectional temperature distribution at that hot-pressing moment. This example is only for illustrating the core layer temperature extraction and cross-sectional temperature field generation method; the actual number of grid nodes, core layer range, and number of measurement zones are determined by the slab structure and temperature measurement arrangement.

[0086] Once the core layer temperature field is generated, the system writes the hot-pressing time, slab batch, test area number, core layer thickness coordinates, core layer temperature value, local temperature gradient, core layer moisture content, upper and lower surface heat flux density, interfacial latent heat flux, thermal conductivity correction factor, surface observation residuals, and data quality identifier into the output data structure. The output data structure can be stored in the form of a matrix, curve data, data table, or hot-pressing controller communication messages. In the matrix form, rows represent core layer thickness coordinates or test area locations, columns represent hot-pressing time nodes, and matrix elements represent corrected temperature values. In the data table form, each record includes time, test area, thickness coordinates, temperature, temperature gradient, moisture content, and thermal conductivity correction information.

[0087] Furthermore, the system sends the predicted core layer cross-sectional temperature field results to the hot-pressing control system or process parameter optimization system. For real-time control scenarios, the hot-pressing control system can adjust the subsequent heating plate temperature, heating rate, or holding time by considering the adhesive curing temperature requirements, the material's allowable temperature operating boundary, the core layer's moisture content, and the core layer temperature evolution trend over time. For offline process optimization scenarios, the system can compare the core layer cross-sectional temperature field and moisture content field of different batches of slabs to analyze the impact of batch moisture content, density fluctuations, texture differences, interlayer contact state, and moisture migration processes on the core layer heating hysteresis. Thus, the system completes the process of generating real-time predicted core layer cross-sectional temperature field results from measurable surface temperature and process parameters.

[0088] In summary, the method for predicting the temperature field of the plywood hot-pressed core layer based on the embodiments of the present invention has been clarified. It involves time-series analysis of the heating plate temperature, slab surface temperature, hot-pressing pressure, and initial physical properties of the slab at different time points during the hot-pressing process. Dynamic modeling is then performed using the thermo-humidity coupling heat transfer equation and a recursive parameter correction algorithm to determine the real-time state of the temperature field in the core layer cross-section and the hydrothermal evolution trend. This improves the accuracy and timeliness of core layer temperature prediction without damaging the slab structure, helping to identify early risks of insufficient curing or overheating degradation caused by batch fluctuations in physical properties. This allows for timely adjustment of heating curves and process parameters, reducing product scrap rates and improving batch quality consistency.

[0089] Here, those skilled in the art will understand that the specific operations of each step in the above-described plywood hot-pressed core layer temperature field prediction system have been referenced above. Figure 1 and Figure 2 The method for predicting the temperature field of the hot-pressed core layer of plywood has been described in detail, and therefore, its repeated description will be omitted.

[0090] In summary, the plywood hot-pressing core layer temperature field prediction system based on embodiments of the present invention has been clarified. It performs time-series analysis on the heating plate temperature, slab surface temperature, hot-pressing pressure, and initial physical properties of the slab at different time points during the hot-pressing process. Combined with dynamic modeling using the thermo-humidity coupling heat transfer equation and a recursive parameter correction algorithm, it determines the real-time state of the core layer cross-section temperature field and the hydrothermal evolution trend. This improves the accuracy and timeliness of core layer temperature prediction without damaging the slab structure, helping to identify early risks of insufficient curing or overheating degradation caused by batch fluctuations in physical properties. This allows for timely adjustment of heating curves and process parameters, reducing product scrap rates and improving batch quality consistency.

Claims

1. A method for predicting the temperature field of a hot-pressed core layer of plywood, characterized in that, include: Acquire heating plate temperature data, slab surface temperature data, hot pressing pressure data, and initial physical property parameters of the slab during the hot pressing process, and align them according to the same hot pressing time axis to construct a multi-source input matrix; Based on the hot pressing pressure data and heating plate temperature data, the interfacial heat flow boundary conditions of the slab surface are determined. Several spatial grid nodes are divided along the slab thickness direction. The preliminary temperature response of each grid node is calculated based on the heat transfer equation, and a preliminary temperature field distribution matrix is ​​generated. The observation residual is formed by the measured surface temperature of the slab at the current moment and the calculated temperature of the corresponding surface node in the preliminary temperature field distribution matrix. The heat transfer parameters are corrected by the parameter recursion correction algorithm, and the temperature field is updated based on the corrected heat transfer parameters to obtain the corrected temperature field distribution matrix. The temperature values ​​of the corresponding grid nodes in the core region are extracted from the corrected temperature field distribution matrix to generate the core temperature gradient curve or the core cross-sectional temperature field, and the core temperature field prediction result is output.

2. The method for predicting the temperature field of the hot-pressed core layer of plywood according to claim 1, characterized in that, The interfacial heat flux boundary conditions include the contact heat conduction flux determined by the pressure contact state and the phase change-assisted heat transfer flux determined by the outward migration and condensation of water vapor. The interfacial heat flux boundary conditions are determined according to the following relationship: ； in, The heat flux density entering the slab surface, For contact heat transfer coefficient, The surface temperature of the heating plate. The temperature is either the temperature corrected at the previous time step or the temperature iterated at the current time step. The latent heat of condensation distribution coefficient, For latent heat of vaporization, This refers to the mass flux of water vapor migrating from the interior of the slab to the surface interface.

3. The method for predicting the temperature field of the hot-pressed core layer of plywood according to claim 2, characterized in that, The interfacial heat flow boundary conditions of the slab surface are determined by determining the contact heat transfer coefficient based on the hot-pressing pressure data to quantify the interfacial thermal resistance. The contact heat transfer flux is calculated by multiplying the contact heat transfer coefficient by the difference between the heating plate temperature and the slab surface temperature, and the phase change-assisted heat transfer flux formed by the outward migration and condensation of water vapor is superimposed to jointly determine the interfacial heat flow boundary conditions of the slab surface.

4. The method for predicting the temperature field of the hot-pressed core layer of plywood according to claim 3, characterized in that, The thermal conductivity in the thickness direction corresponding to the spatial grid node is determined based on the texture direction, water content, and temperature of the single-layer plate. When the material calibration data includes longitudinal thermal conductivity, radial thermal conductivity, and chordal thermal conductivity, the anisotropic thermal conductivity tensor is projected onto the thickness direction of the slab to obtain the basic thermal conductivity of each slab layer in the thickness direction. The interlayer thermal resistance between adjacent grid nodes is then determined based on the adhesive layer thickness, adhesive thermal conductivity, and degree of curing.

5. The method for predicting the temperature field of the hot-pressed core layer of plywood according to claim 4, characterized in that, The heat transfer equation includes moisture migration flux, local moisture evaporation rate, water content time-progression term, and temperature time-progression term. The water content time-progression term is determined according to the following relationship: ； in, This represents the initial water content at the next time point. This represents the water content after correction at the previous time point. and This represents the water mass flux at the adjacent interface. The density of absolutely dry wood. To control the thickness of the body, This represents the localized rate of water evaporation.

6. The method for predicting the temperature field of the hot-pressed core layer of plywood according to claim 5, characterized in that, The heat transfer equation is solved using a sequential operator splitting method, which includes: First, solve the water content transport equation based on the temperature field of the previous moment or the current iteration, and then update the water content field and water vapor outflow flux. Then, based on the updated moisture content field, the thermal conductivity, moisture evaporation rate, interlayer thermal resistance, and interfacial heat flow boundary are calculated; finally, the temperature field is implicitly solved.

7. The method for predicting the temperature field of the hot-pressed core layer of plywood according to claim 6, characterized in that, The heat transfer parameters are corrected based on an ensemble Kalman filter framework. The direction and magnitude of the correction are determined by the ensemble covariance between the surface temperature observation residual and the heat transfer parameters. An adaptive observation noise matrix is ​​constructed by subtracting the observation increment residual after the predicted temperature rise trend to suppress the initial temperature rise interference of hot pressing. At the same time, the corrected parameters are constrained within the physical operating boundary of the material calibration to realize the online recursive update of the heat transfer parameters.

8. A system for predicting the temperature field of a hot-pressed core layer of plywood, characterized in that, include: The data acquisition module is used to acquire heating plate temperature data, slab surface temperature data, hot pressing pressure data, and initial physical property parameters of the slab during the hot pressing process, and align them according to the same hot pressing time axis to construct a multi-source input matrix; The forward thermal-humidity coupling solution module is used to determine the interfacial heat flow boundary conditions on the slab surface based on the hot-pressing pressure data and the heating plate temperature data, divide a number of spatial grid nodes along the slab thickness direction, calculate the preliminary temperature response of each grid node based on the heat transfer equation, and generate a preliminary temperature field distribution matrix. The recursive correction module is used to form an observation residual based on the measured surface temperature of the slab at the current moment and the calculated temperature of the corresponding surface node in the preliminary temperature field distribution matrix. The module then uses a parameter recursive correction algorithm to correct the heat transfer parameters and updates the temperature field based on the corrected heat transfer parameters to obtain the corrected temperature field distribution matrix. The core temperature extraction and output module is used to extract the temperature values ​​of the corresponding grid nodes in the core region from the corrected temperature field distribution matrix, generate the core temperature gradient curve or the core cross-sectional temperature field, and output the core temperature field prediction result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for predicting the temperature field of the hot-pressed core layer of plywood as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for predicting the temperature field of the hot-pressed core layer of plywood as described in any one of claims 1 to 7.

Images