Hot air fusion sealing method for color steel plate and magnesium base material

Abstract

The present application provides a kind of hot air melting sealing method of color steel plate and magnesium base material, belong to color steel plate and magnesium base material production technical field, the present application is by infrared thermal imager acquisition color steel plate edge temperature field and input green function convolution hot air non-uniform heating heat transfer inverse problem solving module obtains optimal heat flow distribution, the optimal heat flow distribution is converted into hot air gun scanning speed and power joint control curve after executing hot air heating timing decoupling process, synchronously with variation level set function to glue flow front 3D morphology is real-time forecasted and adjusted hot air gun scanning speed and coating rate accordingly, then brush the fireproof sealant containing silane coupling agent and complete three-stage gradient curing treatment, finally interlaminar peel strength and edge sealing quality are tested, solve the technical problem that interface bonding quality is uncontrollable in the process of hot air melting sealing of color steel plate and magnesium base material hot air temperature field non-uniform distribution leads to.

Description

Technical Field

[0001] This invention belongs to the field of production technology of color steel sheet and magnesium substrate, specifically, it relates to a hot air melting and sealing method for color steel sheet and magnesium substrate. Background Technology

[0002] Color-coated steel-magnesium composite panels are integrated fireproof air duct panels formed by bonding color-coated steel sheets and magnesium substrates with adhesives. Their edge sealing process typically employs hot air melting, where a hot air gun heats the edges of the color-coated steel sheet, melting residual adhesive to fill edge gaps before applying fire-retardant sealant to create a complete edge seal. However, in current hot air melting sealing processes, the heat flux density at the hot air gun outlet exhibits a Gaussian distribution radially, resulting in a significant temperature gradient within the limited width of the color-coated steel sheet edge. Existing processes rely on operator experience or fixed parameters for open-loop temperature control, failing to dynamically adjust the hot air gun scanning speed and output power based on real-time temperature field distribution. This leads to localized overheating or underheating in the heated area, resulting in defects such as excessive adhesive melting and loss or incomplete filling. In the existing technology, the hot air melting and sealing process lacks real-time perception and closed-loop control of the spatial distribution of the hot air temperature field. The problem of non-uniform heat flux density distribution cannot be effectively compensated, and the bonding strength and sealing integrity of the edge sealing area are difficult to be stably guaranteed. In other words, the existing technology has the technical problem of uncontrollable interface bonding quality due to the non-uniform distribution of the hot air temperature field during the hot air melting and sealing process of color steel plate and magnesium substrate. Summary of the Invention

[0003] In view of this, the present invention provides a hot air melting and sealing method for color steel sheet and magnesium substrate, which can solve the technical problem in the prior art where the non-uniform distribution of hot air temperature field during the hot air melting and sealing process of color steel sheet and magnesium substrate leads to uncontrollable interface bonding quality.

[0004] This invention is implemented as follows: This invention provides a hot air melting and sealing method for color steel sheets and magnesium substrates, comprising the following steps:

[0005] The color steel magnesium composite sheet is placed in an environment with a preset temperature and relative humidity for a preset time for pretreatment. At the same time, a near-infrared transmission spectroscopy online detection device is used to evaluate the moisture content field of the sheet in real time. Areas with moisture content higher than a preset moisture content threshold are marked as high humidity areas, and the pretreatment time of the high humidity areas is extended until the moisture content is lower than the preset moisture content threshold.

[0006] The edge temperature field of the color steel plate is acquired by an infrared thermal imager at a preset frame rate. The edge temperature field of the color steel plate is input into the Green function convolution hot air non-uniform heating heat transfer inverse problem solving module. After forward Fourier transform, Tikhonov regularization Green function frequency domain inverse matrix multiplication and inverse Fourier transform, the optimal heat flux distribution is obtained. The optimal heat flux distribution is then converted into a joint control curve of hot air gun scanning speed and power through piecewise linear interpolation.

[0007] According to the hot air gun scanning speed and power joint control curve, the edge of the color steel plate is heated with hot air at a preset temperature for a preset width area for a preset time to melt and activate the flow of residual glue, and the pressing is completed within a preset pressing time. After pressing, the temperature field of the edge of the color steel plate is collected again. If the temperature uniformity deviation exceeds the preset temperature deviation threshold, the temperature field of the edge of the color steel plate is re-input into the Green's function convolution hot air non-uniform heating heat transfer inverse problem solving module, and the steps of obtaining the optimal heat flow distribution and converting it into the hot air gun scanning speed and power joint control curve and this step are executed again.

[0008] The glue flow front is defined as the zero level set of the variational level set function. An energy functional is constructed using interfacial energy, glue viscosity dissipation, and capillary pressure difference. The variational level set function is iteratively evolved through gradient descent. An adaptive mesh refinement technique is used to refine the mesh to a preset mesh resolution in the region near the glue flow front. The three-dimensional morphology of the glue flow front is predicted in real time. If the offset of the glue flow front exceeds the preset offset threshold, the hot air gun scanning speed is reduced and the glue application rate is decreased before the heating step is repeated.

[0009] Fire-retardant sealant containing silane coupling agent is applied along the edge of the color steel plate to form a triangular support structure with the color steel plate and magnesium substrate. The sealed color steel magnesium composite plate is then cured sequentially at a first curing temperature for a first curing time, at a second curing temperature for a second curing time, and at a third curing temperature for a third curing time to complete a three-stage gradient curing process.

[0010] For color steel magnesium composite panels that have completed the three-stage gradient curing process, interlayer peel strength and edge sealing quality are tested: panels with interlayer peel strength lower than the preset peel strength threshold are returned to the gradient pre-drying process for reprocessing; panels with water seepage at the edge seal after immersion test are returned to the gradient pre-drying process for reprocessing.

[0011] The preset temperature is 45℃, the preset relative humidity is 30%RH, the preset duration is 4 to 6 hours, and the preset moisture content threshold is 0.5%.

[0012] The near-infrared transmission spectroscopy online detection device detects the spectral absorption differences at a preset wavelength, quantitatively inverts the internal moisture content and spatial distribution of the board, and outputs a two-dimensional distribution map of the moisture content field of the board.

[0013] The preset frame rate is 100 frames per second, and the single solution time of the Green's function convolution hot air non-uniform heating heat transfer inverse problem solving module does not exceed 50ms.

[0014] The regularization parameter of the Tikhonov regularization is adaptively selected by the L-curve method. The L-curve method determines the optimal regularization parameter by plotting a parameterized curve between the regularized residual norm and the regularized subnorm in a double logarithmic coordinate system and selecting the point of maximum curvature.

[0015] The piecewise linear interpolation connects the discrete spatial node values ​​of the optimal heat flux distribution piecewise using a linear function between adjacent nodes, converting them into the hot air gun scanning speed and output power values ​​corresponding to each position on the continuous scanning path.

[0016] The preset temperature hot air is 120℃ hot air, the preset width area is a 5mm wide area, the heating time is 5 seconds, the preset pressing time is 5 seconds, and the preset temperature deviation threshold is ±5℃.

[0017] The time step of the gradient descent iteration satisfies the CFL condition, which requires that the ratio of the flow field characteristic velocity, the time step, and the spatial grid size does not exceed the upper limit of stability. In each iteration, the Fourier heat conduction equation is solved simultaneously to obtain the updated temperature field value, the simplified solution of the Stokes equation is solved to obtain the updated flow velocity field value, and a re-initialization operation is performed to maintain the sign distance characteristic of the variational level set function.

[0018] The preset grid resolution is 5μm, and the preset offset threshold is 0.2mm.

[0019] The fire-retardant sealant contains 0.5% silane coupling agent by mass, and has a coating width of 8 mm and a coating thickness of 3 mm.

[0020] The triangular support structure refers to the geometric constraint shape in which the fireproof sealant forms an approximately triangular cross section between the edge of the color steel plate and the magnesium substrate, so that the fireproof sealant layer forms an adhesive force in two directions with the side of the color steel plate and the upper surface of the magnesium substrate at the same time.

[0021] The first curing temperature is 30°C and the first curing time is 8 hours; the second curing temperature is 45°C and the second curing time is 12 hours; the third curing temperature is 25°C and the third curing time is 24 hours.

[0022] In the three-stage gradient curing process, the first curing temperature stage allows the adhesive to initially form and fix the position of the color steel plate, the second curing temperature stage accelerates the chemical reaction between the adhesive and the magnesium substrate, and the third curing temperature stage avoids the internal stress introduced by high-temperature rapid curing, which could cause the color steel magnesium composite plate to warp.

[0023] The preset peel strength threshold is 1.5 kN / m, and the edge sealing quality inspection is determined by observing whether water seepage occurs at the edge seal after a water immersion test.

[0024] The reinitialization operation refers to periodically resetting the variational level set function to the signed distance function to the glue flow front surface during the iterative evolution of the variational level set function, in order to correct the function gradient distortion caused by numerical accumulation error.

[0025] This invention introduces a Green's function convolution module for solving the inverse problem of non-uniform heating heat transfer in hot air. It uses an infrared thermal imager to collect the temperature field at the edge of the color-coated steel sheet in real time. After converting the temperature field into the frequency domain via Fourier transform, the optimal heat flux distribution is inversely derived using the frequency domain inverse matrix of the Green's function with Tikhonov regularization. This optimal heat flux distribution is then converted into a joint control curve of the hot air gun scanning speed and power through piecewise linear interpolation, achieving closed-loop precise control of the hot air temperature field. Based on a linear partial differential equation model of heat transfer in thin-walled color-coated steel sheets, this invention uses the Green's function to describe the impact response of a point heat source. Through inverse convolution, the deviation between the measured temperature field and the target temperature field is directly mapped into a heat flux control command, eliminating the non-uniform temperature gradient effect caused by the Gaussian distribution of heat flux density. This makes the heating state of each location in the edge sealing area more consistent, thereby ensuring the uniformity of glue melting and filling and the stability of the interface bonding strength. In summary, this invention solves the technical problem mentioned in the background art where the non-uniform distribution of the hot air temperature field during the hot air melting and sealing process of color-coated steel sheets and magnesium substrates leads to uncontrollable interface bonding quality. Attached Figure Description

[0026] Figure 1 This is a flowchart of the method of the present invention.

[0027] Figure 2 This is a curve showing the combined control of power and speed for hot air scanning in the edge area of ​​the color steel plate.

[0028] Figure 3 This is a graph showing the evolution of the sealant flow front offset over time. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below.

[0030] like Figure 1 The diagram shows a flowchart of a hot air melting and sealing method for color steel sheet and magnesium substrate provided by the present invention. This method includes the following steps:

[0031] S01. Pre-treat the color steel magnesium composite sheet in an environment with a temperature of 45℃ and a relative humidity of 30%RH for 4 to 6 hours. Simultaneously, use a near-infrared transmission spectroscopy online detection device to evaluate the moisture content field of the sheet in real time. Mark the area with a moisture content higher than 0.5% as a high humidity area, and extend the pre-treatment time for the high humidity area until the moisture content is lower than 0.5%.

[0032] S02. The edge temperature field of the color steel plate is acquired by an infrared thermal imager at a frame rate of 100 frames per second. The edge temperature field of the color steel plate is input into the Green function convolution hot air non-uniform heating heat transfer inverse problem solving module. After forward Fourier transform, Tikhonov regularization Green function frequency domain inverse matrix multiplication and inverse Fourier transform, the optimal heat flow distribution is obtained. The optimal heat flow distribution is then converted into a joint control curve of hot air gun scanning speed and power through piecewise linear interpolation.

[0033] S03. According to the hot air gun scanning speed and power joint control curve, heat the 5mm wide area of ​​the color steel plate edge with 120℃ hot air for 5 seconds to melt and activate the flow of residual glue, and complete the pressing within 5 seconds; after pressing, collect the temperature field of the color steel plate edge again. If the temperature uniformity deviation exceeds ±5℃, re-input the temperature field of the color steel plate edge into the Green's function convolution hot air non-uniform heating heat transfer inverse problem solving module, and re-execute S02 and S03.

[0034] S04. Define the sealant flow front as the zero level set of the variational level set function. Construct an energy functional using interfacial energy, adhesive viscosity dissipation, and capillary pressure difference. Iteratively evolve the variational level set function through gradient descent. Use adaptive mesh refinement technology to refine the mesh to 5μm resolution in the region near the adhesive flow front. Real-time prediction of the three-dimensional morphology of the adhesive flow front. If the offset of the adhesive flow front exceeds 0.2mm, reduce the hot air gun scanning speed and the adhesive application rate, and then re-execute S03.

[0035] S05. Apply a fireproof sealant containing 0.5% silane coupling agent by mass along the edge of the color steel plate, with a width of 8mm and a thickness of 3mm, so that the fireproof sealant forms a triangular support structure with the color steel plate and the magnesium substrate; place the sealed color steel magnesium composite plate in a 30℃ environment for 8 hours, a 45℃ environment for 12 hours, and a 25℃ environment for 24 hours in sequence to complete the three-stage gradient curing treatment;

[0036] S06. Conduct interlayer peel strength and edge sealing quality inspections on color steel magnesium composite panels that have completed the three-stage gradient curing treatment: panels with an interlayer peel strength lower than 1.5kN / m are returned to S01 for reprocessing; panels with water seepage at the edge seal after the immersion test are returned to S01 for reprocessing.

[0037] Among them, the near-infrared transmission spectroscopy online detection device refers to an online detection device that utilizes the transmission characteristics of magnesium substrate to the near-infrared band spectrum, and quantitatively inverts the internal moisture content and spatial distribution of the board by detecting the spectral absorption differences at a preset wavelength. The output result is a two-dimensional distribution map of the moisture content field of the board, which is used to identify high humidity areas and guide the extension of pre-drying time.

[0038] The principle of the Green's function convolution module for solving the inverse problem of non-uniform heating heat transfer in color steel plates is as follows: The heat transfer process of thin-walled color steel plates is modeled as a system of linear partial differential equations. The Green's function describes the temperature impact response generated by an arbitrary heat source acting on the thin-walled structure, i.e., the transfer kernel function between the point heat source input and the temperature field output. The forward problem is to find the temperature field given the heat flux distribution, and the inverse problem is to inversely deduce the required heat flux distribution given the target temperature field. The algorithm transforms the temperature field at the edge of the color steel plate into the frequency domain through a forward Fourier transform. In the frequency domain, the inverse problem is solved using Wiener filtering regularization. The frequency domain signal of the target temperature field is divided by the frequency domain response matrix of the Green's function, and Tikhonov regularization is applied to suppress the ill-conditioned amplification noise inherent in the inverse problem. The regularization parameter is adaptively selected by the L-curve method to ensure the stability of the solution. After the frequency domain solution is completed, the optimal heat flux distribution is restored by inverse Fourier transform, and then mapped to the joint control curve of the hot air gun scanning speed and power through piecewise linear interpolation, which is then input into the hot air gun actuator. This solver module transforms hot air heating control from passive open-loop compensation to active physical-driven closed-loop optimization, fundamentally solving the problem of non-uniform temperature gradient caused by Gaussian distribution of heat flux density within 5mm of the edge of the color steel plate. This substantially improves the temperature uniformity of the target area and controls the single solution time to within 50ms, meeting the real-time control requirements of the process.

[0039] Tikhonov regularization is a method that constrains the norm of the solution by introducing a regularization parameter in the frequency domain inverse matrix solution. This is used to prevent computational divergence caused by noise amplification during the inverse problem solution. The larger the regularization parameter, the smoother the solution but the greater the deviation. The L-curve method adaptively determines the optimal regularization parameter by plotting the trade-off between the norm of the solution and the norm of the residual and selecting the point of maximum curvature. The formula is as follows:

[0040] ;

[0041] in , , For the residual vector, For regularization, For regularization parameters, For the curvature of the curve, The optimal regularization parameter is .

[0042] Piecewise linear interpolation refers to converting the discrete spatial node values ​​of the optimal heat flux distribution into the corresponding hot air gun scanning speed and output power values ​​at each position on the continuous scanning path of the hot air gun by connecting adjacent nodes piecewise with a linear function, in order to generate an executable joint control curve for the hot air gun scanning speed and power.

[0043] The principle of the variational level set function is as follows: During the sealing adhesive melting and filling process, the adhesive flow front is a surface that evolves over time in three-dimensional space. The variational level set method implicitly defines this surface as an isosurface where a scalar function has a value of zero in three-dimensional space. The variational level set function value is negative for the adhesive-filled area and positive for the unfilled area, thus transforming the surface tracking problem into a scalar field evolution problem. An energy functional is constructed using three physical mechanisms: interfacial energy, adhesive viscous dissipation, and capillary pressure difference. The variational level set function is iteratively updated along the negative gradient direction at a time step that satisfies the CFL condition, allowing the adhesive flow front surface to evolve naturally according to physical laws. In each iteration, the Fourier heat conduction equation is solved simultaneously to obtain the updated temperature field value, and the simplified solution of the Stokes equation is solved to obtain the updated velocity field value. A re-initialization operation is performed to maintain the sign distance characteristic of the variational level set function and prevent numerical dispersion from distorting the adhesive flow front. An adaptive mesh refinement technique is used in the region near the glue flow front to refine the mesh to a resolution of 5μm, ensuring the accuracy of the three-dimensional morphology prediction of the glue flow front. This method transforms the sealant filling process from one that relies on experience-based judgment to a quantitatively predictable real-time monitoring process. When the predicted offset of the glue flow front exceeds a set threshold, the hot air gun scanning speed and glue application rate are intervened in a timely manner, fundamentally suppressing the occurrence of edge glue shortages and local blistering defects.

[0044] The CFL condition refers to the constraint that the ratio of the time step to the spatial grid size must satisfy to ensure the stability of the numerical solution of partial differential equations. The formula is as follows:

[0045] ;

[0046] in The characteristic velocity of the flow field, in units ; For time step, unit ; Spatial grid size, unit ; The upper limit for stability is set to 1 for the explicit format; all terms on the left side of the equation are dimensionless to ensure consistency of the formula dimensions.

[0047] Reinitialization refers to periodically resetting the variational level set function to the sign distance function to the glue flow front surface during the iterative evolution of the variational level set function, in order to correct the function gradient distortion caused by numerical accumulation error and prevent numerical diffusion or spurious splitting of the glue flow front surface during iteration.

[0048] Among them, the triangular support structure refers to the geometric constraint shape of the fireproof sealant forming an approximately triangular cross section between the edge of the color steel plate and the magnesium substrate. This allows the fireproof sealant layer to form adhesion forces in two directions with both the side of the color steel plate and the upper surface of the magnesium substrate. This not only blocks the path of moisture penetration along the interface, but also provides bending moment support when the edge is subjected to bending load.

[0049] The three-stage gradient curing process refers to a curing process in which the curing process of color steel magnesium composite panels is divided into three stages: a 30°C initial curing stage, a 45°C stabilization stage, and a 25°C full curing stage. The 30°C initial curing stage allows the adhesive to initially form and fix the position of the color steel panel. The 45°C stabilization stage accelerates the chemical reaction between the adhesive and the magnesium substrate to improve the bonding strength. The 25°C full curing stage avoids the internal stress introduced by high-temperature rapid curing, which could cause the color steel magnesium composite panel to warp. The three stages work together to ensure that the interlayer peel strength reaches the requirement of not less than 1.5kN / m.

[0050] It should be noted that the first key technical idea of ​​this invention is to solve the inverse heat transfer problem based on Green's function convolution. Traditional hot air heating processes use open-loop control with fixed power and speed parameters, which cannot respond to local temperature deviations caused by the Gaussian distribution of heat flux density. This invention models the heat transfer of thin-walled color steel plates as a linear partial differential equation system, uses Green's function to establish an analytical transfer relationship between heat flux distribution and temperature field in the frequency domain, and directly inversely deduces the optimal heat flux distribution by performing inverse convolution on the measured temperature field, then maps it to a joint control curve of hot air gun scanning speed and power, thus transforming heating control from experience-driven to closed-loop regulation driven by a physical model. The second key technical idea is to dynamically track the glue flow front based on variational level set functions. In traditional processes, the sealant filling status cannot be detected in real time, and defects such as insufficient glue and bubbling can only be inspected afterward. This invention defines the glue flow front as the zero level set of a variational level set function. An energy functional is constructed using interfacial energy, glue viscosity dissipation, and capillary pressure difference. Real-time prediction of the three-dimensional morphology of the glue flow front is achieved through gradient descent iterative evolution, and the hot air gun scanning speed and glue application rate are dynamically adjusted based on the prediction results. When these two technical approaches work synergistically, the inverse heat transfer problem-solving module ensures heating uniformity, and the variational level set function tracking module ensures filling integrity. Together, they constitute a closed-loop sensing and control system for the hot air melting and sealing process, ensuring interfacial bonding quality from multiple physical dimensions simultaneously. Compared to traditional single open-loop control, this system offers higher stability and reliability.

[0051] It should be noted that this invention also solves the following technical problem: During the hot air melting and sealing process of color steel magnesium composite panels, the moisture absorbed by the magnesium substrate during storage or construction rapidly vaporizes during the hot air heating stage, generating local vapor pressure and forcing the interfacial adhesive layer to bubble. Existing processes cannot assess the internal moisture content distribution of the panel in advance, resulting in random bubbling defects that are difficult to prevent. In other words, existing technologies suffer from the technical problem of difficulty in preventing interfacial bubbling defects caused by the vaporization of residual moisture in the magnesium substrate. This invention adds a gradient pre-drying treatment step before the hot air melting and sealing process, placing the panel in an environment with a temperature of 45°C and a relative humidity of 30%RH for pretreatment. Simultaneously, a near-infrared transmission spectroscopy online detection device is used to assess the moisture content field of the panel in real time. For high-humidity areas with a moisture content higher than 0.5%, the pretreatment time is preferentially extended, reducing the overall moisture content of the panel to a safe range, thus eliminating the root cause of interfacial bubbling at the material condition level.

[0052] Furthermore, during the hot air heating process, the galvanized layer of the color steel sheet undergoes accelerated oxidation at high temperatures, forming a zinc oxide film. Since the surface energy of zinc oxide is lower than that of zinc, the wettability of the fire-retardant sealant at the edge of the color steel sheet drops sharply, significantly reducing its bonding strength. However, hot air heating is a necessary condition for activating the fluidity of the adhesive, creating an inherent contradiction. Existing processes cannot simultaneously satisfy both the requirements of sufficient adhesive melting and controllable oxide film thickness. In other words, existing technologies suffer from a technical problem of unstable bonding strength in fire-retardant sealant due to the contradiction between hot air heating and galvanized layer oxidation. This invention addresses this contradiction through a hot air heating sequence decoupling process. First, concentrated heating rapidly melts the adhesive and activates its flow; then, immediate heat removal and pressing within a specified time minimizes the total heating time. Simultaneously, a silane coupling agent is added to the fire-retardant sealant to improve its wetting and spreading rate on the zinc oxide surface. This collaborative approach addresses the aforementioned contradiction from both process sequence and material perspectives.

[0053] Specifically, the principle of this invention is as follows: The fundamental reason why this invention can solve the above-mentioned technical problems is that the heat transfer process of the thin-walled structure of color steel plate satisfies the principle of linear superposition, that is, the temperature field generated by any heat flux distribution can be expressed as the superposition of the responses of heat sources at each point, and the Green's function is precisely the transfer kernel function describing the impact response of the heat source at this point. Based on this physical characteristic, the inverse problem of deriving the required heat flux distribution from the measured temperature field is equivalent to the inversion operation of the frequency domain response matrix of the Green's function in the frequency domain. Applying Tikhonov regularization can effectively suppress the inherent ill-conditioned amplification noise problem of the inverse problem, ensuring the stability and physical realizability of the solution results. After the obtained optimal heat flux distribution is converted into a joint control curve of the hot air gun scanning speed and power through piecewise linear interpolation, the hot air gun can output a heat flux density that matches the target temperature field at different positions, thereby eliminating the non-uniform temperature gradient effect caused by the Gaussian heat flux distribution at the physical mechanism level, making the glue melting process consistent throughout the edge area, and thus ensuring the stable bonding quality of the interface.

[0054] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0055] The specific implementation of step S01 is as follows: The color-coated magnesium composite sheet is pretreated in an environment with a temperature of 45℃ and a relative humidity of 30%RH for 4 to 6 hours. Simultaneously, a near-infrared transmission spectroscopy online detection device is used to evaluate the moisture content field of the sheet in real time. The near-infrared transmission spectroscopy online detection device utilizes the transmission characteristics of the magnesium substrate to the near-infrared spectrum. By detecting the spectral absorption differences at a preset wavelength, it quantitatively inverts the internal moisture content and its spatial distribution of the sheet, outputting a two-dimensional distribution map of the sheet's moisture content field. Areas with a moisture content higher than 0.5% are marked as high-humidity areas, and the pretreatment time for these high-humidity areas is extended until their moisture content is lower than 0.5% before proceeding to subsequent steps.

[0056] The specific implementation of step S02 is as follows: The temperature field at the edge of the color steel plate is acquired using an infrared thermal imager at a frame rate of 100 frames per second. The acquired temperature field is then input into the Green's function convolution hot air non-uniform heating heat transfer inverse problem solving module. The principle of the Green's function convolution hot air non-uniform heating heat transfer inverse problem solving module is as follows: The heat transfer process of the thin-walled color steel plate is modeled as a linear partial differential equation system. The Green's function describes the temperature impact response generated by an arbitrary point heat source acting on the thin-walled structure, i.e., the transfer kernel function between the point heat source input and the temperature field output. Let the discrete spatial nodes at the edge of the plate have a total of... indivual, The empirical value is to arrange a node every 5mm along the edge of the board. Temperature response at each node With heat flux density at each node The relationship between them is determined by the Green's function matrix. The equation for the positive problem is described as follows:

[0057] ;

[0058] In the formula, For the first Temperature of each node, in °C; For reference temperature, the target temperature is set at 120℃; Green's function matrix The Line number Column element, representing the first When a unit heat flux density is applied to the node, the first Temperature response generated by each node, in units Its value was obtained by numerically solving the linear partial differential equation for heat transfer in thin walls by applying a unit pulse heat source boundary condition. After applying unit pulses to each node, a complete matrix is ​​assembled. ; Let the reference value of the Green's function be taken. The largest absolute value among matrix elements, in units ; For the first Heat flux density at each node, in units ; For reference heat flux density, the empirical value is taken as... ; For the first Measurement noise term for each node, in °C. Green's function matrix. The matrix form is as follows:

[0059] ;

[0060] In the formula, for Viglin function matrix, unit .

[0061] The temperature field column vector is obtained by using a forward Fourier transform. By switching to the frequency domain, the frequency domain temperature field signal is obtained. ,in for A 3D real column vector, in °C; for A complex column vector, in °C; a synchronous pair of Green's functions matrix. The frequency domain response matrix is ​​obtained by performing a Fourier transform. , for 3D complex matrix, unit Introducing Tikhonov regularization constraints in the frequency domain, the optimal heat flux distribution frequency domain solution is obtained. The solution formula is expressed as follows:

[0062] ;

[0063] In the formula, The frequency domain solution vector of the optimal heat flux distribution, in units. ,for 3D complex column vector; for The conjugate transpose of the unit matrix. ; for 3D identity matrix, dimensionless; The optimal regularization parameter, in units The optimal regularization parameter is adaptively determined by the L-curve method. The L-curve method determines the optimal regularization parameter by plotting a trade-off curve between the norm of the solution and the norm of the residual and selecting the point of maximum curvature. The formula is as follows:

[0064] ;

[0065] In the formula, ,in For residual vectors The Euclidean norm; ,in For regularization vectors The Euclidean norm; For the residual vector, The column vector of the target temperature field, in °C; For the corresponding regular expression parameters The heat flux density solution vector at time , in units ; For regularization parameters, in units ; Let be the curvature of the L-curve, which is dimensionless; To make the curvature The optimal regularization parameter for finding the maximum value, in units .

[0066] After solving in the frequency domain, the optimal heat flux distribution is restored by inverse Fourier transform. ,in For continuous spatial coordinates along the edge of the board, units ; The optimal heat flux density is continuously distributed along the edge, per unit Then, piecewise linear interpolation is used to map the heat flux distribution at discrete nodes to the scanning speed of the hot air gun. With output power The joint control curve and the piecewise linear interpolation formula for the scan rate are expressed as follows:

[0067] ;

[0068] The piecewise linear interpolation formula for output power is expressed as follows:

[0069] ;

[0070] In the formula, For position Hot air gun scanning speed, unit ; For reference scanning speed, the empirical value is taken as... ; and The first With the Scanning speed at discrete nodes, in units , based on optimal heat flux distribution The calibration curve was obtained by converting the heat flux-velocity calibration curve of the hot air gun. The calibration curve was obtained during the equipment debugging phase by controlling different scanning speeds and measuring the corresponding heat flux density. and The first With the Spatial coordinates of discrete nodes, in units ; For position Output power of hot air gun, unit ; For reference output power, an empirical value of 500 is used. ; and The first With the Output power at discrete nodes, in units The value is also derived from the calibration curve. The single solution time is controlled within 50ms, meeting the requirements of real-time process control.

[0071] The specific implementation of step S03 is as follows: According to the joint control curve of the scanning speed and power of the hot air gun, a 5mm wide area on the edge of the color steel plate is heated with 120℃ hot air for 5 seconds to melt and activate the flow of residual glue, and the pressing is completed within 5 seconds. After pressing, the temperature field of the edge of the color steel plate is collected again. If the temperature uniformity deviation exceeds ±5℃, the temperature field is re-input into the Green's function convolution hot air non-uniform heating heat transfer inverse problem solving module, and steps S02 and S03 are executed again until the temperature uniformity meets the requirements.

[0072] The specific implementation of step S04 is as follows: the sealant flow front is defined as a variational level set function. The zero level set, i.e., the flow front surface satisfies The glue-filled area meets the requirements. The unfilled area satisfies ,in , , Three-dimensional spatial coordinates, units ; Evolution time, in units ; For variational level set functions, unit Its absolute value represents the signed distance from any point in space to the surface of the glue flow front. A generalized energy functional is constructed using three physical mechanisms: interfacial energy, glue viscosity dissipation rate, and capillary pressure difference. The viscous dissipation term characterizes the energy dissipated by the flow of glue per unit time, with dimensions of . The capillary pressure difference term and the interfacial energy term both have dimensions. These three terms together constitute the generalized energy functional describing the driving force of the glue flow front evolution. The generalized energy functional is expressed as follows:

[0073] ;

[0074] In the formula, For generalized energy functionals, units ; For reference energy values, empirical values ​​are taken as follows: ; The interfacial energy density between the adhesive and the substrate, in units of... The contact angle was determined through an experiment. The experimental steps were as follows: Step 1, fire-retardant sealant was melted at 120℃ and then dripped onto the surface of a magnesium substrate; Step 2, the droplet profile image was acquired using a contact angle measuring instrument and the contact angle was fitted. Step 3: Calculate the interfacial energy density based on the Yang-Dupré equation. ,in For solid-gas interface energy, The solid-liquid interface energy was obtained by fitting contact angle data using the Owens-Wendt method. The smoothed Dirac function in the variational level set method is concentrated in... Nearby, workplace Its function is to confine the interfacial energy integral to the curved surface at the front of the glue flow; The gradient magnitude of the variational level set function, in units. ; For volume infinitesimal elements, unit ; This refers to the dynamic viscosity of the adhesive, expressed in units of... The values ​​were determined by a rotational rheometer under heating conditions. For reference time, the empirical value is taken as... This is used to multiply the viscous dissipation rate term by the dimension of time, thereby unifying the dimensions of the term. ; For the Herveside function, in the glue-filled area ( Take 1 in the unfilled area ( (Take 0, dimensionless) For glue flow rate field gradient magnitude, unit ; For glue flow velocity field, unit The solution is given by the simplified solution of the Stokes equation; Capillary pressure difference, unit This is calculated using the Laplace-Young equation; It is a three-dimensional computational domain.

[0075] Generalized energy functional Taking variational values ​​and iteratively updating the time step along the negative gradient direction to satisfy the Courant-Friedrich-Lévy condition. The evolution equation of the variational level set function is expressed as follows:

[0076] ;

[0077] In the formula, For reference length, the empirical value is... ; For generalized energy functionals to dimensionless variational level set functions The variational derivative, unit This makes the right side of the equation dimensionless, consistent with the left side. Each iteration simultaneously solves the Fourier heat conduction equation to obtain updated temperature field values, and solves the simplified Stokes equation to obtain the glue flow velocity field. Update the values ​​and perform a reinitialization operation to maintain the signed distance property of the variational level set function. The reinitialization equation is expressed as follows:

[0078] ;

[0079] In the formula, To reinitialize pseudo-time, units ; For reference, the pseudo-time is taken as the empirical value. ; For symbolic functions, Take 1 at time. Take -1 at time, When the time is 0, it is dimensionless; For the dimensionless gradient modulus, this term equals 1 when the variational level set function maintains its sign distance characteristic. Reinitialization corrects this term to 1 to prevent numerical dispersion from distorting the glue flow front. Time step. The Courant-Friedrich-Lévy condition must be satisfied, as expressed in the following formula:

[0080] ;

[0081] In the formula, here Take the characteristic velocity scalar value of the glue flow velocity field, in units ; For time step, unit ; Spatial grid size, unit ; The upper limit for stability is set to 1 by default for explicit formats. Adaptive mesh refinement technology is used to refine the mesh to a resolution of 5μm in the region near the glue flow front, enabling real-time prediction of the 3D morphology of the glue flow front. If the glue flow front offset exceeds 0.2mm, the hot air gun scanning speed is reduced and the glue application rate is decreased before re-executing step S03.

[0082] The specific implementation of step S05 is as follows: A fire-retardant sealant containing 0.5% silane coupling agent is applied along the edge of the color steel plate, with a width of 8mm and a thickness of 3mm. This creates a triangular support structure with the color steel plate and the magnesium substrate, forming an approximately triangular cross-section. Simultaneously, it establishes adhesion in two directions with the side of the color steel plate and the upper surface of the magnesium substrate, both blocking moisture penetration and providing bending moment support. After sealing, the sealant is sequentially cured at 30℃ for 8 hours for initial curing, at 45℃ for 12 hours for stabilization, and at 25℃ for 24 hours for full curing. This three-stage gradient curing treatment works synergistically to ensure that the interlayer peel strength reaches a requirement of not less than 1.5kN / m.

[0083] The specific implementation method of step S06 is as follows: The interlayer peel strength and edge sealing quality of the color-coated magnesium composite panels that have undergone the three-stage gradient curing process are tested. Panels with an interlayer peel strength lower than 1.5 kN / m are returned to step S01 for reprocessing; panels showing water seepage at the edge seal after the water immersion test are returned to step S01 for reprocessing to ensure product quality is up to standard.

[0084] To better understand and implement this invention, the following is a specific application scenario of the invention, Example 2: In order to solve the problem of interlayer water seepage and local blistering in the edge sealing of color steel magnesium composite panels in actual engineering, the technicians built a complete hot air melting sealing test environment. By implementing the full process of this invention on a batch of color steel magnesium composite panels with a specification of 1200mm×600mm, the synergistic control effect of each key process parameter was verified.

[0085] A total of 20 sheets were prepared for processing. The steel surface was 0.5 mm thick, and the magnesium substrate was 12 mm thick. After scanning the initial moisture content of the sheets using an online near-infrared transmission spectroscopy device, it was found that 7 sheets had localized high-humidity areas, with the highest moisture content reaching 1.3%. These high-humidity areas were mainly concentrated at the four corners and the central edge of the sheets. Technicians placed all the sheets in a pretreatment environment at 45℃ and 30% RH, extending the pretreatment time for the high-humidity areas. Ultimately, the moisture content of all sheets was reduced to below 0.5%, with the area with the highest moisture content decreasing to 0.38% after extended treatment, meeting the requirements for proceeding to the next process.

[0086] In the inverse problem of non-uniform heating and heat transfer in Green's function convolution, technicians used an infrared thermal imager to collect the temperature field at the edge of the sheet metal at a frame rate of 100 frames per second, and input the collected temperature field data into the solution module. Due to the difference in thermal conductivity of the material and the heat dissipation effect at the end of the sheet metal, the initial temperature field in the edge area of ​​the color steel sheet exhibits a significant non-uniform distribution, with a maximum local temperature difference of 18℃, which does not meet the temperature uniformity requirements for subsequent pressing. The solution module transforms the temperature field into the frequency domain through a forward Fourier transform, performs Tikhonov regularization, and automatically determines the optimal regularization parameters using the L-curve method. The single solution time is 43ms, which meets the real-time control requirements. The output heat flux distribution is converted into a joint control curve of the hot air gun scanning speed and power after piecewise linear interpolation. The hot air gun power distribution range is 850~1200. The scanning speed range is 8–22. .

[0087] The heating operation was performed according to the joint control curve, using 120℃ hot air to heat a 5mm wide area on the edge of the color steel plate for 5 seconds. After heating, pressing was completed within 5 seconds. After pressing, the edge temperature field was collected again. Upon inspection, the temperature uniformity deviation of 18 out of 20 plates was controlled within ±5℃. The temperature uniformity deviations of the other two plates were ±6.2℃ and ±7.8℃, respectively, exceeding the threshold. The temperature field of these two plates was re-input into the solution module, and the heating and pressing process was repeated. After the second processing, the temperature uniformity deviations were reduced to ±3.9℃ and ±4.5℃, respectively, meeting the requirements.

[0088] In the glue flow front prediction stage, the variational level set function defines the glue flow front as a zero level set. Adaptive mesh refinement technology is used in the region near the glue flow front to refine the mesh to a resolution of 5μm, enabling real-time three-dimensional morphological prediction of the glue filling process. For example... Figure 3 As shown, the offset of the glue flow front at different positions evolved over time. Three out of 20 boards exhibited an offset exceeding 0.2 mm during the glue application process, with the largest offset reaching 0.31 mm. Based on this, technicians promptly adjusted the hot air gun scanning speed and reduced the glue application rate. After re-performing the pressing operation, the offsets all returned to within 0.12 mm. The measured results of key process parameters for each board are shown in Table 1.

[0089] Table 1. Measured Results of Key Process Parameters for Each Plate

[0090]

[0091] A fire-retardant sealant containing 0.5% silane coupling agent was applied along the edges of the color steel plate, with a width of 8mm and a thickness of 3mm. This created a triangular support structure between the sealant, the color steel plate, and the magnesium substrate. This triangular support structure formed a bidirectional adhesive force on the cross-section, both blocking moisture penetration and providing bending moment support for the edges. Subsequently, all plates underwent a three-stage gradient curing treatment: initial curing at 30℃ for 8 hours, stabilization at 45℃ for 12 hours, and full curing at 25℃ for 24 hours. The synergistic effect of these three stages fully released the internal stress of the adhesive, avoiding the risk of warping caused by rapid high-temperature curing.

[0092] After curing, all boards were subjected to interlayer peel strength test and edge sealing quality test. The test results of interlayer peel strength are shown in Table 2.

[0093] Table 2 Statistical Table of Interlayer Peel Strength Test Results

[0094]

[0095] The interlaminar peel strength of all 20 boards is not less than 1.5. No need to return them for reprocessing. After the water immersion test, the edge seals were inspected one by one, and none of the 20 boards showed any water leakage, indicating that the sealing quality was up to standard.

[0096] This invention represents a fundamental improvement in technical principles compared to traditional hot air melting and sealing methods. Traditional methods rely on operators manually controlling the scanning speed and power of the hot air gun based on experience, resulting in an open-loop operation of heat application. This makes it impossible to perceive the non-uniformity of the actual temperature field at the edge of the board, leading to difficulties in eliminating local overheating or underheating defects. This invention introduces a Green's function convolution inverse problem solving module, transforming temperature field control from open-loop empirical compensation to a physically driven closed-loop optimization. Essentially, it utilizes the Green's function kernel of the linear partial differential equation for thin-wall heat transfer to describe the heat source-to-temperature transfer relationship. In the frequency domain, it uses Tikhonov regularization to stably solve the inverse problem, allowing any non-uniform temperature distribution to be accurately back-calculated into the corresponding heat flow control strategy. This fundamentally eliminates the problem of non-uniform edge temperature gradients caused by the Gaussian distribution of heat flux density. Traditional methods rely entirely on visual inspection and experience during the glue filling process, making it impossible to predict the occurrence of local glue shortages or bubbling. This invention transforms the three-dimensional surface evolution problem of the adhesive flow front into a scalar field energy functional minimization problem using a variational level set method. Adaptive mesh refinement near the front improves the spatial resolution to 5μm, enabling quantitative prediction of the adhesive filling morphology. Real-time intervention is possible once the front deviation exceeds a threshold, mechanistically suppressing edge defects. The three-stage gradient curing process, through phased temperature control, allows the adhesive to form a stable framework in the initial curing stage, complete the chemical reaction in the stabilization stage, and fully release internal stress in the full curing stage. This solves the warping and adhesion failure problems caused by thermal stress accumulation in traditional single-temperature curing. The triangular support structure, through geometric constraints, transforms the stress path of the sealant layer from unidirectional tension to bidirectional synergistic bearing, outperforming traditional flat-mounted sealing structures in both moisture blocking and bending support directions.

[0097] It should be noted that the variables involved in this invention are explained in detail in Tables 3 and 4.

[0098] Table 3. Variable Explanation Table (Part 1)

[0099]

[0100] Table 4. Variable Explanation Table (Part Two)

[0101]

[0102] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A hot air melting and sealing method for color steel sheet and magnesium substrate, characterized in that, Includes the following steps: The color steel magnesium composite sheet is placed in an environment with a preset temperature and relative humidity for a preset time for pretreatment. At the same time, a near-infrared transmission spectroscopy online detection device is used to evaluate the moisture content field of the sheet in real time. Areas with moisture content higher than a preset moisture content threshold are marked as high humidity areas, and the pretreatment time of the high humidity areas is extended until the moisture content is lower than the preset moisture content threshold. The edge temperature field of the color steel plate is acquired by an infrared thermal imager at a preset frame rate. The edge temperature field of the color steel plate is input into the Green function convolution hot air non-uniform heating heat transfer inverse problem solving module. After forward Fourier transform, Tikhonov regularization Green function frequency domain inverse matrix multiplication and inverse Fourier transform, the optimal heat flux distribution is obtained. The optimal heat flux distribution is then converted into a joint control curve of hot air gun scanning speed and power through piecewise linear interpolation. According to the hot air gun scanning speed and power joint control curve, the color steel plate edge area of ​​preset width is heated by hot air at preset temperature for preset duration to melt and activate the flow of residual glue, and the pressing is completed within the preset pressing time. After pressing, the temperature field at the edge of the color steel plate is collected again. If the temperature uniformity deviation exceeds the preset temperature deviation threshold, the temperature field at the edge of the color steel plate is re-input into the Green's function convolution hot air non-uniform heating heat transfer inverse problem solving module, and the steps of obtaining the optimal heat flow distribution and converting it into the hot air gun scanning speed and power joint control curve and this step are executed again. The glue flow front is defined as the zero level set of the variational level set function. An energy functional is constructed using interfacial energy, glue viscosity dissipation, and capillary pressure difference. The variational level set function is iteratively evolved through gradient descent. An adaptive mesh refinement technique is used to refine the mesh to a preset mesh resolution in the region near the glue flow front, and the three-dimensional morphology of the glue flow front is predicted in real time. If the offset of the glue flow front exceeds the preset offset threshold, reduce the hot air gun scanning speed and reduce the glue application rate before re-executing the heating step. Fire-retardant sealant containing silane coupling agent is applied along the edge of the color steel plate to form a triangular support structure with the color steel plate and magnesium substrate. The sealed color steel magnesium composite plate is then cured sequentially at a first curing temperature for a first curing time, at a second curing temperature for a second curing time, and at a third curing temperature for a third curing time to complete a three-stage gradient curing process. The interlayer peel strength and edge sealing quality of the color steel magnesium composite panels that have completed the three-stage gradient curing process are tested: the panels with interlayer peel strength lower than the preset peel strength threshold are returned to the gradient pre-drying process for reprocessing. Boards that show water seepage at the edge seal after immersion testing are returned to the gradient pre-drying process for reprocessing.

2. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 1, characterized in that, The preset temperature is 45℃, the preset relative humidity is 30%RH, the preset duration is 4 to 6 hours, and the preset moisture content threshold is 0.5%.

3. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 2, characterized in that, The near-infrared transmission spectroscopy online detection device detects the spectral absorption differences at preset wavelengths, quantitatively inverts the internal moisture content and spatial distribution of the board, and outputs a two-dimensional distribution map of the moisture content field of the board.

4. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 3, characterized in that, The preset frame rate is 100 frames per second, and the single solution time of the Green's function convolution hot air non-uniform heating heat transfer inverse problem solving module does not exceed 50ms.

5. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 4, characterized in that, The regularization parameter of the Tikhonov regularization is adaptively selected by the L-curve method, which determines the optimal regularization parameter by plotting a parameterized curve between the regularized residual norm and the regularized subnorm in a double logarithmic coordinate system and selecting the point of maximum curvature.

6. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 5, characterized in that, The piecewise linear interpolation connects the discrete spatial node values ​​of the optimal heat flux distribution piecewise with a linear function between adjacent nodes, and converts them into the hot air gun scanning speed and output power values ​​corresponding to each position on the continuous scanning path.

7. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 6, characterized in that, The preset temperature hot air is 120℃, the preset width area is 5mm wide, the heating time is 5 seconds, the preset pressing time is 5 seconds, and the preset temperature deviation threshold is ±5℃.

8. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 7, characterized in that, The time step of the gradient descent iteration satisfies the CFL condition, which requires that the ratio of the flow field characteristic velocity, the time step, and the spatial grid size does not exceed the upper limit of stability. Each iteration simultaneously solves the Fourier heat conduction equation to obtain updated temperature field values, solves the simplified Stokes equation to obtain updated velocity field values, and performs a re-initialization operation to maintain the sign distance property of the variational level set function.

9. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 8, characterized in that, The preset grid resolution is 5μm, and the preset offset threshold is 0.2mm.

10. The hot air melting and sealing method for color steel sheet and magnesium substrate according to claim 9, characterized in that, The fire-retardant sealant contains 0.5% silane coupling agent by mass, and has a coating width of 8 mm and a coating thickness of 3 mm.

Images