Color photovoltaic module lamination intelligent control method and system based on multi-parameter feedback

Abstract

The invention relates to the technical field of intelligent manufacturing control, and discloses a color photovoltaic module lamination intelligent control method and system based on multi-parameter feedback, and the method comprises the steps: constructing a global objective function containing the curing degree and the color difference; monitoring the deformation acceleration of the coloring layer microstructure in real time by using a spectrum sensor, calculating an optical damage risk factor, dynamically adjusting the weight of a target function according to the optical damage risk factor, and calculating the optimal oscillation frequency in combination with a rheological shear thinning mechanism; and controlling an execution mechanism to apply micro-oscillation pressure waves with the optimal oscillation frequency, and performing quality judgment according to the accumulated cost of the whole process. By monitoring the microstructure deformation acceleration of the coloring layer, the problem of traditional temperature feedback lag is solved, meanwhile, the contradiction between bubble discharge and microstructure protection is solved by utilizing an adhesive film shear thinning mechanism, self-adaptive intelligent regulation and control in the laminating process are realized, optical chromatic aberration caused by excessive pressurization is avoided, and the product quality is improved. And the yield and the reliability of the assembly are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing control technology, specifically to an intelligent control method and system for laminating color photovoltaic modules based on multi-parameter feedback. Background Technology

[0002] With the rapid development of Building Integrated Photovoltaics (BIPV) technology, the application scenarios of photovoltaic modules have expanded from simple power plants to building curtain walls, roof tiles, and other fields, and the market has placed higher demands on the aesthetics of photovoltaic modules. To meet the diverse needs of building appearance, colored photovoltaic modules have emerged. These typically use nano-optical thin films or photonic crystal color-coding layers introduced into the encapsulation structure to display specific colors using the principles of light interference and diffraction. Unlike traditional modules, these modules with fine optical microstructures place extremely high demands on the precision of the lamination process; any minute physical deformation can lead to color differences or appearance patches, affecting the overall building effect.

[0003] Existing photovoltaic module lamination control technologies mainly rely on preset temperature-vacuum-pressure curves, using PID algorithms to adjust the heating plate temperature or airbag pressure. During the lamination process, thermocouples are typically used to collect temperature data, and fixed pressurization timing and pressure values ​​are set based on experience. High temperatures are used to reduce the viscosity of the adhesive film, and high pressure is used to expel air from the gaps between the solar cells, ensuring that the cross-linking and curing degree of the adhesive film meets the standard.

[0004] However, existing lamination control technologies have limitations for color photovoltaic modules with intricate optical structures. Specifically, current lamination processes often struggle to balance the conflict between bubble removal and microstructure protection: typically, high lamination pressure is required to remove bubbles from the cell gaps, but this can easily damage the fragile nanostructures of the coloring layer, leading to severe optical color aberrations or visible patches in the module; conversely, reducing pressure to protect the coloring layer results in insufficient encapsulant filling power and residual bubbles. Furthermore, traditional temperature-feedback-based PID control exhibits significant hysteresis, failing to perceive the dynamic relationship between the encapsulant rheological state and the stress-induced deformation of the microstructure in real time, making precise pressure adjustment difficult at the critical edge of microstructure collapse. Therefore, a multi-parameter feedback-based intelligent lamination control method and system for color photovoltaic modules is urgently needed to address these issues. Summary of the Invention

[0005] To address the problems in related technologies, this invention provides a multi-parameter feedback-based intelligent control method for color photovoltaic module lamination, thereby overcoming the aforementioned technical problems in existing related technologies.

[0006] To solve the aforementioned technical problem, the present invention is achieved through the following technical solution: In a first aspect, embodiments of the present invention provide a method for intelligent control of color photovoltaic module lamination based on multi-parameter feedback, specifically including: reading the raw material parameters of the module to be laminated and constructing a global objective function for model predictive control; during the lamination process, using sensors to collect spectral data of the color-causing layer of the module in real time and calculating an optical damage risk factor; dynamically adjusting the weight allocation in the global objective function in combination with the optical damage risk factor; calculating the optimal oscillation frequency that minimizes the global objective function based on the rheological shear thinning mechanism and damage suppression mechanism; controlling the laminator actuator to apply a micro-oscillation pressure wave with a frequency equal to the optimal oscillation frequency, and calculating the cumulative cost convergence value after the lamination process is completed, comparing it with a preset quality acceptance threshold to determine the module quality.

[0007] As a preferred embodiment of the intelligent control method for color photovoltaic module lamination based on multi-parameter feedback described in this invention, the expression of the global objective function is: ; In the formula, Let be the total cost function. To predict the time domain, for The degree of crosslinking of the adhesive film at any given time. For the target degree of crosslinking, For the predicted color difference value, Here is the color difference penalty function. The frequency of the applied pressure oscillation. To control the rate of change of frequency, The maximum operating frequency allowed by the actuator; These are quality weights, optical weights, and control smoothing weights, respectively. .

[0008] As a preferred embodiment of the intelligent control method for lamination of color photovoltaic modules based on multi-parameter feedback described in this invention, the color difference penalty function adopts an exponential barrier function, expressed as: ; In the formula, The color difference threshold; the color difference value Based on real-time collected color coordinates Color coordinates of the standard chromogenic layer The Euclidean distance between them was calculated.

[0009] As a preferred embodiment of the intelligent control method for color photovoltaic module lamination based on multi-parameter feedback described in this invention, the formula for calculating the optical damage risk factor is as follows: ; In the formula, for Optical damage risk factors at any time The reference center wavelength of the chromogenic layer; The characteristic wavelength peak, The acceleration of the characteristic wavelength drift. Let be the system response time constant. The initial wavelength, This is the maximum permissible wavelength offset. This is the deviation weighting coefficient.

[0010] As a preferred embodiment of the intelligent control method for color photovoltaic module lamination based on multi-parameter feedback described in this invention, the relationship between the optical weight and the optical damage factor is expressed as follows: ; In the formula, for Optical weights at time, The safety threshold inflection point, The sensitivity coefficient, To control the smoothing weights.

[0011] As a preferred embodiment of the intelligent control method for color photovoltaic module lamination based on multi-parameter feedback described in this invention, the formula for calculating the optimal oscillation frequency is: ; In the formula, Based on the base frequency, This is the gain factor calculated based on the shear-thinning principle. This is the braking coefficient calculated based on damage risk.

[0012] As a preferred embodiment of the intelligent control method for lamination of color photovoltaic modules based on multi-parameter feedback described in this invention, the gain factor... The calculation formula is: ; In the formula, The zero-shear viscosity at the current temperature. For ideal viscosity, To adjust the viscosity gain, To obtain the maximum value function, This indicates the lower limit of the safety threshold.

[0013] As a preferred embodiment of the intelligent control method for color photovoltaic module lamination based on multi-parameter feedback described in this invention, the braking coefficient... The calculation formula is: ; In the formula, To protect the attenuation coefficient.

[0014] As a preferred embodiment of the intelligent control method for color photovoltaic module lamination based on multi-parameter feedback described in this invention, the determination of module quality includes: calculating the cumulative cost convergence value based on a global objective function. Set quality acceptance thresholds ;like The component is deemed qualified; if The component was deemed unqualified.

[0015] Secondly, embodiments of the present invention provide a multi-parameter feedback-based intelligent control system for color photovoltaic module lamination, comprising: a model building module for reading raw material parameters of the module to be laminated and constructing a global objective function for model predictive control; a monitoring and evaluation module for real-time acquisition of spectral data and calculation of optical damage risk factors during the lamination process; a weight adjustment module for dynamically adjusting the weight allocation in the global objective function in combination with the optical damage risk factors; a frequency calculation module for calculating the optimal oscillation frequency that minimizes the global objective function; and a collaborative control and judgment module for controlling the laminator actuator to output pressure waves and calculating the cumulative cost convergence value for quality judgment.

[0016] The present invention has the following beneficial effects: 1. This invention transforms the complex nonlinear coupling problem in the lamination process of colored photovoltaic modules into a real-time solvable optimal control problem by constructing a global objective function for multi-objective collaborative optimization. This method overcomes the limitations of traditional control methods that can only adjust single variables such as temperature or pressure. It automatically and dynamically adjusts the control weights of curing quality and optical appearance based on the real-time physical state of the color-forming layer microstructure. This allows the system to apply full pressure during the film curing period to expel air bubbles, and to promptly reduce pressure during the softening danger period to protect the microstructure. Thus, while ensuring no air bubble residue in the module, it effectively avoids the collapse of the color-forming layer microstructure and color difference problems caused by excessive pressure.

[0017] 2. This invention calculates the optimal oscillation frequency by combining the shear thinning principle of rheology with the damage threshold of fracture mechanics. In the low-temperature, high-viscosity stage, the system automatically outputs high-frequency oscillations, utilizing shear force to reduce the viscosity of the adhesive film, achieving rapid leveling and bubble removal under relatively low static pressure. Conversely, in the high-temperature softening stage where microstructure risks are high, the system forcibly suppresses the oscillation amplitude. This control method avoids the shortcomings of traditional single oscillation modes that cannot adapt to the full-temperature range characteristics of the adhesive film, ensuring that in complex and variable lamination environments, it can both accelerate liquefaction and degassing using physical properties and prevent physical damage, thereby improving production efficiency.

[0018] 3. This invention utilizes an embedded spectral sensor to collect the characteristic wavelengths of the chromogenic layer in real time and constructs a quantified optical damage risk factor by calculating the second derivative of wavelength drift. Compared to traditional passive control relying on thermocouple temperature feedback, this invention can directly monitor the physical deformation acceleration of the chromogenic layer microstructure in a pressure field, thereby identifying whether the microstructure is on the critical edge of collapse. This advanced sensing capability provides timely decision-making basis for the control system, enabling it to intervene before irreversible optical damage occurs, thereby reducing the generation of substandard products with color differences.

[0019] 4. This invention not only focuses on the static appearance of the final product, but also uses numerical integration to calculate the cumulative cost convergence value of the entire process for quality judgment. This method can quantitatively assess the risk journey experienced by photovoltaic modules during processing, thereby identifying and intercepting those defective products that, although passing the final appearance inspection, may have hidden internal microcracks or residual internal stress due to excessive high-risk vibrations or critical deformations during processing. Compared with traditional visual inspection, quality control is more stringent, which helps to improve the long-term weather resistance and reliability of photovoltaic modules leaving the factory.

[0020] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 The present invention provides a flowchart of a smart control method for laminating color photovoltaic modules based on multi-parameter feedback.

[0023] Figure 2 This invention provides a schematic diagram of a module for a color photovoltaic module lamination intelligent control system based on multi-parameter feedback.

[0024] Figure 3 This is a cross-sectional schematic diagram of the internal structure and sensor arrangement of the laminator provided by the present invention.

[0025] Figure 4 The dynamic adjustment curve of the global objective function weights provided by this invention.

[0026] Figure 5 The shear-thinning gain curve provided by this invention.

[0027] Figure 6 The damage braking mechanism curve provided by the present invention.

[0028] Figure 7 A comparison diagram of typical lamination process control timing waveforms provided by the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Example 1 In the field of building-integrated photovoltaics (BIPV), colored photovoltaic modules achieve color display by introducing nano-optical thin films or photonic crystal structures. However, existing lamination processes often struggle to balance the conflict between bubble removal and microstructure protection: the high pressure applied to remove bubbles can easily damage the microstructure of the coloring layer, leading to color differences, while reducing pressure to protect the coloring layer can result in bubble residue.

[0031] To solve the above technical problems, such as Figure 1 As shown, Embodiment 1 of the present invention provides an intelligent control method for laminating colored photovoltaic modules based on multi-parameter feedback. Specifically, Embodiment 1 takes the lamination process of a colored photovoltaic module for a BIPV curtain wall as an example: the colored photovoltaic module structure includes ultra-white glass, POE film, photonic crystal coloring layer, solar cells, and a backsheet. The on-site laminator is equipped with a distributed fiber optic spectral sensor and a piezoelectric ceramic micro-oscillation pressurization module.

[0032] In the specific implementation of Example 1: First, the raw material parameters of the module to be laminated are read, and a global objective function for model predictive control is constructed. This method transforms the complex nonlinear and multi-physics coupling problem in the lamination process of colored photovoltaic modules into an optimal control problem that can be solved in real time. This overcomes the shortcomings of traditional single-variable regulation, which cannot quantitatively assess the damage to the microstructure caused by aggressive pressurization, and achieves the best balance between curing quality and optical appearance. Second, during the lamination process, the spectral data of the coloring layer of the module is collected in real time using sensors to calculate the optical damage risk factor. This method overcomes the hysteresis of traditional thermocouple temperature feedback by monitoring the deformation acceleration of the coloring layer microstructure and uses the spectral differential signal to proactively perceive the microstructure damage trend, providing an objective and timely decision-making basis for the control system. Then, the weight allocation in the global objective function is dynamically adjusted in conjunction with the optical damage risk factor. This invention utilizes the nonlinear switching characteristics of the Sigmoid function to automatically allocate control weights according to the real-time physical state of the microstructure, overcoming the rigidity problem caused by traditional fixed weights that prevent the system from applying full pressure during the film hardening period and from timely releasing pressure during the softening danger period. It can achieve a smooth switch of control strategy from efficiency priority to appearance priority. Subsequently, based on the rheological shear thinning mechanism and damage suppression mechanism, the optimal oscillation frequency that minimizes the global objective function is calculated. This method combines the shear thinning principle of rheology with the damage threshold of fracture mechanics to achieve intelligent control of accelerated liquefaction venting and risk avoidance protection, dynamically finding the optimal balance between flow efficiency and structural safety, and ensuring the dual optimization of bubble discharge rate and appearance yield. Finally, the laminator actuator is controlled to apply a micro-oscillation pressure wave at the optimal oscillation frequency, and the cumulative cost convergence value is calculated after the lamination process is completed. This value is then compared with the preset quality acceptance threshold to determine the component quality. This method identifies and intercepts hidden defects that may appear to be qualified but have internal micro-cracks or residual stress by quantitatively evaluating the cumulative cost of the entire process. This achieves a more stringent quality acceptance than traditional visual inspection and can improve the long-term reliability of the products leaving the factory.

[0033] Furthermore, to better illustrate the technical solution of Embodiment 1 of the present invention, a detailed description of the intelligent control method for lamination of color photovoltaic modules based on multi-parameter feedback is provided, specifically including the following: S1. Construct the predictive control model and the global objective function, which includes the following sub-steps: S11. System initialization: Read the raw material parameters of the components to be laminated and construct the global objective function of model predictive control (MPC). : ; In the formula, For prediction of the time domain The total cost function within; for The degree of crosslinking of the adhesive film at any given time; Target degree of crosslinking; The predicted color difference value; Here is the color difference penalty function; The frequency of the applied pressure oscillation; To control the rate of change of frequency; This refers to the maximum permissible operating frequency of the micro-oscillation actuator in the laminator; These are quality weights, optical weights, and control smoothing weights, respectively. .

[0034] S12, Set the color difference penalty function The specific form is as follows. In this embodiment, an exponential barrier function is used. When the color difference approaches the color difference threshold, the penalty value increases exponentially, forcing the system to adjust its strategy: ; in, This is the color difference threshold.

[0035] In this embodiment, for any time The system calculates the real-time color difference value based on the following formula. : ; in, Real-time color coordinates acquired and converted by a spectral sensor; The color coordinates of the standard colorimetric layer pre-stored in the system under ideal conditions.

[0036] Furthermore, color difference threshold The colorimetric limit is determined based on the human eye's perception limit using the International Commission on Illumination (CIE) standard colorimetric system. This is based on CIE standards and the human eye's ability to distinguish color differences. In a color space, color difference Represents the Euclidean distance between the sample and the standard, such as At that time, the human eye can hardly distinguish them (no color difference); At that time, only a professionally trained eye can detect subtle differences under specific light sources (BIPV high-end curtain wall standard). At times, the difference is noticeable to the average person, but it is acceptable when the modules are not adjacent (standard for ordinary photovoltaic modules). At that time, there were obvious color difference patches (unacceptable). Therefore, in this embodiment, a color difference threshold can be set. .

[0037] In this embodiment 1, a global objective function for multi-objective collaborative optimization is constructed. This method transforms the complex nonlinear and multi-physics coupling problem in the lamination process of color photovoltaic modules into an optimal control problem that can be solved in real time. It overcomes the limitations of traditional PID control or single-logic control, which can only adjust a single variable and cannot quantitatively assess the damage to the microstructure caused by aggressive pressurization. Instead, it achieves model-based proactive planning of the optimal control path, ensuring that the control system maintains the best balance between curing quality and optical appearance throughout the entire time domain.

[0038] S2. Real-time monitoring and optical damage risk assessment, specifically including the following sub-steps: S21. During the lamination heating and pressure holding stages, an embedded spectral sensor is used to collect the spectral response curves of key locations of the component in real time, and the characteristic wavelength peaks are extracted. Specifically, the probe of the spectral sensor adopts a 5-point layout, covering the geometric center and four corner points of the coverage component; combined with the attached... Figure 3 The diagram shows the internal structure of the laminator and the cross-sectional layout of the sensors. The probes need to be installed facing the color-causing layer area of ​​the component to ensure that the deformation state of the color-causing layer microstructure can be directly sensed.

[0039] S22. The system calculates the second derivative of the spectral drift and, in conjunction with the current wavelength deviation, calculates the optical damage risk factor. : ; In the formula, The reference center wavelength of the chromogenic layer; The acceleration of the characteristic wavelength drift represents the tendency of the microstructure to collapse; The system response time constant; The initial wavelength; This is the maximum permissible wavelength offset; This is the deviation weighting coefficient.

[0040] For example: In this embodiment 1, the system response time constant The step response system identification method is used to determine it, as follows: First, set the base pressure with the laminator unloaded. Send an amplitude of [value] to the piezoelectric ceramic actuator. A step pressure command, for example, an instantaneous step from 0.05 MPa to 0.06 MPa. Using the sampling frequency. The high-frequency pressure sensor records the response curve of the actual pressure. Secondly, the pressure response curve was extracted to reach a steady-state value. (Right now The time required is denoted as the physical time constant. Finally, considering the sampling period of the control system... To determine the final .

[0041] In this embodiment 1, the reference center wavelength of the chromogenic layer is... The method used is the online self-calibration method for material input, as follows: exist Before the laminator closes, a spectral sensor performs a rapid multi-point scan of the coloring layer of the component to be laminated, for example, a 5-point scan: the center and four corners. Gaussian fitting is applied to the spectral curve of each scan point to extract the peak wavelength with the highest reflectance. Calculate the arithmetic mean of the wavelengths at multiple points, and use it as the reference center wavelength for the chromogenic layer of this batch of components: .

[0042] In this embodiment 1, the deviation weighting coefficient The risk is determined using a risk normalization method based on destructive experiments, as follows: First, the critical failure acceleration of the material is determined through preliminary experiments. and critical color difference deviation In a laboratory setting, high-frequency, high-amplitude vibrations are applied to the chromogenic layer. The maximum value of the spectral drift acceleration at the instant when irreversible microstructural collapse occurs (i.e., the spectrum cannot be restored after the pressure is removed) is recorded and denoted as the critical failure acceleration. This value represents the physical limit of dynamic risk. Meanwhile, according to the CIEDE2000 colorimetric standard, the color difference is measured... The corresponding average wavelength shift is denoted as the critical chromatic difference deviation. This value represents the acceptable minimum for static appearance.

[0043] Secondly, when the system's dynamic acceleration reaches At that time, its risk contribution value should reach the static deviation. If the risk contribution values ​​are equal, then they should all be considered as 100% risk.

[0044] Finally, substitute the two terms into the formula to achieve balance: Dynamic term limit values: ; Static term limit values: ; make Solving for Objective calculation formula: .

[0045] In this embodiment 1, the system sampling period is set. Response time constant Chromogenic reference center wavelength Maximum permissible wavelength offset Deviation weighting coefficient Specifically, for example: at a certain moment The sensor measures the characteristic wavelength. 552nm, initial wavelength The wavelength was 551 nm at the previous moment and 550.2 nm at the moment before that. The system uses the finite difference method to calculate the wavelength drift acceleration: Simultaneously calculate the static wavelength deviation: Calculate the optical damage risk factor. The calculation results indicate that the current wavelength has shifted and the drift rate is accelerating, and the chromogenic layer microstructure is in a state of slight stress deformation.

[0046] The above example illustrates that, in the lamination heating and pressure holding process, the present invention, such as Figure 4 As shown, a characteristic wavelength of the chromogenic layer is acquired in real time using an embedded fiber optic spectral sensor. By calculating the weighted combination of the second derivative of wavelength drift and static deviation, a quantified optical damage risk factor is constructed. It is used to monitor the physical deformation state of the chromogenic layer microstructure in a pressure field in real time and identify whether the microstructure is on the critical edge of collapse. This overcomes the lag of traditional thermocouple temperature feedback and uses spectral differential signals to perceive the damage trend of microstructure in advance. This provides an objective and timely decision-making basis for the subsequent control system to achieve a dynamic balance between pursuing curing quality and protecting optical appearance.

[0047] S3, combined with optical damage risk factors Dynamically adjust the global objective function weight Specifically, it includes the following sub-steps: S31, Controlling smoothing weights The smoothness of the control quantity is determined to prevent damage to the piezoelectric actuator from high-frequency, severe vibrations. This is determined using the frequency response test boundary method. Specifically, the system's cutoff frequency is measured by performing a frequency sweep test on the laminator's pressurization system. Based on the bandwidth limitation in control theory, set... If the value is constant to ensure that the rate of change of the control command always remains within the linear response region, then: .

[0048] S32, Optical Weight Should be based on the risk of damage The increase is not linear but rather rapid. When the risk approaches its limit... The objective function should dominate, forcing the controller to sacrifice the curing speed. In exchange for optical safety. (Building) and The functional relationship is as follows: ; In the formula, Indicates leaving and Total weight space allocated; This represents the inflection point of the safety threshold. is a sensitivity coefficient used to control the steepness of weight switching. This formula characterizes: when... (Safe) time, Approaching 0, the system ignores minute color differences and accelerates curing at full speed; when near (In dangerous) situations, Tend to ,at this time When the system is compressed to 0, it completely switches to protection mode.

[0049] Furthermore: .

[0050] For example: In this embodiment 1, the safety threshold inflection point This represents the critical physical boundary point where the chromogenic layer microstructure transitions from the elastic deformation region to the plastic deformation region or the damage region; below this value, the microstructure deformation can self-recover; above this value, irreversible optical damage will occur. The determination method is as follows: First, a coloring layer sample from the same batch as the production line was selected and placed on a precision pressure testing bench in the laboratory. The sample was tested in extremely small increments. Pressure was gradually applied, held for 5 seconds after each application, and then completely released to 0. After each release, the spectral wavelength was checked to see if it completely returned to its initial value. Calculate the residual wavelength deviation. .

[0051] Then, plot the "load pressure - residual deviation" curve. When For the first time, the system measurement noise floor of the spectrometer exceeded that of the spectrometer (e.g. When the value of a material is 0, the state corresponding to that moment is determined to be the elastic limit point of the material.

[0052] Finally, the second derivative of the spectrum and the wavelength deviation measured at the elastic limit point are substituted into the equation in step S2. The calculation formula yields a result that is objectively defined by physics. .

[0053] In this embodiment 1, the sensitivity coefficient This determines the steepness of the control weight switching. To objectively determine... , introduce 3 Full protection criterion: This requires that when the risk of damage reaches the theoretical maximum value ( When optical weights are used, The normalization coefficient must reach a statistically complete saturation state (i.e., This ensures the system can provide maximum protection when on the verge of collapse. The specific method is as follows: First, define This is the point of absolute physical destruction, where the microstructure completely collapses. At this point, the Sigmoid function is required to... The output value reached Based on the normal distribution Principle: Confidence level ,set up .

[0054] Then, substitute the boundary conditions into the equations for the Sigmoid part: In the formula, for The maximum normalized value.

[0055] Solve the above equation by performing a logarithmic transformation. : ; ; ; .

[0056] In this embodiment 1, by combining optical damage risk factors Dynamically adjust the global objective function weight By leveraging the nonlinear switching properties of the sigmoid function, the controller's attention is automatically allocated based on the real-time physical state of the microstructure. Specifically, for example, control smoothing weights are set. Then the weight space for quality and appearance is: Set a safety threshold In the early stages of lamination At that time, the microstructural deformation of the chromogenic layer was detected to be small, and the risk factor was low. The optical weights were calculated. At this time, the quality weight The objective function is based on the degree of solidification. As the primary driver, the system will output high-frequency oscillations to achieve high curing efficiency; while During the high-temperature softening period, if a sudden increase in risk factors is detected... , It will rise rapidly to approximately ,lead to sudden drop The objective function is transformed into color difference. Under the dominant mode, the system will switch to a conservative mode to actively suppress oscillation amplitude to ensure the safety of the optical appearance. This method overcomes the rigidity problem in traditional lamination control, where fixed weight parameters prevent the system from applying full pressure during the film curing period and from releasing pressure in time during the softening danger period. This achieves a smooth switch in control strategy from efficiency-first to appearance-first, ensuring maximum production efficiency within the safety window while preventing microstructure collapse.

[0057] S4. Dynamically calculate the optimal oscillation frequency based on the shear-thinning mechanism. This is to achieve the global objective function described in step S1 within the control period. Minimize the global objective function using analytical decomposition. Minimize the optimal oscillation frequency Specifically, it includes the following steps: S41, such as Figure 5 As shown, to minimize the cost of uncured or bubble-residual substances in the objective function, the control law must exhibit a positive gain characteristic; that is, the higher the viscosity, the higher the shear oscillation frequency required to thin the film. The system reads the zero-shear viscosity at the current temperature. Calculate the current viscosity and the ideal viscosity. The deviation ratio is used to calculate the gain factor required to thin the film using the Carreau shear thinning principle. : ; In the formula, To adjust the viscosity gain, To obtain the maximum value function, This represents the lower safety threshold, used to prevent control failure caused by calculating negative or zero values.

[0058] For example, viscosity adjustment gain It depends directly on the power law exponent of the film material. That is, if the material has a very strong shear thinning effect ( (Very small), with a slight increase in frequency, the viscosity drops rapidly; at this point, controlling the gain... It should be relatively small; if the material shear thinning effect is weak ( (Approaching 1), a significant increase in frequency is needed to reduce viscosity, at which point the control gain... It should be relatively large. Therefore, This can be reflected by the following formula: .

[0059] For example: The flow behavior index of the current batch of POE film was measured. It exhibits significant shear-thinning properties, therefore This means that if the viscosity is detected to be 10% too high, the system will automatically increase the oscillation frequency by 20% to accurately match the physical thinning law of the material, thus ensuring the stability of control accuracy between different batches of materials.

[0060] S42, such as Figure 6 As shown, to minimize the chromatic aberration risk cost in the objective function, the control law must exhibit negative suppression characteristics; that is, the greater the damage risk, the lower the allowable oscillation frequency should be to suppress chromatic aberration. The system reads the current optical damage risk factor. Calculate the braking coefficient using an exponential function : ; In the formula, To protect the attenuation coefficient.

[0061] For example, the protection attenuation coefficient The limits are determined jointly by the hardware physical limits of the actuator and the material limits of the coloring layer, using the following method: First, the system reads the hardware specifications of the laminator's micro-oscillation actuator to obtain its maximum output frequency. (e.g., 50Hz) and minimum effective resolution frequency (For example, 0.5Hz). The ratio of the two. The system's hardware dynamic range is defined; secondly, the optical damage risk factor is set. The critical damage threshold of the material coloring layer calibrated in step S2 is reached. At that time, regardless of the rheological requirements, the control system must force the output frequency to be reduced from its maximum value. Attenuation to minimum value This ensures that no destructive vibrational energy is generated under critical conditions; finally, based on the above boundary conditions, the system utilizes the formula... Automatically calculates the attenuation coefficient.

[0062] For example: assuming the maximum frequency of the piezoelectric ceramic vibration system is... The minimum starting frequency is Then the hardware dynamic range If the critical damage threshold of the chromogenic layer is calibrated as follows: The system will then automatically calculate the protection attenuation coefficient: .

[0063] S43. Through mathematical composition operations, find the intersection of the above positive demands and negative constraints in the multidimensional physical space. This point is the point at the current moment that enables the global objective function to be satisfied. The minimum solution. The fundamental frequency. With gain factor Braking coefficient Multiplication: , bring in and The formula yields: .

[0064] Specifically, in this embodiment 1: setting , Example scenarios are as follows: Scenario A, Low-temperature melting period: At this time, the film temperature is low. (Very thick), and at this time the microstructure is stable. Substitute into the formula, Significantly increased, system output approximately The high-frequency vibration causes the film to thin under high-frequency shearing, rapidly filling the gaps between the solar cells under relatively low static pressure and expelling air bubbles.

[0065] Scenario B, High Temperature and High Pressure Period: At this time, the film has softened, but the spectrum shows accelerated wavelength drift. Substitute into the formula, , Since this value is lower than the minimum starting frequency of the system hardware (0.5Hz), the control system triggers the lower limit amplitude protection, forcing the system to maintain operation at a safe idle speed of 0.5Hz, thereby automatically reducing the force and avoiding physical damage to the coloring layer.

[0066] In this embodiment 1, the optimal oscillation frequency is calculated by combining the shear thinning principle of rheology with the damage threshold of fracture mechanics. The optimal balance between flow efficiency and structural safety is dynamically sought. Specifically, for example, when the adhesive film is in the low-temperature melting stage with high viscosity and low risk, the gain coefficient... Braking coefficient The system automatically outputs frequencies higher than the baseline value. High-frequency shear force is used to forcibly reduce the viscosity of the adhesive film to accelerate venting; however, when the adhesive film is in a high-temperature softening stage with low viscosity and high risk, although the rheological requirements decrease, if microstructural risks are detected... Increase braking coefficient It will decay exponentially, forcibly suppressing... The gain effect makes the final output oscillation frequency Reduced to a safe range. This method overcomes the problem that traditional single oscillation modes are difficult to adapt to changes in the rheological properties of the film across the entire temperature range, and achieves intelligent control that is strong when needed (accelerating liquefaction) and weak when needed (risk avoidance and protection), ensuring the dual optimization of bubble removal rate and appearance yield in complex and variable lamination environments.

[0067] S5. Perform collaborative control and quality assessment, specifically including the following steps: S51. The actuator of the laminator superimposes a micro-oscillation pressure wave with a frequency of and an amplitude of while applying the basic static pressure . This process forms a real-time closed loop: the physical oscillation changes the rheological state of the adhesive film and the stress state of the color-forming layer, and these state changes are captured again by the sensors in step S2 to update the and at the next moment, thereby triggering the recalculation of steps S3 - S4 until the end of the lamination process cycle.

[0068] S52. After the lamination process is completed, the system retrieves the data during the process, including the real-time crosslinking degree , the real-time color difference and the control frequency , and substitutes them back into the global objective function constructed in step S1, and calculates the final cumulative cost convergence value through numerical integration: ; Further, a quality acceptance threshold is set, and the cumulative cost convergence value is compared with the preset quality acceptance threshold to determine whether the colored photovoltaic module is qualified: If , it is determined that the module is qualified and recorded in the good product library; if , it is determined that the module is unqualified, marked as unqualified, and an alarm is triggered to prompt manual re-inspection.

[0069] In this embodiment 1, the quality acceptance threshold is the statistical upper limit of the qualified product distribution, and the determination method is as follows: First, before formal mass production, a small batch of trial production is carried out (for example, pieces). These modules are subjected to strict off-line physical inspections, including measuring the color difference of the whole plate using a desktop colorimeter and measuring the crosslinking degree using the Soxhlet extraction method. Select all the golden sample modules with all physical indicators qualified; retrieve the calculated values of all the golden sample modules recorded by the system during the lamination process to form a sample set ; Then, calculate the arithmetic mean and the standard deviation of this sample set: ; ; ​Finally, using Formula for calculating objective threshold: .

[0070] For example, during the trial production phase, 50 colored photovoltaic modules that passed physical testing were selected. System calculations revealed that these 50 modules... The mean is The standard deviation is The system will then automatically set: Therefore, in subsequent production, any component's... If it exceeds This means it is judged as abnormal.

[0071] Specifically, for example, during a lamination cycle lasting 600 seconds, the system records the following data: Quality performance: Measured final degree of crosslinking (Target The process fits well, and the cumulative cost of the first item is extremely low. Optical performance: Although a slight deformation trend of the microstructure was observed during the peak melting period of the film at around 200 seconds: instantaneous value reached However, due to step S4 The mechanism intervened quickly, forcibly reducing the oscillation frequency, causing the color difference to return to normal rapidly. Within this timeframe, no exponential penalties were triggered throughout the process, and the cumulative cost of the second item was relatively small. Control item performance: The frequency adjustment process was smooth, and no abnormalities were observed. Jump to Despite the dramatic fluctuations, the third item's value remains normal.

[0072] Finally, the total cost convergence value of this component is calculated through integration. Conclusion: Due to The system automatically determines that the photovoltaic module is a qualified product. This step utilizes... By quantifying the quality of complex processes into a single digital indicator, it is possible to identify defective components that may appear fine on the outside but may have internal micro-cracks or long-term weather resistance issues due to high-risk vibrations during processing. This enables a more stringent quality acceptance process than traditional visual inspection.

[0073] In this embodiment 1, the calculated optimal oscillation frequency is used... Real-time mapping is performed as physical micro-oscillation pressure waves to execute closed-loop control, and the convergence value is calculated by accumulating costs. With quality acceptance threshold Comparison is used to achieve quality acceptance. Specifically, for example: during lamination, if external disturbances cause cumulative costs during the process... Exceeded the quality acceptance threshold This means that although the final appearance of the component may seem acceptable, it has undergone excessive high-risk vibrations or critical microstructural deformations during processing, and the system will determine it as a defective product. This method overcomes the limitations of traditional manual or visual inspection, which can only identify visible surface defects and cannot detect hidden internal damage such as microstructural fatigue or residual internal stress caused by process fluctuations. This allows for the interception of potentially risky defective components, helping to improve the reliability of outgoing products.

[0074] Example 2 As a second embodiment of the present invention, such as Figure 2 As shown in Example 1, this example also discloses a multi-parameter feedback-based intelligent control system for color photovoltaic module lamination, which specifically includes a physical execution layer, a sensing and detection layer, and a central control layer.

[0075] The physical execution layer includes a laminator body and a micro-oscillation actuator. The laminator body is equipped with a heating platform and a vacuum chamber to support the photovoltaic modules to be laminated. The micro-oscillation actuator uses piezoelectric ceramic (PZT) micro-oscillators, which are evenly distributed below the laminator heating platform. This actuator can respond to high-frequency control signals and generate micro-pressure waves with adjustable frequency range from 0.1Hz to 100Hz and precisely controllable amplitude to apply shear force to the adhesive film.

[0076] The sensing layer serves as a data acquisition unit, comprising an embedded fiber optic spectral sensor, a high-frequency pressure sensor, and a temperature sensor. The embedded fiber optic spectral sensor is pre-embedded inside the flexible silicone plate of the laminator, with the fiber grating (FBG) probe positioned directly opposite the colorimetric layer region of the component, for real-time acquisition of the colorimetric layer's reflectance spectral data. and characteristic wavelength The high-frequency pressure sensor and temperature sensor are used to monitor real-time pressure. With film temperature .

[0077] The central control layer is a PLC intelligent control terminal, which stores computer-readable instructions. When the instructions are executed by the processor, the following functional modules are run: The model building module is primarily used for system initialization configuration. It contains a pre-built raw material database storing rheological curves (such as zero-shear viscosity) of different batches of POE / EVA films. ) and the standard optical parameters of the chromogenic layer (such as the reference center wavelength) Before lamination begins, this module reads the current work order information and constructs the global objective function for model predictive control as described in Example 1. And set the target crosslinking degree (e.g., 95%) and maximum operating frequency (e.g., 50Hz).

[0078] A monitoring and evaluation module, which is communicatively connected to the aforementioned fiber optic spectral sensor, is used during the lamination process. The module reads spectral data at a specific sampling period. It is equipped with a differential operation unit for real-time calculation of the second derivative of wavelength drift. Simultaneously, by combining the static wavelength deviation, the optical damage risk factor at the current moment is calculated using a formula. .

[0079] The weighting adjustment module receives data from the monitoring and evaluation module. The signal has a built-in Sigmoid nonlinear mapping algorithm, based on optical damage risk factors. Dynamically adjust the global objective function Weight allocation in the process. When detected Below the safety threshold At that time, the module automatically adjusts the quality weight. Reduce optical weight When detected As the threshold approaches, the module rapidly increases the optical weights. This allows the system control strategy to switch from efficiency-first to protection-first.

[0080] The frequency calculation module, the core computing unit, is responsible for calculating the optimal control parameters. It simultaneously receives current film temperature data and optical damage risk data. On one hand, based on the Carreau shear-thinning principle, it calculates the flow gain factor. To determine the frequency increment required to reduce viscosity; on the other hand, based on the damage suppression mechanism, the braking coefficient is calculated. Finally, the optimal oscillation frequency for the current moment is calculated and output. .

[0081] The collaborative control and decision-making module is used to calculate... The signal is converted into an analog voltage signal to drive the piezoelectric ceramic actuator, applying micro-oscillations to the component. After lamination is complete, historical data from the entire process is retrieved to optimize the global objective function. Perform numerical integration to calculate the cumulative cost convergence value. Set a preset threshold. ,like The module sends a "qualified" signal to the production line's MES system, and the component flows to the next process; if... The component is marked as defective and an "abnormal" signal is sent to the sorting robot, which then moves the component to the re-inspection area.

[0082] This system uses a monitoring and evaluation module to perceive the microstructural deformation acceleration of the color-causing layer in real time, overcoming the lag problem of traditional thermocouple temperature feedback and achieving early warning of microstructural collapse trends. Through a weight adjustment module and a frequency calculation module, it intelligently considers whether to aggressively increase pressure or conservatively avoid risks, utilizing the shear thinning principle to achieve adaptive adjustment—strengthening when needed and weakening when necessary—ensuring sufficient film leveling and venting while effectively avoiding optical damage caused by excessive pressure. Finally, a collaborative control and judgment module executes micro-oscillation operations and conducts rigorous acceptance testing, intercepting hidden defects—those that appear acceptable but harbor internal microcracks or residual stress—based on the cumulative cost throughout the process, thus improving the long-term weather resistance and reliability of photovoltaic modules leaving the factory.

[0083] Furthermore, in order to verify the actual effect of the method described in Embodiment 1 of the present invention in resolving the contradiction between bubble expulsion and microstructure protection, a comparative experiment was constructed in this embodiment.

[0084] Experimental subjects: Colored BIPV photovoltaic module laminates produced in the same batch were selected, and a POE film system with a nano-imprinted coloring layer was used.

[0085] Experimental equipment: Intelligent laminator equipped with a micro-vibration pressurization system and a fiber optic spectral monitoring system.

[0086] Interference settings: In order to simulate extreme working conditions in production, “local thickness unevenness” (deviation +0.5mm) is artificially introduced in the component pre-lay stage. This sudden change in thickness will cause abnormal flow resistance of the adhesive film during lamination, which can easily induce the collapse of the microstructure of the coloring layer during the high-temperature softening stage.

[0087] Experimental Groups: This experiment included a control group and an experimental group, each containing 10 component samples. Among them: Control group (using existing technology): based on traditional constant frequency vibration lamination process. The oscillation frequency is fixed at [value missing]. It does not have real-time feedback and adjustment capabilities.

[0088] Experimental group (using the scheme of this invention): Intelligent control method based on multi-parameter feedback. Initial reference frequency is set. Maximum allowed frequency Safety threshold Activate damage braking mechanism and shear thinning frequency compensation.

[0089] The experiment focuses on observing the high-temperature softening stage during lamination. to The system response during this period corresponds to the point of lowest film viscosity, which is a high-risk area for microstructure collapse. For example... Figure 7 As shown, the experimental process of the control group and the experimental group was as follows: Control group: At that time, due to localized stress concentration caused by uneven thickness, the optical damage risk factor increases. It rapidly climbed and broke through 0.6 (entering the danger zone). However, due to the lack of a feedback mechanism, the system remained... The high-frequency vibrations and continuous high-frequency shear forces directly damage the microstructure of the chromogenic layer, leading to risk factors. It has remained at a high level for a long time. Continue until curing is complete.

[0090] Experimental group: In At that time, the monitoring module detected Sudden increase to (Exceed The system immediately adjusts the weights of the global objective function, optical weights. from Upgraded to Damage suppression mechanisms intervene, and the calculated braking coefficient is determined. It decays exponentially, and the optimal oscillation frequency of the command output. exist From the inside sudden drop (Idle state). Maintain approximately [speed] at low frequency. Subsequently, the microstructure deformation was detected to have stopped. Falling back to The system then increased its frequency again. The residual bubbles are then discharged by utilizing the shear-thinning effect.

[0091] After data lamination, the curing quality (degree of cross-linking) and optical appearance (color difference) of the two sets of components were evaluated. The results of the detection are shown in Table 1 below.

[0092] Table 1 Conclusion: The above experimental results show that, compared to the control group, the high stress concentration caused by local stress concentration significantly reduces the stress level of the control group. Average color difference ( The experimental group reduced the average color difference to [a certain value] by real-time monitoring of optical damage risk factors and intervention of a braking mechanism. ( This successfully controlled color difference within a range imperceptible to the human eye, effectively preventing visual patches caused by the collapse of the color-causing layer's microstructure. Furthermore, data shows that although the system forcibly reduced the oscillation frequency during the high-temperature softening danger period to protect the microstructure, the experimental group's film crosslinking degree still reached a high level thanks to frequency compensation using a shear-thinning mechanism during the safe period. And the residual bubble rate is controlled within All physical properties fully meet the shipment standards for Grade A components (crosslinking degree). And bubble rate This demonstrates that the system can dynamically find the optimal balance between curing quality and optical appearance, resolving the conflict between bubble removal and the protection of the coloring layer's microstructure in the lamination process of colored photovoltaic modules. Furthermore, under extreme operating conditions with uneven raw material thickness, the yield rate of the control group was only [percentage missing]. The yield rate of the experimental group using this invention was increased to [percentage missing]. This proves that the present invention can effectively cope with sudden process fluctuations during production, prevent hidden damage, and improve the reliability of finished products.

[0093] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0094] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for intelligent lamination control of color photovoltaic modules based on multi-parameter feedback, characterized in that, include: Read the raw material parameters of the component to be laminated, and construct a global objective function for model predictive control; During the lamination process, sensors are used to collect spectral data of the coloring layer of the component in real time to calculate the optical damage risk factor; The weight allocation in the global objective function is dynamically adjusted based on the optical damage risk factors. Based on the rheological shear thinning mechanism and damage suppression mechanism, the optimal oscillation frequency that minimizes the global objective function is calculated; The laminator actuator is controlled to apply a micro-oscillation pressure wave with the optimal oscillation frequency. After the lamination process is completed, the cumulative cost convergence value is calculated and compared with the preset quality acceptance threshold to determine the component quality.

2. The intelligent control method for laminating color photovoltaic modules based on multi-parameter feedback according to claim 1, characterized in that, The expression for the global objective function is: ； In the formula, Let be the total cost function. To predict the time domain, for The degree of crosslinking of the adhesive film at any given time. For the target degree of crosslinking, For the predicted color difference value, Here is the color difference penalty function. The frequency of the applied pressure oscillation. To control the rate of change of frequency, The maximum operating frequency allowed by the actuator; These are quality weights, optical weights, and control smoothing weights, respectively. .

3. The intelligent control method for laminating color photovoltaic modules based on multi-parameter feedback according to claim 2, characterized in that, The color difference penalty function adopts an exponential barrier function, expressed as: ； In the formula, The color difference threshold; the color difference value Based on real-time collected color coordinates Color coordinates of the standard chromogenic layer The Euclidean distance between them was calculated.

4. The intelligent control method for laminating color photovoltaic modules based on multi-parameter feedback according to claim 1, characterized in that, The formula for calculating the optical damage risk factor is as follows: ； In the formula, for Optical damage risk factors at any time The reference center wavelength of the chromogenic layer; The characteristic wavelength peak, The acceleration of the characteristic wavelength drift. Let be the system response time constant. The initial wavelength, This is the maximum permissible wavelength offset. This is the deviation weighting coefficient.

5. The intelligent control method for laminating color photovoltaic modules based on multi-parameter feedback according to claim 2, characterized in that, The relationship between the optical weight and the optical damage factor is expressed as follows: ； In the formula, for Optical weights at time, The safety threshold inflection point, The sensitivity coefficient, To control the smoothing weights.

6. The intelligent control method for laminating color photovoltaic modules based on multi-parameter feedback according to claim 1, characterized in that, The formula for calculating the optimal oscillation frequency is as follows: ； In the formula, Based on the base frequency, This is the gain factor calculated based on the shear-thinning principle. This is the braking coefficient calculated based on damage risk.

7. The intelligent control method for laminating color photovoltaic modules based on multi-parameter feedback according to claim 6, characterized in that, The gain factor The calculation formula is: ； In the formula, The zero-shear viscosity at the current temperature. For ideal viscosity, To adjust the viscosity gain, To be a function that maximizes the value, This indicates the lower limit of the safety threshold.

8. The intelligent control method for laminating color photovoltaic modules based on multi-parameter feedback according to claim 6, characterized in that, The braking coefficient The calculation formula is: ； In the formula, To protect the attenuation coefficient.

9. The intelligent control method for laminating color photovoltaic modules based on multi-parameter feedback according to claim 1, characterized in that, The quality assessment component includes: calculating the cumulative cost convergence value based on the global objective function. Set quality acceptance thresholds ;like The component is deemed qualified; if The component was deemed unqualified.

10. A multi-parameter feedback-based intelligent control system for color photovoltaic module lamination, employing the multi-parameter feedback-based intelligent control method for color photovoltaic module lamination as described in any one of claims 1 to 9, characterized in that, include: The model building module is used to read the raw material parameters of the component to be laminated and construct the global objective function for model predictive control. The monitoring and evaluation module is used to acquire spectral data in real time and calculate optical damage risk factors during the lamination process; The weighting adjustment module is used to dynamically adjust the weight allocation in the global objective function in combination with optical damage risk factors; The frequency calculation module is used to calculate the optimal oscillation frequency that minimizes the global objective function; The collaborative control and judgment module is used to control the output pressure wave of the laminator actuator and calculate the cumulative cost convergence value for quality judgment.

Images