Lamination process parameter self-adaptive regulation and control system and method

Abstract

The invention relates to the technical field of photovoltaic module manufacturing, in particular to a lamination process parameter self-adaptive regulation and control system and method. Comprising a multi-source sensing module, an edge calculation module and a dynamic execution module, and temperature, pressure and thickness data of a lamination process are collected in real time through an infrared thermal imaging unit, a distributed pressure sensing unit and an ultrasonic thickness monitoring unit; the edge calculation module adopts a CNN-LSTM hybrid model to perform spatial-temporal feature fusion analysis on the multi-source data, predicts a defect risk level and generates a regulation and control instruction; the dynamic execution module achieves dynamic parameter adjustment through the partition temperature control unit, the hydraulic servo unit and the emergency braking unit, the problems of parameter stiffness, defect detection lag and the like of a traditional lamination process are solved, the AOI spot check defect rate is reduced to 0.6% from 1.3%, the lamination period is shortened by 3 minutes, and the production efficiency and the product quality are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic module manufacturing, and in particular to a lamination process parameter adaptive control system and method. Background Art

[0002] In the photovoltaic module lamination process, traditional techniques rely on preset fixed parameters or manual experience to adjust the lamination temperature, pressure, and time to achieve EVA (ethylene-vinyl acetate copolymer) film curing and cell encapsulation. The benefits of existing technologies are as follows: First, standardized process parameters ensure basic module encapsulation quality, reducing consistency risks in mass production; second, combined with AOI (automated optical inspection) spot checks, it can screen out some appearance defects such as bubbles and offset, keeping the defect rate below 1.3%; third, the coordinated temperature and pressure regulation based on a PID control algorithm can shorten the lamination cycle by approximately 8%-12%.

[0003] However, in existing technologies, traditional laminators rely on a fixed temperature-pressure curve and are unable to sense changes in the fluidity of the EVA film in real time and dynamically adjust parameters, resulting in a bubble defect rate as high as 3% to 5%. Secondly, existing AOI equipment only performs inspections after lamination is completed. By the time defects are discovered, the components have already been degraded or even scrapped, and there is a lack of a real-time intervention mechanism. In addition, closed control protocols are used in laminators of different brands, making it difficult to achieve universal deployment of existing optimization systems, which restricts process upgrades. Summary of the Invention

[0004] The purpose of the present invention is to provide a system and method for adaptively controlling lamination process parameters, aiming to solve the technical problems in the prior art where traditional laminators rely on a fixed temperature-pressure curve and are unable to sense changes in the fluidity of the EVA film in real time and dynamically adjust parameters, resulting in a bubble defect rate as high as 3% to 5%. Secondly, existing AOI equipment only performs inspections after lamination is completed. When defects are discovered, the components have already been degraded or even scrapped, and there is a lack of a real-time intervention mechanism. In addition, closed control protocols are used in laminators of different brands, making it difficult for existing optimization systems to achieve universal deployment, thus restricting process upgrades.

[0005] To achieve the above objectives, the present invention adopts an adaptive control system for lamination process parameters, including a multi-source perception module, an edge computing module and a dynamic execution module. The multi-source perception module includes an infrared thermal imaging unit, a distributed pressure sensing unit and an ultrasonic thickness monitoring unit. The infrared thermal imaging unit is installed on the top of the lamination cavity, the distributed pressure sensing unit is embedded in the laminator transmission belt, and the ultrasonic thickness monitoring unit is set at the end of the transmission belt. The dynamic execution module includes a partitioned temperature control unit, a hydraulic servo unit and an emergency brake unit. The multi-source perception module is connected to the edge computing module, and the edge computing module is connected to the dynamic execution module.

[0006] The infrared thermal imaging unit is used to collect the surface temperature distribution of the laminate in real time, the distributed pressure sensing unit is used to monitor the pressure changes in the lamination area, and the ultrasonic thickness monitoring unit is used to dynamically detect the thickness data of the component;

[0007] The edge computing module is used to receive multi-source perception data, perform spatiotemporal feature fusion analysis on the data, predict the defect risk level and generate control instructions;

[0008] The zone temperature control unit is used to adjust the temperature of the heating zone, the hydraulic servo unit is used to adjust the lamination pressure, and the emergency brake unit is used to immediately stop the machine and issue an alarm when a level III risk is detected.

[0009] Among them, the resolution of the infrared thermal imaging unit is ≥640×480, and the frame rate is ≥30Hz; the distributed pressure sensing unit includes a 16×16 flexible film array, with a measuring range of 0-200kPa and a sampling rate of ≥100Hz; the accuracy of the ultrasonic thickness monitoring unit is ±1.5μm, and the measurement frequency is 5MHz.

[0010] Among them, the zoned temperature control unit includes multiple heating zones, and each heating zone is equipped with a bidirectional thyristor power regulator with an adjustment accuracy of ±0.5%; the hydraulic servo unit is a high-precision proportional valve with a response time of ≤20ms.

[0011] Among them, the edge computing module is NVIDIA Jetson AGX Xavier; it has a built-in spatiotemporal feature fusion model unit, the spatiotemporal feature fusion model unit adopts a CNN-LSTM hybrid model, and the timestamp error of the spatiotemporal feature fusion model unit is ≤1ms; the spatiotemporal feature fusion model unit includes a convolutional neural network branch unit and a long short-term memory network branch unit, the input data size of the spatiotemporal feature fusion model unit is 224×224×3 pixels, the sliding window length is 60s, and the output risk level is level I-III.

[0012] The multi-source sensing module further includes an environment compensation unit and a pressure sensing array unit, and both the environment compensation unit and the pressure sensing array unit are connected to the edge computing module;

[0013] The environmental compensation unit is used to monitor the lamination cavity environment in real time through temperature and humidity sensors, and dynamically correct the temperature measurement value in combination with thermal imaging data to eliminate environmental interference;

[0014] The pressure sensing array unit adopts an electromagnetic shielding coating to reduce the influence of the high-frequency driving motor on signal acquisition.

[0015] The dynamic execution module further includes a fault-tolerant mechanism unit and a pressure closed-loop control unit, and the edge computing module is further connected to the fault-tolerant mechanism unit and the pressure closed-loop control unit;

[0016] The fault-tolerant mechanism unit is used to automatically enable redundant heating in adjacent areas when the temperature control of a certain heating area fails;

[0017] The pressure closed-loop control unit adjusts the proportional valve opening in real time in combination with the PID algorithm.

[0018] The present invention also provides a method for adaptively controlling laminating process parameters, which is applied to the above-mentioned adaptive control system for laminating process parameters and comprises the following steps:

[0019] First, the infrared thermal imaging unit collects real-time temperature distribution data on the laminate surface. The distributed pressure sensing unit continuously monitors pressure changes in the laminate area. The ultrasonic thickness monitoring unit dynamically detects changes in component thickness. The environmental compensation unit monitors cavity environmental parameters in real time through temperature and humidity sensors and fuses them with thermal imaging data to eliminate interference from environmental factors on temperature measurement.

[0020] The edge computing module receives multi-source sensor data from various sensors and achieves time synchronization of the data through a built-in multi-modal data synchronization and alignment algorithm unit. The system segments the thermal imaging data into regions of interest, focusing on extracting the temperature gradient characteristics of the laminate edge. At the same time, the pressure sensor data undergoes electromagnetic shielding processing.

[0021] The CNN branch then processes the 224×224×3 pixel thermal imaging data from the multi-source data to extract features of the bubble formation area. The LSTM branch analyzes the pressure and thickness time series data within a 60-second sliding window to predict the probability of defects within the next five minutes. The model integrates temporal and spatial features to output risk warnings at levels I, II, III, and III.

[0022] Then, based on different risk levels, the system implements corresponding control strategies. The zoned temperature control unit precisely adjusts the temperature of each heating zone through a bidirectional thyristor power regulator: for level I risk, temperature difference fine-tuning is performed, and for level II risk, local temperature rise compensation is implemented. The hydraulic servo unit uses a high-precision proportional valve combined with a PID algorithm to achieve closed-loop pressure control. At level II risk, a 3.5kPa boost is implemented and maintained for 30 seconds. When temperature control fails, the fault-tolerant mechanism automatically activates redundant heating in adjacent areas. When a level III risk is detected, the emergency brake unit immediately shuts down and triggers an audible and visual alarm.

[0023] Finally, the system continuously monitors the adjusted process parameters to ensure temperature uniformity and pressure stability. At the same time, it conducts random inspections on finished products through automatic optical inspection equipment and feeds data such as the actual defect rate into the CNN-LSTM model, achieving continuous optimization of the model and continuous improvement of system performance.

[0024] The present invention provides a system and method for adaptively controlling lamination process parameters, including a multi-source perception module, an edge computing module and a dynamic execution module. The temperature, pressure and thickness data of the lamination process are collected in real time through an infrared thermal imaging unit, a distributed pressure sensing unit and an ultrasonic thickness monitoring unit; the edge computing module uses a CNN-LSTM hybrid model to perform spatiotemporal feature fusion analysis on multi-source data, predict the defect risk level and generate control instructions; the dynamic execution module realizes dynamic parameter adjustment through a partitioned temperature control unit, a hydraulic servo unit and an emergency braking unit. The present invention solves the problems of rigid parameters and delayed defect detection in traditional lamination processes, reduces the AOI spot check defect rate from 1.3% to 0.6%, shortens the lamination cycle by 3 minutes, and significantly improves production efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0026] Figure 1 It is a schematic diagram of the principle of the lamination process parameter adaptive control system of the present invention.

[0027] Figure 2 It is a flow chart of the steps of a method for adaptively controlling lamination process parameters of the present invention.

[0028] 1-Multi-source perception module, 2-Edge computing module, 3-Dynamic execution module, 4-Infrared thermal imaging unit, 5-Distributed pressure sensing unit, 6-Ultrasonic thickness monitoring unit, 7-Partitioned temperature control unit, 8-Hydraulic servo unit, 9-Emergency braking unit, 10-Heating zone, 11-Bidirectional thyristor power regulator, 12-Spatiotemporal feature fusion model unit, 13-Convolutional neural network branch unit, 14-Long short-term memory network branch unit, 15-Environmental compensation unit, 16-Pressure sensing array unit, 17-Fault-tolerant mechanism unit, 18-Pressure closed-loop control unit. DETAILED DESCRIPTION

[0029] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0030] See also Figure 1 The present invention provides a lamination process parameter adaptive control system, including a multi-source perception module 1, an edge computing module 2 and a dynamic execution module 3. The multi-source perception module 1 includes an infrared thermal imaging unit 4, a distributed pressure sensing unit 5 and an ultrasonic thickness monitoring unit 6. The infrared thermal imaging unit 4 is installed on the top of the lamination cavity, the distributed pressure sensing unit 5 is embedded in the lamination machine transmission belt, and the ultrasonic thickness monitoring unit 6 is arranged at the end of the transmission belt. The dynamic execution module 3 includes a partitioned temperature control unit 7, a hydraulic servo unit 8 and an emergency brake unit 9. The multi-source perception module 1 is connected to the edge computing module 2, and the edge computing module 2 is connected to the dynamic execution module 3.

[0031] The infrared thermal imaging unit 4 is used to collect the surface temperature distribution of the laminate in real time, the distributed pressure sensing unit 5 is used to monitor the pressure change in the lamination area, and the ultrasonic thickness monitoring unit 6 is used to dynamically detect the thickness data of the component;

[0032] The edge computing module 2 is used to receive multi-source perception data, perform spatiotemporal feature fusion analysis on the data, predict the defect risk level and generate control instructions;

[0033] The zone temperature control unit 7 is used to adjust the temperature of the heating zone 10, the hydraulic servo unit 8 is used to adjust the lamination pressure, and the emergency brake unit 9 is used to immediately stop the machine and issue an alarm when a level III risk is detected.

[0034] In this embodiment, the infrared thermal imaging unit 4 is used to collect the surface temperature distribution of the laminate in real time, the distributed pressure sensing unit 5 is used to monitor the pressure changes in the lamination area, and the ultrasonic thickness monitoring unit 6 is used to dynamically detect the component thickness data; the edge computing module 2 is used to receive multi-source perception data, perform spatiotemporal feature fusion analysis on the data, predict the defect risk level and generate control instructions; the partitioned temperature control unit 7 is used to adjust the temperature of the heating zone 10, the hydraulic servo unit 8 is used to adjust the lamination pressure, and the emergency braking unit 9 is used to immediately shut down and alarm when a level III risk is detected.

[0035] Furthermore, the resolution of the infrared thermal imaging unit 4 is ≥640×480, and the frame rate is ≥30Hz; the distributed pressure sensing unit 5 includes a 16×16 flexible film array, with a measuring range of 0-200kPa and a sampling rate of ≥100Hz; the accuracy of the ultrasonic thickness monitoring unit 6 is ±1.5μm, and the measurement frequency is 5MHz.

[0036] Furthermore, the zoned temperature control unit 7 includes multiple heating zones 10, and each heating zone 10 is equipped with a bidirectional thyristor power regulator 11 with an adjustment accuracy of ±0.5%; the hydraulic servo unit 8 is a high-precision proportional valve with a response time of ≤20ms.

[0037] Furthermore, the edge computing module 2 is NVIDIA Jetson AGX Xavier; it has a built-in spatiotemporal feature fusion model unit 12, the spatiotemporal feature fusion model unit 12 adopts a CNN-LSTM hybrid model, and the timestamp error of the spatiotemporal feature fusion model unit 12 is ≤1ms; the spatiotemporal feature fusion model unit 12 includes a convolutional neural network branch unit 13 and a long short-term memory network branch unit 14, the input data size of the spatiotemporal feature fusion model unit 12 is 224×224×3 pixels, the sliding window length is 60s, and the output risk level is level I-III.

[0038] Furthermore, the multi-source sensing module 1 further includes an environment compensation unit 15 and a pressure sensing array unit 16, and both the environment compensation unit 15 and the pressure sensing array unit 16 are connected to the edge computing module 2;

[0039] The environmental compensation unit 15 is used to monitor the lamination cavity environment in real time through temperature and humidity sensors, and dynamically correct the temperature measurement value in combination with thermal imaging data to eliminate environmental interference;

[0040] The pressure sensing array unit 16 uses an electromagnetic shielding coating to reduce the impact of the high-frequency drive motor on signal acquisition.

[0041] In this embodiment, the environmental compensation unit 15 is used to monitor the laminate cavity environment in real time through temperature and humidity sensors, and dynamically correct the temperature measurement value in combination with thermal imaging data to eliminate environmental interference; the pressure sensing array unit 16 uses an electromagnetic shielding coating to reduce the impact of the high-frequency drive motor on signal acquisition.

[0042] Furthermore, the dynamic execution module 3 further includes a fault tolerance mechanism unit 17 and a pressure closed-loop control unit 18, and the edge computing module 2 is also connected to the fault tolerance mechanism unit 17 and the pressure closed-loop control unit 18;

[0043] The fault-tolerant mechanism unit 17 is used to automatically enable redundant heating in adjacent areas when the temperature control of a certain heating area 10 fails;

[0044] The pressure closed-loop control unit 18 adjusts the proportional valve opening in real time in combination with the PID algorithm.

[0045] In this embodiment, the fault-tolerant mechanism unit 17 is used to automatically enable redundant heating of adjacent areas when the temperature control of a certain heating area 10 fails; the pressure closed-loop control unit 18 adjusts the proportional valve opening in real time in combination with the PID algorithm.

[0046] Based on the present invention, please refer to Figure 2 The present invention also provides a method for adaptively controlling laminating process parameters, which is applied to the above-mentioned adaptively controlling laminating process parameters system, and comprises the following steps:

[0047] S1: First, the infrared thermal imaging unit 4 collects the surface temperature distribution data of the laminate in real time, the distributed pressure sensing unit 5 continuously monitors the pressure changes in the lamination area, and the ultrasonic thickness monitoring unit 6 dynamically detects the thickness changes of the component. The environmental compensation unit 15 monitors the cavity environmental parameters in real time through the temperature and humidity sensor and fuses them with the thermal imaging data to eliminate the interference of environmental factors on the temperature measurement;

[0048] S2: Based on the edge computing module 2 receiving multi-source sensing data from various sensors, the system achieves time synchronization of the data through the built-in multi-modal data synchronization and alignment algorithm unit. The system segments the thermal imaging data into regions of interest, focusing on extracting the temperature gradient characteristics of the laminate edge. At the same time, the pressure sensing data is processed through electromagnetic shielding;

[0049] S3: The CNN branch then processes the 224×224×3 pixel thermal imaging data from the multi-source data to extract features of the bubble formation area. The LSTM branch analyzes the pressure and thickness time series data within a 60-second sliding window to predict the probability of defect occurrence within the next 5 minutes. The model integrates the temporal and spatial features and outputs three levels of risk warnings: I, II, III, and III.

[0050] S4: Based on different risk levels, the system executes corresponding control strategies. The zoned temperature control unit 7 accurately adjusts the temperature of each heating zone 10 through the bidirectional thyristor power regulator 11: for level I risk, temperature difference fine-tuning is performed, and for level II risk, local temperature rise compensation is implemented. The hydraulic servo unit 8 uses a high-precision proportional valve combined with a PID algorithm to achieve closed-loop control of pressure. At level II risk, a 3.5 kPa boost is executed and maintained for 30 seconds. When temperature control fails, the fault-tolerant mechanism automatically activates redundant heating in adjacent areas. When a level III risk is detected, the emergency brake unit 9 immediately shuts down and triggers an audible and visual alarm.

[0051] S5: Finally, the system continuously monitors the adjusted process parameters to ensure temperature uniformity and pressure stability. At the same time, it conducts random inspections on finished products through automatic optical inspection equipment and feeds data such as the actual defect rate into the CNN-LSTM model to achieve continuous optimization of the model and continuous improvement of system performance.

[0052] In the present invention, three embodiments are also provided, as follows:

[0053] Example 2:

[0054] Taking the lamination of 166mm polysilicon modules as an example: the initial temperature is set to 140°C, the pressure is 80kPa, and the vacuum is maintained for 280s. The defect prediction process (time series example) is as follows: T+100s: The thermal imaging detects an abnormal low temperature of 2.5°C in the lower right corner (normal fluctuation should be ≤1.5°C); T+160s: The pressure sensor detects a pressure drop of 1.8kPa in the corresponding area; the model predicts that the area will have a temperature of ≥1.5mm at T+280s. 2 Bubble generation (Level I risk); Dynamic adjustment: Command issued: Increase power in the lower right heating zone 10 by 8% (140→151.2°C), and fine-tune the pressure in this area by 0.5 kPa; Verification Results: At T+280s, thermal imaging showed good temperature uniformity (ΔT≤0.6°C), and the final component was found to be defect-free after AOI inspection.

[0055] Example 2:

[0056] Take the lamination of 182mm polysilicon modules as an example: the initial temperature is 145°C, the pressure is 85kPa, and the vacuum holding time is 300s. During the lamination process, the thermal imaging detected an abnormal low temperature of 3°C in the upper left corner of the module at T+120s (normal fluctuation should be ≤1.5°C). At T+180s, the pressure sensor detected a pressure drop of 2.3kPa in the corresponding area. Based on this, the model predicts that the area will have a temperature of ≥2mm at T+300s. 2 Bubble generation (Level II risk); the system then issued a dynamic adjustment command, increasing the power of the upper left heating zone 10 by 12% (i.e., the temperature was adjusted from 145°C to 162.4°C), and at the same time increasing the pressure in this area by 3.5kPa; finally, at T+300s, thermal imaging showed that temperature uniformity had been restored (ΔT≤0.8°C), and after inspection by automatic optical inspection equipment (AOI), the component was found to be defective.

[0057] Example 3:

[0058] Taking the lamination of 210mm monocrystalline silicon modules as an example: the initial temperature is set at 150°C, the pressure is 90kPa, and the vacuum is maintained for 320s. The defect prediction process is as follows: T+150s: thermal imaging detects a 4°C low temperature abnormality in the central area (normal fluctuation should be ≤1.5°C); T+200s: the pressure sensor detects a 2.8kPa pressure drop in the corresponding area; the model predicts that the area will have a ≥3mm high temperature at T+350s. 2 Bubble generation (Level II risk); Dynamic adjustment: Command issued: Increase power in central heating zone 10 by 15% (150→172.5°C), and simultaneously increase pressure in this area by 3.5 kPa; Verification results: At T+350s, thermal imaging showed that temperature uniformity had recovered (ΔT≤0.7°C), and the final component was found to be defect-free after AOI inspection.

[0059] The above disclosure is only a preferred embodiment of the present invention, and certainly cannot be used to limit the scope of the rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiment and equivalent changes made in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A lamination process parameter adaptive control system, characterized in that: It includes a multi-source perception module, an edge computing module and a dynamic execution module. The multi-source perception module includes an infrared thermal imaging unit, a distributed pressure sensing unit and an ultrasonic thickness monitoring unit. The infrared thermal imaging unit is installed on the top of the lamination cavity, the distributed pressure sensing unit is embedded in the lamination machine transmission belt, and the ultrasonic thickness monitoring unit is set at the end of the transmission belt. The dynamic execution module includes a partitioned temperature control unit, a hydraulic servo unit and an emergency brake unit. The multi-source perception module is connected to the edge computing module, and the edge computing module is connected to the dynamic execution module; The infrared thermal imaging unit is used to collect the surface temperature distribution of the laminate in real time, the distributed pressure sensing unit is used to monitor the pressure changes in the lamination area, and the ultrasonic thickness monitoring unit is used to dynamically detect the thickness data of the component; The edge computing module is used to receive multi-source perception data, perform spatiotemporal feature fusion analysis on the data, predict the defect risk level and generate control instructions; The zone temperature control unit is used to adjust the temperature of the heating zone, the hydraulic servo unit is used to adjust the lamination pressure, and the emergency brake unit is used to immediately stop the machine and issue an alarm when a level III risk is detected.

2. The lamination process parameter adaptive control system according to claim 1, characterized in that: The resolution of the infrared thermal imaging unit is ≥640×480, and the frame rate is ≥30Hz; the distributed pressure sensing unit includes a 16×16 flexible film array, with a measuring range of 0-200kPa and a sampling rate of ≥100Hz; the accuracy of the ultrasonic thickness monitoring unit is ±1.5μm, and the measurement frequency is 5MHz.

3. The lamination process parameter adaptive control system according to claim 2, characterized in that: The zoned temperature control unit includes multiple heating zones, and each heating zone is equipped with a bidirectional thyristor power regulator with an adjustment accuracy of ±0.5%; the hydraulic servo unit is a high-precision proportional valve with a response time of ≤20ms.

4. The lamination process parameter adaptive control system according to claim 3, characterized in that: The edge computing module is NVIDIA Jetson AGX Xavier; It has a built-in spatiotemporal feature fusion model unit, which adopts a CNN-LSTM hybrid model. The timestamp error of the spatiotemporal feature fusion model unit is ≤1ms. The spatiotemporal feature fusion model unit includes a convolutional neural network branch unit and a long short-term memory network branch unit. The input data size of the spatiotemporal feature fusion model unit is 224×224×3 pixels, the sliding window length is 60s, and the output risk level is level I-III.

5. The lamination process parameter adaptive control system according to claim 4, characterized in that: The multi-source sensing module further includes an environment compensation unit and a pressure sensing array unit, both of which are connected to the edge computing module; The environmental compensation unit is used to monitor the lamination cavity environment in real time through temperature and humidity sensors, and dynamically correct the temperature measurement value in combination with thermal imaging data to eliminate environmental interference; The pressure sensing array unit adopts an electromagnetic shielding coating to reduce the influence of the high-frequency driving motor on signal acquisition.

6. The lamination process parameter adaptive control system according to claim 5, characterized in that: The dynamic execution module further includes a fault tolerance mechanism unit and a pressure closed loop control unit, and the edge computing module is further connected to the fault tolerance mechanism unit and the pressure closed loop control unit; The fault-tolerant mechanism unit is used to automatically enable redundant heating in adjacent areas when the temperature control of a certain heating area fails; The pressure closed-loop control unit adjusts the proportional valve opening in real time in combination with the PID algorithm.

7. A method for adaptively controlling laminating process parameters, applied to the adaptively controlling laminating process parameters system according to claim 6, characterized in that: The steps include: First, the infrared thermal imaging unit collects real-time temperature distribution data on the laminate surface. The distributed pressure sensing unit continuously monitors pressure changes in the laminate area. The ultrasonic thickness monitoring unit dynamically detects changes in component thickness. The environmental compensation unit monitors cavity environmental parameters in real time through temperature and humidity sensors and fuses them with thermal imaging data to eliminate interference from environmental factors on temperature measurement. The edge computing module receives multi-source sensor data from various sensors and achieves time synchronization of the data through a built-in multi-modal data synchronization and alignment algorithm unit. The system segments the thermal imaging data into regions of interest, focusing on extracting the temperature gradient characteristics of the laminate edge. At the same time, the pressure sensor data undergoes electromagnetic shielding processing. The CNN branch then processes the 224×224×3 pixel thermal imaging data from the multi-source data to extract features of the bubble formation area. The LSTM branch analyzes the pressure and thickness time series data within a 60-second sliding window to predict the probability of defects within the next five minutes. The model integrates temporal and spatial features to output risk warnings at levels I, II, III, and III. Then, based on different risk levels, the system implements corresponding control strategies. The zoned temperature control unit precisely adjusts the temperature of each heating zone through a bidirectional thyristor power regulator: for level I risk, temperature difference fine-tuning is performed, and for level II risk, local temperature rise compensation is implemented. The hydraulic servo unit uses a high-precision proportional valve combined with a PID algorithm to achieve closed-loop pressure control. At level II risk, a 3.5kPa boost is implemented and maintained for 30 seconds. When temperature control fails, the fault-tolerant mechanism automatically activates redundant heating in adjacent areas. When a level III risk is detected, the emergency brake unit immediately shuts down and triggers an audible and visual alarm. Finally, the system continuously monitors the adjusted process parameters to ensure temperature uniformity and pressure stability. At the same time, it conducts random inspections on finished products through automatic optical inspection equipment and feeds data such as the actual defect rate into the CNN-LSTM model, achieving continuous optimization of the model and continuous improvement of system performance.

Images