Optimization Method for Process Curve of Resin Infiltration Thermoplastic Prepreg Forming Thermal Control

Abstract

Optimization method for the process curve of resin infiltration thermoplastic prepreg forming for thermal control. In the preheating stage, a camera or other image acquisition device is used to capture images of the preheating process in real time, and the preheating effect is evaluated through image analysis algorithms. In the heating and infiltration stage, the method uses thermocouples to control and maintain the set heating temperature. At the same time, the impregnation effect of the resin and fibers is evaluated in real time through image analysis and deep learning models. To more precisely meet specific quality standards, the objective function is also optimized so as to increase the loss when predicting low gloss but actually having high gloss. Finally, after curing and forming, the method further includes necessary post-processing and comprehensive quality inspection. Overall, the invention provides a highly integrated and adaptive solution aimed at improving the efficiency and quality of the thermoplastic prepreg forming process.

Description

Technical Field

[0001] The present invention relates to the fields of composite material processing, image analysis, and deep learning, and particularly to a method for optimizing the thermal control process curve for the formation of resin-infiltrated thermoplastic prepreg filaments. Background Art

[0002] In the field of composite material manufacturing, the thermoplastic prepreg forming technology has become an increasingly concerned method. This technology has high potential application value, such as in multiple industries including aviation, automotive, and renewable energy. However, the thermoplastic prepreg forming process involves multiple complex and precise steps, such as preheating, heating and infiltration, curing and forming, and post-treatment and quality inspection, and each step requires precise control to ensure the quality of the final product. Traditionally, these control processes mostly rely on manual operations or simple automatic control systems. For example, preheating and heating are usually carried out using thermocouples, but thermocouples can only provide limited temperature control accuracy and cannot comprehensively evaluate the temperature distribution and impregnation effect of the material. In addition, the existing quality inspection methods mainly rely on post-inspection, which is not only inefficient but also difficult to achieve real-time quality control and optimization. More importantly, traditional methods are difficult to handle the dynamic changes of material properties during the forming process. For example, physical properties such as the gloss and hue of the prepreg will change during the heating and infiltration process, but the existing technologies do not have effective methods to monitor and adjust these changes in real time. This leads to more iterations and adjustments during the forming process, thus increasing costs and time. At the same time, although some advanced monitoring systems attempt to improve these problems by using image recognition or machine learning algorithms, these methods are usually offline, lack real-time performance, or require a large amount of training data and computing resources. In summary, the current technologies have multiple limitations and deficiencies, and there is an urgent need for a thermoplastic prepreg forming method that can achieve high-precision, real-time monitoring, and adaptive control. Summary of the Invention

[0003] To solve the above problems, the present invention proposes a method for optimizing the thermal control process curve for the formation of resin-infiltrated thermoplastic prepreg filaments, which can achieve high-precision, real-time monitoring, and adaptive control of thermoplastic prepreg forming.

[0004] To achieve this purpose, the present invention provides a method for optimizing the thermal control process curve for the formation of resin-infiltrated thermoplastic prepreg filaments, including the following steps:

[0005] 1) Measurement and determination of the physical and chemical properties of fibers and resins;

[0006] The first step is material preparation and preheating. Operators select the appropriate types of fibers and resins as raw materials. These materials should meet the preset physical and chemical property requirements. After selecting the materials, necessary measurements are performed, including the melting point, viscosity, and thermal conductivity of the fibers and resins. These data will be used to determine the optimal preheating and heating parameters.

[0007] 2) Image acquisition and analysis during the warm-up phase;

[0008] Image algorithms are used to capture and analyze the preheating phase to achieve non-destructive testing of the process. First, the camera is positioned to clearly capture the preheating image and cover the entire mold area. Thermal imaging is used to determine the optimal camera configuration parameters. In addition, the image acquisition frequency must ensure sufficient density, that is, it must be adjusted according to the fiber's operating linear speed, and the preheating image sampling formula is given.

[0009] A global preheating optimization method is proposed to evaluate the preheating quality in real time. The preheated fiber is analyzed using images and given a global preheating judgment formula.

[0010] 3) Heating and penetration;

[0011] During the heating and infiltration phase, the preset heating temperature is controlled and maintained to achieve optimal resin and fiber impregnation. After preheating is completed, the thermocouple begins to monitor the temperature of the working environment in real time. This temperature data is sent to a control system, namely a PLC or microcontroller, which can automatically adjust the power and working status of the heating equipment to maintain the predetermined temperature level.

[0012] All real-time data from the heating process, including temperature, time, and possible anomalies, is recorded and analyzed to facilitate subsequent quality control and process optimization. By integrating high-precision temperature monitoring and real-time data feedback, the quality of the impregnation process is maximized, while enhancing the stability and repeatability of the entire process.

[0013] 4) Prepreg forming feedback optimization;

[0014] The entire heating and penetration process is controlled by using image analysis and the BP network algorithm. First, parameters are extracted from the image after heating and penetration are completed, including the average hue value, the variance of the hue, the average wire material glossiness, and the variance of the glossiness in the image. The difference between the maximum pixel value and the minimum pixel value of each column in the image is used as one of the input features for learning. That is, through image analysis, the average hue value, the variance of the hue, the average wire material glossiness, the variance of the glossiness, and the difference between the maximum pixel value and the minimum pixel value of each column in the image are collected and input into the BP network to predict and score the glossiness of the sample after the prepreg wire is formed. The glossiness loss function is used as the overall evaluation criterion;

[0015] 5) Integration of prepreg wire forming and quality feedback;

[0016] According to the real-time data of the above-mentioned image analysis, the heating temperature and heating time of the thermocouple are adaptively adjusted. In addition, the relevant parameters of preheating are adjusted at this stage, so as to control the entire heating and penetration process, and further enhance the internal quality and appearance characteristics of the composite material. After reaching the preset heating time and temperature, curing treatment will be carried out, then the mold will be removed, and the formed composite material will be taken out for further post-treatment. Finally, necessary post-treatment steps will be carried out, and a comprehensive quality inspection of the final product will be carried out, including the evaluation of mechanical properties and surface quality.

[0017] As a further improvement of the present invention, the preheating image sampling formula in step 2) is expressed as:

[0018] Among them, the preheating image sampling formula is expressed as:

[0019]

[0020] Among them, s represents the number of image acquisitions per second, v represents the distance that the fiber rolls forward per second, with the unit of millimeter, and α is the sampling hyperparameter;

[0021] The global preheating determination formula in step 2) is expressed as:

[0022] Among them, the global preheating determination formula is expressed as:

[0023]

[0024]

[0025] Among them, is the average grayscale value, N is the total number of pixels, L iis the pixel value, γ is the uniformity index of the pre-preheating process, ΔT is the temperature change amount that needs to be adjusted in real time, and k1 and k2 are two weighting factors used to balance the two objectives of gray mean and uniformity. is the target gray mean. is the gray mean of the current image, U tar is the target uniformity index.

[0026] As a further improvement of the present invention, the glossiness loss function in step 4) is expressed as:

[0027] Among them, the glossiness loss function is expressed as:

[0028]

[0029]

[0030] Among them, Loss is the loss function of the network, w i is the loss weight adjustment function. ω is the pre-amplification hyperparameter and the exponential amplification slot, used to control the amplification degree of the weight, N is the total number of samples, y i , are the actual glossiness and the predicted glossiness respectively, and ρ is the error determination factor.

[0031] The present invention is used for the optimization method of the thermal control process curve of resin infiltration thermoplastic prepreg forming, and the beneficial effects are as follows:

[0032] 1) A method for optimizing the thermal control process curve of resin infiltration thermoplastic prepreg forming provided by the present application adopts image analysis and deep learning technologies to realize real-time quality monitoring. It can not only detect multiple attributes such as the temperature, glossiness, and hue of the prepreg in real time, but also optimize these parameters through an adaptive algorithm, so as to achieve efficient and high-quality forming.

[0033] 2) A method for optimizing the thermal control process curve of resin infiltration thermoplastic prepreg forming provided by the present application, due to the adoption of adaptive adjustment and real-time monitoring, can greatly reduce rework caused by manual intervention and quality problems, thereby reducing production costs and improving production efficiency.

[0034] 3) A method for optimizing the thermal control process curve of resin infiltration thermoplastic prepreg forming provided by the present application introduces a set of real-time parameter adjustment algorithms, which not only consider the current image and sensor data, but also compare with the preset targets, and then automatically adjust the heating temperature and time of the thermocouple. This greatly improves the self-adaptability and accuracy of the process.

[0035] 4) A method for optimizing the thermal control process curve of resin infiltration thermoplastic prepreg filament forming provided by this application uses a weighted objective function and particularly emphasizes the gloss error. Especially when predicting a finished product with low gloss as a high-gloss one, the model becomes more accurate in predicting low gloss. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Flowchart of a method for optimizing the thermal control process curve of resin infiltration thermoplastic prepreg filament forming provided by an embodiment of this application;

[0037] Figure 2 Schematic diagram of the preheating treatment image of a method for optimizing the thermal control process curve of resin infiltration thermoplastic prepreg filament forming provided by an embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0039] As Figure 1 shown is the flowchart of a method for optimizing the thermal control process curve of resin infiltration thermoplastic prepreg filament forming provided by this application.

[0040] Step S1: Measurement and determination of the physical and chemical properties of fibers and resins.

[0041] In the initial stage of this process method, first is the material preparation and preheating process, and these two steps are integrated into a coherent operation link. Specifically, the operator selects appropriate types of fibers and resins as raw materials. Among them, these materials should meet the preset physical and chemical property requirements to ensure excellent performance in the subsequent forming and curing processes. After selecting the materials, necessary measurements are carried out, including the melting point, viscosity, and thermal conductivity of the fibers and resins, and these data will be used to determine the optimal preheating and heating parameters.

[0042] Step S2: Image acquisition and analysis in the preheating stage.

[0043] As Figure 2 shown is the schematic diagram of the preheating treatment image of a method for optimizing the thermal control process curve of resin infiltration thermoplastic prepreg filament forming provided by this application.

[0044] In this application, an image algorithm is used to collect and analyze the preheating stage to achieve non-destructive detection of this process. In this step, the camera is first positioned at a location where it can clearly capture the preheating and cover the entire die area, and the optimal configuration parameters of the camera are determined through thermal imaging. In addition, for the image acquisition frequency, this application needs to ensure sufficient density, that is, it is adjusted according to the running linear speed of the fiber and given by the preheating image sampling formula.

[0045] Among them, the preheating image sampling formula is expressed as:

[0046]

[0047] Among them, s represents the number of image acquisitions per second, v represents the distance that the fiber rolls forward per second, the unit is millimeter, and α is the sampling hyperparameter.

[0048] Through this formula, the frequency of image acquisition in the preheating stage can be completed. When the running linear speed of the fiber is faster, the sampling frequency is higher, and when the running linear speed of the fiber is slower, the sampling frequency is a lower value, thereby reasonably reducing the burden on the equipment.

[0049] This application proposes a global preheating optimization method to evaluate the quality of preheating in real time. During the preheating process, over-preheating or under-preheating often occurs. In the traditional preheating process, it is necessary to analyze the temperature of the fiber through something like a temperature sensor, etc. However, the response time of the temperature sensor is not fast enough to accurately reflect the actual temperature inside or on the surface of the material. This may lead to temperature fluctuations, thus affecting the quality of the final product, and traditional temperature sensors usually can only measure the temperature at a single position or a few positions, and it is difficult to comprehensively understand the temperature distribution of the entire object or material. Therefore, this application uses images to analyze the preheated fiber filaments and gives them according to the global preheating determination formula.

[0050] Among them, the global preheating determination formula is expressed as:

[0051]

[0052]

[0053] Among them, is the average gray value, N is the total number of pixels, L i is the pixel point value, γ is the uniformity index of the pre-preheating process, ΔT is the temperature change amount that needs to be adjusted in real time, and k1 and k2 are two weight factors used to balance the two objectives of gray mean and uniformity, is the target gray mean, is the current image gray mean, U taris the target uniformity index.

[0054] Through formula (2) and formula (3), the preheating temperature can be adaptively adjusted to approach the preset target, thereby optimizing the preheating quality and the entire process. The temperature change amount that needs to be adjusted in real time can directly output the interpolation that needs to be adjusted. This parameter is adaptively adjusted according to real-time image analysis and the average values of two target gray levels and uniformity. By balancing the importance of these two targets through weights k1 and k2, that is, through this company, the preheating temperature can be adaptively adjusted to achieve a production process with higher quality and higher efficiency, while reducing energy consumption and the need for manual adjustment.

[0055] Step S3: Heating and impregnation.

[0056] In the heating and impregnation stage, the key is to precisely control and maintain the preset heating temperature to achieve the best resin and fiber impregnation effect. After preheating is completed, in this step, the thermocouple starts to monitor the temperature of the working environment in real time. These temperature data are sent to an advanced control system, namely a PLC or a microcontroller, which can then automatically adjust the power and working state of the heating equipment to maintain the predetermined temperature level. In addition, precise temperature control can ensure that the viscosity and fluidity of the resin are in the best state, thereby improving its contact and impregnation quality with the fiber. All real-time data of the heating process, including temperature, time, and possible abnormalities, are recorded and analyzed for subsequent quality control and process optimization. Therefore, this stage realizes the maximization of the quality of the impregnation process through the integration of high-precision temperature monitoring and real-time data feedback, while enhancing the stability and repeatability of the entire process.

[0057] Step S4: Pre-impregnated filament forming feedback optimization.

[0058] In this application, image analysis and BP network algorithm are used to control the entire heating and impregnation process. First, parameters of the image after heating and impregnation are extracted, including the average value of hue in the image, the variance of hue, the average value of wire material glossiness, and the variance of glossiness. In addition, the difference between the maximum pixel value and the minimum pixel value of each column in the image in this application is used as one of the input features for learning, that is, through image analysis, the average value of hue in the image, the variance of hue, the average value of wire material glossiness, the variance of glossiness, and the difference between the maximum pixel value and the minimum pixel value of each column are collected and input into the BP network to predict and score the glossiness of the sample after pre-impregnated filament forming. In addition, the glossiness loss function is used as the overall evaluation criterion in this application.

[0059] Among them, the glossiness loss function is expressed as:

[0060]

[0061]

[0062] Among them, Loss is the loss function of the network, and w i is the loss weight adjustment function, ω is the pre-amplification hyperparameter and the exponential amplification slot, used to control the amplification degree of the weight. N is the total number of samples, and y i , are the actual glossiness and the predicted glossiness respectively, and ρ is the error determination factor.

[0063] Through this formula, the loss function of the entire network can be determined. And when the glossiness predicted by the model is higher than the actual glossiness, this error will be amplified, so as to give more attention during the optimization process, enabling the network to give timely feedback and regulation when judging that the formed glossiness of the prepreg is low.

[0064] Step S5: Integration of prepreg forming and quality feedback.

[0065] According to the real-time data of the above-mentioned image analysis, adaptively adjust the heating temperature and heating time of the thermocouple. In addition, the relevant parameters of preheating can also be adjusted in this stage, so as to more precisely control the entire heating and penetration process. Furthermore, the internal quality and appearance characteristics of the composite material are enhanced. After reaching the preset heating time and temperature, curing treatment will be carried out, then the mold will be removed, and the formed composite material will be taken out for further post-treatment. Finally, necessary post-treatment steps such as trimming and polishing are carried out, and a comprehensive quality inspection of the final product is carried out, including the evaluation of mechanical properties and surface quality.

[0066] The above is only a preferred embodiment of the present invention, and it is not a limitation to the present invention in any other form. Any modification or equivalent change made according to the technical essence of the present invention still belongs to the scope protected by the present invention.

Claims

1. Method for optimizing thermo-control process curve for forming resin infiltration type thermoplastic prepreg wire, characterized by: It includes the following steps: 1) Measurement and determination of the physical and chemical properties of fibers and resins; First is the material preparation and preheating process. The operator selects appropriate types of fibers and resins as raw materials, and these materials should meet the preset physical and chemical property requirements. After selecting the materials, measurements are carried out, including the melting point, viscosity, and thermal conductivity of the fibers and resins. These data will be used to determine the optimal preheating and heating parameters; 2) Image acquisition and analysis during the preheating stage; Image algorithms are used to acquire and analyze the preheating stage to achieve non-destructive detection of this process. First, the camera is positioned at a location where it can clearly capture the preheating and cover the entire mold area, and the best configuration parameters of the camera are determined through thermal imaging. In addition, for the image acquisition frequency, sufficient density needs to be ensured, that is, it is adjusted according to the running linear speed of the fiber and given by the preheating image sampling formula; A global preheating optimization method is proposed to evaluate the quality of preheating in real time. Images are used to analyze the preheated fiber filaments and given according to the global preheating determination formula; The preheating image sampling formula in step 2) is expressed as: （1） Among them, represents the number of image acquisitions per second, represents the distance that the fiber rolls forward per second, in millimeters, is the sampling hyperparameter; The global preheating determination formula in step 2) is expressed as: （2） （3） Among them, is the average gray value, is the total number of pixels, is the pixel value, is the uniformity index of the preheating process, is the temperature change amount that needs to be adjusted in real time, and These two are weight factors, used to balance the two objectives of gray mean and uniformity, is the target average gray value, is the current average gray value of the image, is the target uniformity index; 3) Heating and penetration; During the heating and penetration stage, by controlling and maintaining the preset heating temperature to achieve the best resin and fiber impregnation effect. After preheating, in this step, the thermocouple starts to monitor the temperature of the working environment in real time. These temperature data are sent to a control system, namely a PLC or a microcontroller, which can then automatically adjust the power and working state of the heating equipment to maintain the predetermined temperature level; All real-time data during the heating process, including temperature and time, are recorded and analyzed for subsequent quality control and process optimization. By integrating high-precision temperature monitoring and real-time data feedback, the quality of the impregnation process is maximized, and at the same time, the stability and repeatability of the entire manufacturing process are enhanced; 4) Optimization of the feedback of the pre-impregnated filament forming; Image analysis and BP network algorithm are used to control the entire heating and penetration process. First, parameters of the image after heating and penetration are extracted, including the average value of the hue in the image, the variance of the hue, the average value of the wire material glossiness, and the variance of the glossiness. The difference between the maximum pixel value and the minimum pixel value of each column in the image is used as one of the input features for learning, that is, through image analysis, the average value of the hue in the image, the variance of the hue, the average value of the wire material glossiness, the variance of the glossiness, and the difference between the maximum pixel value and the minimum pixel value of each column are collected and input into the BP network to predict and score the sample glossiness after the pre-impregnated filament is formed, and the glossiness loss function is used as the overall evaluation criterion; 5) Integration of the pre-impregnated filament forming and quality feedback; According to the real-time data of the above-mentioned image analysis, the heating temperature and heating time of the thermocouple are adaptively adjusted. In addition, the relevant parameters of preheating are adjusted at this stage, so as to control the entire heating and penetration process, thereby enhancing the internal quality and appearance characteristics of the composite material. After reaching the preset heating time and temperature, curing treatment will be carried out, then the mold will be removed, and the formed composite material will be taken out for further post-treatment. Finally, post-treatment steps will be carried out, and a comprehensive quality inspection of the final product will be carried out, including the evaluation of mechanical properties and surface quality.

2. The method for optimizing the thermal control process curve for resin infiltration type thermoplastic prepreg filament forming according to claim 1, wherein: The glossiness loss function in step 4) is expressed as: （4） （5） Among them, is the loss function of the network, is the loss weight adjustment function, is the pre-amplification hyperparameter and the exponential amplification slot, used to control the amplification degree of the weight, is the total number of samples, are the actual glossiness and the predicted glossiness respectively, is the error determination factor.

Images