A deformation prediction method and system for automobile injection molding parts
The digital twin model with real-time monitoring and machine learning addresses inefficiencies in predicting automobile part deformation, improving accuracy and efficiency in injection molding processes.
Patent Information
- Application Number
- CN202411332737.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-24
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-09-24
AI Technical Summary
Existing methods for controlling injection molding deformation in automobile parts rely on manual adjustments and lack real-time monitoring, leading to inefficiencies, high costs, and inaccurate predictions, especially for complex shapes, and traditional simulation methods lack real-time data feedback and accuracy.
A deformation prediction method and system using a digital twin model, combined with real-time monitoring and machine learning algorithms, to simulate and predict deformation by adjusting process parameters dynamically based on real-time data feedback.
Improves prediction accuracy and response speed, reduces scrap rates, and optimizes the injection molding process, enhancing product quality and production efficiency.
Smart Images

Figure CN118862371B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of automobile parts manufacturing, and in particular to a deformation prediction method and system for automobile injection molding parts. Background Art
[0002] As an important pillar of the national economy, the innovation and development of the manufacturing technology of the automobile industry has always been an important symbol of industrial progress. Driven by the lightweight and precision manufacturing of automobiles, plastic materials are widely used in the manufacturing of automobile parts due to their light weight, high cost-effectiveness and strong plasticity. Among them, injection molding technology, as one of the mainstream methods of plastic processing, has been continuously expanded in the production of automobile parts. However, improper control of factors such as temperature, pressure, and cooling speed during the injection molding process can easily lead to deformation of parts, which not only affects the appearance quality of the product, but more importantly, may reduce the mechanical properties and service life of the parts.
[0003] In the prior art, the solution to the problem of injection deformation mainly relies on experience and trial and error, and the deformation is reduced by manually adjusting the mold design and process parameters. However, this method has the defects of low efficiency, high cost and difficulty in adapting to complex-shaped parts. In addition, traditional measurement and detection methods often can only discover problems after the production process is completed, and real-time monitoring and prediction cannot be achieved. In addition, some improved methods attempt to introduce computer-aided engineering software for simulation analysis, but due to the lack of real-time data feedback and accurate model calibration, its prediction accuracy and practicality still need to be improved. Summary of the invention
[0004] The purpose of the present invention is to provide a deformation prediction method and system for automobile injection molding parts to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:
[0005] In a first aspect, the present application provides a deformation prediction method for automotive injection molded parts, comprising:
[0006] Obtain the injection molding process parameters of automotive injection molding parts, where the injection molding process parameters include production history data and material property parameters, perform data cleaning and standardization on the injection molding process parameters, and obtain the initial data set for establishing the digital twin model;
[0007] Based on the initial data set, a digital twin model is constructed, where the digital twin model is used to simulate the physical behavior and response of automotive injection molding parts during the production process. After model training and verification, the deformation rate is adjusted to obtain the optimal digital twin model that reflects the actual production process.
[0008] The injection molding process is simulated using the optimal digital twin model to obtain simulation results. Based on the simulation results and the surface state and dimensional changes of automotive injection molding parts captured in real time by high-resolution cameras during the production process, real-time key feature data of automotive injection molding parts are obtained through image processing and feature extraction algorithms. The real-time key feature data includes thickness, warpage and surface defects.
[0009] Based on real-time key feature data, the deformation trend of automotive injection molded parts is analyzed in combination with the random forest algorithm. The random forest algorithm consists of multiple decision trees, each of which is trained according to feature data and target variables to obtain the prediction results of each decision tree. After algorithm training and feature importance evaluation, the features that contribute most to deformation prediction are selected to obtain a deformation prediction model.
[0010] According to the deformation prediction model, an adaptive algorithm is used to dynamically adjust the model parameters. Based on the deviation between the real-time monitoring data and the prediction results, the final prediction results of the deformation of automotive injection molding parts are obtained through feedback loop optimization.
[0011] Preferably, the production history data is used to record the actual operation data of the injection molding machine during the production process, including injection temperature, mold temperature, injection speed, holding pressure and cooling time; material property parameters include material melting point and fluidity.
[0012] Preferably, a digital twin model is constructed based on the initial data set, wherein the digital twin model is used to simulate the physical behavior and response of automotive injection molding parts during the production process, and after model training and verification, the deformation rate is adjusted to obtain the optimal digital twin model reflecting the actual production process, which includes:
[0013] According to an initial data set including injection molding process parameters and material property parameters, the injection molding process parameters are interpolated and calculated using a data interpolation technique to obtain a first data set, the material property parameters are calculated using a principal component analysis method to obtain a second data set, and the first data set and the second data set are interpolated and dimensionally reduced to obtain an initial state of the digital twin model;
[0014] Based on the principles of physics and the initial state of the digital twin model, the finite difference method is used to discretize the physical equations. Combined with the constitutive relationship of the material, a dynamic equation describing the stress-strain relationship of automotive injection molded parts at different time steps is established. The constitutive relationship of the material includes the thermal expansion coefficient and viscoelastic properties.
[0015] The dynamic equations are input into the simulation software for numerical solution, and the model parameters are adjusted by the gradient descent method to minimize the prediction error and obtain a verified and adjusted optimal digital twin model.
[0016] Preferably, the surface state and dimensional changes of the automobile injection molded parts in the production process captured in real time by the simulation results and the high-resolution camera are processed by image processing and feature extraction algorithms to obtain real-time key feature data of the automobile injection molded parts, including:
[0017] According to the simulation results of the digital twin model, the high-risk deformation areas of the parts during the injection molding process are identified. After the semantic segmentation algorithm of deep learning, the high-risk deformation areas in the simulation results are separated from the overall view to obtain the target areas for machine vision inspection. The high-risk deformation areas include areas with uneven cooling speed or stress concentration.
[0018] Use a high-resolution camera to capture the 2D image of the target area in real time, and use 3D reconstruction technology to restore the 3D geometric information of the parts from the captured 2D image to build a geometric model of the surface state of the parts;
[0019] Machine vision algorithms are used to extract and analyze features of geometric models. After morphological operations and deep learning classifiers, real-time key feature data of automotive injection molded parts are obtained.
[0020] Preferably, the deviation between the real-time monitoring data and the prediction result is optimized through a feedback loop to obtain the final prediction result of the deformation of the automobile injection molded parts, which includes:
[0021] The key feature data of injection molding parts obtained through real-time monitoring are compared with the prediction results of the digital twin model. The deviation between the prediction results and the actual monitoring data is quantified using statistical methods to obtain the prediction deviation data.
[0022] Based on the prediction deviation data, an adaptive control algorithm is used to determine the direction and amplitude of the model parameters that need to be adjusted, and a genetic algorithm is used to iteratively update the direction and amplitude of the model parameters to minimize the prediction deviation and obtain the updated value of the model parameters;
[0023] The updated values of the model parameters are input into the optimal digital twin model, the injection molding process is re-simulated, and the deformation trend of the parts is comprehensively analyzed and predicted in combination with the machine learning model to obtain the final prediction results of the deformation of automotive injection molding parts.
[0024] In a second aspect, the present application also provides a deformation prediction system for automotive injection molded parts, comprising:
[0025] Processing module: used to obtain the injection molding process parameters of automotive injection molding parts, where the injection molding process parameters include production history data and material property parameters, and perform data cleaning and standardization on the injection molding process parameters to obtain the initial data set for establishing the digital twin model;
[0026] Model building module: used to build a digital twin model based on the initial data set. The digital twin model is used to simulate the physical behavior and response of automotive injection molding parts during the production process. After model training and verification, the deformation rate is adjusted to obtain the optimal digital twin model that reflects the actual production process.
[0027] Capture module: used to simulate the injection molding process using the optimal digital twin model to obtain simulation results; and based on the simulation results and the surface state and dimensional changes of automotive injection molding parts captured in real time by high-resolution cameras during the production process, through image processing and feature extraction algorithms, obtain real-time key feature data of automotive injection molding parts, including thickness, warpage and surface defects;
[0028] Training and evaluation module: used to analyze the deformation trend of automotive injection molded parts based on real-time key feature data combined with the random forest algorithm. The random forest algorithm consists of multiple decision trees, each of which is trained according to feature data and target variables to obtain the prediction results of each decision tree. After algorithm training and feature importance evaluation, the features that contribute most to deformation prediction are selected to obtain a deformation prediction model.
[0029] Optimization module: It is used to dynamically adjust the model parameters according to the deformation prediction model using an adaptive algorithm, and obtain the final prediction results of the deformation of automotive injection molded parts through feedback loop optimization based on the deviation between real-time monitoring data and prediction results.
[0030] In a third aspect, the present application also provides a deformation prediction device for automotive injection molded parts, comprising:
[0031] Memory for storing computer programs;
[0032] A processor is used to implement the steps of the deformation prediction method of automobile injection molded parts when executing the computer program.
[0033] In a fourth aspect, the present application further provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned deformation prediction method based on automotive injection molded parts are implemented.
[0034] The beneficial effects of the present invention are:
[0035] The present invention realizes dynamic prediction and real-time feedback of component deformation during the injection molding process by combining real-time monitoring data, advanced image processing technology and machine learning algorithms, which not only improves the prediction accuracy and response speed, but also optimizes the injection molding process through an adaptive adjustment mechanism, reduces the scrap rate in production, and improves product quality and production efficiency. In addition, the application of the present invention also provides a new technical path for the intelligent production of automotive injection molding parts, which is of great significance to promoting the innovative development of the automotive manufacturing industry.
[0036] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.
[0038] Figure 1 It is a schematic flow chart of a deformation prediction method for automobile injection molded parts according to an embodiment of the present invention;
[0039] Figure 2 Schematic diagram of the structure of the deformation prediction system for automobile injection molding parts described in an embodiment of the present invention;
[0040] Figure 3 It is a schematic diagram of the structure of the deformation prediction device for automobile injection molded parts described in an embodiment of the present invention.
[0041] In the figure: 701, processing module; 702, model building module; 703, capture module; 704, training evaluation module; 705, optimization module; 800, deformation prediction equipment for automobile injection molding parts; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component. DETAILED DESCRIPTION
[0042] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0043] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.
[0044] Embodiment 1:
[0045] This embodiment provides a deformation prediction method for automobile injection molded parts.
[0046] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300, step S400 and step S500.
[0047] S100. Obtain injection molding process parameters of automotive injection molding parts, where the injection molding process parameters include production history data and material property parameters, perform data cleaning and standardization on the injection molding process parameters, and obtain an initial data set for establishing a digital twin model.
[0048] It can be understood that in this step S100, the production history data is used to record the actual operation data of the injection molding machine during the production process, including injection temperature, mold temperature, injection speed, holding pressure and cooling time; material property parameters include material melting point and fluidity.
[0049] S200. Based on the initial data set, a digital twin model is constructed, wherein the digital twin model is used to simulate the physical behavior and response of automotive injection molding parts during the production process. After model training and verification, the deformation rate is adjusted to obtain the optimal digital twin model reflecting the actual production process. The calculation formula of the simulated injection molding process of the digital twin model is as follows:
[0050] In the formula, St+1 represents the shape state after the next time step, St Indicates time t The shape state of the injection molded parts, Pinj and Pmold Represent the injection pressure and the pressure inside the mold, α is the material deformation rate, Δ t is the time step.
[0051] In the actual research of this embodiment, the process of building a digital twin model may involve the problem of multi-physics field coupling, such as thermal-mechanical coupling, fluid-solid coupling, etc. These coupling effects are crucial to the accuracy of the model, but they also increase the complexity of model construction. To solve this problem, researchers may use multi-scale modeling technology to combine the macro injection molding process with the micro material structure characteristics to achieve a more comprehensive simulation.
[0052] In addition, high-performance computing resources, such as cloud computing platforms, may be used in the construction and optimization of digital twin models to process large amounts of data and complex computing tasks. The elastic computing capabilities of cloud computing platforms can provide necessary support for model training and simulation.
[0053] It can be understood that step S200 includes S201, S202 and S203, wherein:
[0054] S201. According to an initial data set including injection molding process parameters and material property parameters, the injection molding process parameters are interpolated and calculated using a data interpolation technique to obtain a first data set, the material property parameters are calculated using a principal component analysis method to obtain a second data set, and the first data set and the second data set are interpolated and dimensionally reduced to obtain an initial state of the digital twin model;
[0055] It should be noted that in the early stage of building a digital twin model, the initial data set needs to be processed in detail to ensure that the model can accurately reflect the actual production conditions. In this step, injection molding process parameters, such as temperature and pressure, may have missing or discontinuous data during the production process, that is, data interpolation technology, such as linear interpolation or spline interpolation, is used to calculate these parameters to fill the data gaps and ensure the integrity and continuity of the first data set. Among them, the material property parameters may contain multiple variables, and there may be correlations between these variables. PCA can be used to reduce the dimensions, extract the main influencing factors, form the second data set, reduce the complexity of the model and highlight the key variables. The interpolated first data set and the second data set processed by PCA are integrated to form the initial state of the digital twin model, ensuring that data from different sources and different dimensions can be used consistently in the model.
[0056] S202. Based on the principles of physics and the initial state of the digital twin model, the finite difference method is used to discretize the physical equations, and combined with the constitutive relationship of the material, a dynamic equation describing the stress-strain relationship of the automotive injection molded parts at different time steps is established, where the constitutive relationship of the material includes the thermal expansion coefficient and viscoelastic properties;
[0057] It should be noted that continuous physical equations, such as the heat and mass conservation equations, are converted into discrete forms through the finite difference method so that they can be numerically solved on a computer; the constitutive relationship of the material is established by combining the thermal expansion coefficient and viscoelastic properties of the material, which describes the behavior of the material under different temperatures and stress states, and based on the principles of physics and the constitutive relationship of the material, a dynamic equation describing the stress-strain relationship of automotive injection molded parts at different time steps is established.
[0058] S203. Input the dynamic equation into the simulation software for numerical solution, adjust the model parameters by gradient descent method to minimize the prediction error, and obtain the verified and adjusted optimal digital twin model.
[0059] It should be noted that after numerical solution and parameter optimization, the optimal digital twin model obtained can simulate the actual production process with high accuracy, providing a reliable tool for deformation prediction of automotive injection molding parts.
[0060] S300: Use the optimal digital twin model to simulate the injection molding process and obtain simulation results. Based on the simulation results and the surface state and dimensional changes of automotive injection molding parts captured in real time by high-resolution cameras during the production process, the real-time key feature data of automotive injection molding parts are obtained through image processing and feature extraction algorithms. The real-time key feature data includes thickness, warpage and surface defects.
[0061] It can be understood that step S300 includes S301, S302 and S303, wherein:
[0062] S301. According to the simulation results of the digital twin model, identify the high-risk deformation areas of the parts in the injection molding process, separate the high-risk deformation areas in the simulation results from the overall view through the deep learning semantic segmentation algorithm, and obtain the target area for machine vision inspection, where the high-risk deformation areas include areas with uneven cooling speed or stress concentration;
[0063] It should be noted that, first of all, the digital twin model simulates the entire injection molding process, including the filling, holding, cooling and curing stages of the plastic melt. Its simulation results provide detailed physical states of parts during the production process, including temperature distribution, pressure distribution and stress state; and uses deep learning technology, especially convolutional neural network (CNN), to perform semantic segmentation on the simulation results, where the semantic segmentation algorithm can identify and distinguish different objects or areas in the image, and separate the high-risk deformation area from the overall view, that is, it is manifested as uneven cooling rate or stress concentration, which is a potential source of deformation of injection molded parts; and the simulation results are processed by the semantic segmentation algorithm to obtain the precise position and shape of the high-risk deformation area. These areas are marked as target areas for machine vision detection, providing a basis for subsequent real-time monitoring and quality control. This step can automatically identify and accurately locate the key areas where deformation may occur during the injection molding process, providing a clear target for subsequent real-time monitoring, thereby improving the accuracy of deformation prediction and the controllability of the production process.
[0064] In actual production environments, the identified high-risk deformation areas can be combined with machine vision systems to achieve real-time monitoring and feedback. For example, a camera can be installed on an injection molding machine to capture the production status of parts in real time and perform online analysis through deep learning algorithms to achieve real-time quality control.
[0065] S302, using a high-resolution camera to capture a two-dimensional image of the target area in real time, and using a three-dimensional reconstruction technology to restore the three-dimensional geometric information of the component from the captured two-dimensional image, and construct a geometric model of the surface state of the component;
[0066] S303. Apply machine vision algorithms to extract and analyze features of geometric models, and obtain real-time key feature data of automotive injection molding parts through morphological operations and deep learning classifiers.
[0067] It should be noted that the three-dimensional geometric information of parts is restored from the captured two-dimensional images through three-dimensional reconstruction technology, such as stereo vision or time-of-flight (ToF) technology. This process involves the extraction of depth information and the application of stereo matching algorithms, so as to construct an accurate geometric model of the surface state of the parts. The constructed geometric model is subjected to feature extraction using machine vision algorithms, including but not limited to edge detection, corner detection, and texture analysis methods to identify key features of the surface of parts; the extracted features are classified and identified using deep learning classifiers, such as convolutional neural networks (CNN), including automatic identification of features such as warpage, thickness changes, and surface defects.
[0068] In this embodiment, the application of 3D reconstruction technology may be combined with multi-view imaging or structured light technology to improve reconstruction accuracy and speed, and these technologies can work stably in different production environments and provide reliable 3D data even at high speeds or under complex lighting conditions. Among them, the training of machine vision algorithms and deep learning models may require a large amount of labeled data, and in actual research, a rich training set can be constructed through data collected during simulation and actual production to train and optimize the algorithm. Therefore, in the context of intelligent manufacturing, the implementation of this step can also be combined with the Industrial Internet of Things (IIoT) platform to achieve data integration and analysis, and further improve the level of intelligence in the production process. In this way, the entire production process can be optimized, and production efficiency and product quality can be improved.
[0069] S400. Based on real-time key feature data, the deformation trend of automotive injection molded parts is analyzed in combination with the random forest algorithm, wherein the random forest algorithm is composed of multiple decision trees, each decision tree is trained according to feature data and target variables to obtain the prediction result of each decision tree; after algorithm training and feature importance evaluation, the feature that contributes most to deformation prediction is selected to obtain a deformation prediction model.
[0070] It can be understood that the calculation formula for obtaining the prediction result of each decision tree in this step is as follows:
[0071] In the formula, fi ( X ) indicates the i A decision tree for feature data X The prediction results are the prediction results of each decision tree.
[0072] It should be noted that the implementation of step S400 may involve a large amount of data processing and computing tasks. With the development of industrial big data technology, distributed computing and cloud computing platforms can be used to handle these tasks and improve computing efficiency. As a mature integrated learning method, the random forest algorithm has great potential in the deformation prediction of automotive injection molding parts. In actual research, the model performance can be optimized by adjusting the parameters of the random forest, such as the number of decision trees, the method of feature selection, etc.; feature importance evaluation can not only help understand the impact of different features on deformation prediction, but also guide the optimization of process parameters. For example, if a specific material property or process parameter is identified as a key feature, the manufacturer can adjust these parameters in a targeted manner to reduce the risk of deformation. Therefore, the deformation prediction model constructed in this step can be integrated into the automatic control system to realize real-time monitoring and intelligent adjustment of the injection molding process, improve production efficiency and product quality, and promote the digitalization and intelligent development of the automotive manufacturing industry.
[0073] S500: According to the deformation prediction model, an adaptive algorithm is used to dynamically adjust the model parameters, and according to the deviation between the real-time monitoring data and the prediction result, a final prediction result of the deformation of the automobile injection molding parts is obtained through feedback loop optimization.
[0074] It can be understood that step S500 includes S501, S502 and S503, wherein:
[0075] S501. Compare the key feature data of the injection molding parts obtained by real-time monitoring with the prediction results of the digital twin model, and use statistical methods to quantify the deviation between the prediction results and the actual monitoring data to obtain prediction deviation data;
[0076] S502, based on the prediction deviation data, using an adaptive control algorithm to determine the direction and amplitude of the model parameters that need to be adjusted, and using a genetic algorithm to iteratively update the direction and amplitude of the model parameters to minimize the prediction deviation and obtain updated values of the model parameters;
[0077] S503. Input the updated values of the model parameters into the optimal digital twin model, re-simulate the injection molding process, and combine the machine learning model to comprehensively analyze and predict the deformation trend of the parts to obtain the final prediction results of the deformation of the automotive injection molding parts.
[0078] It should be noted that statistical methods, such as mean square error (MSE) or absolute deviation, are used to compare the key feature data obtained from real-time monitoring with the prediction results of the digital twin model to quantify the deviation between the two; and based on the prediction deviation data, an adaptive control algorithm, such as a PID controller, is used to determine the direction and amplitude of the model parameter adjustment. In this step, the dynamic adjustment and optimization of the model parameters are achieved through real-time monitoring data feedback, which improves the prediction accuracy of the digital twin model, enables the model to more accurately reflect the actual production process, and provides a reliable basis for the deformation prediction of automotive injection molding parts. In practical applications, the implementation of this step may be integrated into the intelligent manufacturing system to achieve closed-loop control of the production process. The quantification of the prediction deviation and the optimization of the model parameters can be carried out in real time to adapt to real-time changes in the production process. The combined use of adaptive control algorithms and genetic algorithms can improve the efficiency and accuracy of model parameter optimization. In actual research, the parameter settings and operating mechanisms of these two algorithms can be further explored to achieve faster and more accurate optimization. With the development of machine learning technology, more complex models, such as deep learning networks, can be used to conduct comprehensive analysis and prediction of component deformation trends. The model in this step can process larger-scale data and capture more complex nonlinear relationships. In addition, the simulation and prediction results of the digital twin model can be further integrated with actual production data to achieve real-time monitoring and intelligent adjustment of the production process, which will help improve production efficiency, reduce scrap rates, and promote the manufacturing of automotive injection molding parts to a higher level of automation and intelligence.
[0079] Embodiment 2:
[0080] like Figure 2 As shown, this embodiment provides a deformation prediction system for automobile injection molding parts, see Figure 2 The system comprises:
[0081] Processing module 701: used to obtain injection molding process parameters of automobile injection molding parts, where the injection molding process parameters include production history data and material property parameters, and perform data cleaning and standardization on the injection molding process parameters to obtain an initial data set for establishing a digital twin model;
[0082] Model building module 702: used to build a digital twin model based on the initial data set, wherein the digital twin model is used to simulate the physical behavior and response of automobile injection molding parts in the production process. After model training and verification, the deformation rate is adjusted to obtain the optimal digital twin model reflecting the actual production process. The calculation formula of the simulated injection molding process of the digital twin model is as follows:
[0083] In the formula, St +1 represents the shape state after the next time step, St Indicates time t The shape state of the injection molded parts, Pinj and Pmold Represent the injection pressure and the pressure inside the mold, α is the material deformation rate, Δ t is the time step;
[0084] Capture module 703: used to simulate the injection molding process using the optimal digital twin model to obtain simulation results; and based on the simulation results and the surface state and dimensional changes of the automotive injection molding parts in the production process captured in real time by the high-resolution camera, obtain the real-time key feature data of the automotive injection molding parts through image processing and feature extraction algorithms, wherein the real-time key feature data includes thickness, warpage and surface defects;
[0085] Training and evaluation module 704: used to analyze the deformation trend of automobile injection molded parts based on real-time key feature data in combination with random forest algorithm, wherein the random forest algorithm is composed of multiple decision trees, each decision tree is trained according to feature data and target variables to obtain the prediction result of each decision tree; after algorithm training and feature importance evaluation, the feature that contributes most to deformation prediction is selected to obtain a deformation prediction model;
[0086] Optimization module 705: used to dynamically adjust model parameters using an adaptive algorithm based on the deformation prediction model, and obtain the final prediction result of the deformation of automobile injection molding parts through feedback loop optimization based on the deviation between the real-time monitoring data and the prediction result.
[0087] Specifically, the model building module 702 includes:
[0088] A calculation unit: used for obtaining an initial state of the digital twin model by performing data interpolation and dimension reduction processing on the first and second data sets based on an initial data set including injection molding process parameters and material property parameters, wherein the injection molding process parameters are interpolated and calculated using a data interpolation technique to obtain a first data set, and the material property parameters are calculated using a principal component analysis method to obtain a second data set;
[0089] Establishing unit: It is used to discretize the physical equations based on the physical principles and the initial state of the digital twin model using the finite difference method, and to establish the dynamic equations describing the stress-strain relationship of automotive injection molding parts at different time steps in combination with the constitutive relationship of the material, where the constitutive relationship of the material includes the thermal expansion coefficient and viscoelastic properties;
[0090] Solving unit: used to input the dynamic equations into the simulation software for numerical solution, adjust the model parameters through the gradient descent method to minimize the prediction error, and obtain the verified and adjusted optimal digital twin model.
[0091] Specifically, the capture module 703 includes:
[0092] Identification unit: used to identify high-risk deformation areas of parts during the injection molding process based on the simulation results of the digital twin model. Through the semantic segmentation algorithm of deep learning, the high-risk deformation areas in the simulation results are separated from the overall view to obtain the target area for machine vision inspection. The high-risk deformation areas include areas with uneven cooling speed or stress concentration.
[0093] Construction unit: used to capture the two-dimensional image of the target area in real time using a high-resolution camera, and restore the three-dimensional geometric information of the parts from the captured two-dimensional image through three-dimensional reconstruction technology to build a geometric model of the surface state of the parts;
[0094] Processing unit: It is used to apply machine vision algorithms to extract and analyze features of geometric models. Through morphological operations and deep learning classifiers, real-time key feature data of automotive injection molded parts are obtained.
[0095] Specifically, the optimization module 705 includes:
[0096] Quantification unit: used to compare the key feature data of injection molding parts obtained through real-time monitoring with the prediction results of the digital twin model, and use statistical methods to quantify the deviation between the prediction results and the actual monitoring data to obtain the prediction deviation data;
[0097] Iteration unit: used to determine the direction and amplitude of the model parameters that need to be adjusted based on the prediction deviation data by using an adaptive control algorithm, and iteratively update the direction and amplitude of the model parameters by using a genetic algorithm to minimize the prediction deviation and obtain the updated value of the model parameters;
[0098] Prediction unit: It is used to input the updated values of the model parameters into the optimal digital twin model, re-simulate the injection molding process, and combine the machine learning model to comprehensively analyze and predict the deformation trend of the parts to obtain the final prediction results of the deformation of the automotive injection molding parts.
[0099] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0100] Embodiment 3:
[0101] Corresponding to the above method embodiment, this embodiment also provides a deformation prediction device for an automobile injection molded part. The deformation prediction device for an automobile injection molded part described below and the deformation prediction method for an automobile injection molded part described above can refer to each other.
[0102] Figure 3 FIG. 8 is a block diagram of a deformation prediction device 800 for an automobile injection molded component according to an exemplary embodiment. Figure 3 As shown, the deformation prediction device 800 for automobile injection molded parts includes: a processor 801 and a memory 802. The deformation prediction device 800 for automobile injection molded parts also includes one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.
[0103] The processor 801 is used to control the overall operation of the automobile injection molded component deformation prediction device 800 to complete all or part of the steps in the above-mentioned automobile injection molded component deformation prediction method. The memory 802 is used to store various types of data to support the operation of the automobile injection molded component deformation prediction device 800, and these data may include, for example, instructions for any application or method used to operate on the automobile injection molded component deformation prediction device 800, and application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 802 or sent via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules can be a keyboard, a mouse or buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the deformation prediction device 800 of the automotive injection molding parts and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 805 can include: a Wi-Fi module, a Bluetooth module or an NFC module.
[0104] In an exemplary embodiment, the deformation prediction device 800 for automobile injection molded parts can be implemented by one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned deformation prediction method for automobile injection molded parts.
[0105] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned deformation prediction method for automobile injection molded parts are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 802 including program instructions, and the above-mentioned program instructions can be executed by the processor 801 of the deformation prediction device 800 for automobile injection molded parts to complete the above-mentioned deformation prediction method for automobile injection molded parts.
[0106] Embodiment 4:
[0107] Corresponding to the above method embodiment, a readable storage medium is also provided in this embodiment. The readable storage medium described below and the deformation prediction method of an automobile injection molded component described above can be referred to each other.
[0108] A computer program is stored on the readable storage medium, and when the computer program is executed by the processor, the steps of the method for predicting deformation of automobile injection molded parts in the above method embodiment are implemented.
[0109] The readable storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other readable storage medium that can store program codes.
[0110] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
[0111] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
Claims
1. A deformation prediction method for automotive injection molded parts. It is characterized in that include: Obtain the injection molding process parameters of automotive injection molding parts, where the injection molding process parameters include production history data and material property parameters, perform data cleaning and standardization on the injection molding process parameters, and obtain the initial data set for establishing the digital twin model; Based on the initial data set, a digital twin model is constructed, where the digital twin model is used to simulate the physical behavior and response of automotive injection molding parts during the production process. After model training and verification, the deformation rate is adjusted to obtain the optimal digital twin model that reflects the actual production process. The calculation formula for the simulated injection molding process of the digital twin model is as follows: In the formula, St +1 represents the shape state after the next time step, St Indicates time t The shape state of the injection molded parts, Pinj and Pmold Represent the injection pressure and the pressure inside the mold, α is the material deformation rate, Δ t is the time step; Use the optimal digital twin model to simulate the injection molding process and obtain simulation results; Based on the simulation results and the surface state and dimensional changes of automotive injection molded parts captured in real time by high-resolution cameras during the production process, the real-time key feature data of automotive injection molded parts are obtained through image processing and feature extraction algorithms. The real-time key feature data includes thickness, warpage and surface defects. Based on real-time key feature data, the deformation trend of automotive injection molded parts is analyzed in combination with the random forest algorithm. The random forest algorithm consists of multiple decision trees, each of which is trained according to feature data and target variables to obtain the prediction results of each decision tree. After algorithm training and feature importance evaluation, the features that contribute most to deformation prediction are selected to obtain a deformation prediction model. According to the deformation prediction model, an adaptive algorithm is used to dynamically adjust the model parameters. Based on the deviation between the real-time monitoring data and the prediction results, the final prediction results of the deformation of automotive injection molding parts are obtained through feedback loop optimization.
2. The method for predicting deformation of automotive injection molded parts according to claim 1, It is characterized in that The production history data is used to record the actual operation data of the injection molding machine during the production process, including injection temperature, mold temperature, injection speed, holding pressure and cooling time; material property parameters include material melting point and fluidity.
3. The method for predicting deformation of automotive injection molded parts according to claim 1, It is characterized in that According to the initial data set, a digital twin model is constructed, wherein the digital twin model is used to simulate the physical behavior and response of automotive injection molding parts during the production process. After model training and verification, the deformation rate is adjusted to obtain the optimal digital twin model reflecting the actual production process, which includes: According to an initial data set including injection molding process parameters and material property parameters, the injection molding process parameters are interpolated and calculated using a data interpolation technique to obtain a first data set, the material property parameters are calculated using a principal component analysis method to obtain a second data set, and the first data set and the second data set are interpolated and dimensionally reduced to obtain an initial state of the digital twin model; Based on the principles of physics and the initial state of the digital twin model, the finite difference method is used to discretize the physical equations. Combined with the constitutive relationship of the material, a dynamic equation describing the stress-strain relationship of automotive injection molded parts at different time steps is established. The constitutive relationship of the material includes the thermal expansion coefficient and viscoelastic properties. The dynamic equations are input into the simulation software for numerical solution, and the model parameters are adjusted by the gradient descent method to minimize the prediction error and obtain the verified and adjusted optimal digital twin model.
4. The method for predicting deformation of automotive injection molded parts according to claim 1, It is characterized in that The surface state and dimensional changes of the automotive injection molded parts in the production process captured in real time by the simulation results and high-resolution cameras are processed by image processing and feature extraction algorithms to obtain real-time key feature data of the automotive injection molded parts, including: According to the simulation results of the digital twin model, the high-risk deformation areas of the parts during the injection molding process are identified. After the semantic segmentation algorithm of deep learning, the high-risk deformation areas in the simulation results are separated from the overall view to obtain the target areas for machine vision inspection. The high-risk deformation areas include areas with uneven cooling speed or stress concentration. Use a high-resolution camera to capture the 2D image of the target area in real time, and use 3D reconstruction technology to restore the 3D geometric information of the parts from the captured 2D image to build a geometric model of the surface state of the parts; Machine vision algorithms are used to extract and analyze features of geometric models. After morphological operations and deep learning classifiers, real-time key feature data of automotive injection molded parts are obtained.
5. The method for predicting deformation of automotive injection molded parts according to claim 1, It is characterized in that The calculation formula for obtaining the prediction result of each decision tree is as follows: In the formula, fi ( X ) indicates the i A decision tree for feature data X The prediction results are the prediction results of each decision tree.
6. The method for predicting deformation of automotive injection molded parts according to claim 1, It is characterized in that According to the deviation between the real-time monitoring data and the prediction results, the final prediction results of the deformation of the automotive injection molding parts are obtained through feedback loop optimization, including: The key feature data of injection molding parts obtained through real-time monitoring are compared with the prediction results of the digital twin model. The deviation between the prediction results and the actual monitoring data is quantified using statistical methods to obtain the prediction deviation data. Based on the prediction deviation data, an adaptive control algorithm is used to determine the direction and amplitude of the model parameters that need to be adjusted, and a genetic algorithm is used to iteratively update the direction and amplitude of the model parameters to minimize the prediction deviation and obtain the updated value of the model parameters; The updated values of the model parameters are input into the optimal digital twin model, the injection molding process is re-simulated, and the deformation trend of the parts is comprehensively analyzed and predicted in combination with the machine learning model to obtain the final prediction results of the deformation of automotive injection molding parts.
7. A deformation prediction system for automobile injection molded parts, based on the deformation prediction method for automobile injection molded parts according to claim 1, It is characterized in that include: Processing module: used to obtain the injection molding process parameters of automotive injection molding parts, where the injection molding process parameters include production history data and material property parameters, and perform data cleaning and standardization on the injection molding process parameters to obtain the initial data set for establishing the digital twin model; Model building module: used to build a digital twin model based on the initial data set. The digital twin model is used to simulate the physical behavior and response of automotive injection molding parts during the production process. After model training and verification, the deformation rate is adjusted to obtain the optimal digital twin model that reflects the actual production process. The calculation formula of the digital twin model's simulated injection molding process is as follows: In the formula, St +1 represents the shape state after the next time step, St Indicates time t The shape state of the injection molded parts, Pinj and Pmold Represent the injection pressure and the pressure inside the mold, α is the material deformation rate, Δ t is the time step; Capture module: used to simulate the injection molding process using the optimal digital twin model and obtain simulation results; Based on the simulation results and the surface state and dimensional changes of automotive injection molded parts captured in real time by high-resolution cameras during the production process, the real-time key feature data of automotive injection molded parts are obtained through image processing and feature extraction algorithms. The real-time key feature data includes thickness, warpage and surface defects. Training and evaluation module: used to analyze the deformation trend of automotive injection molded parts based on real-time key feature data combined with the random forest algorithm. The random forest algorithm consists of multiple decision trees, each of which is trained according to feature data and target variables to obtain the prediction results of each decision tree. After algorithm training and feature importance evaluation, the features that contribute most to deformation prediction are selected to obtain a deformation prediction model. Optimization module: It is used to dynamically adjust the model parameters according to the deformation prediction model using an adaptive algorithm, and obtain the final prediction results of the deformation of automotive injection molded parts through feedback loop optimization based on the deviation between real-time monitoring data and prediction results.
8. The deformation prediction system for automobile injection molded parts according to claim 7, It is characterized in that The model building module includes: A calculation unit: used for obtaining an initial state of the digital twin model by performing data interpolation and dimension reduction processing on the first and second data sets based on an initial data set including injection molding process parameters and material property parameters, wherein the injection molding process parameters are interpolated and calculated using a data interpolation technique to obtain a first data set, and the material property parameters are calculated using a principal component analysis method to obtain a second data set; Establishing unit: It is used to discretize the physical equations based on the physical principles and the initial state of the digital twin model using the finite difference method, and to establish the dynamic equations describing the stress-strain relationship of automotive injection molding parts at different time steps in combination with the constitutive relationship of the material, where the constitutive relationship of the material includes the thermal expansion coefficient and viscoelastic properties; Solving unit: used to input the dynamic equations into the simulation software for numerical solution, adjust the model parameters through the gradient descent method to minimize the prediction error, and obtain the verified and adjusted optimal digital twin model.
9. The deformation prediction system for automobile injection molded parts according to claim 7, It is characterized in that The capture module includes: Identification unit: used to identify high-risk deformation areas of parts during the injection molding process based on the simulation results of the digital twin model. Through the semantic segmentation algorithm of deep learning, the high-risk deformation areas in the simulation results are separated from the overall view to obtain the target area for machine vision inspection. The high-risk deformation areas include areas with uneven cooling speed or stress concentration. Construction unit: used to capture the two-dimensional image of the target area in real time using a high-resolution camera, and restore the three-dimensional geometric information of the parts from the captured two-dimensional image through three-dimensional reconstruction technology to build a geometric model of the surface state of the parts; Processing unit: It is used to apply machine vision algorithms to extract and analyze features of geometric models. Through morphological operations and deep learning classifiers, real-time key feature data of automotive injection molded parts are obtained.
10. The deformation prediction system for automobile injection molded parts according to claim 7, It is characterized in that The optimization module includes: Quantification unit: used to compare the key feature data of injection molding parts obtained through real-time monitoring with the prediction results of the digital twin model, and use statistical methods to quantify the deviation between the prediction results and the actual monitoring data to obtain the prediction deviation data; Iteration unit: used to determine the direction and amplitude of the model parameters that need to be adjusted based on the prediction deviation data by using an adaptive control algorithm, and iteratively update the direction and amplitude of the model parameters by using a genetic algorithm to minimize the prediction deviation and obtain the updated value of the model parameters; Prediction unit: It is used to input the updated values of the model parameters into the optimal digital twin model, re-simulate the injection molding process, and combine the machine learning model to comprehensively analyze and predict the deformation trend of the parts to obtain the final prediction results of the deformation of the automotive injection molding parts.
