A plastic part strength detection method and system

Abstract

The present application relates to a plastic parts strength detection method and system, which includes obtaining multi-dimensional sensor data of plastic parts; pre-processing the multi-dimensional sensor data to obtain key feature data of plastic parts; using a pre-trained strength model to analyze and evaluate the key feature data to obtain a strength prediction result of the plastic parts; based on the strength prediction result, generating a strength report of the plastic parts, and optimizing the production process parameters of the plastic parts according to the feedback of the strength report. The present application has the effect of improving the accuracy of plastic parts strength detection.

Description

Technical Field

[0001] The present application relates to the technical field of plastic component strength detection, and in particular to a plastic component strength detection method and system. Background Art

[0002] At present, most traditional plastic parts strength testing methods rely on a single test method, such as mechanical tensile testing, compression testing, etc. These methods can provide certain strength data, but their testing process is usually destructive and cannot fully reflect the strength performance of plastic parts during actual use. In addition, existing methods often ignore the heterogeneity of materials during the production process and the structural non-uniformity of plastic parts. Especially in terms of mold design, injection molding process and material selection, it is very easy to cause tiny defects or uneven stress distribution inside plastic parts. These defects are often not discovered in time through traditional testing methods, which may eventually affect the overall strength and service life of the product.

[0003] Although existing strength testing methods have introduced some advanced sensor technologies, such as laser scanning and ultrasonic testing, these technologies are often used independently, and it is difficult to effectively consider multiple factors such as material surface, internal defects, and production environment. In addition, traditional analysis methods have weak modeling capabilities for heterogeneous materials and cannot comprehensively and accurately evaluate the strength performance of plastic parts. This leads to greater uncertainty in the results of strength testing, and it is impossible to provide timely feedback and adjustments in the production process, which affects the final quality and reliability of plastic parts.

[0004] The above-mentioned existing technical solutions have the following defects: the existing detection methods fail to fully consider the heterogeneity and structural non-uniformity of plastic parts during the production process, resulting in low accuracy and reliability of strength test results, so there is room for improvement. Summary of the invention

[0005] In order to improve the accuracy of plastic part strength detection, the present application provides a plastic part strength detection method and system.

[0006] The above-mentioned invention objective of the present application is achieved through the following technical solutions: A plastic part strength detection method, the plastic part strength detection method comprising: Obtain multi-dimensional sensor data of plastic parts; Preprocessing the multi-dimensional sensor data to obtain key feature data of the plastic part; Analyzing and evaluating the key characteristic data using a pre-trained strength model to obtain a strength prediction result of the plastic part; Based on the strength prediction result, a strength report of the plastic part is generated, and the production process parameters of the plastic part are optimized according to feedback from the strength report.

[0007] By adopting the above technical solution and acquiring multi-dimensional sensor data of plastic parts, the physical property information of plastic parts in different environments and states can be comprehensively collected, including surface morphology, internal defects, thermal distribution and other parameters, which helps to more accurately capture the diverse characteristics of plastic parts and avoid information omissions or deviations that may be caused by a single data source; by pre-processing the multi-dimensional sensor data, noise can be removed, abnormal data can be eliminated, and the data can be reduced in dimension to extract more accurate and representative key features, which can greatly improve the quality and availability of the data and ensure that subsequent analysis will not be disturbed by redundant or irrelevant data, thereby improving the accuracy of strength prediction results; by using pre-trained The trained strength model analyzes and evaluates key feature data and can process large amounts of high-dimensional data through deep learning and machine learning algorithms to obtain accurate strength prediction results. It can automatically identify key factors that affect the strength of plastic parts and quickly generate reliable strength predictions when new data is input, providing a scientific basis for subsequent process optimization. The strength report generated based on the strength prediction results can intuitively reflect the strength performance and potential risks of plastic parts, provide detailed data support for adjustments during the design and production process, help determine whether the strength of plastic parts meets the requirements, and further identify potential structural weaknesses and failure modes, thereby providing feedback to optimize production process parameters to improve the overall quality and performance of plastic parts.

[0008] In one example, the present application may be further configured as follows: the step of obtaining multi-dimensional sensor data of a plastic part includes: Acquiring surface geometric data of the plastic part by means of a laser scanning sensor, wherein the surface geometric data includes wall thickness distribution, surface smoothness and surface defect data; Acquiring internal defect data of the plastic part by means of an ultrasonic sensor, wherein the internal defect data includes the location, size and density of the defect; Acquiring thermal distribution data of the plastic part by using an infrared sensor; Acquiring production environment data related to the plastic part through an environmental sensor, wherein the production environment data includes humidity, air pressure and temperature change data; The surface geometry data, the internal defect data, the thermal distribution data and the production environment data are fused using a data fusion algorithm to obtain the multi-dimensional sensor data.

[0009] By adopting the above technical solution, the surface geometric data of plastic parts can be obtained by laser scanning sensors, and the surface profile, wall thickness distribution, smoothness and surface defects of plastic parts can be accurately measured, so as to fully understand the geometric characteristics of plastic parts, help identify surface irregularities and potential structural defects, and provide basic data for subsequent strength evaluation; the internal defect data of plastic parts can be obtained by ultrasonic sensors, and the hidden defects such as bubbles, cracks, and voids that may exist inside the plastic parts can be deeply detected, and the location, size and density of these defects can be accurately obtained, which is helpful to reveal internal problems that may affect the overall strength and reliability of plastic parts, and further improve the comprehensiveness of detection; the thermal distribution data of plastic parts can be obtained by infrared sensors, and the thermal behavior of plastic parts under different temperature conditions can be monitored, revealing possible thermal stress and temperature non-uniformity; the production environment data of plastic parts can be obtained by environmental sensors, and the changes in environmental conditions such as humidity, air pressure, temperature, etc. during the production process of plastic parts can be grasped in real time; the use of data fusion algorithms to fuse these multi-dimensional sensor data can complement the advantages of various data sources, eliminate the deviations and limitations that may be caused by single data, and obtain a more comprehensive and accurate description of the state of plastic parts by fusing information from different sources.

[0010] In one example, the present application may be further configured as follows: the preprocessing of the multi-dimensional sensor data to obtain the key feature data of the plastic part includes: Using Kalman filtering to denoise the multi-dimensional sensor data to obtain denoised data; The denoised data is processed by principal component analysis to reduce the dimensionality and extract the key feature data.

[0011] By adopting the above technical solution, by using Kalman filtering to denoise the multi-dimensional sensor data, random noise and unnecessary interference signals in the data can be effectively removed, thereby improving the quality and accuracy of the data and reducing the negative impact of noise on intensity prediction and model analysis; by using the principal component analysis method to reduce the dimensionality of the denoised data, high-dimensional sensor data can be compressed into fewer principal components while retaining the core information of the data, thereby removing redundant information, reducing computational complexity, and improving the efficiency of subsequent processing and analysis. At the same time, the most representative key feature data can be extracted, thereby providing a more effective input for the intensity model.

[0012] In one example, the present application may be further configured as follows: before analyzing and evaluating the key feature data using the pre-trained strength model to obtain the strength prediction result of the plastic part, the plastic part strength detection method further includes: Based on historical production data, sample data of plastic parts with known strength values ​​are collected, and the sample data are labeled to form a training set; Analyze the historical production data, identify and extract the influencing factors of the strength of the plastic parts, quantify the influencing factors, and determine the contribution value of the influencing factors to the strength of the plastic parts; Based on the contribution value, the strength model constructed based on the convolutional neural network and the support vector machine network is trained using the training set, and during the training process, the hyperparameters of the strength model are optimized through K-fold validation to obtain the pre-trained strength model; During the model training process, the transfer learning method is used to preliminarily optimize the intensity model.

[0013] By adopting the above technical solution, by collecting sample data of plastic parts with known strength values ​​based on historical production data and labeling the sample data, it is possible to ensure that the training set contains accurate labels and sample pairs, so that the training model can better learn the relationship between strength and different variables, and improve the accuracy and reliability of model prediction; by analyzing historical production data, identifying and extracting the influencing factors of the strength of plastic parts, and quantitatively describing these factors, it is possible to accurately locate the key variables affecting the strength of plastic parts, ensure that the model can take into account all important influencing factors, and help understand the contribution of different production conditions and parameters to strength, thereby enhancing the interpretability and prediction ability of the model; by training the strength model constructed by the convolutional neural network and the support vector machine network based on the contribution value and the training set, and optimizing the hyperparameters of the model through K-fold verification, it is possible to effectively improve the generalization ability of the model and avoid overfitting, thereby making the strength prediction more accurate and stable; by adopting the transfer learning method in the model training process, the existing pre-trained model can be used to optimize the new task, significantly speed up the training speed of the model, and improve its performance on specific tasks, so that the model can quickly adapt in the initial stage and provide higher prediction accuracy.

[0014] In one example, the present application may be further configured as follows: the key feature data is analyzed and evaluated using a pre-trained strength model to obtain a strength prediction result of the plastic part, including: Inputting the key feature data into the pre-trained strength model, wherein the pre-trained strength model includes a convolutional neural network and a fully connected layer; The convolutional neural network layer performs in-depth analysis on the key feature data through multiple convolutional layers, pooling layers and activation functions to extract local features and global features of the plastic part; Based on the key feature data, using the fully connected layer to integrate and analyze the local features and the global features to generate the strength prediction result; During the prediction process, if it is detected that the intensity prediction value in the intensity prediction result exceeds a set threshold, an alarm mechanism is triggered, and an alarm signal is generated and transmitted to the user end.

[0015] By adopting the above technical solution, by inputting the key feature data into the pre-trained strength model, it can be ensured that the model accurately analyzes the input data based on the existing knowledge and experience, so as to generate the strength prediction results that are consistent with the actual situation, avoiding the errors and uncertainties of the traditional manual analysis; by deeply analyzing the key feature data through the convolutional neural network layer through multiple convolutional layers, pooling layers and activation functions, valuable local features and global features can be extracted from the original data, which can reflect the performance and strength performance of plastic parts under different production conditions, and further enhance the prediction ability and accuracy of the model; by integrating and analyzing the local features and global features based on the key feature data and using the fully connected layer to generate the strength prediction results, the complex multi-dimensional features can be effectively integrated into an overall evaluation, and the comprehensive analysis ability of the model can be improved, so that the strength prediction results are more representative and accurate; by triggering the alarm mechanism and generating an alarm signal to transmit to the user end if the strength prediction value exceeds the set threshold during the prediction process, the potential quality problems of the plastic parts can be monitored in real time, and timely feedback can be given to the user so that measures can be taken to prevent or improve them, and avoid quality accidents or product failures caused by insufficient strength.

[0016] In one example, the present application may be further configured as follows: generating a strength report of the plastic part based on the strength prediction result includes: Filling the strength prediction result into a preset report template, wherein the report template includes a key information page of the plastic part, a strength prediction result page, an environmental adaptability assessment page, a risk assessment page, and an optimization suggestion page; Based on the strength prediction results, a failure mode analysis method is used to identify potential failure areas, and a predicted stress concentration area and a load range of failure are generated; According to the strength prediction results, the failure probability under different usage conditions is calculated and the corresponding risk level is given.

[0017] By adopting the above technical solution, by filling the strength prediction results into the preset report template, the content and format of the report can be standardized, so that the report has a unified structure and clear information presentation, which is convenient for users to understand and further analyze. The design of each module ensures comprehensive coverage of strength, environmental adaptability, risks and optimization suggestions, which helps to provide a complete strength analysis view of plastic parts; by using the failure mode analysis method to identify potential failure areas, the key parts of plastic parts where failures may occur can be accurately located, and the performance of these areas under different working conditions can be predicted, thereby providing users with the identification of high-risk areas, generating predicted stress concentration areas and failure load ranges, which helps to identify and avoid potential structural weaknesses or failure points in advance, thereby improving product reliability; by calculating the failure probability under different operating conditions and giving the corresponding risk level, the failure risk under different situations can be quantified, providing a basis for subsequent decision-making and optimization, helping users to better understand the performance risks of products under different conditions, thereby optimizing product design or production processes and improving product safety and stability.

[0018] In one example, the present application may be further configured as follows: optimizing the production process parameters of the plastic part according to the strength report feedback includes: adjusting the design parameters of the plastic part mold according to the stress concentration area identified in the strength prediction result; Optimizing process parameters during the injection molding of the plastic part according to the analysis of failure modes in the strength report; By analyzing the data within a preset production cycle, the production process of the plastic parts is reviewed and summarized regularly to generate an optimization plan.

[0019] By adopting the above technical solution, by adjusting the design parameters of the plastic part mold according to the stress concentration area identified in the strength prediction result, the mold structure can be improved in a targeted manner, the potential defects or damage risks caused by stress concentration can be reduced, and the overall strength and service life of the plastic part can be improved; by optimizing the process parameters in the injection molding process of the plastic part according to the analysis of the failure mode in the strength report, the parameters can be adjusted in a targeted manner during the production process to avoid or reduce defects caused by improper injection molding process, thereby improving production efficiency and product quality; by analyzing the data within the preset production cycle, regularly reviewing and summarizing the production process, and generating an optimization plan, the problems in the production process can be continuously monitored and adjusted in time, the production process can be continuously optimized, production stability can be improved, waste can be reduced, production efficiency can be improved, and the consistency and reliability of the final product can be ensured.

[0020] The second object of the invention is achieved by the following technical solutions: A plastic part strength detection system, the plastic part strength detection system comprising: A sensor data acquisition module is used to acquire multi-dimensional sensor data of plastic parts; A preprocessing module, used for preprocessing the multi-dimensional sensor data to obtain key feature data of the plastic part; An evaluation module, used to analyze and evaluate the key characteristic data using a pre-trained strength model to obtain a strength prediction result of the plastic part; A report generation module is used to generate a strength report of the plastic part based on the strength prediction result, and optimize the production process parameters of the plastic part according to the feedback of the strength report.

[0021] By adopting the above technical solution and acquiring multi-dimensional sensor data of plastic parts, the physical property information of plastic parts in different environments and states can be comprehensively collected, including surface morphology, internal defects, thermal distribution and other parameters, which helps to more accurately capture the diverse characteristics of plastic parts and avoid information omissions or deviations that may be caused by a single data source; by pre-processing the multi-dimensional sensor data, noise can be removed, abnormal data can be eliminated, and the data can be reduced in dimension to extract more accurate and representative key features, which can greatly improve the quality and availability of the data and ensure that subsequent analysis will not be disturbed by redundant or irrelevant data, thereby improving the accuracy of strength prediction results; by using pre-trained The trained strength model analyzes and evaluates key feature data and can process large amounts of high-dimensional data through deep learning and machine learning algorithms to obtain accurate strength prediction results. It can automatically identify key factors that affect the strength of plastic parts and quickly generate reliable strength predictions when new data is input, providing a scientific basis for subsequent process optimization. The strength report generated based on the strength prediction results can intuitively reflect the strength performance and potential risks of plastic parts, provide detailed data support for adjustments during the design and production process, help determine whether the strength of plastic parts meets the requirements, and further identify potential structural weaknesses and failure modes, thereby providing feedback to optimize production process parameters to improve the overall quality and performance of plastic parts.

[0022] In summary, this application includes the following beneficial technical effects: 1. By acquiring multi-dimensional sensor data of plastic parts, the physical property information of plastic parts in different environments and states can be comprehensively collected, including surface morphology, internal defects, thermal distribution and other parameters, which helps to more accurately capture the diverse characteristics of plastic parts and avoid information omissions or deviations that may be caused by a single data source; by pre-processing the multi-dimensional sensor data, noise can be removed, abnormal data can be eliminated, and the data can be reduced in dimension, so as to extract more accurate and representative key features, which can greatly improve the quality and availability of data and ensure that subsequent analysis will not be interfered by redundant or irrelevant data, thereby improving the accuracy of strength prediction results; 2. By using the pre-trained strength model to analyze and evaluate key feature data, a large amount of high-dimensional data can be processed through deep learning and machine learning algorithms to obtain accurate strength prediction results. It can automatically identify the key factors affecting the strength of plastic parts and quickly generate reliable strength predictions when new data is input, providing a scientific basis for subsequent process optimization; the strength report generated based on the strength prediction results can intuitively reflect the strength performance and potential risks of plastic parts, provide detailed data support for adjustments in the design and production process, help determine whether the strength of plastic parts meets the requirements, and further identify potential structural weaknesses and failure modes, and then feedback to optimize production process parameters to improve the overall quality and performance of plastic parts. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flow chart of a plastic part strength detection method in one embodiment of the present application; Figure 2 This is a flowchart for implementing step S10 in a plastic part strength detection method in one embodiment of the present application; Figure 3 This is a flowchart for implementing step S20 in a plastic part strength detection method in one embodiment of the present application; Figure 4 This is a flowchart for implementing step S30 in a plastic part strength detection method in one embodiment of the present application; Figure 5 This is another implementation flow chart of step S30 in a plastic part strength detection method in an embodiment of the present application; Figure 6 This is a flowchart for implementing step S40 in a plastic component strength detection method in an embodiment of the present application; Figure 7 This is another implementation flow chart of step S40 in a plastic component strength detection method in an embodiment of the present application; Figure 8 It is a principle block diagram of a plastic component strength detection system in one embodiment of the present application. DETAILED DESCRIPTION

[0024] The present application is further described in detail below in conjunction with the accompanying drawings.

[0025] In one embodiment, if Figure 1 As shown, the present application discloses a plastic part strength detection method, which specifically includes the following steps: S10: Acquire multi-dimensional sensor data of the plastic part.

[0026] Specifically, first, select appropriate sensors to collect comprehensive data on plastic parts. These sensors are responsible for collecting multi-dimensional data on the surface, interior and external environment of plastic parts. For example, use laser scanning sensors to scan the surface of plastic parts and collect their geometric data, including wall thickness distribution, surface smoothness and possible defects such as microcracks or pores; ultrasonic sensors are used to obtain defect data inside plastic parts, and determine the location, size and density of defects by analyzing the reflection of ultrasonic waves; infrared sensors are used to collect temperature distribution on the surface and inside of plastic parts. These temperature changes can reflect stress concentration or material unevenness that may occur in the manufacturing process of plastic parts; environmental sensors are responsible for recording parameters such as temperature, humidity, and air pressure in the production environment. These environmental factors have a direct impact on the strength and stability of plastic parts. Through the collaborative work of these sensors, data from all aspects of plastic parts can be fully collected to ensure the diversity and comprehensiveness of the data.

[0027] S20: Preprocess the multi-dimensional sensor data to obtain key feature data of the plastic part.

[0028] Specifically, after acquiring the multi-dimensional sensor data, the data must first be denoised and processed to improve data quality. The collected raw data can be filtered using the Kalman filter algorithm, which is a recursive estimation method that can predict the state at the current moment based on the state estimate at the previous moment and the current observation value, and correct the error caused by noise, thereby obtaining smooth and stable data. After denoising, the next step is to reduce the dimensionality of the data using the principal component analysis method. PCA maps the multi-dimensional sensor data to a new space through linear transformation, retaining the features that best represent the strength of the plastic parts, while removing redundant and irrelevant dimensions. The data after dimensionality reduction is the key feature data mentioned above. These data can more concentratedly reflect the main quality characteristics of plastic parts, remove unnecessary noise and redundant information, and provide more accurate input for subsequent strength analysis.

[0029] S30: Analyze and evaluate the key feature data using the pre-trained strength model to obtain a strength prediction result of the plastic part.

[0030] Specifically, the pre-processed key feature data is input into the pre-trained strength prediction model, and the model analyzes the data according to the learned rules to generate the strength prediction results of the plastic parts. The strength prediction model is usually based on deep learning or machine learning algorithms. The model can learn the complex relationship between various features and the strength of plastic parts through training with a large amount of historical data. For example, the model may have learned how to predict the strength of plastic parts based on their wall thickness, surface smoothness, internal defects and environmental conditions. Through calculation, the model can output a strength prediction value that reflects the maximum load that the plastic parts may bear in actual use. In this process, cross-validation technology and hyperparameter tuning methods are used during model training to improve the accuracy and stability of the model. Ultimately, the obtained strength prediction results can be used as a basis for decision-making to help determine whether the plastic parts meet the design and use requirements. For example, if the model detects that the strength prediction value of a thin-walled area of ​​the bumper exceeds the set safety threshold, it will predict that the area may break or deform in actual use, thereby affecting the safety of the entire vehicle. The result that the model may output is: "The bumper may experience crack propagation when a 1000 Newton impact force is applied, especially in the lower right corner area.

[0031] S40: Based on the strength prediction result, a strength report of the plastic part is generated, and the production process parameters of the plastic part are optimized according to the feedback of the strength report.

[0032] Specifically, after obtaining the strength prediction results of plastic parts, a detailed strength report is generated based on the prediction results. The report includes basic information of plastic parts, strength prediction results, possible environmental adaptability issues, and risk assessment. For example, if the strength prediction results show that there is high stress concentration in certain areas, the report will recommend a detailed analysis of these areas, and identify possible failure areas through failure mode analysis (FMEA), while evaluating the failure probability of plastic parts under different working conditions. The report will also provide optimization suggestions for the production process based on the strength prediction results, such as adjusting mold design parameters to reduce stress concentration, or modifying process parameters (such as injection pressure, temperature, speed, etc.) during the injection molding process to optimize the strength and stability of plastic parts. In addition, based on the feedback of prediction results and process parameters, the production process is reviewed regularly, summarized and improvement plans are generated to continuously optimize the production process and improve the overall quality of plastic parts.

[0033] By adopting the above technical solution and acquiring multi-dimensional sensor data of plastic parts, the physical property information of plastic parts in different environments and states can be comprehensively collected, including surface morphology, internal defects, thermal distribution and other parameters, which helps to more accurately capture the diverse characteristics of plastic parts and avoid information omissions or deviations that may be caused by a single data source; by pre-processing the multi-dimensional sensor data, noise can be removed, abnormal data can be eliminated, and the data can be reduced in dimension to extract more accurate and representative key features, which can greatly improve the quality and availability of the data and ensure that subsequent analysis will not be disturbed by redundant or irrelevant data, thereby improving the accuracy of strength prediction results; by using pre-trained The trained strength model analyzes and evaluates key feature data and can process large amounts of high-dimensional data through deep learning and machine learning algorithms to obtain accurate strength prediction results. It can automatically identify key factors that affect the strength of plastic parts and quickly generate reliable strength predictions when new data is input, providing a scientific basis for subsequent process optimization. The strength report generated based on the strength prediction results can intuitively reflect the strength performance and potential risks of plastic parts, provide detailed data support for adjustments during the design and production process, help determine whether the strength of plastic parts meets the requirements, and further identify potential structural weaknesses and failure modes, thereby providing feedback to optimize production process parameters to improve the overall quality and performance of plastic parts.

[0034] In one embodiment, if Figure 2 As shown, in step S10, the multi-dimensional sensor data of the plastic part is obtained, which specifically includes: S11: Acquire surface geometry data of the plastic part through a laser scanning sensor, where the surface geometry data includes wall thickness distribution, surface smoothness, and surface defect data.

[0035] Specifically, first, the plastic part is scanned using a laser scanning sensor to obtain its surface geometry data. The laser scanning sensor emits a laser beam and obtains surface information by measuring the time difference of the reflected light. The data includes the distribution of wall thickness, surface smoothness, and surface defects. For example, if the wall thickness of a certain area of ​​the plastic part is thin, or there are scratches or dents on the surface during the scanning process, these data will be recorded, reflecting potential problems in the area, such as weak structures or uneven surfaces that may affect the strength and service life of the plastic part.

[0036] S12: Acquire internal defect data of the plastic part through an ultrasonic sensor, where the internal defect data includes the location, size, and density of the defect.

[0037] Specifically, the ultrasonic sensor then detects whether there are pores, cracks or other types of defects inside the plastic part by emitting high-frequency sound waves and receiving the reflected signals. For example, if an ultrasonic scan finds a 2 mm diameter pore in a corner of a plastic part, or cracks are densely distributed in a certain area, these data will record the location, size and density of the defect, providing a basis for subsequent strength assessment.

[0038] S13: Obtaining thermal distribution data of the plastic part through an infrared sensor.

[0039] Specifically, infrared sensors are used to detect the thermal distribution data of plastic parts. Infrared sensors detect the thermal radiation on the surface and inside of an object to identify temperature differences and obtain thermal distribution data. For example, during the production process, a certain part of a plastic part may experience thermal stress concentration due to unreasonable mold design or uneven cooling. Infrared sensors can record this data and help analyze the possible thermal damage or stress concentration problems in that part, thereby providing a reference for subsequent strength prediction and design optimization.

[0040] S14: Acquire production environment data related to plastic parts through environmental sensors, where the production environment data includes humidity, air pressure, and temperature change data.

[0041] Specifically, environmental sensors are used to obtain data related to the production environment of plastic parts, including humidity, air pressure, and temperature changes. If the temperature or humidity fluctuates greatly during the production process, it may affect the physical properties of the plastic parts, such as its hardness and toughness. Environmental sensors can monitor and record this data in real time as important background information for evaluating the strength of plastic parts.

[0042] S15: Utilize a data fusion algorithm to fuse surface geometry data, internal defect data, thermal distribution data, and production environment data to obtain multi-dimensional sensor data.

[0043] Specifically, the data fusion algorithm integrates the data from different sensors through weighted average, maximum entropy method or other advanced algorithms, so that each data point can be effectively combined to produce a comprehensive multi-dimensional data set. It helps to integrate the influencing factors of different dimensions, such as surface defects, internal pores, temperature changes, etc., so as to provide more comprehensive and accurate data support for the strength prediction model. After data fusion, a comprehensive data model can be generated that can reflect the geometric shape of plastic parts, and can also combine the impact of their internal defects, thermal stress and production environment on their performance, and finally form a comprehensive multi-dimensional sensor data set to further improve the accuracy and reliability of plastic parts strength assessment.

[0044] In one embodiment, if Figure 3 As shown, in step S20, the multi-dimensional sensor data is preprocessed to obtain key feature data of the plastic part, specifically including: S21: Use Kalman filtering to denoise the multi-dimensional sensor data to obtain denoised data.

[0045] Specifically, Kalman filtering is a recursive estimation method that can gradually correct the observed signal data by combining the dynamic model of the system and the noise characteristics of the sensor, thereby reducing the error caused by noise. For example, during the sensor acquisition process, due to equipment accuracy or environmental interference, the measurement data may fluctuate or become unstable. Through Kalman filtering, the difference between the predicted value and the measured value is weighted, so that the result of each measurement is more accurate than the original data, and finally the denoised data is obtained, reflecting a more stable and reliable sensor output. For example, if a sensor generates a fluctuating signal due to changes in ambient temperature, Kalman filtering can correct this fluctuation by dynamically tracking historical data, making the final output data more accurate and removing irrelevant interference information.

[0046] S22: Perform dimensionality reduction processing on the denoised data through principal component analysis to extract key feature data.

[0047] Specifically, principal component analysis (PCA) is then used to reduce the dimensionality of the denoised data. PCA is a commonly used dimensionality reduction technique that converts the original data into a set of new uncorrelated variables (principal components) through linear transformation. These principal components are sorted according to the variance they explain, and the first few principal components can usually explain most of the changes in the data. For example, assuming that the denoised data contains information in multiple dimensions, such as temperature, pressure, surface defects, etc., there may be a high correlation between data of different dimensions. PCA calculates the covariance matrix and performs eigenvalue decomposition on it, and selects the first few eigenvectors with the largest variance as new principal components. This can effectively reduce the dimension of the data while retaining most of the information and extracting key feature data. If the original data contains 10 variables, after PCA processing, it may only be necessary to retain the first two or three principal components to explain more than 80% of the information in the original data, thereby simplifying the subsequent analysis and modeling process without losing too much important information, and further improving computational efficiency and accuracy.

[0048] In one embodiment, if Figure 4 As shown, before step S30, that is, before using the pre-trained strength model to analyze and evaluate the key feature data to obtain the strength prediction result of the plastic part, a plastic part strength detection method further includes: S301: Based on historical production data, sample data of plastic parts with known strength values ​​are collected, and the sample data are labeled to form a training set.

[0049] Specifically, representative samples of plastic parts are selected from historical production records to ensure that these sample data contain different strength values ​​and related production parameters (such as temperature, humidity, injection pressure, etc.). These data can come from plastic parts in different batches, different production environments or different production processes. The strength value of each plastic part (usually obtained through experimental testing) is used as the label of the sample, and it is combined with other production parameters to form a data pair containing input features and output labels. For example, if the temperature and pressure of a plastic part during the injection molding process are 180°C and 50MPa, respectively, and its strength test result is 150MPa, then this data will be part of the training set, where the input features are temperature and pressure, and the output label is a strength value of 150MPa. In this way, a training set with clear labels is formed.

[0050] S302: Analyze historical production data, identify and extract factors affecting the strength of plastic parts, quantify the factors, and determine the contribution of the factors to the strength of plastic parts.

[0051] Specifically, through data analysis, statistical analysis or correlation analysis methods, identify which production factors have a significant impact on the strength of plastic parts, such as temperature, humidity, injection pressure, cooling rate, etc. The relationship between these factors and the strength of plastic parts can be analyzed by means of correlation coefficient calculation, regression analysis, etc. By quantitatively describing these influencing factors, for example, it can be calculated that the effect of temperature on strength is that for every increase of 1°C, the strength increases by 0.2MPa, and the effect of pressure on strength is that for every increase of 1MPa, the strength increases by 0.5MPa, etc. Further, through this analysis method, the contribution value of each influencing factor to the strength of plastic parts is determined, and based on the contribution value, a more accurate strength model is constructed.

[0052] S303: Based on the contribution value, the strength model constructed based on the convolutional neural network and the support vector machine network is trained using the training set, and the hyperparameters of the strength model are optimized through K-fold verification during the training process to obtain a pre-trained strength model.

[0053] Specifically, during the model training process, the collected historical production data and labels are first input into the convolutional neural network and support vector machine network for training. The convolutional neural network can extract local features in the sample data through its multi-layer structure, and the support vector machine performs data classification or regression prediction by constructing a hyperplane. During the training process, the data set is divided into K subsets by the K-fold cross-validation method, each of which is used as a validation set in turn, and the other K-1 subsets are used as training sets. Multiple training and validation are performed to prevent overfitting and improve the generalization ability of the model. In each training process, the network hyperparameters (such as the convolution kernel size of the convolution layer, the C value of the support vector machine, and the selection of the kernel function, etc.) are adjusted to obtain the optimal hyperparameter combination, thereby constructing an accurate intensity prediction model.

[0054] S304: During the model training process, a transfer learning method is used to preliminarily optimize the strength model.

[0055] Specifically, transfer learning is a technique that accelerates the training of new models by borrowing knowledge from existing models, and is particularly suitable for situations where sample data is limited. In this embodiment, a pre-trained model that has been trained on a similar task can be selected, such as a CNN or SVM model that is widely used in the field of material strength prediction, and fine-tuned by applying it to the current task. First, pre-training is performed on the existing model to learn certain features, and then the last few layers of the model are fine-tuned or trained based on the new training data. In this way, the empirical knowledge of the existing model can be used to improve the prediction performance of the new task and reduce the time and computing resources required for training from scratch. Assuming that there is already a CNN model for strength prediction in a certain field, in a new production environment, due to the small amount of data, through transfer learning, the convolutional layer of the existing model is used as a preliminary feature extractor, and only the last fully connected layer is fine-tuned, which can significantly improve the effect of the model on the new task.

[0056] In one embodiment, if Figure 5 As shown, in step S30, the key feature data is analyzed and evaluated using the pre-trained strength model to obtain the strength prediction result of the plastic part, which specifically includes: S31: Input the key feature data into a pre-trained strength model, where the pre-trained strength model includes a convolutional neural network and a fully connected layer.

[0057] Specifically, the key feature data is input into the pre-trained strength model, which includes a convolutional neural network and a fully connected layer. At this stage, the key feature data (such as wall thickness distribution, surface defects, temperature, etc.) extracted from sensors or data analysis are first input into the trained model as input. The core part of the model includes a convolutional neural network (CNN) and a fully connected layer. CNN is responsible for feature extraction and local pattern recognition of the data, while the fully connected layer is used to integrate these features for global reasoning. Assuming that the model has been trained for strength prediction of different plastic parts and has good generalization ability, the input key feature data will be deeply analyzed by the strength model to obtain the corresponding strength prediction value.

[0058] S32: The convolutional neural network layer performs in-depth analysis on key feature data through multiple convolutional layers, pooling layers, and activation functions to extract local and global features of plastic parts.

[0059] Specifically, the convolution layer extracts features from the input data through convolution operations, extracting key local features such as local defects, textures, and shapes; the pooling layer performs dimensionality reduction processing on the convolved data through downsampling operations to reduce the complexity of the data while maintaining the significance of the features. For example, assuming that the input is a two-dimensional image representing the geometric morphology of the surface of a plastic part, the convolution layer can detect local features such as edges and textures in the image, while the pooling layer helps reduce the spatial distribution range of these features and extract more significant and concise features; the activation function is applied after each convolution layer to increase the nonlinear ability of the network, allowing the network to learn more complex feature patterns. Through multiple convolutions and pooling, CNN can automatically learn and extract representative local features from the input data, such as tiny cracks, bubbles, and thermal unevenness on the surface of plastic parts; global features include information such as the overall morphology and stress distribution of the material.

[0060] S33: Based on the key feature data, the fully connected layer is used to integrate and analyze the local features and global features to generate the intensity prediction results.

[0061] Specifically, the fully connected layer combines the local features extracted from the convolutional layer and the pooling layer with the overall information, and generates the final strength prediction result through a series of weighted summation and activation functions. In the fully connected layer, all the feature maps output by the convolutional layer will be expanded and processed by the nodes of the fully connected layer. These nodes obtain a prediction value through weighted summation and application of nonlinear activation functions (such as ReLU or sigmoid). For example, if the local defect information such as cracks and bubbles on the surface of the plastic part is extracted by the convolutional layer in the previous step, the fully connected layer will combine this information with the global characteristics of the material such as temperature and pressure, and obtain the predicted strength of the plastic part after comprehensive analysis, for example, the predicted value is 150MPa.

[0062] S34: During the prediction process, if it is detected that the intensity prediction value in the intensity prediction result exceeds the set threshold, an alarm mechanism is triggered, and an alarm signal is generated and transmitted to the user end.

[0063] Specifically, suppose the model predicts that the strength of a certain plastic part is 200MPa, and the set threshold is 180MPa. When the predicted value exceeds the set threshold, the system will trigger an alarm mechanism, generate an alarm and notify the user in time. When the predicted strength value exceeds the threshold, an internal control logic is immediately triggered to generate an alarm signal, which can be a visual or sound alarm, or sent to the user through a communication network. For example, through SMS, email or other instant messaging tools, the relevant operator or engineer is notified of the possible quality risks of the plastic part, and it is recommended to take corresponding actions, such as redesigning the mold or adjusting the production process, so as to avoid potential product failure risks.

[0064] In one embodiment, if Figure 6 As shown, in step S40, a strength report of the plastic part is generated based on the strength prediction result, which specifically includes: S41: Fill the strength prediction results into a preset report template, which includes a key information page of the plastic part, a strength prediction result page, an environmental adaptability assessment page, a risk assessment page, and an optimization suggestion page.

[0065] Specifically, the report template is a pre-set format that contains different types of information columns. First, fill in the basic data of the plastic parts such as production batch, material type, size, and usage scenario in the basic information page; then, fill in the strength prediction results (for example, the bearing strength and stress distribution of the plastic parts) into the strength prediction result page, where the predicted maximum bearing capacity, strength value and stress state of each part can be intuitively displayed; the environmental adaptability assessment page will evaluate the strength performance of the plastic parts in different environments according to their use environment (such as environmental factors such as temperature and humidity), and associate the prediction results with environmental adaptability; the risk assessment page combines the strength prediction results with potential failure modes to evaluate the risk points of plastic parts in actual use, especially marking high-risk areas; finally, the optimization suggestion page makes improvement suggestions based on the evaluation results, such as adjusting the production process or design optimization suggestions.

[0066] S42: Based on the strength prediction results, use the failure mode analysis method to identify potential failure areas and generate predicted stress concentration areas and failure load ranges.

[0067] Specifically, based on the strength prediction results, stress analysis models (such as finite element analysis) can be used to predict the stress distribution of different parts of plastic parts under specific working conditions. By looking at the stress concentration areas, it can be determined which areas may fail under load, and then the potential failure areas can be inferred. For example, suppose that during the strength prediction process, the joint part of a plastic part has stress concentration under load, and the stress in this area exceeds the yield strength of the material. At this time, the joint part can be identified as a potential failure area, and the failure load range of this area can be determined.

[0068] S43: Based on the strength prediction results, calculate the failure probability under different operating conditions and give the corresponding risk level.

[0069] Specifically, first, based on the strength prediction results, consider the impact of different operating conditions on plastic parts, such as temperature changes, load changes, etc., then simulate the stress state of the plastic parts according to these conditions, and then calculate the failure probability under each condition. The failure probability is usually estimated through historical data or simulation analysis. For example, the Monte Carlo method or Bayesian inference is used to simulate the stress distribution and failure mode of the material under different conditions, and the possibility of failure is obtained based on the simulation results, and finally the risk level is obtained. For example, if the failure probability of a plastic part exceeds 50% under a certain operating condition, the risk level under this condition can be calibrated as high risk; if the failure probability is low, the risk level can be low or medium risk. The assessment of risk levels can help decision makers take corresponding preventive measures in practical applications based on set standards.

[0070] In one embodiment, if Figure 7 As shown, in step S40, the production process parameters of the plastic parts are optimized according to the strength report feedback, specifically including: S44: Adjust design parameters of the plastic part mold according to the stress concentration area identified in the strength prediction result.

[0071] Specifically, first determine the stress distribution of plastic parts in different parts through strength prediction results, especially in stress concentration areas, which are usually weak links that are prone to cracking or deformation. For example, if the strength prediction shows that there is stress concentration at the joints or certain protrusions of the plastic parts, which exceeds the maximum load-bearing capacity of the material, then the mold design can be adjusted based on this information to optimize the shape or thickness of these areas to reduce stress concentration. For example, by thickening a part of the mold or adding support structures to certain parts, the stress distribution in these areas can be made more uniform, thereby improving the overall strength and reliability and reducing potential failure risks. By modifying the mold design parameters to optimize these stress concentration areas, the quality and service life of plastic parts can be effectively improved.

[0072] S45: Optimize the process parameters during the injection molding of plastic parts based on the analysis of failure modes in the strength report.

[0073] Specifically, failure mode analysis can reveal potential problems of defects and stress concentration that plastic parts may encounter during the injection molding process. Based on these analysis results, the process parameters used in the injection molding process, such as injection pressure, injection temperature, cooling time, injection speed, etc., can be adjusted to optimize the injection molding process of plastic parts. Specifically, if the failure mode analysis shows that a part of the plastic part has uneven structure or internal stress due to uneven injection pressure or too short cooling time, these problems can be improved by adjusting the injection pressure or extending the cooling time, thereby improving the molding quality and mechanical properties of the plastic part. For example, for certain thin-walled areas, if stress concentration is predicted and too fast cooling speed may cause deformation or cracks, adjusting the cooling rate and injection temperature can effectively reduce stress and ensure the overall stability of the plastic part during the molding process.

[0074] S46: By analyzing the data within the preset production cycle, the production process of plastic parts is reviewed and summarized regularly to generate optimization solutions.

[0075] Specifically, first collect and store various types of production data within the production cycle, such as temperature, humidity, production rate, injection pressure, cooling time, etc., conduct data analysis regularly, and evaluate the impact of various parameters in the production process on the strength, appearance and dimensional stability of plastic parts. For example, if it is found that a certain injection pressure or cooling time setting in a certain production cycle causes a large number of plastic parts to have defects or insufficient strength, then the production situation of this period can be regularly summarized to find out the root cause of the problem and generate corresponding optimization plans. The content of the optimization plan may include adjusting production parameters, improving equipment settings, optimizing material selection or changing production processes. Specifically, if it is found that the temperature fluctuations during the production process are large, resulting in unstable molding, it can be proposed to use a temperature control system in this production stage to maintain a stable production environment, thereby reducing defects in the production process. Ultimately, through this continuous review and optimization, the stability of the production process and the overall quality of plastic parts can be gradually improved.

[0076] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0077] In one embodiment, a plastic part strength detection system is provided, and the plastic part strength detection system corresponds one-to-one to a plastic part strength detection method in the above embodiment. Figure 8 As shown, the plastic parts strength detection system includes a sensor data acquisition module, a preprocessing module, an evaluation module and a report generation module. The detailed description of each functional module is as follows: A sensor data acquisition module is used to acquire multi-dimensional sensor data of plastic parts; A preprocessing module is used to preprocess multi-dimensional sensor data to obtain key feature data of plastic parts; An evaluation module is used to analyze and evaluate key feature data using a pre-trained strength model to obtain strength prediction results for plastic parts; The report generation module is used to generate a strength report of the plastic part based on the strength prediction results, and optimize the production process parameters of the plastic part according to the strength report feedback.

[0078] Optionally, the module for acquiring sensor data includes: The laser detection submodule is used to obtain the surface geometry data of the plastic part through a laser scanning sensor. The surface geometry data includes wall thickness distribution, surface smoothness and surface defect data. Ultrasonic detection submodule, used to obtain internal defect data of plastic parts through ultrasonic sensors, the internal defect data includes the location, size and density of defects; Infrared detection submodule, used to obtain thermal distribution data of plastic parts through infrared sensors; The environment detection submodule is used to obtain production environment data related to plastic parts through environmental sensors. The production environment data includes but is not limited to humidity, air pressure and temperature change data; The data integration submodule is used to fuse surface geometry data, internal defect data, thermal distribution data and production environment data using a data fusion algorithm to obtain multi-dimensional sensor data.

[0079] Optionally, the preprocessing module includes: The filtering submodule is used to perform denoising on the multi-dimensional sensor data by using Kalman filtering to obtain denoised data; The dimensionality reduction submodule is used to perform dimensionality reduction processing on the denoised data through principal component analysis to extract key feature data.

[0080] Optionally, the plastic part strength detection system further includes: A training set acquisition module is used to collect sample data of plastic parts with known strength values ​​based on historical production data, and label the sample data to form a training set; The contribution value determination module is used to analyze historical production data, identify and extract the factors affecting the strength of plastic parts, and quantitatively describe the influencing factors to determine the contribution value of the influencing factors to the strength of plastic parts; A training module is used to train the strength model constructed based on the convolutional neural network and the support vector machine network using the training set based on the contribution value, and optimize the hyperparameters of the strength model through K-fold verification during the training process to obtain a pre-trained strength model; The optimization module is used to perform preliminary optimization of the strength model using the transfer learning method during the model training process.

[0081] Optional assessment modules include: An input submodule, used to input key feature data into a pre-trained intensity model, the pre-trained intensity model includes a convolutional neural network and a fully connected layer; The deep analysis submodule is used for the convolutional neural network layer to perform deep analysis on key feature data through multiple convolutional layers, pooling layers and activation functions to extract local and global features of plastic parts; The integrated analysis submodule is used to integrate and analyze local features and global features based on key feature data using a fully connected layer to generate strength prediction results; The alarm submodule is used to trigger the alarm mechanism and generate an alarm signal to transmit to the user end if it is detected that the intensity prediction value in the intensity prediction result exceeds the set threshold during the prediction process.

[0082] Optionally, report generation modules include: The filling submodule is used to fill the strength prediction results into the preset report template. The report template includes the basic information page of the plastic part, the strength prediction result page, the environmental adaptability assessment page, the risk assessment page and the optimization suggestion page; The identification area submodule is used to identify potential failure areas based on the strength prediction results using the failure mode analysis method, and generate predicted stress concentration areas and failure load ranges; The calculation submodule is used to calculate the failure probability under different use conditions based on the strength prediction results and give the corresponding risk level.

[0083] The design parameter adjustment submodule is used to adjust the design parameters of the plastic part mold according to the stress concentration area identified in the strength prediction result; The process parameter adjustment submodule is used to optimize the process parameters during the injection molding process of plastic parts based on the analysis of failure modes in the strength report; The summary submodule is used to analyze the data within the preset production cycle, regularly review and summarize the production process of plastic parts, and generate optimization solutions.

[0084] The specific definition of a plastic part strength detection system can be found in the definition of a plastic part strength detection method mentioned above, which will not be repeated here. Each module in the above-mentioned plastic part strength detection system can be implemented in whole or in part through software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0085] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0086] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for testing the strength of a plastic part, characterized in that: The plastic part strength detection method comprises: Obtain multi-dimensional sensor data of plastic parts; Preprocessing the multi-dimensional sensor data to obtain key feature data of the plastic part; Analyzing and evaluating the key characteristic data using a pre-trained strength model to obtain a strength prediction result of the plastic part; Based on the strength prediction result, a strength report of the plastic part is generated, and the production process parameters of the plastic part are optimized according to feedback from the strength report.

2. A plastic part strength detection method according to claim 1, characterized in that: The obtaining of multi-dimensional sensor data of the plastic part comprises: Acquiring surface geometric data of the plastic part by means of a laser scanning sensor, wherein the surface geometric data includes wall thickness distribution, surface smoothness and surface defect data; Acquiring internal defect data of the plastic part by means of an ultrasonic sensor, wherein the internal defect data includes the location, size and density of the defect; Acquiring thermal distribution data of the plastic part by using an infrared sensor; Acquiring production environment data related to the plastic part through an environmental sensor, wherein the production environment data includes humidity, air pressure and temperature change data; The surface geometry data, the internal defect data, the thermal distribution data and the production environment data are fused using a data fusion algorithm to obtain the multi-dimensional sensor data.

3. A plastic part strength detection method according to claim 1, characterized in that: The preprocessing of the multi-dimensional sensor data to obtain the key feature data of the plastic part includes: Using Kalman filtering to denoise the multi-dimensional sensor data to obtain denoised data; The key feature data is extracted by performing dimensionality reduction processing on the denoised data using principal component analysis.

4. A plastic part strength detection method according to claim 1, characterized in that: Before analyzing and evaluating the key feature data using the pre-trained strength model to obtain the strength prediction result of the plastic part, the plastic part strength detection method further includes: Based on historical production data, sample data of plastic parts with known strength values ​​are collected, and the sample data are labeled to form a training set; Analyze the historical production data, identify and extract the influencing factors of the strength of the plastic parts, quantify the influencing factors, and determine the contribution value of the influencing factors to the strength of the plastic parts; Based on the contribution value, the strength model constructed based on the convolutional neural network and the support vector machine network is trained using the training set, and during the training process, the hyperparameters of the strength model are optimized through K-fold validation to obtain the pre-trained strength model; During the model training process, the transfer learning method is used to preliminarily optimize the strength model.

5. A plastic part strength detection method according to claim 1, characterized in that: The key feature data is analyzed and evaluated by using the pre-trained strength model to obtain the strength prediction result of the plastic part, including: Inputting the key feature data into the pre-trained strength model, wherein the pre-trained strength model includes a convolutional neural network and a fully connected layer; The convolutional neural network layer performs in-depth analysis on the key feature data through multiple convolutional layers, pooling layers and activation functions to extract local features and global features of the plastic part; Based on the key feature data, using the fully connected layer to integrate and analyze the local features and the global features to generate the strength prediction result; During the prediction process, if it is detected that the intensity prediction value in the intensity prediction result exceeds a set threshold, an alarm mechanism is triggered, and an alarm signal is generated and transmitted to the user end.

6. A plastic part strength detection method according to claim 1, characterized in that: Generating a strength report of the plastic part based on the strength prediction result includes: Filling the strength prediction result into a preset report template, wherein the report template includes a key information page of the plastic part, a strength prediction result page, an environmental adaptability assessment page, a risk assessment page, and an optimization suggestion page; Based on the strength prediction results, a failure mode analysis method is used to identify potential failure areas, and a predicted stress concentration area and a load range of failure are generated; According to the strength prediction results, the failure probability under different usage conditions is calculated and the corresponding risk level is given.

7. A plastic part strength detection method according to claim 1, characterized in that: Optimizing the production process parameters of the plastic part according to the strength report feedback includes: adjusting the design parameters of the plastic part mold according to the stress concentration area identified in the strength prediction result; Optimizing process parameters during the injection molding of the plastic part according to the analysis of failure modes in the strength report; By analyzing the data within a preset production cycle, the production process of the plastic parts is reviewed and summarized regularly to generate an optimization plan.

8. A plastic part strength detection system, characterized in that: The plastic part strength detection system comprises: A sensor data acquisition module is used to acquire multi-dimensional sensor data of plastic parts; A preprocessing module, used for preprocessing the multi-dimensional sensor data to obtain key feature data of the plastic part; An evaluation module, used to analyze and evaluate the key characteristic data using a pre-trained strength model to obtain a strength prediction result of the plastic part; A report generation module is used to generate a strength report of the plastic part based on the strength prediction result, and optimize the production process parameters of the plastic part according to the feedback of the strength report.

Images