A Process Parameter Optimization Control Method and System for Automobile Injection Molded Parts

Abstract

The present invention discloses a method and system for optimizing and controlling process parameters of automotive injection molded parts, belonging to the technical field of data processing. The method includes: based on the historical processing records of the target injection molded part, collecting sample data to train a feedforward neural network and constructing a product quality prediction plug-in; analyzing the influence of parameter fluctuations according to the historical processing records and constructing a fluctuation influence analyzer; determining the product demand characteristics of the target injection molded part in combination with the automotive usage scenario; using the product quality prediction plug-in and the fluctuation influence analyzer with the product demand characteristics as the expectation to optimize the process parameters, and outputting the optimal process parameters for processing control. This application solves the problems in the prior art that the control accuracy of injection molding process parameters is insufficient, resulting in large fluctuations in product quality and low production efficiency. By combining historical records, product quality prediction plug-ins and fluctuation analysis, the precise optimization of process parameters is realized, the product quality is stabilized, and the controllability and efficiency of the production process are improved.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a process parameter optimization control method and system for automobile injection molding parts. Background Art

[0002] With the rapid development of the automotive industry, the production demand for automotive parts is gradually developing towards high precision and high efficiency. As one of the important processes in the production of automotive parts, injection molding is widely used in the manufacture of body, interior, engine and other parts. The complexity and sophistication of the injection molding process requires that every parameter in the production process needs to be precisely controlled to ensure the quality of the final product.

[0003] At present, with the development of automation and intelligent manufacturing technology, how to improve the control accuracy of injection molding process through data analysis and optimization algorithms has become a technical problem that needs to be solved urgently in the industry. The existing technology often lacks in-depth mining and comprehensive utilization of historical processing data, and fails to fully consider the impact of fluctuations in various process parameters on product quality. Therefore, this application proposes a method for optimizing the process parameters of automotive injection molding parts based on neural networks and fluctuation analysis, aiming to improve the stability of the production process and the quality of the product. Summary of the invention

[0004] The present application provides a method and system for optimizing the control of process parameters of automotive injection molded parts, aiming to solve the problem in the prior art that the control accuracy of injection molding process parameters is insufficient, resulting in large fluctuations in product quality and low production efficiency.

[0005] In view of the above problems, the present application provides a method and system for optimizing and controlling process parameters of automotive injection molded parts.

[0006] The first aspect disclosed in the present application provides a method for optimizing and controlling process parameters of automotive injection molded parts, the method comprising collecting sample data to train a feedforward neural network based on historical processing records of target injection molded parts, and constructing a product quality prediction plug-in; performing parameter fluctuation impact analysis based on the historical processing records, and constructing a fluctuation impact analyzer; determining product demand characteristics of the target injection molded parts in combination with automobile usage scenarios; and optimizing process parameters based on the product demand characteristics, using the product quality prediction plug-in and the fluctuation impact analyzer, and outputting the optimal process parameters for processing control.

[0007] Another aspect disclosed in this application provides a process parameter optimization control system for automotive injection molded parts. The system includes a product quality prediction plug-in construction module for collecting sample data based on the historical processing records of the target injection molded parts to train a feedforward neural network and construct a product quality prediction plug-in; a fluctuation impact analyzer construction module for analyzing the impact of parameter fluctuations according to the historical processing records and constructing a fluctuation impact analyzer; a product requirement feature determination module for determining the product requirement features of the target injection molded parts in combination with the automotive usage scenarios; and an optimal process parameter output module for using the product quality prediction plug-in and the fluctuation impact analyzer with the product requirement features as the expectation to optimize the process parameters, output the optimal process parameters for processing control.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] Due to the technical solution of training a feedforward neural network with sample data based on historical processing records to construct a product quality prediction plug-in and combining a fluctuation impact analyzer to optimize and control the process parameters, the problems in the prior art of insufficient control accuracy of injection molding process parameters, resulting in large fluctuations in product quality and low production efficiency, are solved. By accurately predicting and optimizing the process parameters, human intervention can be reduced, the stability of the production process can be improved, the inconsistency of product quality can be reduced, and thus the production efficiency can be increased and the scrap rate can be reduced, achieving the technical effects of stabilizing product quality, increasing production efficiency, and improving process control accuracy.

[0010] The above description is only an overview of the technical solutions of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of this application more obvious and understandable, the specific embodiments of this application are specifically given below. Description of the Drawings

[0011] Figure 1 This is a schematic flow chart of a process parameter optimization control method for automotive injection molded parts provided by an embodiment of this application.

[0012] Figure 2 This is a schematic flow chart of constructing a product quality prediction plug-in in a process parameter optimization control method for automotive injection molded parts provided by an embodiment of this application.

[0013] Figure 3 This is a schematic structural diagram of a process parameter optimization control system for automotive injection molded parts provided by an embodiment of this application.

[0014] Description of the reference numerals: Product quality prediction plug-in construction module 11, Fluctuation impact analyzer construction module 12, Product requirement feature determination module 13, Optimal process parameter output module 14. Specific embodiments

[0015] The general idea of the technical solution provided by this application is as follows:

[0016] The embodiments of this application provide a method and system for optimizing and controlling the process parameters of automotive injection molded parts. By training a feedforward neural network with sample data based on historical processing records, a product quality prediction plug-in is constructed, and the process parameters are optimized in combination with a fluctuation impact analyzer. This method accurately predicts and analyzes the impact of process parameters on product quality, finds key parameters, optimizes their settings, reduces quality fluctuations during production, and thus improves the stability of product quality and production efficiency. Ultimately, it ensures that the injection molding process can meet the quality requirements of the product and improves the accuracy and efficiency of the overall manufacturing process.

[0017] After introducing the basic principle of this application, the various non-limiting embodiments of this application will be specifically introduced below in conjunction with the accompanying drawings of the specification.

[0018] Embodiment 1

[0019] As Figure 1 shown, the embodiments of this application provide a method for optimizing and controlling the process parameters of automotive injection molded parts, and the method includes:

[0020] Step S100: Based on the historical processing records of the target injection molded parts, collect sample data to train a feedforward neural network and construct a product quality prediction plug-in.

[0021] Specifically, the target injection molded parts refer to the automotive injection molded parts that ultimately need to be produced. These injection molded parts have specific design requirements and quality standards, such as the interior panels of cars, engine covers, etc. The historical processing records refer to the historical data recorded during the production process, including the process parameters during production (such as injection temperature, pressure, speed, etc.) and product quality data (such as dimensions, strength, surface quality, etc.). These data reflect the performance of each link in the past production process. The sample data refers to the data set extracted from the historical processing records and is used to train and optimize the model. The sample data contains a set of inputs and corresponding outputs. A feedforward neural network is an artificial neural network that performs calculations through forward propagation. The network consists of multiple layers, and the nodes of each layer are connected by weights. Each time the input data passes through the network layer by layer and finally outputs a prediction result. In this solution, it is used to predict the product quality under different process parameters. The product quality prediction plug-in is a model obtained by training a feedforward neural network and can predict the quality indicators of injection molded parts, such as dimensional accuracy, surface quality, etc., according to the input process parameters. The role of this plug-in is to predict the quality of the final product based on the process parameters and provide a basis for subsequent optimization.

[0022] First, it is necessary to collect and analyze data based on the historical processing records of the target injection-molded part. Specifically, the historical processing records include key process parameters during the injection molding process, such as injection temperature, injection pressure, cooling time, etc., as well as the corresponding product quality data, such as dimensional accuracy, strength, surface finish, etc. These data represent the relationship between various process parameters and product quality in the past production process.

[0023] Next, sample data is extracted from these historical processing records. Each piece of sample data contains a set of process parameters and their corresponding quality indicators. These sample data will be used as training data to train a feedforward neural network. The feedforward neural network is trained through these sample data to learn how to predict the product quality from the given process parameters. This process usually involves multiple iterations. In each iteration, the neural network adjusts its internal parameters (such as weights and biases) so that the output result of the network gradually approaches the actual quality data. After training is completed, the neural network will generate a product quality prediction plugin that can receive new process parameters as input and predict the quality of the injection-molded part under these parameters.

[0024] Through this step, the process parameters can be adjusted in real time during the production process without relying on manual experience. For example, assume that in a certain production, new injection temperatures and pressures are used. The system quickly evaluates the quality problems under this setting through the prediction plugin, so that the process parameters can be adjusted in time to avoid the production of defective products. By continuously optimizing the prediction plugin, more precise process control can be achieved, thereby improving production efficiency and product consistency.

[0025] Step S200: Analyze the influence of parameter fluctuations based on the historical processing records and construct a fluctuation influence analyzer.

[0026] Specifically, parameter fluctuations refer to the range of changes in process parameters (such as injection temperature, injection pressure, etc.) during the production process. Usually, there will be certain fluctuations in process parameters, and these fluctuations will affect the quality of the final product. The fluctuation influence analyzer can analyze the influence of the fluctuations of process parameters on product quality. By quantifying the fluctuation degrees of different parameters, it helps to identify which parameter fluctuations have a significant impact on product quality and provides a basis for optimizing process control. Fluctuation influence analysis refers to analyzing the relationship between process parameter fluctuations and product quality to identify the specific influence degree of each parameter fluctuation on product quality. Through fluctuation analysis, key parameters can be identified and their fluctuations can be controlled to avoid the production of defective products.

[0027] First, conduct an analysis of the impact of parameter fluctuations based on historical processing records. The first step is to screen out the process parameters and corresponding product quality data in the historical processing records. For example, injection molding production data from the past few months can be selected, which records process parameters such as injection temperature, pressure, and cooling time in different production cycles, and corresponds to the product quality data for each cycle, such as dimensional error, warpage, etc. Next, based on this historical data, conduct a parameter fluctuation analysis. First, select a specific sequence of process parameters (for example, the injection temperature in all production cycles). Then, calculate the fluctuations of this process parameter in different production cycles. Specifically, calculate the fluctuation characteristics of each parameter. Commonly used methods include calculating statistical indicators such as standard deviation and variance.

[0028] After obtaining the fluctuation characteristics of each process parameter, conduct a quality impact analysis. By correlating the fluctuation characteristics of each parameter with the product quality data, determine which process parameter fluctuations have a significant impact on quality. Based on these analysis results, construct a fluctuation impact analyzer.

[0029] Through this method, it is possible to accurately analyze and identify the specific impacts of different process parameter fluctuations on product quality. This will help manufacturers optimize parameter settings during the production process, reduce unnecessary fluctuations, and ensure the quality of the final product is more stable.

[0030] Step S300: Determine the product requirement characteristics of the target injection molded part in combination with the automotive usage scenario.

[0031] Specifically, the automotive usage scenario refers to various environments and operating conditions that an automobile faces during daily use, such as driving speed, temperature changes, collisions, or vibrations. The design and manufacture of automotive injection molded parts need to consider these usage conditions to ensure the reliability and durability of the components in actual use. The target injection molded part refers to a specific injection molded part, such as a body panel, interior trim, engine hood, etc. These components need to meet certain design and functional requirements. The product requirement characteristics refer to the product design standards and quality requirements defined based on the usage requirements and performance requirements of the target injection molded part. These characteristics include dimensional accuracy, strength, surface quality, high temperature resistance, corrosion resistance, etc., and reflect the various technical indicators and performance requirements of the product. The target product characteristics refer to all the key performance indicators determined based on the functions and usage requirements of the target injection molded part, such as strength, stiffness, dimensional accuracy, impact resistance, heat resistance, etc. These characteristics are the standards that must be met during the design and manufacturing process.

[0032] First, consider the environment and working conditions that an automobile faces during actual use. For example, the engine hood of an automobile needs to withstand large temperature differences, so it needs to have good high-temperature resistance; while the interior trim parts of the vehicle body require high scratch resistance and UV resistance. To ensure the performance of the target injection molded part in these usage scenarios, we need to combine the actual situation and clarify the demand characteristics of the product. Specifically, analyze the various working conditions in the automobile usage scenario. Taking the engine hood as an example, it will be affected by factors such as high temperature, vibration, and collision. Therefore, its design needs to meet the following characteristics: 1) High-temperature resistance: It is required to maintain the strength and stability of the material at high temperatures; 2) Vibration resistance: The material is required to withstand the vibration from the engine without breaking or deforming; 3) Corrosion resistance: Due to the heat and chemical substances generated during engine operation, the material is required to have good corrosion resistance. Next, according to the requirements of these actual usage scenarios, clarify the product demand characteristics of the target injection molded part and specify them into a series of performance indicators. By analyzing the usage scenario, clarify the product demand characteristics of the engine hood, such as high-temperature resistance, impact resistance, dimensional accuracy, and corrosion resistance. These characteristics determine the design standards and manufacturing requirements of the target injection molded part.

[0033] By combining the automobile usage scenario to determine the product demand characteristics of the target injection molded part, it is possible to ensure that the components have sufficient performance and reliability in the actual working environment.

[0034] Step S400: Using the product demand characteristics as the expectation, utilize the product quality prediction plug-in and the variation impact analyzer to optimize the process parameters and output the optimal process parameters for processing control.

[0035] Specifically, the optimal process parameters refer to a set of process parameters obtained through the optimization process. These parameters can produce the best product quality and production efficiency during the production process and meet the product demand characteristics.

[0036] First, by clarifying the product demand characteristics of the target injection molded part, determine the key quality indicators to be achieved. Next, use the product quality prediction plug-in and input the historical processing records and target process parameters into the plug-in. The product quality prediction plug-in predicts the quality of the final product based on the given process parameters through technologies such as feedforward neural networks. At the same time, the variation impact analyzer also plays a role in the work. By analyzing the historical data, the variation impact analyzer helps to identify the degree of influence of the variation of different process parameters on the quality of the final product. For example, when the temperature fluctuates greatly, the warpage has a significant impact on the product quality, while the pressure fluctuation has a greater impact on the strength. The variation impact analyzer will evaluate the stability of different process parameters based on this analysis result and provide a basis for optimization.

[0037] On this basis, by combining the use of the product quality prediction plug-in and the fluctuation impact analyzer, the process parameters are optimized. The core of this process is to input different process parameters, simulate their impacts on product quality, and find a set of process parameter combinations that best meet the product demand characteristics. The finally output optimal process parameters will ensure that the product quality reaches the expected standard and improve production efficiency.

[0038] Through this optimization process, a set of optimal process parameters can be found during the production process, enabling the product quality to always be within the expected standard and making the production process more stable and efficient.

[0039] Furthermore, as Figure 2 shown, based on the historical processing records of the target injection molded part, sample data is collected to train a feedforward neural network, and a product quality prediction plug-in is constructed, including: based on the historical processing records of the target injection molded part, the injection temperature, injection pressure, injection speed, cooling time, and cooling temperature under K consecutive monitoring nodes within the historical processing cycle are collected to obtain multiple sample parameter sequence sets, where K is an integer greater than 10; the product quality indicators under different sample parameter sequence sets are obtained to obtain multiple sample quality data sets, where the product quality indicators at least include dimensional accuracy, structural strength, surface quality, molding quality, and material properties; the multiple sample parameter sequence sets and multiple sample quality data sets are used to train the feedforward neural network to generate the product quality prediction plug-in.

[0040] Specifically, the K monitoring nodes refer to a set of time series points selected during the historical processing. These nodes record the process parameters at different time points and are used to establish the relationship between different parameters and product quality, where K is a positive integer greater than or equal to 1. The sample parameter sequence set refers to the process parameters under K consecutive monitoring nodes extracted from the historical processing records, and the combination of these parameters forms a set of sample data. The sample quality data set refers to obtaining the corresponding product quality data set based on each set of sample parameter sequence sets. The quality data includes dimensional accuracy, structural strength, surface quality, molding quality, etc., and can be obtained through testing or measurement.

[0041] First, sample data is collected from the historical processing records of the target injection molded part. The historical processing records include the process parameters and product quality data in each production cycle. By selecting K consecutive monitoring nodes in the historical processing cycle, that is, the process parameters and quality data corresponding to each node during an injection molding operation, a sample parameter sequence set is constructed. Assuming K is an integer greater than 10, each node records a series of process parameters such as the injection temperature, injection pressure, and cooling time at that time, and these consecutive process data form a sample parameter sequence set.

[0042] Next, input these sets of sample parameter sequences and the sample quality datasets into a feedforward neural network for training. The feedforward neural network will learn how to predict the final product quality metrics (output data) based on these process parameters (input data). Through multiple training iterations, the network will continuously adjust its internal weights and parameters to reduce the error between the predicted results and the actual product quality, thereby generating a product quality prediction plugin.

[0043] Through this method, accurate product quality prediction can be achieved, ensuring that the quality of each batch of products during the production process is within the design range. It improves the stability and efficiency of production, reduces the generation of defective products, and enhances the overall quality and appearance performance of automotive injection molded parts.

[0044] Furthermore, training the feedforward neural network using the multiple sets of sample parameter sequences and multiple sample quality datasets to generate the product quality prediction plugin includes: using the multiple sets of sample parameter sequences and multiple sample quality datasets as training data, and equally dividing them into P parts to obtain P training sets, where P is an integer greater than 15; training the feedforward neural network using the P training sets respectively until convergence, obtaining P product quality prediction units, and constructing the product quality prediction plugin based on the combination of the P product quality prediction units.

[0045] Specifically, the P training sets refer to dividing the training data into P parts, with each part becoming a separate training set, where P is an integer greater than 15. Each training set is a randomly selected part from the entire training data. The product quality prediction unit refers to a small module trained by the feedforward neural network, specifically used to predict certain product quality metrics based on the input process parameters. Each prediction unit is a functional unit obtained by the neural network through training. The product quality prediction plugin refers to combining multiple product quality prediction units to finally obtain a complete model that can be used to predict the quality of injection molded parts. This plugin can predict product quality based on the input of new process parameters.

[0046] First, use the multiple sets of sample parameter sequences and multiple sample quality datasets as training data, and divide them into P training sets, where P is an integer greater than 15. For example, if we have 500 sets of sample data, divide these data evenly into 20 training sets, with each training set containing 25 sets of sample data. The sample data in each training set includes multiple sets of process parameters and the corresponding quality data.

[0047] Then, use these P training sets to train the feedforward neural network respectively. Each training set will be independently trained through the feedforward neural network until the network converges, that is, the network has learned a set of parameters that can better predict the product quality. This process is achieved through multiple iterations. In each iteration, the neural network adjusts its weights according to the error until it obtains a relatively accurate prediction ability.

[0048] After training each training set, a product quality prediction unit will be obtained. After P times of training, the network will generate P such prediction units. These prediction units respectively represent the learning results of the neural network on different training sets and have the ability to predict the product quality based on the given process parameters. Finally, combine these P product quality prediction units to construct a complete product quality prediction plug-in. This plug-in can accept new process parameter inputs and predict the product quality under these parameters.

[0049] By equally dividing the training data into multiple training sets and training them separately, it is possible to avoid over-reliance on a certain part of the data and improve the generalization ability of the neural network. The independent training of each training set enables the network to learn the product quality prediction model from multiple perspectives and multiple data sets, thereby improving the prediction accuracy and robustness.

[0050] This step can not only improve the stability of the production process but also optimize the process parameter control in the production process, thereby improving the consistency and controllability of the product quality, ultimately reducing the scrap rate and rework rate, and enhancing the overall production efficiency.

[0051] Furthermore, perform parameter fluctuation impact analysis based on the historical processing records, and construct a fluctuation impact analyzer, including: according to the historical processing records, screen and obtain multiple historical processing record data sets under multiple predetermined identical parameter sequences, where the historical processing record data includes the actual parameter sequence and the product quality data; randomly select the first predetermined identical parameter sequence, as well as the first set of actual parameter sequences and the first product quality data set; calculate the parameter fluctuation characteristics of multiple first actual parameter sequences in the first set of actual parameter sequences to obtain the first set of parameter fluctuation coefficients; based on the expected product quality as the benchmark, perform quality impact analysis according to the first product quality data set, and output the first set of quality impact coefficients; based on the first predetermined identical parameter sequence, the first set of parameter fluctuation coefficients, and the first set of quality impact coefficients, map and construct the first fluctuation impact analysis branch, and sequentially generate multiple fluctuation impact analysis branches to construct the fluctuation impact analyzer.

[0052] Specifically, the actual parameter sequence refers to the sequence of actual process parameters recorded during the production process, such as injection temperature, pressure, injection speed, etc. These data can be used for subsequent analysis of the impact of different parameter settings on product quality. The parameter fluctuation characteristics refer to the amplitude and pattern of change of a certain process parameter in different production cycles. Usually, statistical measures such as standard deviation and variance are calculated to describe the degree of parameter fluctuation. The parameter fluctuation coefficient set is a set composed of the fluctuation characteristics of multiple process parameters. The fluctuation coefficient of each process parameter reflects the amplitude of change of this parameter during the production process. The larger the fluctuation coefficient, the greater the impact of this parameter on quality. The quality impact coefficient is used to represent the specific impact of the fluctuation of a certain process parameter on product quality. By analyzing historical data, the degree of impact of fluctuations in different process parameters on quality can be obtained. The fluctuation impact analysis branch refers to the multiple analysis branches generated during the fluctuation impact analysis process based on different process parameters, fluctuation coefficients, and quality impact coefficients. Each branch represents an analysis path under a specific set of process conditions, helping to better understand the specific impact of different parameters on product quality.

[0053] First, select multiple data sets containing a predetermined identical parameter sequence from historical processing records. For example, records where the injection temperature remains within the same range (such as 250°C - 255°C) in all production cycles, or data where the injection pressure fluctuates within a specific range (such as 800 bar - 850 bar). In this way, by selecting data sets under the same process conditions, the influence of other factors can be eliminated, and the focus can be on analyzing the impact of the fluctuation of a single process parameter on quality.

[0054] Next, randomly select the first predetermined identical parameter sequence and the corresponding actual parameter sequence set and product quality data set from these identical parameter sequences. For example, assume that the temperature in a certain process cycle is 252°C, the pressure is 820 bar, and the cooling time is 14 seconds. The product quality data includes the warpage and dimensional accuracy of this cycle. These data will serve as the basis for analysis and be used for subsequent fluctuation calculation and quality impact analysis.

[0055] Then, calculate the parameter fluctuation characteristics of multiple process parameters in the selected first actual parameter sequence. For example, the standard deviation of the injection temperature, the fluctuation amplitude of the injection pressure, etc. These fluctuation characteristics reflect the change situation of each process parameter during the production process. Through the calculated first parameter fluctuation coefficient set, we can quantify the fluctuation amplitude of each process parameter and thus understand which process parameters have a greater impact on product quality.

[0056] Subsequently, based on the expected product quality, quality impact analysis is performed according to the selected first product quality dataset. At this time, the impact of fluctuations in different process parameters on product quality is analyzed. For example, fluctuations in injection temperature have a greater impact on warpage, while injection pressure has a more significant impact on product strength. These results will be output in the form of a quality impact coefficient set, which is used to evaluate the actual impact of fluctuations in each process parameter on product quality.

[0057] Finally, based on the analysis results, a fluctuation impact analysis branch is constructed through the first predetermined identical parameter sequence, the first parameter fluctuation coefficient set, and the first quality impact coefficient set. These branches show how fluctuations in different process parameters affect the quality of the final product, and multiple analysis branches are generated accordingly, and finally a complete fluctuation impact analyzer is constructed.

[0058] By constructing the fluctuation impact analyzer, the specific impact of fluctuations in different process parameters on product quality can be better understood. This analyzer can provide precise guidance to help manufacturers control key process parameters during production, avoid unnecessary quality fluctuations, and thus improve product consistency and stability.

[0059] Furthermore, the parameter fluctuation characteristics of multiple first actual parameter sequences are calculated to obtain the first parameter fluctuation coefficient set, including: calculating the standard deviation and mean of multiple parameters in multiple first actual parameter sequences respectively to obtain multiple first parameter standard deviation sets and multiple first parameter mean sets; based on the multiple first parameter standard deviation sets and multiple first parameter mean sets, weighted fusion is performed to obtain multiple first comprehensive parameter standard deviations and multiple first comprehensive parameter means, where the weight is positively correlated with the impact degree of the parameter on product quality; setting the ratio of the first comprehensive parameter standard deviation to the first comprehensive parameter mean as the first parameter fluctuation coefficient, and calculating the first parameter fluctuation coefficient set according to the multiple first comprehensive parameter standard deviations and multiple first comprehensive parameter means.

[0060] Specifically, the parameter mean refers to the average value of the process parameter in multiple production cycles, which represents the normal level of the parameter. The mean is used to describe the typical setting of the process parameter. The comprehensive parameter standard deviation refers to the result of weighted fusion of the standard deviations of multiple process parameters. This can reflect the comprehensive impact of fluctuations in different parameters on the quality of the final product. The comprehensive parameter mean refers to the result of weighted fusion of the means of multiple process parameters, representing the impact of the average level of different process parameters on product quality.

[0061] First, statistical analysis needs to be performed on each process parameter in multiple first actual parameter sequences. For each process parameter, its standard deviation and mean are calculated. The standard deviation reflects the fluctuation range of the process parameter in multiple production cycles, while the mean represents the normal level of the parameter during the production process. Next, using the calculated multiple first parameter standard deviation sets and first parameter mean sets, weighted fusion is carried out to obtain the first comprehensive parameter standard deviation and the first comprehensive parameter mean. Weighted fusion means assigning different weights according to the influence degree of each process parameter on product quality. Usually, parameters with a greater impact on product quality are assigned higher weights, while parameters with a smaller impact are assigned lower weights. The purpose of this weighted fusion is to combine the fluctuation characteristics of multiple process parameters to form a comprehensive fluctuation index.

[0062] Then, the fluctuation coefficient of each process parameter is calculated, that is, the ratio of the first comprehensive parameter standard deviation to the first comprehensive parameter mean is used as the fluctuation coefficient. The larger the fluctuation coefficient, the more significant the impact of the fluctuation of the process parameter on product quality. Finally, through the above steps, a first parameter fluctuation coefficient set is obtained, which includes the fluctuation impacts of each process parameter on product quality. These fluctuation coefficients will help us identify which process parameter fluctuations have a greater impact on product quality and provide a basis for subsequent process optimization.

[0063] Through this method, the impact of the fluctuation of each process parameter on product quality can be quantified and analyzed. The finally obtained parameter fluctuation coefficient set provides a basis for process optimization. Based on these fluctuation coefficients, the process parameters with a greater impact on quality can be focused on, and corresponding optimization measures can be taken to reduce parameter fluctuations and improve the stability of product quality.

[0064] Furthermore, with the product demand characteristics as the expectation, using the product quality prediction plug-in and the fluctuation impact analyzer, process parameter optimization is carried out to output the optimal process parameters, including: obtaining the processing parameter thresholds of the target injection molding part, randomly selecting multiple process parameters within the processing parameter thresholds, where each process parameter includes multiple parameter sequences; inputting the multiple process parameters into the fluctuation impact analyzer for similarity matching and outputting multiple quality impact coefficients with the highest similarity; with the product demand characteristics as the expectation, using the product quality prediction plug-in, based on the multiple quality impact coefficients and multiple process parameters, process parameter optimization is carried out to output the optimal process parameters.

[0065] Specifically, the processing parameter threshold refers to the maximum and minimum acceptable values of process parameters during the production process. It defines the operating range of process parameters to ensure that the process operations within this range can guarantee product quality. The process parameter sequence refers to the continuous changes experienced by each process parameter over multiple production cycles during the production process. For example, the continuous change of injection molding temperature from 250°C to 260°C within a production cycle is a process parameter sequence. Similarity matching refers to comparing and matching among multiple process parameters to find the historical data most similar to the current process conditions. In this way, the quality results that may occur under the new process conditions can be predicted.

[0066] First, determine the processing parameter threshold of the target injection molded part, that is, the acceptable range of process parameters. For example, assume the target injection molded part is an automotive interior part, with the required dimensional accuracy of ±0.1mm, the injection molding temperature range set from 250°C to 260°C, and the injection pressure range from 800 bar to 850 bar. These parameter thresholds help to define the operating range of process parameters, ensuring that the products produced within this range can meet the expected quality requirements.

[0067] Next, randomly select multiple process parameters within the processing parameter threshold, where each process parameter includes multiple parameter sequences. For example, the sequence of injection molding temperature includes multiple settings (250°C, 255°C, 260°C), and the sequence of injection pressure includes 800 bar, 810 bar, 820 bar, etc.

[0068] Input these randomly selected process parameters into the fluctuation impact analyzer for similarity matching. During this process, the fluctuation impact analyzer finds the historical data most similar to the current process parameter settings based on historical data and the fluctuations of process parameters, and obtains the quality impact coefficient with the highest similarity to the current settings by comparing the quality performance in the historical data. For example, assume the selected process parameters include a temperature of 250°C and a pressure of 800 bar, and the fluctuation impact analyzer finds the historical data similar to these parameter settings and outputs a quality impact coefficient for temperature and pressure.

[0069] Finally, based on these quality impact coefficients and process parameter sequences, use the product quality prediction plug-in to optimize the process parameters. The plug-in optimizes the process parameter combination according to the input parameters and quality impact coefficients, and outputs a set of optimal process parameters that can best meet the product demand characteristics and improve production efficiency. This process helps to find the best production parameter settings by simulating the impact of different process parameters on product quality.

[0070] Through this optimization process, it is possible to ensure that the target injection-molded part meets the optimal quality standards during the production process while improving production efficiency. This method can accurately find the most suitable process parameters based on historical data, fluctuations in process parameters, and product demand characteristics, thereby reducing quality fluctuations during production and ensuring product consistency and high quality.

[0071] Furthermore, with the product demand characteristics as the expectation, using the product quality prediction plug-in, process parameter optimization is carried out according to the multiple quality impact coefficients and multiple process parameters, including: adjusting the product demand characteristics according to the multiple quality impact coefficients respectively to obtain multiple corrected product demand characteristics; using the product quality prediction plug-in to predict the product quality according to the multiple quality impact coefficients and multiple process parameters, and outputting multiple predicted product quality data; carrying out process parameter optimization according to the multiple corrected product demand characteristics and multiple predicted product quality data, and outputting the optimal process parameters.

[0072] Specifically, the corrected product demand characteristics refer to adjusting the original product demand characteristics to better adapt to fluctuations in process parameters and actual production conditions. The corrected demand characteristics can reflect the ideal state that the product should reach under given process conditions.

[0073] First, use multiple quality impact coefficients to adjust the product demand characteristics according to the influence degree of each process parameter on product quality. For example, assume that the target injection-molded part is a car engine hood, and its product demand characteristics include dimensional accuracy of ±0.1 mm, strength ≥50 J, and high-temperature resistance ≥120 °C. During the actual production process, it is found that temperature fluctuations have a greater impact on dimensional accuracy, while pressure fluctuations have a more significant impact on strength. Therefore, based on these quality impact coefficients, we can correct the original product demand characteristics. For example, temperature fluctuations will change the dimensional accuracy requirement to ±0.15 mm, and pressure fluctuations will require the strength to reach 55 J to adapt to fluctuations in process parameters.

[0074] Next, use the product quality prediction plug-in to predict the product quality according to these adjusted corrected product demand characteristics and multiple different process parameters. For example, after inputting multiple process parameters, the product quality prediction plug-in will output the predicted product quality data under different process parameter combinations. These data include dimensional error, warpage, surface quality, etc., to evaluate the quality performance of the product under given process conditions.

[0075] Finally, by comparing multiple calibrated product requirement features with multiple predicted product quality data and using an optimization algorithm to optimize process parameters, the optimal process parameters that can best meet product requirements are obtained. For example, among multiple combinations of process parameters, a certain set of process parameters can make the dimensional error of the product be ±0.1 mm, the warpage degree is the smallest, and the strength reaches the target requirement, so these process parameters are output as the optimal production settings.

[0076] Through this step, the process control in the production process can be optimized through the precise matching of multiple quality impact coefficients and process parameters, ensuring that the product quality meets the predetermined requirement features. It can effectively reduce the scrap rate, reduce the production of non-conforming products, improve the consistency of product quality, and ensure that the production results of each batch can meet the requirements and technical standards.

[0077] Furthermore, using the product quality prediction plug-in, product quality prediction is performed according to the multiple quality impact coefficients and multiple process parameters, and multiple predicted product quality data are output, including: randomly selecting a first process parameter and a first quality impact coefficient; multiplying the ratio of the first quality impact coefficient to the historical maximum quality impact coefficient by P and rounding up to obtain Q, which is set as the number of selected units in the first unit. Randomly select Q product quality prediction units in the product quality prediction plug-in, perform product quality prediction on the first process parameter, output Q prediction results, calculate the mean value to obtain the first predicted product quality data, and add it to the multiple predicted product quality data.

[0078] Specifically, the historical maximum quality impact coefficient refers to the maximum value among the quality impact coefficients generated by the fluctuations of all process parameters in historical data. This value is used to measure the degree of influence of the most significant process parameter fluctuations on quality in history. The first process parameter refers to a specific process parameter randomly selected during the process parameter optimization. For example, injection temperature, injection pressure, cooling time, etc. The first quality impact coefficient refers to the initial value of the quality impact coefficient under the currently selected process parameter. This value is calculated based on historical data and reflects the degree of influence of the fluctuation of this process parameter on product quality. The number of selected units in the first unit (Q) is calculated and used to determine the number of product quality prediction units to be selected from multiple prediction units. This number is obtained by multiplying the ratio of the first quality impact coefficient to the historical maximum quality impact coefficient by P and rounding up. This number is used to determine how many product quality prediction units to use in the prediction process. The product quality prediction unit is an independent module in the product quality prediction plug-in and is used to predict product quality based on the given process parameters. Each unit provides a prediction result of product quality by inputting process parameters.

[0079] First, randomly select a first process parameter and a first quality impact coefficient from multiple quality impact coefficients and process parameters. Suppose we select the injection temperature as 250°C, the injection pressure as 800 bar, etc. as the first process parameter, and obtain its corresponding first quality impact coefficient (such as a specific coefficient reflecting the influence degree of temperature on warpage).

[0080] Next, multiply the ratio of the first quality impact coefficient to the historical maximum quality impact coefficient by P (P is a constant, usually greater than 1), and then round it to obtain the quantity (Q) of the first unit selected. The calculation formula of Q is:

[0081]

[0082] The function of this step is to determine how many product quality prediction units need to be selected according to the ratio of the current quality impact coefficient to the historical maximum value.

[0083] For example, if the first quality impact coefficient we selected is 0.05, and the historical maximum quality impact coefficient is 0.1, assuming P is 10, then the calculation of Q is:

[0084]

[0085] Therefore, 5 prediction units need to be randomly selected from the product quality prediction plug-in for quality prediction. Then, apply the Q product quality prediction units to the selected first process parameter, and predict the product quality through these units. For example, if we select the temperature of 250°C and the pressure of 800 bar, after the prediction of 5 different prediction units, 5 different prediction results are obtained.

[0086] Finally, take the average of these prediction results to obtain a final first predicted product quality data. For example, if the 5 prediction results are ±0.12mm, ±0.15mm, ±0.10mm, ±0.11mm, ±0.13mm respectively, then calculate their average value as 0.12mm. This average value will be used as the first predicted product quality data and added to the existing multiple predicted product quality data for further optimizing the process parameters.

[0087] Through this step, with the help of multiple prediction units, the error caused by a single prediction result can be reduced, and the accuracy and reliability of the prediction result can be improved. By calculating the average value of multiple prediction results, the product quality under the current process conditions can be more accurately reflected, providing more accurate data support for the subsequent optimization of process parameters.

[0088] Furthermore, according to the multiple corrected product demand characteristics and multiple predicted product quality data, conduct optimization of process parameters and output the optimal process parameters, including:

[0089] Perform deviation analysis on the multiple corrected product requirement characteristics and the multiple predicted product quality data respectively, and determine multiple process fitness levels according to the multiple deviation values. Among them, the smaller the deviation value, the higher the process fitness level.

[0090] Arrange the multiple process parameters in descending order of fitness to generate a process parameter sequence.

[0091] Eliminate 15% of the process parameters at the end of the process parameter sequence and randomly supplement them using the processing parameter threshold.

[0092] Perform iterative optimization analysis until a predetermined convergence number is reached, output the current process parameter sequence, and select the process parameter at the forefront of the current process parameter sequence as the optimal process parameter.

[0093] Specifically, the process fitness level measures the matching degree between the process parameters and the product requirements through the deviation value. The smaller the deviation value, the higher the fitness level. That is, the more the process parameters meet the quality requirements of the product, the greater the fitness level. The process parameter sequence refers to a sequence formed by arranging all process parameters in a certain order during the process optimization. For example, multiple process parameter sets formed by different temperature and pressure combinations. The end of the process parameter sequence refers to the worst 15% of the parameter combinations during the process optimization.

[0094] First, the deviation analysis is to analyze the differences between the multiple corrected product requirement characteristics and the multiple predicted product quality data. The smaller the deviation value, the closer the predicted quality is to the requirement characteristics, and the more ideal the process parameters are. According to these deviation values, the process fitness level of each process parameter is calculated. The higher the fitness level of the parameters means that they can better meet the expected product quality requirements. Next, according to the process fitness level, the multiple process parameters are arranged in descending order of fitness to form a process parameter sequence. Then, by eliminating 15% of the process parameters at the end of the process parameter sequence, the parameter combinations with the worst fitness are removed. For example, if there are 10 parameter combinations, during the optimization process, the worst 15% are eliminated, that is, the parameter combination with the lowest fitness is eliminated. The remaining 90% will continue to participate in the optimization and be randomly supplemented in combination with the processing parameter threshold to ensure that the process parameters are still within the acceptable range during the optimization process.

[0095] Next, perform iterative optimization analysis. Through multiple optimization iterations, the process parameters will be adjusted and new fitness levels will be calculated in each round until the predetermined convergence number is reached. That is, after several rounds of iteration, the optimization results of the process parameters tend to be stable and there are no large changes. At this time, output the current process parameter sequence and select the process parameter at the forefront of the sequence as the final optimal process parameter.

[0096] Through this optimization process, the consistency and stability of product quality can be significantly improved, while the production efficiency is enhanced. Deviation analysis helps to identify the differences between process parameters and requirement characteristics, and through fitness optimization, it ensures that each round of optimization moves towards the target quality standard. By eliminating the worst process parameters and supplementing them with machining parameter thresholds, invalid process settings and parameter fluctuations are avoided, ensuring that the process parameters are within the feasible range.

[0097] In summary, the process parameter optimization control method for an automotive injection molding part provided by the embodiment of the present application has the following technical effects:

[0098] 1. By optimizing and controlling process parameters based on historical processing records, it is possible to effectively predict the product quality under different process conditions, reduce human intervention, and improve the stability of the production process. This method can precisely control each process parameter in production through the combination of a feedforward neural network and a fluctuation analyzer, optimize the production process, improve product consistency and quality, and ultimately reduce the scrap rate and enhance the production efficiency.

[0099] 2. The fluctuation impact analyzer can analyze the impact of process parameter fluctuations on product quality, identify the key parameters with large fluctuations, and thus conduct targeted process control. By analyzing the mapping relationship between the fluctuation coefficient and the quality impact coefficient, unnecessary fluctuations can be effectively reduced, ensuring stable product quality and enhancing the reliability and controllability of the production process.

[0100] 3. By optimizing process parameters according to the product requirement characteristics, it can ensure that the process parameters used in the production process best match the product quality requirements. This method can find the most suitable process conditions among multiple parameter combinations through the combination of fluctuation impact analysis and a quality prediction plug-in, thereby improving the quality consistency of the product and optimizing the production efficiency.

[0101] Embodiment 2

[0102] Based on the same inventive concept as the process parameter optimization control method for an automotive injection molding part in the foregoing embodiment, as Figure 3 shown, the embodiment of the present application provides a process parameter optimization control system for an automotive injection molding part, and the system includes:

[0103] The product quality prediction plug-in construction module 11 is used to collect sample data based on the historical processing records of the target injection molded part to train a feedforward neural network and construct a product quality prediction plug-in; the fluctuation impact analyzer construction module 12 is used to analyze the impact of parameter fluctuations according to the historical processing records and construct a fluctuation impact analyzer; the product demand feature determination module 13 is used to determine the product demand features of the target injection molded part in combination with the automotive usage scenario; the optimal process parameter output module 14 is used to use the product quality prediction plug-in and the fluctuation impact analyzer with the product demand features as the expectation to optimize the process parameters and output the optimal process parameters for processing control.

[0104] Further, the product quality prediction plug-in construction module 11 is also used to perform the following steps: based on the historical processing records of the target injection molded part, collect the injection temperature, injection pressure, injection speed, cooling time, and cooling temperature at K consecutive monitoring nodes within the historical processing cycle to obtain multiple sample parameter sequence sets, where K is an integer greater than 10; obtain the product quality indicators under different sample parameter sequence sets to obtain multiple sample quality data sets, where the product quality indicators at least include dimensional accuracy, structural strength, surface quality, molding quality, and material properties; use the multiple sample parameter sequence sets and multiple sample quality data sets to train a feedforward neural network to generate the product quality prediction plug-in.

[0105] Further, the product quality prediction plug-in construction module 11 is also used to perform the following steps: use the multiple sample parameter sequence sets and multiple sample quality data sets as training data and divide them equally into P parts to obtain P training sets, where P is an integer greater than 15; use the P training sets to train a feedforward neural network respectively until convergence, obtain P product quality prediction units, and construct the product quality prediction plug-in according to the combination of the P product quality prediction units.

[0106] Further, the fluctuation impact analyzer construction module 12 is also used to perform the following steps: according to the historical processing records, screen and obtain multiple historical processing record data sets under multiple predetermined identical parameter sequences, where the historical processing record data includes the actual parameter sequence and the product quality data; randomly select the first predetermined identical parameter sequence, as well as the first actual parameter sequence set and the first product quality data set; calculate the parameter fluctuation characteristics of multiple first actual parameter sequences in the first actual parameter sequence set to obtain the first parameter fluctuation coefficient set; based on the expected product quality as the benchmark, perform quality impact analysis according to the first product quality data set and output the first quality impact coefficient set; map and construct the first fluctuation impact analysis branch based on the first predetermined identical parameter sequence, the first parameter fluctuation coefficient set, and the first quality impact coefficient set, and sequentially generate multiple fluctuation impact analysis branches to construct the fluctuation impact analyzer.

[0107] Further, the process parameter optimization control system for an automotive injection molded part is further configured to perform the following steps: calculate the standard deviation and mean value of multiple parameters in multiple first actual parameter sequences respectively to obtain multiple first parameter standard deviation sets and multiple first parameter mean value sets; based on the multiple first parameter standard deviation sets and multiple first parameter mean value sets, perform weighted fusion to obtain multiple first comprehensive parameter standard deviations and multiple first comprehensive parameter mean values, wherein the weight is positively correlated with the influence degree of the parameter on the product quality; set the ratio of the first comprehensive parameter standard deviation to the first comprehensive parameter mean value as the first parameter fluctuation coefficient, and calculate a first parameter fluctuation coefficient set based on the multiple first comprehensive parameter standard deviations and multiple first comprehensive parameter mean values.

[0108] Further, the optimal process parameter output module 14 is further configured to perform the following steps: obtain the processing parameter thresholds of the target injection molded part, randomly select multiple process parameters within the processing parameter thresholds, wherein each process parameter includes multiple parameter sequences; input the multiple process parameters into the fluctuation influence analyzer for similarity matching, and output multiple quality influence coefficients with the highest similarity; with the product demand characteristics as the expectation, use the product quality prediction plug-in to optimize the process parameters according to the multiple quality influence coefficients and multiple process parameters, and output the optimal process parameters.

[0109] Further, the optimal process parameter output module 14 is further configured to perform the following steps: adjust the product demand characteristics respectively according to the multiple quality influence coefficients to obtain multiple corrected product demand characteristics; use the product quality prediction plug-in to predict the product quality according to the multiple quality influence coefficients and multiple process parameters, and output multiple predicted product quality data; perform process parameter optimization according to the multiple corrected product demand characteristics and multiple predicted product quality data, and output the optimal process parameters.

[0110] Further, the optimal process parameter output module 14 is further configured to perform the following steps: randomly select a first process parameter and a first quality influence coefficient; multiply the ratio of the first quality influence coefficient to the historical maximum quality influence coefficient by P and round it to obtain Q, which is set as the number of selected units in the first unit. Randomly select Q product quality prediction units in the product quality prediction plug-in to predict the product quality of the first process parameter, output Q prediction results, calculate the mean value to obtain the first predicted product quality data, and add it to the multiple predicted product quality data.

[0111] Further, the optimal process parameter output module 14 is further configured to perform the following steps: perform deviation analysis on the multiple corrected product demand characteristics and the multiple predicted product quality data respectively, determine multiple process fitnesses according to the multiple deviation values, wherein the smaller the deviation value, the higher the process fitness; arrange the multiple process parameters in descending order of fitness to generate a process parameter sequence; eliminate 15% of the process parameters at the end of the process parameter sequence and randomly supplement them using the processing parameter threshold; perform iterative optimization analysis until a predetermined number of convergences is reached, output the current process parameter sequence, and select the process parameter at the forefront of the current process parameter sequence as the optimal process parameter.

[0112] Any step of the method described above can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor to implement any method in the embodiments of the present application, and no redundant limitations are made here.

[0113] Further, the first or second mentioned above does not only represent an order relationship, but also represents a specific concept, and / or means that multiple elements can be selected individually or in whole. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and variations.

Claims

1. A method for optimizing and controlling process parameters of an automotive injection molding part, characterized in that, The method includes: Based on the historical processing records of the target injection molded part, sample data is collected to train a feedforward neural network, and a product quality prediction plug-in is constructed, including: Based on the historical processing records of the target injection molded part, the injection temperature, injection pressure, injection speed, cooling time, and cooling temperature under K consecutive monitoring nodes within the historical processing cycle are collected to obtain multiple sample parameter sequence sets, where K is an integer greater than 10; The product quality indicators under different sample parameter sequence sets are obtained to obtain multiple sample quality data sets, where the product quality indicators at least include dimensional accuracy, structural strength, surface quality, molding quality, and material properties; The feedforward neural network is trained using the multiple sample parameter sequence sets and multiple sample quality data sets to generate the product quality prediction plug-in; According to the historical processing records, a parameter fluctuation impact analysis is carried out to construct a fluctuation impact analyzer, including: According to the historical processing records, multiple historical processing record data sets under multiple predetermined identical parameter sequences are screened and obtained, where the historical processing record data includes the actual parameter sequence and the product quality data; Randomly select the first predetermined identical parameter sequence, as well as the first actual parameter sequence set and the first product quality data set; Calculate the parameter fluctuation characteristics of multiple first actual parameter sequences in the first actual parameter sequence set to obtain the first parameter fluctuation coefficient set; Based on the expected product quality as a benchmark, perform a quality impact analysis according to the first product quality data set, and output the first quality impact coefficient set; Based on the first predetermined identical parameter sequence, the first parameter fluctuation coefficient set, and the first quality impact coefficient set, map and construct the first fluctuation impact analysis branch, and sequentially generate multiple fluctuation impact analysis branches to construct the fluctuation impact analyzer; Combine the automotive usage scenario to determine the product demand characteristics of the target injection molded part; Taking the product demand characteristics as the expectation, use the product quality prediction plug-in and the fluctuation impact analyzer to optimize the process parameters, and output the optimal process parameters for processing control, including: Obtain the processing parameter threshold of the target injection molded part, and randomly select multiple process parameters within the processing parameter threshold, where each process parameter includes multiple parameter sequences; Input the multiple process parameters into the fluctuation impact analyzer for similarity matching, and output the multiple quality impact coefficients with the highest similarity; Taking the product demand characteristics as the expectation, use the product quality prediction plug-in, and optimize the process parameters according to the multiple quality impact coefficients and multiple process parameters, and output the optimal process parameters.

2. The process parameter optimization control method for an automotive injection molded part according to claim 1, characterized in that, Using the multiple sample parameter sequence sets and multiple sample quality data sets to train a feedforward neural network to generate the product quality prediction plug-in, including: Taking the multiple sample parameter sequence sets and multiple sample quality data sets as training data, and equally dividing them into P parts to obtain P training sets, where P is an integer greater than 15; Use the P training sets to train the feedforward neural network respectively until convergence, obtain P product quality prediction units, and construct the product quality prediction plug-in according to the combination of the P product quality prediction units.

3. The process parameter optimization control method for an automotive injection molding part according to claim 1, characterized in that, Calculate the parameter fluctuation characteristics of multiple first actual parameter sequences to obtain the first parameter fluctuation coefficient set, including: Calculate the standard deviation and mean of multiple parameters in multiple first actual parameter sequences respectively to obtain multiple first parameter standard deviation sets and multiple first parameter mean sets; Based on the multiple first parameter standard deviation sets and multiple first parameter mean sets, weighted fusion is performed to obtain multiple first comprehensive parameter standard deviations and multiple first comprehensive parameter means, where the weight is positively correlated with the influence degree of the parameter on the product quality; Set the ratio of the first comprehensive parameter standard deviation to the first comprehensive parameter mean as the first parameter fluctuation coefficient, and calculate the first parameter fluctuation coefficient set according to the multiple first comprehensive parameter standard deviations and multiple first comprehensive parameter means.

4. The process parameter optimization control method for an automotive injection molding part according to claim 1, characterized in that, Taking the product demand characteristics as the expectation, using the product quality prediction plug-in, process parameter optimization is carried out according to the multiple quality influence coefficients and multiple process parameters, including: Adjust the product demand characteristics respectively according to the multiple quality influence coefficients to obtain multiple corrected product demand characteristics; Using the product quality prediction plug-in, product quality prediction is carried out according to the multiple quality influence coefficients and multiple process parameters, and multiple predicted product quality data are output; According to the multiple corrected product demand characteristics and multiple predicted product quality data, process parameter optimization is carried out, and the optimal process parameters are output.

5. The process parameter optimization control method for an automotive injection molded part according to claim 4, characterized in that, Using the product quality prediction plug-in, product quality prediction is carried out according to the multiple quality influence coefficients and multiple process parameters, and multiple predicted product quality data are output, including: Randomly select the first process parameter and the first quality influence coefficient; Multiply the ratio of the first quality influence coefficient to the historical maximum quality influence coefficient by P and round it to obtain Q, and set it as the first unit selection quantity, Randomly select Q product quality prediction units in the product quality prediction plug-in, carry out product quality prediction on the first process parameter, output Q prediction results, and calculate the mean to obtain the first predicted product quality data, and add it to the multiple predicted product quality data.

6. The process parameter optimization control method for an automotive injection molded part according to claim 4, wherein, According to the multiple corrected product demand characteristics and multiple predicted product quality data, process parameter optimization is carried out, and the optimal process parameters are output, including: Perform deviation analysis on the multiple corrected product demand characteristics and multiple predicted product quality data respectively, and determine multiple process fitnesses according to multiple deviation values, where the smaller the deviation value, the higher the process fitness; Arrange multiple process parameters in descending order of fitness to generate a process parameter sequence; Eliminate 15% of the process parameters at the end of the process parameter sequence, and perform random supplementation using the processing parameter threshold; Perform iterative optimization analysis until the predetermined convergence number is reached, output the current process parameter sequence, and select the process parameter at the forefront of the current process parameter sequence as the optimal process parameter.

7. A process parameter optimization control system for automotive injection molded parts, characterized in that, For implementing the process parameter optimization control method of an automotive injection molding part according to any one of claims 1 to 6, the system includes: A product quality prediction plug-in construction module for collecting sample data to train a feedforward neural network based on the historical processing records of the target injection molding part and constructing a product quality prediction plug-in; A fluctuation influence analyzer construction module for performing parameter fluctuation influence analysis according to the historical processing records and constructing a fluctuation influence analyzer; A product requirement feature determination module, which is used to determine the product requirement features of the target injection molded part in combination with the vehicle usage scenarios; An optimal process parameter output module, which is used to take the product requirement features as the expectation, utilize the product quality prediction plug-in and the fluctuation impact analyzer to optimize the process parameters, and output the optimal process parameters for processing control.

Images