An injection molding machine processing parameter optimization method, server, medium and product

Abstract

The application provides an injection molding machine processing parameter optimization method, a server, a medium and a product, and relates to the field of parameter optimization. The method first collects core process parameters such as injection pressure, barrel temperature and mold temperature in real time, and associates with equipment number data; inputs these data into a product defect prediction model trained by deep learning to obtain a predicted yield rate; based on the prediction result, an optimization adjustment strategy is formulated and pushed to the operation end, and after responding to the instruction, it is issued to the injection molding machine for execution. At the same time, the actual yield of a certain number of products is obtained through quality monitoring as feedback, and the optimization strategy is iteratively adjusted accordingly to form a closed-loop control, effectively reducing the subjectivity of parameter setting, improving parameter matching degree, and thus stabilizing the yield rate of the injection molding machine, and improving the accuracy and stability of process control.

Description

Technical Field

[0001] This application relates to the field of parameter optimization technology, and in particular to a method, server, medium and product for optimizing injection molding machine processing parameters. Background Technology

[0002] Injection molding, as a core process for molding plastic products, has been widely used in industrial manufacturing and consumer goods production. In actual production, multiple injection molding machines are often used to process products in parallel. The product yield is highly dependent on the matching degree of process parameters such as injection pressure, barrel temperature, and mold temperature, and is also affected by factors such as equipment operating conditions and environmental conditions.

[0003] Currently, injection molding generally adopts a single-machine independent control mode. Process parameters are mainly set and adjusted based on the operator's experience and operation manual. Product quality is generally based on the yield rate. In the end, some injection molding machines produce high yield rates, while others produce low yield rates. The yield rate of injection molded products is highly dependent on the worker's experience and familiarity with the equipment.

[0004] However, with the scaling up of production and the increasing demands for product precision, the drawbacks of relying solely on manual experience for parameter adjustment have become increasingly prominent. Differences in the experience levels and familiarity with the equipment among different operators easily lead to strong subjectivity and poor consistency in parameter settings. Therefore, how to stably control product quality and reduce reliance on manual experience has become a common need in the injection molding industry. Summary of the Invention

[0005] This application provides a method, server, medium, and product for optimizing injection molding machine processing parameters, which can reduce reliance on manual experience and steadily improve the yield of injection molded products.

[0006] Firstly, this application provides a method for optimizing injection molding machine processing parameters. The method includes: real-time acquisition of process parameters of a target injection molding machine, including at least injection pressure, barrel temperature, and mold temperature; obtaining the serial number data of the target injection molding machine and inputting the process parameters and the serial number data into a product defect prediction model to obtain the predicted yield rate of the target injection molding machine. The product defect prediction model is trained in advance using deep learning based on multiple historical process parameter sample sets labeled with actual yield rates; determining an optimization adjustment strategy based on the predicted yield rate and sending it to an operating terminal; responding to a process parameter adjustment command from the operating terminal, sending the optimization adjustment strategy to the target injection molding machine for execution; real-time monitoring of quality feedback after executing the optimization adjustment strategy, where the quality feedback refers to the actual yield rate of a set number of products; and adjusting the optimization adjustment strategy based on the quality feedback.

[0007] By adopting the above technical solution, core process parameters such as injection pressure and barrel temperature are collected in real time. Combined with the injection molding machine serial number, precise equipment association is achieved, avoiding parameter confusion between different machines. The parameters and serial number are input into a product defect prediction model trained by deep learning. This model, based on a large number of historical samples labeled with actual yield rates, can accurately output predicted yield rates, providing a scientific basis for parameter optimization and replacing traditional manual experience-based judgment. Optimization strategies are formulated and implemented based on the prediction results. Simultaneously, the strategies are dynamically adjusted through real-time quality feedback, forming a closed-loop control of "collection-prediction-optimization-execution-feedback-readjustment." This effectively reduces the subjectivity of parameter setting, improves parameter matching, thereby stabilizing the injection molding machine's production yield rate, reducing product defects caused by unreasonable parameters, and improving the accuracy and stability of process control.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, adjusting the optimization strategy based on the quality feedback includes: if there are non-conforming products in the quality feedback, collecting defect performance characteristics of multiple non-conforming products within the same processing batch; comparing multiple defect performance characteristics, extracting common defect characteristics as target defects; and tracing back the main process parameters that caused the target defect based on a pre-established defect and process parameter knowledge graph, and determining the main process parameters as preliminary parameters to be adjusted.

[0009] By adopting the above technical solution, when quality feedback indicates non-conforming products, the defect characteristics of multiple non-conforming products from the same batch are collected and compared. Common defects are extracted as target defects, accurately pinpointing the core quality issues in current production and avoiding misjudgments due to the specificity of individual defect samples. Combined with a pre-established defect-process parameter knowledge graph, the main process parameters leading to the target defect are traced back to clarify the initial direction for adjustment, replacing the traditional method of blindly adjusting parameters manually. This targeted tracing can quickly locate the root cause of the problem, reduce the number of ineffective parameter adjustments, shorten the parameter adjustment cycle, and ensure the accuracy of the adjustment direction.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, after determining the main process parameter as the preliminary parameter to be adjusted, the method further includes: obtaining actual test data of the injection molding raw material corresponding to the current processing batch, the actual test data including at least the raw material moisture content and melt index; determining whether the deviation of the actual test data from the preset standard material parameter exceeds the allowable range; if the deviation exceeds the allowable range, determining that the current target defect is affected by the material's own abnormality, and generating a material abnormality warning; if the deviation does not exceed the allowable range, confirming that the target defect is caused by the process parameter, and making trial adjustments based on the preliminary parameter to be adjusted.

[0011] By adopting the above technical solution, after determining the initial parameters to be adjusted, actual test data of the injection molding raw materials are further obtained, focusing on the two key material indicators affecting injection molding: moisture content and melt flow index. By comparing the deviation between the actual test data and the preset standard material parameters, it is possible to determine whether there are any abnormalities in the material. This effectively distinguishes whether the defects are caused by process parameters or by abnormalities in the material itself, avoiding misjudging defects caused by material abnormalities as parameter problems, thus avoiding ineffective parameter adjustments and saving adjustment costs and time. If the material is abnormal, an early warning is generated in a timely manner to remind staff to handle the material problem; if the material is normal, the parameters are confirmed as the root cause of the problem, ensuring that subsequent parameter adjustments are targeted, further improving the accuracy of parameter optimization, and guaranteeing the stability of production quality.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, the step of making trial adjustments based on the preliminary parameters to be adjusted specifically includes: extracting multiple preliminary parameters to be adjusted that cause the target defect; determining associated process parameters related to each of the preliminary parameters to be adjusted based on a defect and process parameter knowledge graph, wherein the associated process parameters refer to parameters that have a constraint relationship with the preliminary parameters to be adjusted; generating trial data for each of the preliminary parameters to be adjusted and corresponding associated process parameters according to a set fine-tuning step size; and determining a set of trial process parameters containing multiple trial data.

[0013] By employing the aforementioned technical solution, after extracting multiple preliminary adjustable parameters that lead to the target defect, the associated process parameters of each preliminary adjustable parameter are mined based on a knowledge graph. This clarifies the constraints between parameters—that is, other related parameters that may be affected by adjusting a certain parameter, thus avoiding new process imbalances caused by adjusting a single parameter. Trial data is generated according to a set fine-tuning step size, forming a trial process parameter set containing multiple sets of trial data. The fine-tuning mode avoids batch defects caused by parameter mutations, and the multiple sets of trial data cover different parameter ratios, providing sufficient samples for subsequent selection of the optimal parameter. This approach ensures the safety of parameter adjustment while improving the comprehensiveness of optimal parameter selection, effectively enhancing the scientific nature of parameter adjustment.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, after determining the set of trial process parameters containing multiple trial data, the method further includes: sequentially sending each trial data in the set of trial process parameters to the target injection molding machine for small-batch trial production; obtaining the actual product test results of each trial production batch; comparing the actual product test results with the target defect; if there is target trial data that makes the target defect disappear or reduce to the qualified standard, but a new defect appears in the actual product test results, then it is determined that the target trial data has a process conflict, and the target trial data is recorded to the trial process blacklist.

[0015] By adopting the above technical solution, for target trial data that can resolve the target defect but induce new defects, process conflicts are identified and recorded in the trial process blacklist. This effectively avoids the subsequent use of such parameters, prevents the occurrence of new defects, and reduces ineffective trial production and product loss. Simultaneously, the establishment of the blacklist accumulates experience in process conflicts, providing a basis for avoidance in subsequent parameter optimization, preventing repeated pitfalls, improving the efficiency of parameter optimization, ensuring the orderliness and effectiveness of the trial production process, and further improving the quality of the final parameter optimization.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, after comparing the actual product inspection results with the target defect, the method further includes: screening out trial data that did not trigger process conflicts and whose target defect disappeared or decreased to the qualified standard, as the final optimization adjustment strategy; if there are multiple final optimization adjustment strategies, selecting the target optimization adjustment strategy with the highest actual yield rate; issuing the target optimization adjustment strategy to the target injection molding machine for formal continuous production; setting a stable observation period during the formal continuous production process, and continuously monitoring the real-time yield rate during the stable observation period to verify the effectiveness of the trial adjustment.

[0017] By adopting the above technical solution, after the optimal strategy is issued to formal production, a stable observation period is set and the real-time yield rate is continuously monitored. This can verify the effectiveness of the trial adjustment, promptly detect fluctuations in parameters during continuous production, avoid quality decline caused by parameter deviations due to long-term production, further ensure the stability and reliability of parameter optimization, and ensure that the quality of mass-produced products meets the standards and remains stable.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, after the step of recording the target trial data to the trial process blacklist, the method further includes: extracting the target trial data and the corresponding new defect recorded in the trial process blacklist; parsing the parameter ratio characteristics in the target trial data and establishing a conflict mapping relationship between the parameter ratio characteristics and the new defect; based on the conflict mapping relationship, adding an exclusion constraint rule for the target injection molding machine in the defect and process parameter knowledge graph; when the step of generating preliminary trial data is triggered again, calling the exclusion constraint rule to pre-filter the preliminary trial data to intercept trial data that causes process conflict.

[0019] By adopting the above technical solution, target trial data with process conflicts and corresponding emerging defects are blacklisted. Parameter ratio characteristics are extracted, and a conflict mapping relationship is established between these parameters and emerging defects, clearly defining the correlation between parameter ratios and defects. Based on this mapping relationship, exclusion constraint rules are added to the knowledge graph, enabling dynamic updates and improvements to the knowledge graph and enriching the experience reserves for process optimization. When generating preliminary trial data, this constraint rule is invoked for pre-filtering, directly intercepting trial data that may lead to process conflicts, reducing the number of invalid trial runs, shortening the parameter optimization cycle, and preventing the generation of emerging defects.

[0020] In conjunction with some embodiments of the first aspect, in some embodiments, the server includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory being used to store computer program code including computer instructions, the one or more processors invoking the computer instructions to cause the server to perform the methods described in the first aspect and any possible implementation thereof.

[0021] In a second aspect, this application provides a computer-readable storage medium including instructions that, when executed on a server, cause the server to perform the method described in the first aspect and any possible implementation thereof.

[0022] Thirdly, this application provides a computer program product, including a computer program that, when run on a server, causes the server to perform the method described in the first aspect and any possible implementation thereof.

[0023] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0024] 1. By employing a technical approach that involves real-time acquisition of core process parameters, combined with the injection molding machine serial number, inputting the predicted yield rate into a product defect prediction model trained by deep learning, and then formulating optimization strategies based on the prediction results and dynamically adjusting them through quality feedback to form a closed-loop control, this approach effectively solves the technical problems in existing technologies where injection molding machine process parameters rely on manual experience setting, are highly subjective, and have poor consistency, leading to unstable yield rates. This achieves scientific and accurate process parameter setting, stabilizes the injection molding machine's production yield rate, reduces product defects, and improves the stability of process control.

[0025] 2. By employing a technical approach that involves collecting actual test data of injection molding raw materials after determining the initial parameters to be adjusted, comparing the deviations with preset standard material parameters, distinguishing whether defects are caused by process parameters or material abnormalities, and generating warnings when material abnormalities occur and confirming parameters as the root cause when materials are normal, this technology effectively solves the technical problem in existing technologies where defects caused by material abnormalities are easily misjudged as process parameter problems, leading to ineffective parameter adjustments, wasted costs and time. This achieves precise and targeted parameter adjustments, saves adjustment costs and time, promptly handles material abnormalities, and ensures the stability of production quality.

[0026] 3. By employing a technical approach that involves sequentially conducting small-batch trial production using trial data from a set of trial process parameters, comparing the trial production test results with the target defects, and identifying trial data that resolves the target defects but induces new defects as process conflicts and recording them in the trial process blacklist, this approach effectively solves the technical problem in existing technologies where parameters with process conflicts are easily used during trial production, leading to new defects, increased product losses, and ineffective trial production. This achieves the technical effects of avoiding the generation of new defects, reducing product losses and ineffective trial production, accumulating experience in process conflicts, improving parameter optimization efficiency, and ensuring the orderly and effective nature of trial production. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of a scenario for the injection molding machine processing parameter optimization method in the embodiments of this application;

[0028] Figure 2 This is a flowchart illustrating a method for optimizing injection molding machine processing parameters in an embodiment of this application;

[0029] Figure 3 This is another flowchart illustrating the method for optimizing injection molding machine processing parameters in the embodiments of this application;

[0030] Figure 4 This is a schematic diagram of the physical device structure of a server in an embodiment of this application. Detailed Implementation

[0031] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0032] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0033] To facilitate understanding, the method provided in this implementation is described in a scenario below. Please refer to [link / reference]. Figure 1 This is a schematic diagram of a scenario for the injection molding machine processing parameter optimization method in the embodiments of this application.

[0034] exist Figure 1 The paper demonstrates a complete closed-loop application scenario of the injection molding machine processing parameter optimization method in the embodiments of this application. It mainly consists of four core parts working together: the target injection molding machine, the data processing and strategy center, the operation terminal, and the quality monitoring module, to achieve automated optimization of process parameters and stable control of production quality.

[0035] As the main production unit, the target injection molding machine collects core process parameters (injection pressure, barrel temperature, mold temperature, etc.) in real time during operation, along with its own unique serial number. This data is then synchronously transmitted to the data processing and strategy center to achieve precise binding between the equipment and the data.

[0036] As the core of optimization decision-making, the data processing and strategy center first inputs the received real-time process parameters and equipment number into a product defect prediction model that has been pre-trained by deep learning. The model analyzes the current production status based on historical sample data and outputs the predicted yield rate of the target injection molding machine. Subsequently, based on the predicted yield rate and combined with preset quality standards, the center automatically generates corresponding optimization adjustment strategies, clarifying the direction and magnitude of parameter adjustments.

[0037] The generated optimization and adjustment strategies will be pushed to the operator's terminal for viewing, review, and confirmation. After the operator sends instructions through the terminal, the strategy center will send the optimization and adjustment strategies back to the target injection molding machine, driving the injection molding machine to automatically adjust process parameters and execute the new production plan.

[0038] After the target injection molding machine implements the optimization and adjustment strategy, the produced products are sent to the quality monitoring module, which detects the actual quality of the products and obtains quality feedback data.

[0039] The quality monitoring module transmits the detected quality feedback data back to the data processing and strategy center. Based on the actual product quality, the center iteratively modifies the original optimization and adjustment strategy to form a closed-loop control process, continuously improving the yield and stability of injection molding machine production.

[0040] The following describes the process of the method provided in this implementation, using the above scenario as an example. Please refer to [link / reference]. Figure 2 This is a flowchart illustrating a method for optimizing injection molding machine processing parameters in an embodiment of this application.

[0041] S101. Real-time acquisition of the process parameters of the target injection molding machine, including at least the injection pressure, barrel temperature and mold temperature;

[0042] The target injection molding machine refers to the specific injection molding machine for which the server needs to optimize process parameters, implement monitoring and control, and is used to represent the specific equipment object to be optimized.

[0043] Specifically, this step is executed after the injection molding machine starts processing and throughout the entire production process. The execution scenario is the initial data acquisition stage when the server performs real-time monitoring and parameter optimization of the target injection molding machine. It is applicable to all injection molding production scenarios that need to improve yield through parameter optimization. This step must be executed regardless of whether the target injection molding machine is in the mass production, trial production, or process debugging stage.

[0044] The server establishes a communication connection with the target injection molding machine to collect process parameters during the machine's operation in real time. The collection frequency can be set according to the production accuracy requirements to ensure that the collected parameters accurately reflect the real-time operating status of the injection molding machine. Among these parameters, the process parameters include at least three core parameters: injection pressure, barrel temperature, and mold temperature. Additional relevant process parameters such as holding pressure, holding time, and injection speed can also be collected based on product type and raw material characteristics.

[0045] S102. Obtain the serial number data of the target injection molding machine, and input the process parameters and the serial number data into the product defect prediction model to obtain the current predicted yield rate of the target injection molding machine. The product defect prediction model is trained in advance by deep learning based on multiple historical process parameter sample sets with actual yield rate labels.

[0046] The historical process parameter sample set refers to the set of process parameters collected in advance by the server from different production batches of each injection molding machine. Each sample is labeled with the actual yield rate of the corresponding batch, which is used to represent the basic data for model training. For example, the sample set may contain sample data of injection molding machine M001 with an actual yield rate of 98% when the injection pressure is 90MPa, the barrel temperature is 220℃, and the mold temperature is 85℃.

[0047] Specifically, this step is executed after the server completes the real-time acquisition of process parameters in step S101. This step is required every time a complete set of process parameters is acquired. The execution scenario is when the server makes a prediction on the current production quality of the target injection molding machine.

[0048] The server first obtains the unique serial number of the current target injection molding machine through the preset equipment information management module, ensuring that the serial number corresponds one-to-one with the collected process parameters and avoiding confusion of parameters from different injection molding machines. Then, the server associates and integrates the real-time process parameters (injection pressure, barrel temperature, mold temperature, etc.) collected in step S101 with the serial number data to form a complete set of input data, which is then input into the pre-trained product defect prediction model.

[0049] The training process of the product defect prediction model is pre-processed by the server: the server collects historical process parameters from multiple injection molding machines (covering different equipment numbers and production batches), labels each historical process parameter sample with the actual yield rate of the corresponding batch, forming a large set of labeled historical samples; deep learning algorithms (such as convolutional neural networks and recurrent neural networks) are used to train the sample set, mining the inherent correlation between process parameters, equipment numbers, and actual yield rates. After training until the model's prediction error is lower than a preset threshold (such as 5%), the model training is completed and put into use. After receiving input data, the model calculates and outputs the predicted yield rate of the current target injection molding machine under the set of process parameters through the correlation formed by internal training.

[0050] This step solves the problem in existing technologies that make it difficult to predict production quality in advance and that yield rates can only be obtained through inspection after production is completed, leading to a large number of defective products and increased losses. By predicting the yield rate in advance, the server can promptly identify the deficiencies in the current process parameters, providing a clear basis for subsequent parameter optimization, reducing the generation of defective products, reducing production losses, and replacing manual experience judgment to improve the accuracy and efficiency of quality prediction.

[0051] S103. Based on the predicted yield rate, determine the optimization and adjustment strategy and send it to the operation terminal;

[0052] The server first presets a yield rate threshold (this threshold can be flexibly adjusted according to product precision requirements and production needs, such as setting the threshold to 95% for ordinary plastic products and 99% for precision plastic products); then it compares the predicted yield rate obtained in step S102 with this threshold: if the predicted yield rate reaches or exceeds the threshold, the server determines that the current process parameters do not need to be adjusted, and only sends the predicted yield rate to the operator so that the operator can monitor the expected production quality in real time; if the predicted yield rate is lower than the threshold, the server, based on the correlation between process parameters and yield rate mined during the product defect prediction model training process, combined with the target injection molding machine's serial number data (to distinguish the characteristic differences of different equipment), formulates a targeted optimization and adjustment strategy—the adjustment strategy must clearly define the process parameters to be adjusted, the specific adjustment direction (increase or decrease), and the adjustment range (e.g., adjusting the injection pressure by 5MPa each time, and the temperature by 5℃ each time), while avoiding excessive parameter adjustment ranges that could lead to production abnormalities. Once the optimization and adjustment strategy is formulated, the server sends the strategy to the operator through the communication module. After receiving it, the operator displays the strategy to the operator in a clear and visible form, such as text and charts, for the operator to view and review, ensuring that the operator understands the specific content of the parameter adjustment.

[0053] This step provides operators with clear adjustment guidance, reduces reliance on manual experience, and improves the pertinence and efficiency of parameter adjustments. At the same time, by synchronizing to the operating terminal, it enables operators to supervise and confirm the optimization process, ensuring the safety and rationality of the adjustment process.

[0054] S104. In response to the process parameter adjustment command from the operating terminal, the optimization adjustment strategy is sent to the target injection molding machine for execution;

[0055] This step is executed immediately after the server sends the optimization and adjustment strategy to the operator and the operator issues a process parameter adjustment command through the operator. The execution scenario is when the server applies the optimization strategy to the target injection molding machine and realizes the parameter adjustment. It is suitable for scenarios where the operator confirms that the optimization and adjustment strategy is correct and agrees to execute the adjustment. If the operator finds that the optimization and adjustment strategy is unreasonable (such as the adjustment range is too large or the adjustment parameters are inappropriate), the operator can issue a modification command through the operator. After the server receives the command, it will formulate a new optimization and adjustment strategy and repeat steps S103 and S104.

[0056] The server continuously listens for commands from the operator. Upon receiving a process parameter adjustment command from the operator, it immediately responds by converting the optimized adjustment strategy into a format recognizable control signal for the target injection molding machine's control system, ensuring the machine can accurately interpret the adjustment requirements. Subsequently, the server, through its communication connection with the target injection molding machine, sends the converted control signal (i.e., the optimized adjustment strategy) to the machine's control system. Upon receiving the signal, the target injection molding machine's control system gradually adjusts the corresponding process parameters according to the strategy requirements. For example, if the strategy requires adjusting the injection pressure from 90MPa to 95MPa, the control system controls the injection system's pressure valve to slowly increase the pressure to 95MPa and stabilize it. During the adjustment process, the target injection molding machine provides real-time feedback on the parameter adjustment progress to the server, which simultaneously displays the progress on the operator's terminal for real-time monitoring.

[0057] This step enables automated and precise parameter adjustment by having the server respond to operator commands and remotely distribute optimization strategies. This reduces the workload and errors of manual on-site operations, improves the efficiency and accuracy of parameter adjustment, and allows operators to remotely monitor the adjustment process, ensuring its smooth progress.

[0058] S105. Real-time monitoring of quality feedback after implementing the optimization and adjustment strategy, where the quality feedback refers to the actual good product status of the set quantity of products.

[0059] This step is performed when the server verifies the optimization and adjustment effects in real time and collects feedback data. It is applicable to the effect detection after the execution of all optimization and adjustment strategies, ensuring that every parameter adjustment can obtain effective feedback.

[0060] The server first presets a quantity for detecting optimization effects. This quantity can be flexibly adjusted according to product characteristics and production efficiency. It also presets product quality inspection standards (such as dimensional error range, flawless appearance, and performance compliance). Subsequently, the server establishes a communication connection with the target injection molding machine's associated inspection equipment (such as vision inspection equipment and dimensional inspection equipment), receiving real-time inspection data from the equipment on the products produced by the target injection molding machine. The inspection equipment must inspect each product, distinguishing between qualified and unqualified products, and recording the defect types of unqualified products (such as insufficient material, shrinkage, and burrs). When the number of inspected products reaches the preset quantity, the server performs statistical analysis on the inspection data, generating quality feedback: counting the number of qualified and unqualified products, calculating the actual yield rate, and summarizing the defect types and quantities of unqualified products to identify current quality problems in production. During the monitoring process, the server sends real-time quality feedback data (such as the number of inspected products, current yield rate, and unqualified defect types) to the operator, allowing operators to monitor the optimization effects in real time.

[0061] In this step, by monitoring the quality feedback in real time, the actual effect of the optimization and adjustment strategy can be grasped in a timely manner, providing a true and accurate basis for the further adjustment of subsequent strategies, ensuring that the adjustment effect meets the expectations, reducing ineffective adjustments, and lowering production losses.

[0062] S106. Adjust the optimization and adjustment strategy according to the quality feedback.

[0063] The server first analyzes the quality feedback obtained in step S105 to determine whether there are unqualified products. If not, the execution of this step is terminated.

[0064] If there are unqualified products, the defect feature collection process is immediately started. The server establishes a communication connection with the detection equipment (such as vision detection equipment and manual detection terminals) supporting the target injection molding machine to collect the defect manifestation features of multiple unqualified products within the same processing batch. The number of collected samples should meet the statistical requirements to ensure that the collected defect features are representative and avoid feature deviation caused by过少采集数量. During the collection process, the server records the defect manifestations of each unqualified product in detail, including information such as defect type, defect location, and defect severity. For example, record specific features such as "material shortage at the top of the product, shortage area is about 2mm²" and "burr at the edge of the product, burr thickness is about 0.1mm". Then the server compares and analyzes the collected multiple defect manifestation features. Through the feature extraction algorithm, the defect manifestations common to all (or the vast majority) of the unqualified products are screened out and determined as the target defects. If the defect manifestations of multiple unqualified products vary greatly and there are no obvious common features, the server takes all defect manifestations as the target defects, or further increases the number of collected samples and re - compares and extracts to ensure that the target defects can accurately reflect the main quality problems of the current batch.

[0065] Finally, the server calls the pre - established knowledge graph of defects and process parameters and conducts reverse tracing based on the determined target defects. The server uses the target defects as retrieval keywords to search for all associated process parameters in the knowledge graph, and then through weight analysis (according to the influence degree of each parameter on this defect in historical data), screens out the main process parameters that play a dominant role in the generation of the target defects and determines these main process parameters as the preliminary parameters to be adjusted. For example, if the target defect is "shrinkage", the associated process parameters in the knowledge graph include mold temperature, holding pressure, and holding time. The server analyzes historical data and determines that the too - low mold temperature is the most important factor causing shrinkage, and the insufficient holding time is a secondary factor. Then the mold temperature and holding time are determined as the preliminary parameters to be adjusted.

[0066] In this embodiment, by employing a closed-loop technology that involves real-time acquisition of core process parameters of the target injection molding machine, obtaining equipment number data, and inputting it into a product defect prediction model trained by deep learning to obtain a predicted yield rate, formulating optimization and adjustment strategies based on the predicted yield rate and sending them to the operating terminal, and responding to the operating terminal's adjustment instructions by sending the strategies to the target injection molding machine for execution, and monitoring the quality feedback after execution in real time, and then continuously iterating and adjusting the optimization and adjustment strategies based on the quality feedback, the process parameter optimization is automated, precise, and dynamic. This effectively solves the problem of the dependence on manual experience and strong subjectivity in the setting of injection molding machine process parameters in the prior art, thereby reducing the dependence on manual experience, improving the scientificity and pertinence of process parameter setting and adjustment, stabilizing and improving the yield rate of injection molded products, reducing the generation of defective products, reducing production losses, and improving the stability and production efficiency of injection molding.

[0067] In some embodiments, during the process of determining the initial parameters to be adjusted based on the reverse tracing of the target defect, it may occur that the target defect is not caused by abnormal process parameters, but by the characteristics of the injection molding raw material itself not meeting production requirements. In this case, steps S201-S204 can be performed:

[0068] Following the above embodiments, the method provided in this embodiment will now be described in more detail. Please refer to [link / reference]. Figure 3 This is another flowchart illustrating the method for optimizing injection molding machine processing parameters in this application.

[0069] S201. Obtain the actual test data of the injection molding raw material corresponding to the current processing batch. The actual test data shall include at least the moisture content and melt flow index of the raw material.

[0070] Actual test data refers to the specific characteristic data obtained after testing the injection molding raw materials of the current batch. It is used to reflect the actual quality state of the raw materials and to provide a basis for judging whether the raw materials are abnormal. The moisture content of the raw materials refers to the proportion of the mass of water contained in the injection molding raw materials to the total mass of the raw materials. It is used to indicate the degree of dryness of the raw materials. The melt flow index refers to the mass of the injection molding raw materials passing through a standard capillary tube per unit time under specified temperature and pressure. It is used to indicate the melt flowability of the raw materials.

[0071] The server first obtains basic information about the injection molding raw materials of the current batch through the raw material management module (such as raw material model and batch number), and then establishes a communication connection with the raw material testing equipment (such as moisture content meter and melt flow indexer) to obtain the actual test data of the raw materials of the current batch in real time. The testing equipment needs to sample and test the raw materials of the current batch. The number of samples needs to meet the testing standards (such as taking 3-5 samples) to ensure the representativeness of the test data. The server takes the average value of the test data of multiple samples as the final actual test data.

[0072] S202. Determine whether the deviation between the actual test data and the preset standard material parameters exceeds the allowable range;

[0073] The allowable range refers to the range of acceptable deviations pre-set by the server. It is used to determine whether the deviation of the actual test data of the raw materials is within a reasonable range, so as to avoid misjudging the raw materials as abnormal due to minor deviations. The allowable range can be flexibly adjusted according to the characteristics of the raw materials and the precision requirements of the products.

[0074] The server first retrieves preset standard material parameters from its own storage module that perfectly match the model of the injection molding raw material in the current batch, ensuring the accuracy of the standard parameters. Then, it calculates the deviations of each indicator (raw material moisture content, melt flow index) in the actual test data from the corresponding standard parameters. For example, the deviation between an actual moisture content of 0.08% and a standard moisture content of 0.05% is 0.03%, and the deviation between an actual melt flow index of 7g / 10min and a standard melt flow index of 8-12g / 10min is -1g / 10min. Finally, the calculated deviations of each indicator are compared with their corresponding allowable ranges to determine whether the deviation of each indicator exceeds the allowable range. If any indicator deviation exceeds the allowable range, the overall deviation between the actual test data and the standard material parameters is determined to be outside the allowable range. If all indicator deviations are within the allowable range, it is determined that the deviation is within the allowable range.

[0075] S203. If the deviation exceeds the allowable range, it is determined that the current target defect is affected by the material's own abnormality, and a material abnormality warning is generated.

[0076] This step is executed immediately after the server determines, through step S202, that the deviation between the actual test data of the raw material and the standard material parameters exceeds the allowable range. The server first clarifies, based on the judgment result of step S202, that the current raw material has an inherent anomaly. Then, combining this with the previously identified target defect, it determines that the cause of the target defect is the material's inherent anomaly, rather than an abnormality in the process parameters—because the characteristics of the raw material directly affect the plastic melting, filling, and molding process. If the raw material has problems such as excessive moisture content or abnormal melt flow index, even if the process parameters are reasonable, product defects will still occur. Finally, the server generates a material anomaly warning. The warning information must clearly indicate the batch of the abnormal raw material, the abnormal indicators (such as moisture content and melt flow index), the actual test value, the standard value, and the allowable range. Simultaneously, the warning information is sent to the operating terminal and the raw material management terminal, reminding operators to promptly handle the abnormal raw material (such as drying the raw material or replacing it with qualified raw material), and to suspend production on the target injection molding machine to avoid using abnormal raw materials to produce more defective products.

[0077] S204. If the deviation does not exceed the allowable range, it is confirmed that the target defect is caused by the process parameters, and trial adjustments are made based on the initial parameters to be adjusted.

[0078] This step is executed immediately after the server determines, through step S202, that the deviation between the actual raw material test data and the standard material parameters is within the allowable range. The server first, based on the judgment result of step S202, eliminates the influence of material anomalies on the target defect, clarifying that the current target defect is caused by abnormal process parameters.

[0079] Subsequently, the previously determined preliminary adjustment parameters are invoked, and the adjustment direction (increase or decrease) of each preliminary adjustment parameter is clarified by combining the defect and process parameter knowledge graph. More specifically, the server first invokes the previously determined preliminary adjustment parameters, extracts and sorts them, and clarifies all preliminary adjustment parameters that cause the target defect. If there are multiple preliminary adjustment parameters, the current value and adjustment direction of each parameter (determined based on the defect and process parameter knowledge graph, such as increasing the mold temperature if the target defect is shrinkage) must be recorded one by one to ensure that no core adjustment parameter is missed. If there is only one preliminary adjustment parameter, this step still needs to be followed to avoid ignoring the related process parameters.

[0080] Next, the server calls the pre-established defect and process parameter knowledge graph. For each initially adjustable parameter, it queries the related process parameters that have a constraint relationship with it. During the query, the server will filter out the process parameters that are directly related to the initially adjustable parameter and must be considered simultaneously during adjustment, based on the parameter association rules stored in the knowledge graph, and exclude parameters that are not related. For example, if the initially adjustable parameter is the barrel temperature (melting section), and the related process parameters stored in the knowledge graph are injection speed and holding time, the server will extract these two parameters as related process parameters, and at the same time clarify the constraint relationship between the barrel temperature and these two related parameters (if the barrel temperature is increased, the injection speed can be appropriately increased and the holding time can be appropriately shortened).

[0081] Next, the server calls the pre-set fine-tuning step size (the fine-tuning step size can be flexibly adjusted according to the product precision and process characteristics; for precision products, a smaller step size can be set, such as 3MPa / cycle or 3℃ / cycle), and generates the adjusted value of each initial parameter to be adjusted based on the adjustment direction of each parameter. At the same time, according to the constraints between parameters, the adjusted values ​​of the corresponding related process parameters are generated synchronously. The two are combined to form a set of trial data. If there are multiple initial parameters to be adjusted, the server will generate trial data for each initial parameter to be adjusted and its related process parameters, and ensure that the adjustment range and ratio of the multiple sets of trial data do not overlap, covering different adjustment scenarios.

[0082] Finally, the server will summarize and organize all the generated trial data, remove duplicate and unreasonable trial data (such as trial data with parameter values ​​that exceed the equipment's operating range), and finally determine the trial process parameter set containing multiple trial data. The size of the trial process parameter set can be set according to actual production needs, and usually contains 3-5 sets of trial data. This ensures the comprehensiveness of selecting the optimal solution, while avoiding excessive trial data that could lead to increased trial production losses and extended trial production cycles.

[0083] In this embodiment, the technical means of obtaining actual test data of the current batch of injection molding raw materials after determining the initial parameters to be adjusted, and judging whether the deviation of the actual test data from the preset standard material parameters exceeds the allowable range, determines that the target defect is caused by material abnormality and generates an early warning if the deviation exceeds the allowable range, and confirms that the target defect is caused by process parameters and conducts trial adjustments if the deviation does not exceed the allowable range, thus achieving accurate attribution of the root cause of the target defect, distinguishing the impact of material abnormality and process parameter abnormality, and effectively solving the problems in the prior art that cannot distinguish whether the target defect is caused by injection molding raw material abnormality or process parameter abnormality, blindly adjusting process parameters resulting in ineffective parameter adjustment, wasting production costs and time, and failing to detect material abnormalities in time, leading to the continuous generation of unqualified products. Therefore, it achieves the technical effect of improving the accuracy and comprehensiveness of defect attribution, avoiding ineffective parameter adjustment, timely checking of material abnormalities, reducing production losses, providing accurate basis for subsequent trial adjustments of process parameters, and ensuring the stability of injection molding production quality and production efficiency.

[0084] In some embodiments, after determining a set of trial process parameters containing multiple trial data, during the small-batch trial production verification of each trial data, it may occur that some trial data can solve the target defect but cause new product defects. In this case, the following implementation method can be performed:

[0085] The server sorts all the trial data in the trial process parameter set to determine the order of distribution (which can be based on the parameter adjustment range from smallest to largest, or the priority of the parameters to be adjusted initially) to avoid confusion in the distribution order, which could lead to a mismatch between the trial production batch and the trial data. Subsequently, according to the sorting result, the first set of trial data is distributed to the target injection molding machine. During distribution, the trial data must be converted into control signals that can be recognized by the target injection molding machine's control system to ensure that the target injection molding machine can accurately interpret and execute the set of parameters.

[0086] After the target injection molding machine receives the trial data, it starts small-batch trial production and produces according to the process parameters set by the trial data. During the trial production, the server monitors the operating status of the target injection molding machine in real time to ensure that the trial production process is stable and there are no equipment failures. After the trial production is completed, the server establishes a communication connection with the supporting testing equipment (visual inspection equipment, dimensional inspection equipment, etc.) to obtain the actual product inspection results of all products in the trial production batch. The testing equipment needs to conduct a comprehensive inspection of each trial production product and record the pass status, defect type and quantity to ensure the accuracy and comprehensiveness of the inspection results.

[0087] The server compares and analyzes the actual product inspection results with the target defect, focusing on two key dimensions: first, whether the trial data eliminates or reduces the target defect to a preset acceptable level (e.g., the target defect incidence rate is below 1%); and second, whether any new defects appear in the actual product inspection results. During the comparison process, the server systematically checks the defect types of the trial production products, distinguishing between target defects and new defects to avoid misjudging minor residual target defects as new defects.

[0088] Finally, a judgment is made based on the comparison results: if a set of trial data (i.e., target trial data) meets the condition of "making the target defect disappear or reduce it to the qualified standard, but causing new defects to appear", the server determines that the target trial data has a process conflict. It is believed that although the data can solve the core defect, the parameter ratio is unreasonable and will cause new quality problems, and is not suitable for formal production. Subsequently, the server records the target trial data and its corresponding new defect type and trial production batch information together into the trial process blacklist, and establishes the correspondence between "trial data-new defects" to facilitate subsequent query and filtering.

[0089] The server, following a preset distribution order, repeats steps one through four for the remaining trial data in the trial process parameter set. It completes small-batch trial production, results acquisition, comparative analysis, and process conflict resolution for each set of trial data, ensuring all trial data is verified and complete. If a set of trial data fails to improve the target defect, or if no new defects appear but the target defect does not meet the standard, it is not considered a process conflict; only the trial production effect is recorded for subsequent selection of the optimal solution.

[0090] The above embodiments ensure the accuracy and relevance of trial production results by conducting small-batch trial production on each set of trial data sequentially; by comparing the test results with the target defects, trial data with process conflicts are accurately identified; by recording conflicting trial data to a blacklist, the reuse of such data is avoided, reducing the generation of new defects and production losses, while eliminating invalid data for subsequent selection of the optimal optimization strategy, improving the efficiency and safety of parameter optimization, and further ensuring product quality stability.

[0091] In some embodiments, during the process of conducting small-batch trial production and comparing the actual product test results with the target defects using trial data from the trial process parameter set, some trial data may not trigger process conflicts and can effectively resolve the target defects. In this case, the following implementation can be performed:

[0092] The server checks each set of trial data according to the screening criteria (no process conflict triggered, target defect disappeared or reduced to the qualified standard). Trial data that meets the criteria is selected as the final optimization and adjustment strategy. The server then determines whether there are multiple final optimization and adjustment strategies: if only one final optimization and adjustment strategy exists, the server directly determines that set of data as the final execution plan; if multiple final optimization and adjustment strategies exist, the server calls the actual yield rate data of each trial production batch, compares the actual yield rates corresponding to all final optimization and adjustment strategies, and selects the set with the highest actual yield rate as the final target optimization and adjustment strategy. If multiple sets of actual yield rates are the same, the strategy with the smallest parameter adjustment range and closer to the normal operating parameters of the equipment can be selected first to reduce the equipment's operating load.

[0093] Subsequently, the server converts the final optimization and adjustment strategy (including the target optimization and adjustment strategy) into a format recognizable control signal for the target injection molding machine's control system. Then, through communication with the target injection molding machine, the strategy is sent to the machine, along with a formal continuous production command. This command instructs the target injection molding machine to stop trial production and switch to formal production mode, initiating continuous production according to the process parameters set in the final optimization and adjustment strategy. Simultaneously with the start of formal continuous production, the server initiates a stable observation period timer, presets the observation period duration, and synchronizes it to the operator's terminal for real-time monitoring. During the observation period, the server continuously collects production and product testing data through real-time communication with the target injection molding machine and its associated testing equipment, continuously compiling and updating the real-time yield rate, and monitoring the target injection molding machine's operating conditions (such as equipment temperature and pressure stability) to ensure stable production.

[0094] After the stable observation period ends, the server summarizes and analyzes the real-time yield data during the observation period to determine whether the real-time yield rate has been maintained above the preset qualified threshold without significant fluctuations. If this condition is met, the effectiveness of the trial adjustment is verified, and the final optimized adjustment strategy can be used for formal production in the long term. If the real-time yield rate fluctuates significantly or falls below the qualified threshold during the observation period, the server determines that the trial adjustment effect has not met expectations, immediately suspends formal continuous production, regenerates the trial process parameter set, and conducts trial production verification again until an effective final optimized adjustment strategy is obtained.

[0095] The above embodiments ensure the effectiveness and safety of the final optimization and adjustment strategy by accurately screening trial data that meets the conditions; maximize the improvement of the quality of formal production by selecting the strategy with the highest actual yield rate; and verify the long-term effectiveness of trial adjustments by setting a stable observation period and monitoring the real-time yield rate, thereby avoiding batch defects caused by parameter issues in formal production, ensuring the stability and continuity of large-scale production, reducing production losses, and improving production efficiency and product quality consistency.

[0096] In some embodiments, during the process of recording target trial data with process conflicts to the trial process blacklist, it is possible that when subsequent preliminary trial data is generated, the same or similar process conflicts may recur, leading to invalid trial production and production losses. In this case, the following implementation can be performed:

[0097] The server calls the database interface of the trial process blacklist to extract the latest recorded target trial data, and simultaneously extracts the newly generated defect information corresponding to the target trial data, ensuring a one-to-one correspondence and preventing data misalignment. Next, the server performs parameter parsing on the extracted target trial data, focusing on the parameter ratio characteristics: the server extracts the specific values ​​of all process parameters (including initial adjustment parameters and their related process parameters) from the target trial data, calculates the numerical ratios and adjustment differences between parameters, and combines this with existing parameter association rules in the defect and process parameter knowledge graph to identify the core contradiction of the parameter ratio characteristic (such as the ratio of two parameters exceeding the reasonable range). For example, if the target trial data is "injection pressure 120MPa, holding pressure 80MPa" and the newly generated defect is "product burrs," the server parsing reveals the parameter ratio characteristic as "injection pressure to holding pressure ratio is 1.5:1," and determines that this ratio exceeds the reasonable range, which is the core cause of the burrs.

[0098] Afterwards, based on the parameter ratio characteristics obtained from parsing and the corresponding new defects, the server establishes a conflict mapping relationship: the mapping relationship must clearly define the unique correspondence or many-to-one correspondence between "a certain parameter ratio characteristic" and "a certain new defect", and at the same time mark the target injection molding machine number corresponding to the mapping relationship to ensure the relevance of the mapping relationship; after the establishment is completed, the server stores the conflict mapping relationship in a temporary database for subsequent updates to the knowledge graph. The server calls the update interface of the defect and process parameter knowledge graph. Based on the established conflict mapping relationship, it adds an exclusion constraint rule for the target injection molding machine. The content of the exclusion constraint rule must correspond precisely to the conflict mapping relationship, explicitly prohibiting the generation of trial data containing the parameter ratio characteristics. At the same time, it marks the target injection molding machine number to which the rule applies to avoid applying the rule to other injection molding machines (because different injection molding machines have different equipment characteristics, the reasonable range of parameter ratios varies). For example, based on the conflict mapping relationship of "injection pressure to holding pressure ratio of 1.5:1 corresponds to product burrs", an exclusion constraint rule is added: "For target injection molding machine M003, it is prohibited to generate trial data with injection pressure to holding pressure ratio > 1.4:1". After the rule is added, the server saves the knowledge graph to ensure that the rule can be called in subsequent steps.

[0099] When the target injection molding machine experiences the target defect again, and the server triggers the generation of preliminary trial data again, it will first call the exclusion constraint rules for that target injection molding machine in the defect and process parameter knowledge graph to pre-filter the preliminary trial data to be generated: After generating each set of preliminary trial data, the server first parses its parameter ratio characteristics and compares them with the prohibited ratio characteristics in the exclusion constraint rules. If the parameter ratio characteristics of the preliminary trial data meet the prohibited conditions of the exclusion constraint rules, the server will immediately intercept the data and not include it in the trial process parameter set; if the parameter ratio characteristics of the preliminary trial data do not violate the exclusion constraint rules, it will be allowed to enter the subsequent steps to generate the trial process parameter set.

[0100] Finally, after each pre-filtering operation, the server records the filtering results (the number of intercepted trial data and specific parameter ratios) and synchronizes them to the operator's terminal for review. Simultaneously, if subsequent misjudgments are found in the intercepted trial data (e.g., due to changes in equipment operating conditions rendering the original exclusion constraint rules inapplicable), the operator can instruct the server via the terminal to modify or delete the corresponding exclusion constraint rules, ensuring the rules' applicability. Furthermore, the server periodically summarizes data from the trial process blacklist, updates conflict mapping relationships and exclusion constraint rules, and continuously improves the defect and process parameter knowledge graph.

[0101] This step identified the core causes of process conflicts by extracting target trial data and emerging defects, and analyzing parameter ratio characteristics. By establishing conflict mapping relationships and adding exclusionary constraint rules, the knowledge graph was dynamically updated and improved. Subsequent pre-filtering intercepted conflict trial data, preventing the recurrence of similar process conflicts from the source, reducing the generation of emerging defects and ineffective trial production losses, shortening the parameter optimization cycle, and improving the efficiency and safety of parameter optimization. Simultaneously, it achieved continuous improvement in process optimization capabilities, further ensuring product quality stability and production efficiency.

[0102] The server in this application embodiment is described below from a hardware processing perspective. Please refer to [link / reference]. Figure 4 This is a schematic diagram of the physical device structure of a server in an embodiment of this application.

[0103] It should be noted that, Figure 4 The server structure shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0104] like Figure 4 As shown, the server includes a Central Processing Unit (CPU) 401, which can perform various appropriate actions and processes based on a program stored in Read-Only Memory (ROM) 402 or a program loaded from storage portion 408 into Random Access Memory (RAM) 403, such as performing the methods described in the above embodiments. The RAM 403 also stores various programs and data required for system operation. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An Input / Output (I / O) interface 405 is also connected to the bus 404.

[0105] The following components are connected to I / O interface 405: input section 406 including audio input devices, push-button switches, etc.; output section 407 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 408 including a hard disk, etc.; and communication section 409 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0106] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs the various functions defined in this application.

[0107] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0108] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0109] Specifically, the server in this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, it implements the injection molding machine processing parameter optimization method provided in the above embodiment.

[0110] In another aspect, this application also provides a computer-readable storage medium, which may be included in the server described in the above embodiments; or it may exist independently and not assembled into the server. The storage medium carries one or more computer programs that, when executed by a processor of the server, cause the server to implement the injection molding machine processing parameter optimization method provided in the above embodiments.

[0111] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0112] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0113] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for optimizing injection molding machine processing parameters, characterized in that, The method includes: The process parameters of the target injection molding machine are collected in real time, including at least the injection pressure, barrel temperature and mold temperature. The serial number data of the target injection molding machine is obtained, and the process parameters and the serial number data are input into the product defect prediction model to obtain the predicted yield rate of the current target injection molding machine. The product defect prediction model is trained in advance by deep learning based on multiple historical process parameter sample sets labeled with actual yield rates. Based on the predicted yield rate, an optimization and adjustment strategy is determined and sent to the operation terminal; In response to the process parameter adjustment command from the operating terminal, the optimization adjustment strategy is sent to the target injection molding machine for execution; Real-time monitoring of quality feedback after implementing the optimization and adjustment strategy, wherein the quality feedback refers to the actual good product status of a set quantity of products; The optimization strategy is adjusted based on the quality feedback. The step of adjusting the optimization strategy based on the quality feedback includes: If there are non-conforming products in the quality feedback, the defect characteristics of multiple non-conforming products in the same processing batch will be collected. The common defect features are extracted by comparing multiple defect manifestation features and used as target defects. Based on a pre-established knowledge graph of defects and process parameters, the main process parameters that caused the target defect are traced back and identified as preliminary parameters to be adjusted.

2. The method according to claim 1, characterized in that, After the step of determining the main process parameters as preliminary parameters to be adjusted, the method further includes: Obtain the actual test data of the injection molding raw material corresponding to the current processing batch, wherein the actual test data includes at least the moisture content and melt flow index of the raw material; Determine whether the deviation between the actual test data and the preset standard material parameters exceeds the allowable range; If the deviation exceeds the allowable range, it is determined that the current target defect is affected by the material's own abnormality, and a material abnormality warning is generated. If the deviation does not exceed the allowable range, it is confirmed that the target defect is caused by process parameters, and trial adjustments are made based on the initial parameters to be adjusted.

3. The method according to claim 2, characterized in that, The step of making trial adjustments based on the initial parameters to be adjusted specifically includes: Extract several preliminary parameters to be adjusted that cause the target defect; Based on the defect and process parameter knowledge graph, associated process parameters are determined that are related to each of the initial parameters to be adjusted. The associated process parameters are those that have a constraint relationship with the initial parameters to be adjusted. According to the set fine-tuning step size, generate trial data for each of the initial parameters to be adjusted and the corresponding associated process parameters; Determine a set of trial process parameters that includes multiple sets of the trial data.

4. The method according to claim 3, characterized in that, After determining the set of trial process parameters containing multiple trial data, the method further includes: The test data from each of the test process parameters in the set are sequentially sent to the target injection molding machine for small-batch trial production; Obtain the actual product test results for each trial production batch; Compare the actual product test results with the target defect; If target test data exists that makes the target defect disappear or reduce it to the qualified standard, but a new defect appears in the actual product test results, then it is determined that the target test data has a process conflict, and the target test data is recorded to the test process blacklist.

5. The method according to claim 4, characterized in that, After the step of comparing the actual product test results with the target defect, the method further includes: The trial data that did not trigger process conflicts and whose target defects disappeared or decreased to the qualified standard were selected as the final optimization and adjustment strategy. If multiple final optimization and adjustment strategies exist, the target optimization and adjustment strategy with the highest actual yield rate shall be selected. The target optimization and adjustment strategy is then sent to the target injection molding machine for formal continuous production. A stable observation period is set during the formal continuous production process, and the real-time yield rate during the stable observation period is continuously monitored to verify the effectiveness of the trial adjustment.

6. The method according to claim 4, characterized in that, After the step of recording the target test data to the test process blacklist, the method further includes: Extract the target test data and the corresponding new defects recorded in the test process blacklist; Analyze the parameter ratio characteristics in the target trial data and establish a conflict mapping relationship between the parameter ratio characteristics and the newly formed defects; Based on the conflict mapping relationship, a new exclusion constraint rule for the target injection molding machine is added to the defect and process parameter knowledge graph; When the step of generating preliminary test data is triggered again in a subsequent step, the exclusion constraint rule is invoked to pre-filter the preliminary test data in order to intercept test data that may cause process conflicts.

7. A server, characterized in that, The server includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code including computer instructions, and the one or more processors invoking the computer instructions to cause the server to perform the method as described in any one of claims 1-6.

8. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on the server, it causes the server to perform the method as described in any one of claims 1-6.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is run on the server, it causes the server to perform the method as described in any one of claims 1-6.

Images