Injection molding optimization control system and method based on neural network

Abstract

The application belongs to the technical field of injection molding. An injection molding optimization control system and method based on a neural network are provided. The method comprises determining a plurality of key quality indicators in the injection molding production process and acquiring the key quality indicators in real time; inputting the acquired key quality indicators into an injection product anomaly prediction model constructed based on a neural network to obtain an injection product anomaly prediction set; determining a data acquisition time period according to the injection product anomaly prediction set and acquiring a key quality indicator set according to the data acquisition time period; optimizing an injection process parameter model using the key quality indicator set to obtain new process parameters, and controlling the injection molding production equipment to produce the next batch of injection products according to the new process parameters. The application realizes online automatic optimization of the process parameters of the injection molding production equipment.

Description

Technical Field

[0001] This invention relates to the field of injection molding technology, and more specifically, to an injection molding optimization control method, system, electronic device, computer storage medium, and computer program product based on neural networks. Background Technology

[0002] Injection molding, also known as injection molding, is a molding method that combines injection and molding. The advantages of injection molding include high production speed and efficiency, automated operation, a wide variety of designs and shapes (from simple to complex), and sizes ranging from large to small. It also produces dimensionally accurate products, facilitates product updates and replacements, and can create complex shapes. Injection molding is suitable for mass production and molding processes involving complex shapes.

[0003] To ensure the quality of injection-molded products meets standards, the process parameters during injection molding production need to be continuously optimized. Existing injection molding parameter optimization methods are all obtained offline, resulting in low efficiency and failing to meet practical needs. This invention aims to solve or partially improve this technical problem. Summary of the Invention

[0004] To address this issue, the present invention provides a neural network-based injection molding optimization control method, system, electronic device, computer storage medium, and computer program product to solve the aforementioned technical problems.

[0005] A first aspect of the present invention provides a neural network-based injection molding optimization control method, the method comprising the following steps:

[0006] Several key quality indicators in the injection molding production process are identified, and the key quality indicators are acquired in real time.

[0007] The obtained key quality indicators are input into the injection molding product anomaly prediction model based on neural network to predict the injection molding product anomaly prediction set, which includes several injection molding product anomaly prediction values ​​sorted by time.

[0008] The data acquisition period is determined based on the injection molding product anomaly prediction set, and a key quality indicator set is obtained based on the data acquisition period; wherein, the key quality indicator set includes the recorded key quality indicators located within the acquisition period;

[0009] The injection molding process parameter model is optimized using the set of key quality indicators to obtain new process parameters, and the injection molding production equipment is controlled to produce the next batch of injection molded products according to the new process parameters.

[0010] In some embodiments, the real-time acquisition of each of the key quality indicators includes:

[0011] In multiple stages of injection molding product production, testing equipment is used to test injection molding product samples using contact and / or non-contact testing methods to obtain the aforementioned key quality indicators.

[0012] In some embodiments, inputting the acquired key quality indicators into an injection molding product anomaly prediction model constructed based on a neural network to predict an injection molding product anomaly prediction set includes:

[0013] Obtain the transmission rate statistics of each production stage of the injection molding production equipment, and determine the estimated transmission time between each pair of production stages based on the transmission rate statistics.

[0014] If the initial injection molded product sample in the production stage is set as Y1, then the injection molded product samples Yn in each production stage are selected according to the estimated transmission time values, and the key quality indicators corresponding to the injection molded product samples Y1-Yn are constructed as key quality indicator primitives.

[0015] Repeat the above steps to obtain multiple key quality indicator primitives, and sort each key quality indicator primitive based on the acquisition time of injection molded product sample Y1.

[0016] The sorted key quality indicator primitives are input into the injection molding product anomaly prediction model constructed based on a neural network to predict the injection molding product anomaly prediction set; wherein, the injection molding product anomaly prediction set includes several injection molding product anomaly prediction values ​​ordered sequentially according to future time.

[0017] In some embodiments, the prediction of the injection-molded product anomaly prediction set includes:

[0018] The injection molding product anomaly prediction model outputs a number of initial prediction values ​​for injection molding product anomalies and their corresponding confidence values, ordered sequentially according to future time. Each of the initial prediction values ​​for injection molding product anomalies is located between a set upper limit and a lower limit of the prediction value.

[0019] The correction weight of each of the initial predicted values ​​of the injection molded product anomalies is calculated based on the number of the initial predicted values ​​of the injection molded product anomalies, each of the future times and the confidence values.

[0020] The modified weights are multiplied by the corresponding initial predicted values ​​of the injection molded product anomalies to obtain a number of predicted values ​​of injection molded product anomalies ordered sequentially according to future time.

[0021] In some embodiments, calculating the correction weight of each initial predicted value of an anomaly of the injection molded product based on the number of anomaly initial predicted values, each future time, and the confidence value includes: In the formula, , For the first Correction weights for the initial predicted values ​​of anomalies in each injection molded product; The number of initial predicted abnormal values ​​for injection molded products. As the baseline number; This represents the average of the maximum time spans between each pair of the stated future moments. In order to be with the first The time span between two adjacent future moments corresponding to the initial predicted value of an anomaly for an injection molded product; For the first Confidence value of the initial predicted value of anomalies in an injection molded product.

[0022] In some embodiments, determining the data acquisition period based on the injection molded product anomaly prediction set, and obtaining the key quality indicator set based on the data acquisition period, includes:

[0023] Calculate the sum of the products of the predicted anomaly value of each injection molded product and the corresponding confidence value, and determine the data acquisition period based on the magnitude of the sum; the data acquisition period ends at the current time.

[0024] Extract the key quality indicators recorded within the data acquisition period, and construct the key quality indicator set accordingly.

[0025] A second aspect of the present invention provides an injection molding optimization control system based on a neural network, comprising an acquisition module, a processing module, and a storage module, wherein the processing module is electrically connected to the storage module and the acquisition module, respectively.

[0026] The acquisition module is used to acquire several key quality indicators during the injection molding production process and transmit them to the processing module.

[0027] The storage module is used to store computer programs;

[0028] The processing module is used to retrieve and execute the computer program in the storage module to perform the method described in the preceding item, obtain new process parameters, and control the injection molding production equipment to produce the next batch of injection molded products according to the new process parameters.

[0029] A third aspect of the present invention also discloses an electronic device comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, the processor executing the computer program to implement the method as described in any of the preceding claims.

[0030] A fourth aspect of the present invention provides a computer storage medium storing a computer program, characterized in that the computer program is executed by a processor to implement the method as described in any of the preceding claims.

[0031] The fifth aspect of the present invention provides a computer program product that, when run on a terminal, causes the terminal to execute in order to implement the method as described in any of the preceding claims.

[0032] The beneficial effects of this invention are as follows:

[0033] This invention achieves online optimization of process parameters during the injection molding process by combining an injection molding process parameter model and an injection molding product anomaly prediction model, significantly improving the efficiency of process parameter optimization and meeting actual production needs. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a schematic flowchart of an injection molding optimization control method based on neural networks disclosed in an embodiment of the present invention;

[0036] Figure 2 This is a schematic diagram of the structure of an injection molding optimization control system based on a neural network disclosed in an embodiment of the present invention. Detailed Implementation

[0037] The following specific embodiments illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] Furthermore, the technical features involved in the different embodiments of this application described below can be combined with each other as long as they do not conflict with each other.

[0039] like Figure 1 As shown in the figure, an embodiment of the present invention discloses an injection molding optimization control method based on neural networks, the method comprising the following steps:

[0040] Several key quality indicators in the injection molding production process are identified, and the key quality indicators are acquired in real time.

[0041] The obtained key quality indicators are input into the injection molding product anomaly prediction model based on neural network to predict the injection molding product anomaly prediction set, which includes several injection molding product anomaly prediction values ​​sorted by time.

[0042] The data acquisition period is determined based on the injection molding product anomaly prediction set, and a key quality indicator set is obtained based on the data acquisition period; wherein, the key quality indicator set includes the recorded key quality indicators located within the acquisition period;

[0043] The injection molding process parameter model is optimized using the set of key quality indicators to obtain new process parameters, and the injection molding production equipment is controlled to produce the next batch of injection molded products according to the new process parameters.

[0044] In this embodiment of the invention, two models are designed: an injection molding process parameter model and an injection molding product anomaly prediction model. The injection molding process parameter model is used to determine the optimal set of all process parameters involved in the injection molding process (such as injection speed, injection temperature, and injection pressure) based on the key quality indicators of the product to be injection molded (such as product dimensions, strength, and heat deformation) and the equipment parameters of the injection molding production equipment. The injection molding production equipment then starts production of the first batch of injection molded products based on the optimal parameter set. The injection molding product anomaly prediction model is used to predict the anomaly trend of this batch of injection molded products based on the key quality indicators obtained in real time. Based on the prediction result, a set of key quality indicators for a suitable time period is obtained. This set of key quality indicators can be used by the injection molding process parameter model to re-optimize the process parameters, and the optimized process parameters can be used for the production of the next batch of injection molded products. Therefore, this invention, through the cooperation of the above two models, achieves online optimization of process parameters during the injection molding product production process, significantly improving the efficiency of process parameter optimization and meeting actual production needs.

[0045] It should be noted that the injection molding process parameter model can also be built based on neural networks. The training set used for training includes key quality indicators and their corresponding process parameters (including all process parameters), labeled data representing the correlation between at least one process parameter and the key quality indicator. After training, the injection molding process parameter model can analyze the merits of the process parameters used in the current batch of production based on the key quality indicators characterizing the quality of each stage of injection molding, and automatically optimize at least some of the process parameters, thereby ensuring better product quality in the next batch of injection molded products.

[0046] In some embodiments, the real-time acquisition of each of the key quality indicators includes:

[0047] In multiple stages of injection molding product production, testing equipment is used to test injection molding product samples using contact and / or non-contact testing methods to obtain the aforementioned key quality indicators.

[0048] In this embodiment of the invention, existing quality inspections are all conducted manually after product production is completed, which inevitably means that the optimization of process parameters can only be achieved offline. However, this invention deploys corresponding testing equipment at each stage of injection molding production. This equipment can extract an appropriate number of samples from the corresponding production stage and test their key quality indicators, thus providing a basis for online optimization of the process parameters of this invention. The testing equipment can use contact and / or non-contact testing methods, and can be image recognition equipment, hardness testers, strength testers, thermal strain testers, etc., depending on the actual characteristics of different production stages and the key quality indicators to be tested; this invention does not impose any limitations.

[0049] In some embodiments, inputting the acquired key quality indicators into an injection molding product anomaly prediction model constructed based on a neural network to predict an injection molding product anomaly prediction set includes:

[0050] Obtain the transmission rate statistics of each production stage of the injection molding production equipment, and determine the estimated transmission time between each pair of production stages based on the transmission rate statistics.

[0051] If the initial injection molded product sample in the production stage is set as Y1, then the injection molded product samples Yn in each production stage are selected according to the estimated transmission time values, and the key quality indicators corresponding to the injection molded product samples Y1-Yn are constructed as key quality indicator primitives.

[0052] Repeat the above steps to obtain multiple key quality indicator primitives, and sort each key quality indicator primitive based on the acquisition time of injection molded product sample Y1.

[0053] The sorted key quality indicator primitives are input into the injection molding product anomaly prediction model constructed based on a neural network to predict the injection molding product anomaly prediction set; wherein, the injection molding product anomaly prediction set includes several injection molding product anomaly prediction values ​​ordered sequentially according to future time.

[0054] In this embodiment of the invention, injection molding production includes several main production stages such as mold closing, pre-plasticizing, injection, holding pressure, cooling, mold opening, ejection, and mold adjustment. The invention uses a control and testing device to sample and test at least several of these stages. Assuming a batch of 100 injection-molded products, products numbered 16, 30, 45, 58, 74, 85, and 96 are selected as samples. Injection-molded product 16 is selected as a sample and removed by the testing device after the pre-plasticizing stage; injection-molded product 30 is selected as a sample and removed by the testing device after the injection stage; injection-molded product 45 is selected as a sample and removed by the testing device after the holding pressure stage… The testing device performs specific key quality index tests on the injection-molded product samples removed at the corresponding production stages and records the corresponding acquisition time. Furthermore, injection molding machines produce plastic products in batches according to a certain process; injection molding is a typical batch and intermittent process. Since this invention does not sample and test the final injection-molded product, but rather samples the semi-finished and finished products at each stage of the injection-molded product process, and these semi-finished and finished products are not sampled from the same sample, the key quality indicators obtained by the above method cannot be accurately divided according to the same batch, and therefore cannot reflect the key quality differences between batches.

[0055] To address the aforementioned problems, this invention statistically analyzes the transmission rates of each production stage in an injection molding production line to obtain statistical values ​​for each transmission rate. Here, the injection molding production line is modeled as a continuous production line, with each line possessing its own transmission rate. This transmission rate includes the production rate (i.e., the rate at which a worker or machine completes the full production action of the corresponding production stage) and the conveying rate (i.e., the transmission rate of the conveyor belt between adjacent production lines). Based on these rates and the physical distance between each production stage, the estimated transmission time between any two production stages can be calculated. Of course, the transmission rate in this invention can be the action rate parameter of the moving equipment in the corresponding production stage, and the transmission time is the time between the completion of the previous production stage and the completion of the next production stage. The aforementioned transmission rates can be obtained through statistical calculations, the specific process of which will not be elaborated further.

[0056] After determining the estimated transmission time, the expected time for injection molded product sample Y1 to arrive at the next production stage can be calculated. This expected time is also roughly the time when other injection molded products adjacent to sample Y1, such as injection molded product number 17, arrive at the next production stage (11:25:20). If the recorded acquisition time includes injection molded product sample Y2 with a time close to this (11:25:26), it can be determined that injection molded product sample Y2 and injection molded product sample Y1 are adjacent in the production sequence, i.e., belong to the same batch. By repeating the above method, Y2-Yn (e.g., numbers 17, 18, 19, 20) belonging to the same batch as Y1 and adjacent in the production sequence can be determined from each production stage. These Y2-Yn are then integrated with their corresponding key quality indicators to obtain key quality indicator primitives. The multiple key quality indicator primitives obtained in the above method reflect the quality fluctuations of key quality indicators in the injection molding production equipment at intermediate production stages and in the final injection molded product. Then, by conducting in-depth analysis using the injection molded product anomaly prediction model, the predicted values ​​of injection molded product anomalies at different future times can be obtained. The predicted values ​​of several injection-molded products, ordered sequentially by future time points, represent the estimation of abnormal conditions of injection molding production equipment in the short term.

[0057] In some embodiments, the prediction of the injection-molded product anomaly prediction set includes:

[0058] The injection molding product anomaly prediction model outputs a number of initial prediction values ​​for injection molding product anomalies and their corresponding confidence values, ordered sequentially according to future time. Each of the initial prediction values ​​for injection molding product anomalies is located between a set upper limit and a lower limit of the prediction value.

[0059] The correction weight of each of the initial predicted values ​​of the injection molded product anomalies is calculated based on the number of the initial predicted values ​​of the injection molded product anomalies, each of the future times and the confidence values.

[0060] The modified weights are multiplied by the corresponding initial predicted values ​​of the injection molded product anomalies to obtain a number of predicted values ​​of injection molded product anomalies ordered sequentially according to future time.

[0061] In this embodiment of the invention, the injection molding product anomaly prediction model, through in-depth analysis of the key quality indicator primitives obtained above, can predict several initial predicted values ​​of injection molding product anomalies and their corresponding confidence values, ordered sequentially according to future times: [(a1,b1,t1),(a2,b2,t2),..., (an,bn,tn)], where an is the initial predicted value of the injection molding product anomaly, bn is the confidence value of the initial predicted value of the injection molding product anomaly, and tn is the corresponding future time. Due to limitations in the accuracy of the injection molding product anomaly prediction model's construction and training level, the accuracy of the above-mentioned initial predicted values ​​of injection molding product anomalies and their corresponding confidence values ​​is not high enough, and corrections need to be made considering other practical factors.

[0062] Specifically, based on the number of initial predicted values ​​of injection molded product anomalies (i.e., n above), each future time (i.e., tn above), and the confidence value (i.e., bn above), a correction weight is calculated for each initial predicted value of injection molded product anomalies. Then, the correction weight is used to multiply the corresponding initial predicted value of injection molded product anomalies to obtain the final predicted value of injection molded product anomalies. Compared with the initial predicted value of injection molded product anomalies, this predicted value of injection molded product anomalies takes into account the actual characteristics of the preliminary prediction results of the injection molded product anomaly prediction model and has higher reliability.

[0063] In some embodiments, calculating the correction weight of each initial predicted value of an anomaly of the injection molded product based on the number of anomaly initial predicted values, each future time, and the confidence value includes: In the formula, , For the first Correction weights for the initial predicted values ​​of anomalies in each injection molded product; The number of initial predicted abnormal values ​​for injection molded products. As the baseline number; This represents the average of the maximum time spans between each pair of the stated future moments. In order to be with the first The time span between two adjacent future moments corresponding to the initial predicted value of an anomaly for an injection molded product; For the first Confidence value of the initial predicted value of anomalies in an injection molded product.

[0064] In this embodiment of the invention, the injection molding product anomaly prediction model can directly predict multiple initial predicted values ​​of injection molding product anomalies ordered sequentially according to future time. However, some initial predicted values ​​of injection molding product anomalies will exceed the range defined by the preset upper and lower limits of the predicted value, and these will be deleted. Predicted values ​​within the range are considered to have higher reliability and will be retained. As a result, the number of remaining initial predicted values ​​of injection molding product anomalies decreases and the time span is no longer uniform.

[0065] The number of initial predicted values ​​for injection molded product anomalies is related to the prediction accuracy of the injection molded product anomaly prediction model. A higher number of predicted values ​​indicates more predictions falling within the aforementioned reasonable range (i.e., the range between the upper and lower limits of the predicted value), thus increasing the overall reliability of the prediction. A larger time span between future moments of the initial predicted values ​​for injection molded product anomalies indicates that the prediction model can achieve reliable predictions over a longer period, also indicating higher overall reliability of the prediction. Furthermore, the closer the local time span is to the average of the maximum time span (i.e., the maximum time span divided by n), the higher the accuracy of the local prediction. Additionally, considering the confidence value of the initial predicted value for injection molded product anomalies compared to the sum of all confidence values, and the ratio of the current time span to the maximum time span, the correction weight for the i-th initial predicted value for injection molded product anomalies can be calculated.

[0066] In some embodiments, determining the data acquisition period based on the injection molded product anomaly prediction set, and obtaining the key quality indicator set based on the data acquisition period, includes:

[0067] Calculate the sum of the products of the predicted anomaly value of each injection molded product and the corresponding confidence value, and determine the data acquisition period based on the magnitude of the sum; the data acquisition period ends at the current time.

[0068] Extract the key quality indicators recorded within the data acquisition period, and construct the key quality indicator set accordingly.

[0069] In this embodiment of the invention, the predicted anomaly values ​​of injection molded products obtained above are still associated with the corresponding confidence values. All predicted anomaly values ​​of injection molded products are multiplied by their corresponding confidence values ​​and then summed, i.e., a weighted summation operation is performed to obtain a representation of the overall "anomaly" degree of the output results of the injection molded product anomaly prediction model. The acquisition time span of key quality indicators is then determined based on the degree of "anomaly" reflected by this representation. Specifically, the greater the degree of "anomaly," i.e., the larger the sum, the longer the time span of recorded key quality indicators is acquired. More data representing the quality of injection molded products is used to optimize process parameters, which can improve the accuracy of the optimization results. Otherwise, a shorter time span of recorded key quality indicators is acquired, using less data representing the quality of injection molded products to optimize process parameters. Less data allows for a faster optimization rate.

[0070] like Figure 2 As shown, this embodiment of the invention also discloses an injection molding optimization control system based on a neural network, including an acquisition module, a processing module, and a storage module, wherein the processing module is electrically connected to the storage module and the acquisition module respectively;

[0071] The acquisition module is used to acquire several key quality indicators during the injection molding production process and transmit them to the processing module.

[0072] The storage module is used to store computer programs;

[0073] The processing module is used to retrieve and execute the computer program in the storage module to perform the method described in the preceding item, obtain new process parameters, and control the injection molding production equipment to produce the next batch of injection molded products according to the new process parameters.

[0074] This invention also discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, characterized in that: the processor executes the computer program to implement the method described in the foregoing embodiments.

[0075] This invention also discloses a computer storage medium storing a computer program, characterized in that the computer program is executed by a processor to implement the method described in the foregoing embodiments.

[0076] This invention also discloses a computer program product that, when run on a terminal, causes the terminal to execute the method described in the foregoing embodiments.

[0077] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0078] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0079] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0080] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention.

Claims

1. A neural network-based injection molding optimization control method, characterized in that, The method includes the following steps: Several key quality indicators in the injection molding production process are identified, and the key quality indicators are acquired in real time. The obtained key quality indicators are input into the injection molding product anomaly prediction model based on neural network to predict the injection molding product anomaly prediction set, which includes several injection molding product anomaly prediction values ​​sorted by time. The data acquisition period is determined based on the injection molding product anomaly prediction set, and a key quality indicator set is obtained based on the data acquisition period; wherein, the key quality indicator set includes the recorded key quality indicators located within the acquisition period; The injection molding process parameter model is optimized using the set of key quality indicators to obtain new process parameters, and the injection molding production equipment is controlled to produce the next batch of injection molded products according to the new process parameters.

2. The injection molding optimization control method based on neural networks according to claim 1, characterized in that: The real-time acquisition of each of the key quality indicators includes: In multiple stages of injection molding product production, testing equipment is used to test injection molding product samples using contact and / or non-contact testing methods to obtain the aforementioned key quality indicators.

3. The injection molding optimization control method based on neural networks according to claim 1, characterized in that: The process involves inputting the acquired key quality indicators into a neural network-based injection molding product anomaly prediction model to predict an injection molding product anomaly prediction set, including: Obtain the transmission rate statistics of each production stage of the injection molding production equipment, and determine the estimated transmission time between each pair of production stages based on the transmission rate statistics. If the initial injection molded product sample in the production stage is set as Y1, then the injection molded product samples Yn in each production stage are selected according to the estimated transmission time values, and the key quality indicators corresponding to the injection molded product samples Y1-Yn are constructed as key quality indicator primitives. Repeat the above steps to obtain multiple key quality indicator primitives, and sort each key quality indicator primitive based on the acquisition time of injection molded product sample Y1. The sorted key quality indicator primitives are input into the injection molding product anomaly prediction model constructed based on a neural network to predict the injection molding product anomaly prediction set; wherein, the injection molding product anomaly prediction set includes several injection molding product anomaly prediction values ​​ordered sequentially according to future time.

4. The injection molding optimization control method based on neural networks according to claim 3, characterized in that: The prediction yields a set of anomaly predictions for the injection-molded product, including: The injection molding product anomaly prediction model outputs a number of initial prediction values ​​for injection molding product anomalies and their corresponding confidence values, ordered sequentially according to future time. Each of the initial prediction values ​​for injection molding product anomalies is located between a set upper limit and a lower limit of the prediction value. The correction weight of each of the initial predicted values ​​of the injection molded product anomalies is calculated based on the number of the initial predicted values ​​of the injection molded product anomalies, each of the future times and the confidence values. The modified weights are multiplied by the corresponding initial predicted values ​​of the injection molded product anomalies to obtain a number of predicted values ​​of injection molded product anomalies ordered sequentially according to future time.

5. The injection molding optimization control method based on neural networks according to claim 4, characterized in that: The step of calculating the correction weight of each initial predicted value of an anomaly in the injection molded product based on the number of anomalies in the initial predicted values, each future time, and the confidence value includes: In the formula, is the adjustment weight for the initial abnormal prediction value of the i-th injection molded product; n is the number of initial abnormal prediction values ​​for injection molded products, and N is the baseline number; This represents the average of the maximum time spans between each pair of the stated future moments. This represents the time span between two adjacent future moments corresponding to the initial predicted value of the anomaly of the i-th injection molded product. is the confidence value of the initial predicted value of the anomaly for the i-th injection molded product.

6. The injection molding optimization control method based on neural networks according to claim 4, characterized in that: The step of determining the data acquisition period based on the injection molded product anomaly prediction set, and obtaining the key quality indicator set based on the data acquisition period, includes: Calculate the sum of the products of the predicted anomaly value of each injection molded product and the corresponding confidence value, and determine the data acquisition period based on the magnitude of the sum; the data acquisition period ends at the current time. Extract the key quality indicators recorded within the data acquisition period, and construct the key quality indicator set accordingly.

7. A neural network-based injection molding optimization control system, comprising an acquisition module, a processing module, and a storage module, wherein the processing module is electrically connected to the storage module and the acquisition module, respectively; The acquisition module is used to acquire several key quality indicators during the injection molding production process and transmit them to the processing module. The storage module is used to store computer programs; Its features are: The processing module is used to retrieve and execute the computer program in the storage module to perform the method as described in any one of claims 1-6, obtain new process parameters, and control the injection molding production equipment to produce the next batch of injection molded products according to the new process parameters.

8. An electronic device, comprising: At least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, characterized in that: the processor executes the computer program to implement the method as claimed in any one of claims 1-6.

9. A computer storage medium storing a computer program, characterized in that: The computer program is executed by a processor to implement the method as described in any one of claims 1-6.

10. A computer program product comprising a computer program stored on a non-transitory computer-readable medium, characterized in that: When the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

Images