Control method of injection molding system and injection molding system

Abstract

The invention relates to a control method of an injection molding system and the injection molding system.The method comprises the steps that in response to the situation that injection molding equipment enters a pressure maintaining stage, pressure maintaining is conducted on the basis of one value from 80% to 90% of the maximum filling pressure, and pressure change information in the pressure maintaining duration time is collected; responding to a plastic product forming trigger signal of injection molding equipment, and controlling an image acquisition unit to continuously acquire N frames of plastic product images; on the basis of the pressure change information, the N frames of plastic product images and a pre-trained neural network, determining the adjustment directions of the dwell pressure and the dwell time; and based on the adjusting direction, the pressure maintaining pressure and the pressure maintaining time of the next process of the first-stage injection molding equipment are adjusted, and the first-stage pressure maintaining pressure is adjusted to be 4%-7% of the maximum filling pressure. According to the pressure maintaining device, self-adaptive adjustment of the pressure maintaining pressure is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of injection molding, and in particular to a control method of an injection molding system and an injection molding system. Background Art

[0002] During plastic injection molding, the melt injected into the cavity will shrink due to cooling, so the screw must continue to move forward slowly so that the melt in the material tube continues to be injected into the cavity to supplement the shrinkage of the part. The above process is called holding pressure.

[0003] During the pressure holding process, the nozzle of the injection molding machine continuously adds material to the cavity to fill the volume vacated by the shrinkage of the part. If the cavity is filled without pressure holding, the part will shrink by about 25%, and shrinkage marks will form at the ribs due to excessive shrinkage.

[0004] Choosing the appropriate holding time and holding pressure can effectively prevent the shrinkage of the product, and the production of the product is relatively stable, achieving ideal finished product quality control. However, in actual applications, the holding pressure and holding time are set once by the technicians based on their experience and analysis of the finished product, and then the set holding pressure and holding time are continuously used for injection molding. However, since the state of the injection molding machine, the environment, and the state of the raw materials may change, if the holding time and pressure are not updated at this time, the quality of the finished product will be reduced. Summary of the invention

[0005] Aiming at the problem of how to automatically adjust the holding pressure and the holding time, the present invention provides a control method of an injection molding system and an injection molding system.

[0006] In a first aspect, the present application provides a control method for an injection molding system, which is applicable to an injection molding system, wherein the injection molding system comprises: an injection molding device, a control panel, an image acquisition unit, and a pressure sensor; the method comprises: In response to the injection molding device entering a holding pressure stage, holding pressure is performed based on a value between 80% and 90% of the maximum filling pressure, and pressure change information within the holding pressure duration is collected; In response to a trigger signal of plastic product molding of the injection molding device, the image acquisition unit is controlled to continuously acquire N frames of plastic product images; Based on the pressure change information, N frames of the plastic product image and the pre-trained neural network, determine the adjustment direction of the holding pressure and the holding time; Based on the adjustment direction, the holding pressure and holding time of the next process of the injection molding equipment at the first level are adjusted, wherein the first level holding pressure is adjusted to 4% to 7% of the maximum filling pressure.

[0007] In some embodiments, the step of determining the adjustment direction of the holding pressure and the holding time based on the pressure change information, N frames of the plastic product image, and a pre-trained neural network includes: Sending the pressure change information to a spatiotemporal feature extraction network; Sending N frames of plastic product images to a spatial feature extraction network; The outputs of the spatiotemporal feature extraction network and the spatial feature extraction network are concatenated and sent to a fully connected neural network; Based on the output of the fully connected neural network, the adjustment direction of the holding pressure and the holding time is confirmed.

[0008] In some embodiments, the training data of the pre-trained neural network includes: Pressure change information, corresponding plastic product image, and adjustment direction of the holding pressure and holding time, wherein the adjustment direction includes: one or more combinations of increasing the holding pressure, reducing the holding pressure, increasing the holding time, and reducing the holding time.

[0009] In some embodiments, when the training data is produced, the displayed annotation page information includes: a plastic product image at a previous moment, a curve graph corresponding to the pressure change information, and a plastic product image at a next moment, a curve graph corresponding to the pressure change information; The method further includes: using a pentagonal tree model for annotation by an annotator, wherein the annotator selects child nodes of the pentagonal tree based on the annotated page information, and the child nodes of the pentagonal tree include: Increase the holding pressure and increase the holding time; Increase holding pressure and reduce holding time; Reduce the holding pressure and increase the holding time; Reduce holding pressure and reduce holding time; Maintain the holding pressure and holding time.

[0010] In some embodiments, when the training data is produced, the displayed annotation page information includes: a plastic product image at a previous moment, a curve graph corresponding to the pressure change information, and a plastic product image at a next moment, a curve graph corresponding to the pressure change information; The method further includes: using a ternary tree model for annotation by an annotator, wherein the annotator selects a child node of the ternary tree based on the annotated page information, and the child node of the ternary tree includes: Increase the holding pressure and increase the holding time; Reduce holding pressure and reduce holding time; Maintain the holding pressure and holding time.

[0011] In some embodiments, the spatiotemporal feature extraction network comprises: An imaging module, used for dividing the pressure change information into multiple segments of sub-time sequence signals, and converting each segment of the sub-time sequence signal into a two-dimensional image; A convolutional neural network, configured to receive the two-dimensional image and extract a spatial feature map of the two-dimensional image; The recurrent neural network receives the spatial feature map in time sequence and outputs a spatiotemporal feature map.

[0012] In some embodiments, the convolutional neural network includes multiple convolutional layers connected in sequence, and the convolution kernel sizes of the multiple convolutional layers decrease along the connection order.

[0013] In some embodiments, the imaging module includes: A sliding window module, used for intercepting the sub-time sequence signals according to a preset time interval, wherein the overlap of adjacent sub-time sequence signals is 70% to 75%, and the time length of the sub-time sequence signals is 0.75 to 1.25 seconds; The Gram angle field module is used to convert the sub-time series signal into a Gram angle field diagram.

[0014] In some embodiments, the injection molding system further includes a fill light, and the method further includes: The maximum filling pressure is adjusted according to the adjusted holding pressure and a preset ratio of the holding pressure to the maximum filling pressure.

[0015] In some embodiments, the injection molding system further includes a fill light, and the method further includes: The raw material type or plastic product type input by the user is obtained, and the fill light of the corresponding color is controlled to light up, wherein the fill light includes red light, yellow light, blue light and white light.

[0016] In a second aspect, the present application provides an injection molding system, which includes an injection molding device, a control board, an image acquisition unit, and a pressure sensor; the control board is respectively connected to the injection molding device, the image acquisition unit, and the pressure sensor, and the control board stores an injection molding control program. When the injection molding control program is executed, the above-mentioned control method of the injection molding system is implemented.

[0017] This application confirms the adjustment direction of the holding pressure and holding time based on the pressure change information in the holding stage, the image of the plastic product, and the pre-trained neural network; then based on the adjustment direction, the holding pressure and holding time of the next process of the injection molding equipment are adjusted step by step, thereby realizing automatic adjustment of the holding pressure and reducing shrinkage marks. In addition, since the pressure change information in the holding stage is used, the holding pressure can be adjusted in time after the holding pressure in the holding stage changes. For example, in the holding stage, if the holding pressure gradually decreases due to equipment or other reasons, the holding pressure can be increased in the next process. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of an embodiment of a control method for an injection molding system of the present invention; Figure 2 A flowchart of another embodiment of the control method of the injection molding system of the present invention; Figure 3 It is a schematic diagram of the structure of the pre-trained neural network of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0020] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0021] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0022] Example 1 Reference Figure 1 The present application provides a control method for an injection molding system, which is applicable to the injection molding system. The injection molding system includes: an injection molding device, a control panel, an image acquisition unit, and a pressure sensor; the method includes: S100, in response to the injection molding equipment entering a pressure holding stage, performing pressure holding based on a value between 80% and 90% of the maximum filling pressure, and collecting pressure change information during the pressure holding duration; In the injection molding equipment, the injection molding equipment will automatically detect the end of injection molding and then enter the pressure holding stage. The control board only needs to collect the pressure after this moment through the pressure sensor to obtain the pressure change information during the pressure holding duration.

[0023] S200, in response to a trigger signal of plastic product molding of an injection molding device, controlling the image acquisition unit to continuously acquire N frames of plastic product images; In some implementations, the injection molding equipment will automatically detect the molding of the plastic product and then output a trigger signal to the control board; in some embodiments, the camera can also continuously capture images, and after confirming the molding of the plastic product based on the continuously captured images, the next N frames of plastic product images are recorded.

[0024] S300, confirming the adjustment direction of the holding pressure and the holding time based on the pressure change information, N frames of the plastic product image and the pre-trained neural network; The adjustment direction may include increasing the holding pressure, decreasing the holding pressure, increasing the holding time, and decreasing the holding time, or a combination thereof.

[0025] S400, based on the adjustment direction, adjusting the holding pressure and holding time of the next process of the injection molding equipment at the first level, wherein the first level holding pressure is adjusted to 4% to 7% of the maximum filling pressure.

[0026] In this embodiment, the holding pressure is first preset to about 85% of the maximum filling pressure, and then the amplitude of the first-level adjustment of the holding pressure and the amplitude of the first-level adjustment of the holding time are designed and preset, for example, 5% of the maximum filling pressure and 5% of the initial holding time. With this setting, the user only needs to enter the maximum filling pressure and the initial holding time. In actual application, after adjusting the holding pressure, you can go back to adjust the maximum filling pressure. For example, if the holding pressure is increased by 5%, the maximum filling pressure can be adjusted to: (holding pressure*105%) / 85%.

[0027] By adjusting the settings of the first level, step-by-step adjustments are achieved, which can reduce the difficulty of labeling training data and reduce the amount of computation of the neural network, as will be explained in detail later.

[0028] This application confirms the adjustment direction of the holding pressure and holding time based on the pressure change information in the holding stage, the image of the plastic product, and the pre-trained neural network; then based on the adjustment direction, the holding pressure and holding time of the next process of the injection molding equipment are adjusted step by step, thereby realizing automatic adjustment of the holding pressure and reducing shrinkage marks. In addition, since the pressure change information in the holding stage is used, the holding pressure can be adjusted in time after the holding pressure in the holding stage changes. For example, in the holding stage, if the holding pressure gradually decreases due to equipment or other reasons, the holding pressure can be increased in the next process.

[0029] Reference Figure 2 In some embodiments, S300, based on the pressure change information, N frames of the plastic product image and the pre-trained neural network, confirming the adjustment direction of the holding pressure and the holding time includes: S301, sending the pressure change information to a spatiotemporal feature extraction network; The spatiotemporal feature extraction network refers to a neural network that extracts spatiotemporal features. It should be noted that the time series signal of the pressure change information is specifically a continuous digital signal within the pressure holding time period, so it contains time information, and the spatial relationship between adjacent signals contains spatial information; therefore, the spatiotemporal feature information can be extracted; specifically, it can be implemented using a combination of the convolutional neural network and the recurrent neural network mentioned later.

[0030] S302, sending N frames of plastic product images to a spatial feature extraction network; The plastic product image is used to confirm the size of the shrinkage mark, so only a spatial feature extraction network is needed. However, in this embodiment, N frames of plastic product images are used, so the sizes of N shrinkage marks can be confirmed, and then the change curve of the shrinkage mark can be drawn as a feature provider and a fully connected neural network.

[0031] The spatial feature extraction network may be implemented as a multi-layer convolutional layer, and specifically, an existing convolutional neural network, such as LeNet-5, may be used.

[0032] S303, splicing the outputs of the spatiotemporal feature extraction network and the spatial feature extraction network, and sending the spliced ​​outputs to a fully connected neural network; It should be noted that the spatiotemporal feature extraction network outputs a feature vector, and the spatial feature extraction network also outputs a feature vector. In practical applications, in order to align two different feature vectors, the feature vectors can be flattened first and then concatenated to obtain a concatenated feature vector.

[0033] The fully connected neural network can integrate the information of the spliced ​​feature vectors and then output the classification result; the classification result is the adjustment direction, which includes: increasing the holding pressure, reducing the holding pressure, increasing the holding time, and reducing the holding time, one or more combinations thereof. The specifics can be confirmed based on the labeled training data.

[0034] S304: Based on the output of the fully connected neural network, determine the adjustment direction of the holding pressure and the holding time.

[0035] The output of the fully connected neural network is the probability of multiple adjustment directions. At this time, you only need to select the adjustment direction with the highest probability as the final output result.

[0036] In some embodiments, the training data of the pre-trained neural network includes: The pressure change information, the corresponding plastic product image, and the adjustment direction of the holding pressure and the holding time, the adjustment direction including: increasing the holding pressure, reducing the holding pressure, increasing the holding time, and reducing the holding time, one or more combinations thereof.

[0037] In some embodiments, when the training data is produced, the displayed annotation page information includes: a plastic product image at a previous moment, a curve graph corresponding to the pressure change information, and a plastic product image at a next moment, a curve graph corresponding to the pressure change information; The method further includes: using a pentagonal tree model for annotation by an annotator, wherein the annotator selects child nodes of the pentagonal tree based on the annotated page information, and the child nodes of the pentagonal tree include: Pressure change information, corresponding plastic product image, corresponding increase in holding pressure and increase in holding time; Pressure change information, corresponding plastic product images, corresponding increase in holding pressure and decrease in holding time; Pressure change information, corresponding plastic product images, corresponding reduction of holding pressure and increase of holding time; Pressure change information, corresponding plastic product image, corresponding reduction of holding pressure and reduction of holding time; Pressure change information, corresponding plastic product image, corresponding holding pressure and holding time.

[0038] In this embodiment, multi-dimensional adjustment of holding pressure and holding time is provided. Compared with the three-branch tree, the adjustment methods are more abundant and suitable for more delicate injection molded parts. In some other embodiments, when the training data is produced, the displayed annotation page information includes: a plastic product image at a previous moment, a curve graph corresponding to the pressure change information, and a plastic product image at a next moment, a curve graph corresponding to the pressure change information; The method further includes: using a ternary tree model for annotation by an annotator, wherein the annotator selects a child node of the ternary tree based on the annotated page information, and the child node of the ternary tree includes: Pressure change information, corresponding plastic product image, corresponding increase in holding pressure and increase in holding time; Pressure change information, corresponding plastic product image, corresponding reduction of holding pressure and reduction of holding time; Pressure change information, corresponding plastic product image, corresponding holding pressure and holding time.

[0039] In this embodiment, a single-dimensional adjustment of the holding pressure and the holding time is provided. Compared with the pentagonal tree, the adjustment method is simpler and suitable for relatively rough injection molded parts.

[0040] In the above two embodiments, considering that the labeling personnel have no idea about the size of the shrinkage mark, a curve graph corresponding to the plastic product image and pressure change information at the previous moment, and a curve graph corresponding to the plastic product image and pressure change information at the next moment are provided.

[0041] What the annotator needs to mark is not the size of the shrinkage mark, because the annotator has no idea about the size of the shrinkage mark.

[0042] In this embodiment, the annotator only needs to confirm whether the shrinkage mark is larger or smaller than the previous moment. If it is larger than the previous moment, the holding pressure and / or the holding time can be reduced; if it is smaller than the previous moment, the holding pressure and / or the holding time can be reduced; further, if the curve graph corresponding to the pressure change information shows that the pressure continues to decrease, the holding pressure can be increased. If only the curve graph at the current moment shows that the pressure continues to decrease, the holding pressure can also be increased.

[0043] Reference Figure 3 In some embodiments, the spatiotemporal feature extraction network includes: An imaging module, used for dividing the pressure change information into multiple segments of sub-time sequence signals, and converting each segment of the sub-time sequence signal into a two-dimensional image; A convolutional neural network, configured to receive the two-dimensional image and extract a spatial feature map of the two-dimensional image; The recurrent neural network receives the spatial feature map in time sequence and outputs a spatiotemporal feature map.

[0044] It should be noted that the timing signal of the pressure change information is specifically a continuous digital signal within the pressure holding time period, and the neural network is not good at extracting the spatial features of the timing signal. In this embodiment, the timing signal is first segmented and then converted into a two-dimensional image, and then a two-dimensional convolutional neural network is used for feature extraction, which can more effectively extract spatial features.

[0045] Furthermore, since the convolutional neural network can output the spatial feature map in time sequence, the time series signals corresponding to adjacent spatial feature maps are continuous. At this time, the recurrent neural network receives the spatial feature map in time sequence, and thus the spatiotemporal feature map can be obtained. The recurrent neural network may use LSTM or the updated GRU, which is not limited here.

[0046] In some embodiments, the convolutional neural network includes multiple convolutional layers connected in sequence, and the convolution kernel sizes of the multiple convolutional layers decrease along the connection order.

[0047] It should be noted that in order to reduce the amount of calculation, the convolution kernel of the convolutional neural network can be designed to be gradually reduced, for example, 256*256, 128*128, 64*64, and 32*32; In some embodiments, the imaging module includes: A sliding window module, used for intercepting the sub-time sequence signals according to a preset time interval, wherein the overlap of adjacent sub-time sequence signals is 70% to 75%, and the time length of the sub-time sequence signals is 0.75 to 1.25 seconds; The Gram angle field module is used to convert the sub-time series signal into a Gram angle field diagram.

[0048] It should be noted that the holding time is not long enough. In order to achieve higher frequency recognition, the present embodiment designs adjacent sub-timing signals to have a certain degree of overlap, so that more sub-timing signals can be obtained, thereby increasing the recognition frequency and improving the accuracy.

[0049] The diagonal line of the Gram angular field diagram is the original sub-time series signal. Therefore, the use of the Gram angular field module can not only extract spatial information, but also retain the timing information inside the original sub-time series signal to a certain extent.

[0050] In some embodiments, the injection molding system further includes a fill light, and the method further includes: The raw material type or plastic product type input by the user is obtained, and the fill light of the corresponding color is controlled to light up, wherein the fill light includes red light, yellow light, blue light and white light.

[0051] It should be noted that the image acquisition unit is to capture the shrinkage marks of plastic products. Since the light in the production environment is generally not good, it is necessary to design a fill light. However, different plastic products have different colors. Controlling the fill light of the corresponding color to light up can better capture images that can identify shrinkage marks. In addition, because a neural network is used, it can not only identify shrinkage marks, but also capture some information that the labelers have not paid attention to, thus providing high-quality images of plastic products and effectively improving accuracy.

[0052] Example 2 The present application provides an injection molding system, which includes an injection molding device, a control board, an image acquisition unit, and a pressure sensor; the control board is respectively connected to the injection molding device, the image acquisition unit, the fill light, and the pressure sensor, and the control board stores an injection molding control program. When the injection molding control program is executed, the following method is implemented: S100, in response to the injection molding equipment entering a pressure holding stage, performing pressure holding based on a value between 80% and 90% of the maximum filling pressure, and collecting pressure change information during the pressure holding duration; S200, in response to a trigger signal of plastic product molding of an injection molding device, controlling the image acquisition unit to continuously acquire N frames of plastic product images; S300, confirming the adjustment direction of the holding pressure and the holding time based on the pressure change information, N frames of the plastic product image and the pre-trained neural network; S400, based on the adjustment direction, adjusting the holding pressure and holding time of the next process of the injection molding equipment at the first level, wherein the first level holding pressure is adjusted to 4% to 7% of the maximum filling pressure.

[0053] In some embodiments, S300, based on the pressure change information, N frames of the plastic product image and the pre-trained neural network, confirming the adjustment direction of the holding pressure and the holding time includes: S301, sending the pressure change information to a spatiotemporal feature extraction network; S302, sending N frames of plastic product images to a spatial feature extraction network; S303, splicing the outputs of the spatiotemporal feature extraction network and the spatial feature extraction network, and sending the spliced ​​outputs to a fully connected neural network; S304: Based on the output of the fully connected neural network, determine the adjustment direction of the holding pressure and the holding time.

[0054] In some embodiments, the training data of the pre-trained neural network includes: Pressure change information, corresponding plastic product image, and adjustment direction of the holding pressure and holding time, wherein the adjustment direction includes: one or more combinations of increasing the holding pressure, reducing the holding pressure, increasing the holding time, and reducing the holding time.

[0055] In some embodiments, when the training data is produced, the displayed annotation page information includes: a plastic product image at a previous moment, a curve graph corresponding to the pressure change information, and a plastic product image at a next moment, a curve graph corresponding to the pressure change information; The method further includes: using a pentagonal tree model for annotation by an annotator, wherein the annotator selects child nodes of the pentagonal tree based on the annotated page information, and the child nodes of the pentagonal tree include: Pressure change information, corresponding plastic product image, corresponding increase in holding pressure and increase in holding time; Pressure change information, corresponding plastic product images, corresponding increase in holding pressure and decrease in holding time; Pressure change information, corresponding plastic product images, corresponding reduction of holding pressure and increase of holding time; Pressure change information, corresponding plastic product image, corresponding reduction of holding pressure and reduction of holding time; Pressure change information, corresponding plastic product image, corresponding holding pressure and holding time.

[0056] In some embodiments, when the training data is produced, the displayed annotation page information includes: a plastic product image at a previous moment, a curve graph corresponding to the pressure change information, and a plastic product image at a next moment, a curve graph corresponding to the pressure change information; The method further includes: using a ternary tree model for annotation by an annotator, wherein the annotator selects a child node of the ternary tree based on the annotated page information, and the child node of the ternary tree includes: Pressure change information, corresponding plastic product image, corresponding increase in holding pressure and increase in holding time; Pressure change information, corresponding plastic product image, corresponding reduction of holding pressure and reduction of holding time; Pressure change information, corresponding plastic product image, corresponding holding pressure and holding time.

[0057] In some embodiments, the spatiotemporal feature extraction network comprises: An imaging module, used for dividing the pressure change information into multiple segments of sub-time sequence signals, and converting each segment of the sub-time sequence signal into a two-dimensional image; A convolutional neural network, configured to receive the two-dimensional image and extract a spatial feature map of the two-dimensional image; The recurrent neural network receives the spatial feature map in time sequence and outputs a spatiotemporal feature map.

[0058] In some embodiments, the convolutional neural network includes multiple convolutional layers connected in sequence, and the convolution kernel sizes of the multiple convolutional layers decrease along the connection order.

[0059] In some embodiments, the imaging module includes: A sliding window module, used for intercepting the sub-time sequence signals according to a preset time interval, wherein the overlap of adjacent sub-time sequence signals is 70% to 75%, and the time length of the sub-time sequence signals is 0.75 to 1.25 seconds; The Gram angle field module is used to convert the sub-time series signal into a Gram angle field diagram.

[0060] In some embodiments, the injection molding system further includes a fill light, and the method further includes: The maximum filling pressure is adjusted according to the adjusted holding pressure and a preset ratio of the holding pressure to the maximum filling pressure.

[0061] In some embodiments, the injection molding system further includes a fill light, and the method further includes: The raw material type or plastic product type input by the user is obtained, and the fill light of the corresponding color is controlled to light up, wherein the fill light includes red light, yellow light, blue light and white light.

[0062] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0063] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0064] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0065] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0066] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0067] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0068] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A control method for an injection molding system, characterized in that: Applicable to an injection molding system, the injection molding system comprises: an injection molding device, a control panel, an image acquisition unit, and a pressure sensor; the method comprises: In response to the injection molding device entering a holding pressure stage, holding pressure is performed based on a value between 80% and 90% of the maximum filling pressure, and pressure change information within the holding pressure duration is collected; In response to a trigger signal of plastic product molding of the injection molding device, the image acquisition unit is controlled to continuously acquire N frames of plastic product images; Based on the pressure change information, N frames of the plastic product image and the pre-trained neural network, determine the adjustment direction of the holding pressure and the holding time; Based on the adjustment direction, the holding pressure and holding time of the next process of the injection molding equipment at the first level are adjusted, wherein the first level holding pressure is adjusted to 4% to 7% of the maximum filling pressure.

2. The control method of the injection molding system according to claim 1, characterized in that: The step of determining the adjustment direction of the holding pressure and the holding time based on the pressure change information, N frames of the plastic product image, and a pre-trained neural network includes: Sending the pressure change information to a spatiotemporal feature extraction network; Sending N frames of plastic product images to a spatial feature extraction network; The outputs of the spatiotemporal feature extraction network and the spatial feature extraction network are concatenated and sent to a fully connected neural network; Based on the output of the fully connected neural network, the adjustment direction of the holding pressure and the holding time is confirmed.

3. The control method of the injection molding system according to claim 2, characterized in that: The training data of the pre-trained neural network includes: Pressure change information, corresponding plastic product image, and adjustment direction of the holding pressure and holding time, wherein the adjustment direction includes: one or more combinations of increasing the holding pressure, reducing the holding pressure, increasing the holding time, and reducing the holding time.

4. The control method of the injection molding system according to claim 3, characterized in that: When the training data is produced, the displayed annotation page information includes: the image of the plastic product at the previous moment, the curve graph corresponding to the pressure change information, and the image of the plastic product at the next moment, the curve graph corresponding to the pressure change information; The method further includes: using a pentagonal tree model for annotation by an annotator, wherein the annotator selects child nodes of the pentagonal tree based on the annotated page information, and the child nodes of the pentagonal tree include: Increase the holding pressure and increase the holding time; Increase holding pressure and reduce holding time; Reduce the holding pressure and increase the holding time; Reduce holding pressure and reduce holding time; Maintain the holding pressure and holding time.

5. The control method of the injection molding system according to claim 3, characterized in that: When the training data is produced, the displayed annotation page information includes: the image of the plastic product at the previous moment, the curve graph corresponding to the pressure change information, and the image of the plastic product at the next moment, the curve graph corresponding to the pressure change information; The method further includes: using a ternary tree model for annotation by an annotator, wherein the annotator selects a child node of the ternary tree based on the annotated page information, and the child node of the ternary tree includes: Increase the holding pressure and increase the holding time; Reduce holding pressure and reduce holding time; Maintain the holding pressure and holding time.

6. The control method of the injection molding system according to claim 3, characterized in that: The spatiotemporal feature extraction network comprises: An imaging module, used for dividing the pressure change information into multiple segments of sub-time sequence signals, and converting each segment of the sub-time sequence signal into a two-dimensional image; A convolutional neural network, configured to receive the two-dimensional image and extract a spatial feature map of the two-dimensional image; The recurrent neural network receives the spatial feature map in time sequence and outputs a spatiotemporal feature map.

7. The control method of the injection molding system according to claim 6, characterized in that: The convolutional neural network includes multiple convolutional layers connected in sequence, and the convolution kernel sizes of the multiple convolutional layers decrease along the connection order; The imaging module comprises: A sliding window module, used for intercepting the sub-time sequence signals according to a preset time interval, wherein the overlap of adjacent sub-time sequence signals is 70% to 75%, and the time length of the sub-time sequence signals is 0.75 to 1.25 seconds; The Gram angle field module is used to convert the sub-time series signal into a Gram angle field diagram.

8. The control method of the injection molding system according to any one of claims 1 to 7, characterized in that: Also includes: The maximum filling pressure is adjusted according to the adjusted holding pressure and a preset ratio of the holding pressure to the maximum filling pressure.

9. The control method of the injection molding system according to claim 1, characterized in that: The injection molding system further includes a fill light, and the method further includes: The raw material type or plastic product type input by the user is obtained, and the fill light of the corresponding color is controlled to light up, wherein the fill light includes red light, yellow light, blue light and white light.

10. An injection molding system, characterized in that: Including injection molding equipment, control board, image acquisition unit, pressure sensor; The control board is connected to the injection molding equipment, the image acquisition unit, and the pressure sensor respectively. The control board stores an injection molding control program. When the injection molding control program is executed, the control method of the injection molding system according to any one of claims 1 to 9 is implemented.

Images