Box-type injection molding filling simulation method and system based on machine learning

Abstract

The present invention relates to the field of injection molding, and in particular to a box-matching injection molding filling simulation method and system based on machine learning, wherein the method comprises: obtaining historical data on box-matching injection molding, wherein the historical data comprises an injection mold image, corresponding injection molding parameters and injection molding results; obtaining an adjustment sequence of each injection molding parameter; obtaining a qualified adjustment interval and an unqualified adjustment interval of each injection molding parameter according to the data distribution of each injection molding parameter; calculating the possibility that each parameter value within the qualified adjustment interval of the injection molding parameter satisfies the qualified injection molding, and determining its optimal adjustment parameter set and candidate adjustment parameter set; photographing an injection mold image to be simulated, and determining a mold image in historical data that matches the mold image; simulating according to the optimal adjustment parameter set and candidate adjustment parameter set of the injection molding parameters corresponding to the mold image in the matched historical data, and obtaining the injection molding parameters with the best injection molding effect, thereby reducing the number of experiments for simulated injection molding and reducing the cost of injection molding simulation.

Description

Technical Field

[0001] The present invention relates to the field of injection molding, and more specifically to a box-matching injection molding filling simulation method and system based on machine learning. Background Art

[0002] Injection molding technology is widely used in various fields. For example, injection molding technology can be used to produce various plastic household products, auto parts, medical equipment, and electronic products. By using injection molding technology, high-precision, high-quality plastic products can be produced in large quantities and efficiently, reducing the error and labor intensity of manual operation. Before mass production of plastic products, enterprises will first determine the injection molding parameters to adapt to the injection mold, and continuously adjust the injection molding parameters for injection molding tests to determine the best injection molding parameters to improve the quality of plastic products.

[0003] Numerical simulation technology can simulate physical phenomena through computers and obtain numerical solutions of physical parameters through mathematical methods. Therefore, the prior art usually obtains the optimal injection molding parameters by establishing a three-dimensional model of the injection mold and performing multiple numerical simulations. For example, a method for determining the injection molding process parameters of plastic products and a method for establishing a coefficient library thereof are proposed in the patent document with publication number CN111898265B: a three-dimensional model is established for various types of plastic products and multiple numerical simulation analyses are performed to obtain the first theoretical injection molding process parameters of various types of plastic products, the first theoretical injection molding process parameters of various types of plastic products are adjusted, and multiple mold trials are performed on various types of plastic products to determine the actual injection molding process parameters of various types of plastic products. However, although this method reduces the waste of mold trial time and mold trial resources to a certain extent, the numerical simulation technology has a large amount of calculation, and for products with complex structures, it takes a long time to calculate the optimal injection molding parameters. Summary of the invention

[0004] In order to solve the problem of long simulation calculation time and high cost in the process of simulating injection molding of products with complex structures, the present invention proposes a box-matching injection molding filling simulation method and system based on machine learning. To this end, the present invention provides solutions in the following aspects.

[0005] In the first aspect, a box-matching injection molding filling simulation method based on machine learning includes: obtaining historical data about box-matching injection molding, the historical data including an injection mold image, corresponding injection molding parameters and injection molding results, the injection molding parameters including temperature, pressure and push speed; obtaining the adjustment order of each injection molding parameter; calculating the data distribution of each injection molding parameter when the injection molding results are qualified and unqualified, and obtaining the qualified adjustment interval and unqualified adjustment interval of each injection molding parameter; calculating the possibility of each parameter value in the qualified adjustment interval of the injection molding parameter satisfying the qualified injection molding, determining the parameter value with the possibility of qualified injection molding greater than a threshold as the optimal adjustment parameter set, and the remaining parameter values ​​as the candidate adjustment parameter set; photographing the injection mold image to be simulated, and determining the mold image in the historical data that matches it; determining the optimal adjustment parameter set and candidate adjustment parameter set of the corresponding injection molding parameters according to the mold image in the historical data, determining the value of the last injection molding parameter according to the adjustment order, and adjusting the injection molding parameters before the last in the optimal adjustment parameter set and the candidate adjustment parameter set for simulation.

[0006] The present invention obtains the adjustment order of each injection molding parameter and the optimal adjustment parameter set and the candidate adjustment parameter set, and determines the optimal adjustment parameter set and the candidate adjustment parameter set of the simulated injection mold through the similarity between the simulated injection mold and the historical injection mold. Therefore, the obtained simulation parameters have a higher probability of obtaining a better simulation effect and the data range is more concentrated, thereby reducing the time of simulated injection molding and reducing costs accordingly.

[0007] In one embodiment, the injection molding parameters further include: cooling speed.

[0008] In one embodiment, obtaining the adjustment order of each injection molding parameter includes: calculating the mutual information entropy between the injection molding parameters and the injection molding results, determining the decisiveness of each injection molding parameter on the injection molding results according to the mutual information entropy, and determining the adjustment order of each injection molding parameter in descending order according to the size of the decisiveness, wherein the decisiveness of the injection molding result is positively correlated with the mutual information entropy.

[0009] By calculating the mutual information entropy to determine the adjustment order of injection molding parameters, the injection molding parameters that have a greater impact on the injection molding results can be more accurately identified.

[0010] In one embodiment, calculating the mutual information entropy between the injection molding parameters and the injection molding results includes: calculating the difference between the information entropy of the injection molding results and the conditional entropy of the injection molding results under the condition of determined injection molding parameters, and using the difference as the mutual information entropy of the injection molding parameters and the injection molding results.

[0011] By calculating the mutual information entropy of injection molding parameters and injection molding results through the difference between the information entropy of injection molding results and the conditional entropy of injection molding results under determined injection molding parameters, the mutual information between injection molding parameters and injection molding results can be obtained, which more intuitively reflects the decisiveness of injection molding parameters on injection molding results.

[0012] In one embodiment, the data distribution of each injection molding parameter when the injection molding result is qualified and unqualified is calculated to obtain the qualified adjustment interval of each injection molding parameter, including: recording the data with qualified injection molding results in the historical data as qualified data, and recording the unqualified data as unqualified data, respectively calculating the data distribution of the qualified data, first performing a normality test on the data to test whether it is a normal distribution, if the data distribution is a normal distribution, taking the mean of the qualified injection molding parameter data as its first center point, according to 3 The qualified adjustment range of the injection molding parameters is obtained by the rule ,in Indicates the first center point of the injection molding parameters, Represents the first variance of the injection molding parameters;

[0013] The mean of the unqualified injection molding parameter data is taken as the second center point. The unqualified adjustment range of the injection molding parameters is obtained by the rule. ,in The second center point representing the injection molding parameters, Represents the second variance of the injection molding parameters.

[0014] If the data distribution is not a normal distribution, calculate the third center point of the injection molding parameters. With the third center point as the center of the circle, when 90% of the data falls within the circle, the radius R1 at this time is used to determine the adjustment range of the injection molding parameters. The qualified adjustment range is , where the third center point Calculate the average of the median, mode and mean of the qualified injection molding parameter data.

[0015] Calculate the fourth center point of the injection molding parameters, take the fourth center point as the center of the circle, when 90% of the data falls within the circle, the radius at this time That is to determine the adjustment range of injection molding parameters, the unqualified adjustment interval is , where the fourth center point The average is the sum of the median, mode and mean of the unqualified injection molding parameter data.

[0016] The present invention adopts different means to determine the adjustment interval for different types of data distribution (normal distribution or non-normal distribution), which is more reasonable and conducive to the final simulation.

[0017] In one embodiment, the possibility of each parameter value in the injection molding parameter qualified adjustment interval satisfying the injection molding qualification is calculated, and the parameter value with the injection molding qualification greater than the threshold is determined as the optimal adjustment parameter set, and the remaining parameter values ​​are the candidate adjustment parameter set, including: the calculation formula for the possibility of each parameter value in the injection molding parameter qualified adjustment interval satisfying the injection molding qualification is:

[0018]

[0019] Indicates injection molding parameters The possibility of meeting the requirements of injection molding. Indicates the value of the injection molding parameter. Indicates the center point of the unqualified injection molding parameter x, Indicates the adjustment range of unqualified injection molding parameters. Represents the center point of the qualified injection molding parameter x, Indicates the adjustment range of qualified injection molding parameters.

[0020] The optimal adjustment parameter set and the candidate adjustment parameter set are divided by the above calculation method, which reflects the influence of the injection molding parameters on the injection molding results, and the division result is more objective and accurate.

[0021] In the second aspect, a box-matching injection molding filling simulation system based on machine learning includes: a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the box-matching injection molding filling simulation method based on machine learning is implemented.

[0022] By simulating each injection molding parameter through a computer program, the result is more accurate and the error probability is relatively small. The injection molding result corresponding to each injection molding parameter can be obtained more quickly, and the injection molding parameters with the best injection molding effect can be obtained.

[0023] The beneficial effects of the present invention are as follows: the present invention more accurately determines the decisive size of each injection molding parameter on the injection molding result by calculating the mutual information entropy of the injection molding parameters on the injection molding result, thereby obtaining the adjustment order of each injection molding parameter, and determines the optimal adjustment parameter set and the candidate adjustment parameter set by calculating the data distribution of the injection molding parameters, further narrowing the parameter adjustment interval, and determines the optimal adjustment parameter set and the candidate adjustment parameter set of the simulated injection mold by the similarity between the simulated injection mold and the historical injection mold. Therefore, the obtained simulation parameters have a higher probability of obtaining a better simulation effect and the data range is more concentrated. Therefore, simulated injection molding of products with complex structures can reduce the time of simulated injection molding and can also reduce costs accordingly. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0025] Figure 1 is a flowchart schematically illustrating the steps of a box assembly injection molding simulation method based on machine learning according to an embodiment of the present invention;

[0026] Figure 2 It is a structural diagram of the box-matching injection molding filling simulation system based on machine learning in this embodiment. DETAILED DESCRIPTION

[0027] 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 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.

[0028] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0029] Figure 1 It is a step diagram of a box-matching injection molding filling simulation method based on machine learning according to an embodiment of the present invention.

[0030] like Figure 1 As shown, the box-assembly injection molding filling simulation method based on machine learning includes steps S1 to S6.

[0031] S1: Acquire historical data about box-mounted injection molding, wherein the historical data includes an injection mold image, corresponding injection molding parameters, and injection molding results.

[0032] In one embodiment, historical box-mounted injection mold images, corresponding injection molding parameters and injection molding results stored in a computer are obtained, wherein the injection molding parameters include temperature, pressure and advancement speed, the parameter values ​​contained in each injection molding parameter are a set, and each parameter value has a corresponding injection molding result. The injection molding result under each injection molding parameter is given a label according to the historical injection molding situation, and the label is qualified / unqualified, and the parameter value of each injection molding parameter is divided into two types: qualified and unqualified.

[0033] For example:

[0034]

[0035] That is to say, the set of all parameter values ​​contained in the injection molding parameter temperature is {T1, T2, ..., Tn}, T1 is just one of the parameter values. The parameter values ​​of each injection molding parameter when the injection molding result is qualified form a set, and the parameter values ​​when the injection molding result is unqualified form another set.

[0036] Among them, T1, N1, V1 is a group of data, the corresponding injection molding result is qualified, and the injection mold image is Z; T2, N2, V2 is a group of data, the corresponding injection molding result is unqualified, and the injection mold image is Z. Of course, in fact, there will be multiple groups of data in the form of the above tables in the historical data.

[0037] S2: Obtain the adjustment sequence of each injection molding parameter.

[0038] The adjustment order of the injection molding parameters is determined according to the decisiveness of each injection molding parameter on the injection molding result, wherein the decisiveness of each injection molding parameter on the injection molding result is obtained by calculating the mutual information entropy between each historical injection molding parameter and the injection molding result. The mutual information entropy between each injection molding parameter and the injection molding result reflects the reduction of the uncertainty of the injection molding result under the condition of known injection molding parameters. The more the uncertainty of the injection molding result is reduced, the greater the certainty of the injection molding result is under the condition of the known injection molding parameters, and the greater the decisiveness of the injection molding parameters on the injection molding result. Therefore, when adjusting the injection molding parameters to adapt to the new injection molding environment, the injection molding parameters with large decisiveness are adjusted first to ensure the qualified injection molding products.

[0039] In one embodiment, the information entropy of the injection molding result is first calculated, and the information entropy of the injection molding result reflects the uncertainty of the injection molding result. The larger the information entropy, the greater the uncertainty of the injection molding result. Then, the conditional entropy of the injection molding parameters on the injection molding result is calculated, and the conditional entropy of the injection molding parameters on the injection molding result reflects the uncertainty of the injection molding result when the injection molding parameters are determined. The mutual information entropy of the injection molding parameters on the injection molding result is calculated based on the information entropy of the injection molding result and the conditional entropy of the injection molding result when the injection molding parameters are determined, wherein the mutual information entropy of the injection molding parameters on the injection molding result is the difference between the information entropy of the injection molding result and the conditional entropy of the injection molding result when the injection molding parameters are determined.

[0040] Specifically, step S2 includes step S201 to step S203.

[0041] Step S201: calculating the information entropy of the injection molding result according to the occurrence probability of the injection molding result.

[0042] Among them, the information entropy of the injection molding result for:

[0043]

[0044] Indicates the injection molding result (for example, indicates qualified or unqualified), Indicates the injection molding result is Probability of occurrence.

[0045] Step S202: Calculate the conditional entropy of the injection molding parameters on the injection molding results according to the historical data.

[0046] Among them, the conditional entropy of the injection molding parameters on the injection molding results for:

[0047]

[0048] represents various injection molding parameters (for example, x1 represents temperature, x2 represents pressure, and x3 represents push speed), represents the joint probability of injection molding parameters and injection molding results, Indicates the injection molding parameters are probability.

[0049] Step S203: obtaining the mutual information entropy of the injection molding parameters on the injection molding results according to the information entropy of the injection molding results and the conditional entropy of the injection molding parameters on the injection molding results.

[0050] Among them, the mutual information entropy of the injection molding parameters on the injection molding results for:

[0051]

[0052] in, represents the mutual information entropy of the j-th injection molding parameter on the injection molding result, Represents the information entropy of the injection molding result, It represents the conditional entropy of the j-th injection molding parameter on the injection molding result. The larger the value is, the more the uncertainty of the j-th injection molding parameter on the injection molding result is reduced, and the greater its decisiveness on the injection molding result is. The more decisive the injection molding parameter is on the injection molding result, when adjusting the injection molding parameter to adapt to the new injection molding environment, the injection molding parameter with greater decisiveness should be adjusted first to ensure the qualified injection molding product.

[0053] S3: According to the data distribution of each injection molding parameter, the qualified adjustment interval and the unqualified adjustment interval of each injection molding parameter are obtained.

[0054] The data distribution of each injection molding parameter when the injection molding result is qualified or unqualified is calculated by acquiring the historical injection molding parameters and injection molding result data. According to the data distribution of each injection molding parameter when the injection molding result is qualified or unqualified, the mean is used as the center point, and the set of injection molding parameters whose probability of qualified and unqualified injection molding results is greater than the threshold is selected as the qualified adjustment interval and unqualified adjustment interval of each injection molding parameter. Based on this, the present invention calculates the possibility of each parameter value in the qualified adjustment interval of the injection molding parameter meeting the qualified injection molding in the next step and determines the optimal adjustment parameter set and candidate adjustment parameter set of each injection molding parameter.

[0055] Specifically, step S3 includes step S301 to step S302.

[0056] Step S301: Perform normal distribution test on historical injection molding data.

[0057] Draw the frequency histogram of each injection molding parameter and injection molding result, and judge whether it is a normal distribution by comparing the image of the normal distribution.

[0058] Step S302: Calculate the data distribution of each injection molding parameter when the injection molding result is qualified or unqualified, and obtain the qualified adjustment interval and unqualified adjustment interval of each injection molding parameter.

[0059] If the data distribution is normal, take the mean of qualified injection molding parameter data as its first center point, according to 3 The qualified adjustment range of the injection molding parameters is obtained by the rule ,in Indicates the first center point of the injection molding parameters, Represents the first variance of the injection molding parameters;

[0060] The mean of the unqualified injection molding parameter data is taken as the second center point. The unqualified adjustment range of the injection molding parameters is obtained by the rule. ,in The second center point representing the injection molding parameters, Represents the second variance of the injection molding parameters.

[0061] If the data distribution is not a normal distribution, calculate the third center point of the injection molding parameters. With the third center point as the center of the circle, when 90% of the data falls within the circle, the radius R1 at this time is used to determine the adjustment range of the injection molding parameters. The qualified adjustment range is , where the third center point Calculate the average of the median, mode and mean of the qualified injection molding parameter data.

[0062] Calculate the fourth center point of the injection molding parameters, take the fourth center point as the center of the circle, when 90% of the data falls within the circle, the radius at this time That is to determine the adjustment range of injection molding parameters, the unqualified adjustment interval is , where the fourth center point The average is the sum of the median, mode and mean of the unqualified injection molding parameter data.

[0063] S4: Calculate the possibility of each parameter value in the injection molding parameter qualified adjustment range meeting the injection molding qualification, and determine the optimal adjustment parameter set and candidate adjustment parameter set.

[0064] The closer the injection molding parameters are to the parameter values ​​of the qualified center point, the greater the possibility of satisfying the injection molding qualification. The closer the injection molding parameters are to the parameter values ​​of the unqualified center point, the smaller the possibility of satisfying the injection molding qualification. Therefore, when adjusting the injection molding parameters, it is necessary to be close to the qualified adjustment interval and away from the unqualified adjustment interval. In addition, the larger the adjustment range of the qualified center point, the greater the tolerance of the injection molding qualification to the parameter. Under the condition that the distance between different injection molding parameters and the parameter values ​​of the qualified center point is the same, the injection molding parameters with a large adjustment range are more likely to satisfy the injection molding qualification. By calculating the possibility of each parameter value in the injection molding parameter qualified adjustment interval satisfying the injection molding qualification, the parameter value with the possibility of injection molding qualification greater than the threshold is determined as the optimal adjustment parameter set, and the remaining parameter values ​​are the candidate adjustment parameter set.

[0065] S5: photographing an image of the injection mold to be simulated, and determining a mold image in historical data that matches the image.

[0066] The injection mold image to be simulated taken by the camera is compared with the mold image in the historical data to find a matching mold image in the historical data. For example, the similarity between the two can be calculated, and the historical mold image with a similarity greater than a threshold is selected as the mold image matching the injection mold to be simulated.

[0067] S6: Simulating according to the optimal adjustment parameter set and candidate adjustment parameter set of the injection molding parameters corresponding to the mold image in the matched historical data.

[0068] According to the matched historical mold images, the best adjustment parameter set and candidate adjustment parameter set corresponding to the historical mold image with the greatest similarity can be used for injection molding simulation, or the union of the best adjustment parameter sets of the historical mold images with a similarity greater than a threshold value can be used as a new best adjustment parameter set, and the union of the candidate adjustment parameter sets can be used as a new candidate adjustment parameter set, and the new best adjustment parameter set and the new candidate adjustment parameter set can be used for injection molding simulation. During the simulation, the last injection molding parameter is determined according to the adjustment sequence of the injection molding parameters obtained in step S2, a group of parameter values ​​of each injection molding parameter is selected from the best adjustment parameter set and the candidate adjustment parameter set of each injection molding parameter, and the injection molding results are recorded, and the parameter values ​​of the injection molding parameters before the last injection molding parameter are adjusted in sequence according to the adjustment sequence and the injection molding results are recorded, and after simulating the combination of all parameter values ​​of each injection molding parameter in the best adjustment parameter set and the candidate adjustment parameter set, the injection molding parameter with the best injection molding effect is determined.

[0069] For example, if the order of injection molding parameters temperature, pressure and push speed is pressure, push speed, temperature, then temperature is the last injection molding parameter. First, select a parameter value from the optimal adjustment parameter set and candidate adjustment parameter set of pressure, push speed and temperature to form a set of simulated injection molding parameter data, record the injection molding results, and then adjust the parameter value of pressure first, then adjust the parameter value of push speed in sequence, and record the injection molding results. After simulating the combination of all parameter values ​​of each injection molding parameter in the optimal adjustment parameter set and candidate adjustment parameter set of pressure and push speed, determine the injection molding parameters with the best injection molding effect.

[0070] Figure 2 It is a structural diagram of the box-matching injection molding filling simulation system based on machine learning in this embodiment.

[0071] like Figure 2 As shown, the box-matching injection molding filling simulation system based on machine learning includes: a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the box-matching injection molding filling simulation method based on machine learning according to the first aspect of the present invention is implemented.

[0072] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that may be stored or otherwise maintained by such a computer-readable medium.

[0073] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly defined.

[0074] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A box-type injection molding simulation method based on machine learning, characterized in that: include: Acquire historical data about box-mounted injection molding, the historical data including an injection mold image, corresponding injection molding parameters and injection molding results, the injection molding parameters including temperature, pressure and advancement speed; Obtaining the adjustment order of each injection molding parameter, including: calculating the mutual information entropy between the injection molding parameters and the injection molding result, including: calculating the difference between the information entropy of the injection molding result and the conditional entropy of the injection molding result under the condition of determining the injection molding parameters, and using the difference as the mutual information entropy of the injection molding parameter and the injection molding result; determining the determinacy of each injection molding parameter on the injection molding result according to the mutual information entropy, and determining the adjustment order of each injection molding parameter in descending order according to the size of the determinacy, wherein the determinacy is positively correlated with the mutual information entropy; Calculate the data distribution of each injection molding parameter when the injection molding result is qualified or unqualified, and obtain the qualified adjustment interval and unqualified adjustment interval of each injection molding parameter; Calculate the possibility of each parameter value in the injection molding parameter qualified adjustment interval meeting the injection molding qualification, determine the parameter value with the injection molding qualification probability greater than the threshold as the optimal adjustment parameter set, and the remaining parameter values ​​as the candidate adjustment parameter set; Taking pictures of the injection mold image to be simulated, and determining the mold image in the historical data that matches the mold image; Determine the optimal adjustment parameter set and candidate adjustment parameter set of the corresponding injection molding parameters according to the mold image in the historical data, determine the value of the last injection molding parameter according to the adjustment sequence, and adjust the injection molding parameters before the last in the optimal adjustment parameter set and candidate adjustment parameter set for simulation.

2. The box-matching injection molding filling simulation method based on machine learning according to claim 1 is characterized in that: The injection molding parameters also include cooling rate.

3. The box-matching injection molding filling simulation method based on machine learning according to claim 1 is characterized in that: Calculate the data distribution of each injection molding parameter when the injection molding results are qualified and unqualified, and obtain the qualified adjustment range and unqualified adjustment range of each injection molding parameter, including: If the data distribution is normal, take the mean of qualified injection molding parameter data as its first center point, according to 3 The qualified adjustment range of the injection molding parameters is obtained by the rule. ,in Indicates the first center point of the injection molding parameters, Represents the first variance of the injection molding parameters; The mean of the unqualified injection molding parameter data is taken as the second center point. The unqualified adjustment range of the injection molding parameters is obtained by the rule. ,in The second center point representing the injection molding parameters, Represents the second variance of the injection molding parameters.

4. The box-matching injection molding filling simulation method based on machine learning according to claim 1 is characterized in that: Calculate the data distribution of each injection molding parameter when the injection molding results are qualified and unqualified, and obtain the qualified adjustment range and unqualified adjustment range of each injection molding parameter, including: If the data distribution is not a normal distribution, calculate the third center point of the injection molding parameters, take the third center point as the center of the circle, and when 90% of the data falls within the circle, the radius at this time is R1, and the qualified adjustment interval is , where the third center point The median, mode and mean of the qualified injection molding parameter data are averaged; the fourth center point of the injection molding parameter is calculated, and the fourth center point is taken as the center of the circle. When 90% of the data falls into the circle, the radius at this time is , the unqualified adjustment interval is , where the fourth center point The average is the sum of the median, mode and mean of the unqualified injection molding parameter data.

5. The box-matching injection molding filling simulation method based on machine learning according to claim 1 is characterized in that: The possibility of each parameter value in the injection molding parameter qualified adjustment interval meeting the injection molding qualification is calculated, and the parameter value with the injection molding qualification probability greater than the threshold is determined as the optimal adjustment parameter set, and the remaining parameter values ​​are candidate adjustment parameter sets, including: Calculate injection molding parameters Possibility of meeting the requirements of injection molding : Indicates the value of the injection molding parameter. Indicates the center point of the unqualified injection molding parameter x, Indicates the adjustment range of unqualified injection molding parameters. Represents the center point of the qualified injection molding parameter x, Indicates the adjustment range of qualified injection molding parameters.

6. The box-type injection molding simulation system based on machine learning is characterized by: include: A memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the box-assembly injection molding filling simulation method based on machine learning as described in any one of claims 1 to 5 is implemented.

Images