A method and device for predicting quality of industrial injection molding products

Abstract

The present application discloses a method and device for predicting the quality of industrial injection molding products. The technical solution of the present application is first based on historical production record data of real injection molding product samples, and the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample are extracted from the historical production record data, and the product sample data are divided into several groups according to different injection molding product samples as training data. Then, model training is performed based on a preset initial fuzzy inference model and a heuristic training optimization algorithm to obtain a quality prediction model. In actual application, by obtaining the injection molding machine process setting parameters of actual production conditions and inputting them into the quality prediction model, the prior knowledge of the mapping relationship between the process setting parameters and product defects contained in the quality prediction model can be used to perform prediction operations to obtain injection molding product quality prediction results, without the need for a long debugging and trial-and-error process, thereby reducing the time cost and material cost of debugging the injection molding machine before production.

Description

Technical Field

[0001] The present application relates to the field of computer and industrial injection molding production technology, and in particular to a method and device for predicting the quality of industrial injection molding products. Background Art

[0002] Industrial plastic products are widely used in many fields due to their light weight and high strength, such as daily life, aviation, etc. Injection molding is one of the most important industrial plastic processing methods, with high efficiency and manufacturability.

[0003] Although there are a wide range of setting parameters and sensors that feedback various injection molding process parameters in the injection molding machine, the influence of interference variables still exists in the ongoing production process, causing repeated fluctuations in the injection molding process, affecting process stability, and thus leading to deviations in the quality of injection molded products. And due to the diversity of injection molding processes, the setting of injection molding process parameters mainly relies on the experience of on-site personnel for debugging. This method requires a long time of trial and error, and there is a technical problem of high trial and error costs. Summary of the invention

[0004] The present application provides a method and device for predicting the quality of industrial injection molding products, which is used to solve the technical problem that the debugging of injection molding process parameters in the existing industrial injection molding product production needs to rely on the experience and trial and error of on-site personnel, and there is a high trial and error cost.

[0005] In order to solve the above technical problems, the first aspect of the present application provides a method for predicting the quality of industrial injection molding products, comprising:

[0006] Obtain historical production record data for injection molding product samples;

[0007] Extracting the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample according to the historical production record data, and classifying the various injection molding machine process setting parameters and defect evaluation information according to different injection molding product sample individuals to obtain several groups of product sample data;

[0008] Based on a preset initial fuzzy reasoning model and a heuristic training optimization algorithm, the product sample data is used as training data, and the initial fuzzy reasoning model is trained in combination with the heuristic training optimization algorithm to obtain a quality prediction model;

[0009] The process setting parameters of the injection molding machine under actual production conditions are obtained, and the process setting parameters of the injection molding machine are input into the quality prediction model, so as to obtain the injection molding product quality prediction result through the operation of the quality prediction model.

[0010] Preferably, the process setting parameters of the injection molding machine specifically include: injection position, injection pressure, holding time, pressure transfer position, cooling time, injection temperature and three barrel section temperatures.

[0011] Preferably, based on a preset initial fuzzy inference model and a heuristic training optimization algorithm, the product sample data is used as training data, and the initial fuzzy inference model is trained in combination with the heuristic training optimization algorithm to obtain a quality prediction model, specifically including:

[0012] Using the product sample data as training data, the initial fuzzy inference model is trained, and when the convergence degree of the training result of the initial fuzzy inference model meets the preset conditions, an initial quality prediction model is obtained;

[0013] The quality prediction model is iteratively optimized for parameters using the heuristic training optimization algorithm, and when a preset number of iterations is reached, the current quality prediction model is output.

[0014] Preferably, the fuzzy reasoning model is specifically: ANFIS fuzzy reasoning model; the heuristic training optimization algorithm is specifically: IDTBO optimization algorithm.

[0015] Preferably, the iterative optimization of parameters of the quality prediction model by the heuristic training optimization algorithm and, when a preset number of iterations is reached, outputting the current quality prediction model specifically includes:

[0016] Initializing multiple groups of IDTBO populations according to the parameter combination of the quality prediction model;

[0017] According to the fitness of different IDTBO populations, each IDTBO population is divided into a coach population and a student population;

[0018] According to the coach population and the student population, a global search is performed based on a preset adaptive change operator to calculate a better solution of the quality prediction model in the search space;

[0019] According to a preset first student position update formula, the position of the student population is updated, and based on the fitness of the updated student population, the student population whose fitness is lower than the preset condition is eliminated;

[0020] In the local search phase, the positions of the student population are updated according to the preset second student position update formula, and based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated;

[0021] According to the Euclidean distance between the current position of the student population and the position of the candidate solution, the neighboring population of the student population is determined, and then according to the preset neighboring individual learning and searching method, a number of candidate solutions are determined from different neighboring individuals and individuals randomly selected from the population and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than the preset condition is eliminated;

[0022] The process of iteratively executing a global search based on the first student position update formula until several candidate solutions are determined and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than a preset condition is eliminated. When the number of iterations reaches the preset number of iterations, the current quality prediction model is output.

[0023] At the same time, the second aspect of the present application provides an industrial injection molding product quality prediction device, comprising:

[0024] A historical production record acquisition unit, used to acquire historical production record data of injection molding product samples;

[0025] A sample data generating unit is used to extract the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample according to the historical production record data, and classify the various injection molding machine process setting parameters and defect evaluation information according to different injection molding product sample individuals to obtain several groups of product sample data;

[0026] A model training unit, for performing model training on the initial fuzzy reasoning model based on a preset initial fuzzy reasoning model and a heuristic training optimization algorithm, using the product sample data as training data, and combining the heuristic training optimization algorithm to obtain a quality prediction model;

[0027] The product quality prediction execution unit is used to obtain the injection molding machine process setting parameters of the actual production conditions, and input the injection molding machine process setting parameters into the quality prediction model to obtain the injection molding product quality prediction result through the operation of the quality prediction model.

[0028] Preferably, the process setting parameters of the injection molding machine specifically include: injection position, injection pressure, holding time, pressure transfer position, cooling time, injection temperature and three barrel section temperatures.

[0029] Preferably, the model training unit specifically includes:

[0030] An initial model training subunit is used to perform model training on the initial fuzzy inference model using the product sample data as training data, and to obtain an initial quality prediction model when the convergence degree of the training result of the initial fuzzy inference model meets a preset condition;

[0031] The initial model optimization subunit is used to perform iterative parameter optimization on the quality prediction model through the heuristic training optimization algorithm, and output the current quality prediction model when a preset number of iterations is reached.

[0032] Preferably, the fuzzy reasoning model is specifically: ANFIS fuzzy reasoning model; the heuristic training optimization algorithm is specifically: IDTBO optimization algorithm.

[0033] Preferably, the initial model optimization subunit is specifically used for:

[0034] Initializing multiple groups of IDTBO populations according to the parameter combination of the quality prediction model;

[0035] According to the fitness of different IDTBO populations, each IDTBO population is divided into a coach population and a student population;

[0036] According to the coach population and the student population, a global search is performed according to a preset first student position update formula to calculate a better solution of the quality prediction model in the search space, the position of the student population is updated, and then based on the fitness of the updated student population, the student population with a fitness lower than a preset condition is eliminated;

[0037] In the local search phase, the position of the student population is updated according to the preset second student position update formula, and then based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated;

[0038] According to a preset third student position update formula, the position of the student population is updated again, and based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated, wherein the third student position update formula is a DTBO algorithm after replacing the constant R with a sine variable;

[0039] According to the Euclidean distance between the current position of the student population and the position of the candidate solution, the neighboring population of the student population is determined, and then according to the preset neighboring individual learning and searching method, a number of candidate solutions are determined from different neighboring individuals and individuals randomly selected from the population and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than the preset condition is eliminated;

[0040] The process of iteratively executing a global search based on the first student position update formula until several candidate solutions are determined and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than a preset condition is eliminated. When the number of iterations reaches the preset number of iterations, the current quality prediction model is output.

[0041] It can be seen from the above technical solutions that this application has the following advantages:

[0042] The technical solution of the present application is first based on the historical production record data of real injection molding product samples, and the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample are extracted from the historical production record data, and the product sample data are divided into several groups according to different injection molding product samples as training data, and then the model is trained based on the preset initial fuzzy inference model and heuristic training optimization algorithm to obtain a quality prediction model. In actual application, by obtaining the injection molding machine process setting parameters of the actual production conditions and inputting them into the quality prediction model, the prior knowledge of the mapping relationship between the process setting parameters and product defects contained in the quality prediction model can be used to perform prediction operations to obtain the injection molding product quality prediction results, without the need for a long debugging and trial and error process, thereby reducing the time cost and material cost of debugging the injection molding machine before production. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0044] Figure 1 A schematic flow chart of an embodiment of a method for predicting quality of industrial injection molding products provided in this application;

[0045] Figure 2 A schematic diagram of a flow chart of an embodiment of step 103 of a method for predicting quality of industrial injection molding products provided in this application;

[0046] Figure 3 A schematic structural diagram of an embodiment of an industrial injection molding product quality prediction device provided in this application. DETAILED DESCRIPTION

[0047] The present application provides a method and device for predicting the quality of industrial injection molding products, which is used to solve the technical problem that the debugging of injection molding process parameters in the existing industrial injection molding product production needs to rely on the experience and trial and error of on-site personnel, and there is a high trial and error cost.

[0048] In order to make the purpose, features, and advantages of the invention of this application more obvious and easy to understand, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described below are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] First, a detailed description of an embodiment of an industrial injection molding product quality prediction method provided by the present application is as follows:

[0050] See also Figure 1 , this embodiment provides a method for predicting the quality of industrial injection molding products, comprising:

[0051] Step 101, obtaining historical production record data of injection molded product samples;

[0052] It should be noted that first, the historical production record data of real injection molding product samples are obtained through the cloud platform or database, including the process setting parameters of the injection molding machine, sensor molding parameters, historical parameter settings, and product quality information.

[0053] Step 102: extracting the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample according to the historical production record data, and classifying the various injection molding machine process setting parameters and defect evaluation information according to different injection molding product sample individuals to obtain several groups of product sample data;

[0054] It should be noted that, according to the historical production record data obtained in the previous step, the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample are extracted respectively, wherein the injection molding machine process setting parameters mentioned in this embodiment are several important influencing parameters for injection molding process production, specifically including the following items: injection position, injection pressure, holding time, transfer holding position, cooling time, injection temperature and three barrel section temperatures; the defect evaluation information is used to measure the degree of shrinkage and warpage defects of the injection molding product, and the shrinkage and warpage degree can be divided into 3 levels: normal, slight shrinkage, and severe shrinkage;

[0055] Next, the above-mentioned injection molding machine process setting parameters and defect evaluation information are classified according to different injection molding product sample individuals to obtain several groups of product sample data, each group of product sample data represents a combination of injection molding machine process setting parameters and defect evaluation information of an injection molding product sample individual.

[0056] Step 103: Based on the preset initial fuzzy reasoning model and the heuristic training optimization algorithm, the product sample data is used as training data, and the initial fuzzy reasoning model is trained in combination with the heuristic training optimization algorithm to obtain a quality prediction model;

[0057] It should be noted that based on the product sample data obtained in the previous step, it is used as training data, and combined with the heuristic training optimization algorithm to train and optimize the initial fuzzy reasoning model, so as to obtain the corresponding quality prediction model.

[0058] Furthermore, the fuzzy reasoning model mentioned in this embodiment is specifically: ANFIS fuzzy reasoning model; the heuristic training optimization algorithm is specifically: IDTBO optimization algorithm.

[0059] More specifically, see Figure 2 The quality prediction model building process in step 103 of this embodiment can be referred to as follows:

[0060] Step 1031: Using product sample data as training data, the initial fuzzy inference model is trained. When the convergence degree of the training result of the initial fuzzy inference model meets the preset conditions, an initial quality prediction model is obtained.

[0061] It should be noted that the collected process parameters are used as input and the defect severity evaluated by expert experience is used as output. They are paired into input-output pairs as the training set of the ANFIS model;

[0062] Set the appropriate number and type of membership functions, and set the fuzzy inference system in the ANFIS model to the Mamdani model. Construct the objective function of the ANFIS model so that ANFIS can continuously approach the corresponding output defect severity through fuzzy inference of input parameters. And give the number of ANFIS model training iterations, and the optimization method is back propagation.

[0063] The training set is input into the ANFIS model for iterative training. The model fitting is repeated until the training result or the objective function budget is reached, and then a preliminary ANFIS model is obtained.

[0064] More specifically, using represents the training data set, First, the training data set is sampled to obtain the real injection molding process parameter data samples, which are input into the ANFIS model. In the first layer of the ANFIS model, the fuzzy membership value of each input is calculated by the previously set membership function, which is between 0 and 1. The calculation formula is shown in (1):

[0065]

[0066] in and Yes Input The mean and standard deviation parameters of the j-th Gaussian membership function are the pre-factors of ANFIS.

[0067] After entering the second layer, the product of all the membership values ​​obtained in the first layer is calculated using formula (2) and named the trigger strength of rule R(j).

[0068]

[0069] The main function of the third layer is to normalize the trigger strength (the output of the second layer), that is, the ratio of the trigger strength of a single rule to that of all input rules, calculated using formula (3).

[0070]

[0071] The original input and the subsequent parameters and normalized trigger strength are used to calculate the output. Formula (4) is used to calculate the contribution of each rule to the ANFIS result.

[0072]

[0073] The last layer calculates the sum of all inputs from the fourth layer through formula (5) and gives the final result.

[0074]

[0075] This results in a fuzzy rule-based ANFIS model, where the two inputs and and a linear output function ( Some examples of fuzzy rules of ) are shown below:

[0076] Rule #1. IF is AND is THEN

[0077]

[0078] Rule #2. IF is AND is THEN

[0079]

[0080] Step 1032: perform iterative optimization of parameters of the quality prediction model through a heuristic training optimization algorithm, and when a preset number of iterations is reached, output the current quality prediction model.

[0081] It should be noted that after step 1031, the parameters of the preliminarily trained ANFIS model are further optimized using the improved IDTBO optimization algorithm. Through the optimization of the IDTBO optimization algorithm, the quality prediction model can provide higher accuracy and better stability in the injection molding quality prediction scenario. The IDTBO population consists of individual drivers and individual trainees. After initialization according to the following formula (6), the individuals with higher fitness values ​​are used as coaches, and the remaining individuals are used as trainees.

[0082]

[0083] in is the i-th candidate solution, is the jth variable of the ith candidate solution. N is the population size of DTBO, m is the number of problem variables. r is a random number between [0,1]. and are the upper and lower limits of the j-th problem variable. Members serve as driving instructors, as shown in formula (7):

[0084]

[0085] in For the coaching matrix, For the i-th coach, is the jth dimension of the i-th coach. is the number of driving instructors, t is the current iteration number, and T is the maximum iteration number.

[0086] After initializing and determining the individual drivers and trainees in IDTBO, we first perform a global search in the search space using the adaptive operator in the first stage to calculate the optimal solution of ANFIS in the search space. The adaptive operator is designed as follows:

[0087]

[0088] Where t is the current iteration number and T is the maximum iteration number.

[0089] Combined with the practice of the second-stage students under the guidance of the coach, the positions of the second-stage students are updated according to formula (9).

[0090]

[0091] in The values ​​are from a continuous uniform distribution [ 0 , γ ] A random number selected from , and is the upper and lower limits of the j-th problem variable. When the current number of iterations reaches the maximum number of iterations .

[0092] During local search, this application replaces the constant in the original DTBO algorithm with The scaling factor is The sine variable is used to expand the local search range to improve the local search capability. The position of the trainees in the third stage is updated using formula (10).

[0093]

[0094] This application proposes a new ONL search strategy as the fourth stage of IDTBO. In the prototype DTBO, for each driving learner, the position is updated under the guidance of the instructor. This approach leads to slow population convergence and easy to fall into local optimal solutions. To address this weakness, the ONL search strategy is adopted as the fourth stage to provide learners with the ability to learn from other excellent learners and improve their utilization skills.

[0095] Learning the ONL search strategy for adjacent individuals generates another candidate solution This candidate solution is learned from different neighboring individuals and individuals randomly selected from the population. Assume that the original position before the three-stage IDTBO is , the candidate solution position obtained after three stages is First, the Euclidean distance between the current position and the candidate solution position is calculated, and then the radius is calculated using formula (11) :

[0096]

[0097] Where t is the current iteration number. Then, use express The neighboring individuals are considered by the radius It is calculated by formula (12).

[0098]

[0099] in for and The Euclidean distance between them is constructed After the neighboring individuals are obtained, multiple neighboring individuals are learned according to formula (13).

[0100]

[0101] in The dth dimension is the neighboring individuals A random individual selected from and individuals randomly selected from the population The dth dimension of is calculated. r is a random number between [0,1].

[0102] And when selecting and updating the population at each stage, the elimination strategy is used to update the population. Its fitness , together with the original population and the fitness value of each individual, form a synthetic matrix as shown in formula (14): .

[0103]

[0104] The matrix Reorder them according to the fitness value and take the first m variables as the new population, so as to realize the test elimination strategy to update the population and keep more excellent information. This elimination strategy is used to update the population in all stages of IDTBO.

[0105] Finally, after executing the above four stages for all individuals, the iteration counter increases by 1 until the predetermined number of iterations is reached. After the iteration is completed, the optimized ANFIS model is obtained, and the ANFIS model at this time can achieve quality prediction of injection molding products with higher accuracy and stability.

[0106] Step 104: Obtain the process setting parameters of the injection molding machine under actual production conditions, and input the process setting parameters of the injection molding machine into the quality prediction model to obtain the injection molding product quality prediction result through the operation of the quality prediction model.

[0107] It should be noted that, in actual application, by obtaining the process setting parameters of the injection molding machine under actual production conditions and inputting them into the quality prediction model, the prior knowledge of the mapping relationship between the process setting parameters and product defects contained in the quality prediction model can be used to perform prediction operations, and the neural network and the fuzzy reasoning system can be combined, and the reasoning ability of the neural network and the interpretability of the fuzzy system can be combined to achieve high-precision and high-stability injection molding product quality prediction results with a small amount of data. The obtained quality prediction results can be fed back to engineers to provide them with a reference for adjusting the process parameters of the injection molding machine, which can shorten the debugging and trial-and-error process, and reduce the production costs and time of engineers' manual observation and machine adjustment.

[0108] The above is a detailed description of an embodiment of an industrial injection molding product quality prediction method provided by the present application. The following is a detailed description of an embodiment of an industrial injection molding product quality prediction device provided by the present application.

[0109] See also Figure 3 , this embodiment provides an industrial injection molding product quality prediction device, comprising:

[0110] A historical production record acquisition unit 201 is used to acquire historical production record data of injection molding product samples;

[0111] The sample data generating unit 202 is used to extract the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample according to the historical production record data, and classify the various injection molding machine process setting parameters and defect evaluation information according to different injection molding product sample individuals to obtain several groups of product sample data;

[0112] The model training unit 203 is used to perform model training on the initial fuzzy reasoning model based on a preset initial fuzzy reasoning model and a heuristic training optimization algorithm, using product sample data as training data and combining the heuristic training optimization algorithm to obtain a quality prediction model;

[0113] The product quality prediction execution unit 204 is used to obtain the process setting parameters of the injection molding machine under actual production conditions, and input the process setting parameters of the injection molding machine into the quality prediction model to obtain the injection molding product quality prediction result through the operation of the quality prediction model.

[0114] Furthermore, the process setting parameters of the injection molding machine specifically include: injection position, injection pressure, holding time, transfer holding position, cooling time, injection temperature and three barrel section temperatures.

[0115] Furthermore, the model training unit 203 specifically includes:

[0116] The initial model training subunit 2031 is used to perform model training on the initial fuzzy inference model using product sample data as training data, and when the convergence degree of the training result of the initial fuzzy inference model meets the preset conditions, an initial quality prediction model is obtained;

[0117] The initial model optimization subunit 2032 is used to perform iterative optimization of parameters of the quality prediction model through a heuristic training optimization algorithm, and output the current quality prediction model when a preset number of iterations is reached.

[0118] Furthermore, the fuzzy reasoning model is specifically: ANFIS fuzzy reasoning model; the heuristic training optimization algorithm is specifically: IDTBO optimization algorithm.

[0119] Furthermore, the initial model optimization subunit is specifically used for:

[0120] Initialize multiple groups of IDTBO populations according to the parameter combination of the quality prediction model;

[0121] According to the fitness of different IDTBO populations, each IDTBO population is divided into a coach population and a student population;

[0122] According to the coach population and the student population, a global search is performed based on a preset adaptive change operator to calculate a better solution of the quality prediction model in the search space;

[0123] According to the preset first student position update formula, the position of the student population is updated, and based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated;

[0124] In the local search phase, the position of the student population is updated according to the preset second student position update formula, and based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated;

[0125] According to the Euclidean distance between the current position of the student population and the position of the candidate solution, the neighboring population of the student population is determined, and then according to the preset neighboring individual learning and searching method, several candidate solutions are determined from different neighboring individuals and individuals randomly selected from the population and the student population is updated, and based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated;

[0126] The process of iteratively executing a global search based on the first student position update formula until several candidate solutions are determined and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than a preset condition is eliminated. When the number of iterations reaches the preset number of iterations, the current quality prediction model is output.

[0127] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the terminals, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0128] In the several embodiments provided in the present application, it should be understood that the disclosed terminals, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0129] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0130] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0131] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0132] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0133] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

[0134] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for predicting the quality of industrial injection molding products, characterized in that: include: Obtain historical production record data for injection molding product samples; Extracting the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample according to the historical production record data, and classifying the various injection molding machine process setting parameters and defect evaluation information according to different injection molding product sample individuals to obtain several groups of product sample data; Using the product sample data as training data, the initial fuzzy inference model is trained, and when the convergence degree of the training result of the initial fuzzy inference model meets the preset conditions, an initial quality prediction model is obtained; Initializing multiple groups of IDTBO populations according to the parameter combination of the quality prediction model; According to the fitness of different IDTBO populations, each IDTBO population is divided into a coach population and a student population; According to the coach population and the student population, a global search is performed according to a preset first student position update formula to calculate a better solution of the quality prediction model in the search space, the position of the student population is updated, and then based on the fitness of the updated student population, the student population with a fitness lower than a preset condition is eliminated; In the local search phase, the position of the student population is updated according to the preset second student position update formula, and then based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated; According to a preset third student position update formula, the position of the student population is updated again, and based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated, wherein the third student position update formula is a DTBO algorithm after replacing the constant R with a sine variable; According to the Euclidean distance between the current position of the student population and the position of the candidate solution, the neighboring population of the student population is determined, and then according to the preset neighboring individual learning and searching method, a number of candidate solutions are determined from different neighboring individuals and individuals randomly selected from the population and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than the preset condition is eliminated; Iteratively executing a global search based on the first student position update formula until a number of candidate solutions are determined and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than a preset condition is eliminated, and when the number of iterations reaches a preset number of iterations, the current quality prediction model is output; The process setting parameters of the injection molding machine under actual production conditions are obtained, and the process setting parameters of the injection molding machine are input into the quality prediction model, so as to obtain the injection molding product quality prediction result through the operation of the quality prediction model.

2. The method for predicting the quality of industrial injection molding products according to claim 1, characterized in that: The process setting parameters of the injection molding machine specifically include: injection position, injection pressure, holding time, transfer holding position, cooling time, injection temperature and three barrel section temperatures.

3. The method for predicting the quality of industrial injection molding products according to claim 1, characterized in that: The initial fuzzy reasoning model is specifically: ANFIS fuzzy reasoning model.

4. An industrial injection molding product quality prediction device, characterized in that: include: A historical production record acquisition unit, used to acquire historical production record data of injection molding product samples; A sample data generating unit is used to extract the injection molding machine process setting parameters and defect evaluation information corresponding to each injection molding product sample according to the historical production record data, and classify the various injection molding machine process setting parameters and defect evaluation information according to different injection molding product sample individuals to obtain several groups of product sample data; A model training unit, for performing model training on the initial fuzzy reasoning model based on a preset initial fuzzy reasoning model and a heuristic training optimization algorithm, using the product sample data as training data, and combining the heuristic training optimization algorithm to obtain a quality prediction model; A product quality prediction execution unit, used for obtaining the process setting parameters of the injection molding machine under actual production conditions, and inputting the process setting parameters of the injection molding machine into the quality prediction model, so as to obtain the injection molding product quality prediction result through the operation of the quality prediction model; The model training unit specifically includes: An initial model training subunit is used to perform model training on the initial fuzzy inference model using the product sample data as training data, and to obtain an initial quality prediction model when the convergence degree of the training result of the initial fuzzy inference model meets a preset condition; The initial model optimization subunit is specifically used for: Initializing multiple groups of IDTBO populations according to the parameter combination of the quality prediction model; According to the fitness of different IDTBO populations, each IDTBO population is divided into a coach population and a student population; According to the coach population and the student population, a global search is performed based on a preset adaptive change operator to calculate a better solution of the quality prediction model in the search space; According to a preset first student position update formula, the position of the student population is updated, and based on the fitness of the updated student population, the student population whose fitness is lower than the preset condition is eliminated; In the local search phase, the positions of the student population are updated according to the preset second student position update formula, and based on the fitness of the updated student population, the student population with fitness lower than the preset condition is eliminated; According to the Euclidean distance between the current position of the student population and the position of the candidate solution, the neighboring population of the student population is determined, and then according to the preset neighboring individual learning and searching method, a number of candidate solutions are determined from different neighboring individuals and individuals randomly selected from the population and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than the preset condition is eliminated; The process of iteratively executing a global search based on the first student position update formula until several candidate solutions are determined and the student population is updated, and based on the fitness of the updated student population, the student population with a fitness lower than a preset condition is eliminated. When the number of iterations reaches the preset number of iterations, the current quality prediction model is output.

5. The industrial injection molding product quality prediction device according to claim 4, characterized in that: The process setting parameters of the injection molding machine specifically include: injection position, injection pressure, holding time, transfer holding position, cooling time, injection temperature and three barrel section temperatures.

6. The industrial injection molding product quality prediction device according to claim 4, characterized in that: The fuzzy reasoning model is specifically: ANFIS fuzzy reasoning model.

Images