CN111319206B - Parameter optimization method and device in injection molding system - Google Patents
Parameter optimization method and device in injection molding system Download PDFInfo
- Publication number
- CN111319206B CN111319206B CN201811521647.3A CN201811521647A CN111319206B CN 111319206 B CN111319206 B CN 111319206B CN 201811521647 A CN201811521647 A CN 201811521647A CN 111319206 B CN111319206 B CN 111319206B
- Authority
- CN
- China
- Prior art keywords
- injection molding
- setting parameter
- molding system
- feedback data
- parameters
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C45/00—Injection moulding, i.e. forcing the required volume of moulding material through a nozzle into a closed mould; Apparatus therefor
- B29C45/17—Component parts, details or accessories; Auxiliary operations
- B29C45/76—Measuring, controlling or regulating
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C2945/00—Indexing scheme relating to injection moulding, i.e. forcing the required volume of moulding material through a nozzle into a closed mould
- B29C2945/76—Measuring, controlling or regulating
- B29C2945/76929—Controlling method
- B29C2945/76979—Using a neural network
Landscapes
- Engineering & Computer Science (AREA)
- Manufacturing & Machinery (AREA)
- Mechanical Engineering (AREA)
- Injection Moulding Of Plastics Or The Like (AREA)
Abstract
The embodiment of the invention provides a parameter optimization method and device in an injection molding system. The method comprises the following steps: acquiring a system description of the injection molding system, which is called a first system description; determining at least one second system description match that matches the first system description; acquiring a first setting parameter for producing the current product according to the setting parameter and the feedback data corresponding to the at least one second system description; and automatically adjusting and optimizing the production parameters of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system. The purposes of automatically acquiring the first setting parameter of the production parameter of each device in the injection molding system and automatically optimizing the setting parameter are achieved, the production efficiency is improved, and the defective rate is reduced.
Description
Technical Field
The embodiment of the invention relates to the field of intelligent manufacturing, in particular to a parameter optimization method and device in an injection molding system.
Background
The manufacturing industry directly reflects the productivity level of a country, is an important factor for distinguishing developing countries and developed countries, and with continuous innovation in the fields of artificial intelligence technology, block chain technology, Internet of things technology, 5G and the like, intelligent manufacturing becomes the development direction of the manufacturing industry. The intellectualization of the injection molding equipment as a typical large-scale device is not slow. The injection molding system is composed of a plurality of devices, and parameters of the devices in the injection molding system are adjusted, so that the devices in the injection molding system are coordinated and matched, which is the most core content for achieving the optimal production.
However, currently, the adjustment of parameters of injection molding equipment is set independently by a single device, and there is no coordinated adjustment of parameters of devices involved in the entire injection molding system. And the adjustment of the parameters of each device in the injection molding system in the production of the product is completed manually according to the parameter adjustment recorded by the standard process card.
However, the standard process card only records the mode of individual key parameters, and the key parameters recorded in the standard process card are invariable, but under the condition of long-time production, the quality of products produced by equipment adjusted according to the key parameters recorded in the standard process card is unstable due to the abrasion of the equipment, and the defective rate is high.
Disclosure of Invention
The embodiment of the invention provides a parameter optimization method and device in an injection molding system, which are used for automatically acquiring first setting parameters of production parameters of each device in the injection molding system and automatically optimizing the setting parameters
In a first aspect, an embodiment of the present invention provides a parameter optimization method in an injection molding system, including:
obtaining a first system description of the injection molding system, the first system description being a basic system case for each device in the injection molding system in relation to a current product production, wherein the system description comprises at least one of: production product information, injection molding equipment information, mold information, various peripheral auxiliary machine information and environment information;
determining at least one second system description matched with the first system description, wherein the second system description is stored in a cloud and/or a local database and is produced by each device in the injection molding system;
acquiring a first setting parameter for producing the current product according to a setting parameter and feedback data corresponding to the at least one second system description, wherein the setting parameter is a production parameter of each device in the injection molding system when the product is produced, and the feedback data is feedback data of a sensor in each device in the injection molding system and manual judgment data of a worker;
and optimizing the production parameters of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system.
Optionally, the optimizing the production parameter of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system includes:
acquiring a second setting parameter according to the first setting parameter and the neural network;
updating the production parameters of each device in the injection molding system by using the second setting parameters, and producing the product to be produced at present by using the updated injection molding system to obtain feedback data corresponding to the second setting parameters and the reward value of the product;
and training the neural network according to the feedback data of the second setting parameter and the reward value of the product so as to optimize the production parameter of each device in the injection molding system.
Optionally, before updating the production parameter of each device in the injection molding system by using the second setting parameter, the method includes:
comparing the second setting parameter with the setting parameter of the injection molding system model to obtain a reward value corresponding to the second setting parameter;
and pre-training the neural network according to the reward value corresponding to the second setting parameter.
Optionally, the comparing the second setting parameter with the setting parameter of the injection molding system model to obtain the reward value corresponding to the second setting parameter includes:
calculating the cross entropy of the second setting parameter and the setting parameter of the injection molding system model;
acquiring feedback data corresponding to the setting parameters of the injection molding system model according to the cross entropy;
determining the feedback data as the feedback data corresponding to the second setting parameter;
the obtaining the reward value comprises: and acquiring the reward value according to the feedback data.
Optionally, the obtaining the bonus value according to the feedback data includes:
judging whether the product corresponding to the feedback data is a genuine product or not;
if so, the reward value corresponding to the second set parameter is the reward value when the product corresponding to the feedback data is a genuine product;
if not, the reward value corresponding to the second setting parameter is the reward value when the product corresponding to the feedback data is defective.
Optionally, the obtaining feedback data corresponding to the setting parameters of the injection molding system model according to the cross entropy includes:
acquiring feedback data corresponding to at least one setting parameter of the injection molding system model with the cross entropy smaller than a preset value;
calculating the deviation degree of the feedback data corresponding to the at least one setting parameter;
and determining the feedback data with the minimum deviation degree as the feedback data corresponding to the setting parameters of the injection molding system model.
Optionally, after optimizing the production parameter of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system, the method further includes:
and uploading the first system description, the second setting parameter and the feedback data corresponding to the second setting parameter to a cloud end or storing the feedback data in a local database.
In a second aspect, an embodiment of the present invention provides an apparatus for optimizing parameters in an injection molding system, including:
a first obtaining module for obtaining a system description of the injection molding system, referred to as a first system description, the first system description being a basic system situation of each device in the injection molding system in relation to a current product production, wherein the system description comprises at least one of: production product information, injection molding equipment information, mold information, various peripheral auxiliary machine information and environment information;
a matching module, configured to determine that at least one second system description matches the first system description, where the second system description is stored in a cloud and/or a local database of system descriptions of each device in the injection molding system during production;
the first obtaining module is further configured to obtain a first setting parameter for producing the current product according to a setting parameter and feedback data corresponding to the at least one second system description, where the setting parameter is a production parameter of each device in the injection molding system when the product is produced, and the feedback data is feedback data of a sensor in each device in the injection molding system and human judgment data of a worker;
and the adjusting and optimizing module is used for optimizing the production parameters of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system.
Optionally, the tuning module includes:
the second acquisition module is used for acquiring a second setting parameter according to the first setting parameter and the neural network;
the second obtaining module is further configured to update the production parameters of each device in the injection molding system by using the second setting parameters, and produce the product to be produced currently by using the updated injection molding system, so as to obtain feedback data corresponding to the second setting parameters and a reward value of the product;
and the first updating module is used for updating the neural network according to the feedback data of the second setting parameter and the reward value of the product so as to optimize the production parameter of each device in the injection molding system.
Optionally, before the second obtaining module, the method includes:
the third acquisition module is used for comparing the second setting parameter with the setting parameter of the injection molding system model to acquire a reward value corresponding to the second setting parameter;
and the training module is used for pre-training the neural network according to the reward value corresponding to the second setting parameter.
Optionally, the third obtaining module is specifically configured to:
calculating the cross entropy of the second setting parameter and the setting parameter of the injection molding system model;
acquiring feedback data corresponding to the setting parameters of the injection molding system model according to the cross entropy;
determining the feedback data as the feedback data corresponding to the second setting parameter;
and acquiring the reward value according to the feedback data.
Optionally, when the third obtaining module obtains the bonus value according to the feedback data, the third obtaining module is specifically configured to:
judging whether the product corresponding to the feedback data is a genuine product or not;
if so, the reward value corresponding to the second set parameter is the reward value when the product corresponding to the feedback data is a genuine product;
if not, the reward value corresponding to the second setting parameter is the reward value when the product corresponding to the feedback data is defective.
Optionally, when the third obtaining module obtains the feedback data corresponding to the setting parameter of the injection molding system model according to the cross entropy, the third obtaining module is specifically configured to: acquiring feedback data corresponding to at least one setting parameter of the injection molding system model with the cross entropy smaller than a preset value;
calculating the deviation degree of the feedback data corresponding to the at least one setting parameter;
and determining the feedback data with the minimum deviation degree as the feedback data corresponding to the setting parameters of the injection molding system model.
Optionally, after the tuning module, the tuning module further includes:
and the storage module is used for uploading the first system description, the second setting parameter and the feedback data corresponding to the second setting parameter to a cloud end or storing the feedback data in a local database.
In a third aspect, an embodiment of the present invention provides an apparatus for optimizing parameters in an injection molding system, where the apparatus includes: at least one processor and memory;
the memory stores computer-executable instructions; the at least one processor executes computer-executable instructions stored by the memory to perform the method of any one of the first aspect of the inventive embodiments.
In a fourth aspect, the present invention provides a computer-readable storage medium, in which program instructions are stored, and when the program instructions are executed by a processor, the method according to any one of the first aspect of the present invention is implemented.
In a fifth aspect, embodiments of the present application provide a program product, which includes a computer program stored in a readable storage medium, from which the computer program can be read by at least one processor of a parameter optimization device in an injection molding system, and the computer program can be executed by the at least one processor to cause the parameter optimization device in the injection molding system to implement an inventive embodiment method provided in any one of the first aspects of the inventive embodiments of the present application.
The embodiment of the invention provides a parameter optimization method and device in an injection molding system, which is called as a first system description by obtaining the system description of a product needing to be produced at present; determining at least one second system description match that matches the first system description; acquiring a first setting parameter for producing the product required to be produced currently according to the setting parameter and the feedback data corresponding to the at least one second system description; and automatically adjusting and optimizing the production parameters of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system. The purposes of automatically acquiring the first setting parameter of the production parameter of each device in the injection molding system and automatically optimizing the setting parameter are achieved, the production efficiency is improved, and the defective rate is reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flowchart illustrating a method for automatically adjusting parameters of an injection molding system according to an embodiment of the present invention;
FIG. 2 is a flowchart of step S104 according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of an automatic parameter tuning device of an injection molding system according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an automatic parameter tuning device of an injection molding system according to another embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a flowchart of a parameter optimization method in an injection molding system according to an embodiment of the present invention.
As shown in fig. 1, the method of this embodiment may include:
s101, obtaining a system description of a product which needs to be produced at present, and calling the system description as a first system description.
The first system description is an attribute value of each device in the injection molding system when a product to be produced currently is produced, and the system description may include at least one of production product information, injection molding device information, mold information, various peripheral auxiliary device information, and environment information, for example. Table 1 shows a system description of the injection molding machine. Wherein a "value" in the third column in table 1 represents an attribute value of each name in the first column of each row corresponding to the "value" in that row in the first system description; the "matching coefficient" in the fourth column indicates the degree of matching of the "value" in the system description with the "value" in the first system description when other products are produced.
TABLE 1 System description schematic of an injection molding machine
S102, determining at least one second system description matched with the first system description.
Finding a system description, called a second system description, in the cloud and/or the database, in which the degree of matching between the "value" in the system description and the "value" in the first system description satisfies the "matching coefficient" in the first system description, wherein at least one second system description can be found in the database.
Wherein the second system description is a system description stored in the database for each device in the injection molding system at the time of production.
In particular, the at least one second system description may be determined according to the following matching model:
the correspondence between the "value" and the "matching coefficient" in the first system description may be represented as (I)i,αi) Where I denotes a "value" in the third column of table 1, α denotes a "matching coefficient" in the fourth column of table 1, and I denotes the ith "value" or ith. Matching coefficient "
The correspondence between the "value" and the "matching coefficient" in the second system description may be expressed as
Wherein the content of the first and second substances,
indicating a "value",
denotes "matching coefficient", i denotes the ith "value".
Determining whether the second system description matches the first system description according to equation (1):
when the matching degree of any "value" in the second system description and the corresponding "value" in the first system description meets the formula (1), the second system description is illustrated to be matched with the first system description.
S103, obtaining a first setting parameter for producing the product needing to be produced at present according to the setting parameter and the feedback data corresponding to the at least one second system description.
The set parameters are production parameters of each device in the injection molding system when the product is produced, and the feedback data are feedback data of a sensor in each device in the injection molding system and data obtained by manual judgment of workers when the production parameters of each device in the injection molding system are current set parameters. Table 2 exemplarily shows production parameters of a feeder in the injection molding system, and table 3 exemplarily shows feedback data of a mold temperature controller in the injection molding system.
TABLE 2 production parameters of a feeder in an injection molding system
TABLE 3 mold temperature feedback data in injection molding systems
| Serial number | Name of variable | Value of variable | Relevance |
| 1 | Temperature of flow channel 1 | 220.5 | 50 |
| 2 | Temperature of flow channel 2 | 210.7 | 50 |
| 3 | Temperature of flow channel 3 | 215.3 | 50 |
| 4 | Temperature of flow channel 4 | 220.7 | 50 |
| 5 | Temperature of flow passage 5 | 220.5 | 50 |
| 6 | Temperature of flow channel 6 | 210.8 | 50 |
| 7 | Temperature of flow channel 7 | 215.3 | 50 |
| 8 | Temperature of flow channel 8 | 210.6 | 50 |
| 9 | Temperature of the flow channel 9 | 220.1 | 50 |
| 10 | Temperature of flow channel 10 | 220.3 | 50 |
| 11 | Temperature of the flow channel 11 | 210.4 | 50 |
| 12 | Temperature of the flow channel 12 | 215.4 | 50 |
Wherein the correlation in table 3 represents the relationship between the feedback data and the product quality.
Each second system description corresponds to M setting parameters, each setting parameter corresponds to N feedback data, wherein M is larger than 1, and N is larger than 1, so that each second system has (M x N) group data in common, and each group data comprises three parts, namely the system description, the setting parameters and the feedback data. Wherein the system description is indicated by the letter I, the setting parameters are indicated by the letter S, and the feedback data are indicated by the letter F. For example, a second system description I corresponds to two setting parameters S1、S2Each setting parameter corresponds to three feedback data F1、F2、F3Then, six sets of data can be composed: (I, S)1，F1)、(I，S1，F2)、(I，S1，F3)、(I，S2，F1)、(I，S2，F2)、(I，S2，F3)。
And deleting the invalid data meeting the deletion condition according to the deletion condition. Specifically, the deleting condition may be that, in N feedback data corresponding to any one setting parameter, when N is less than or equal to 5 (that is, the number of times of producing the product using the setting parameter is less than 5 times) and/or the product using the setting parameter cannot achieve continuous production of 5 genuine products, the setting parameter and the N feedback data corresponding to the setting parameter are deleted.
And calculating the deviation degree of each group of effective data according to the relevance, wherein the deviation degree is represented by W. The calculation formula of the deviation degree of each group of finite data is formula (2):
wherein, WiDenotes the degree of deviation, β, of the ith set of valid datajRepresenting the correlation of the jth parameter in the feedback data in any set of valid data, fijThe method comprises the steps of representing specific values of j-th parameters in feedback data in the ith group of effective data, representing the number of effective data by m, and representing the number of parameters contained in the feedback data in any group of effective data by n.
And comparing the deviation degree of each group of effective data with a first threshold value, and reserving the effective data with the deviation degree smaller than the first threshold value, namely the effective data.
And performing weighted average on parameters at the same position in the setting parameters in the actual effective data, and calculating the initial value of the parameter at the position in the initial setting parameters according to the deviation of each group of actual effective data. Wherein, the initial value of each parameter in the initial setting parameters can be calculated according to formula (3):
wherein s isqDenotes an initial value, t, of the q-th parameter of the initially set parameterspqDenotes a specific value of the q-th parameter among the setting parameters in the p-th group of actual effective data, WpDenotes the degree of deviation of the p-th set of actual effective data, x denotes the number of actual effective data,
and an worth coefficient ψ representing the p-th group of actual valid data.
It should be noted that the initial setting parameters in this step are also referred to as first setting parameters.
S104, optimizing the production parameters of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system.
Specifically, a second setting parameter is obtained and output from the first setting parameter according to the neural network, the production parameter of each device in the injection molding system is set as the second setting parameter, and product production is carried out, so that feedback data and reward values of the product production are obtained. The neural network is then trained based on the reward value, the second set of parameters, and the first set of parameters. And then inputting the second setting parameter as the first setting parameter into the trained neural network to obtain the second setting parameter. And training the neural network according to the process, and when the second setting parameter output by the trained neural network is used as the production parameter of each device in the injection molding system, the rate of certified products of the produced products reaches a preset value.
Optionally, fig. 2 is a flowchart of step S104 according to an embodiment of the present invention. As shown in fig. 2, the implementation manner of step S104 may be:
s1041, obtaining a second setting parameter according to the first setting parameter and the neural network.
And taking the first setting parameter as the input of the neural network, and obtaining the adjustment value of any parameter value in the first setting parameter after the parameter value is adjusted through the neural network. Wherein the adjustment of any one parameter value in the first setting parameter of the adjustment value contributes to the automatic tuning of the setting parameter. And acquiring at least one adjusting value with larger adjusting value according to an intelligent optimization algorithm or an optimization algorithm (such as an epsilon greedy algorithm), and executing corresponding parameter value adjustment so as to acquire a second setting parameter. The input of the neural network is referred to as a first setting parameter, and the output is referred to as a second setting parameter.
It should be noted that, when the neural network adjusts any one parameter value in the first setting parameters, the number of the adjusted parameter values each time cannot exceed the setting value, the size of the setting value is determined according to actual production, the upper and lower limits of the adjustment range for each parameter value are 10%, and if the adjusted parameter value exceeds the upper and lower limits of the adjustment setting for the parameter value, the upper and lower limits of the parameter value are taken as the standard.
S1042, updating the production parameters of each device in the injection molding system by using the second setting parameters, and producing the products which need to be produced currently by using the updated injection molding system to obtain the feedback data and the reward value corresponding to the second setting parameters.
The reward value of the product is represented by V, the reward value can be obtained according to formula (4) if the produced product is a good product, and the reward value can be obtained according to formula (5) if the produced product is a defective product.
Formula (4) of Y-psi C
V is-Y + psi C formula (5)
Wherein Y is a positive integer, ψ represents a value coefficient, C is a positive integer, and the numerical values of Y and C are selected by the user as needed, which is not limited in this embodiment.
And S1043, updating the neural network according to the feedback data and the reward value of the second setting parameter so as to optimize the production parameter of each device in the injection molding system.
And optimizing the parameters of the neural network according to the reward value, the second setting parameter, the parameter value adjustment with the maximum current adjustment value, the first setting parameter and the parameter value adjustment with the maximum adjustment value when the first setting parameter is obtained, and updating the neural network. The loss function used in optimizing the parameters of the neural network is, for example, formula (6):
L(ω)＝(VT+1+λ×maxa′Q(ST+1,a′,ω)-Q(ST,abest,ω))2formula (6)
Wherein, VT+1Indicating the prize value of the product produced when the production parameter of each device in the injection molding system is the second set parameter, ST+1Representing a second setting parameter, a' representing the maximum adjustment value when the second setting parameter is obtained, ω representing a parameter of the neural network, STDenotes a first setting parameter, abestWhich indicates the maximum adjustment value when the first setting parameter is obtained, and lambda indicates the discount factor.
And inputting the second setting parameter as the first setting parameter into the updated neural network, updating the neural network according to the process, and stopping updating when the updating end condition is met. The update condition is, for example, the number of updates to reach the neural network.
It should be noted that, when the reward value used in the neural network is updated to use the second setting parameter as the production parameter of each device in the injection molding system, the reward value is obtained according to whether the product is a good product or a bad product by actually producing the product that needs to be produced currently.
Optionally, after S104, the method further includes:
and uploading the first system description, the second setting parameter and the feedback data corresponding to the second setting parameter to a cloud end or storing the feedback data in a local database.
In the embodiment, a system description of a product to be produced at present is acquired, which is called a first system description; determining at least one second system description match that matches the first system description; acquiring a first setting parameter for producing the product required to be produced currently according to the setting parameter and the feedback data corresponding to the at least one second system description; and automatically adjusting and optimizing the production parameters of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system. The purposes of automatically acquiring the first setting parameter of the production parameter of each device in the injection molding system and automatically optimizing the setting parameter are achieved, the production efficiency is improved, and the defective rate is reduced.
In some embodiments, before S1042, further comprising: comparing the second setting parameter with the setting parameter of the injection molding system model to obtain a reward value corresponding to the second setting parameter; and pre-training the neural network according to the reward value corresponding to the second setting parameter.
In this embodiment, an injection molding system model for training a deep neural network is designed. Wherein the injection molding system model simulation is used to simulate the injection molding system. The injection molding system model is designed according to setting parameters and feedback data corresponding to at least one second system description. Accordingly, a reward value corresponding to the second set of parameters may be determined from the injection molding system model.
The neural network is then pre-trained based on the feedback data and the reward value for the second set of parameters. The specific implementation process of the pre-training neural network may refer to S1043 according to the feedback data and the reward value of the second setting parameter, which is not described herein again.
It should be noted that the reward value corresponding to the second setting parameter used when the neural network is pre-trained is obtained through the injection molding system model, rather than the reward value corresponding to the second setting parameter obtained according to the produced product after the production parameter of each production device of the injection molding system is updated by using the second setting parameter.
It should be noted that, before obtaining the reward value pre-training neural network corresponding to the second setting parameter by using the injection molding system model, it is necessary to determine whether a condition for stopping the pre-training neural network is satisfied, and when the condition is satisfied, it is indicated that the neural network can be trained by using the reward value corresponding to the second setting parameter during actual production, that is, the second setting parameter updates the production parameters of each production device of the injection molding system, and then the reward value corresponding to the second setting parameter obtained according to the produced product is used to train the neural network. And when the condition is not met, continuously obtaining the corresponding reward value of the second setting parameter through the injection molding system model, and pre-training the neural network by using the reward value. Then, the second setting parameter is used as an input to the trained neural network, i.e., the first setting parameter, and the second setting parameter is obtained according to the trained neural network.
In this embodiment, the condition for judging the end of the neural network training may be that a preset training time is reached, or that the reward value obtained for K consecutive times is a reward value corresponding to a genuine product, where K is selected by the user as needed, and this embodiment is not limited thereto.
The pre-training neural network in the embodiment of the invention obtains the corresponding reward value of the second setting parameter through the injection molding system model, and pre-trains the neural network by using the reward value. And updating the neural network by using the reward value corresponding to the second set parameter obtained according to the produced product after updating the production parameter of each production device of the injection molding system by using the second set parameter.
In this embodiment, before the second setting parameter output by the neural network is used for actual production, the neural network is trained, that is, after the second setting parameter is obtained, the feedback data and the reward value corresponding to the second setting parameter are obtained through the injection molding system model, and the neural network is trained according to the obtained feedback data and the reward value corresponding to the second setting parameter. Updating the second setting parameter output by the trained neural network to the injection molding system, continuously updating the neural network by using the corresponding feedback data and reward value of the second setting parameter in actual production, and optimizing the production parameters (namely the setting parameters) of each production device in the injection molding system. Therefore, the situation that a large number of defective products are produced due to the fact that the production parameters of each device in the injection molding system are updated by using the setting parameters capable of producing the defective products can be avoided, and the defective rate is further reduced.
In some embodiments, one possible way to obtain the reward value corresponding to the second setting parameter by comparing the second setting parameter with the setting parameter of the injection molding system model is to:
calculating the cross entropy of the second setting parameter and the setting parameter of the injection molding system model; acquiring feedback data corresponding to the setting parameters of the injection molding system model according to the cross entropy; determining the feedback data as the feedback data corresponding to the second setting parameter; and acquiring the reward value according to the feedback data.
In this embodiment, after the second setting parameter is obtained, the cross entropy between the second setting parameter and the setting parameter of the injection molding system model is calculated. The method for calculating the cross entropy may refer to the prior art, and is not described herein again.
And acquiring one feedback data in the feedback data corresponding to the setting parameter of the injection molding system model according to the cross entropy, and determining the feedback data as the feedback data of the second setting parameter.
In some embodiments, one possible implementation of obtaining the reward value according to the feedback data is:
judging whether the product corresponding to the feedback data is a genuine product or not; if so, the reward value corresponding to the second set parameter is the reward value when the product corresponding to the feedback data is a genuine product; if not, the reward value corresponding to the second setting parameter is the reward value when the product corresponding to the feedback data is defective.
In this embodiment, the quality of the product corresponding to the feedback data (i.e., whether the product is a genuine product or a defective product) may be obtained, and when the product is a genuine product, the reward value of the second setting parameter is the reward value when the product corresponding to the feedback data is a genuine product; and when the product is a defective product, the reward value of the second setting parameter is the reward value when the product corresponding to the feedback data is a defective product.
In some embodiments, one possible implementation manner of obtaining the feedback data corresponding to the setting parameters of the injection molding system model according to the cross entropy is as follows:
acquiring feedback data corresponding to at least one setting parameter of the injection molding system model with the cross entropy smaller than a preset value; calculating the deviation degree of the feedback data corresponding to the at least one setting parameter; and determining the feedback data with the minimum deviation degree as the feedback data corresponding to the setting parameters of the injection molding system model.
In this embodiment, the cross entropy and the preset value of the second setting parameter and the setting parameter of the injection molding system model are determined, and feedback data corresponding to at least one setting parameter of the injection molding system model with the cross entropy smaller than the preset value is obtained, where the preset value is selected by a user according to actual requirements, and this embodiment is not limited thereto. Calculating the deviation degree of the feedback data corresponding to the at least one setting parameter, wherein the method for the deviation degree may refer to formula (2), which is not described herein again.
It should be noted that, when calculating the deviation degree based on the feedback data, it is not necessary to delete the invalid data satisfying the deletion condition based on the deletion condition, and here, the deviation degree of all the feedback data corresponding to the setting parameter whose cross entropy between the second setting parameters is greater than or equal to the preset value is calculated.
And obtaining feedback data with the minimum deviation degree, and determining the feedback data as feedback data corresponding to the second setting parameter.
Fig. 3 is a schematic structural diagram of a parameter optimization apparatus in an injection molding system according to an embodiment of the present invention. As shown in fig. 3, the apparatus of this embodiment may include a first obtaining module 31, a matching module 32, and an adjusting and optimizing module 33. Optionally, the apparatus may further include: a save module 34.
A first obtaining module 31, configured to obtain a system description of the injection molding system, referred to as a first system description, where the first system description is a basic system condition of each device in the injection molding system related to current product production, where the system description may include production product information, injection molding device information, mold information, various peripheral accessory information, and environment information;
a matching module 32 for determining at least one second system description match with the first system description, the second system description being stored in a database as a system description of each device in the injection molding system at the time of production;
the first obtaining module 31 is further configured to obtain a first setting parameter for producing the current product according to a setting parameter and feedback data corresponding to the at least one second system description, where the setting parameter is a production parameter of each device in the injection molding system when a product is produced, and the feedback data is feedback data of a sensor in each device in the injection molding system and human judgment data of a worker;
and the tuning module 33 is used for automatically tuning the production parameters of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system.
The saving module 34 is configured to upload the first system description, the second setting parameter, and the feedback data corresponding to the second setting parameter to the cloud or save the feedback data in the database.
The parameter optimization apparatus in the injection molding system described above in this embodiment may be configured to implement the technical solutions in the above method embodiments, and the implementation principles and technical effects are similar, where the functions of each module may refer to the corresponding descriptions in the method embodiments, and are not described herein again.
Fig. 4 is a schematic structural diagram of a parameter optimization apparatus in an injection molding system according to another embodiment of the present invention. As shown in fig. 4, the parameter optimizing apparatus in the injection molding system may be a network device or a chip of the network device, and the apparatus may include: at least one processor 41 and a memory 42. Fig. 4 shows a parameter optimization device in an injection molding system, as an example of a processor, in which,
and a memory 42 for storing programs. In particular, the program may include program code comprising computer operating instructions. The memory 42 may comprise high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
The processor 41 is configured to execute the computer execution instruction stored in the memory 52 to implement the parameter optimization method in the injection molding system in the foregoing embodiment, which has similar implementation principles and technical effects and is not described herein again.
The processor 41 may be a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits configured to implement the embodiments of the present Application.
Alternatively, in particular implementations, if the memory 42 and the processor 41 are implemented independently, then
The memory 42 and the processor 41 may be connected to each other via a bus and communicate with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended ISA (EISA) bus, or the like. The buses may be divided into address buses, data buses, control buses, etc., but do not represent only one bus or one type of bus.
Alternatively, in a specific implementation, if the memory 42 and the processor 41 are integrated on a chip, the memory 42 and the processor 41 may perform the same communication through an internal interface.
The parameter optimization device in the injection molding system described above in this embodiment may be used to implement the technical solutions in the above method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
Those of ordinary skill in the art will understand that: all or a portion of the steps of implementing the above-described method embodiments may be performed by hardware associated with program instructions. The program may be stored in a computer-readable storage medium. When executed, the program performs steps comprising the method embodiments described above; and the aforementioned storage medium includes: various media capable of storing program codes, such as Read-Only Memory (ROM), Random Access Memory (RAM), magnetic or optical disk, and the like.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. An injection molding system setting parameter optimization method is characterized by comprising the following steps:
obtaining a first system description of the injection molding system, the first system description being a basic system case for each device in the injection molding system in relation to a current product production, wherein the system description comprises at least one of: production product information, injection molding equipment information, mold information, various peripheral auxiliary machine information and environment information;
finding a second system description in a cloud and/or a database, wherein the matching degree of the value in at least one system description and the value in the first system description meets a matching coefficient in the first system description, and the second system description is stored in the cloud and/or the local database and used for producing each device in the injection molding system;
acquiring a first setting parameter for producing the current product according to a setting parameter and feedback data corresponding to the at least one second system description, wherein the setting parameter is a production parameter of each device in the injection molding system when the product is produced, and the feedback data is feedback data and artificial judgment data of each device in the injection molding system;
acquiring a second setting parameter according to the first setting parameter and the neural network;
updating the production parameters of each device in the injection molding system by using the second setting parameters, and producing the products to be produced at present by using the updated injection molding system to obtain feedback data and reward values corresponding to the second setting parameters;
updating the neural network according to the feedback data and the reward value of the second setting parameter so as to optimize the production parameters of each device in the injection molding system;
the acquiring a first setting parameter for producing the current product according to the setting parameter and the feedback data corresponding to the at least one second system description specifically includes: deleting the invalid data meeting the deleting condition according to the deleting condition, calculating the deviation degree of each group of valid data according to the relevance, comparing the deviation degree of each group of valid data with a first threshold value, keeping the valid data with the deviation degree smaller than the first threshold value, called as actual valid data, carrying out weighted average on the parameters at the same position in the setting parameters in the actual valid data, and calculating the initial value of the parameter at the position in the first setting parameter according to the deviation degree of each group of actual valid data.
2. The method of claim 1, wherein prior to updating the production parameters for each device in the injection molding system using the second set of parameters, comprising:
comparing the second setting parameter with the setting parameter of the injection molding system model to obtain a reward value corresponding to the second setting parameter;
and pre-training the neural network according to the reward value corresponding to the second setting parameter.
3. The method of claim 2, wherein the comparing the second set of parameters with the set of parameters of the injection molding system model to obtain the reward value corresponding to the second set of parameters comprises:
calculating the cross entropy of the second setting parameter and the setting parameter of the injection molding system model;
acquiring feedback data corresponding to the setting parameters of the injection molding system model according to the cross entropy;
determining the feedback data as the feedback data corresponding to the second setting parameter;
and acquiring the reward value according to the feedback data.
4. The method of claim 3, wherein obtaining the reward value based on the feedback data comprises:
judging whether the product corresponding to the feedback data is a genuine product or not;
if so, the reward value corresponding to the second set parameter is the reward value when the product corresponding to the feedback data is a genuine product;
if not, the reward value corresponding to the second setting parameter is the reward value when the product corresponding to the feedback data is defective.
5. The method according to claim 3, wherein the obtaining feedback data corresponding to the setting parameters of the injection molding system model according to the cross entropy comprises:
acquiring feedback data corresponding to at least one setting parameter of the injection molding system model with the cross entropy smaller than a preset value;
calculating the deviation degree of the feedback data corresponding to the at least one setting parameter;
and determining the feedback data with the minimum deviation degree as the feedback data corresponding to the setting parameters of the injection molding system model.
6. The method of claim 1, wherein the optimizing the production parameters of each device in the injection molding system according to the first setup parameters, the neural network, and the injection molding system further comprises:
and uploading the first system description, the second setting parameter and the feedback data corresponding to the second setting parameter to a cloud end or storing the feedback data in a local database.
7. An apparatus for optimizing parameters in an injection molding system, comprising:
a first obtaining module for obtaining a system description of the injection molding system, referred to as a first system description, the first system description being a basic system situation of each device in the injection molding system in relation to a current product production, wherein the system description comprises at least one of: production product information, injection molding equipment information, mold information, various peripheral auxiliary machine information and environment information;
a matching module, configured to determine that at least one second system description matches the first system description, where the second system description is stored in a cloud and/or a local database of system descriptions of each device in the injection molding system during production;
the first obtaining module is further configured to obtain a first setting parameter for producing the current product according to a setting parameter and feedback data corresponding to the at least one second system description, where the setting parameter is a production parameter of each device in the injection molding system when the product is produced, and the feedback data is feedback data of a sensor in each device in the injection molding system and human judgment data of a worker;
the adjusting and optimizing module is used for optimizing the production parameters of each device in the injection molding system according to the first setting parameter, the neural network and the injection molding system
The tuning module is specifically configured to obtain a second setting parameter according to the first setting parameter and the neural network;
updating the production parameters of each device in the injection molding system by using the second setting parameters, and producing the products to be produced at present by using the updated injection molding system to obtain feedback data and reward values corresponding to the second setting parameters;
and updating the neural network according to the feedback data and the reward value of the second setting parameter so as to optimize the production parameters of each device in the injection molding system.
8. An apparatus for optimizing parameters in an injection molding system, comprising: a memory for storing program instructions and a processor for calling the program instructions in the memory to perform the method of parameter optimization in an injection molding system according to any of claims 1-6.
9. A readable storage medium, characterized in that the readable storage medium has stored thereon a computer program; the computer program, when executed, implements a method of parameter optimization in an injection molding system according to any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811521647.3A CN111319206B (en) | 2018-12-13 | 2018-12-13 | Parameter optimization method and device in injection molding system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811521647.3A CN111319206B (en) | 2018-12-13 | 2018-12-13 | Parameter optimization method and device in injection molding system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111319206A CN111319206A (en) | 2020-06-23 |
| CN111319206B true CN111319206B (en) | 2021-11-23 |
Family
ID=71170502
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811521647.3A Active CN111319206B (en) | 2018-12-13 | 2018-12-13 | Parameter optimization method and device in injection molding system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111319206B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113741189B (en) * | 2021-08-31 | 2022-04-12 | 广东工业大学 | Self-adaptive learning rapid control method for injection molding machine |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0897786B1 (en) * | 1997-08-21 | 2006-03-22 | Kistler Holding AG | Regulation process for an injection moulding machine for plastics |
| JP4499601B2 (en) * | 2005-04-01 | 2010-07-07 | 日精樹脂工業株式会社 | Control device for injection molding machine |
| DE102006014475A1 (en) * | 2006-03-29 | 2007-10-04 | Rieter Ingolstadt Spinnereimaschinenbau Ag | Procedure for controlling a spinning preparation machine e.g. carding engine, drawing frame/rotor spinning machine, by determining input variables of a control device of the spinning machine so that parameter of the machine is optimized |
| CN101320400B (en) * | 2008-07-16 | 2010-04-21 | 桂林电子科技大学 | Optimization design method of micro-electron packaging device based on artificial neural network |
| CN104527009A (en) * | 2015-01-06 | 2015-04-22 | 广西大学 | Injection moulding process parameter optimization regulation and control method based on Gaussian prediction |
| AT519491A1 (en) * | 2016-12-23 | 2018-07-15 | Engel Austria Gmbh | Method for optimizing a process optimization system and method for simulating a shaping process |
| JP6779316B2 (en) * | 2017-01-17 | 2020-11-04 | 日精樹脂工業株式会社 | Molding condition estimation method for injection molding machine |
| CN107944147A (en) * | 2017-11-28 | 2018-04-20 | 柳州市城中区聚宝机械冲压厂 | Shooting Technique optimization method and Shooting Technique based on GRNN neutral nets |
-
2018
- 2018-12-13 CN CN201811521647.3A patent/CN111319206B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111319206A (en) | 2020-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI785227B (en) | Method for batch normalization layer pruning in deep neural networks | |
| TW202139045A (en) | Privacy protection-based target service model determination | |
| CN108170739A (en) | Problem matching process, terminal and computer readable storage medium | |
| JP6950756B2 (en) | Neural network rank optimizer and optimization method | |
| CN106709565A (en) | Optimization method and device for neural network | |
| WO2020224297A1 (en) | Method and device for determining computer-executable integrated model | |
| CN108416187A (en) | A kind of method and device of determining pruning threshold, model pruning method and device | |
| CN111967971B (en) | Bank customer data processing method and device | |
| CN113037531A (en) | Flow prediction method, device and storage medium | |
| CN110264407B (en) | Image super-resolution model training and reconstruction method, device, equipment and storage medium | |
| CN113642938B (en) | Intelligent production management method and system | |
| CN115249315B (en) | Heterogeneous computing device-oriented deep learning image classification method and device | |
| JP2012518834A (en) | Method and system for calculating website visitor ratings | |
| CN111319206B (en) | Parameter optimization method and device in injection molding system | |
| CN109102468B (en) | Image enhancement method and device, terminal equipment and storage medium | |
| CN113361698A (en) | Processing method and device of neural network model, and data processing method and device | |
| CN110874635B (en) | Deep neural network model compression method and device | |
| CN111949640A (en) | Intelligent parameter adjusting method and system based on industrial big data | |
| CN109783769B (en) | Matrix decomposition method and device based on user project scoring | |
| CN107203916B (en) | User credit model establishing method and device | |
| CN111524023A (en) | Greenhouse adjusting method and system | |
| US20210397962A1 (en) | Effective network compression using simulation-guided iterative pruning | |
| CN113326832B (en) | Model training method, image processing method, electronic device, and storage medium | |
| CN113010762B (en) | Data enhancement method, device, storage medium and electronic equipment | |
| CN114385876A (en) | Model search space generation method, device and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |