CN109101698B - Feature selection method and device based on injection molding model and storage medium - Google Patents
Feature selection method and device based on injection molding model and storage medium Download PDFInfo
- Publication number
- CN109101698B CN109101698B CN201810796642.5A CN201810796642A CN109101698B CN 109101698 B CN109101698 B CN 109101698B CN 201810796642 A CN201810796642 A CN 201810796642A CN 109101698 B CN109101698 B CN 109101698B
- Authority
- CN
- China
- Prior art keywords
- injection molding
- characteristic
- variable
- target
- variables
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention discloses a feature selection method based on an injection molding model, which comprises the steps of firstly obtaining sample data corresponding to each feature variable of the injection molding model, then calculating feature values corresponding to the feature variables according to the sample data, finally selecting target feature values from the feature values according to a predefined rule, and taking the feature variables corresponding to the target feature values as target feature variables. When determining the characteristic variables of the injection molding model, the method firstly calculates the characteristic values corresponding to the characteristic variables, finally selects the target characteristic values from the characteristic values, and takes the characteristic variables corresponding to the target characteristic values as the target characteristic variables, so that the number of the target characteristic variables of the final injection molding model is small, and the structure of the injection molding model is simplified. The invention also discloses a characteristic selection device and a storage medium based on the injection molding model, and the effects are as above.
Description
Technical Field
The invention relates to the field of injection molding models, in particular to a feature selection method and device based on an injection molding model and a storage medium.
Background
The quality of injection-molded products is affected by various process conditions including injection-molding material properties, injection-molding process parameters, injection molds, filling and cooling of the melt, and the like. The property of the injection molding material is not easy to change, so that the quality of injection molding products is difficult to adjust by changing the property of the injection molding material; the production quality and the production efficiency of injection products can be easily changed by optimizing injection molding process parameters, injection molds, filling and cooling of the solution and other factors.
In the injection molding process, many process parameters are included, such as injection pressure, holding pressure, melt temperature, maximum mold clamping force, maximum wall shear stress, cooling time, mold temperature, filling time, holding time, and the like. The influence degree of different process parameters on the production quality and the production efficiency of injection products is different, and the traditional method is that a plurality of process parameters are used as design variables of an injection molding model, and the structure of the injection molding model is complex due to more design variables, and the calculation process is complex when the objective function of the injection molding product is solved through the injection molding model; the objective function can be the production quality and the production efficiency of the injection molding product.
Therefore, how to reduce the design variables in the injection molding model to simplify the structure of the injection molding model is a problem to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide a feature selection method, a feature selection device and a storage medium based on an injection molding model, which reduce design variables in the injection molding model and simplify the structure of the injection molding model.
In order to achieve the above purpose, the embodiment of the present invention provides the following technical solutions:
first, an embodiment of the present invention provides a feature selection method based on an injection molding model, including:
acquiring sample data corresponding to each characteristic variable of the injection molding model;
calculating a characteristic value corresponding to each characteristic variable according to each sample data;
and selecting a target characteristic value from each characteristic value according to a predefined rule, and taking a characteristic variable corresponding to the target characteristic value as a target characteristic variable.
Preferably, the calculating a feature value corresponding to each feature variable according to each sample data includes:
forming a sample matrix by each sample data and calculating a covariance matrix according to the sample matrix;
and calculating the eigenvalue corresponding to each of the characteristic variables by using the covariance matrix.
Preferably, the forming each sample data into a sample matrix and calculating a covariance matrix according to the sample matrix includes:
composing each sample data into the sample matrix;
calculating an average eigenvalue value by using the sample data corresponding to the characteristic variables in the sample matrix and the number of the characteristic variables;
calculating the difference value between each sample data and the average characteristic numerical value;
and combining the difference values to obtain the covariance matrix.
Preferably, after the selecting a target feature value from each feature value according to a predefined rule and taking a feature variable corresponding to the target feature value as a target feature variable, the method further includes:
taking the target characteristic variable as an input variable;
taking the production quality and the production efficiency of the injection molding product as output variables;
constructing the injection molding model by using the input variable and the output variable and optimizing the injection molding model;
and solving the optimized injection molding model to obtain the optimal solution of each target characteristic variable.
Preferably, the constructing and optimizing an injection molding model by using the input variables and the output variables includes:
constructing the injection molding model by using the input variables and the output variables;
setting the number of various neurons of a hidden layer of the injection molding model;
selecting the optimal neuron number of the injection molding model from the neuron numbers by using test sample data and verification sample data corresponding to the input variable;
and obtaining the optimized injection molding model by using the input variable, the output variable and the hidden layer corresponding to the optimal neuron number.
Preferably, the selecting a target feature value from each feature value according to a predefined rule, and using a feature variable corresponding to the target feature value as a target feature variable includes:
and selecting a characteristic value larger than a target threshold value from all the characteristic values, and taking a characteristic variable corresponding to the characteristic value larger than the target threshold value as a target characteristic variable of the injection molding model.
Preferably, the target characteristic variables include a dwell pressure, a melt temperature, a cooling time, a mold temperature, and a dwell time.
Then, the embodiment of the invention discloses a feature selection device based on an injection molding model, which comprises the following components:
the sample data acquisition module is used for acquiring sample data corresponding to each characteristic variable of the injection molding model;
the characteristic value calculating module is used for calculating a characteristic value corresponding to each characteristic variable according to each sample data;
and the target characteristic variable setting module is used for selecting a target characteristic value from each characteristic value according to a predefined rule and taking the characteristic variable corresponding to the target characteristic value as a target characteristic variable.
Secondly, the embodiment of the invention discloses another feature selection device based on an injection molding model, which comprises the following components:
a memory for storing a computer program;
a processor for executing the computer program stored in the memory to implement the steps of the injection molding model-based feature selection method as described in any one of the above.
Finally, the embodiment of the invention discloses a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of the feature selection method based on an injection molding model as described in any one of the above.
The feature selection method based on the injection molding model comprises the steps of firstly obtaining sample data corresponding to each feature variable of the injection molding model, then calculating feature values corresponding to the feature variables according to the sample data, finally selecting target feature values from the feature values according to predefined rules, and taking the feature variables corresponding to the target feature values as target feature variables. Therefore, in the method, when the characteristic variables of the injection molding model are determined, all the characteristic variables related to the injection molding model are not used as the target characteristic variables, but the characteristic values corresponding to the characteristic variables are calculated first, the target characteristic values are selected from the characteristic values finally, and the characteristic variables corresponding to the target characteristic values are used as the target characteristic variables, so that the number of the target characteristic variables of the final injection molding model is small, and the structure of the injection molding model is simplified. The invention also discloses a characteristic selection device and a storage medium based on the injection molding model, and the effects are as above.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic flow chart of a feature selection method based on an injection molding model according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a feature selection apparatus based on an injection molding model according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of another feature selection apparatus based on an injection molding model according to an embodiment of the present invention;
FIG. 4 (a) is a front view of an injection mold of an automotive interior panel according to an embodiment of the present invention;
FIG. 4 (b) is a sectional view of an injection mold for an automotive interior panel according to an embodiment of the present invention;
FIG. 5 is a graph illustrating the variance accumulation of various characteristic variables according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of an ELM extreme learning machine model of an automotive interior panel according to an embodiment of the present invention;
FIG. 7 (a) is a graph of the training results of the first ELM extreme learning machine model disclosed in the embodiment of the present invention;
FIG. 7 (b) is a graph of the training results of the second ELM extreme learning machine model disclosed in the embodiment of the present invention;
FIG. 7 (c) is a graph of the training results of the third ELM extreme learning machine model disclosed in the embodiment of the present invention;
FIG. 7 (d) is a graph of the training results of a fourth ELM extreme learning machine model disclosed in the embodiment of the present invention;
fig. 8 (a) is a distribution curve diagram of predicted values and experimental values of two objective function values disclosed in the embodiment of the present invention;
FIG. 8 (b) is a graph of the output error of two objective functions disclosed in the present invention;
FIG. 9 is a schematic flow chart of an ELM learning machine-non-dominated sorting genetic algorithm according to an embodiment of the present invention;
FIG. 10 is a randomly generated initial population spatial distribution scattergram disclosed by an embodiment of the present invention;
FIG. 11 is a graph illustrating convergence of the evolution of the randomly generated initial population according to an embodiment of the present invention;
fig. 12 is a distribution scatter diagram of Pareto optimal solutions of the first front end of population evolution to 100 generations, which is disclosed by the embodiment of the present invention.
Detailed Description
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 only a part of the embodiments of the present invention, and not all of the embodiments. 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.
The embodiment of the invention discloses a feature selection method, a feature selection device and a storage medium based on an injection molding model, which reduce design variables in the injection molding model and simplify the structure of the injection molding model.
Referring to fig. 1, fig. 1 is a schematic flow chart of a feature selection method based on an injection molding model according to an embodiment of the present invention, where the method includes:
s101, sample data corresponding to each characteristic variable of the injection molding model is obtained.
Specifically, in this embodiment, the injection molding model is a mathematical model that is established by using the injection molding process parameters as input variables, and the product quality and production efficiency of the injection molded product as output variables, and combining the input variables, the output variables, the injection molding constraint conditions, and the like. The quality of the injection molding product can be called as warping deformation, and the production efficiency can be called as a forming period; the input variable is any one of the characteristic variables in this embodiment, and each of the characteristic variables mainly includes melt temperature, mold temperature, injection pressure, injection time, holding pressure, holding time, cooling time, and the like. For example, g 1 (X)，g 2 (X) production quality (warpage) W of the injection molding model as a constraint condition of the injection molding model apg (X) and production efficiency (Molding cycle) T of injection Molding die cyc (X) as an output variable (objective function) of the injection molding model, and a plurality of characteristic variables as input variables X of the injection molding model, establishing the following injection molding model:
wherein the objective function W apg (X) and T cyc (X) is obtained by Moldflow Adviser2017 modular flow analysis, and the constraint condition g 1 (X) indicates that the plastic must be filled in all typesCavity, constraint g 2 (X) indicates that the plastic must fill all the cavities within a predetermined time, N t_cav And N f_cav Respectively representing the number of all cavities and the number of cavities that have been filled, t f And t f_pre Respectively representing the filling time and the required time. The model can be established by utilizing Moldflow software to perform a simulation injection molding experiment and process data so as to establish a multi-objective function and multi-input variable injection molding process nonlinear optimization model. Of course, the injection molding model may be established in different modeling manners according to different objective functions, and is not limited herein.
Further, the sample data is a data value corresponding to each characteristic variable, for example, the sample data is 20s, 21s, 22s, 21.5s corresponding to the characteristic variable dwell time. Of course, the number of sample data and the size of the sample data of the characteristic variables in the embodiment of the present invention are not limited herein according to the actual injection molding environment. The data for each sample can be obtained by orthogonal and modeling by a Mold flow advisor 2017.
And S102, calculating characteristic values corresponding to the characteristic variables according to the sample data.
Specifically, in the present embodiment, the magnitude of the characteristic value of each characteristic variable represents the degree of influence of the characteristic variable on the injection molding model. That is, the larger the characteristic value is, the larger the influence degree of the corresponding characteristic variable on the production quality and the production efficiency of the injection-molded product by injection molding is. The specific calculation process will be described in detail in the following embodiment.
S103, selecting a target characteristic value from the characteristic values according to a predefined rule, and taking a characteristic variable corresponding to the target characteristic value as a target characteristic variable.
Specifically, in this embodiment, the predefined rule may have several modes, the first mode may be that, first, each eigenvalue is sorted from large to small, then, each eigenvalue is superimposed to obtain a sum of total eigenvalues, then, an occupation ratio of each eigenvalue to the sum of the total eigenvalues is calculated, then, each occupation ratio is sorted from large to small, and a larger occupation ratio is superimposed, and if a superposition result of the larger occupation ratio is already close to 100%, and occupation ratios of the remaining eigenvalues to the sum of the total eigenvalues are small, a eigenvalue corresponding to the eigenvalue of the larger occupation ratio may be used as a target eigenvalue. For example, if the number of the characteristic variables is 9, and the characteristic values corresponding to the 9 characteristic variables are 4.988, 2.459, 0.971, 0.458, 0.097, 0.012, 0.010, 0.003, 0.000, and 0.001, respectively, the ratio of the characteristic values corresponding to the first two characteristic variables to the sum of all the characteristic values is calculated as 82.743%, and at this time, the ratio of the third characteristic value to the fourth characteristic value to the ratio of the first characteristic value to the second characteristic value is calculated as 98.619%. At this time, the ratios of the first four eigenvalues to the sum of the total eigenvalues are already close to 100%, and the ratios of the last five eigenvalues to the sum of all eigenvalues add up to 1.381%. Therefore, at this time, the first four feature values, namely 4.988, 2.459, 0.971 and 0.458, can be used as target feature values, and the feature variables corresponding to these four feature values can be used as target feature variables. In a second manner, a target ratio threshold may be set, the ratios of all eigenvalues to the sum of all eigenvalues are sorted from large to small, and accumulated in the order from large to small, and when the accumulation result is greater than or equal to the target ratio threshold, the eigenvalue corresponding to the accumulated eigenvalue is used as the target eigenvalue, for example, the target ratio threshold is set to 98% by taking the above-mentioned numerical value as an example; calculating the ratio of each characteristic value to the sum of all the characteristic values, and then overlapping the ratios from large to small, wherein the ratio of the first four characteristic values to the sum of all the characteristic values is 98.619%, which is 98% larger than the target ratio threshold, so that the characteristic variables corresponding to the first four characteristic values are used as target characteristic variables. In the third embodiment, the third embodiment is taken as a preferred embodiment, that is, the feature value greater than the target threshold is selected from the feature values, and the feature variable corresponding to the feature value greater than the target threshold is taken as the target feature variable of the injection molding model. Of course, the predefined rule may have other rules according to the actual environment of injection molding, and the target ratio threshold, the target threshold and the size of the characteristic value are not limited herein.
The feature selection method based on the injection molding model comprises the steps of firstly obtaining sample data corresponding to each feature variable of the injection molding model, then calculating feature values corresponding to the feature variables according to the sample data, finally selecting target feature values from the feature values according to predefined rules, and taking the feature variables corresponding to the target feature values as target feature variables. Therefore, when determining the characteristic variables of the injection molding model, the method does not take all the characteristic variables related to the injection molding model as target characteristic variables, but calculates the characteristic values corresponding to the characteristic variables, selects the target characteristic values from the characteristic values, and takes the characteristic variables corresponding to the target characteristic values as the target characteristic variables, so that the number of the target characteristic variables of the final injection molding model is small, and the structure of the injection molding model is simplified.
Based on the above embodiment, in this embodiment, calculating the feature value corresponding to each feature variable from each sample data includes:
forming a sample matrix by using each sample data and calculating a covariance matrix according to the sample matrix;
and calculating the eigenvalue corresponding to each eigenvalue by using the covariance matrix.
As a preferred embodiment, the forming of the sample matrix from the sample data and the calculation of the covariance matrix according to the sample matrix include:
forming a sample matrix by using each sample data;
calculating an average characteristic numerical value by using each sample data corresponding to each characteristic variable in the sample matrix and the number of the characteristic variables;
calculating the difference value between each sample data and the average characteristic value;
and combining the difference values to obtain a covariance matrix.
Specifically, in this embodiment, the sample matrix is composed of sample data corresponding to each feature variable, and the covariance matrix is a covariance matrixThe array is obtained by combining the difference values between the sample data corresponding to each characteristic variable and the average value of the sample data corresponding to all the characteristic variables. For example, if the number of characteristic variables is 9, x is respectively 1 ，x 2 ，x 3 ，x 4 ，x 5 ，x 6 ，x 7 ，x 8 ，x 9 (ii) a Each feature variable corresponds to 49 sample data, and the formed sample matrix is X = [ X = [ [ X ]) 1 ,x 2 ,...,x 9 ]Wherein x is i ＝[x i1 ,x i2 ,...,x i49 ]I =1,2, …,9. Then, calculating average characteristic numerical value according to sample data corresponding to all characteristic variables and the number of the characteristic variables
Mean characteristic value->
The calculation formula of (c) is as follows:
then calculating the sample data and average characteristic value of each characteristic variable
The difference between them is calculated as follows:
in the above two formulas, i =1,2, …,9, each characteristic variable x i Containing 49 sample data. Thus, the covariance matrix C can be expressed as:
get helpAnd calculating the eigenvalue of each characteristic variable and the eigenvector corresponding to each characteristic variable after the variance matrix. Besides directly representing the degree of influence of each characteristic variable on the injection-molded product by the size of the characteristic value, the degree of influence on the injection-molded product can also be represented by the variance of each characteristic vector. Each characteristic variable is a principal component. For example, the eigenvalue of each eigenvalue is calculated to be λ by the above formula i And a feature vector η i Then each principal component can be represented as:
y i ＝η i T X
for each principal component y i Its variance can be calculated by:
Var(y i )＝Var(η i T X)＝η i T Var(X)η i ＝λ i η i T η i
wherein i =0,1, …,9; it can be seen that the variance of each principal component is proportional to its corresponding eigenvalue, that is, the magnitude of the eigenvalue of each eigenvalue determines the magnitude of the variance of each eigenvalue. Therefore, the characteristic value or the variance of the characteristic variable can represent the influence of the characteristic variable on the production quality and the production cycle of the injection molding product.
In the present embodiment, the number of feature variables and the number of sample data corresponding to each feature variable are not limited to these.
Based on the foregoing embodiment, in this embodiment, after selecting a target feature value from each feature value according to a predefined rule, and taking a feature variable corresponding to the target feature value as a target feature variable, the method further includes:
taking the target characteristic variable as an input variable;
taking the production quality and the production efficiency of the injection molding product as output variables;
constructing an injection molding model by using the input variable and the output variable, and optimizing the injection molding model;
and solving the optimized injection molding model to obtain the optimal solution of each target characteristic variable.
As a preferred embodiment, the constructing and optimizing an injection molding model using input variables and output variables includes:
constructing an injection molding model by using the input variable and the output variable;
setting the number of various neurons of a hidden layer of an injection molding model;
selecting the optimal number of neurons of the injection molding model from the number of the neurons by using test sample data and verification sample data corresponding to the input variable;
and obtaining the optimized injection molding model by using the input variable, the output variable and the hidden layer corresponding to the optimal neuron number.
Specifically, in this embodiment, the injection molding model may be an ELM model established by an ELM extreme learning machine, and the injection molding model mainly has a three-layer structure. Namely an input layer, an intermediate hidden layer and an output layer. The process of establishing the injection molding model and the process of solving the injection molding model will be described below by way of specific examples. For example, if the finally selected target characteristic variables are the mold temperature, the melt temperature, the holding pressure, the holding time, and the cooling time, respectively. The five characteristic variables are used as input variables of the injection molding model, and the product quality and the production period are used as output variables of the injection molding model. And then determining the number of the neurons of the middle hidden layer to obtain the final number of the neurons, and combining the input variable and the output variable to obtain the optimized injection molding model. The optimization process of the injection molding model is as follows: the number of hidden layer neurons is set to be 5,6,7,8 respectively, and then an injection molding model with an input variable of 5, an output variable of 2 and the number of hidden layer neurons of 5,6,7,8 is obtained respectively. Then, the four injection molding models are tested by using test sample data, verification sample data and training sample data, then training errors output by each injection molding model are obtained, and the number of hidden layer neurons corresponding to the minimum training error is selected as the optimal number of neurons.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a feature selection device based on an injection molding model according to an embodiment of the present invention, and the device includes:
a sample data obtaining module 201, configured to obtain sample data corresponding to each feature variable of the injection molding model;
a feature value calculation module 202, configured to calculate, according to each sample data, a feature value corresponding to each feature variable;
and the target characteristic variable setting module 203 is configured to select a target characteristic value from each characteristic value according to a predefined rule, and use a characteristic variable corresponding to the target characteristic value as the target characteristic variable.
The characteristic selecting device based on the injection molding model disclosed by the embodiment of the invention comprises a sample data acquiring module, a characteristic value calculating module, a target characteristic variable setting module, a characteristic variable selecting module and a characteristic variable selecting module, wherein the sample data acquiring module is used for acquiring sample data corresponding to each characteristic variable of the injection molding model, the characteristic value calculating module is used for calculating a characteristic value corresponding to each characteristic variable according to each sample data, and the target characteristic variable setting module is used for selecting a target characteristic value from each characteristic value according to a predefined rule and taking the characteristic variable corresponding to the target characteristic value as a target characteristic variable. Therefore, when determining the characteristic variables of the injection molding model, the method does not take all the characteristic variables related to the injection molding model as target characteristic variables, but calculates the characteristic values corresponding to the characteristic variables, selects the target characteristic values from the characteristic values, and takes the characteristic variables corresponding to the target characteristic values as the target characteristic variables, so that the number of the target characteristic variables of the final injection molding model is small, and the structure of the injection molding model is simplified.
The embodiment of the present invention further discloses another feature selection device based on an injection molding model, please refer to fig. 3, fig. 3 is a schematic structural diagram of another feature selection device based on an injection molding model disclosed in the embodiment of the present invention, and the device includes:
a memory 301 for storing a computer program;
a processor 302 for executing the computer program stored in the memory to implement the steps of the injection molding model-based feature selection method as mentioned in any of the above embodiments.
The effect of the other feature selection device based on the injection molding model disclosed in the embodiment of the present invention is as the effect of the above-mentioned feature selection method based on the injection molding model, and details are not repeated herein.
In order to better understand the present solution, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the feature selection method based on an injection molding model as mentioned in any of the above embodiments.
The effect of the computer-readable storage medium disclosed in the embodiment of the present invention is as described above with respect to the feature selection method based on an injection molding model, and details are not repeated here.
The technical solution provided by the present invention will be described in detail below by taking an interior panel of an automobile as an example. Referring to fig. 4 (a) and 4 (b), fig. 4 (a) is a front view of an injection mold of an automotive interior panel according to an embodiment of the present invention, and fig. 4 (b) is a cross-sectional view of an injection mold of an automotive interior panel according to an embodiment of the present invention. In the embodiment, an automobile interior panel is taken as a research main body, an injection molding model is designed through Pro/E Creo Parametric2.0 three-dimensional design software, and in the aspect of the integral structure, an irregular oval opening is formed in the middle of a plastic part, and a plurality of protruding bolts are arranged at the bottom of the plastic part and are used for connecting a plurality of accessories. The overall external dimension is as follows: 240.00mm x 138.651mm x 3.495mm, the entire thickness distribution is relatively uniform except for the raised portions.
And (3) introducing the three-dimensional part model into CAE software Moldflow advisor 2017, designing a forming die, a runner system and a cooling system, and carrying out filling, pressure maintaining, cooling and warpage analysis on the three-dimensional part model. We set up here as a multi-cavity gating system. In addition, in a pouring system of an injection mold, whether a gate position is properly selected directly affects the molding quality of a molded part, and a series of defects such as short shots, jetting, stagnation, and depressions may be caused by an improper gate position. In order to obtain the optimal gate position, the embodiment adopts a Moldflow gate position analysis function to obtain the optimal gate position of the plastic part, and the coordinates of the gate position are X:249.982mm Y:10.695mm Z:10.000mm.
The dimensions of the die are defined as 125mm for A plate thickness, 100mm for B plate thickness, and 800mm x 900mm x 225mm for the entire die, depending on the specific dimensions of the part. The material selected for the mould is Tool Steel P-20: generic. The runner and gate related parameters are shown in Table 1-1.
TABLE 1-1 Runner System parameters
The cooling system is very important in the temperature regulation system of the entire mold. Generally, in the whole injection molding cycle of a plastic part, the cooling time can occupy more than half of the whole molding cycle, so the quality of the design of the cooling system directly affects the production efficiency of the product. The cooling system is designed to regulate the temperature of the whole die mainly through cooling water, and the system mainly relates to the position, the size and the layout form of the cooling water, and also comprises various parameters such as the temperature, the flow rate and the like of a cooling medium. 8 cooling circuits are designed in the whole mould, 8 pipes and 4 circuits (in the mould cavity and the mould core respectively) in each part.
Firstly, establishing a three-dimensional injection molding model, and after the three-dimensional injection molding model is established, acquiring sample data of the injection molding model, wherein the acquisition process of the sample data is as follows: set the optimization target to the mass (warp deformation) W apg (X) and efficiency (forming cycle) t cyc (X), 9 characteristic variables include die temperature T mold Temperature T of the melt melt Pressure maintaining pressure P p Pressure maintaining time t p Cooling time t c Maximum mold clamping force F clp Maximum wall shear stressP w Filling time t fil And injection pressure P inj And the like.
According to the characteristics of the mold flow analysis of the Moldflow advisor 2017, the other corresponding parameters can be obtained by simulation only by setting part of characteristic variables, and the temperature T of the mold is selected in the specification mold Temperature T of the melt melt And dwell time t p Three variables, 6 levels were divided and orthogonal experiments were performed. Considering the property of the PP material, setting the temperature of the die to be 50-75 ℃, and taking 6 horizontal points at intervals of 5 ℃; the temperature of the melt ranges from 230 ℃ to 255 ℃, and 6 horizontal points at intervals of 5 ℃ are taken; and 7-12 s of dwell time, 6 horizontal points at intervals of 1s are taken, an orthogonal table is designed, and corresponding experimental data of other characteristic variables and target variables can be obtained through simulation by the Moldflow advisor 2017. Tables 1-2 are tables of partial orthogonal test sample data.
Table 1-2 partial orthogonal test sample data table
After the sample data is obtained, a target characteristic variable needs to be selected from the characteristic variables, a specific selection method adopts a principal component analysis method, and the specific selection process is as follows:
the present document relates to a total of 9 decision variables, each of which has 49 experimental values, i.e. a sample matrix of X = [ X ] 1 ,x 2 ,...,x 9 ]Wherein x is i ＝[x i1 ,x i2 ,...,x i49 ] T I =1,2. Before principal component analysis, samples need to be centralized, and each sample data
I.e. is>
The covariance matrix of the sample matrix is
Then, the characteristic value lambda of the covariance matrix is obtained i And corresponding feature vector eta i Thus, the mapped principal component y can be obtained i ＝η i T X, for each component y i The variance is calculated as follows,
Var(y i )＝Var(η i T X)＝η i T Var(X)η i ＝η i T C 49×49 η i ＝λ i η i T η i
the variance of each component is in direct proportion to the corresponding eigenvector through the formula, namely the variance of each principal component can be reflected by the size of the eigenvalue, the weight sequence of each principal component can be obtained according to the sequence of the eigenvalue, and then p principal components are selected in an accumulated manner according to the maximum accumulated variance contribution rate criterion and the weight from large to small. The total variance interpretation obtained by PCA analysis is shown in tables 1-3, the variance accumulation graph is shown in FIG. 5, and FIG. 5 is a variance accumulation graph of various characteristic variables disclosed by the embodiment of the invention.
TABLE 1-3 interpretation of Total variance by PCA analysis
By the formula Var (y) i )＝Var(η i T X)＝η i T Var(X)η i ＝η i T C 49×49 η i ＝λ i η i T η i Note that the eigenvalues of the covariance matrix after normalization lambda i Is numerically equal to the variance Var (y) of its corresponding feature vector i ) In the initial feature variable space we takeAll 9 eigenvalues and corresponding eigenvectors are presented. As can be seen from tables 1 to 3, the eigenvalues of the first and second principal components are both greater than 1, the variances (eigenvalues) of the two principal components are 4.988 and 2.459, respectively, and the cumulative percentage is 82.743%. In order to make the principal component matrix cover the information of the original data space as much as possible, the 3 rd and 4 th principal components need to be taken into consideration, the total variance ratio of the principal components is 10.792% and 5.084%, respectively, at this time, the cumulative variance ratio reaches 98.619%, the characteristic values of the fifth to ninth components are all less than 0.1, and the cumulative variance ratio is 1.381% of the total. Therefore, only the first 4 principal components are taken, and the feature vector space formed by the first 4 principal components is approximately considered to be capable of expressing all information of the original variable space.
In addition, fig. 5 shows the eigenvalues corresponding to the eigenvectors of the covariance matrix, that is, the variances of the principal components. In particular, the accumulated value (accumulated variance ratio) up to the 4 th principal component has already approached the total value, which more intuitively explains that the first 4 principal components have sufficient reliability to replace the total sample information.
More specifically, tables 1-4 list the component matrices for each principal component consisting of a linear combination of 9 variables, and from the previous analysis, 4 principal components containing sufficient raw variable data information have been obtained, to which are added the 5 th to 9 th components for comparison, and the specific component matrices and vector coefficients are shown in tables 1-4.
TABLE 1-4 component matrices and vector coefficients
Tables 1-4 show principal component loading matrices, each column of loading values showing the correlation of the respective variable with the relevant principal component, exemplified by column 1, with 0.990 actually being the correlation of dwell pressure with the 1 st principal component. In addition, for after normalizationThe determined coefficient R is obtained by regression of the column vector data of the holding pressure and the 1 st principal component 2 =0.98, the corresponding load R =0.99, i.e. the load of the holding pressure on the 1 st principal component.
Based on the above analysis, it can be noted that the load of the 5 variables of the first 4 main components, including the holding pressure, the melt temperature, the cooling time, the mold temperature and the holding time, is considerably larger than that of the other 4 variables, especially in the 1 st main component, the correlation coefficient is close to 1, which indicates that the 1 st main component has a strong correlation relationship, and the load of the other 4 variables is smaller, and there are two explanations: first, it has less impact on the population, similar to noise factor; second, it loses more information in the dimension reduction process, and at this time, the 4 principal components obtained in the foregoing can already express almost all the information of the original sample space, so the information lost by the 4 variables is larger than itself, but can be ignored compared with the total. These two possible conclusions are consistent: i.e. injection pressure, maximum clamping force, maximum wall shear stress, filling time, the 4 variables have less influence on the overall system, and from another point of view, we see that the mold temperature T mold Temperature T of the melt melt Pressure maintaining pressure P p Pressure maintaining time t p And a cooling time t c These 5 variables are the main influencing factors (components) of the original sample space. So in the following analysis we will analyze these 5 variables in all original sample spaces and these 5 variables can compare the complete representation of the total amount of information.
After 5 target characteristic variables are obtained, the target characteristic variables are used as input variables, the product quality and the product period are used as output variables to establish an injection molding model, the number of hidden layer neurons is preliminarily set to be 6, and then a three-layer single-hidden-layer ELM extreme learning machine is established, as shown in FIG. 6, FIG. 6 is an ELM model schematic diagram of an automotive interior panel disclosed by the embodiment of the invention.
The input variables are the 5 design variables mentioned above in the orthogonal test after characteristic variable extraction and dimension reduction, namely the mold temperature T mold Melt and fuseBody temperature T melt Pressure maintaining pressure P p Pressure maintaining time t p And a cooling time t c The output variable contains two objective functions: quality (warpage) and efficiency (forming cycle).
In order to determine the number of hidden layer neurons more accurately, the hidden layer neurons are set to 5,6,7,8, and training analysis is performed, and the result is shown in fig. 7 (a), 7 (b), 7 (c), and 7 (d), where 4 graphs correspond to the situations where the hidden layer neurons are 5,6,7, and 8, respectively, and fig. 7 (a) is a training result graph of the first ELM model; FIG. 7 (b) is a graph of the training results of the second ELM model; FIG. 7 (c) is a graph of the training results of the third ELM model; FIG. 7 (d) is a graph of the training results of the fourth ELM model; in the figure, train is training sample data, test is verification sample data, validity is Test sample data, and Y-axis is output error. By comparing the graphs as shown in fig. 7 (a), 7 (b), 7 (c), 7 (d) and 4, it can be seen that premature convergence occurs in the graphs 7 (a), 7 (b) and 7 (c), especially the graph 7 (a) cannot be trained, and the training errors of 7 (c) and 7 (d) are 4.3138 and 2.5654, respectively, which are large. Moreover, 7 (c) and 7 (d) have overfitting phenomena. When the number of hidden layer neurons is 6, i.e., the 7 (b) graph error falls to 0.32977, this error is within an acceptable range, so the number of hidden layer neurons is finally determined to be 6.
In the application, the ELM extreme learning machine mainly functions to perform fitness evaluation on subsequent NSGA ii (non-dominated sorting genetic algorithm) through errors between predicted values and experimental values after neural network training, wherein fig. 8 (a) is a distribution curve diagram of predicted values and experimental values of two objective function values, tc is a forming period experimental value, tcp is a forming period predicted value, warp is a warp deformation experimental value, warp is a warp deformation predicted value, and fig. 8 (b) is an output error curve diagram of two objective functions, for a forming period, the average value of the overall predicted errors is 9.6%, and the average error except extreme points is only 0.016; for the warpage deformation, the overall prediction error is 8.7%, and as can be seen from fig. 8 (b), except for the extreme point with a large individual error, the predicted values and the experimental values of most parts have good compatibility.
Referring to fig. 9, fig. 9 is a schematic flow chart of an ELM learning machine-non-dominated sorting genetic algorithm according to an embodiment of the present invention; as shown in fig. 9, firstly, performing a modular flow analysis in a Moldflow advisor 2017 according to an orthogonal table to obtain initial training data; then, training and predicting objective function values by importing training data into an ELM extreme learning machine, carrying out PARITO non-dominated sorting on samples according to training error results of the ELM extreme learning machine, calculating a fitness value, initializing a population to carry out NSGAII genetic operation until the requirement of an evolution algebra is met, and stopping evolution. Fig. 10 is a randomly generated initial population spatial distribution scattergram, fig. 11 is a randomly generated initial population evolution convergence curve, after the population evolution iteration process has progressed to 100 generations, an error fitness value of the population evolution iteration process begins to tend to be stable, that is, the Pareto optimal solution set at this time also tends to be stable gradually, and fig. 12 is a distribution scattergram of Pareto optimal solutions at the first front end of the population evolution to 100 generations.
On the basis of fig. 11, designers can conveniently find corresponding design variables, i.e., forming process parameters, from various points in the Pareto optimal solution set in the figure, and then substitute Moldflow Adviser2017 for modular flow analysis to check the superiority of the design scheme, and then select relatively optimal process design parameters according to the particularity and requirements of the designed product.
Tables 1 to 5 show the results of the Moldflow simulation analysis of the corresponding 4 Pareto optimal solutions in fig. 12, and it is apparent from fig. 12 that the forming cycle of the product is inevitably prolonged under the condition of ensuring the quality (warpage), so that a compromise scheme is selected as far as possible without special requirements.
TABLE 1-5Pareto optimal solutions (parts)
Selecting a fourth group of solutions in Pareto optimal solution sets in tables 1-5 as optimized injection molding process parameters, and comparing the fourth group of solutions with forming results of unoptimized process parameters, wherein emphasis is placed on differences of quality (buckling deformation) defects and production efficiency (forming period) of products, on one hand, from the aspect of buckling deformation, namely quality problems of the products, buckling variables of the unoptimized process products exceeding the maximum nominal deviation are 21.2%, the buckling deformation exceeding the maximum nominal deviation after optimization is 4.95%, and the ratio of low buckling deformation areas is increased from 57.3% to 74.4%; on the other hand, the molding cycle of the whole product is not optimized, the molding cycle is 40.55s, after the process parameters are optimized, the molding cycle of the product is 32.41s, the whole molding cycle is reduced by about 20.07%, and the production efficiency is improved on the premise of ensuring the product quality. Besides the two main aspects, the simulation results of the product before and after optimization have more differences, and specific parameters are shown in tables 1-6.
Tables 1-6 comparison of the optimization results
The applicability of the algorithm model is verified by analyzing and comparing only a plurality of optimal solutions in the Pareto solution set, and when the Pareto solution set deals with specific design problems, a product designer can choose the Pareto optimal solution with emphasis according to specific design requirements to obtain an optimal design scheme.
The feature selection method, device and storage medium based on the injection molding model provided by the present application are described in detail above. The principles and embodiments of the present application are explained herein using specific examples, which are provided only to help understand the method and the core idea of the present application. It should be noted that, for those skilled in the art, without departing from the principle of the present application, the present application can also make several improvements and modifications, and those improvements and modifications also fall into the protection scope of the claims of the present application.
The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
It should also be noted that, in this specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
Claims (9)
1. A feature selection method based on an injection molding model is characterized by comprising the following steps:
acquiring sample data corresponding to each characteristic variable of an injection molding model, wherein the injection molding model is a mathematical module combining an input variable, an output variable and constraint conditions of injection molding, the input variable is a process parameter for representing injection molding, and the output variable is data for representing quality and efficiency of an injection molding product;
calculating a characteristic value corresponding to each characteristic variable according to each sample data;
selecting a target characteristic value from each characteristic value according to a predefined rule, and taking a characteristic variable corresponding to the target characteristic value as a target characteristic variable;
wherein the target characteristic variables include dwell pressure, melt temperature, cooling time, mold temperature, and dwell time.
2. The injection molding model-based feature selection method according to claim 1, wherein the calculating a feature value corresponding to each of the feature variables according to each of the sample data comprises:
forming a sample matrix by each sample data and calculating a covariance matrix according to the sample matrix;
and calculating the eigenvalue corresponding to each of the characteristic variables by using the covariance matrix.
3. The method of claim 2, wherein the forming each sample data into a sample matrix and calculating a covariance matrix from the sample matrix comprises:
forming the sample data into a sample matrix;
calculating an average eigenvalue value by using the sample data corresponding to the characteristic variables in the sample matrix and the number of the characteristic variables;
calculating the difference value between each sample data and the average characteristic numerical value;
and combining the difference values to obtain the covariance matrix.
4. The injection molding model-based feature selection method of claim 1, wherein after selecting a target feature value from each feature value according to a predefined rule and using a feature variable corresponding to the target feature value as a target feature variable, the method further comprises:
taking the target characteristic variable as an input variable;
taking the production quality and the production efficiency of the injection molding product as output variables;
constructing the injection molding model by using the input variable and the output variable and optimizing the injection molding model;
and solving the optimized injection molding model to obtain an optimal solution of each target characteristic variable.
5. The injection molding model-based feature selection method according to claim 4, wherein the constructing and optimizing the injection molding model using the input variables and the output variables comprises:
constructing the injection molding model by using the input variables and the output variables;
setting the number of various neurons of a hidden layer of the injection molding model;
selecting the optimal neuron number of the injection molding model from the neuron numbers by using test sample data and verification sample data corresponding to the input variable;
and obtaining the optimized injection molding model by using the input variable, the output variable and the hidden layer corresponding to the optimal neuron number.
6. The injection molding model-based feature selection method according to claim 1, wherein the selecting a target feature value from the feature values according to a predefined rule, and using a feature variable corresponding to the target feature value as a target feature variable comprises:
and selecting a characteristic value larger than a target threshold value from all the characteristic values, and taking a characteristic variable corresponding to the characteristic value larger than the target threshold value as a target characteristic variable of the injection molding model.
7. A device is selected to characteristic based on injection moulding model, its characterized in that includes:
the injection molding system comprises a sample data acquisition module, a parameter analysis module and a parameter analysis module, wherein the sample data acquisition module is used for acquiring sample data corresponding to each characteristic variable of an injection molding model, the injection molding model is a mathematical module which combines an input variable, an output variable and constraint conditions of injection molding, the input variable is a process parameter for representing injection molding, and the output variable is data for representing quality and efficiency of an injection molding product;
the characteristic value calculating module is used for calculating a characteristic value corresponding to each characteristic variable according to each sample data;
the target characteristic variable setting module is used for selecting a target characteristic value from each characteristic value according to a predefined rule and taking a characteristic variable corresponding to the target characteristic value as a target characteristic variable;
wherein the target characteristic variables include dwell pressure, melt temperature, cooling time, mold temperature, and dwell time.
8. A device is selected to characteristic based on injection moulding model, its characterized in that includes:
a memory for storing a computer program;
a processor for executing a computer program stored in the memory for implementing the steps of the injection molding model based feature selection method according to any one of claims 1 to 6.
9. A computer-readable storage medium, on which a computer program is stored, which computer program is executable by a processor for implementing the steps of the method for injection molding model-based feature selection according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810796642.5A CN109101698B (en) | 2018-07-19 | 2018-07-19 | Feature selection method and device based on injection molding model and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810796642.5A CN109101698B (en) | 2018-07-19 | 2018-07-19 | Feature selection method and device based on injection molding model and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109101698A CN109101698A (en) | 2018-12-28 |
| CN109101698B true CN109101698B (en) | 2023-04-18 |
Family
ID=64846710
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810796642.5A Active CN109101698B (en) | 2018-07-19 | 2018-07-19 | Feature selection method and device based on injection molding model and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109101698B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11403535B2 (en) * | 2018-12-21 | 2022-08-02 | Industrial Technology Research Institute | Model-based machine learning system |
| CN112115579A (en) * | 2020-08-12 | 2020-12-22 | 江苏师范大学 | Multi-target optimization method for injection molding process parameters of glass fiber reinforced plastics |
| CN113103535A (en) * | 2021-03-17 | 2021-07-13 | 贵州大学 | GA-ELM-GA-based injection molding part mold parameter optimization method |
| CN115619050B (en) * | 2022-12-19 | 2023-03-28 | 绵阳丰瑞尚雅木业有限公司 | Method and system for predicting processing quality of fireproof aluminum-wood veneer and storage medium |
| CN117532885A (en) * | 2024-01-10 | 2024-02-09 | 成都航空职业技术学院 | Intelligent auxiliary system, method and storage medium for 3D printing |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008072047A (en) * | 2006-09-15 | 2008-03-27 | Omron Corp | Model preparing device, analyzing device for process abnormality, and those method and program |
| CN105550744A (en) * | 2015-12-06 | 2016-05-04 | 北京工业大学 | Nerve network clustering method based on iteration |
| CN106597260A (en) * | 2016-12-29 | 2017-04-26 | 合肥工业大学 | Simulation circuit fault diagnosis method based on continuous wavelet analysis and ELM network |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8527223B2 (en) * | 2009-09-11 | 2013-09-03 | University Of Cincinnati | Methods and systems for energy prognosis |
| US8855804B2 (en) * | 2010-11-16 | 2014-10-07 | Mks Instruments, Inc. | Controlling a discrete-type manufacturing process with a multivariate model |
-
2018
- 2018-07-19 CN CN201810796642.5A patent/CN109101698B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008072047A (en) * | 2006-09-15 | 2008-03-27 | Omron Corp | Model preparing device, analyzing device for process abnormality, and those method and program |
| CN105550744A (en) * | 2015-12-06 | 2016-05-04 | 北京工业大学 | Nerve network clustering method based on iteration |
| CN106597260A (en) * | 2016-12-29 | 2017-04-26 | 合肥工业大学 | Simulation circuit fault diagnosis method based on continuous wavelet analysis and ELM network |
Non-Patent Citations (1)
| Title |
|---|
| "PCA和GA-ELM在立体图像质量客观评价中的应用";王光华;《中国优秀硕士学位论文全文数据库 信息科技辑》;20170315;I138-4487，正文第1-57页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109101698A (en) | 2018-12-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109101698B (en) | Feature selection method and device based on injection molding model and storage medium | |
| CN110377948A (en) | A kind of injection parameters Multipurpose Optimal Method | |
| CN110640982A (en) | Multi-target optimization method for injection molding process parameters of thin-wall injection molding part | |
| CN111563347B (en) | Injection molding process parameter optimization method for fiber reinforced composite material | |
| CN112115579A (en) | Multi-target optimization method for injection molding process parameters of glass fiber reinforced plastics | |
| Matin et al. | A CAD/CAE-integrated injection mold design system for plastic products | |
| AlKaabneh et al. | A combined analytical hierarchical process (AHP) and Taguchi experimental design (TED) for plastic injection molding process settings | |
| Meiabadi et al. | Optimization of plastic injection molding process by combination of artificial neural network and genetic algorithm | |
| CN112069610B (en) | Injection molding process parameter optimization method for transparent complex multi-cavity plastic part | |
| US8983878B2 (en) | Computer-implemented method for optimizing an injection-molding process for producing thick-walled components | |
| CN107972243A (en) | Shooting Technique optimization method and Shooting Technique based on SOM neutral nets | |
| Li et al. | Statistical characterization and robust design of RTM processes | |
| CN113103535A (en) | GA-ELM-GA-based injection molding part mold parameter optimization method | |
| Sedighi et al. | Optimisation of gate location based on weld line in plastic injection moulding using computer-aided engineering, artificial neural network, and genetic algorithm | |
| Rai et al. | An intelligent system for predicting HPDC process variables in interactive environment | |
| CN110502779B (en) | Intelligent design method of injection molding mold based on fuzzy reasoning | |
| Kwak et al. | Application of neural network and computer simulation to improve surface profile of injection molding optic lens | |
| CN112884236A (en) | Short-term load prediction method and system based on VDM decomposition and LSTM improvement | |
| Cheng et al. | Optimization of injection mold based on fuzzy moldability evaluation | |
| CN113722992A (en) | Injection molding process parameter multi-target optimization method for injection molding part with insert | |
| Yang et al. | Research on Optimization of Injection Molding Process Parameters of Automobile Plastic Front-End Frame | |
| JP4785587B2 (en) | Molding condition setting method, program, and injection molding machine | |
| CN101853318A (en) | Method for quickly forecasting cooling time in injection molding of plastics | |
| Sandu et al. | Prediction of polymer flow length by coupling finite element simulation with artificial neural network | |
| EP3511149A1 (en) | Curvature deformation prevention design method for resin molded article, program, recording medium, and curvature deformation prevention design apparatus for resin molded article |
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 |