CN114535142A - Data-driven intelligent determination method for dimension qualification of injection molding product - Google Patents
Data-driven intelligent determination method for dimension qualification of injection molding product Download PDFInfo
- Publication number
- CN114535142A CN114535142A CN202210024752.6A CN202210024752A CN114535142A CN 114535142 A CN114535142 A CN 114535142A CN 202210024752 A CN202210024752 A CN 202210024752A CN 114535142 A CN114535142 A CN 114535142A
- Authority
- CN
- China
- Prior art keywords
- injection molding
- data
- model
- classification model
- data set
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 67
- 238000001746 injection moulding Methods 0.000 title claims abstract description 63
- 238000012797 qualification Methods 0.000 title claims abstract description 17
- 238000012216 screening Methods 0.000 claims abstract description 52
- 238000013145 classification model Methods 0.000 claims abstract description 44
- 238000001514 detection method Methods 0.000 claims abstract description 18
- 238000005457 optimization Methods 0.000 claims abstract description 9
- 238000002790 cross-validation Methods 0.000 claims abstract description 8
- 238000004140 cleaning Methods 0.000 claims abstract description 6
- 238000004364 calculation method Methods 0.000 claims description 12
- 238000002347 injection Methods 0.000 claims description 12
- 239000007924 injection Substances 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 4
- 238000013024 troubleshooting Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 description 9
- 230000002950 deficient Effects 0.000 description 4
- 229920003023 plastic Polymers 0.000 description 4
- 239000004033 plastic Substances 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000002844 melting Methods 0.000 description 1
- 230000008018 melting Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000465 moulding Methods 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/34—Sorting according to other particular properties
- B07C5/342—Sorting according to other particular properties according to optical properties, e.g. colour
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/36—Sorting apparatus characterised by the means used for distribution
- B07C5/361—Processing or control devices therefor, e.g. escort memory
-
- 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
Landscapes
- Injection Moulding Of Plastics Or The Like (AREA)
Abstract
The invention discloses an intelligent determination method for the dimensional qualification of an injection molding product based on data driving, which comprises the following steps: s1, collecting process data in the injection molding process, cleaning the collected data set, standardizing the data set and reconstructing a label; s2, determining feature screening dimensionality by using a tree model-based feature screening method and a chi-square detection feature screening method for the obtained data set, screening features, and reducing dimensionality of input data; s3, screening out a classification model by adopting 5-fold cross validation, and evaluating the classification performance of the screened classification model by using an ROC curve and an AUC value; s4, adding L1 regularization into a loss function of the classification model; s5, carrying out Bayesian optimization on the classification model, selecting a hyper-parameter most suitable for the injection molding data set, and improving the classification accuracy of the model; s6, apply to actual determination. The invention is used for solving the problems of low efficiency and low accuracy of manual detection on the dimension qualification rate of injection molding products.
Description
Technical Field
The invention belongs to the technical field of injection molding, and particularly relates to an intelligent determination method for the dimensional qualification of an injection molding product based on data driving.
Background
The plastic has the characteristics of good plasticity, higher impact strength, good chemical stability and the like, so that the plastic products are widely applied to various industries and fields, wherein the plastic products manufactured by the injection molding process account for more than 60 percent. With the continuous update of internet technology, the concept of internet plus intelligent manufacturing is more and more closely related to the injection molding process.
The quality of the injection molding product involves a plurality of factors, and the mutual influence among the factors is that it is desired to directly judge a certain index of the injection molding product, and the traditional method relying on limited experience and a simple formula has the disadvantages of poor stability, low accuracy and low efficiency. Therefore, an intelligent method based on data driving appears in the environment of Internet plus intelligent manufacturing, and the intelligent method based on data driving can not only predict the machine fault of the injection molding machine, but also directly judge some indexes of injection molding products.
The size of injection moulding article is the key point that injection moulding article quality detected, domestic judgement to injection moulding article size qualification mainly relies on the manual work to detect at present stage, the manual work detects and needs the manual work to use the measuring tool to measure injection moulding article's size, not only speed is slow, manpower and material resources are invested in greatly, cause easily to miss and erroneous judgement, consequently regard as the drive with the data in the injection moulding course of working, it is the problem that awaits the solution to establish the classification model that is fit for the data set of moulding plastics to classify injection moulding article's size.
Disclosure of Invention
The invention mainly aims to overcome the defects of the prior art and provides an intelligent determination method for the dimensional qualification of an injection molding product based on data driving, so as to solve the problems of low efficiency and low accuracy of manual detection of the dimensional qualification of the injection molding product.
In order to achieve the purpose, the invention adopts the following technical scheme:
the data-driven intelligent determination method for the dimensional qualification of the injection molding product comprises the following steps:
s1, collecting process data in the injection molding process, cleaning the collected data set, standardizing the data set and reconstructing a label;
s2, determining feature screening dimensionality and screening features by using the acquired data set and using a tree model-based feature screening method and a chi-square detection feature screening method;
s3, screening out a classification model by adopting 5-fold cross validation, and evaluating the classification performance of the screened size classification model by using an ROC curve and an AUC value to obtain an optimal size classification model;
s4, adding L1 regularization into a loss function of the optimal size classification model to prevent overfitting;
s5, carrying out Bayesian optimization on the optimal size classification model, selecting the best hyper-parameter suitable for the injection molding data set, and improving the classification accuracy of the classification model;
s6, the optimal size classification model is applied to intelligent judgment of the size qualification of the injection molding products, the importance degrees of the characteristics influencing the sizes of the injection molding products are sorted, when the sizes of the injection molding products are in problem, the judgment of a fault area can be facilitated, and the time for troubleshooting is shortened.
Further, in step S1, the data cleansing includes filling or deleting missing values, constructing difference features, and deleting invariant features.
Further, in step S1, the normalization process specifically includes:
linearly varying the raw acquisition data using min-max normalization to map the resulting values between [0,1], the calculation formula being:
where x denotes the original value, x*The normalized values are shown, min represents the smallest value of the entire feature, and max represents the largest value of the entire feature.
Further, in step S1, the reconfiguration label specifically includes:
when the dimension data of the injection molded article is within the allowable tolerance range, the label is marked as 1, and when the dimension data of the injection molded article is outside the allowable tolerance range, the label is marked as 0.
Further, in step S2, the feature screening based on the tree model specifically includes:
the importance degree of the features is sorted by the variable quantity of the purity of the kini, and the calculation formula is as follows:
ΔGini(A)＝Gini(D)-Gini(D-A)
wherein Gini (D) represents the degree of purity of the data set D, PiRepresenting the probability of the total number of samples of the category i, Δ Gini (D-A) is the decrement of the degree of purity of the kiney after the feature A is added, the larger the value is, the larger the correlation of the feature A and the label is, and Gini (D-A) determines the degree of purity of the kiney after the feature A for the data set D;
the chi-square detection feature screening method specifically comprises the following steps:
the importance degree of the features is sorted according to the magnitude of the chi-square value, and the calculation formula is as follows:
wherein ,xiRepresenting the actual value and E the theoretical value.
Further, in step S2, the dimension of feature screening is specifically determined as:
when a tree model-based feature screening method is used, the average value of Δ Gini is taken as a threshold; when the chi-square detection feature screening method is used, the average value of chi-square values is taken as a threshold.
Further, in step S3, the 5-fold cross validation method specifically includes:
the injection molding data set was divided into 5 data sets, the average of 5 MSEs was calculated, and the average MSE was used to screen existing classification models commonly used for industrial data.
Further, in step S3, the classification performance of the screened size classification model is evaluated by using the ROC curve and the AUC value, and specifically:
the following indices were evaluated: true class TP, missing class FP, false class FN, true negative class TN, the ratio TPR of samples classified as non-defective to all non-defective, the ratio FPR of samples classified as non-defective to all non-defective, and the Accuracy Accuracy.
Further, step S4 is specifically:
adding L1 regularization to the loss function of the LR model to prevent overfitting of the model, wherein the formula of the loss function is as follows:
wherein | ω | purple1For the L1 regularization term, α is the coefficient before the regularization term.
Further, in step S5, a bayesian optimization method is used to optimize the hyperparameters of the size classification model to obtain the hyperparameters most suitable for the size data set of the injection molded product, and the best size classification model is obtained, so that the classification accuracy is highest.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the method takes an injection molding process data set as a drive, and combines a machine learning classification model to directly and intelligently judge the qualification of the dimension of the injection molded product, wherein the dimension data in the data set is reconstructed on the basis, and qualified data and unqualified data are taken as labels; comparing different feature screens and the combination of the models, and screening out a classification model most suitable for the size of the injection molding product; for the phenomenon that the monitoring parameters in the injection molding data set can be linearly or nonlinearly represented by process parameters to cause model overfitting, L1 regularization is added into a loss function of the model, and a sparse matrix is used for preventing model overfitting; finally, Bayesian optimization is carried out on the whole model, and a hyper-parameter value most suitable for the injection molding data set is selected, so that the classification accuracy of the classification model is improved; the importance of the features influencing the size is sorted by the feature screening method based on the tree model, when the size of the injection molding product is unqualified, factors influencing the large size of the injection molding product can be preferentially checked, and the overhaul time is shortened.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a flow chart of 5-fold cross validation;
FIG. 3 is a classification model ROC curve combined with tree model based feature screening;
FIG. 4 is a classification model ROC curve combined with chi-square detection feature screening;
FIG. 5 is a ranking of feature importance based on tree model feature screening.
Detailed Description
The present invention will be described in further detail with reference to examples and drawings, but the present invention is not limited thereto.
Examples
As shown in FIG. 1, the method for intelligently judging the dimensional qualification of the injection molding product based on data driving comprises the following steps:
s1, collecting process data in the injection molding process, cleaning the collected data set, standardizing the data set and reconstructing a label;
the data cleaning comprises filling or deleting missing values, constructing difference characteristics, deleting invariant characteristics and the like. In this embodiment, the missing values in each feature are counted, when the missing value in the feature is greater than 50%, the feature is deleted, and when the missing value is less than 50%, the missing value is filled by using the average value.
The standardization treatment specifically comprises the following steps:
linearly varying the raw data using min-max normalization to map the resulting values between [0,1], the calculation formula being:
where x denotes the original value, x*The normalized values are shown, min represents the smallest value of the entire feature, and max represents the largest value of the entire feature.
The reconstruction tag specifically comprises:
when the dimension data of the injection molded article is within the allowable tolerance range, the label is marked as 1, and when the dimension data of the injection molded article is outside the allowable tolerance range, the label is marked as 0.
S2, determining feature screening dimensions by using the collected data set and using a feature screening method based on a tree model and a chi-square detection feature screening method, screening features, and reducing the dimensions of input data;
the characteristic screening based on the tree model specifically comprises the following steps:
the importance degree of the features is sorted by the variable quantity of the purity of the kini, and the calculation formula is as follows:
ΔGini(A)＝Gini(D)-Gini(D-A)
wherein Gini (D) represents the degree of purity of the data set D, PiRepresenting the probability of the total number of samples of the category i, Δ Gini (D-A) is the decrement of the degree of purity of the kiney after the feature A is added, the larger the value is, the larger the correlation of the feature A and the label is, and Gini (D-A) determines the degree of purity of the kiney after the feature A for the data set D;
the chi-square detection feature screening method specifically comprises the following steps:
the importance degree of the features is sorted according to the magnitude of the chi-square value, and the calculation formula is as follows:
wherein ,xiRepresenting the actual value and E the theoretical value.
In this embodiment, the dimension for determining feature screening specifically is:
when a feature screening method based on a tree model is used, screening out features larger than a threshold value by taking the average value of delta Gini as the threshold value to obtain 19-dimensional features; when the chi-square detection feature screening method is used, features with chi-square values larger than a threshold value are screened out by taking the average value of the chi-square values as the threshold value to obtain 19-dimensional features, feature dimensions obtained by combining the two feature screening methods are all 19, and finally the screened feature dimension is determined to be 19.
S3, screening out a classification model by adopting 5-fold cross validation, and evaluating the classification performance of the screened classification model by using an ROC curve and an AUC value to obtain an optimal classification model;
the 5-fold cross validation method specifically comprises the following steps:
and (3) cutting the injection molding data set into 5 data sets, calculating the average value of MSE for 5 times, and screening the classification model by using the average MSE.
In this embodiment, as shown in fig. 2, the training set is divided into 5 data sets, one data set is taken out each time as the test set, the other 4 data sets are taken as the training set, 5 times of calculation are performed in total, one MSE can be obtained each time, and finally, an average value of the 5 MSEs is calculated, where the calculation formula is:
the average MSE of the 7 classification models was obtained as shown in table 1 below.
TABLE 1
And calculating the average value of the seven models to be 0.951, taking 0.951 as a threshold value, selecting the classification model larger than the threshold value, and finishing primary screening.
And evaluating the classification performance of the screened classification model by using the ROC curve and the AUC value, specifically comprising the following steps:
evaluating the following indexes, namely a true class TP, a missing judgment FP, a false judgment FN, a true negative class TN, a ratio TPR of samples of qualified products classified into qualified products to all qualified products, a ratio FPR of samples of unqualified products classified into qualified products to all unqualified products and an Accuracy Accuracy; wherein, the index calculation formula is as follows:
in step S4, it is found that the ROC curve of the LR model under the feature screening method based on the tree model and the feature screening combination based on the chi-square detection is the one shown in fig. 3, and the ROC curve of the LR model under the feature screening method based on the chi-square detection and the feature screening combination based on the chi-square detection is the one shown in fig. 4, and comparing the two sets of ROC curves, the ROC curve of the LR model under the feature screening method based on the tree model and the feature screening combination based on the chi-square detection is the closest (0, 1) point, which indicates that the LR model is more suitable for the size classification of the injection molded product. The obtained AUC values based on the two feature screens are shown in the following table 2, and the comparison of the AUC values under the two feature screens shows that the AUC values of each model under the feature screen based on the tree model are generally higher than the feature screen based on chi-square detection, so that the feature screening method based on the tree model is more suitable for the classification of the dimension qualification of the injection molding products.
TABLE 2
S4, adding L1 regularization into a loss function of the classification model to prevent overfitting;
because the injection molding data set is easy to generate the problem of multiple collinearity, the monitoring parameters in the injection molding data can be linearly or nonlinearly represented by machine parameters, such as the pressure of a nozzle related to an injection starting point, the maximum injection pressure, the glue melting time and the like; the temperature in the mold cavity is related to the mold temperature, the barrel temperature, the hot runner temperature, etc. Since mutual interference among these parameters is likely to cause overfitting of the model when the model is fitted, L1 regularization is added to the loss function of the LR model, and overfitting of the model is effectively prevented by the sparse matrix, where the formula of the loss function is:
wherein | ω | purple1For the L1 regularization term, α is the coefficient before the regularization term.
S5, carrying out Bayesian optimization on the classification model, selecting a hyper-parameter most suitable for an injection molding data set, and improving the classification accuracy of the model;
and optimizing the final super-parameters of the model, such as a regular term coefficient, a type weight parameter and the like, by using a Bayesian optimization method, mainly calculating the mean value and the variance of a function value at each point through Gaussian process regression, constructing an acquisition function according to the mean value and the variance, determining the point at which each iteration is to be sampled, and obtaining the super-parameter most suitable for the size data set of the injection molding product through continuous iteration to obtain the best size classification model so as to ensure that the classification accuracy is highest. As shown in table 3 below, is a comparison before and after optimization.
TABLE 3
S6, the optimal classification model is applied to intelligent judgment of the dimension qualification of the injection molding products, the importance degrees of the features influencing the dimension of the injection molding products are sorted, when the dimension of the injection molding products is in a problem, the fault area can be judged, and the fault troubleshooting time is shortened.
According to the obtained optimal injection molding product size classification model, the importance degree of the features is sorted by a feature screening method based on a tree model in the model, as shown in fig. 5, it can be seen that the data collected by a sensor is closer to the actual production, so that the size is more sensitive to the data collected by the sensor, and the nozzle pressure can be seen from the graph as the factor influencing the maximum size, and then the pressure in a mold cavity and the circulating water temperature of a mold temperature machine, so that when the size of the injection molding product is in a problem, whether the nozzle is damaged or not, whether the mold is damaged or not and whether the mold temperature machine normally operates or not are preferentially checked to help troubleshooting, and the overhaul time is shortened.
It should also be noted that in this specification, terms such as "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 an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. The data-driven intelligent determination method for the dimensional qualification of the injection molding product is characterized by comprising the following steps of:
s1, collecting process data in the injection molding process, cleaning the collected data set, standardizing the data set and reconstructing a label;
s2, determining feature screening dimensionality and screening features by using the acquired data set and using a tree model-based feature screening method and a chi-square detection feature screening method;
s3, screening out a classification model by adopting 5-fold cross validation, and evaluating the classification performance of the screened size classification model by using an ROC curve and an AUC value to obtain an optimal size classification model;
s4, adding L1 regularization into a loss function of the optimal size classification model to prevent overfitting;
s5, carrying out Bayesian optimization on the optimal size classification model, selecting the best hyper-parameter suitable for the injection molding data set, and improving the classification accuracy of the classification model;
s6, the optimal size classification model is applied to intelligent judgment of the size qualification of the injection molding products, the importance degrees of the characteristics influencing the sizes of the injection molding products are sorted, when the sizes of the injection molding products are in problem, the judgment of a fault area can be facilitated, and the time for troubleshooting is shortened.
2. The intelligent determination method for the dimensional acceptability of injection molding products based on data driving as claimed in claim 1 wherein in step S1, the data cleaning includes filling or deleting missing values, constructing difference features and deleting invariant features.
3. The method for intelligently determining the dimensional acceptability of an injection molded product based on data driving according to claim 1, wherein in step S1, the standardization process is specifically:
linearly varying the raw acquisition data using min-max normalization to map the resulting values between [0,1], the calculation formula being:
where x denotes the original value, x*The normalized values are shown, min represents the smallest value of the entire feature, and max represents the largest value of the entire feature.
4. The intelligent determination method for the dimensional acceptability of the injection molding product based on the data driving as claimed in claim 1, wherein in step S1, the reconfiguration label is specifically:
when the dimension data of the injection molded article is within the allowable tolerance range, the label is marked as 1, and when the dimension data of the injection molded article is outside the allowable tolerance range, the label is marked as 0.
5. The intelligent determination method for the dimensional acceptability of the injection molding product based on the data driving as claimed in claim 1, wherein in step S2, the feature screening based on the tree model is specifically:
the importance degree of the characteristics is ranked through the variable quantity of the purity of the kunity, and the calculation formula is as follows:
ΔGini(A)＝Gini(D)-Gini(D-A)
wherein Gini (D) represents the degree of purity of the data set D, PiRepresenting the probability of the total number of samples of the category i, Δ Gini (D-A) is the decrement of the degree of purity of the kiney after the feature A is added, the larger the value is, the larger the correlation of the feature A and the label is, and Gini (D-A) determines the degree of purity of the kiney after the feature A for the data set D;
the chi-square detection feature screening method specifically comprises the following steps:
the importance degree of the features is sorted according to the magnitude of the chi-square value, and the calculation formula is as follows:
wherein ,xiIndicating the actual value and E the theoretical value.
6. The intelligent determination method for the dimensional acceptability of the injection molding product based on the data driving as claimed in claim 5, wherein in step S2, the dimension for determining the feature screening is specifically:
when a tree model-based feature screening method is used, the average value of Δ Gini is taken as a threshold; when the chi-squared detection feature screening method is used, the average of chi-squared values is taken as a threshold.
7. The intelligent determination method for the dimensional qualification of the injection-molded product based on the data driving as claimed in claim 1, wherein in step S3, the 5-fold cross validation method specifically comprises:
the injection molding data set was divided into 5 data sets, the average of 5 MSEs was calculated, and the average MSE was used to screen existing classification models commonly used for industrial data.
8. The intelligent determination method for the dimensional acceptability of the injection molding product based on the data driving as claimed in claim 1, wherein in step S3, the classification performance of the screened size classification model is evaluated by using the ROC curve and the AUC values, specifically:
the following indices were evaluated: the method comprises the following steps of real class TP, missed judgment FP, false judgment FN, true and negative class TN, the ratio TPR of samples of qualified products to all qualified products, the ratio FPR of samples of unqualified products to qualified products and Accuracy Accuracy.
9. The intelligent determination method for the dimensional acceptability of the injection molding product based on the data driving as claimed in claim 1, wherein the step S4 is specifically:
adding L1 regularization to the loss function of the LR model to prevent overfitting of the model, wherein the formula of the loss function is as follows:
wherein | ω | calucity1For the L1 regularization term, α is the coefficient before the regularization term.
10. The intelligent determination method for the dimensional acceptability of the injection molding product based on data driving as claimed in claim 1, wherein in step S5, the super parameters of the dimensional classification model are optimized by using a bayesian optimization method to obtain the super parameters most suitable for the dimensional data set of the injection molding product, and the best dimensional classification model is obtained, so that the classification accuracy is highest.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210024752.6A CN114535142B (en) | 2022-01-11 | 2022-01-11 | Intelligent judgment method for size qualification of injection molding product based on data driving |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210024752.6A CN114535142B (en) | 2022-01-11 | 2022-01-11 | Intelligent judgment method for size qualification of injection molding product based on data driving |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114535142A true CN114535142A (en) | 2022-05-27 |
| CN114535142B CN114535142B (en) | 2023-09-26 |
Family
ID=81669969
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210024752.6A Active CN114535142B (en) | 2022-01-11 | 2022-01-11 | Intelligent judgment method for size qualification of injection molding product based on data driving |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114535142B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100153332A1 (en) * | 2008-12-17 | 2010-06-17 | Rollins John B | Data mining model interpretation, optimization, and customization using statistical techniques |
| CN109257383A (en) * | 2018-11-09 | 2019-01-22 | 中国人民解放军战略支援部队信息工程大学 | A kind of BGP method for detecting abnormality and system |
| CN113379135A (en) * | 2021-06-18 | 2021-09-10 | 浙江大学 | Intelligent production line product quality low-delay integrated prediction method and system based on cloud edge collaborative computing |
-
2022
- 2022-01-11 CN CN202210024752.6A patent/CN114535142B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100153332A1 (en) * | 2008-12-17 | 2010-06-17 | Rollins John B | Data mining model interpretation, optimization, and customization using statistical techniques |
| CN109257383A (en) * | 2018-11-09 | 2019-01-22 | 中国人民解放军战略支援部队信息工程大学 | A kind of BGP method for detecting abnormality and system |
| CN113379135A (en) * | 2021-06-18 | 2021-09-10 | 浙江大学 | Intelligent production line product quality low-delay integrated prediction method and system based on cloud edge collaborative computing |
Non-Patent Citations (1)
| Title |
|---|
| 宋建等: "基于特征选择和贝叶斯优化LightGBM 的注塑制品尺寸预测", 工程塑料应用 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114535142B (en) | 2023-09-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111539553B (en) | Wind turbine generator fault early warning method based on SVR algorithm and off-peak degree | |
| CN111967486A (en) | Complex equipment fault diagnosis method based on multi-sensor fusion | |
| US11966847B2 (en) | Method for constructing efficient product surface defect detection model based on network collaborative pruning | |
| CN108875772B (en) | Fault classification model and method based on stacked sparse Gaussian Bernoulli limited Boltzmann machine and reinforcement learning | |
| CN111768000A (en) | Industrial process data modeling method for online adaptive fine-tuning deep learning | |
| CN109783898A (en) | A kind of intelligent optimization method of injection molding manufacturing technique parameter | |
| CN113821976A (en) | Lithium battery fault diagnosis modeling method based on integrated algorithm | |
| CN112287980B (en) | Power battery screening method based on typical feature vector | |
| CN113255591A (en) | Bearing fault diagnosis method based on random forest and fusion characteristics | |
| CN113515512A (en) | Quality control and improvement method for industrial internet platform data | |
| CN112749840A (en) | Method for acquiring reference value of energy efficiency characteristic index of thermal power generating unit | |
| CN110263380B (en) | Spinning process cascade modeling subsection interval parameter configuration method | |
| CN115358297A (en) | Injection molding machine abnormity detection method and system based on improved MKECA method | |
| CN115157601A (en) | Injection molding machine and detection method thereof | |
| CN110032069B (en) | Polyester fiber spinning process segmented parameter configuration method based on error compensation | |
| CN114535142A (en) | Data-driven intelligent determination method for dimension qualification of injection molding product | |
| CN1975706A (en) | Cigarette organoleptic quality qualitative index estimating method | |
| CN113449792A (en) | Method for nondestructive rapid detection of food quality | |
| US11816556B1 (en) | Method for predicting air quality index (AQI) based on a fusion model | |
| CN110910528B (en) | Method and device for predicting tensile strength of paper sheet | |
| CN113177578A (en) | Agricultural product quality classification method based on LSTM | |
| CN111696631A (en) | Method for quantitatively measuring relative greenness of various plastic packaging materials | |
| CN114239378A (en) | Injection molding product size prediction method based on custom LightGBM model loss | |
| CN115510740A (en) | Aero-engine residual life prediction method based on deep learning | |
| CN114266337A (en) | Intelligent tobacco leaf grading model based on residual error network and grading method using model |
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 |