CN113255778A - Welding spot quality detection method and device based on multi-model fusion and storage medium - Google Patents
Welding spot quality detection method and device based on multi-model fusion and storage medium Download PDFInfo
- Publication number
- CN113255778A CN113255778A CN202110590395.5A CN202110590395A CN113255778A CN 113255778 A CN113255778 A CN 113255778A CN 202110590395 A CN202110590395 A CN 202110590395A CN 113255778 A CN113255778 A CN 113255778A
- Authority
- CN
- China
- Prior art keywords
- welding spot
- model
- quality detection
- welding
- training
- 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.)
- Pending
Links
- 238000003466 welding Methods 0.000 title claims abstract description 271
- 238000001514 detection method Methods 0.000 title claims abstract description 163
- 230000004927 fusion Effects 0.000 title claims abstract description 46
- 238000003860 storage Methods 0.000 title claims abstract description 14
- 238000000034 method Methods 0.000 claims abstract description 109
- 238000012549 training Methods 0.000 claims abstract description 105
- 230000008569 process Effects 0.000 claims abstract description 87
- 238000012360 testing method Methods 0.000 claims abstract description 43
- 238000011156 evaluation Methods 0.000 claims abstract description 34
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 21
- 238000010801 machine learning Methods 0.000 claims abstract description 12
- 229910000679 solder Inorganic materials 0.000 claims description 26
- 238000012795 verification Methods 0.000 claims description 24
- 238000012545 processing Methods 0.000 claims description 18
- 238000005070 sampling Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 17
- 238000010606 normalization Methods 0.000 claims description 11
- 238000007689 inspection Methods 0.000 claims description 10
- 238000010276 construction Methods 0.000 claims description 5
- 238000010200 validation analysis Methods 0.000 claims 1
- 230000006872 improvement Effects 0.000 abstract description 6
- 230000009286 beneficial effect Effects 0.000 description 6
- 230000007547 defect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000007637 random forest analysis Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000012804 iterative process Methods 0.000 description 4
- 238000007477 logistic regression Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000007500 overflow downdraw method Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000001808 coupling effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Investigating And Analyzing Materials By Characteristic Methods (AREA)
Abstract
The invention discloses a method, a device and a storage medium for detecting the quality of a welding spot based on multi-model fusion, wherein the method comprises the following steps: acquiring a welding spot process parameter data set in the resistance spot welding process, and dividing the welding spot process parameter data set into a training set and a testing set according to a preset proportion; constructing a plurality of single welding spot quality detection models by adopting a machine learning algorithm, and inputting a training set into each single welding spot quality detection model for training; evaluating each trained single welding spot quality detection model according to a preset evaluation index, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; and inputting the test set into each base classifier, and fusing output results of all the base classifiers according to a preset rule to obtain a welding spot quality detection result. The embodiment of the invention can effectively prevent the model from being over-fitted, has good generalization capability, improves the robustness of the model, and further realizes the improvement of the accuracy and the detection efficiency of the quality detection of the welding spot.
Description
Technical Field
The invention relates to the technical field of resistance spot welding, in particular to a method and a device for detecting the quality of a welding spot based on multi-model fusion and a storage medium.
Background
At present, the automobile industry mainly adopts manual spot inspection methods such as chiseling inspection, total damage inspection and ultrasonic inspection to detect the quality of welding points of an automobile body, but the spot inspection methods have the problems of narrow coverage of the detected welding points, low spot inspection frequency and the like, so that a large number of unqualified welding points are easy to flow out, and potential safety hazards are caused. In order to solve the problem, the automobile industry is exploring a method for detecting the quality of welding spots by analyzing welding spot process parameters through machine learning, so that the quality of the welding spots of an automobile body is completely detected, and potential safety hazards caused by unqualified welding spots are eliminated.
Disclosure of Invention
The technical problem to be solved by the embodiments of the present invention is to provide a method, an apparatus and a storage medium for detecting solder joint quality based on multi-model fusion, which can effectively prevent model overfitting, have good generalization capability, improve the robustness of the model, and further improve the accuracy and detection efficiency of solder joint quality detection.
In order to achieve the above object, an embodiment of the present invention provides a method for detecting quality of a solder joint based on multi-model fusion, including:
acquiring a welding spot process parameter data set in the resistance spot welding process, and dividing the welding spot process parameter data set into a training set and a testing set according to a preset proportion;
establishing a plurality of single welding spot quality detection models by adopting a machine learning algorithm, inputting the training set into each single welding spot quality detection model, and training each single welding spot quality detection model;
evaluating each trained single welding spot quality detection model according to a preset evaluation index, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; the evaluation indexes comprise accuracy, missing detection rate and false detection rate;
and inputting the test set into each base classifier, and fusing output results of all the base classifiers according to a preset rule to obtain a welding spot quality detection result.
As an improvement of the above scheme, the acquiring a welding spot process parameter data set in a resistance spot welding process, and dividing the welding spot process parameter data set into a training set and a test set according to a preset proportion specifically includes:
collecting welding spot process parameter data in the resistance spot welding process; the welding spot process parameter data comprise a resistance value, a current value and a heat value;
arranging the resistance value, the current value and the heat value into time sequence data according to a time sequence, and obtaining an expression of each welding spot data x as follows: x ═ xt1,2, ·, n; wherein n represents the number of time series, xtThe t-th feature representing x;
summarizing the obtained data of each welding spot, and establishing a welding spot process parameter data set X, X { (X)1,y1),(x2,y2),...,(xN,yN) }; wherein N represents the number of weld data samples, xiRepresenting single solder joint data, yiThe quality of the solder joint is represented,the quality of the welding spots is divided into qualified welding spots and unqualified welding spots;
and dividing the welding spot process parameter data set X into a training set and a testing set according to a preset proportion.
As an improvement of the above scheme, the method further comprises:
carrying out data normalization processing on the welding spot process parameter data set, and compressing the welding spot process parameter data to a preset range;
performing upsampling processing on the training set to enable the number of positive samples and the number of negative samples in the training set to be equal; the positive sample is a sample with qualified welding spot quality, and the negative sample is a sample with unqualified welding spot quality.
As an improvement of the above scheme, the performing upsampling processing on the training set to make the number of positive samples and the number of negative samples in the training set equal specifically includes:
determining a sampling multiplying power N according to the proportion of the positive sample to the negative sample in the training set;
for each few class sample x, selecting a plurality of samples from K neighbors of the sample x;
randomly selecting a sample x from a plurality of samples selected in K adjacent tonAccording to the formula xnew＝x+rand(0,1)·(xn-x) constructing a new sample;
and repeating the sampling multiplying power for N times in the construction steps so as to enable the number of the positive samples and the number of the negative samples in the training set to be equal.
As an improvement of the above solution, the welding spot process parameter data set further includes a verification set, and then the training of each single welding spot quality inspection model further includes:
inputting the training set into training models with different hyper-parameters for training for each single welding spot quality detection model;
inputting the verification set into a training model of each hyper-parameter for verification, and selecting the hyper-parameters which are well represented in the verification set;
and inputting the test set into a single welding spot quality detection model with good performance and super-parameters so as to evaluate the detection accuracy of the single welding spot quality detection model.
As an improvement of the above scheme, the inputting the test set into each of the base classifiers and fusing the output results of all the base classifiers according to a preset rule to obtain a solder joint quality detection result specifically includes:
inputting the test set into the base classifier to obtain an output result of the base classifier; wherein the output result comprises a pass and a fail;
and counting the output results, and taking the output results which account for most as the detection results of the quality of the welding spots according to a rule that a minority obeys most.
The embodiment of the invention also provides a welding spot quality detection device based on multi-model fusion, which comprises:
the acquisition module is used for acquiring a welding spot process parameter data set in the resistance spot welding process and dividing the welding spot process parameter data set into a training set, a testing set and a verification set;
the single model training module is used for constructing a single welding spot quality detection model by adopting a single machine learning algorithm, inputting the training set into the single welding spot quality detection model and training the single welding spot quality detection model;
the evaluation module is used for evaluating the single welding spot quality detection model through preset evaluation indexes;
and the welding spot quality detection module is used for constructing a multi-model fusion algorithm according to the evaluation result so as to detect the welding spot quality.
Further, the apparatus further comprises:
the normalization module is used for carrying out data normalization processing on the welding spot process parameter data set and compressing the welding spot process parameter data to a preset range;
the up-sampling module is used for performing up-sampling processing on the training set so as to enable the number of positive samples and the number of negative samples in the training set to be equal; the positive sample is a sample with qualified welding spot quality, and the negative sample is a sample with unqualified welding spot quality.
The embodiment of the invention also provides a welding spot quality detection device based on multi-model fusion, which comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the welding spot quality detection device based on multi-model fusion realizes any one of the above welding spot quality detection methods based on multi-model fusion.
The embodiment of the invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program runs, the apparatus where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned weld spot quality detection methods based on multi-model fusion.
Compared with the prior art, the welding spot quality detection method, the welding spot quality detection device and the storage medium based on multi-model fusion have the advantages that: the method comprises the steps of obtaining a welding spot technological parameter data set in the resistance spot welding process, and dividing the welding spot technological parameter data set into a training set and a testing set according to a preset proportion; establishing a plurality of single welding spot quality detection models by adopting a machine learning algorithm, inputting the training set into each single welding spot quality detection model, and training each single welding spot quality detection model; evaluating each trained single welding spot quality detection model according to a preset evaluation index, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; the evaluation indexes comprise accuracy, missing detection rate and false detection rate; and inputting the test set into each base classifier, and fusing output results of all the base classifiers according to a preset rule to obtain a welding spot quality detection result. The embodiment of the invention adopts a multi-feature fusion method, and establishes a model through a plurality of parameters related to the quality of the welding spot, thereby being more beneficial to judging the quality of the welding spot; the detection results of the multiple models are fused, the defect of poor generalization capability of a single model is overcome, the models are not easy to over-fit, the robustness of the models is improved, the noise resistance is good, and the accuracy and the detection efficiency of the quality detection of the welding spots are improved. In addition, the adopted base classifier is simple to realize, and the defects that the neural network algorithm searches along the negative gradient direction in the iterative process, can not converge to the global optimal solution at a higher speed and has a low training speed are avoided.
Drawings
FIG. 1 is a schematic flow chart of a method for detecting quality of a solder joint based on multi-model fusion according to a preferred embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a welding spot quality detection apparatus based on multi-model fusion according to a preferred embodiment of the present invention;
FIG. 3 is a schematic structural diagram of another preferred embodiment of a welding spot quality detection apparatus based on multi-model fusion provided by 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.
Referring to fig. 1, fig. 1 is a schematic flowchart illustrating a method for detecting quality of a solder joint based on multi-model fusion according to a preferred embodiment of the present invention. The welding spot quality detection method based on multi-model fusion comprises the following steps:
s1, acquiring a welding spot process parameter data set in the resistance spot welding process, and dividing the welding spot process parameter data set into a training set and a testing set according to a preset proportion;
s2, constructing a plurality of single welding spot quality detection models by adopting a machine learning algorithm, inputting the training set into each single welding spot quality detection model, and training each single welding spot quality detection model;
s3, evaluating each trained single welding spot quality detection model according to a preset evaluation index, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; the evaluation indexes comprise accuracy, missing detection rate and false detection rate;
and S4, inputting the test set into each base classifier, and fusing output results of all base classifiers according to a preset rule to obtain a welding spot quality detection result.
Specifically, in the embodiment, a welding spot process parameter data set in a resistance spot welding process in a vehicle body manufacturing process is obtained, and the welding spot process parameter data set is divided into a training set and a testing set according to a preset proportion; then, a single machine learning algorithm is adopted to construct a plurality of single welding spot quality detection models with multiple fused characteristics for current characteristics, resistance characteristics and heat characteristics, the training set is input into each single welding spot quality detection model, and each single welding spot quality detection model is trained; secondly, evaluating each trained single welding spot quality detection model according to a preset evaluation index, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; the evaluation indexes comprise accuracy, missing detection rate and false detection rate; and finally, inputting the test set into each base classifier, and fusing output results of all the base classifiers according to a preset rule to obtain a welding spot quality detection result.
It should be noted that, because the resistance spot welding process is influenced by multiple factors, the quality of the welding spot has the characteristic of high nonlinearity and multi-parameter coupling effect, so that the parameters such as resistance, current and the like and the spot welding quality are not simple linear relations, and larger errors are inevitably brought by only carrying out the quality detection of the welding spot through a certain characteristic. Moreover, the quality of the welding spot is detected from a single aspect, which is not beneficial to describing the characteristics of the welding nugget, and from the algorithm point of view, the model is easy to converge to the local optimum, which is not beneficial to the detection of the quality of the welding spot by the model. Therefore, the embodiment of the invention adopts a multi-feature fusion method, namely, a model is established through a plurality of parameters related to the quality of the welding spot, thereby being more beneficial to judging the quality of the welding spot. Meanwhile, due to the poor generalization capability of a single model, the model is easy to over-fit, so that the robustness of the model is poor; if the neural network algorithm is adopted, the model is searched along the negative gradient direction in the iterative process, the global optimal solution cannot be converged at a high speed, the training speed is low, and real-time processing on a production line is not facilitated. Therefore, the embodiment of the invention fuses the detection results of multiple models, improves the defect of poor generalization capability of a single model, ensures that the model is not easy to over-fit, improves the robustness of the model and has good anti-noise capability. In addition, the adopted base classifier is simple to realize, and the defects that the neural network algorithm searches along the negative gradient direction in the iterative process, can not converge to the global optimal solution at a higher speed and has a low training speed are avoided.
In another preferred embodiment, the S1, acquiring a welding spot process parameter data set during resistance spot welding, and dividing the welding spot process parameter data set into a training set and a test set according to a preset ratio, specifically including:
s101, collecting welding spot process parameter data in a resistance spot welding process; the welding spot process parameter data comprise a resistance value, a current value and a heat value;
s102, arranging the resistance value, the current value and the heat value into time sequence data according to a time sequence, and obtaining an expression of each welding spot data x as follows: x ═ xt1,2, ·, n; wherein n represents the number of time series, xtThe t-th feature representing x;
s103, summarizing the obtained welding spot data, and establishing a welding spot process parameter data set X, wherein X is { (X)1,y1),(x2,y2),...,(xN,yN) }; wherein N represents the number of weld data samples, xiRepresenting single solder joint data, yiRepresenting the quality of welding spots, wherein the quality of the welding spots is divided into qualified welding spots and unqualified welding spots;
and S104, dividing the welding spot process parameter data set X into a training set and a testing set according to a preset proportion.
Specifically, welding spot process parameter data in the resistance spot welding process are collected; the welding spot process parameter data comprise a resistance value, a current value and a heat value; supply the electricityArranging the resistance value, the current value and the heat value into time sequence data according to a time sequence, and obtaining an expression of each welding spot data x as follows: x ═ xt1,2, ·, n; wherein n represents the number of time series, xtThe t-th feature representing x; summarizing the obtained data of each welding spot, and establishing a welding spot process parameter data set X, X { (X)1,y1),(x2,y2),...,(xN,yN) }; wherein N represents the number of weld data samples, xiRepresenting single solder joint data, yiRepresenting the quality of welding spots, wherein the quality of the welding spots is divided into qualified welding spots and unqualified welding spots; and dividing the welding spot process parameter data set X into a training set and a testing set according to a preset proportion.
In yet another preferred embodiment, the method further comprises:
carrying out data normalization processing on the welding spot process parameter data set, and compressing the welding spot process parameter data to a preset range;
performing upsampling processing on the training set to enable the number of positive samples and the number of negative samples in the training set to be equal; the positive sample is a sample with qualified welding spot quality, and the negative sample is a sample with unqualified welding spot quality.
Specifically, the welding spot process parameter data set is subjected to data normalization processing, and the welding spot process parameter data are compressed to a preset range; the welding spot data has the problem of unbalance of the proportion of positive samples and negative samples, and then the training set is subjected to up-sampling processing, so that the number of the positive samples and the number of the negative samples in the training set are equal; the positive sample is a sample with qualified welding spot quality, and the negative sample is a sample with unqualified welding spot quality.
It should be noted that the preset range is preferably [0,1], and the welding spot process parameter data is compressed to the range of [0,1], which is beneficial to eliminating the influence caused by the dimensions of the characteristics of resistance, current and heat.
In another preferred embodiment, the upsampling the training set to make the number of positive samples and the number of negative samples in the training set equal specifically includes:
determining a sampling multiplying power N according to the proportion of the positive sample to the negative sample in the training set;
for each few class sample x, selecting a plurality of samples from K neighbors of the sample x;
randomly selecting a sample x from a plurality of samples selected in K adjacent tonAccording to the formula xnew＝x+rand(0,1)·(xn-x) constructing a new sample;
and repeating the sampling multiplying power for N times in the construction steps so as to enable the number of the positive samples and the number of the negative samples in the training set to be equal.
Specifically, a sampling multiplying factor N is determined according to a ratio of the positive sample to the negative sample in the training set; for each few class sample x, selecting a plurality of samples from K neighbors of the sample x; randomly selecting a sample x from a plurality of samples selected in K adjacent tonAccording to the formula xnew＝x+rand(0,1)·(xn-x) a new sample; and repeating the sampling multiplying power for N times in the construction steps so as to enable the number of the positive samples and the number of the negative samples in the training set to be equal. In another preferred embodiment, the weld process parameter data set further includes a verification set, and then training each of the single weld quality inspection models further includes:
inputting the training set into training models with different hyper-parameters for training for each single welding spot quality detection model;
inputting the verification set into a training model of each hyper-parameter for verification, and selecting the hyper-parameters which are well represented in the verification set;
and inputting the test set into a single welding spot quality detection model with good performance and super-parameters so as to evaluate the detection accuracy of the single welding spot quality detection model.
It should be noted that, in the process of model training, in order to optimize the model more so as to obtain a global optimal solution, overfitting is avoided, and the model complexity is constrained by a regular term. The relation between the optimal solution and the regular term can be found in a balanced way through the super-parameter adjustment, and the performance and the effect of the model are improved.
Specifically, for each single welding spot quality detection model, inputting the training set into training models with different hyperparameters for training; inputting the verification set into a training model of each hyper-parameter for verification, and selecting the hyper-parameters which are well represented in the verification set; and inputting the test set into a single welding spot quality detection model with good performance and super-parameters so as to evaluate the detection accuracy of the single welding spot quality detection model.
In another preferred embodiment, the S4, inputting the test set into each of the base classifiers, and fusing output results of all the base classifiers according to a preset rule to obtain a solder joint quality detection result, specifically including:
inputting the test set into the base classifier to obtain an output result of the base classifier; wherein the output result comprises a pass and a fail;
and counting the output results, and taking the output results which account for most as the detection results of the quality of the welding spots according to a rule that a minority obeys most.
Specifically, the test set is input into the base classifier to obtain an output result of the base classifier; wherein the output result comprises a pass and a fail; and counting the output results, and taking the output results which account for most as the detection results of the quality of the welding spots according to a rule that a minority obeys most.
For example, historical welding spot process parameter data in the manufacturing process of a vehicle body is used as raw data, the total number of samples is 408, wherein 323 qualified welding spot samples and 85 unqualified welding spot samples are adopted, the data volume is small, the ratio of positive samples to negative samples is close to 4:1, and the samples have no default. A single weld sample contains 3 types of features: 50 current characteristics, 50 resistance characteristics and 50 heat characteristics. Wherein 50 features comprised by each physical quantity have the same dimensions and are of the same order of magnitude. In this embodiment, the characteristics select three types of characteristics, namely current, resistance, and heat, and the total number of the characteristics is 150, and data normalization processing is performed on the sample. Dividing all samples into a training set and a testing set according to the proportion of 7:3, and increasing the number of unqualified samples in the training set by adopting an up-sampling mode so that the proportion of positive and negative samples in the training set is 1: 1.
Then, a single machine learning algorithm is used to construct a plurality of single weld spot quality detection models with multiple feature fusion for the current feature, the resistance feature and the heat feature, for example, three single models are constructed in this embodiment, which are respectively a random forest model, a naive bayes model and a logistic regression model, as shown in table 1. And inputting the training set into each single welding spot quality detection model, training each single welding spot quality detection model, and performing hyper-parameter adjustment through the verification set and the test set. If the random forest model is constructed, the super parameters have large influence on the model, different values are selected for the parameters to train the model, the verification set is input into the training model of each super parameter for verification, the value of the super parameter well represented in the verification set is selected as the specific value of the parameter, the test set is input into the single welding spot quality detection model of the well represented super parameter, and the detection accuracy of the single welding spot quality detection model is evaluated.
| Model (model) | Hyper-parameter |
| Random forest model | min_samples_split＝4、min_samples_leaf＝8、max_depth＝6 |
| Naive Bayes model | Is free of |
| Logistic regression model | penalty＝'l2'、tol＝0.001、C＝1.1 |
TABLE 1
Evaluating each trained single welding spot quality detection model according to evaluation indexes such as accuracy, missing detection rate and false detection rate, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; and inputting the test set into each base classifier, wherein each base classifier can carry out quality detection on welding spot process parameter data, count the output results, and take the output results which account for most as the detection results of the welding spot quality according to the rules that a minority obeys most. For example, the random forest model is predicted to be qualified, the naive Bayes model is predicted to be qualified, and the logistic regression model is predicted to be unqualified. And if the number of the base classifiers which are predicted to be qualified is 2 and the base classifiers account for the majority, the final welding spot quality detection result is qualified.
In the model fusion algorithm, each base classifier can predict the class of the data, then statistics is carried out on the predictions made by all the base classifiers, and the class with the most prediction is the classification result finally obtained. And evaluating the feasibility of the model by taking the accuracy, the omission factor and the false detection rate as evaluation indexes. Table 2 shows the comparison between the evaluation index of the multi-model fusion algorithm and the single model in this embodiment.
| Model (model) | Rate of accuracy | Rate of missed examination | False detection rate |
| Random forest model | 92％ | 5.3％ | 20.2％ |
| Naive Bayes model | 87％ | 9.2％ | 38.1％ |
| Logistic regression model | 90％ | 6.7％ | 26.2％ |
| Multi-model fusion algorithm | 94％ | 4％ | 14.7％ |
TABLE 2
According to the welding spot quality detection method based on multi-model fusion, provided by the embodiment of the invention, the welding spot quality detection result is superior to that of a single model, and the method has good generalization capability, so that the model is not easy to over-fit, the robustness of the model is improved, and the accuracy of welding spot quality detection is further improved.
Correspondingly, the invention also provides a welding spot quality detection device based on multi-model fusion, which can realize all the processes of the welding spot quality detection method based on multi-model fusion in the embodiment.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a welding spot quality detection apparatus based on multi-model fusion according to a preferred embodiment of the present invention. The welding spot quality detection device based on multi-model fusion comprises:
the acquisition module 201 is configured to acquire a welding spot process parameter data set in a resistance spot welding process, and divide the welding spot process parameter data set into a training set and a test set according to a preset ratio;
the single model training module 202 is configured to construct a plurality of single welding spot quality detection models by using a machine learning algorithm, input the training set into each single welding spot quality detection model, and train each single welding spot quality detection model;
the evaluation module 203 is used for evaluating each trained single welding spot quality detection model according to a preset evaluation index, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; the evaluation indexes comprise accuracy, missing detection rate and false detection rate;
and the welding spot quality detection module 204 is configured to input the test set into each of the base classifiers, and fuse output results of all the base classifiers according to a preset rule to obtain a welding spot quality detection result.
Preferably, the obtaining module 201 is specifically configured to:
collecting welding spot process parameter data in the resistance spot welding process; the welding spot process parameter data comprise a resistance value, a current value and a heat value;
arranging the resistance value, the current value and the heat value into time sequence data according to a time sequence, and obtaining an expression of each welding spot data x as follows: x ═ xt1,2, ·, n; wherein n represents the number of time series, xtThe t-th feature representing x;
summarizing the obtained data of each welding spot, and establishing a welding spot process parameter data set X, X { (X)1,y1),(x2,y2),...,(xN,yN) }; wherein N represents the number of weld data samples, xiRepresenting single solder joint data, yiRepresenting the quality of welding spots, wherein the quality of the welding spots is divided into qualified welding spots and unqualified welding spots;
and dividing the welding spot process parameter data set X into a training set and a testing set according to a preset proportion.
Preferably, the apparatus further comprises:
the normalization module is used for carrying out data normalization processing on the welding spot process parameter data set and compressing the welding spot process parameter data to a preset range;
the up-sampling module is used for performing up-sampling processing on the training set so as to enable the number of positive samples and the number of negative samples in the training set to be equal; the positive sample is a sample with qualified welding spot quality, and the negative sample is a sample with unqualified welding spot quality.
Preferably, the performing the upsampling process on the training set to make the number of the positive samples and the number of the negative samples in the training set equal specifically includes:
determining a sampling multiplying power N according to the proportion of the positive sample to the negative sample in the training set;
for each few class sample x, selecting a plurality of samples from K neighbors of the sample x;
randomly selecting a sample x from a plurality of samples selected in K adjacent tonAccording to the formula xnew＝x+rand(0,1)·(xn-x) constructing a new sample;
and repeating the sampling multiplying power for N times in the construction steps so as to enable the number of the positive samples and the number of the negative samples in the training set to be equal.
Preferably, the welding spot process parameter data set further includes a verification set, and then training each single welding spot quality detection model further includes:
inputting the training set into training models with different hyper-parameters for training for each single welding spot quality detection model;
inputting the verification set into a training model of each hyper-parameter for verification, and selecting the hyper-parameters which are well represented in the verification set;
and inputting the test set into a single welding spot quality detection model with good performance and super-parameters so as to evaluate the detection accuracy of the single welding spot quality detection model.
Preferably, the solder joint quality detection module 204 is specifically configured to:
inputting the test set into the base classifier to obtain an output result of the base classifier; wherein the output result comprises a pass and a fail;
and counting the output results, and taking the output results which account for most as the detection results of the quality of the welding spots according to a rule that a minority obeys most.
In a specific implementation, the working principle, the control flow and the technical effect of the welding spot quality detection apparatus based on multi-model fusion provided in the embodiment of the present invention are the same as those of the welding spot quality detection method based on multi-model fusion in the above embodiment, and are not described herein again.
Referring to fig. 3, fig. 3 is a schematic structural diagram of another preferred embodiment of a welding spot quality detection apparatus based on multi-model fusion according to the present invention. The multi-model fusion-based solder joint quality detection apparatus comprises a processor 301, a memory 302 and a computer program stored in the memory 302 and configured to be executed by the processor 301, wherein the processor 301 implements the multi-model fusion-based solder joint quality detection method according to any of the above embodiments when executing the computer program.
Preferably, the computer program may be divided into one or more modules/units (e.g., computer program 1, computer program 2, … …) that are stored in the memory 302 and executed by the processor 301 to implement the present invention. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used for describing the execution process of the computer program in the terminal device.
The Processor 301 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, a discrete hardware component, etc., the general purpose Processor may be a microprocessor, or the Processor 301 may be any conventional Processor, the Processor 301 is a control center of the terminal device, and various interfaces and lines are used to connect various parts of the terminal device.
The memory 302 mainly includes a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function, and the like, and the data storage area may store related data and the like. In addition, the memory 302 may be a high speed random access memory, a non-volatile memory such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash Card (Flash Card), and the like, or the memory 302 may be other volatile solid state memory devices.
It should be noted that the above-mentioned multi-model fusion-based solder joint quality detection apparatus may include, but is not limited to, a processor and a memory, and those skilled in the art will understand that the schematic diagram of fig. 3 is only an example of the above-mentioned multi-model fusion-based solder joint quality detection apparatus, and does not constitute a limitation of the above-mentioned multi-model fusion-based solder joint quality detection apparatus, and may include more or less components than those shown in the drawings, or may combine some components, or may be different components.
The embodiment of the present invention further provides a computer-readable storage medium, where the computer-readable storage medium includes a stored computer program, and when the computer program runs, a device where the computer-readable storage medium is located is controlled to execute the method for detecting quality of a weld spot based on multi-model fusion according to any of the above embodiments.
The embodiment of the invention provides a welding spot quality detection method, a welding spot quality detection device and a storage medium based on multi-model fusion, wherein a welding spot process parameter data set in a resistance spot welding process is obtained and is divided into a training set and a testing set according to a preset proportion; establishing a plurality of single welding spot quality detection models by adopting a machine learning algorithm, inputting the training set into each single welding spot quality detection model, and training each single welding spot quality detection model; evaluating each trained single welding spot quality detection model according to a preset evaluation index, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; the evaluation indexes comprise accuracy, missing detection rate and false detection rate; and inputting the test set into each base classifier, and fusing output results of all the base classifiers according to a preset rule to obtain a welding spot quality detection result. The embodiment of the invention adopts a multi-feature fusion method, and establishes a model through a plurality of parameters related to the quality of the welding spot, thereby being more beneficial to judging the quality of the welding spot; the detection results of the multiple models are fused, the defect of poor generalization capability of a single model is overcome, the models are not easy to over-fit, the robustness of the models is improved, the noise resistance is good, and the accuracy and the detection efficiency of the quality detection of the welding spots are improved. In addition, the adopted base classifier is simple to realize, and the defects that the neural network algorithm searches along the negative gradient direction in the iterative process, can not converge to the global optimal solution at a higher speed and has a low training speed are avoided.
It should be noted that the above-described system embodiments are merely illustrative, where the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. In addition, in the drawings of the embodiment of the system provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, and may be specifically implemented as one or more communication buses or signal lines. One of ordinary skill in the art can understand and implement it without inventive effort.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention.
Claims (10)
1. A welding spot quality detection method based on multi-model fusion is characterized by comprising the following steps:
acquiring a welding spot process parameter data set in the resistance spot welding process, and dividing the welding spot process parameter data set into a training set and a testing set according to a preset proportion;
establishing a plurality of single welding spot quality detection models by adopting a machine learning algorithm, inputting the training set into each single welding spot quality detection model, and training each single welding spot quality detection model;
evaluating each trained single welding spot quality detection model according to a preset evaluation index, and selecting the single welding spot quality detection model with good evaluation performance as a base classifier; the evaluation indexes comprise accuracy, missing detection rate and false detection rate;
and inputting the test set into each base classifier, and fusing output results of all the base classifiers according to a preset rule to obtain a welding spot quality detection result.
2. The method for detecting the quality of the welding spot based on the multi-model fusion as claimed in claim 1, wherein the obtaining of the welding spot process parameter data set in the resistance spot welding process and the dividing of the welding spot process parameter data set into a training set and a testing set according to a preset ratio specifically comprises:
collecting welding spot process parameter data in the resistance spot welding process; the welding spot process parameter data comprise a resistance value, a current value and a heat value;
arranging the resistance value, the current value and the heat value into time sequence data according to a time sequence, and obtaining an expression of each welding spot data x as follows: x ═ xt1,2, ·, n; wherein n represents the number of time series, xtThe t-th feature representing x;
summarizing the obtained data of each welding spot, and establishing a welding spot process parameter data set X, X { (X)1,y1),(x2,y2),...,(xN,yN) }; wherein N represents the number of weld data samples, xiRepresents a singleSolder joint data, yiRepresenting the quality of welding spots, wherein the quality of the welding spots is divided into qualified welding spots and unqualified welding spots;
and dividing the welding spot process parameter data set X into a training set and a testing set according to a preset proportion.
3. The method for detecting the quality of the welding spot based on the multi-model fusion as claimed in claim 1, wherein the method further comprises:
carrying out data normalization processing on the welding spot process parameter data set, and compressing the welding spot process parameter data to a preset range;
performing upsampling processing on the training set to enable the number of positive samples and the number of negative samples in the training set to be equal; the positive sample is a sample with qualified welding spot quality, and the negative sample is a sample with unqualified welding spot quality.
4. The method according to claim 3, wherein the upsampling is performed on the training set to equalize the number of positive samples and the number of negative samples in the training set, and specifically comprises:
determining a sampling multiplying power N according to the proportion of the positive sample to the negative sample in the training set;
for each few class sample x, selecting a plurality of samples from K neighbors of the sample x;
randomly selecting a sample x from a plurality of samples selected in K adjacent tonAccording to the formula xnew＝x+rand(0,1)·(xn-x) constructing a new sample;
and repeating the sampling multiplying power for N times in the construction steps so as to enable the number of the positive samples and the number of the negative samples in the training set to be equal.
5. The method for solder joint quality inspection based on multi-model fusion of claim 1, wherein the solder joint process parameter data set further comprises a validation set, and then each of the single solder joint quality inspection models is trained, further comprising:
inputting the training set into training models with different hyper-parameters for training for each single welding spot quality detection model;
inputting the verification set into a training model of each hyper-parameter for verification, and selecting the hyper-parameters which are well represented in the verification set;
and inputting the test set into a single welding spot quality detection model with good performance and super-parameters so as to evaluate the detection accuracy of the single welding spot quality detection model.
6. The method according to claim 1, wherein the inputting the test set into each of the base classifiers and fusing the output results of all the base classifiers according to a preset rule to obtain a solder joint quality detection result specifically comprises:
inputting the test set into the base classifier to obtain an output result of the base classifier; wherein the output result comprises a pass and a fail;
and counting the output results, and taking the output results which account for most as the detection results of the quality of the welding spots according to a rule that a minority obeys most.
7. The utility model provides a solder joint quality detection device based on multi-model fuses which characterized in that includes:
the acquisition module is used for acquiring a welding spot process parameter data set in the resistance spot welding process and dividing the welding spot process parameter data set into a training set, a testing set and a verification set;
the single model training module is used for constructing a single welding spot quality detection model by adopting a single machine learning algorithm, inputting the training set into the single welding spot quality detection model and training the single welding spot quality detection model;
the evaluation module is used for evaluating the single welding spot quality detection model through preset evaluation indexes;
and the welding spot quality detection module is used for constructing a multi-model fusion algorithm according to the evaluation result so as to detect the welding spot quality.
8. The apparatus for detecting quality of a solder joint based on multi-model fusion according to claim 1, wherein the apparatus further comprises:
the normalization module is used for carrying out data normalization processing on the welding spot process parameter data set and compressing the welding spot process parameter data to a preset range;
the up-sampling module is used for performing up-sampling processing on the training set so as to enable the number of positive samples and the number of negative samples in the training set to be equal; the positive sample is a sample with qualified welding spot quality, and the negative sample is a sample with unqualified welding spot quality.
9. A multi-model fusion-based solder joint quality detection apparatus, comprising a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the multi-model fusion-based solder joint quality detection method according to any one of claims 1 to 6 when executing the computer program.
10. A computer-readable storage medium, comprising a stored computer program, wherein when the computer program runs, the computer-readable storage medium controls a device to execute the method for detecting the quality of the welding spot based on the multi-model fusion according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110590395.5A CN113255778A (en) | 2021-05-28 | 2021-05-28 | Welding spot quality detection method and device based on multi-model fusion and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110590395.5A CN113255778A (en) | 2021-05-28 | 2021-05-28 | Welding spot quality detection method and device based on multi-model fusion and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113255778A true CN113255778A (en) | 2021-08-13 |
Family
ID=77185215
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110590395.5A Pending CN113255778A (en) | 2021-05-28 | 2021-05-28 | Welding spot quality detection method and device based on multi-model fusion and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255778A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113695713A (en) * | 2021-09-17 | 2021-11-26 | 蕴硕物联技术(上海)有限公司 | Online monitoring method and device for welding quality of inner container of water heater |
| CN113779886A (en) * | 2021-09-18 | 2021-12-10 | 天津商科数控技术股份有限公司 | Method, device and system for detecting welding spot quality abnormity based on deep learning |
| CN114242159A (en) * | 2022-02-24 | 2022-03-25 | 北京晶泰科技有限公司 | Method for constructing antigen peptide presentation prediction model, and antigen peptide prediction method and device |
| CN115266141A (en) * | 2022-07-29 | 2022-11-01 | 广汽本田汽车有限公司 | Spot welding quality detection method and device based on GRU-C network and storage medium |
| CN115310486A (en) * | 2022-08-09 | 2022-11-08 | 重庆大学 | Intelligent detection method for welding quality |
| CN116452588A (en) * | 2023-06-15 | 2023-07-18 | 苏州松德激光科技有限公司 | Welding quality assessment method and system |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1730220A (en) * | 2005-08-24 | 2006-02-08 | 沈阳工业大学 | Quality control device and method for resistance spot welding |
| US20130132311A1 (en) * | 2011-11-18 | 2013-05-23 | Honeywell International Inc. | Score fusion and training data recycling for video classification |
| CN106845717A (en) * | 2017-01-24 | 2017-06-13 | 哈尔滨工业大学 | A kind of energy efficiency evaluation method based on multi-model convergence strategy |
| CN107301604A (en) * | 2017-08-29 | 2017-10-27 | 重庆华龙强渝信用管理有限公司 | Multi-model fusion estimation system |
| US20180300631A1 (en) * | 2015-06-30 | 2018-10-18 | Arizona Technology Enterprises | Method and apparatus for large scale machine learning |
| CN108918309A (en) * | 2018-08-22 | 2018-11-30 | 广州亨龙智能装备股份有限公司 | The solder joint detection method and solder joint detection device of galvanized steel plain sheet |
| US20180346151A1 (en) * | 2017-05-30 | 2018-12-06 | The Boeing Company | Advanced analytic methods and systems utilizing trust-weighted machine learning models |
| CN111209977A (en) * | 2020-01-16 | 2020-05-29 | 北京百度网讯科技有限公司 | Method, apparatus, device and medium for training and using classification model |
| CN111681235A (en) * | 2020-06-11 | 2020-09-18 | 广东工业大学 | IC welding spot defect detection method based on learning mechanism |
| CN111695626A (en) * | 2020-06-10 | 2020-09-22 | 湖南湖大金科科技发展有限公司 | High-dimensional unbalanced data classification method based on mixed sampling and feature selection |
| CN111932489A (en) * | 2020-06-03 | 2020-11-13 | 西安电子科技大学 | Weld defect detection method, system, storage medium, computer device and terminal |
| CN112345982A (en) * | 2020-09-29 | 2021-02-09 | 歌尔科技有限公司 | Method and device for detecting welding condition of circuit element |
| CN112487708A (en) * | 2020-11-24 | 2021-03-12 | 山西三友和智慧信息技术股份有限公司 | Resistance spot welding quality prediction method based on ensemble learning |
-
2021
- 2021-05-28 CN CN202110590395.5A patent/CN113255778A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1730220A (en) * | 2005-08-24 | 2006-02-08 | 沈阳工业大学 | Quality control device and method for resistance spot welding |
| US20130132311A1 (en) * | 2011-11-18 | 2013-05-23 | Honeywell International Inc. | Score fusion and training data recycling for video classification |
| US20180300631A1 (en) * | 2015-06-30 | 2018-10-18 | Arizona Technology Enterprises | Method and apparatus for large scale machine learning |
| CN106845717A (en) * | 2017-01-24 | 2017-06-13 | 哈尔滨工业大学 | A kind of energy efficiency evaluation method based on multi-model convergence strategy |
| US20180346151A1 (en) * | 2017-05-30 | 2018-12-06 | The Boeing Company | Advanced analytic methods and systems utilizing trust-weighted machine learning models |
| CN107301604A (en) * | 2017-08-29 | 2017-10-27 | 重庆华龙强渝信用管理有限公司 | Multi-model fusion estimation system |
| CN108918309A (en) * | 2018-08-22 | 2018-11-30 | 广州亨龙智能装备股份有限公司 | The solder joint detection method and solder joint detection device of galvanized steel plain sheet |
| CN111209977A (en) * | 2020-01-16 | 2020-05-29 | 北京百度网讯科技有限公司 | Method, apparatus, device and medium for training and using classification model |
| CN111932489A (en) * | 2020-06-03 | 2020-11-13 | 西安电子科技大学 | Weld defect detection method, system, storage medium, computer device and terminal |
| CN111695626A (en) * | 2020-06-10 | 2020-09-22 | 湖南湖大金科科技发展有限公司 | High-dimensional unbalanced data classification method based on mixed sampling and feature selection |
| CN111681235A (en) * | 2020-06-11 | 2020-09-18 | 广东工业大学 | IC welding spot defect detection method based on learning mechanism |
| CN112345982A (en) * | 2020-09-29 | 2021-02-09 | 歌尔科技有限公司 | Method and device for detecting welding condition of circuit element |
| CN112487708A (en) * | 2020-11-24 | 2021-03-12 | 山西三友和智慧信息技术股份有限公司 | Resistance spot welding quality prediction method based on ensemble learning |
Non-Patent Citations (1)
| Title |
|---|
| 熊霖;唐万梅;: "基于k-means++的多分类器选择分类研究", 重庆师范大学学报(自然科学版), no. 06, pages 88 - 96 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113695713A (en) * | 2021-09-17 | 2021-11-26 | 蕴硕物联技术(上海)有限公司 | Online monitoring method and device for welding quality of inner container of water heater |
| CN113779886A (en) * | 2021-09-18 | 2021-12-10 | 天津商科数控技术股份有限公司 | Method, device and system for detecting welding spot quality abnormity based on deep learning |
| CN114242159A (en) * | 2022-02-24 | 2022-03-25 | 北京晶泰科技有限公司 | Method for constructing antigen peptide presentation prediction model, and antigen peptide prediction method and device |
| CN115266141A (en) * | 2022-07-29 | 2022-11-01 | 广汽本田汽车有限公司 | Spot welding quality detection method and device based on GRU-C network and storage medium |
| CN115266141B (en) * | 2022-07-29 | 2024-05-14 | 广汽本田汽车有限公司 | Method and device for detecting spot welding quality based on GRU-C network and storage medium |
| CN115310486A (en) * | 2022-08-09 | 2022-11-08 | 重庆大学 | Intelligent detection method for welding quality |
| CN115310486B (en) * | 2022-08-09 | 2023-09-26 | 重庆大学 | Intelligent welding quality detection method |
| CN116452588A (en) * | 2023-06-15 | 2023-07-18 | 苏州松德激光科技有限公司 | Welding quality assessment method and system |
| CN116452588B (en) * | 2023-06-15 | 2023-10-31 | 苏州松德激光科技有限公司 | Welding quality assessment method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113255778A (en) | Welding spot quality detection method and device based on multi-model fusion and storage medium | |
| CN112034789B (en) | Health assessment method, system and assessment terminal for key parts and complete machine of numerical control machine tool | |
| CN107122594B (en) | New energy vehicle battery health prediction method and system | |
| WO2017143919A1 (en) | Method and apparatus for establishing data identification model | |
| CN113792825B (en) | Fault classification model training method and device for electricity information acquisition equipment | |
| CN102135979B (en) | Data cleaning method and device | |
| CN107292330A (en) | A kind of iterative label Noise Identification algorithm based on supervised learning and semi-supervised learning double-point information | |
| CN106228389A (en) | Network potential usage mining method and system based on random forests algorithm | |
| Huang et al. | An effective fault diagnosis method for centrifugal chillers using associative classification | |
| Nkonyana et al. | Performance evaluation of data mining techniques in steel manufacturing industry | |
| CN108780313B (en) | Method, system, and computer-readable medium for performing targeted parameter analysis for an assembly line | |
| CN115268406B (en) | Monolithic chip testing method, monolithic chip testing system, computer equipment and storage medium | |
| CN109255029A (en) | A method of automatic Bug report distribution is enhanced using weighted optimization training set | |
| US20210406727A1 (en) | Managing defects in a model training pipeline using synthetic data sets associated with defect types | |
| CN113646714A (en) | Processing parameter setting method and device for production equipment and computer readable medium | |
| Haghighi et al. | Applying mining schemes to software fault prediction: A proposed approach aimed at test cost reduction | |
| CN114881343B (en) | Short-term load prediction method and device for power system based on feature selection | |
| CN105447520A (en) | Sample classification method based on weighted PTSVM (projection twin support vector machine) | |
| CN115358481A (en) | Early warning and identification method, system and device for enterprise ex-situ migration | |
| CN110208702B (en) | Neural network-based lithium battery ultrasonic information characteristic extraction method | |
| CN114330486A (en) | Power system bad data identification method based on improved Wasserstein GAN | |
| CN114325433A (en) | Lithium ion battery fault detection method and system based on electrochemical impedance spectrum test | |
| CN117630800A (en) | Fault diagnosis method and system for automatic calibrating device of electric energy meter | |
| Vu | Firm’s absorptive capacity: The case of Vietnamese manufacturing firms | |
| CN102255965A (en) | Test platform for multi-sensor fused Internet of things (IOT) based on rough set BP (Back Propagation) neural network |
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 |