Quality assessment of resistance spot welding process based on dynamic resistance signal and random forest based

Xing, Bobin; Xiao, Yi; Qin, Qing Hua; Cui, Hongzhi

doi:10.1007/s00170-017-0889-6

Cited by 70 publications

(32 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The nodal displacements on the centreline of the electrodes can predict the contribution of electrode expansion to total displacement via numerical simulation. Tables Table 1 Chemical composition of CA2S-E (wt %) Table 2 Mechanical properties of CA2S-E Table 3 Welding parameters used in this study [26,27] Highlights:…”

Section: Resultsmentioning

confidence: 99%

Characteristics of shunting effect in resistance spot welding in mild steel based on electrode displacement

2018

Self Cite

View full text Add to dashboard Cite

Shunting effect of resistance spot welding is evaluated based on the electrode displacement signals. The shunted welds in mild steel with different weld spacing were produced. The results showed that the weld spacing and nugget diameter were polynomial-correlated, and the minimum welding spacing of 20 mm can be derived from the results. Both the peak value and gradient of electrode displacement in the weld stage indicated strong correlations with the nugget diameters of shunted welds. Additional shunt path was found to further aggregate the shunting, suggesting the decline in the values of profile features. Furthermore, it is found that the shunting effect led to the decline of the dynamic resistance curves, which is contradictive to the trends between acceptable-sized and undersized welds claimed based on the single weld study. The paper shows that electrode displacement curves of shunting can be incorporated into existing quality monitoring system.

show abstract

Section: Resultsmentioning

confidence: 99%

Characteristics of shunting effect in resistance spot welding in mild steel based on electrode displacement

2018

Self Cite

View full text Add to dashboard Cite

show abstract

“…Many researchers employed RF to predict different welding features. 61 , 62 There has been limited implementation of RF with high energy joining techniques. 54 Reduced Error Pruning Tree (“REPT”) is another tree-based regression model, where information gain/variance reduction-based node statistics are used to develop multiple decision trees.…”

Section: Tools and Techniquesmentioning

confidence: 99%

Prediction of residual stress in electron beam welding of stainless steel from process parameters and natural frequency of vibrations using machine-learning algorithms

Das

Pratihar

et al. 2020

Proceedings of the Institution of Mechanical Engineers, Part C:

View full text Add to dashboard Cite

In the present study, machine learning algorithms have been used to predict residual stress during electron beam welding of stainless steel using the information of input process parameters and natural frequency of vibrations. Accelerating voltage, beam current and welding speed have been considered as input process parameters. Both residual stress and natural frequencies of vibration of the weld obtained using each set of the input parameters are measured experimentally. A number of machine learning algorithms, namely M5 algorithm-based Model Trees Regression, Random forest, Support Vector Regression, Reduced Error Pruning Tree, Multi-layer perceptron, Instance-based k-nearest neighbor algorithm, and Locally weighted learning have been used for the said purpose. Support vector regression and Locally weighted learning are found to perform consistently good and bad, respectively. The predicted welding residual stresses have been validated experimentally through X-ray diffraction (XRD) and good agreements are obtained. In addition, statistical tests are conducted, and the estimated reliability values of the employed models are analyzed through Monte-Carlo simulations.

show abstract

“…On the other side, the incorporation of new technologies such as Industrial Internet of Things (IIoT) and advanced analytics into manufacturing systems aims to produce individualized products at high quality and low costs. In the manufacturing domain, such data-driven approaches have been extensively studied in the past and are based on autoregressive (AR) models, cluster analysis, fuzzy set theory or on supervised learning algorithms such as multivariate regression, multi-layer perceptron and decision trees, as well as k-nearest neighbors [14,15]. Therefore, recent development led to advanced process monitoring systems which integrate machine learning techniques for process control and prediction of critical defects [16,17].…”

Section: Introductionmentioning

confidence: 99%

Deep Learning and Conventional Machine Learning for Image-Based in-Situ Fault Detection During Laser Welding: A Comparative Study

Knaak¹,

Kröger²,

Schulze³

et al. 2021

Preprint

View full text Add to dashboard Cite

An effective process monitoring strategy is a requirement for meeting the challenges posed by increasingly complex products and manufacturing processes. To address these needs, this study investigates a comprehensive scheme based on classical machine learning methods, deep learning algorithms, and feature extraction and selection techniques. In a first step, a novel deep learning architecture based on convolutional neural networks (CNN) and gated recurrent units (GRU) is introduced to predict the local weld quality based on mid-wave infrared (MWIR) and near-infrared (NIR) image data. The developed technology is used to discover critical welding defects including lack of fusion (false friends), sagging and lack of penetration, and geometric deviations of the weld seam. Additional work is conducted to investigate the significance of various geometrical, statistical, and spatio-temporal features extracted from the keyhole and weld pool regions. Furthermore, the performance of the proposed deep learning architecture is compared to that of classical supervised machine learning algorithms, such as multi-layer perceptron (MLP), logistic regression (LogReg), support vector machines (SVM), decision trees (DT), random forest (RF) and k-Nearest Neighbors (kNN). Optimal hyperparameters for each algorithm are determined by an extensive grid search. Ultimately, the three best classification models are combined into an ensemble classifier that yields the highest detection rates and achieves the most robust estimation of welding defects among all classifiers studied, which is validated on previously unknown welding trials.

show abstract

Quality assessment of resistance spot welding process based on dynamic resistance signal and random forest based

Cited by 70 publications

References 12 publications

Characteristics of shunting effect in resistance spot welding in mild steel based on electrode displacement

Characteristics of shunting effect in resistance spot welding in mild steel based on electrode displacement

Prediction of residual stress in electron beam welding of stainless steel from process parameters and natural frequency of vibrations using machine-learning algorithms

Deep Learning and Conventional Machine Learning for Image-Based in-Situ Fault Detection During Laser Welding: A Comparative Study

Contact Info

Product

Resources

About