Application of Machine Learning to Bending Processes and Material Identification

Cruz, Daniel J.; Barbosa, Manuel R.; Santos, Abel D.; Miranda, Sara S.; Amaral, Rui

doi:10.3390/met11091418

Cited by 14 publications

(18 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This phenomena applies on at least a thickness of 3mm attributed to the reduction in elastic sheet recovery due to the increase in bulk deformation of the sheet metals, as sheet thickness decreases the rigidity of material which minimizes the effect of elastic recovery stated by Leu et al [29]. According to Cruz et al [56] the die opening is significant factor in bending operation. Typically, the selection of die opening is dependent on sheet thickness.…”

Section: Resultsmentioning

confidence: 96%

Prediction of Springback Behavior of Vee Bending Process of AA5052 Aluminum Alloy Sheets Using Machine Learning

2023

jjmie

View full text Add to dashboard Cite

This study explores the effect of Vee bending process parameter on the springback (SB) behavior of aluminum (AA5052) alloy at sheet thickness of (2 and 3mm) with die-opening (22, 35, and 50 mm) and punch-holding time (0, 5 and 10 second) which were experimentally examined. Furthermore, to see the relative effect of process parameter on SB behavior, a qualitative approach of analysis of variance (ANOVA) was used, whereas multi linear regression (MLR) and artificial neural network (ANN) were applied to optimize the SB behavior on specified process parameters. The experimental results revealed that as punch holding time and sheet thickness increase, SB behavior reduced, whereas in case of die opening, opposite phenomena observed. ANOVA results revealed that punch-holding time had the greatest effect on SB, followed by die opening and sheet thickness. Two-way parametric interactional effects between punch-holding time and dieopening had a significant effect on SB behavior. By contrast, the interactional effects of sheet thickness were insignificant. The comparative study of MLR and ANN shows that The ANN has better (99% SB predictability) as compared to MLR (73% SB Predictability). Furthermore, the predicted results of both models were compared with actual experimental results. It was observed that the predicted results were approximately near with actual measurements, whereas the performance of MLR and ANN model were measured from sum of absolute error and the sum of the absolute error of ANN was about 12% of that of MLR model. Therefore, ANN produced a superior SB prediction performance compared with MLR. This work demonstrates the formability of AA5052 aluminum alloy in cold work where Vee bending was performed with a punch radius of 0.8 mm.The bend specimens showed no cracks, checking, and surface roughness.

show abstract

Section: Resultsmentioning

confidence: 96%

Prediction of Springback Behavior of Vee Bending Process of AA5052 Aluminum Alloy Sheets Using Machine Learning

2023

jjmie

View full text Add to dashboard Cite

show abstract

“…However, the absence of labelled data poses a significant challenge for developing supervised learning-based solutions. Specifically, supervised learning models, which are commonly used for classification tasks like defect detection, require substantial number of labelled samples for training purposes [10][11][12][13]. Without these labelled samples indicating what constitutes a 'defect' or 'non-defect', a supervised learning model lacks the guidance needed to generalize and make accurate predictions on new, unseen data.…”

Section: Related Work and Proposed Frameworkmentioning

confidence: 99%

Anomaly Detection in the Automotive Stamping Process: An Unsupervised Machine Learning Approach

Zhang,

Creighton,

Lim

et al. 2024

IOP Conf. Ser.: Mater. Sci. Eng.

View full text Add to dashboard Cite

In metal forming, such as stamping of automotive parts, unsupervised machine learning models offer a transformative approach to real-time quality control, especially when labelled data are scarce. Leveraging clustering algorithms and autoencoders, we develop a machine learning system capable of autonomously monitoring sensor data and identifying deviations suggestive of potential defects. The system offers multiple benefits including rapid intervention, reduced part defects and lower stoppages required to rectify defects. The use of unsupervised machine learning models also adds a layer of adaptability, allowing the system to continually refine its understanding of what constitutes a ‘normal’ operation. Empirical evaluation demonstrates the potential of the developed system in detecting anomalies in production data collected from dynamic automotive manufacturing environments.

show abstract

“…Several other areas of application of machine learning can be found in the literature. This includes; additive manufacturing process [19], material design [20,21], process optimization [22], prediction of material quality [23], metal forming processes [24], nanomaterials [25,26], and heavy production process.…”

Section: Other Areasmentioning

confidence: 99%

Application of Machine Learning Tools to Material Science: A Mini-Review

2024

PSBJ

View full text Add to dashboard Cite

Machine learning has become a global trend in artificial intelligence and computing. Its application cuts across different industrial verticals. The application of machine learning algorithms in material science and related fields is becoming a topic of interest to many researchers. However, unlike other fields such as health, sports, communication, and agriculture, the use of machine learning in material science is still in its ideation stage. This mini-review is designed to explore some of those machine learning tools that have been applied to material science, the importance of machine learning in material science, and the challenges limiting the implementation of machine learning techniques in material science. One of the major findings from this review is that the limited availability of material science data is a major challenge to the implementation of machine learning algorithms. A specialized material science data repository was recommended.

show abstract

Application of Machine Learning to Bending Processes and Material Identification

Cited by 14 publications

References 48 publications

Prediction of Springback Behavior of Vee Bending Process of AA5052 Aluminum Alloy Sheets Using Machine Learning

Prediction of Springback Behavior of Vee Bending Process of AA5052 Aluminum Alloy Sheets Using Machine Learning

Anomaly Detection in the Automotive Stamping Process: An Unsupervised Machine Learning Approach

Application of Machine Learning Tools to Material Science: A Mini-Review

Contact Info

Product

Resources

About