Machine Learning-Based Reverse Modeling Approach for Rapid Tool Shape Optimization in Die-Sinking Micro Electro Discharge Machining

2020 ◽  
Vol 20 (3) ◽  
Author(s):  
Anthony Surleraux ◽  
Romain Lepert ◽  
Jean-Philippe Pernot ◽  
Pierre Kerfriden ◽  
Samuel Bigot
Keyword(s):  
Data Augmentation ◽  
Virtual Design ◽  
Computational Optimization ◽  
Modeling Approach ◽  
Electro Discharge Machining ◽  
Simulation Strategy ◽  
Tool Shape ◽  
Reduced Modeling ◽  
Artificial Neural Network Ann ◽  
Optimization Loop

Abstract This paper focuses on efficient computational optimization algorithms for the generation of micro electro discharge machining (µEDM) tool shapes. In a previous paper, the authors presented a reliable reverse modeling approach to perform such tasks based on a crater-by-crater simulation model and an outer optimization loop. Two-dimensional results were obtained but 3D tool shapes proved difficult to generate due to the high numerical cost of the simulation strategy. In this paper, a new reduced modeling optimization framework is proposed, whereby the computational optimizer is replaced by an inexpensive surrogate that is trained by examples. More precisely, an artificial neural network (ANN) is trained using a small number of full reverse simulations and subsequently used to directly generate optimal tool shapes, given the geometry of the desired workpiece cavity. In order to train the ANN efficiently, a method of data augmentation is developed, whereby multiple features from fully simulated EDM cavities are used as separate instances. The performances of two ANN are evaluated, one trained without modification of process parameters (gap size and crater shape) and the second trained with a range of process parameter instances. It is shown that in both cases, the ANN can produce unseen tool shape geometries with less than 6% deviation compared to the full computational optimization process and at virtually no cost. Our results demonstrate that optimized tool shapes can be generated almost instantaneously, opening the door to the rapid virtual design and manufacturability assessment of µEDM die-sinking operations.

scholarly journals Predicting Pressure Sensitivity to Luminophore Content and Paint Thickness of Pressure-Sensitive Paint Using Artificial Neural Network

Sensors ◽  
2021 ◽  
Vol 21 (15) ◽  
pp. 5188
Author(s):  
Mitsugu Hasegawa ◽  
Daiki Kurihara ◽  
Yasuhiro Egami ◽  
Hirotaka Sakaue ◽  
Aleksandar Jemcov
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Data Augmentation ◽  
Percentage Error ◽  
Pressure Sensitivity ◽  
Pressure Sensitive Paint ◽  
Pressure Sensitive ◽  
Artificial Neural ◽  
Paint Thickness ◽  
Artificial Neural Network Ann

An artificial neural network (ANN) was constructed and trained for predicting pressure sensitivity using an experimental dataset consisting of luminophore content and paint thickness as chemical and physical inputs. A data augmentation technique was used to increase the number of data points based on the limited experimental observations. The prediction accuracy of the trained ANN was evaluated by using a metric, mean absolute percentage error. The ANN predicted pressure sensitivity to luminophore content and to paint thickness, within confidence intervals based on experimental errors. The present approach of applying ANN and the data augmentation has the potential to predict pressure-sensitive paint (PSP) characterizations that improve the performance of PSP for global surface pressure measurements.

Development of Parameters Selection System for Ultrasonic Vibration Assisted Electro-Discharge Machining

2004 ◽  
Vol 471-472 ◽  
pp. 687-691
Author(s):  
Yu Jing Hu ◽  
Jian Hua Zhang ◽  
X.F. Wang ◽  
Sheng Feng Ren
Keyword(s):  
Ultrasonic Vibration ◽  
Influencing Factor ◽  
Back Propagation ◽  
Mathematics Model ◽  
Machining Parameters ◽  
Selection System ◽  
Application Range ◽  
Electro Discharge Machining ◽  
Parameters Selection ◽  
Artificial Neural Network Ann

Because the machining of ultrasonic vibration assisted electro-discharge machining (UEDM) is a very complex process and it is too difficult to describe precisely every influencing factor with an accurate mathematics model, the study of parameters selection system is necessary and important for the practical application of machining method, the improvement of machining efficiency and minimizing the tool wear ratio (TWR). In this paper, the model and the corresponding database are built for UEDM based on the back propagation (BP) algorithm artificial neural network (ANN) to optimize machining parameters. Through learning and training, this system realizes the intelligent selection of machining parameters. As shown by the experiment results, the predictions accord with the test results, which shows that the reasonable and reliable project of UEDM can be provided by the system. With the increase of the machining sample, the machining database can be enriched and the application range will be expanded, so this system has the excellent fault-tolerance and extensible quality.

USE OF CONVOLUTIONAL NEURAL NETWORKS FOR X-RAY IMAGE ORIENTATION DETERMINATION

2021 ◽  
Author(s):  
Sandi Baressi Šegota ◽  
◽  
Simon Lysdahlgaard ◽  
Søren Hess ◽  
Ronald Antulov
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
High Performance ◽  
Data Augmentation ◽  
Classification Model ◽  
Orientation Sensitivity ◽  
X Ray ◽  
Artificial Neural Network Ann ◽  
Image Orientation ◽  
Single Orientation

The fact that Artificial Intelligence (AI) based algorithms exhibit a high performance on image classification tasks has been shown many times. Still, certain issues exist with the application of machine learning (ML) artificial neural network (ANN) algorithms. The best known is the need for a large amount of statistically varied data, which can be addressed with expanded collection or data augmentation. Other issues are also present. Convolutional neural networks (CNNs) show extremely high performance on image-shaped data. Despite their performance, CNNs exhibit a large issue which is the sensitivity to image orientation. Previous research shows that varying the orientation of images may greatly lower the performance of the trained CNN. This is especially problematic in certain applications, such as X-ray radiography, an example of which is presented here. Previous research shows that the performance of CNNs is higher when used on images in a single orientation (left or right), as opposed to the combination of both. This means that the data needs to be differentiated before it enters the classification model. In this paper, the CNN-based model for differentiation between left and right-oriented images is presented. Multiple CNNs are trained and tested, with the highest performing being the VGG16 architecture which achieved an Accuracy of 0.99 (+/- 0.01), and an AUC of 0.98 (+/- 0.01). These results show that CNNs can be used to address the issue of orientation sensitivity by splitting the data in advance of being used in classification models.

scholarly journals Quantitative Measure of Concrete Fragment using ANN to Consider Uncertainties Under Impact Loading

2021 ◽  
Author(s):  
Kyeongjin Kim ◽  
WooSeok Kim ◽  
Junwon Seo ◽  
Yoseok Jeong ◽  
Jaeha Lee
Keyword(s):  
Impact Energy ◽  
Impact Loading ◽  
Data Augmentation ◽  
Fragility Curves ◽  
Linear Regression Analysis ◽  
Quantitative Measure ◽  
Multiple Linear Regression Analysis ◽  
Travel Distance ◽  
Distribution Method ◽  
Artificial Neural Network Ann

Abstract In the present study, numerical analysis was performed to predict the amount of concrete fragments generated and the distance travelled by the fragments under impact loading using Smooth Particle Hydrodynamics (SPH). SPH can be used for predicting the amount of fragmentation or the motion of fragmentations. The obtained results of the SPH analysis showed that the amount of fragments and the travel distance can change depending on different velocity-to-mass ratios under same local impact energy. Using the results of the SPH analysis, artificial neural network (ANN) was constructed to consider the uncertainty from the prediction of the fragmentations and travel distance. Furthermore, the results of ANN were compared with the results of Multiple Linear Regression Analysis (MRA). The ANN results showed better correlation coefficient (R2) than the MRA results. Therefore, ANN showed better improvement with consideration of the uncertainty from the prediction of fragmentations and travel distance than the MRA results. Using the constructed ANN, data augmentation was conducted from a limited number of actual data using a statistical distribution method. Finally, the fragility curves of the concrete median barrier were obtained to estimate the probability of occurrence of specific fragmentation amount and travel distance under same impact energy.

scholarly journals Artificial Neural Network for Classifying Injected Materials under Ultrasonography

2021 ◽  
Vol 3 (1) ◽  
pp. 0210101
Author(s):  
Galuh Retno Utari ◽  
Giner Maslebu ◽  
Suryasatriya Trihandaru
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Human Body ◽  
Data Augmentation ◽  
Test Dataset ◽  
Artificial Neural ◽  
Artificial Neural Network Ann ◽  
Human Eyes

We have constructed an artificial neural network (ANN) architecture to classify four different classes of ultrasonography recorded from a jelly box phantom  that was injected by iron, glass, or plastic marble, or without any injection. This jelly box was made as a phantom of a human body, and the injected materials were the cancers. The small size of the injected materials caused only  little disturbances those could not easily distinguished by human eyes. Therefore, ANN was used for classifying the different kind of the injected materials. The number of original images  taken from ultrasonographs were not so many, therefore we did data augmentation for providing large enough dataset that fed into ANN. The data augmentation was constructed by pixel shifting in horizontal and vertical directions. The procedure proposed here produced 98.2% accuracy for predicting test dataset,  though the result was sensitive to the choice of augmentation area.

scholarly journals Data Augmentation and Artificial Neural Networks for Eddy Currents Testing

2020 ◽  
Author(s):  
Romain Cormerais ◽  
Roberto Longo ◽  
Aroune Duclos ◽  
Guillaume Wasselynck ◽  
Gérard Berthiau
Keyword(s):  
Eddy Currents ◽  
Data Augmentation ◽  
Numerical Models ◽  
Simulated Data ◽  
Principal Component ◽  
Physical Models ◽  
Destructive Testing ◽  
Artificial Neural ◽  
Artificial Neural Network Ann ◽  
Non Destructive

Eddy Currents (ECs) Non Destructive Testing (NDT) is widely used to determine the position and size of flaws in metal materials. Due to difficulties in estimating these parameters via inverse algorithms based on physical models, approaches focused on Artificial Neural Network (ANN) are nowadays of great interest. The main drawbacks of these techniques still reside in the complexity of the numerical models and the large number of simulated data needed to train and test the ANN, leading to a considerable amount of calculation time and resources. To overcome these limitations, this article proposes a new approach based on a data augmentation procedure via Principal Component Analysis (PCA) applied to numerical simulations.

Mind wandering as data augmentation: How mental travel supports abstraction

2020 ◽  
Vol 43 ◽  
Author(s):  
Myrthe Faber
Keyword(s):  
Machine Learning ◽  
Data Augmentation ◽  
Mental Content ◽  
Mind Wandering ◽  
Theoretical Framework ◽  
Important Addition

Abstract Gilead et al. state that abstraction supports mental travel, and that mental travel critically relies on abstraction. I propose an important addition to this theoretical framework, namely that mental travel might also support abstraction. Specifically, I argue that spontaneous mental travel (mind wandering), much like data augmentation in machine learning, provides variability in mental content and context necessary for abstraction.

Sign in / Sign up

Export Citation Format

Share Document