scholarly journals Artificial Neural Network Potentials for Mechanics and Fracture Dynamics of Materials

2022 ◽  
Author(s):  
Gang Seob Jung ◽  
Hoon Joo Myung ◽  
Stephan Irle
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Crack Propagation ◽  
Density Functional ◽  
Tight Binding ◽  
Training Data ◽  
Data Generation ◽  
Deformation And Failure ◽  
Deformation And Fracture ◽  
Artificial Neural

Abstract Atomistic understanding of mechanics and failure of materials is the key for engineering and applications. Modeling accurately brittle failure with crack propagation in covalent crystals requires a quantum mechanics-based description of individual bond-breaking events for large system sizes. Machine Learned (ML) potentials have emerged to overcome the traditional, physics-based modeling tradeoff between accuracy and accessible time and length scales. Previous studies have shown successful applications of ML potentials for describing the structure and dynamics of molecular systems and amorphous or liquid phases of materials. However, their application to deformation and failure processes in materials is yet uncommon. In this study, we discuss apparent limitations of ML potentials to describe deformation and fracture under loadings and propose a way to generate and select training data for their employment in simulations of deformation and fracture of crystals. We applied the proposed approach to 2D crystal graphene, utilizing the density-functional tight-binding (DFTB) method for more efficient and extensive data generation in place of density functional theory (DFT). Then, we explore how the data selection affects the accuracy of the developed artificial neural network potential (NNP), indicating that only the errors in total energies and atomic forces are insufficient to judge the NNP’s reliability. Therefore, we evaluate and select NNPs based on their performance in describing physical properties, e.g., stress-strain curves and geometric deformation. In sharp contrast to popular reactive bond order potentials, our optimized NNP predicts straight crack propagation in graphene along both armchair and zigzag lattice directions, as well as higher fracture toughness of zigzag edge direction. Our study provides significant insight into crack propagation mechanisms at atomic scales and highlights strategies for NNP developments of broader materials.

scholarly journals Identification Of Number Using Artificial Neural Network Backpropagation

2018 ◽  
Vol 215 ◽  
pp. 01011
Author(s):  
Sitti Amalia
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Training Data ◽  
Simulation Test ◽  
Backpropagation Algorithm ◽  
Network Parameter ◽  
Artificial Intelligent ◽  
Design And Implementation ◽  
Network Training ◽  
Artificial Neural

This research proposed to design and implementation system of voice pattern recognition in the form of numbers with offline pronunciation. Artificial intelligent with backpropagation algorithm used on the simulation test. The test has been done to 100 voice files which got from 10 person voices for 10 different numbers. The words are consisting of number 0 to 9. The trial has been done with artificial neural network parameters such as tolerance value and the sum of a neuron. The best result is shown at tolerance value varied and a sum of the neuron is fixed. The percentage of this network training with optimal architecture and network parameter for each training data and new data are 82,2% and 53,3%. Therefore if tolerance value is fixed and a sum of neuron varied gave 82,2% for training data and 54,4% for new data

scholarly journals Modeling the concentration of carbonyl of ethylene propylene diene monomer during the thermal aging using artificial neural network

2021 ◽  
Vol 03 (01) ◽  
pp. 45-52
Author(s):  
Hadjira Maouz ◽  
◽  
Asma Adda ◽  
Salah Hanini ◽  
◽  
...  
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Thermal Aging ◽  
Training Data ◽  
Ethylene Propylene Diene Monomer ◽  
Ann Model ◽  
Ethylene Propylene ◽  
Artificial Neural ◽  
Input Variables ◽  
Artificial Neural Network Ann

The concentration of carbonyl is one of the most important properties contributing to the detection of the thermal aging of polymer ethylene propylene diene monomer (EPDM). In this publication, an artificial neural network (ANN) model was developed to predict concentration of carbenyl during the thermal aging of EPDM using a database consisting of seven input variables. The best fitting training data was obtained with the architecture of (7 inputs neurons, 10 hidden neurons and 1 output neuron). A Levenberg Marquardt learning (LM) algorithm, hyperbolic tangent transfer function were used at the hidden and output layer respectively. The optimal ANN was obtained with a high correlation coefficient R= 0.995 and a very low root mean square error RMSE = 0.0148 mol/l during the generalization phase. The comparison between the experimental and calculated results show that the ANN model is able of predicted the concentration of carbonyl during the thermal aging of ethylene propylene diene monomer

scholarly journals Pengaruh Latar Belakang Warna pada Objek Gambar terhadap Hasil Ekstraksi Sinyal EEG

2017 ◽  
Vol 7 (2) ◽  
pp. 161 ◽  
Author(s):  
Catur Atmaji ◽  
Zandy Yudha Perwira
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Training System ◽  
Training Data ◽  
Spectrum Estimation ◽  
Quality Classification ◽  
Eeg Data ◽  
Two Systems ◽  
Artificial Neural ◽  
Color Background

In this study, observation on the differences in features quality of EEG records as a result of training on subjects has been made. The features of EEG records were extracted using two different methods, the root mean square which is acquired from the range between 0.5 and 5 seconds and the average of power spectrum estimation from the frequency range between 20 and 40Hz. All of the data consists of a 4-channel recording and produce good quality classification on artificial neural network, with each of which generates training data accuracy over 90%. However, different results are occured when the trained system is tested on other test data. The test results show that the two systems which are trained using training data with object with color background produce higher accuracy than the other two systems which are trained using training data with object without background color, 63.98% and 60.22% compared to 59.68% and 56.45% accuracy respectively. From the use of the features on the artificial neural network classification system, it can be concluded that the training system using EEG data records derived from the visualization of object with color background produces better features than the visualization of object without color background.

scholarly journals Modeling of daily runoff from a small agricultural watershed using artificial neural network with resampling techniques

2014 ◽  
Vol 17 (1) ◽  
pp. 56-74 ◽  
Author(s):  
Gurjeet Singh ◽  
Rabindra K. Panda ◽  
Marc Lamers
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Agricultural Watershed ◽  
Training Data ◽  
Optimum Number ◽  
Training Process ◽  
Data Set ◽  
Reported Study ◽  
Artificial Neural ◽  
Artificial Neural Network Ann

The reported study was undertaken in a small agricultural watershed, namely, Kapgari in Eastern India having a drainage area of 973 ha. The watershed was subdivided into three sub-watersheds on the basis of drainage network and land topography. An attempt was made to relate the continuously monitored runoff data from the sub-watersheds and the whole-watershed with the rainfall and temperature data using the artificial neural network (ANN) technique. The reported study also evaluated the bias in the prediction of daily runoff with shorter length of training data set using different resampling techniques with the ANN modeling. A 10-fold cross-validation (CV) technique was used to find the optimum number of hidden neurons in the hidden layer and to avoid neural network over-fitting during the training process for shorter length of data. The results illustrated that the ANN models developed with shorter length of training data set avoid neural network over-fitting during the training process, using a 10-fold CV method. Moreover, the biasness was investigated using the bootstrap resampling technique based ANN (BANN) for short length of training data set. In comparison with the 10-fold CV technique, the BANN is more efficient in solving the problems of the over-fitting and under-fitting during training of models for shorter length of data set.

scholarly journals Fast Linear Adaptive Skipping Training Algorithm for Training Artificial Neural Network

2013 ◽  
Vol 2013 ◽  
pp. 1-9 ◽  
Author(s):  
R. Manjula Devi ◽  
S. Kuppuswami ◽  
R. C. Suganthe
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Training Model ◽  
Training Data ◽  
Experimental Result ◽  
Training Algorithms ◽  
Data Set ◽  
Training Time ◽  
Input Sample ◽  
Artificial Neural

Artificial neural network has been extensively consumed training model for solving pattern recognition tasks. However, training a very huge training data set using complex neural network necessitates excessively high training time. In this correspondence, a new fast Linear Adaptive Skipping Training (LAST) algorithm for training artificial neural network (ANN) is instituted. The core essence of this paper is to ameliorate the training speed of ANN by exhibiting only the input samples that do not categorize perfectly in the previous epoch which dynamically reducing the number of input samples exhibited to the network at every single epoch without affecting the network’s accuracy. Thus decreasing the size of the training set can reduce the training time, thereby ameliorating the training speed. This LAST algorithm also determines how many epochs the particular input sample has to skip depending upon the successful classification of that input sample. This LAST algorithm can be incorporated into any supervised training algorithms. Experimental result shows that the training speed attained by LAST algorithm is preferably higher than that of other conventional training algorithms.

scholarly journals A Machine Learning-based Pipeline for the Classification of CTX-M in Metagenomics Samples

Processes ◽  
2019 ◽  
Vol 7 (4) ◽  
pp. 235 ◽  
Author(s):  
Diego Ceballos ◽  
Diana López-Álvarez ◽  
Gustavo Isaza ◽  
Reinel Tabares-Soto ◽  
Simón Orozco-Arias ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Artificial Neural Network ◽  
Bacterial Infections ◽  
Training Data ◽  
Metagenomic Data ◽  
Process Unit ◽  
M Gene ◽  
Central Process ◽  
Artificial Neural

Bacterial infections are a major global concern, since they can lead to public health problems. To address this issue, bioinformatics contributes extensively with the analysis and interpretation of in silico data by enabling to genetically characterize different individuals/strains, such as in bacteria. However, the growing volume of metagenomic data requires new infrastructure, technologies, and methodologies that support the analysis and prediction of this information from a clinical point of view, as intended in this work. On the other hand, distributed computational environments allow the management of these large volumes of data, due to significant advances in processing architectures, such as multicore CPU (Central Process Unit) and GPGPU (General Propose Graphics Process Unit). For this purpose, we developed a bioinformatics workflow based on filtered metagenomic data with Duk tool. Data formatting was done through Emboss software and a prototype of a workflow. A pipeline was also designed and implemented in bash script based on machine learning. Further, Python 3 programming language was used to normalize the training data of the artificial neural network, which was implemented in the TensorFlow framework, and its behavior was visualized in TensorBoard. Finally, the values from the initial bioinformatics process and the data generated during the parameterization and optimization of the Artificial Neural Network are presented and validated based on the most optimal result for the identification of the CTX-M gene group.

scholarly journals Prediction Data Processing Scheme using an Artificial Neural Network and Data Clustering for Big Data

2016 ◽  
Vol 6 (1) ◽  
pp. 330 ◽  
Author(s):  
Se-Hoon Jung ◽  
Jong-Chan Kim ◽  
Chun-Bo Sim
Keyword(s):  
Neural Network ◽  
Principal Component Analysis ◽  
Artificial Neural Network ◽  
Data Clustering ◽  
Principal Component ◽  
Component Analysis ◽  
Learning Model ◽  
Data Generation ◽  
Processing Scheme ◽  
Artificial Neural

Various types of derivative information have been increasing exponentially, based on mobile devices and social networking sites (SNSs), and the information technologies utilizing them have also been developing rapidly. Technologies to classify and analyze such information are as important as data generation. This study concentrates on data clustering through principal component analysis and K-means algorithms to analyze and classify user data efficiently. We propose a technique of changing the cluster choice before cluster processing in the existing K-means practice into a variable cluster choice through principal component analysis, and expanding the scope of data clustering. The technique also applies an artificial neural network learning model for user recommendation and prediction from the clustered data. The proposed processing model for predicted data generated results that improved the existing artificial neural network–based data clustering and learning model by approximately 9.25%.

scholarly journals PREDIKSI BEBAN ENERGI LISTRIK APJ KOTA SEMARANG MENGGUNAKAN METODE RADIAL BASIS FUNCTION (RBF)

eLEKTRIKA ◽  
2019 ◽  
Vol 11 (2) ◽  
pp. 21
Author(s):  
Mukti Dwi Cahyo ◽  
Sri Heranurweni ◽  
Harmini Harmini
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Radial Basis Function ◽  
Basis Function ◽  
Peak Load ◽  
Electricity Demand ◽  
Training Data ◽  
Radial Basis ◽  
Artificial Neural ◽  
Artificial Neural Network Ann

Electric power is one of the main needs of society today, ranging from household consumers to industry. The demand for electricity increases every year. So as to achieve adjustments between power generation and power demand, the electricity provider (PLN) must know the load needs or electricity demand for some time to come. There are many studies on the prediction of electricity loads in electricity, but they are not specific to each consumer sector. One of the predictions of this electrical load can be done using the Radial Basis Function Artificial Neural Network (ANN) method. This method uses training data learning from 2010 - 2017 as a reference data. Calculations with this method are based on empirical experience of electricity provider planning which is relatively difficult to do, especially in terms of corrections that need to be made to changes in load. This study specifically predicts the electricity load in the Semarang Rayon network service area in 2019-2024. The results of this Artificial Neural Network produce projected electricity demand needs in 2019-2024 with an average annual increase of 1.01% and peak load in 2019-2024. The highest peak load in 2024 and the dominating average is the household sector with an increase of 1% per year. The accuracy results of the Radial Basis Function model reached 95%.

scholarly journals Fault Identification by Extreme Pressure in Oil Pipelines Using Artificial Neural Network

2019 ◽  
Vol 8 (6) ◽  
pp. 90-96
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Contextual Information ◽  
Training Data ◽  
Fault Identification ◽  
Pipeline System ◽  
Extreme Pressure ◽  
Feed Forward ◽  
Oil Pipeline ◽  
Artificial Neural

Conventional Artificial Neural Network approaches such as Feed-Forward Networks has been used in diverse applications but are not naturally predictive and also require supervised learning. Feed-forward Artificial Neural Network also trained by backpropagation poses the problem of varnishing gradient. Algorithm using Gaussian membership function with a context-decision gate for detection operations has been proposed as an alternative to the traditional Feed Forward Architecture. The AI monitoring System shows promising results in solving many recurrent problems, particularly those requiring long-term storage dependencies - the Vanishing Gradient problem (VGP) and has the ability to use contextual information when mapping between input and output sequences. The Oil monitoring system employs dynamic data flow modeling to simulate the behavior of probably militant behaviors. The contextual information (context data) includes such context as Pressure from the lab scale experimental setup of oil pipeline system. In this approach, not only the networks are trained to adapt to the given training data, the training data (the expected outputs of fault indices) is also updated to adapt to the neural network. During the training procedure, both the neural networks and training data are updated interactively. Dynamic simulations were performed using a real-time data obtained from the Radial Bias Kernel Network. The data is tested using AI system in MATLAB-SIMULINK environment to verify the performance of the proposed system. The results were promising indicating the real state of fault identification in oil pipeline system caused by extreme pressure during transportation.

Sign in / Sign up

Export Citation Format

Share Document