scholarly journals A Monte Carlo Simulation Approach in Non-linear Structural Dynamics Using Convolutional Neural Networks

2021 ◽  
Vol 7 ◽  
Author(s):  
Franz Bamer ◽  
Denny Thaler ◽  
Marcus Stoffel ◽  
Bernd Markert
Keyword(s):  
Neural Network ◽  
Monte Carlo Simulation ◽  
Monte Carlo ◽  
Convolutional Neural Network ◽  
Ground Motion ◽  
Structural Response ◽  
Structural Failure ◽  
Non Linear ◽  
Time Histories ◽  
Sample Set

The evaluation of the structural response statistics constitutes one of the principal tasks in engineering. However, in the tail region near structural failure, engineering structures behave highly non-linear, making an analytic or closed form of the response statistics difficult or even impossible. Evaluating a series of computer experiments, the Monte Carlo method has been proven a useful tool to provide an unbiased estimate of the response statistics. Naturally, we want structural failure to happen very rarely. Unfortunately, this leads to a disproportionately high number of Monte Carlo samples to be evaluated to ensure an estimation with high confidence for small probabilities. Thus, in this paper, we present a new Monte Carlo simulation method enhanced by a convolutional neural network. The sample-set used for this Monte Carlo approach is provided by artificially generating site-dependent ground motion time histories using a non-linear Kanai-Tajimi filter. Compared to several state-of-the-art studies, the convolutional neural network learns to extract the relevant input features and the structural response behavior autonomously from the entire time histories instead of learning from a set of hand-chosen intensity inputs. Training the neural network based on a chosen input sample set develops a meta-model that is then used as a meta-model to predict the response of the total Monte Carlo sample set. This paper presents two convolutional neural network-enhanced strategies that allow for a practical design approach of ground motion excited structures. The first strategy enables for an accurate response prediction around the mean of the distribution. It is, therefore, useful regarding structural serviceability. The second strategy enables for an accurate prediction around the tail end of the distribution. It is, therefore, beneficial for the prediction of the probability of failure.

Research on fault diagnosis of automobile engines based on the deep learning 1D-CNN method

2022 ◽  
Author(s):  
Canyi Du ◽  
Rui Zhong ◽  
Yishen Zhuo ◽  
Xinyu Zhang ◽  
Feifei Yu ◽  
...  
Keyword(s):  
Neural Network ◽  
Pattern Recognition ◽  
Deep Learning ◽  
Fault Diagnosis ◽  
Convolutional Neural Network ◽  
Pattern Recognition Method ◽  
Recognition Method ◽  
Vibration Signals ◽  
One Dimensional ◽  
Sample Set

Abstract Traditional engine fault diagnosis methods usually need to extract the features manually before classifying them by the pattern recognition method, which makes it difficult to solve the end-to-end fault diagnosis problem. In recent years, deep learning has been applied in different fields, bringing considerable convenience to technological change, and its application in the automotive field also has many applications, such as image recognition, language processing, and assisted driving. In this paper, a one-dimensional convolutional neural network (1D-CNN) in deep learning is used to process vibration signals to achieve fault diagnosis and classification. By collecting the vibration signal data of different engine working conditions, the collected data are organized into several sets of data in a working cycle, which are divided into a training sample set and a test sample set. Then, a one-dimensional convolutional neural network model is built in Python to allow the feature filter (convolution kernel) to learn the data from the training set and these convolution checks process the input data of the test set. Convolution and pooling extract features to output to a new space, which is characterized by learning features directly from the original vibration signals and completing fault diagnosis. The experimental results show that the pattern recognition method based on a one-dimensional convolutional neural network can be effectively applied to engine fault diagnosis and has higher diagnostic accuracy than traditional methods.

scholarly journals Investigating the Electronic Coefficients in Ne-Hg Mixtures: Comparing a Monte Carlo and Artificial Neural Network Model

2020 ◽  
Vol 55 (2) ◽  
Author(s):  
Rafid Abbas Ali ◽  
Faten Sajet Mater ◽  
Asmaa Satar Jeeiad Al-Ragehey
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Monte Carlo Simulation ◽  
Monte Carlo ◽  
Network Architecture ◽  
Vapor Concentration ◽  
Neural Network Architecture ◽  
Marquardt Algorithm ◽  
Artificial Neural ◽  
New Applications

Electron coefficients such as drift velocity, ionization coefficient, mean electron energy and Townsend energy for different concentrations of Hg 0.1%, 1%, 10% and 50% in the Ne-Hg mixture at a reduced electric field were calculated using two approaches taking into account inelastic collisions: The Monte Carlo simulation, and an artificial neural network. The effect of Hg vapor concentration on the electron coefficients showed that insignificant additions of mercury atom impurities to Neon, starting from fractions of a percent, affect the characteristics of inelastic processes and discharge, respectively. The aim of this paper is to explore the new applications of neural networks. The Levenberg-Marquardt algorithm and artificial neural network architecture employed was presented in this work to calculate the electron coefficients for different concentrations of Hg in Ne-Hg mixtures. The artificial neural network has been trained with four models (M1, M2, M3, M4), and analysis of the regression between the values of an artificial neural network and Monte Carlo simulation indicates that the M2 output provided the best perfect correlation at 100 Epochs, and the output data obtained was closest to the target data required through using the different stages of artificial neural network development starting with design, training and testing.

General Regression Neural Network and Monte Carlo Simulation Model for Survival and Growth of Salmonella on Raw Chicken Skin as a Function of Serotype, Temperature, and Time for Use in Risk Assessment†

2009 ◽  
Vol 72 (10) ◽  
pp. 2078-2087 ◽  
Author(s):  
THOMAS P. OSCAR
Keyword(s):  
Neural Network ◽  
Risk Assessment ◽  
Monte Carlo Simulation ◽  
Monte Carlo ◽  
General Regression Neural Network ◽  
Chicken Skin ◽  
Survival And Growth ◽  
Fail Safe ◽  
General Regression ◽  
Monte Carlo Simulation Model

A general regression neural network (GRNN) and Monte Carlo simulation model for predicting survival and growth of Salmonella on raw chicken skin as a function of serotype (Typhimurium, Kentucky, and Hadar), temperature (5 to 50°C), and time (0 to 8 h) was developed. Poultry isolates of Salmonella with natural resistance to antibiotics were used to investigate and model survival and growth from a low initial dose (<1 log) on raw chicken skin. Computer spreadsheet and spreadsheet add-in programs were used to develop and simulate a GRNN model. Model performance was evaluated by determining the percentage of residuals in an acceptable prediction zone from −1 log (fail-safe) to 0.5 log (fail-dangerous). The GRNN model had an acceptable prediction rate of 92% for dependent data (n = 464) and 89% for independent data (n = 116), which exceeded the performance criterion for model validation of 70% acceptable predictions. Relative contributions of independent variables were 16.8% for serotype, 48.3% for temperature, and 34.9% for time. Differences among serotypes were observed, with Kentucky exhibiting less growth than Typhimurium and Hadar, which had similar growth levels. Temperature abuse scenarios were simulated to demonstrate how the model can be integrated with risk assessment, and the most common output distribution obtained was Pearson5. This study demonstrated that it is important to include serotype as an independent variable in predictive models for Salmonella. Had a cocktail of serotypes Typhimurium, Kentucky, and Hadar been used for model development, the GRNN model would have provided overly fail-safe predictions of Salmonella growth on raw chicken skin contaminated with serotype Kentucky. Thus, by developing the GRNN model with individual strains and then modeling growth as a function of serotype prevalence, more accurate predictions were obtained.

Sign in / Sign up

Export Citation Format

Share Document