A Methodology to Estimate Degradation and Reliability Based on Artificial Neural Networks

2005 ◽  
Author(s):  
Enrique A. Susemihl ◽  
Shuzhen Xu
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Failure Modes ◽  
Preliminary Results ◽  
Physical Variables ◽  
Artificial Neural ◽  
Paper Insulation

Failure modes associated with degradation or ageing affect most equipment, and the estimates of failure due to them are particularly important in deciding repair or replacement. A methodology is presented here to relate physical variables with a degradation measure by using artificial neural networks to capture data and experience, and to use this degradation measure to estimate probabilities of failure. This methodology has been applied to transformers to estimate probabilities of failure due to the degradation of paper insulation, and some preliminary results are presented. These results show that the method can provide reasonable estimates.

scholarly journals Improving Full-Body Pose Estimation from a Small Sensor Set Using Artificial Neural Networks and a Kalman Filter

2019 ◽  
Vol 33 ◽  
pp. 10063-10064
Author(s):  
Frank J. Wouda ◽  
Matteo Giuberti ◽  
Giovanni Bellusci ◽  
Bert-Jan F. Van Beijnum ◽  
Peter H. Veltink
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Kalman Filter ◽  
Pose Estimation ◽  
Preliminary Results ◽  
State Prediction ◽  
Artificial Neural ◽  
Body Pose Estimation ◽  
Full Body

Previous research has shown that estimating full-body poses from a minimal sensor set using a trained ANN without explicitly enforcing time coherence has resulted in output pose sequences that occasionally show undesired jitter. To mitigate such effect, we propose to improve the ANN output by combining it with a state prediction using a Kalman Filter. Preliminary results are promising, as the jitter effects are diminished. However, the overall error does not decrease substantially.

Optimal Design for Micromixer Using Macromodel Based on Artificial Neural Networks

2013 ◽  
Vol 339 ◽  
pp. 55-58
Author(s):  
Xue Ye Chen ◽  
Hui Xu
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Navier Stokes ◽  
Model Design ◽  
Design Variables ◽  
Input Layer ◽  
The Neural Networks ◽  
Physical Variables ◽  
Artificial Neural ◽  
Hidden Layer

The micromixer device is modeled using artificial neural networks trained with finite element simulations of the underlying incompressible Navier-Stokes and mass transport PDEs. The neural networks design is based on a three layers perceptron with one input layer, one nonlinear hidden layer and one linear output layer. The neural networks can map the micromixer behavior into a set of analytical performance functions parameterized by the systems physical variables. The macromodel has been extracted from training output of the artificial neural networks. Three design variables, i.e., the flow velocity, the channel width, and the numbers of the mixing unit are selected for model design. The mixing index at the end of the serpentine channels is employed as the objective function. The macromodel has been validated with numerical simulations. It can be demonstrated that this macromodel should facilitate the design of microfluidic device with sophisticated channels networks.

Performance evaluation of artificial neural networks for identification of failure modes in composite plates

2021 ◽  
Vol 63 (6) ◽  
pp. 565-570
Author(s):  
Serkan Balli ◽  
Faruk Sen
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Failure Modes ◽  
Experimental Studies ◽  
Composite Plates ◽  
Geometric Parameters ◽  
Sandwich Composite ◽  
Geometrical Parameters ◽  
Training Algorithms ◽  
Artificial Neural

Abstract The aim of this work is to identify failure modes of double pinned sandwich composite plates by using artificial neural networks learning algorithms and then analyze their accuracies for identification. Mechanically pinned specimens with two serial pins/bolts for sandwich composite plates were used for recognition of failure modes which were obtained in previous experimental studies. In addition, the empirical data of the preceding work was determined with various geometric parameters for various applied preload moments. In this study, these geometric parameters and fastened/bolted joint forms were used for training by artificial neural networks. Consequently, ten different backpropagation training algorithms of artificial neural network were applied for classification by using one hundred data values containing three geometrical parameters. According to obtained results, it was seen that the Levenberg-Marquardt backpropagation training algorithm was the most successful algorithm with 93 % accuracy rate and it was appropriate for modeling of this problem. Additionally, performances of all backpropagation training algorithms were discussed taking into account accuracy and error ratios.

Preliminary results on using artificial neural networks for security assessment (of power systems)

1991 ◽  
Vol 6 (2) ◽  
pp. 890-896 ◽  
Author(s):  
M. Aggoune ◽  
M.A. El-Sharkawa ◽  
D.C. Park ◽  
M.J. Damborg ◽  
R.J. Marks
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Power Systems ◽  
Security Assessment ◽  
Preliminary Results ◽  
Artificial Neural

Prescreening entire mammograms for masses with artificial neural networks: Preliminary results

1997 ◽  
Vol 4 (6) ◽  
pp. 405-414 ◽  
Author(s):  
Barry L. Kalman ◽  
William R. Reinus ◽  
Stan C. Kwasny ◽  
Andrew Laine ◽  
Lawrence Kotner
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Preliminary Results ◽  
Artificial Neural

WEAR PARTICLE TEXTURE CLASSIFICATION USING ARTIFICIAL NEURAL NETWORKS

1999 ◽  
Vol 13 (03) ◽  
pp. 415-428 ◽  
Author(s):  
M. S. LAGHARI ◽  
A. BOUJARWAH
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Visual Inspection ◽  
Failure Modes ◽  
Texture Classification ◽  
Wear Particle ◽  
Metal Particles ◽  
Wear Particles ◽  
Artificial Neural ◽  
Inspection Techniques

Analysis of wear debris carried by a lubricant in an oil-wetted system provides important information about the condition of a machine. This paper describes the analysis of microscopic metal particles generated by wear using computer vision and image processing. The aim is to classify these particles according to their morphology and surface texture and by using the information obtained, to predict wear failure modes in engines and other machinery. This approach obviates the need for specialists and reliance on human visual inspection techniques. The procedure reported in this paper, is used to classify surface features of the wear particles by using artificial neural networks. A visual comparison between cooccurrence matrices representing five different texture classes is described. Based on these comparisons, matrices of reduced sizes are utilized to train a feed-forward neural classifier in order to distinguish between the various texture classes.

Sign in / Sign up

Export Citation Format

Share Document