Modelling of Behavior for Inhibition Corrosion of Bronze Using Artificial Neural Network (ANN)

Millán-Ocampo, D. E.; Parrales, A.; Martı́nez, Susana Silva; Porcayo-Calderón, J.; Hernández-Pérez, José Alfredo

doi:10.3390/e20060409

Cited by 16 publications

(6 citation statements)

References 34 publications

(34 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Isomura [5] applied neural networks to knowledge inference applications, useful information was collected, and an inference rule was extracted using deep neural networks (DNNs). Elusai [6] used ANNs to model behavior to prevent bronze corrosion. In the scheme, corrosion types were classified into different features, and future corrosion behaviors were predicted.…”

Section: Introductionmentioning

confidence: 99%

A Comparison of Regularization Techniques in Deep Neural Networks

Nusrat

Jang

2018

Symmetry

View full text Add to dashboard Cite

Artificial neural networks (ANN) have attracted significant attention from researchers because many complex problems can be solved by training them. If enough data are provided during the training process, ANNs are capable of achieving good performance results. However, if training data are not enough, the predefined neural network model suffers from overfitting and underfitting problems. To solve these problems, several regularization techniques have been devised and widely applied to applications and data analysis. However, it is difficult for developers to choose the most suitable scheme for a developing application because there is no information regarding the performance of each scheme. This paper describes comparative research on regularization techniques by evaluating the training and validation errors in a deep neural network model, using a weather dataset. For comparisons, each algorithm was implemented using a recent neural network library of TensorFlow. The experiment results showed that an autoencoder had the worst performance among schemes. When the prediction accuracy was compared, data augmentation and the batch normalization scheme showed better performance than the others.

show abstract

Section: Introductionmentioning

confidence: 99%

A Comparison of Regularization Techniques in Deep Neural Networks

Nusrat

Jang

2018

Symmetry

View full text Add to dashboard Cite

show abstract

“…Linear models are not sufficient to explain all the sources of variability due to the complex nature of the relationships between molecular structure and activity. Artificial neural networks (ANN) are a type of machine‐learning prediction method with the ability to self‐learn relationships from labeled experimental data and generalize to unlabeled situations [20]. One of the most popular types of ANN used in biological research is multilayer perceptron (MLP).…”

Section: Methodsmentioning

confidence: 99%

Xanthine oxidase targeted model setup and its application for antihyperuricemic compounds prediction by in silico methods

Hou

Shi

Ren

2021

eFood

View full text Add to dashboard Cite

To achieve potential alternatives for hyperuricemia therapeutics, a novel structure-docking energy relationship model was established for high-throughput screening inhibitors of xanthine oxidase (XO). Molecular docking was performed between XO and 311 natural compounds from 6 traditional Chinese herbs. Then, structure-docking energy relationship model was simulated between molecular docking energy and 63 molecular descriptors by multiple linear regressions (MLR), principal component regression (PCR), and artificial neural network (ANN), respectively. The results showed that the ANN model was the best model to predict the docking energy of XO with the coefficient of determination (R2) and mean squared error (MSE) at 0.8746 and 0.9414, respectively. The data of XO inhibitory activity were consistent with the prediction in vitro, which was also further confirmed by hyperuricemia cell model. The results suggested that the structure-docking energy relationship model provides a paradigm framework for the screening of XO inhibitors.

show abstract

“…It is possible to find an adequate architecture from the variation of the number of neurons in the hidden layer with the help of an optimization algorithm, an activation function, and a certain number of iterations [24]. The most suitable optimization algorithm is the Levenberg-Marquardt backpropagation algorithm (LMA), since it allows the decrease of the mean square error (MSE), as opposed to other backpropagation algorithms [25].…”

Section: Ann Modified Wilson Plot Methodologymentioning

confidence: 99%

Heat Transfer Coefficients Analysis in a Helical Double-Pipe Evaporator: Nusselt Number Correlations through Artificial Neural Networks

Parrales

Hernández-Pérez

Flores

et al. 2019

Entropy

Self Cite

View full text Add to dashboard Cite

In this study, two empirical correlations of the Nusselt number, based on two artificial neural networks (ANN), were developed to determine the heat transfer coefficients for each section of a vertical helical double-pipe evaporator with water as the working fluid. Each ANN was obtained using an experimental database of 1109 values obtained from an evaporator coupled to an absorption heat transformer with energy recycling. The Nusselt number in the annular section was estimated based on the modified Wilson plot method solved by an ANN. This model included the Reynolds and Prandtl numbers as input variables and three neurons in their hidden layer. The Nusselt number in the inner section was estimated based on the Rohsenow equation, solved by an ANN. This ANN model included the numbers of the Prandtl and Jackob liquids as input variables and one neuron in their hidden layer. The coefficients of determination were R 2 > 0.99 for both models. Both ANN models satisfied the dimensionless condition of the Nusselt number. The Levenberg-Marquardt algorithm was chosen to determine the optimum values of the weights and biases. The transfer functions used for the learning process were the hyperbolic tangent sigmoid in the hidden layer and the linear function in the output layer. The Nusselt numbers, determined by the ANNs, proved adequate to predict the values of the heat transfer coefficients of a vertical helical double-pipe evaporator that considered biphasic flow with an accuracy of ±0.2 for the annular Nusselt and ±4 for the inner Nusselt.Generally, in the case of heat exchangers with helical pipelines, it is known that the flow behavior has a greater complexity compared to exchangers that include straight pipes [2]. Due to the generation of vortices, induced by the interaction between the viscous forces and the centrifugal force, a disturbance in the flow is generated, which increases the rate of the heat transfer and the friction losses-a behavior referred to as secondary flow [3].If a helical exchanger transports a biphasic flow, specifically gas-liquid, the complexity is even higher due to the sudden changes in the flow pattern. These phenomena are mainly caused by the sliding between the phases, the orientation of the pipe, and the void fraction, among other additional variables and the relationships between them [4]. Therefore, to know and understand these phenomena, it is necessary to estimate the values of the heat transfer coefficient in the helical exchangers that contain biphasic flows, and hence any method that leads to their calculation is considered useful.One of the methods commonly used to estimate heat transfer coefficients is the Wilson plot method, proposed by Wilson in 1915 and subsequently modified in 1957 [5]. This traditional method quantifies the heat flow for each section by the exchanger through the measurement of the global temperature between the two fluids involved and their rate of heat transfer [6]. One of the most significant advantages of this technique is that it avoids the complex dire...

show abstract

Modelling of Behavior for Inhibition Corrosion of Bronze Using Artificial Neural Network (ANN)

Cited by 16 publications

References 34 publications

A Comparison of Regularization Techniques in Deep Neural Networks

A Comparison of Regularization Techniques in Deep Neural Networks

Xanthine oxidase targeted model setup and its application for antihyperuricemic compounds prediction by in silico methods

Heat Transfer Coefficients Analysis in a Helical Double-Pipe Evaporator: Nusselt Number Correlations through Artificial Neural Networks

Contact Info

Product

Resources

About