Search results

The purpose of this paper is to present the results of experimental analysis of the elastic-guided wave mode conversion phenomenon in glass fiber-reinforced polymers. The results of this research presented in this paper are strictly focused on S0/A0’ mode conversion phenomenon caused by discontinuities in the form of circular Teflon inserts (artificial delaminations) and impact damage. Results of this research could be useful in problems of damage detection and localization.

In the research, guided waves are excited using a piezoelectric transducer and sensed in a non-contact manner using a scanning laser Doppler vibrometer. Full wavefield measurements are analyzed. Analysis of the influence of investigated discontinuities on S0/A0’ mode conversion is based on the elastic wave mode filtration in frequency-wavenumber domain. Mode filtration process allows us to remove the effects of the propagation of unwanted type of mode in forward or backward direction. Effects of S0/A0’ mode conversion are characterized by a mode conversion indicator (MCI) based on the amplitude of new mode A0’ and the amplitude of incident S0 mode.

It was noticed that the magnitude of MCI depends on the depth at which the Teflon inserts were located for all analyzed excitation frequencies and diameters of inserts (10 and 20 mm). The magnitude of MCI also increases with increasing impact energies. The S0/A0’ mode conversion phenomenon could be utilized for the detection of surface and internal located discontinuities.

This paper presents the original results of this research related to the influence of discontinuity location with respect to the sample thickness and severity of discontinuity on S0/A0’ mode conversion.

In this manuscript, the authors aimed to demonstrate the influences of influential parameters in mixed-mode crack propagation phenomenon. The authors attempted to cover almost all surrounding issues of this subject as the authors know simulating of propagating cracks as internal strong discontinuity is a complicated issue.

In this manuscript, the authors demonstrated the influences of influential parameters in mixed-mode crack propagation phenomenon. The authors attempted to cover almost all surrounding issues of this subject as the authors know simulating of propagating cracks as internal strong discontinuity is a complicated issue. Furthermore, three different scenarios for crack growth are considered. In reality, edge-cracked plate, center-cracked plate and cracked plate in the presence of void and inclusion are studied. In fact, by designing suitable artificial neural network's (ANN) architectures all the three aforementioned conditions are trained and estimated through those architectures with very good agreement with input data. Also by conducting a series of sensitivity analysis, the most affecting factors in mixed-mode crack propagation in different situations are demonstrated. The obtained results are very interesting and useful for other researchers and also the authors hope the results would be cited by researchers.

The influential parameters on mixed-mode crack propagation were found in this paper.

The computer code using MATLAB was prepared to study the mixed-mode crack paths. Also using ANNs toolbox, the crack path estimation was investigated.

The performance of a welding process determines not only the cost, but also the quality of the product. How to control the welding process in order to ensure good welding performance with less cost and higher Productivity has become critical. The objective of this study is twofold: (1) developing artificial neural networks to predict welding performance using different learning algorithms: back propagation, simulated annealing and tabu search; (2) comparing and discussing the performance of neural networks trained using those algorithms. Statistical analysis shows that back propagation is able to make more accurate prediction than the other algorithms for this particular application. However, all three algorithms demonstrate impressive flexibility and robustness.

This paper aims to discuss the utilization of artificial neural networks (ANNs) and multiple regression method for estimating surface roughness in milling medium density fiberboard (MDF) material with a parallel robot.

In ANN modeling, performance parameters such as root mean square error, mean error percentage, mean square error and correlation coefficients (R²) for the experimental data were determined based on conjugate gradient back propagation, Levenberg–Marquardt (LM), resilient back propagation, scaled conjugate gradient and quasi-Newton back propagation feed forward back propagation training algorithm with logistic transfer function.

In the ANN architecture established for the surface roughness (Ra), three neurons [cutting speed (V), feed rate (f) and depth of cut (a)] were contained in the input layer, five neurons were included in its hidden layer and one neuron was contained in the output layer (3-5-1).Trials showed that LM learning algorithm was the best learning algorithm for the surface roughness. The ANN model obtained with the LM learning algorithm yielded estimation training values R² (97.5 per cent) and testing values R² (99 per cent). The R² for multiple regressions was obtained as 96.1 per cent.

The result of the surface roughness estimation model showed that the equation obtained from the multiple regressions with quadratic model had an acceptable estimation capacity. The ANN model showed a more dependable estimation when compared with the multiple regression models. Hereby, these models can be used to effectively control the milling process to reach a satisfactory surface quality.

The face recognition problem has a long history and a significant practical perspective and one of the practical applications of the theory of pattern recognition, to automatically localize the face in the image and, if necessary, identify the person in the face. Interests in the procedures underlying the process of localization and individual’s recognition are quite significant in connection with the variety of their practical application in such areas as security systems, verification, forensic expertise, teleconferences, computer games, etc. This paper aims to recognize facial images efficiently. An averaged-feature based technique is proposed to reduce the dimensions of the multi-expression facial features. The classifier model is generated using a supervised learning algorithm called a back-propagation neural network (BPNN), implemented on a MatLab R2017. The recognition rate and accuracy of the proposed methodology is comparable with other methods such as the principle component analysis and linear discriminant analysis with the same data set. In total, 150 faces subjects are selected from the Olivetti Research Laboratory (ORL) data set, resulting 95.6 and 85 per cent recognition rate and accuracy, respectively, and 165 faces subjects from the Yale data set, resulting 95.5 and 84.4 per cent recognition rate and accuracy, respectively.

Averaged-feature based approach (dimension reduction) and BPNN (generate supervised classifier).

The recognition rate is 95.6 per cent and recognition accuracy is 85 per cent for the ORL data set, whereas the recognition rate is 95.5 per cent and recognition accuracy is 84.4 per cent for the Yale data set.

Averaged-feature based method.

Gesture recognition plays an important role in many fields such as human–computer interaction, medical rehabilitation, virtual and augmented reality. Gesture recognition using wearable devices is a common and effective recognition method. This study aims to combine the inverse magnetostrictive effect and tunneling magnetoresistance effect and proposes a novel wearable sensing glove applied in the field of gesture recognition.

A magnetostrictive sensing glove with function of gesture recognition is proposed based on Fe-Ni alloy, tunneling magnetoresistive elements, Agilus30 base and square permanent magnets. The sensing glove consists of five sensing units to measure the bending angle of each finger joint. The optimal structure of the sensing units is determined through experimentation and simulation. The output voltage model of the sensing units is established, and the output characteristics of the sensing units are tested by the experimental platform. Fifteen gestures are selected for recognition, and the corresponding output voltages are collected to construct the data set and the data is processed using Back Propagation Neural Network.

The sensing units can detect the change in the bending angle of finger joints from 0 to 105 degrees and a maximum error of 4.69% between the experimental and theoretical values. The average recognition accuracy of Back Propagation Neural Network is 97.53% for 15 gestures.

The sensing glove can only recognize static gestures at present, and further research is still needed to recognize dynamic gestures.

A new approach to gesture recognition using wearable devices.

This study has a broad application prospect in the field of human–computer interaction.

The sensing glove can collect voltage signals under different gestures to realize the recognition of different gestures with good repeatability, which has a broad application prospect in the field of human–computer interaction.

The purpose of this paper is to study the electrochemical properties of electrode material on activated carbon double layer capacitors. It also tries to develop a prediction model to evaluate pore size value.

Back-propagation neural network (BPNN) prediction model is used to evaluate pore size value. Also, an improved heuristic approach genetic algorithm (HAGA) is used to search for the optimal relationship between process parameters and electrochemical properties.

A three-layer ANN is found to be optimum with the architecture of three and six neurons in the first and second hidden layer and one neuron in output layer. The simulation results show that the optimized design model based on HAGA can get the suitable process parameters.

HAGA BPNN is proved to be a practical and efficient way for acquiring information and providing optimal parameters about the activated carbon double layer capacitor electrode material.

recent years, facial expression recognition has been widely used in human machine interaction, clinical medicine and safe driving. However, there is a limitation that conventional recurrent neural networks can only learn the time-series characteristics of expressions based on one-way propagation information.

To solve such limitation, this paper proposes a novel model based on bidirectional gated recurrent unit networks (Bi-GRUs) with two-way propagations, and the theory of identity mapping residuals is adopted to effectively prevent the problem of gradient disappearance caused by the depth of the introduced network. Since the Inception-V3 network model for spatial feature extraction has too many parameters, it is prone to overfitting during training. This paper proposes a novel facial expression recognition model to add two reduction modules to reduce parameters, so as to obtain an Inception-W network with better generalization.

Finally, the proposed model is pretrained to determine the best settings and selections. Then, the pretrained model is experimented on two facial expression data sets of CK+ and Oulu- CASIA, and the recognition performance and efficiency are compared with the existing methods. The highest recognition rate is 99.6%, which shows that the method has good recognition accuracy in a certain range.

By using the proposed model for the applications of facial expression, the high recognition accuracy and robust recognition results with lower time consumption will help to build more sophisticated applications in real world.

Investigates the possibility of applying artificial intelligence to solve practical auditing problems faced by the public sector, namely the tax auditor of the Internal Revenue Services, when targeting firms for further investigation. Suggests that organizations which incorporate an operational artificial neural network system will raise their performance greatly. Proposes that the neural network will overcome problems faced by a direct knowledge acquisition method in building an expert system to preserve the expertise of senior auditors by the IRS in Taiwan. Provides an explanation of the neural network theory with regard to multi‐ and single‐layered neural networks. Statistics reveal the neural network performs favourably, and that three‐layer networks perform better than two‐layer neural networks. The results strongly suggest that neural networks can be used to identify firms requiring further auditing investigation, and also suggest future implications for intelligent auditing machines.

The purpose of this paper is to formulate a specialized artificial neural network algorithm utilizing radial basis function (RBF) for modeling of time to failure of aircraft engine turbines.

The model uses training failure data collected from operators of turboprop aircraft working in harsh desert conditions where sand erosion is a detrimental factor in reducing turbine life. Accordingly, the model is more suited to accurate prediction of life of critical components of such engines. The used RBF employs a closest neighbor type of classifier and the hidden unit’s activation is based on the displacement between the early prototype and the input vector.

The results of the algorithm are compared to earlier work utilizing Weibull regression modeling, as well as Feed Forward Back Propagation NN. The results show that the failure rates estimated by RBF more closely match actual failure data than the estimations by both other models. The trained model showed reasonable accuracy in predicting future failure events. Moreover, the technique is shown to have comparatively higher efficiency even with reduced number of neurons in each layer of ANN. This significantly decreases computation time with minimum effect on the accuracy of results.

Using RBF technique significantly decreases the computational time with minimum effect on the accuracy of results.