Yousef Al-Assaf

Fatigue life prediction of unidirectional glass fiber/epoxy composite laminae using neural networks

Abstract Fatigue behavior of unidirectional glass fiber/epoxy composite laminae under tension–tension and tension–compression loading is predicted using artificial neural networks (ANN). Stress-life experimental data were obtained for fiber orientation angles of 0°, 19°, 45°, 71° and 90°. These tests were performed under stress ratios of 0.5, 0 and −1. The feedforward network used, provided accurate modeling between the input parameters (maximum stress, R-ratio, fiber orientation angle) and the number of cycles to failure. Although a small number of experimental data points were used for training the neural network, the results obtained are comparable to other current fatigue life-prediction methods.

Recognition of control chart patterns using multi-resolution wavelets analysis and neural networks

Control charts pattern recognition is one of the most important tools in statistical process control to identify process problems. Unnatural patterns exhibited by such charts can be associated with certain assignable causes affecting the process. In this paper, multi-resolution wavelets analysis (MRWA) is used to extract distinct features for unnatural patterns by providing distinct time-frequency coefficients. A reduced set of parameters is derived from these coefficients and used as input to an artificial neural network (ANN) classifier. Results show that the performance of the proposed technique in classifying shift, trend and cyclic patterns is superior to that of ANN classifier, which operated on coded observed data.

Features extraction and analysis for classifying causable patterns in control charts

Obtaining adequate features is a critical step in classifying causable patterns in control charts. Various methods were developed to extract features that maximize the inter-class variability while minimizing the intra-class variations. Most of these methods are based on either time or frequency domain analysis. As a multiresolution analysis approach, wavelet transform was considered to exploit the joint time-frequency characteristics of the patterns. However, the effectiveness of the features obtained by multi-resolution wavelet analysis (MRWA) suffers from the frequency leakage among the different spectral bands. This phenomenon is inherent in wavelet analysis regardless of the choice of the mother wavelet. Cross-band frequency leakage smears the band-specific information, which may yield less distinguishing features, especially for short-time observation patterns.In this work we introduce a multi-resolution analysis approach based on discrete cosine transform (DCT) that overcomes the problems associated with MRWA. We also verify that the classification rates of shift, trend, and cyclic causable patterns using multi-resolution DCT (MRDCT) features are higher than those obtained using MRWA features. Furthermore, the computational requirements for MRDCT are notably less than those needed for MRWA. Artificial neural network (ANN) classifier was used with both feature extraction methods.

Surface Myoelectric Signal Analysis: Dynamic Approaches for Change Detection and Classification

Toward the goal of elbow and wrist prostheses control by characterizing events in surface myoelectric signals, this paper presents a dynamic method to simultaneously detect and classify such events. Dynamic cumulative sum of local generalized likelihood ratios using wavelet decomposition of the myoelectric signal is used for on-line detection. Frequency as well as energy changes are detected with this hybrid approach. Classification is composed of using multiresolution wavelet analysis and autoregressive modeling to extract signal features while polynomial classifiers are used for pattern modeling and matching. The results of detecting and classifying four elbow and wrist movements show that, in average, 91% of the events are correctly detected and classified using features obtained from multiresolution wavelet analysis while 95% accuracy is achieved with AR modeling. The classification accuracy decreases, however, if short prostheses response delay is desired. This paper also shows that the performance of the polynomial classifiers is better than that of the commonly used neural networks since it gives higher classification accuracy and consistent classification outcomes. In comparison to the well known support vector machine classification, the polynomial classifier gives similar results without the need to optimize and search for classifier parameters

Energy-based fatigue life prediction of fiberglass/epoxy composites using modular neural networks

The strain energy has successfully been used in the past as a fatigue failure criterion for unidirectional fiber reinforced laminae. This approach has the ability to unify the macro- and microscopic behavior and can allow for extending the failure criterion to incorporate the multiaxial case. In this work, the strain energy will be used as an input to the artificial neural network (ANN) to predict fatigue failure. The results obtained will be compared to those obtained using the maximum applied stress, the fiber orientation angle and the stress ratio as inputs to the ANN.

SLIDING MODE CONTROL OF GENERALIZED FRACTIONAL CHAOTIC SYSTEMS

A sliding mode control technique is introduced for generalized fractional chaotic systems. These systems are governed by a set of fractional differential equations of incommensurate orders. The proposed design method relies on the fact that the stability region of a fractional system contains the stability region of its underlying integer-order model. A sliding mode controller designed for an equivalent integer-order chaotic system is used to stabilize all its corresponding fractional chaotic systems. The design technique is demonstrated using two generalized fractional chaotic models; a chaotic oscillator and the Chen system. The effect of the total fractional order is investigated with respect to the controller effort and the convergence rate of the system response to the origin. Numerical simulations validate the main results of this work.

Neuro-Fuzzy Modeling of Fatigue Life Prediction of Unidirectional Glass Fiber/Epoxy Composite Laminates:

Fatigue behavior of unidirectional glass fiber/epoxy composites under tension–tension and tension–compression loading is important in the design of composite structures. Adaptive neuro-fuzzy modeling was successfully used to model the relationship between the input/output variables of fatigue behavior of unidirectional glass fiber/epoxy composites. The experimental input variables were the maximum stress, fiber orientation, and stress ratio, while the output variable was the number of cycles to failure. In comparison with previous results obtained using neural networks only, the proposed hybrid neuro-fuzzy method gave more accurate fatigue life predictions.

EEG signal modeling using adaptive Markov process amplitude

In this paper, an adaptive Markov process amplitude algorithm is used to model and simulate electroencephalogram (EEG) signals. EEG signal modeling is used as a tool to identify pathophysiological EEG changes potentially useful in clinical diagnosis. The least mean square algorithm is adopted to continuously estimate the parameters of a first-order Markov process model. EEG signals recorded from rodent brains during injury and recovery following global cerebral ischemia are utilized as input signals to the model. The EEG was recorded in a controlled experimental brain injury model of hypoxic-ischemic cardiac arrest. The signals from the injured brain during various phases of injury and recovery were modeled. Results show that the adaptive model is accurate in simulating EEG signal variations following brain injury. The dynamics of the model coefficients successfully capture the presence of spiking and bursting in EEG.

Surface myoelectric signal classification for prostheses control

This paper represents an ongoing investigation for surface myoelectric signal segmentation and classification. The classical moving average technique augmented with principal components analysis and time-frency analysis were used for segmentation. Multiresolution wavelet analysis was adopted as an effective feature extraction technique while artificial neural networks were used for classification. Results of classifying four elbow and wrist movement signals recorded from biceps and triceps gave 5.1% classification error when two channels were used.

Wireless Stand-alone Portable Patient Monitoring and Logging System

In this paper, we propose a wireless stand-alone, embedded system design that integrates the monitoring of three biomedical parameters into a single personal medical device. The three parameters are: blood glucose level, heart rate, and pulse oximetry. The goal of this work is to build a compact and cost-effective device capable of monitoring several medical parameters while patients conduct their normal daily activities, and store these parameter readings in an embedded system-based portable device. A communication protocol and patient monitoring software application are developed to store data that can be later downloaded to a physician’s workstation for analysis and diagnosis.