Abd Kadir Mahamad

Predicting remaining useful life of rotating machinery based artificial neural network

Accurate remaining useful life (RUL) prediction of machines is important for condition based maintenance (CBM) to improve the reliability and cost of maintenance. This paper proposes artificial neural network (ANN) as a method to improve accurate RUL prediction of bearing failure. For this purpose, ANN model uses time and fitted measurements Weibull hazard rates of root mean square (RMS) and kurtosis from its present and previous points as input. Meanwhile, the normalized life percentage is selected as output. By doing that, the noise of a degradation signal from a target bearing can be minimized and the accuracy of prognosis system can be improved. The ANN RUL prediction uses FeedForward Neural Network (FFNN) with Levenberg Marquardt of training algorithm. The results from the proposed method shows that better performance is achieved in order to predict bearing failure.

Improving Elman Network using genetic algorithm for bearing failure diagnosis of induction motor

A bearing is an important component in any rotating machinery especially in induction motors. Thus, timely detection and diagnosis of induction motor bearing (1MB) is crucial to prevent sudden damages. This paper proposes a method to utilize artificial neural network (ANN) by using genetic algorithm (GA) to identify 1MB fault diagnosis. In this case, GA is utilized to find the optimum weights and biases for Elman Network (EN), which is one of ANN families. The vibration signal used in this work is obtained from Case Western Reserve University website. During preprocessing stage, vibration signals are been converted from time domain into frequency domain through Fast Fourier Transform (FFT). Then, enveloping method is used to eliminate the high frequency components from vibration signal. Subsequently, a set of 16 features from vibration and preprocessed signal are extracted. In order to reduce the size of data, a distance evaluation technique is used as features selection. In the development of ANN fault diagnosis, both networks EN (without GA) and GAEN (utilized with GA) in which results are compared and conclusions are drawn.

Development of artificial neural network based fault diagnosis of induction motor dearing

The common component failure of induction motor is bearing. Thus, timely detection and diagnosis of induction motor bearing (IMB) is very crucial in order to prevent sudden damage. This paper proposes developing artificial neural network (ANN) model of IMB fault diagnosis by using Elman Network. The vibration signal obtained from Case Western Reserve University website are been used as input signal. During preprocessing stage, vibration signal have been converted from time domain into frequency domain through fast Fourier transform (FFT). Enveloping method was then, used to eliminate the high frequency components from vibration signal. Subsequently, a set of 16 features from time and frequency domain were extracted. Furthermore, the distance evaluation technique is used in features selection in order to select only informative features. In order to make the ANN model more flexible, the sensitivity analysis of IMB is introduced. Lastly, a graphical user interface (GUI) program is created as a tool for help users determines the situation of IMB conditions.

Implementation of Field Programmable Gate Array based Maximum Power Point Tracking controller of photovoltaic system

Photovoltaic (PV) cell is one of the electrical part to convert photo-light into electricity in order to generate the electrical power. This paper presents the implementation of Field Programmable Gate Array (FPGA) based Maximum Power Point Tracking (MPPT) controller in evaluating the Maximum Power Point (MPP) output voltage of PV system. Matlab Simulink and Quartus II VHDL software tools are used as simulator of this FPGA, while Altera DE2 board is used as a controller. Both simulator and hardware had shown the same results, which prove that the designed system has been successfully extracting the MPP. The system has been evaluated in sunny day and partially shaded conditions to analyze the respective outputs.

FPGA-in-the-Loop Simulation of Cardiac Excitation Modeling towards Real-Time Simulation

Cardiac excitation is a fundamental mechanism within the heart’s function. One way to understand this mechanism is by using numerical modeling techniques. However, an immense amount of computational time has been required in the simulation that generally involves a large number of parameters. In this paper a simulation study of Luo Rudy Phase I (LR-I) mathematical model by using MATLAB Simulink to solve ordinary differential equations (ODEs) using field programmable gate array (FPGA) towards a real-time simulation of cardiac excitation has been presented. The FPGA could be the best solutions because it is able to provide high performance in solving higher order ODEs for real-time hardware implementation. In fact, the FPGA hardware design can be accelerated by using MATLAB Simulink HDL Coder that automates the hardware description language (HDL) code generation from designed MATLAB Simulink blocks. Furthermore, HDL designed implementation can be verified by using HDL Verifier such as co-simulation and FPGA-in-the-Loop (FIL) approaches to simulate the generated HDL code and verify the results. In this paper, results show that the LR-I cardiac excitation modeling is successfully simulated by the MATLAB Simulink and by using the HDL Coder the designed MATLAB Simulink model is successfully converted into VHDL code and verified through the FIL. These have given a positive outlook towards the FPGA hardware implementation for real-time simulation.

Comparison between GLCM and modified Zernike moments for material surfaces identification from photo images

Types of materials are one of an important data for research in acoustic engineering. This paper compares methods for extracting texture data of material surfaces for classification. Gray Level Co-occurrence Matrix (GLCM) and modified Zernike moments that is applied for image extraction are tested and compared with back propagation neural network used for classification. These methods are also applied to the Brodatz texture database as a general comparison. The GLCM method shows a good performance and regression, R>0.9 for the Brodatz database while the collected surfaces datasets using GLCM and modified Zernike moments as well as the Brodatz datasets using modified Zernike moments method had only managed an acceptable performance and regression of R>0.8.

Fault classification performance of induction motor bearing using AI methods

This paper presents an approach of intelligent fault classification of induction motor bearing (IMB) using several artificial intelligent (AI) methods. The most common AI methods are FeedForward Neural Network (FFNN), Elman Network (EN), Radial Basis Function Network (RBFN) and Adaptive Neuro-Fuzzy Inference System (ANFIS). The data of IMB fault is obtained from Case Western Reserve University website in form of vibration signal. For further analysis these datas are converted from time domain into frequency domain through Fast Fourier Transform (FFT) in order to acquire more fault signs during pre-processing stage. Then, during features extraction stage, a set of 16 features from vibration and pre-processing signal are extracted. Subsequently, a distance evaluation technique is used as features selection, in order to select only salient features. Lastly, during fault classification several AI methods are examined, where results are compared and the optimum AI method is selected.

Cardiac excitation modeling: HDL coder optimization towards FPGA stand-alone implementation

The aim of this paper is to discuss the optimization of the hardware description language (HDL) design using fixed-point optimization and speed optimization through a pipelining method. This optimization is very crucial to achieve the best performance in terms of speed, area and power consumption of the generated HDL code before deploying the field programmable gate array (FPGA) stand-alone implementation. As computational mathematical modeling needs immense amounts of simulation time, FPGA could bring the solutions as it provides high performance, and able to perform real-time simulations and compute in parallel mode operation. In this study, in order to ease verification, prototyping, and implementation FPGA, rapid prototyping model-based design approach of HDL Coder from MathWorks has been used to automate HDL codes generation from a designed MATLAB Simulink blocks of Luo-Rudy Phase I (LR-I) model towards FPGA hardware-implemented for numerical solutions of ordinary differential equations (ODEs) responsible in generating the action potential (AP) waveform of mammalian cardiac ventricle cell. By using HDL Coder, the model is successfully converted into an optimal fixed-point VHDL design and the operating frequency is increased from 9.819 MHz to 23. 613MHz by pipelining optimization.

A Simplified Malaysian Vehicle Plate Number Recognition

This paper propose an automatic inspection system of alphabets and numbers to recognize Malaysian vehicles plate number based on digital image processing and Optical Character Recognition (OCR). An intelligent OCR Training Interface has been used as a library and the system has been developed using LabVIEW Software. This software then is used to test with different situation to ensure the proposed system can be applied for real implementation. Based on the results, the proposed system shows good performance for inspection and can recognize an alphabets and numbers of vehicle plate number. To sum up, the proposed system can recognize the alphabets and numbers of the Malaysian vehicles plate number for inspection.

Prediction of Classroom Reverberation Time using Neural Network

In this paper, an alternative method for predicting the reverberation time (RT) using neural network (NN) for classroom was designed and explored. Classroom models were created using Google SketchUp software. The NN applied training dataset from the classroom models with RT values that were computed from ODEON 12.10 software. The NN was conducted separately for 500Hz, 1000Hz, and 2000Hz as absorption coefficient that is one of the prominent input variable is frequency dependent. Mean squared error (MSE) and regression (R) values were obtained to examine the NN efficiency. Overall, the NN shows a good result with MSE < 0.005 and R > 0.9. The NN also managed to achieve a percentage of accuracy of 92.53% for 500Hz, 93.66% for 1000Hz, and 93.18% for 2000Hz and thus displays a good and efficient performance. Nevertheless, the optimum RT value is range between 0.75 – 0.9 seconds