ISSN 2304-7712 (Print) VOLUME 3 NU MBER 1 May 2014 ISSN 2304-7720 (Online) International Journal of Advanced Engineering and Science International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 International Journal of Advanced Engineering and Science ABOUT JOURNAL The International Journal of Advanced Engineering and Science ( Int. j. adv. eng. sci. / IJAES ) was first published in 2012, and is published semi-annually (May and November). IJAES is indexed and abstracted in: ProQuest, Electronic Journals Library, getCITED, ResearchBib, IndexCopernicus, Open J-Gate and JournalSeek. Since 2013, the IJAES has been included into the ProQuest one of the leading full-text databases around the world. The International Journal of Advanced Engineering and Science is an open access peer-reviewed international journal for scientists and engineers involved in research to publish high quality and refereed papers. Papers reporting original research or extended versions of already published conference/journal papers are all welcome. Papers for publication are selected through peer review to ensure originality, relevance, and readability. i ISSN 2304-7712 International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 International Journal of Advanced Engineering and Science CONTENTS 1 Publisher, Editor in Chief, Managing Editor and Editorial Board 2 FPGA-BASED IMPLEMENTATION OF ANN FOR DIRECT TORQUE CONTROL OF INDUCTION MACHINE USING CO-SIMULATION Djalal Eddine KHODJA, Stéphane SIMARD, Rachid BEGUENAN, Aissa KHELDOUN 3 USING INNOVATION DESIGN OF EVALUATIVE CRITERIA SOFTWARE FOR THE BEST SELECTION IN PRODUCT MARKET Shih-Chung Liao 4 CREATING OPTIMAL PRODUCT DESIGN OF EDUCATIONAL MANAGEMENT FOR SOCIAL AND ECONOMICAL DEVELOPMENT Shih-Chung Liao 5 INTRUDER DETECTOR Jonnathan M. Acuzar, Paulo Angelo T. Chan, Gerald Joe M. Datingaling, Philip Vonn P. Malaluan, Princes P. Tumambing, Jake M. Laguador 6 FEASIBILITY STUDY OF COSCIFERA METHYL ESTER Jaquemay D. Malabanan, Keycelyn L. Perlas, Carissa Mae B. Ramirez, Nemy H. Chavez 7 IMPROVED ANY COLONY SEARCH ALGORITHM FOR SOLVING OPTIMAL REACTIVE POWER FOR OPTIMIZATION PROBLEM K. Lenin, Dr.B.Ravindranath Reddy, Dr.M.Surya Kalavathi 8 CREATE AN AUTOMATED QUERY GENERATION BY USING PARSE TREE QUERY LANGUAGE AS A SOLUTION FOR INCREMENTAL INFORMATION EXTRACTION Gayatri Naik, Prf. Amole Pande , Harish Bhabad, Surekha Naik 9 MULTIPLE CHOICE PARTICLE SWARM OPTIMIZATION ALGORITHM FOR SOLVING OPTIMAL REACTIVE POWER DISPATCH PROBLEM K. Lenin, Dr.B.Ravindranath Reddy, Dr.M.Surya Kalavathi 10 DYNAMIC PERFORMANCE OF HVDC LINK BASED 3-LEVEL VSC SUPPLYING A PASSIVE LOAD Djehaf M, Zidi S-A, Djilani Kobibi Y 11 MODELLING A UPFC FOR THE STUDY OF POWER SYSTEM STEADY STATE AND TRANSIENT CHARACTERISTICS Djilani Kobibi Y.I, Hadjeri Samir, Djehaf Mohamed ii ISSN 2304-7712 International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 International Journal of Advanced Engineering and Science Publisher: Elite Hall Publishing House Editor in Chief: Dr. Mohammad Mohsin (India) E-mail: [email protected] Editorial Board: Mr. K. Lenin, Dr. Jake M. Laguador Dr. T. Subramanyam Assistant Professor, Jawaharlal Nehru Professor, Engineering Department FACULTY, MS Quantitative Finance, Department technological university Kukatpally, India Lyceum of the Philippines University, Batangas of Statistics E-mail: [email protected] City, Philippines Pondicherry Central University, India E-mail: [email protected] Email: [email protected] Dr. G. Rajarajan, Miss Gayatri D. Naik. Mr. Rudrarup Gupta Professor in Physics, Centre for Research & Professor, Computer Engg Department, YTIET Academic Researcher, Kolkata, India Development College of Engg, Mumbai University, India E-mail: [email protected] Mahendra Engineering College, India Email: [email protected] Email: [email protected] Mr. Belay Zerga Mrs.Sukanya Roy Mr. Nachimani Charde MA in Land Resources Management, Addis Asst.Professor (BADM), Seth GDSB Patwari Department of Mechanical, Material and Ababa University, Ethiopia College, Rajasthan,India Manufacturing Engineering, The University of E-mail: [email protected] E-mail: [email protected] Nottingham Malaysia Campus E-mail: [email protected] Dr. Sudhansu Sekhar Panda Dr. G Dilli Babu Mr. Jimit R Patel Assistant Professor, Department of Mechanical Assistant Professor, Department of Mechanical Research Scholar, Department of Mathematics, Engineering Engineering, Sardar Patel University, India IIT Patna, India V R Siddhartha Engineering College, Andhra Email: [email protected] Email: [email protected] Pradesh, India Email: [email protected] Dr. Jumah E, Alalwani Assistant Professor, Department of Industrial Engineering, College of Engineering at Yanbu, Yanbu, Saudi Arabia Email: [email protected] Web: http://ijaes.elitehall.com/ ISSN 2304-7712 (Print) ISSN 2304-7720 (Online) iii ISSN 2304-7712 International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 FPGA-based Implementation of ANN for Direct Torque Control of Induction Machine Using Co-simulation Djalal Eddine KHODJA(1),(4) , (1)Faculty of technology, M’sila University, Algeria. [email protected] Tel/Fax: +213 355 519 24 Stéphane SIMARD(2), (2)Applied Sciences Department, UQAC, Canada, [email protected], Tel/Fax:+154185455012 Rachid BEGUENAN(3), (3)Depmt of Electrical, R.Military Cllge, Canada, [email protected], Tel/Fax:+1613544-8107 Aissa KHELDOUN(4) (4)Signals & Systems Lab, Boumerdes Univ, Algeria. [email protected], Tel/Fax:+213355 519 24 Abstract - The aim of this paper is to propose design and implementation of Artificial Neural Network (ANN) on a Field Programmable Gate Array (FPGA). This implementation aim is to contribute in the hardware integration solutions in the areas such as control of power system, where the Direct torque Control (DTC) of induction machine is employed. The specialized Simulink tools used and the design procedure are presented. The results obtained by co-simulation of the induction motor drive in Matlab/Simulink and the ANN on the FPGA are satisfactory and very promising. Keywords - ANN, FPGA, Xilinx, Sigmoid Function, DTC. Djalal Eddine KHODJA is with Msila university, woks on Implementation of Artificial Neural Networks on reconfigurable circuits used in Control and Diagnosis of Induction Machines. 1. Introduction Direct torque control (DTC) of induction motor drives offers high performance in terms of simplicity in control and fast electromagnetic torque response. With dominant characteristics, the direct torque controlled induction motor drive is alternative in industrial applications (Y. Kumsuwan et al 2008, M.S. Carmeli et al 2011). Although in these systems such variables as torque, flux moduls and flux sector are required, the resulting DTC structure is particularly simplistic and therefore, becomes its major advantage. The application of the technique of neural networks in machine control is simple and allowed the resolution of several problems related to controlling these systems. In this work, the conventional DTC 1 ISSN 2304-7712 International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 is controller based on neural networks, where the switch block is replaced with ANN Block. The fast implementation of the developed methods is necessary needs for the industry and the complex (smart) power systems. These methods may be investigated using several techniques that have different characteristics to solve the encountering problems (Y. Kumsuwan et al 2008, M.S. Carmeli et al 2011, Francesco Ricc et al 2003, S. Vaez-Zadeh et al 2007, Antoni Arias et al 2005). The most commonly used techniques are the artificial intelligence-based techniques such as Artificial Neural Networks (ANN) (Antoni Arias et al 2005, A.TISAN et al 2006, M.T. Tommiska 2007)] that they are easier to implement on electronic circuit board such as: Digital Signal Processing (DSP) chips, Application Specific Integrated Circuits (ASICs) or Field programmable gate array (FPGAs) (DJ.Khodja et al 2005, V.A.Tolovka et al 2001, K. Renovell et al 2001). The aim of the present work is to propose an ANN based algorithm to control the induction motor with the DTC technique, then after successful simulation, convert the Simulink model written using Xilinx blockset with the System generator into Hardware Co-simulation. The generated Block will run on the FPGA Vertex4 XC4VSX35 board and the remaining blocks run on the Matlab/Simulink environment. The paper is organized as follows: the next section, the suggested topology; training and the generalization stage of the ANN are investigated, this design must lead to less operating time and a minimum logic usage on the FPGA. Section III undertakes the process of using System generator to generate the corresponding VHDL and the co-simulation of the DTC of induction motor using Xilinx Vertex 4. The paper ends up with a conclusion. 2 Application of artificial neural networks for DTC Fig.1. Scheme of Direct Torque Control-based ANN technique. The application of the technique of neural network in machine control is simple and allowed the resolution of 2 ISSN 2304-7712 International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 several problems related to controlling these systems. In this work with DTC, it is easy to use this technique, where the conventional DTC is replaced by controller based on neural networks as is illustrated in Figure.1. Fig.2 Internal structure of a controller based on neural network. The proposed neural network is a multilayer network (3-6-3) with the architecture illustrated in fig.2. Each neuron is connected to all neurons of the next layer by connections whose weights are randomly chosen real numbers. We notice that w is the weight of the connection between neurons x and y. The xy following steps are necessary to obtain this ANN: . ANN topology. . ANN Learning stage . ANN validation Artificial Neural Network Topology As shown in fig.2, the proposed artificial network consists of three layers, namely: the input layer consists of three neurons, whose function is to transmit the input values that correspond to the input variables to the next layer called hidden layer. The hidden layer is characterized by six neurons with sigmoid shaped activation function. The output layer is composed of three neurons whose output is either 0 or 1. ANN Learning stage The second stage in designing the ANN is the learning process which requires a data base defining the ANN input-output mapping. This data base is mostly given under matrix form as to clarify the inputs and the desired outputs according to the switching table of DTC, see table.1. 3 ISSN 2304-7712 International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 Table.1 switching table of the conventional DTC In this application, the input matrix consists of three inputs (lines) corresponding to : • The first variable is the position of the flux in the reference frame related to the stator. • The second input variable is used the state variable error flux. • The third input variable, the state variable error of the couple is used. This input values are done with this matrix: a= [1 1 1; 1 1 0; 1 0 1; 1 0 0; 2 1 1; 2 1 0; 2 0 1; 2 0 0; 3 1 1; 3 1 0; 3 0 1; 3 0 0; 4 1 1; 4 1 0; 4 0 1; 4 0 0; 5 1 1; 5 1 0; 5 0 1; 5 0 0; 6 1 1; 6 1 0; 6 0 1; 6 0 0]; The output is represented by the pulses of the inverter switches that represent the values zero or one, there matrix called desired output is done by. d= [1 1 0; 0 1 0; 1 0 1; 0 0 1; 0 1 0; 0 1 1; 1 0 0; 1 0 1; 0 1 1; 0 0 1; 1 1 0; 1 0 0; 0 0 1; 1 0 1; 0 1 0; 1 1 0; 1 0 1; 1 0 0; 0 1 1; 0 1 0; 1 0 0; 1 1 0; 0 0 1; 0 1 1]; Using Graphical User Interface (GUI) of the Neural Network Toolboox in Matlab, it’s easy from a given input – output data to train the proposed ANN. The Levenberg – Marquardt backpropagation algorithm is used to train the proposed network topology. To measure the network performance, we have used the Mean Squared Error between the target and the output of the fitting network. The optimal values of the Neural Network weights have been obtained after 56 iterations with an error of 3.58058e-012, see Figure.3. The error is very small however the obtained Network must be checked using validation test. 4 ISSN 2304-7712 International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 Fig.3. Evaluation of the squared error based on the number of iterations of learning (using the method of backpropagation gradient). The design of ANN in Simulink is shown in the figure.4 Fig.4. Artificial neural network made by Simulink. 5 ISSN 2304-7712 International Journal of Advanced Engineering and Science, Vol. 3, No.1, 2014 ANN validation In contrast to the method used in Matlab, which consists to use the learning data base for training and the remaining for the validation test. The results given by the trained network and the behavior of the motor are depicted in figure.5. The test result shows the behavior of the structure of the DTC applied to the induction machine of 1.5kW inverter fed by a two-level voltage with two levels of correction for the couple and the stator flux. We can see clearly that there is a decrease on the electromagnetic torque ripple and we also note that the electromagnetic torque accurately follows its reference and it is its response time (0.4 sec) 140 131 120 130 129 100 Wref 128 Speed (rd/s) Vitesse(rad/s) 24680000 W Speed (rd/s) Vitesse(rad/s)111122224567 0 123 -20 122 0 0.5 1 1.5 2 2.5 1.4 1.6 1.8 2 2.2 Temps(s) Temps(s) Time (s) Time (s) 25 0.5 0.4 20 0.3 15 m) m) 0.2 m) e(N.10 Cem m) e(N. 0.1 e (N Coupl e (N Coupl 0 qu 5 Cr qu r r To To -0.1 0 -0.2 -5 -0.3 0 0.5 1 1.5 2 2.5 1 1.02 1.04 1.06 1.08 1.1 Temps(s) Temps(s) Time (s) Time (s) 1.4 1.5 1.2 1 1 b) 0.5 or flux (Wb) x statorique(We00..68 phis-bittaPhis-Beta 0 StatFlu0.4 -0.5 6 -1 0.2 ISSN 2304-7712 0 -1.5 0 0.5 1 1.5 2 2.5 -2 -1 0 1 2 Temps(s) phis-alpha