0764) Volume 02 Issue 04, July 2013 Performance Analysis Of Hybrid PID .
International Journal of Computer and Information Technology (ISSN: 2279 – 0764)Volume 02– Issue 04, July 2013Performance Analysis of Hybrid PID-ANFIS forSpeed Control of Brushless DC Motor Base onIdentification Model SystemHidayatSasongko, P.H, Sarjiya, SuharyantoDepartment of Electrical and Information Technology,Universitas Gajah MadaDepartment of Electrical, Universitas Bung HattaPadang, Indonesiahdyttanjung@yahoo.comDepartment of Electrical and Information Technology,Universitas Gajah MadaYogyakarta, Indonesiasasongko@te.ugm.ac.idAbstract— This paper presents a performance of hybrid PIDANFIS for the speed control of Brushless Direct Current Motor(BLDCM). The model of BLDCM system is determined byidentification model system based on the measuring of dynamicresponse. The aim of the speed controller is to obtain the speedmotor operating similar to a speed setting. The controller input isa speed that measured by speed sensor encoder. The controlleroutput is the voltage to supply the stator windings of BLDCMthrough pulse wide modulation controller to drive inverter. Thecontroller has three input variables: speed error, actual speed,speed setting and output variable is the signal to determine thevoltage value. There are two controller structures to analyze theperformance of hybrid PID ANFIS controller. The first, outputof PID controller is as a main control and the output of ANFIScontroller is as a recovery control which switched base on speederror. The second, output of PID controller is added by theoutput of ANFIS controller. Using MATLAB Simulink, theperformance analysis is emphasized on the transient parameterof dynamic response. The simulation results show that the bestresponse of speed control is switched from PID to ANFIScontroller at the speed error less equal 10 %.Keywords- BLDCM, speed control, Hybrid PID-ANFISI.INTRODUCTIONBLDCM have been widely used as motor motion recently,because BLDCM has advantage than other motors, such asthe efficiency is more 13 % than induction motor, the volumeis less 40 % than conventional DC motor [1],[2]. The otheradvantages, caused no brush so they require little or nomaintenance, they generate less acoustic and electrical noisethan conventional DC motor, they can be used in hazardousoperation environments (with flammable products)[3].Therefore, many BLDCM had been applied in the industry(e.g. industrial drives ex pump, fans, blower, machine tolls,servo drives, automation process, internal transportationssystems, robots, etc), the public life (e.g. air conditioningsystems, catering equipment, coin laundry machines, autobank machine, etc), the domestic life (kitchen equipment e.g.refrigerators, microwave ovens, mixer, bathroom equipment,washing machines, toys, vision and sound equipment, securitysystems, etc), information and office equipment (e.g.www.ijcit.comcomputers, printers, plotters, scanners, etc), medical andhealthcare equipment (e.g. dentist’s drills, electricwheelchairs, trotters, rehabilitation equipment, artificial heartmotors) and etc [1]. To control the speed of BLDCM, AtmelCorporation have produced the BLDCM using ATmega32M1which applies a classic control. It is PID control. A qualitycontrol of BLDCM depends on the PID constant that is Kp,Ki, Kd. Tune Kp, Ki, Kd is done using the trial and errormethod [3]. Modeling of BLDCM is needed to design andanalyze the performance the speed controller which can beobtined easly by the identification parameter [4]In recent years, new artificial intelligent-based approacheshave been proposed for speed control of BLDCM fuzzy logiccontroller has been applied, which results still show anoscillation on steady state response. Furthermore, to decide onthe domain of membership function is more difficult to bedone [5],[6]. The speed response of BLDCM that uses toANFIS controller which apply the several functions andnumber of membership functions. The best response isobtained on the bell function and five membership functions.[7]. To improve the speed response of BLDCM, appliedparallel fuzzy PID controller which consists of three parallelfuzzy sub controllers that update online the values of theproportional, integral, and derivative gains. The controllerinput is the speed error and the delayed control signal thatrepresented to triangle functions [8].In this paper, a systematic approach for designing hybridPID-ANFIS is developed to find the best speed response ofBLDCM. The model of BLDCM base on identification modelsystem which is determined by dynamic response system. Toanalyse a performance of controller is done by several controlstructures i.e. PID controller, ANFIS controller, hybrid PIDANFIS. The hybrid PID-ANFIS is devided in to two structuresthat are a summing PID and ANFIS output controller and aselecting PID and ANFIS output controller.The control input of ANFIS controller are actual speedand speed setting that represented by the bell function. Theinput of PID controller is the speed error and PID constantdetermined by the close loop Ziegler Nichols method.1
International Journal of Computer and Information Technology (ISSN: 2279 – 0764)Volume 02– Issue 04, July 2013II.where,BLDCM MODELLINGBLDCM is constructed of Permanent Magnet SynchronousMachine (PMSM) 3 phase star connection, 2 poles, inverter 3phase voltage source inverter, rectifier, filter, rotor positionsensor, speed sensor and algorithm control [9],[10]. Theequipment of BLDCM is represented in Fig.1. The AC sourceis rectified to be source of inverter 1 phase, which derives thePMSM. As can be seen, rotor position is an input to thecontroller. Based upon rotor position and other input, thecontroller determines the switching states of each of theinverter semiconductors. The command signal to the controllermay be quite varied depending upon structure of the controlsin the system in which the drive is embedded. Other inputs tothe control the algorithms may include rotor speed, dc linkvoltage, and rectifier.K ci 1 / K c , Tci 1 / ciKp K K K , Kt 1 / Bt , Tt J / Btci b tB. BLDCM Modeling Base on Identification SystemTo model BLDCM base on estimation parameter uses anidentification system. It is a general term used to describemathematical tools and algorithms that build dynamicalmodels forms measured data. A dynamical mode in thiscontext is a mathematical description of the dynamic behaviorof a system or process that emphasized on the identification ofdiscrete-time transfer function from the measured input andoutput signal.A typical discrete-time transfer function is usually given by,as in [12].mm 1b z b z . bz b1m 1m z d(2)G( z ) 0nn 1a z a z . a z a12nn 1and it corresonds to the difference equation, as iny(t ) a1 y(t 1) a 2 y(t 2) . an y(t n) b1u(t d ) b2u(t d 1) . bmu(t d m 1) (t ) (3)Fig.1 The circuit element of BLDCM[9]A. BLDCM Modeling Base on Transfer Function of eachElementThis model is based on the each element of BLDCMwhich refers to Fig.1 that discussed and given in [11]. Blockschematic of speed control system of BLDCM that representsPI controller, driver (inverter 3 phase), speed sensor andPMSM is given by Fig.2.where (t ) can be regarded identification residual. Here theshorthand notation y(t) is used for output signal y(kT), andy(t-1) can be used to describe the output at the previoussample, i.e., y[(k-1)T].dan ai , (i 1,.n) , m n arebi , (i 1,.m 1)constant, z descrete variabel and dan d delay time.Suppose that a set of input and output signals has beenmeasurred and written, as inu [u(1), u(2),.,u(M )]T(4)y [ y(1), y(2),.,y(M )]TFrom equation (3), it can be found that, as iny(1) a1 y(0) . an y(1 n) b1u(1 d ) . bmu(2 m d ) (1)y(2) a1 y(1) . an y(2 n) b1u(2 d ) . bmu(3 m d ) (2). .y(M ) a1y(M 1) . an y(M n) b1u(M d ) . bmu(M 1 m d (M ) (5)where y(t) and u(t) assumed to be zero when t 0 . Thematrix form of (5) can written as iny Fig.2. a. Block schematic of speed control system of BLDCM [11]whereb. Simply block schematic of speed control system of BLDCMThe simply block schematic of speed control systemrepresents the open loop transfer function of BLDCM withoutcontroller represented, as in Kpm *(Tci s 1)(Tt s 1)(Tas s 1)Iaswww.ijcit.com(6)(1). y (1 n)u (1 d ) .u (2 m d ) y (0) y (1).y(2 n)u(2 d).u (3 m d ) (7) . . y(M 1).y(M n)u(M d).u(M 1 m d) T [ a1, a2 ,., an , b1,.,bm ](8)2
International Journal of Computer and Information Technology (ISSN: 2279 – 0764)Volume 02– Issue 04, July 2013 T [ (1),., (M )](9)Minimize the sum of squared residual as inM(10)min 2 (i) i 1The optimum estimation to the undetermined elements in can be written as in(11) [ T ] 1 T ySince the sum of squared residual is minimized, the method isalso known as the leas squares algorithm. The systemidentification is to identify the descrete-time model frommeasured input and output data in Matlab toolbox is providefunction arx(). If the measured input and output signals areexpressed by column vectors u and y the orders of thenumerator and denumerator as asssumed to be m-1 and n,respectively, and the delay tersm is d, the following statementcan be used, as inH arx([y,u],[n,m,d])(12)In this case, input is speed setting and output is actual speedthat can be obtained by measuring dinamic speed respons ofBLDCM.III.HYBRID PID-ANFIS CONTROLLER MODELLINGA. ANFIS PrincipleA typical architecture of an ANFIS which is used isSugeno fuzzy models consist of five layers that every layerhas the node. There are two kind of nodes that called theadaptive node (square symbol) and fixed node (circle symbol)as shown in Fig. 3. The mechanism of Sugeno has two inputsx1 and x2 and one output y. For a first-order Sugeno fuzzymodel [10],[13], a common rule set with two fuzzy if-thenrules is the folowing, as inIf x1 is A1 and x2 is B1 Then y1 c11.x1 c12.x2 c10(13)If x1 is A2 and x2 is B2 Then y2 c21.x1 c22.x2 c20(14)If predicate for two roles are w1 and w2, then can bedetermined the weight average, as iny w1 y1 w2 y2 w1 y1 w2 y2w1 w2(15)The function of every layer is:Layer 1Every node i in this layer is an adaptive node with a nodeactivation function parameter. The output of every node is themembership function degrees which given by inputmembership function, as in A1( x1), B1( x2 ), A2 ( x1) or B 2 ( x2 ) .O1,i Ai ( x1) ,O1,i Bi ( x2 ) ,for i 1, 2, or(16)for i 3, 4,If membership function is given by the generalized bellfunction, as inwww.ijcit.com ( x) 1(17)2bx c1 awhere {a,b,c} is the parameter set. As the value of theseparameters changes, the bell-shaped function variesaccordingly, thus exhibiting various forms of membershipfunctions for fuzzy set A. Parameters in these layers arereferred to as premise parameters.Fig. 3 The architecture ANFIS [10]Layer 2Every node in this layer is fixed node labelled , whoseoutput is the product of all the incoming signals, as in(18)O2,i wi Ai ( x1 ) x Bi ( x2 ) , for i 1, 2Each node output represents the firing strength ( predicate)of a rule. In general, any other T-norm that performs fuzzyAND can be used as the node function in this layer.Layer 3Every node in this layer is a fixed node labelled N. The ithnode calculates the ratio of the gain ratio ith rule firing strength( predicate) to the sum of all rules’ firing strengths, as inO3,i wi wiw1 w2, i 1,2(19)For convenience, outputs of this layer are called normalizedfiring strengths.Layer 4Every node i in this layer is an adaptive node with a nodefunction, as inO4,i wi yi wi (ci1 x1 ci 2 x2 ci 0 )i 1,2 (20)where w i is a normalized firing strength from layer 3 and{ ci1 , ci 2 , ci 0 } is the parameter set of this node. Parameters inthis layer are referred to as consequent parameters.Layer 5The single node in this layer is(50)a fixed node labeled , whichcomputes the overall output as the summation of all incomingsignals, as in i wi yiO5,i wi yi (21) i wii3
International Journal of Computer and Information Technology (ISSN: 2279 – 0764)Volume 02– Issue 04, July 2013The parameter to be trained are a, b and c of the premise andTo discretise the controller is given as inci1, ci 2 and ci 0 of the consequent parameters. ANFIS is trainedusing hybrid learning algorithm that consists of two steps suchas feed forward pass and backward pass. More specifically, inthe forward pass of the hybrid learning algorithm, nodeoutputs go to forward until layer 4 and consequent parametersare identified by the least squares method. In the backwardpass, the error signal propagates backward and the premiseparameters are updated by gradient descent.B. ANFIS ControllerThe ANFIS base control in this paper is the direct inversecontrol because of this is simply method and applicable. Thismethod [10] seems straightforward and only one learning taskis needed to find the inverse model of the plant, which is notvalid in general. Inverse learning or general learning forcontrol purpose is performed in two phases. In the learningphase, the plant ANFIS inverse model is obtained based oninput-output data generated from the former ANFIS model ofthe system as illustrated by Fig. 4. In the application phase, theobtained ANFIS inverse model in used to generated thecontrol action as illustrated by Fig. 5.dudedd 2e Kp K i e.dt K ddtdtdtdt 2duded de Kp K i e K d dtdtdt dt the time rate(25)(26) dis represented bythus:Tsdt U e e Kp Ki e K dTsTsTs T s (27) e U K p . e K i .e.Ts K d . T s the error rate thus: e en en 1(28)(29) (en en 1) (en en 1) (en 1 en 2 ) en 2en 1 en 2(30)The output rate: U U n U n 1(31)thus,U n U n 1 K p .(en en 1) K i .en .Ts K d . e e(32)n n 1TsSubstituting (30) into (32) are obtained as in Fig.4. Block diagram a training phase for inverse control methodThe ANFIS controller generates change in the referencevoltage base on error (e) and derivative error (de). Therefore,the input ANFIS controller is speed error (e) and rate speederror (de). The output controller (u(k)) is suitable signalmatching with the input current into stator windings topreserve BLDCM speed similar a setting. Every input andoutput variables are represented by membership functionfuzzy, which domain of membership function is determined bylearning process as shown above, as in(22)e ref r r (k) – y(k 1)de [d ( ref r )] / dt y(k) – y(k-1)(23) K(33)U n U n 1 K p .(en en 1) Ki .en .Ts d . e 2e enn 1 n 2Tswhere,: controller outputUn: erroren: time samplingTsD. Hybrid PID-ANFIS ControllerThere are two control structures of hybrid PID ANFIScontroller that applied respectively;1) Summing hybridThis structur, where the controller output adds the PID outputand the ANFIS output controller, as represented by (34).U HYBRID U PID U ANFIS(34)Flow chart of the summing hybrid structur is shown by Fig.6.Fig. 5 Block diagram an application phase for inverse control methodC. PID ControllerThe PID controller is defined by the following relationshipbetween the controller input (e) and the controller output (u)that is applied to motor armatur, as in [14][15].deu K p e K i edt K d(24)dtwww.ijcit.comFig.6. The flow chart of the summing hybrid controller output4
International Journal of Computer and Information Technology (ISSN: 2279 – 0764)Volume 02– Issue 04, July 20132) Selecting hybridThis structure, where the controller output selects the PIDoutput controller and the ANFIS output controller base onpersen error, as represented by (35). Three variabels areneeded to determine the selecting hybrid controller, such aserror, UPID and UANFIS. The persen error ( EError ) can bedetermined by trial and error. U,for Error E Error PID(35)U HYBRID for Error E Error U ANFIS , Flow chart of the selecting hybrid structur is shown by Fig.7.Fig. 8. Dynamic speed reponse of BLDCMTABLE I. THE SERIES DATA OF DYNAMIC SPEED RESPONSE OF BLDCMFig.7. The flow chart of the selecting hybrid controller outputIV. SIMULATION AND ANALYSISAccording to the model of ANFIS control systems forBLDCM mentioned above, some vital simulation works havebeen conducted. Motor model parameters used for simulationis such as:Model: ZW60BL120-430Voltage: 48 V (DC)Arus:7APower: 250 WSpeed: 3000 rpmNo.: 110325001Jenankeya Electron Science And Technology Co. LTDTo determine BLDCM model base on identification modelsystem is needed dynamic speed reponse of BLDCM thatshowed in Fig. 8. There are three variabels: time (t), actualspeed (y(t)), setting speed (u(t)). The series data of Fig.8 isshown in Table 1.By using MATLAB programming has been given in (12),further obtained the transfer function of BLDCM model, as in0,05136 z 0,07078TF (36)2z 0,0573z 0,03198Three data is needed to train ANFIS, such as two input dataand an output data. The input data are actual speed and delayactual speed. The output data is training target that representedby ramp function. It represents speed respon of BLDCM whencontroller input increased step by step. The data for training areacquired from the open loop experiment, as shown Fig.9. Forthe evenly distributed grid points of the time input range 3second with time sampling 0,001, maximum value 35, andminimum value 0 is obtained 2001 x 3 training data pairs.PID costants Kp, Ki and Kd are tunned using ZieglerNichols close loop method [14], furher obtained Kp 0,86, Ki 7,2, Kd 0,018www.ijcit.com5
International Journal of Computer and Information Technology (ISSN: 2279 – 0764)Volume 02– Issue 04, July 20130.05136z 0.07078-Kz2 -0.0573z-0.03198GainRamp1DiscretezT ransfer FcnUnit DelayScopekilatbaruT o WorkspaceFig. 9. Learning data of ANFIS controllerThe experiment was done by the bell function with 3membership function for each input and output variables. TheANFIS used here contains 9 (3x3 9) rules, 45 total number offitting parameters, including 18 (3x3 3x3) premise (nonlinier) parameters and 27 (3x9 27) consequent (linier)parameters. The training and root mean square (RMS) errorsobtained from the ANFIS are 0,00072442 for 30 epochs. Theoptimized membership function for input 1 and input 2 aftertrained is shown in Fig. 10.Selecting the PID and ANFIS output is determined by percenterror that done a trial and error method. The simulation circuitof BLDCM speed controller is showed in Fig.11. Thesimulation is done for a constant speed setting and changespeed setting, further determined and evaluated the transientparameter, such as steady state error (error), rise time (tr),delay time (td), over shoot (Mp) and settling time (ts). Thefirst experiment is given the speed setting 2000 rpm and 2500rpm, further the transient responses of all control structures areshown in Fig. 12 and Fig.13. The transient parameter ofBLDCM speed responses for speed setting 2000 rpm wereobtained similar to that for speed setting 2500 rpm, as shownin Table II.Speed respon of BLDCM4000Summing hybrid35003000PIDSpeed, Rpm2500Selecting hybrid20001500Open loop1000ANFIS500000,51,01,52,02,53,0Time, secFig. 12. Speed respon of all control structures for speed setting 2000 RpmFig. 10. The membership function for input 1 and input 2 after trained.4500Summing hybrid0.05136z 0.07078-K-4000z2 -0.0573z-0.03198Gain1Open LoopDiscreteT ransfer Fcn13500PID0.05136z 0.07078PID-Kz2 -0.0573z-0.03198PID ting hybridselfT o Workspace0.05136z 0.07078Mux-K-ANFISz2 -0.0573z-0.03198Gain3Mux2anfismats1Speed, RpmT ransfer Fcn225002000Open loopANFISDiscreteT ransfer Fcn31500StepMuxMux10.05136z 0.07078anfismats71000PID-ANFIS-K-Respon1z2 -0.0573z-0.03198PIDGain4hybridDiscreteDiscreteT o Workspace1T ransfer Fcn4PID Controller250000Mux0,51,01,52,02,53,0Mux30.05136z 0.07078anfismats2-K-Time, secsignalz2 -0.0573z-0.03198PIDDiscretePID ControllerSwitchGain5PID-ANFIS1T o Workspace2DiscreteT ransfer Fcn5Fig. 11. Simulation circuit of BLDCM speed controllerFig. 13. Speed respon of all control structures for speed setting 2500 RpmTABLE II.THE TRANSIENT PARAMETER OF BLDCM SPEED RESPONSEFOR SPEED SET 2000 AND 2500 RPM.There are three control structurs in this experiment, shuch asPID, ANFIS and hybrid PID ANFIS controller. The hybridPID ANFIS consists of a summing output PID and ANFIScontroller, a selecting output PID and ANFIS controller.www.ijcit.com6
International Journal of Computer and Information Technology (ISSN: 2279 – 0764)Volume 02– Issue 04, July 2013According to Table II data, it is shown which is theovershoot of summing hybrid PID ANFIS increased 75%, thefaster a rise time, the oscilation, the slower the settling time 1,4second. This is caused by each of the output signal controllerto strengthen, so needed a long time to settle.The transient response of selecting hybrid PID ANFIS for theerror less equal 10% was resulted the best response. It hascorresponded to what really designed of the selecting hybridPID ANFIS controller. It has ability to select the outputcontroller acurrately.The second experiment is given the speed setting 1000 rpm,then after 1,5 second it is increased to 2000 rpm whichresponses is shown in Fig. 14. The speed setting can befollowed by actual speed for all control structures although theovershoot of summing hybrid controller is higher than others. Itis caused by the output of summing hybrid controller is highertoo, which is shown in Fig. 15. The result of speed error issimilar to zero.Speed response of BLDCM30002500Selecting hybrid2000Speed, RpmSumming hybrid1500Desired speed1000500000,51,01,52,02,53,0Time, secFig. 14. Speed responses hybrid controller for speed setting changesfrom 1000 to 2000 rpm30002500Speed (Rpm) and control output (u(t))Selecting hybrid2000Summing hybrid15001000Speed error5000-50000,51,01,52,02,53,0Timen, secV.CONCLUSIONThis paper has described to control the BLDCM speed, thatcompared the several control structurs. The model of BLDCMwas constructed by identification system that represented ofthe real system that obtained to measure the dynamic responsesystem. The hybrid PID ANFIS controller which applied theselecting error was obtained the best response. It was toobtained when the error less equal than 10% the ANFIScontroller is done but the error great equal 10% the PIDcontroller is done. Other word, the PID controller is as maincontrol and the ANFIS is as recovery control. The hybrid PIDANFIS controller which applied the summing output PIDcontroller and ANFIS controller can be made the rise timeshorter but the overshoot increased.REFERENCES[1]. Gieras, J.F, Wing.M, Permanent Magnet Motor Technology Design &Aplication, 2nd Edition, Printed USA, London UK, 2002.[2]. Hidayat, Sasongko.PH, Srrjiya and Suharyanto,―The Speed and TorqueControl Strategy of the Brushless DC Machine‖, CITEE 2010,Yogyakarta,20 Juli 2010.[3]. Atmel Corporation, ―Brushless DC Motor Control Using ATMega32M1‖ April 2008[4]. Hidayat, Sasongko.PH, Sarjiya and Suharyanto,―Modeling the SpeedControl of Brushless DC Machine Base on Identification ModelSystem‖, Teknos. Padang,vol. 8 pp 60-67, February 2013.[5]. Munawar,I, Hidayat, ‖Design and Implementation Speed Control of the3 Phase Syncronouse Machine to Like Performance of DC MachineUsing Fuzzy Logic‖, SMED 2000, FT UGM, Yogyakarta, 2000.[6]. Hidayat, Munawar, I, ‖Design and Implementation the Speed Control ofBrushless DC Machine Using Fuzzy Logic‖, Thesis Magister Programof Electrical Departement ITB, Bandung, 2000[7]. Hidayat, PH.Sasongko, Sarjiya & Suharyanto, ― Performance Analysisof Adaptive Neuro Fuzzy Inference Systems (ANFIS) for Speed Controlof Brushless DC Motor, in Proc. ICEEI 2011,17-19 July 2011, p. 211.[8]. Rubai.A, Marcel J, Castro.S,Abdul R.O, ― Design and Implementationof Paralel Fuzzy PID Controller for High-Performance Brushless MotorDrives : An Integrated Enviromenent for Rapid Controln Ptototyping‖IEEE Transactions on Industry Applications, Vol. 44, No. 4, July/ Agus2008.[9]. Krause, Paul.C, ― Analysis of Electric Machinery And Drive Systems‖,IEEE Press, 2nd Edition, USA, 2002.[10]. Hidayat, PH.Sasongko, Sarjiya & Suharyanto, ―Modeling andSimulation of Adaptive Neuro Fuzzy Inference Systems (ANFIS) forSpeed Control of Brushless DC Motor‖, Proc. CITEE 2011, Yogyakarta,28 Juli 2011, paper E-4-1.[11]. Crnosija.P, Krishnan.R,‖Transient Performance Base Design Optimationof PM Brushless DC Motor Drive Speed Controller, IEEE ISIE 2005,Dubrovnik, Croatia, June 20-23, 2005, ,p.881.[12]. Xue.D, Chen.Y and Atherton.P.D, ―Linier Feedback Control Analysisand Design with MATLAB, Philadelphia, 2007.[13]. Jang.J-S.R, Sun.C.T, Mizutani.E, ― Neuro-Fuzzy and Soft Computing‖,Prentice-Hall International Inc, USA, 1997[14]. Brian, R ―The Design of PID Controllers Using Ziegler NicholsTuning‖, 2008[15]. Khuntia. S.R, Mohanty.KB, S. Panda and C. Ardil, ‖ A ComparativeStudy of P-I, I-P, Fuzzy and Neuro Fuzzy Controllers for Speed Controlof DC Motor Drive‖ International Journal of Electrical Systems Scienceand Engineering1,Januari 2009.Fig. 15. The control output of hybrid controller for speed setting changesfrom 1000 to 2000 rpmwww.ijcit.com7
International Journal of Computer and Information Technology(IJCIT)About IJCITThe International Journal of Computer and Information Technology (IJCIT) is anacademic journal for Asia and opens to the world. It aims to promote the integration of ComputerScience and Information Technology. It only publishes articles of the highest quality. IJCIT is ascholarly journal that provides a forum to the scholars and advanced level students forexchanging significant information and productive ideas associated with all IT disciplines.ISSN: 2078-5828 (Print), 2218-5224 (Online)Frequency: 2 Issues/Year (July and January)Subject Category: Computer Science and EngineeringPublished by: Research and Publication Unit (RPU)The IJCIT invites contribution in the following categories:1. Original research2. Survey/Review articles, providing a comprehensive review on a scientific topic
TopicsMajor areas of IJCIT include, but not limited to:AlgorithmsArtificial IntelligenceBio-informaticsBengali Language ProcessingCircuits and SystemsComputer VisionComputer Graphics and MultimediaComputer Based EducationComputer NetworksData CommunicationData MiningDatabase SystemsDigital Signal and Image ProcessingDigital System and Logic DesignDistributed and Parallel ProcessingE-commerce and E-governanceGraph Theory and ComputingHuman Computer InteractionInformation SystemsInternet and Web ApplicationsIntelligent SystemsKnowledge and Data EngineeringNeural Networks and Fuzzy LogicPattern RecognitionRoboticsSensorsSignal ProcessingSoftware EngineeringSystem Security and ControlUbiquitous/Mobile ComputingVLSIWireless Communication
Editor-in-ChiefEditorDr. Ahmed Y. Saber, SMIEEESenior EngineerResearch and DevelopmentOperation Technology, Inc., CA, USAE-mail: editor.chief@ijcit.org,aysaber@ieee.orgKazi Shamsul ArefinDepartment of CSEUniversity of Asia PacificBangladeshE-mail: editor@ijcit.org,arefin@uap-bd.eduEditorial PanelDr. Monirul Islam SharifResearcher, Google Inc.Georgia Institute of TechnologyUSADr. Emdad AhmedDepartment of CS, Wayne State UniversityUSADr. M. Julius HossainPostdoctoral Researcher, Dublin City UniversityIrelandDr. Quazi Ehsanul Kabir MamunMonash UniversityAustraliaProf. Dr. M. Lutfar RahmanDepartment of CSE, University of DhakaBangladeshProf. Dr. M. KaykobadDepartment of CSE, Bangladesh University of Engineering and Technology (BUET)BangladeshProf. Dr. M. A. MottalibHead of CIT, Islamic University of Technology (IUT)Bangladesh
Prof. Dr. M. Abdus SobhanDepartment of EECS, Independent University of Bangladesh (IUB)BangladeshDr. M. Fayyaz KhanHead of EEE, United International University (UIU)BangladeshProf. Dr. Hafiz Md. Hasan BabuDepartment of CSE, University of DhakaBangladeshDr. Shafiul AlomDepartment of CSE, University of DhakaBangladeshDr. Masud Al NoorDepartment of BA, University of Asia PacificBangladeshDr. Md. Mamun-ur-RashidDepartment of CSE, University of DhakaBangladeshProf. Dr. Suraiya PervinChairman of CSE, University of DhakaBangladeshProf. Dr. Md. Sekendar AliHead of EEE, University of Asia PacificBangladeshDr. Md. Sultan MahmudCoordinator of Inter Department, University of Asia PacificBangladeshProf. Dr. Subrata Kumar AdityaDepartment of APEC, University of DhakaBangladeshDr. Kazi Muheymin SakibInstitute of Information Technology (IIT), University of DhakaBangladesh
Volume 02- Issue 04, July 2013 www.ijcit.com 2 II. BLDCM MODELLING BLDCM is constructed of Permanent Magnet Synchronous Machine (PMSM) 3 phase star connection, 2 poles, inverter 3 phase voltage source inverter, rectifier, filter, rotor position sensor, speed sensor and algorithm control [9],[10]. . Volume 02- Issue 04, July 2013 , 2
International Journal of Computer and Information Technology (ISSN: 2279 – 0764) Volume 05 – Issue 01, January 2016 www.ijcit.com 139 Accounting Information Systems for A Distribution Company Alexander Setiawan Informatics Engineering Petra Christian University Surabaya, Indonesia Email: Alexander [AT] petra.ac.id Silvia Rostianingsih
Find the volume of each cone. Round the answer to nearest tenth. ( use 3.14 ) M 10) A conical ask has a diameter of 20 feet and a height of 18 feet. Find the volume of air it can occupy. Volume 1) Volume 2) Volume 3) Volume 4) Volume 5) Volume 6) Volume 7) Volume 8) Volume 9) Volume 44 in 51 in 24 ft 43 ft 40 ft 37 ft 27 .
Printable Math Worksheets @ www.mathworksheets4kids.com Find the volume of each triangular prism. 1) Volume 36 cm 25 cm 49 cm 2) Volume 3) Volume 4) Volume 5) Volume 6) Volume 7) Volume 8) Volume 9) Volume 27 ft 35 ft t 34 in 21 in 27 in 34 ft 17 ft 30 ft 20 cm m 53 cm 21
Printable Math Worksheets @ www.mathworksheets4kids.com 1) Volume 2) Volume 3) Volume 4) Volume 5) Volume 6) Volume 7) Volume 8) 9) Volume Find the exact volume of each prism. 10 mm 10 mm 13 mm 7 in 14 in 2 in 5 ft 5
Insurance For The strong Summer /strong Road Trip. Introducing The "At-Home Version" Of Insurance Key Issues. Click here for PDF Archives. Back Issues: strong Volume 2 /strong - strong Issue /strong 20 - October 30, 2013. strong Volume 2 /strong - strong Issue /strong 21 - November 13, 2013: strong Volume 2 /strong - strong Issue /strong 22 - November 27, 2013: strong Volume 2 /strong - strong Issue /strong 23 -
4.3.klinger volume oscillator 8 4.4.volume keltner channels 9 4.5.volume udr 9 4.6.volume tickspeed 10 4.7.volume zone oscillator 11 4.8.volume rise fall 11 4.9.wyckoffwave 12 4.10.volumegraph 13 4.11.volume sentiment long 14 4.12.volume sentiment short 15 5. beschreibung der cond
article objective is to provide a project management approach that enables managers to overcome this challenge. The method proposed herein applies to projects that are managed through Kanban, a popular project management framework for agile projects. Kanban is a visual system in which the various tasks composing a project are displayed as
ACCOUNTING 0452/12 Paper 1 October/November 2019 1 hour 45 minutes Candidates answer on the Question Paper. No Additional Materials are required. READ THESE INSTRUCTIONS FIRST Write your centre number, candidate number and name on all the work you hand in. Write in dark blue or black pen. You may use an HB pencil for any diagrams or graphs. Do not use staples, paper clips, glue or correction .
