TELK OMNIKA Indonesian Journal of Electrical Engineering V ol. 12, No . 5, Ma yl 2014, pp . 3697 3704 DOI: http://dx.doi.org/10.11591/telk omnika.v12.i5.4525 3697 Modelling of a Witricity System Using GSSA Method Lan Jian yu * , T ang Houjun , a nd Geng Xin Shanghai Jiao T ong Univ ersity/Depar tment of Electr ical Engineer ing NO . 800, RD . Dong Chuan, Shanghai, 200240, China, 008613524956457 * e-mail: jian yu lan@163.com Abstract With the r apid de v elopment of mobile app liances , wireless po w er tr ansf er technique has been a hot issue f or researchers . A resonant coupled po w er system called witr icity with high efficiency and the middle r ange tr ansf er distance is presented b y MIT . The main circuit of the witr icity system acts as a resonant con v er ter oper ating in high frequency . The con v er ter is a comple x time v ar iant and non-linear systems . So , it is difficult to obtain its accur ate mathematical m odels . In this paper , the gene r aliz ed state-space a v er aging method is applied to model this con v er ter . With appropr iate v alues f or the circuit par ameters , n umer ical results are compared with which in time domain model. The results sho w that theoretical analysis are w ell ag ree with the sim ulation, and b y proposed method the computational time is remar ka b ly reduced. K e yw or ds: wireless po w er , witr icity , gener aliz ed-state-space-a v er aging Cop yright c 2014 Institute of Ad v anced Engineering and Science . All rights reser v ed. 1. Intr oduction Without cords and plugs , electronic and electr ical products has been more fr iendly and saf er than tr aditi onal method in special o ccasion such as coal mine , under w ater and implanted de vices [1, 2, 3, 4, 5]. The wireless po w er tr ansf er system based on inductiv ely coupled po w er tr ansf er method has the bottlenec k of shor t tr ansf er distance . If the distance betw een energy tr ansmitter and receiv er is more than tens millimetres , the system efficiencies will drop shar ply . Ho w e v er , t he theor y of magnetic resonance coupling pro v es a method of wireless po w er tr ans- f er(WPT) in longer distances . Andre [6] presented his w or k about this method which called witr icity (wireless electr icity). A 100 W b ulb w as lit 1m apar t from the source b y magnetic resonance cou- pling method. Currently , consider ab le eff or ts ha v e been made to the de v elopment of this po w er tr ansf er method. It ha s been sho wn that the distance of wireless po w er tr ansf er is significantly increased b y placing inter mediate resonators betw een tr ansmitter and recei v e r . Though abo v e researches ha v e g reat contr ib utes to the tr ansf er distance and efficiency [7, 8, 9, 10, 11, 12], the studies about mathematical model of the witr icity system are seldom. The circuit of witr icity requires oper ating with a high frequency , nor mally more than 1 MHz, to obtain high tr ansf er effi- ciency in long distance . Usually , a resonant con v er ter with PWM s witches is emplo y ed to dr iv e the witr icity system. The po w er con v er ter model is f ast changing with the time in nature because of the s witching beha viour , so this resonant con v er ter is a time v ar ying, nonlinear and comple x system. Thus , it is difficult to acquire the e xact mathematical model of this con v er ter . There are tw o methods a v ailab le to model a circuit. One is based on the circuit topology , which can be sim ulated b y some soft pac kages lik e SIMULIK or PSIM. The shor tcoming of this method is that it requires v ast resources of computer and a long sim ulation time . The other method is to der iv e its mathematical model directly through mathematical method. The adv a ntage is that an y par ameter can be change to obser v e char acter istics of proposed system in diff erent condition. Based on this idea, the state space a v er aging method is de v eloped f or the analysis and design of PWM con v er ters [13]. Because this me thod considers only the DC components of v ar iab les , it is not v alid f or modelling of resonant con v er ters in which the A C comp onents are the main con- tr ib utions . In addition, this method is based on an assumption that the v ar iab les should be m uch slo w in time domain than s witching frequencies . Ho w e v er , in a resonant con v er ter , the s witching Receiv ed September 26, 2013; Re vised December 26, 2013; Accepted J an uar y 16, 2014 Evaluation Warning : The document was created with Spire.PDF for Python.
3698 ISSN: 2302-4046 frequencies are near to its nature resonant frequencies . After that, a modified state space a v er- aging approach is proposed, which is applied to model the quasi resonant con v er ters with high accur acy . But it till canno t descr ibe the resonant con v er ter in detail. Sanders [14] proposed a gener aliz ed state space a v er aging (GSSA ) method to model resonant con v er ters . Compared to other resonant con v er ter modelling techniques , this GSSA method represents v ar iab les both in a slo w time-v ar ying DC v ar iab les and f ast oscillator y A C v ar iab les . This method is par ticular ly suitab le to f ull resonant con v er ters . Car los[15] has successfully emplo y ed this method in mod- elling a full-b r idge rectifier . Besides , GSSA method w as applied to the har monics estimation of PWM con v er ters [16, 17, 18, 19]. In the application of resonant con v er ters , a contactless po w er supply system is modelled using GSSA method with a completer 9th order system [20]. And then, Xin[21] emplo y ed this method to der iv e the mathematical model of an inductiv ely coupled po w er tr ansf er system, and a rob ust controller has designed based on this model. In addition, a mathematical model to sim ulate tr anscutaneous energy tr ansmission systems b y GSSA method is presented[22], and through selectiv e modal analysis method, the system w as k eeping up to the first-order . Ho w e v er , the mathematical model of the witr icity system based on GSSA or other modelling method is seldom to be f ound up to date . In this paper , the GSSA method is emplo y ed to the modelling of a witr icity system and the steps are presented in detail. The compar ison of the GSSA method and circuit model sim ulation in time do main b y PSIM is analysis , which sho ws that GSSA has a sufficient precision to present the dynamic system of a resonant con v er ter . This paper is organiz ed as f ollo ws . In section 2 the GSSA method is re vie w ed. In section 3, the mathematical model is der iv ed b y GSSA method and n umer ical analysis is presented at last. The conclusions are giv en in section 4. 2. Over vie w of GSSA Method The GSSA method is based on the concept that an y w a v ef or m x ( t ) on a time inter v al [ t; t + T ] can be e xpressed in F our ier ser ies as , x ( t ) = 1 2 A 0 + 1 X n =1 ( a n cos ( ! nt ) + b n sin ( ! nt )) , (1) where t 2 [0 ; T ] , ! = 2 =T , and T is the per iod of a sliding windo w . It is mention to note that the v ar iab le x ( t ) is not necessar ily per iodic and can be imagined so b y summing the windo w repeats o v er all time domain. Thus , a F our ier ser ies e xists and its coefficients can be deter mined from, a n ( t ) = 2 T Z t + T t x ( t ) cos ( n! t ) dt , (2) and b n ( t ) = 2 T Z t + T t x ( t ) sin ( n! t ) dt . (3) It is clear that the a v er age v alue A 0 = 2 and the amplitude p a 2 n + b 2 n of each component change with time . In other w ord, the coefficient a n and b n can reflect the en v elope of the or iginal signal x ( t ) . F or resonant con v er ters , v ar iab les usually ha v e both quasi-sin usoidal and direct com- ponents , and if an ap propr iate sliding windo w is selected, the fundamental ter m of the v ar iab le in each sliding windo w will represented of the or iginal signal x ( t ) w ell. In gener al, the per iod T is chosen near the nature resonant per iod of the con v er ter under studied. The v alue of n represents the accur acy le v el of the signals and if n approaches infinity , the appro ximation error theoretically approaches z eros . 3. Mathematical Modelling of the Witricity System The witr icity system emplo ys a high frequency resonant con v er ter to gener at e the e xcita- tion of the po w er tr ansf er system. In this par t, the oper ational pr inciple of witr icity will be re vie w first, and then the circuit model is discussed. At last the mathematical model is der iv ed b y GSSA. TELK OMNIKA V ol. 12, No . 5, Ma yl 2014 : 3697 3704 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA ISSN: 2302-4046 3699 3.1. Operational Principles of Witricity System Witr icity is based on the near-field, strongly coupled magnetic resonance , and the funda- mental pr inciple is that resonant objects e xchange energy efficiently , while non-resonant objects inter act w eekly . Figure1 sho ws the schematic of a witr icity system using tw o magnetically coupled resonators , which includes the source coil, the de vice coil, the energy source , and the load []. S o u r c e L o a d S o u r c e   c oi l E n e r g y   f l o w D e v i c e   c o i l 1 2 2 1 k k   Figure 1. Basic components of witr icity system Obser v ed from figure 1, energy resonates betw een the source coils and the de vice coils through the electromagnetic fields , though there is a big gap betw een the source coils and the de vice coils which gener ates a lo w coupling coefficient betw een this tw o coils . This ph ysical phenomenon can be e xplained using the coupled mode the or y []. As seen from Figure 1, it can assume that this tw o coils are tw o resonators , and a s ( t ) and a d ( t ) represent the amplitude of t hese resonators . In which, a s ( t ) is the source resonator and a d ( t ) is the de vice resonator respectiv ely . The tw o resonators obe y the f ollo wing equations []. da s ( t ) dt = ( i! s s ) a s ( t ) + ik sd a d ( t ) (4) da d ( t ) dt = ( i! d d ) a d ( t ) + ik ds a s ( t ) (5) Where , ! s and ! d are the individual angular frequency of source coils and de vice coils , respec- tiv ely; k = k 12 = k 21 is the coupling coefficient betw een the source and de vice; s and d are the individual intr insic deca y r ates f or the source and de vice , respectiv ely . Define the coupling f actor as C F = ! k 2 p s d . (6) When this meets the f ollo wing conditions , C F 1 and ! s = ! d , the witr icity system can tr ansf er po w er efficiently from source coils and de vice coils[]. 3.2. Cir cuit Model In this paper , the source de vice emplo ys a half-br idge to e xcite the resonant con v er ter , and the ser ial-par allel compensation net is used which is sho wn in figure 2. D r i v e r M 1 C 2 C L R 1 L 2 L dc U Figure 2. Schematic of proposed half-br idge con v er ter In figure2, L 1 and L 2 denote the pr imar y coil self-inductances and the secondar y coil self-inductances , respectiv ely; C 1 and C 2 represent compensated capacitors on both sides; M is Title of man uscr ipt is shor t and clear , implies research results (First A uthor) Evaluation Warning : The document was created with Spire.PDF for Python.
3700 ISSN: 2302-4046 the m utual inductance betw een the pr imar y coil and secondar y coil and R L is the equiv alent A C resistor of loads . By means of m utual inductance theor y , the equiv alent circuit of this system is deduced and sho wn in Figure3. M 1 C 2 C L R 1 L 2 L r Z ( ) c u t d c + U 0 1 ( ) i t 2 ( ) i t   ! Figure 3. Mutual inductance equiv alent circuit of proposed con v er ter In figure3, Z r is the reflected impedance from the secon dar y side and u c ( t ) is the induced v oltage of secondar y side . Z r can be e xpressed b y equation (7). Z r = ! 2 M 2 Z 2 (7) In which, Z 2 is the par alleled impedance of secondar y side , which is e xpressed in equation (8). Z 2 = j ! L s + R L 1 + j ! C 2 R L (8) Substituting (8) into (7) the reflected resistance and reactance from the secondar y coil to the pr imar y is , respectiv ely , Re Z r = ! 2 M 2 R L R 2 L ( ! 2 C 2 L 2 1) 2 + ! 2 L 2 2 (9) and Im Z r = ! 3 M 2 [ C 2 R 2 L ( ! 2 C 2 L 2 1) + L 2 ] R 2 L ( ! 2 C 2 L 2 1) 2 + ! 2 L 2 2 (10) Then, the equiv alent impedance looking from the input side of the half-br idge in v er ter is , Z eq = j ! L 1 + 1 j ! C 1 + Z r (11) 3.3. Mathematical Models Using GSSA According to Kirchhoffs circuit la ws , the equiv alent circuit equations are as f ollo ws , 8 > > > > > > > > > > < > > > > > > > > > > : s ( t ) U dc = L 1 di 1 ( t ) dt + u c 1 ( t ) + i 1 Z r i 1 ( t ) = C 1 du c 1 ( t ) dt j ! M i 1 ( t ) = L 2 di 2 ( t ) dt + u 2 ( t ) i 2 ( t ) = C 2 du 2 ( t ) dt u c 2 ( t ) R , (12) where i 1 ( t ) and i 2 ( t ) represent the resonant current of pr imar y side and secondar y side respec- tiv ely , and u 1 ( t ) denotes the v oltag e of pr imar y capacitor while u 2 ( t ) is the v olta ge of the sec- ondar y capacitor . Besides , s ( t ) is the s witch function, which can be e xpressed as , s ( t ) = ( 1 nT t < (2 n + 1) T = 2 0(2 n + 1) T = 2 t < ( n + 1) T . (1 3) TELK OMNIKA V ol. 12, No . 5, Ma yl 2014 : 3697 3704 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA ISSN: 2302-4046 3701 Because the circuit model of the witr icity system is a resonant con v er ter which can filter out har monic components , the dynamic system associated with a z ero and first order F our ier coefficient e xpressions can pro vide the enou gh accur acy . So , b y using GSSA method, the state v ar iab les can be e xpressed b y the z ero and first order components and the function is as f ollo ws , 8 > > > > > > > > > > > > > < > > > > > > > > > > > > > : h i 1 i 1 = x 1 + j x 2 h u c 1 i 1 = x 3 + j x 4 h i 2 i 1 = x 5 + j x 6 h u c 2 i 1 = x 7 + j x 8 h i 1 i 0 = x 9 h u c 1 i 0 = x 10 h i 2 i 0 = x 11 h u c 2 i 0 = x 12 . (14) Then, from equation 11, the diff erential equations in state-space ha v e the f ollo wing matr ix f or m, d ~ x dt = A ~ x + U dc . (15) Where , ~ x representing the z ero-order and first-order state v ar iab les , is defined as a column v ector , ~ x = x 1 x 2 x 3 x 4 ::: x 9 x 10 x 11 x 12 T . (16) In addition, the matr ix A can be divided into f our which is e xpressed as , A = A 1 0 0 A 2 , (17) where A 1 and A 2 are sho wn as f ollo ws , A 1 = 2 6 6 6 6 6 6 6 6 6 6 6 4 Z r L 1 1 L 1 0 0 1 C 1 0 0 0 j ! M L 2 0 0 1 L 2 0 0 1 C 2 1 R L C 2 3 7 7 7 7 7 7 7 7 7 7 7 5 (18) Title of man uscr ipt is shor t and clear , implies research results (First A uthor) Evaluation Warning : The document was created with Spire.PDF for Python.
3702 ISSN: 2302-4046 and A 2 = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 Re ( Z r ) L 1 ! + Im Z r L 1 1 L 1 0 0 0 0 0 ! Im Z r L 1 Re Z r L 1 0 1 L 1 0 0 0 0 1 C 1 0 0 ! 0 0 0 0 0 1 C 1 ! 0 0 0 0 0 0 ! M L 2 0 0 0 ! 1 L 2 0 ! M L 2 0 0 0 ! 0 0 1 L 2 0 0 0 0 1 C 2 0 1 R L C 2 ! 0 0 0 0 0 1 C 2 ! 1 R L C 2 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 (19) In addition, the coefficient b is e xpressed as , b = 0 0 0 0 0 2 L 1 ::: 0 0 . (20) T o obtain the n umer ical solution, t he contin uous linear model should be tr ansf er to the discrete mode first. Through the equation (15) and (16), the model is coded in MA TLAB , and the results are compared with the sim ulation results in time domain b y PSIM softw are pac kage . x ( k + 1) = x ( k ) + u ( k ) (21) [ ; ] = f ( A; B ; T s ) (22) where , k = 0 ; 1 ; 2 :::n: The par ameter v alues is sho wn in tab le 1, and the sim ulat ion results are sho wn in figure 4 and figure 5. Figure 4 sho ws the w a v ef or ms of resonant current of pr imar y side , in which the red line represents the solution from GSSA while the b lue lines is the results from the PISM pac kage . T ab le 1. The P erf or mance of the witr icity system V ar iab le Result Unit L 1 18 H C 1 800 pF L 2 18 H C 2 800 pF M 5 H R L 12 Compar ing cur v es in figure 4, the resonant current cur v es based on GSSA method is v er y similar to the sim ulation result from PSIM pac kage . And figure 5 sho ws the errors of these tw o methods . F rom which w e can see that the errors is alw a ys under 0.2 A. thus , it is clear ly that the first appro ximation is sufficiently precise . TELK OMNIKA V ol. 12, No . 5, Ma yl 2014 : 3697 3704 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA ISSN: 2302-4046 3703 0 2 2   1 1 0 4 2 6 T i m e (   s ) C u r r e n t   ( A ) Figure 4. Resonant current of pr imar y side using GSSA and sim ulation method 0 0 . 5 0 . 5 0 . 2 0 . 2 0 1 0 5 1 5 T i m e (   s ) C u r r e n t   ( A ) Figure 5. Errors betw een the GSSA method and the sim ulation method 4. Conc lusion In this paper , a tutor ial mater ial of a method f or modelling the witr icity system is presented. The GSSA me thod has been re vie w ed first. Then the circuit model of the witr icity is introduced. Based on which, the mathematical model is der iv ed using GSSA method, and the steps are pre- sented in detail. The results of GSSA are compared with the sim ulation from PSIM pac kage in time domain, which sho ws that there is little error betw een each other . The n umer ical analyses sho w that a first order appro ximation of GSSA is enough to present the precision of a witr icity sys- tem, and there is no need to include higher-order ter ms in the analysis . The models der iv ed from the GSSA method are suitab le f or dynamic analysis of the witr icity system, and its computational time has been consider ab ly reduced compared to other time domain sim ulation method. Ac kno wledg ement This w or k is suppor ted b y the National Natur al Science Fund of China (51277120) and the ITER project(2011GB113005). The authors also thanks to the help b y State Energy Smar t Gr id R&D Center(Shanghai). Ref erences [1] J . T . Bo ys , G. A. Co vic , and X. Y ongxiang, ”DC analysis technique f or inductiv e po w er tr ansf er pic k-ups , IEEE P o w er Electronics Letters , v ol. 1, pp . 51-53, 2003. [2] W . Guo xing, L. W entai, M. Siv apr akasam, and G. A. K endir , ”Design and analysis of an adap- tiv e tr anscutaneous po w er telemetr y f or biomedical implants , IEEE T r ansactions on Circuits and Systems I: Regular P apers , v ol. 52, pp . 2109-2117, 2005. [3] A. K. RamRakh y ani, S . Mir ab basi, and C . Mu, ”Design and Optimization of Resonance-Based Efficient Wireless P o w er Deliv er y Systems f or Biomedical Implants , IEEE T r ansactions on Biomedical Circuits and Systems , v ol. 5, pp . 48-63, 2011. [4] L. Zou and T . Larsen, ”Dynamic po w er control circuit f or implantab le biomedical de vices , IET Circuits , De vices & Systems , v ol. 5, pp . 297-302, 2011. Title of man uscr ipt is shor t and clear , implies research results (First A uthor) Evaluation Warning : The document was created with Spire.PDF for Python.
3704 ISSN: 2302-4046 [5] J . L. Villa, J . Sallan , J . F . Sanz Osor io , and A. Llombar t, ”High-Misalignment T oler ant Compen- sation T opology F or ICPT Systems , IEEE T r ansactions on Industr ial Electronics , v ol. 59, pp . 945-951, 2012. [6] A. K urs , A. Kar alis , R. Moff att, J . D . Joannopoulos , P . Fisher , and M. Soljacic , ”Wireless po w er tr ansf er via strongly coupled magnetic resonances , Science , v ol. 317, pp . 83-6, J ul 6 2007. [7] N. Shinohar a, ”The wireless po w er tr ansmission: inductiv e coupling, r adio w a v e , and reso- nance coupling, Wile y Interdisciplinar y Re vie ws: Energy and En vironment , v ol. 1, pp . 337- 346, 2012. [8] N. Y . Kim, K. Y . Kim, and C .-W . Kim, ”A utomated frequency tr ac king system f or efficient mid- r ange magnetic resonance wireless po w er tr ansf er , Micro w a v e and Optical T echnology Let- ters , v ol. 54, pp . 1423-1426, 2012. [9] S . Hui, W . Zhong, and C . Lee , ”A Cr itical Re vie w of Recent Prog ress in Mid-Range Wireless P o w er T r ansf er , IEEE T r ansactions on P o w er Electronics , v ol. PP , pp . 1-1, 2013. [10] C . Linhui, L. Shuo , Z. Y ong Chun, and C . Tie J un, ”An Optimizab le Circuit Str ucture f or High- Efficiency Wireless P o w er T r ansf er , IEEE T r ansactions on Industr ial Electronics , v ol. 60, pp . 339-349, 2013. [11] D . J . Thr ima withana, U . K. Mada w ala, and M. Neath, ”A Synchronization T echnique f or Bidi- rectional IPT Systems , IEEE T r ansactions on Industr ial Electronics , v ol. 60, pp . 301-309, 2013. [12] K. W u, D . Choudhur y , and H. Matsumoto , ”Wireless P o w er T r ansmission, T echnology , and Applications [Scanning the Issue], Proceedings of the IEEE , v ol. 101, pp . 1271-1275, 2013. [13] K. J . Kar imi, A. Book er , and A. Mong, ”Modelin g, sim ulation, and v er ification of large DC po w er electronics systems , in P o w er Electronics Specialists Conf erence , 1996. PESC ’96 Record., 27th Ann ual IEEE , 1996, pp . 1731-1737 v ol.2. [14] S . R. Sanders , J . M. No w orolski, X. Z. Liu, and G. C . V erghese , ”Gener aliz ed a v er aging method f or po w er con v ersion circuits , in P o w er Electronics Specialists Conf erence , 1990. PESC ’90 Record., 21st Ann ual IEEE , 1990, pp . 333-340. [15] C . Ga vir ia, E. F ossas , Gr i, x00F , and R. o , ”Rob ust controller f or a full-br idge rectifier using the ID A approach and GSSA modeling, IEEE T r ansactions on Circuits and Systems I: Regular P apers , v ol. 52, pp . 609-616, 2005. [16] X. Jianping and C . Q. Lee , ”A unified a v er aging technique f or the modeling of quasi-resonant con v er ters , IEEE T r ansactions on P o w er Electronics , v ol. 13, pp . 556-563, 1998. [17] A. L. Shenkman, B . Ax elrod, and V . Chudno vsky , ”A ne w simplified model of the dynamics of the current-f ed par allel resonant in v er ter , IEEE T r ansactions on Industr ial Electronics , v ol. 47, pp . 282-286, 2000. [18] F . Musa vi, K. Al-Haddad, and H. Y . Kanaan, ”A large signal a v er aged modelling and control of par alleled DC/DC con v er ters with automatic load shar ing, in Applied P o w er Electronics Conf erence and Exposition , 2005. APEC 2005. T w entieth Ann ual IEEE, 2005, pp . 1353-1358 V ol. 2. [19] A. Rameshkumar and S . Ar um ugam, ”Design of fuzzy controller f or z ero current s witching quasi resoant b uc k con v er ter , in Inf or mation and Comm unication T echnology in Electr ical Sciences (ICTES 2007) , 2007. ICTES . IET -UK Inter national Conf erence on, 2007, pp . 245- 248. [20] A. P . Hu, ”Modeling a contactless po w er supply using GSSA method, in Industr ial T echnol- ogy , 2009. ICIT 2009. IEEE Inter national Conf erence on , 2009, pp . 1-6. [21] D . Xin, T . Chengxuan, T . Chunsen, A. P . Hu, and A. Abdolkhani, ”Study on rob ust controller design f or inductiv ely coupled po w er tr ansf er system, in IECON 2011 - 37th Ann ual Conf er- ence on IEEE Industr ial Electronics Society , 2011, pp . 1614-1619. [22] W an F ang, W ei Liu, Jie Qian, Houjun T ang, Modeling and sim ulation of a tr anscutaneous [1 energy tr ansmission system used in ar tificial organ implants ”, Ar tificial Organs , v ol. 32, no . 12, pp . 1069-1074, 2009. TELK OMNIKA V ol. 12, No . 5, Ma yl 2014 : 3697 3704 Evaluation Warning : The document was created with Spire.PDF for Python.