Sunteți pe pagina 1din 8

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO.

9, SEPTEMBER 2009

3709

Analysis and Design of Maximum Power Point Tracking Scheme for Thermoelectric Battery Energy Storage System
Rae-Young Kim, Student Member, IEEE, Jih-Sheng Lai, Fellow, IEEE, Ben York, Student Member, IEEE, and Ahmed Koran, Student Member, IEEE

AbstractThe analysis and design of an adaptive maximum power point tracking (MPPT) scheme using incremental impedance are presented. A small-signal model is mathematically derived, and the impact of two major design parameters, which are scaling factor and sampling interval, is analyzed in the frequency domain. Four factors which specically affect the MPPT response are also clearly addressed. Based on this analysis, a design methodology to achieve a desirable transient response, while retaining system stability, is developed. The design methodology is implemented and veried with hardware experiments on a thermoelectric generator battery energy storage system, which indicate agreement between dynamic response and target bandwidth. Index TermsBattery energy storage system (BESS), frequencydomain design, maximum power point tracking (MPPT), thermoelectric generator (TEG).

I. I NTRODUCTION

thermoelectric module (TEM), which is a solid-state device that directly converts a temperature difference into electric power, has recently attracted increased attention as an independent, clean, and renewable energy source. In some areas where a considerable amount of energy is unused or wasted in the form of exhaust heat, a thermoelectric-based recovery system has led to reported improvements in overall system efciency [1][3]. However, the utilization of the thermoelectric generator (TEG) technology has been impeded by low energy-conversion efciency, inconsistent output power due to temperature uctuation, and high cost. To maximize energy utilization, maximum power point tracking (MPPT) is implemented to move the operating point of a TEG toward its optimal location. For MPPT, many schemes have been proposed in photovoltaic applications. The perturbation-and-observation scheme [4] and incremental conductance scheme [5], [6] have been widely used due to their simplicity and ease of implementation. However, these schemes often struggle with steady-state

Manuscript received August 20, 2008; revised December 16, 2008 and April 10, 2009. First published June 26, 2009; current version published August 12, 2009. This work was supported in part by Visteon Corporation and in part by BSST, LLC. An earlier version of this paper was presented at the 34th Annual Conference of the IEEE Industrial Electronics Society (IECON 2008), Orlando, FL, November 1013, 2008. The authors are with the Future Energy Electronics Center, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061-0111 USA (e-mail: rykim@vt.edu). Color versions of one or more of the gures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identier 10.1109/TIE.2009.2025717

oscillation and slow dynamic performance. Adaptive MPPT schemes, which employ a variable step size, overcome these difculties. In [7], the authors proposed an MPPT scheme where the step size was modied by integrating the sum of the instantaneous and incremental conductances using an error amplier and simple integral circuitry. In [8][12], the step size was adjusted according to the gradient of the measured power curve. Reference [13] proposed a direct duty cycle control approach with a variable step size. Fuzzy control and neural networks have also been reported for adaptive MPPT schemes [14], [15]. In the adaptive MPPT schemes, a scaling factor, which typically adjusts the input signal to a proper magnitude prior to determining the proceeding step size, is the essential parameter by which to judge dynamic performance and stability [11]. In digital implementations, the sampling interval of the scheme also signicantly impacts its performance [16]. Thus, the selection of these parameters requires proper analysis and theoretical rationale. Although many papers have been published on the subject of adaptive MPPT, the scaling factor and sampling interval were often given arbitrary values, and a research effort for an optimal choice has not been reported extensively. In [11], a simple guideline for selecting the scaling factor was proposed to avoid step size saturation. Reference [16] described the optimization of the sampling interval in order to reduce the impact of measurement error on system stability. This paper discusses the analysis and design of an MPPT scheme for a battery energy storage system (BESS). The scaling factor and sampling interval are theoretically analyzed to clarify their impacts on dynamic performance and stability, specically illustrating their effect on the dc gain of the MPPT scheme. The four factors which affect the MPPT response are also addressed. In order to verify these characteristics, the small-signal model of the MPPT scheme is derived, and the frequency response plots of the loop gain are evaluated, including the dynamics of the TEG and power converter. Based on that analysis, a design procedure is proposed in order to obtain the necessary stability margin while maximizing dynamic response. Experimental results are then provided to verify the performance of MPPT scheme and its design procedure. II. TEG C HARACTERISTICS The TEM is based on the Seeback effect, which states that an electromotive force is introduced between two semiconductors

0278-0046/$26.00 2009 IEEE

3710

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 9, SEPTEMBER 2009

Fig. 1. Electrical characteristics of TEG.

when a temperature difference exists [17]. The electromotive force etem is expressed as etem = (p n ) (Th Tc ) = P N T (1)
Fig. 2. TEG-based BESS with dcdc boost converter.
limit max , in order to protect the battery. The capacitors Cs and Co are selected to provide attenuation of the switching ripple. The A/D and D/A converters are used to provide quantization and restoration of analog signals. The MPPT scheme presented in [7] is used in this paper, while slightly adapted for a wide current range and digital implementation. As described in (5), the power curve of Fig. 1 shows that the derivative of pteg is positive before reaching the MPP, zero at the MPP, and negative after passing the MPP. From (4), the derivative of the pteg is given as (6), and then divided by the iteg . The resultant equation, designated as the error e, can be rewritten as

where Th and Tc are the temperatures of the hot and cold sides, and p and n are the Seeback coefcients. From the electrical circuit point of view, the terminal voltage of the TEM tem is simplied as tem = etem Rtem item (2)

where Rtem is the electrical resistance of TEM, and item is the electric current of TEM. Note that both etem and Rtem are highly dependent on temperature difference. A TEG consists of several TEMs, which are electrically connected in a seriesparallel arrangement. The terminal voltage teg and generated power pteg are expressed as teg = (Ns etem ) (Ns /Np Rtem ) (Np item ) = eteg Rteg iteg pteg = teg iteg = Ns Np tem item (3) (4)

dpteg /diteg

> 0, = 0, < 0,

before the MPP at the MPP after the MPP

(5)

dpteg d(teg iteg ) dteg = = iteg + teg diteg diteg diteg e= teg dteg + . diteg iteg

(6) (7)

where eteg , iteg , and Rteg are the electromotive force, terminal current, and total electrical resistance, respectively. Ns and Np are the number of TEMs in series and in parallel. Fig. 1 shows the actual measurements of TEG electrical characteristics employed in a BESS under three different temperature differences. The intersection of the dashed line with a given power curve species the maximum power point for that temperature difference. The MPPT scheme is designed to determine this maximum point and adjust the input current of the power converter accordingly. III. MPPT S CHEME FOR T HERMOELECTRIC BESS BESS is shown in Fig. 2, where Rs is a connection resistance. When the TEG current iteg is regulated, the battery current is also regulated. Thus, the charging process is controlled via the iteg . The MPPT scheme implemented by digital circuitry calculates an optimal input current i opt from the measured teg and iteg . The current compensator Ci generates the control voltage c based upon the error between the i opt and ii , which is a scaled version of the iteg . The battery voltage bat is monitored, and i opt is switched to zero when bat exceeds the

Tracking the MPP requires the following behavior: when e > 0 1) increasing i opt , when e < 0 2) decreasing i opt , 3) maintaining i , when e = 0. opt

(8)

Equation (8) is implemented by a simple integral compensator with e as an input. The scaling factor k adjusts e to a proper range before the integral compensator. Since e becomes smaller as the operating point approaches the MPP, an adaptive and smooth tracking can be achieved. Applying the forward Euler method, the discrete-time expression of the adaptive MPPT scheme is written as
i opt [n + 1] = iopt [n] + k e[n]

(9)

e[n] = =

teg [n] teg [n 1] teg [n] + iteg [n] iteg [n 1] iteg [n] teg [n] teg [n] + iteg [n] iteg [n] (10)

KIM et al.: ANALYSIS AND DESIGN OF POWER POINT TRACKING SCHEME FOR ENERGY STORAGE SYSTEM

3711

In Fig. 3, the relation between the i opt and e is derived in the z -domain, and the resulting equation is substituted into (15) Eteg k 2 iteg (z ). i opt (z ) = z 1 Iteg (16)

Finally, the continuous-time expression of Gpp (s) is derived by substituting z = esT s into (16), and then multiplying (1 esT s )/(s Ts ), in order to represent the ZOH effect, where Ts represents the sampling interval Gpp (s) =
Fig. 3. Block diagram of thermoelectric BESS with adaptive MPPT scheme.

i opt iteg

k Eteg 1 2 . Ts Iteg s esTs

(17)

where [n + 1], [n], and [n 1] represent the quantized values at the (n + 1)th , nth , and (n 1)th sampling intervals, respectively. IV. M ODELING AND A NALYSIS Fig. 3 shows the BESS control diagram using transfer function blocks: Gid duty-to-current transfer function, Fm pulse width modulation gain, Ci current compensator, and Hi current sensor gain. Equations (9) and (10) are illustrated in z -domain with a sampler and a zero-order hold (ZOH). Two feedback loops are shown in Fig. 3. The current loop Ti is the inner loop, and the MPPT scheme Tpp is the outer loop. It is clear that the dynamics and stability of the MPPT scheme depend on the inner loops dynamics. The loop gains of Ti and Tpp can be written as Ti (s) = Ci (s) Fm Gid (s) Hi Tpp (s) = Gpp (s) Ticl (s) = Gpp (s) 1 Ti (s) Hi 1 + Ti (s) (11)

Using small switching ripple approximation on the input current and output voltage [19] in continuous conduction mode, the duty-to-current transfer function Gid (s) is derived as a thirdorder system, which is provided in the Appendix. The battery is modeled as an electrochemical voltage Eb with internal resistance Rb [18]. Considering the strong damping effect of Rb and neglecting a high frequency pole and zero caused from the Cs , the Gid (s) is derived as (18)(20) with the low-Q approximation [19], where D is a duty cycle, D = (1 D), and Req = Rb RL /(Rb + RL ). As seen from (19), the large L and Co reduce the pg1 and pg2 , often resulting in slow dynamic response in a power converter Gid (s) gdc pg1 (1 + s/zg ) (1 + s/pg1 )(1 + s/pg2 ) pg2 1 Req Co (18)

D 2 Req L

(19)

zg = (12) gdc =

2Vteg Rb D 2 Eb Req Vteg Req Rb Co 2Vteg Req D Eb /Rb . D 3 Req (20)

where Gpp (s) represents the current-to-reference transfer function, and Ticl (s) is the closed-loop reference-to-output transfer function. Note that Tpp has the negative loop gain caused by a positive feedback. Deriving Gpp (s) requires the linearization of the error e around a quasi-steady-state point (Iteg , Eteg ) because of the nonlinearity of the impedance computation of (7). From (3), instantaneous impedance around the point (Iteg , Eteg ) is given as dteg /diteg = Rteg . (13)

Assuming that Ci (s) has two poles and one zero, where cdc is a dc gain, and wpc and wzc are the pole and zero, respectively, Ci (s) = cdc (1 + s/zc ) . s(1 + s/pc ) (21)

Adapting the Taylors series expansion to (7) and retaining only the linear term give e=2 Etem Rteg Item Eteg iteg . 2 Iteg (14)

If we design the zc around pg1 , and the pc nearby zg , the effects of poles and zeroes roughly cancel out. Therefore, the Ticl (s) is approximated as a second-order system. Assuming a small quality factor (usually true due to the desirability of the smooth response of an overdamped design), the Ticl (s) is factorized once again by using the low-Q approximation. Shown in (22) and (23), the dominant pole of pt1 is mainly determined by the current loop parameters, and pt2 depends on Co and Req Ticl (s) 1 1 Hi (1 + s/pt1 )(1 + s/pt2 ) pt2 1 . Req Co (22) (23)

teg Substituting perturbations e = E + e and iteg = Iteg + into (14), the linearized error is derived as e = Eteg iteg . 2 Iteg (15)

pt1 Fm Hi gdc cdc

3712

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 9, SEPTEMBER 2009

loop [16]. Assuming pt1 as the dominant pole of Ticl (s), it becomes pt1 (5 s ). (26)

V. D ESIGN M ETHODOLOGY Based on the aforementioned analysis, a design methodology is presented as follows. It enables requirements on the dynamic response to be met while preserving system stability. 1) The rst step is to select the proper quasi-steady-state point representing the worst case conditions of the design. Typically, at the maximum converter duty cycle, the dominant pole of Tpp (s), pg1 , reaches its lowest frequency, while the gain of Tpp (s) becomes the smallest at the minimum duty. The general rule is to use a maximum duty as the quasi-steady-state point, except in step 5). 2) The next step is to decide the crossover frequency c,target , which represents the dynamics of an MPPT scheme. It is roughly decided by a specied settling time (denoted as Tss ), as shown in (27), under a sufcient gain margin in Tpp (s) [20] c,target (2/Tss ). (27)

Fig. 4. Gain/phase asymptotes in the loop gain of adaptive MPPT scheme. (a) Without digital sampling effect. (b) With digital sampling effect.

Replacing (17) and (22) into (12), the loop gain of the MPPT scheme Tpp (s) is obtained as Tpp (s) 1 k Eteg . 2 s(1 + s/ Ts Hi Iteg )(1 + s/pt2 ) esTs pt1 (24)

Tpp (s) is a third-order system with a time delay caused by digital sampling. The scaling factor k and sampling interval Ts affect the dc gain. Decreasing the k , or increasing the Ts , results in a lower dc gain. This reduces the crossover frequency and generally improves system stability. The magnitude and phase asymptotes of the Tpp (s) are shown in Fig. 4, with and without the digital sampling effect. In Fig. 4(a), the phase plot begins at 90 and drops to 180 near pt1 . Consequently, the pt1 is the rst factor affecting the MPPT crossover frequency (denoted as c ), which indicates the speed of the MPPT dynamic response. Increasing cdc is desirable for a high pt1 , as long as a current loop is stable. Refer to (23). The second factor is the output capacitor Co . A large Co makes the pt2 move close to pt1 , which causes a reduced phase margin m accordingly. Thus, the risk of instability becomes greater as Co increases. To prevent instability, k and Ts are selected to secure a phase margin at the expense of dynamic response. The third factor is the digital sampling, as shown in Fig. 4(b). The additional phase drop starts at the vicinity of one decade lower than the sampling frequency s , where s = 2/Ts . For stability, the s should be selected as s (10 c ). (25)

3) Next is to set the target sampling frequency s,target . The low boundary is given by the relationship of (25), which is represented in (28). The high boundary is determined by noise-shaping methods. A good noise lter is preferred, but it limits the s,target , due to computational intensity. In this paper, a median lter of size N = 6 is employed. As a result, the s,target is limited to a few hundred hertz s,target (10 c,target ). (28)

4) The following step is to locate the dominant pole of Tpp (s)pt1,target . The low boundary condition is given by (26), and the high boundary is restricted by the power converter switching frequency sw . As shown in (23), pt1,target determines the closed-loop bandwidth of the current loop. For a stable current loop, the bandwidth is usually selected as veten times smaller than the switching frequency. This is the last factor, which becomes a major limitation in medium and high power applications because of their slow switching frequency (5 s,target ) pt1,target (1/10 1/5) sw . (29)

A lower boundary of pt1 is given to ensure the reduction of a sensing error caused from the transient behavior of the current

5) Next, the k of (24) is selected by using gain/phase plots, to match the c,target . Note that the minimum duty is used as the quasi-steady-state point for this step. 6) Finally, the Ci (s) is designed. The zc and pc are selected to cancel pg2 and zg given by (19) and (20). The cdc is then adjusted to get to pt1,target according (23). Using the circuit parameters shown in Table I, the MPPT scheme and current compensator are designed step by step. The specied settling time is 500 ms, i.e., Tss = 500 ms. Fig. 5

KIM et al.: ANALYSIS AND DESIGN OF POWER POINT TRACKING SCHEME FOR ENERGY STORAGE SYSTEM

3713

TABLE I E LECTRICAL PARAMETERS OF BESS

Fig. 6. Test setup for BESS. (a) Block diagram and (b) photograph of the experimental setup.

Fig. 5. Gain/phase plot of the loop gain in adaptive MPPT scheme. The solid line is the theoretical loop gain, and the dotted line is the approximated one. Under (a) maximum and (b) minimum duty cycles. TABLE II D ESIGN R ESULTS OF BESS Fig. 7. Inductor current iL , TEG current iteg , and battery voltage bat (5 s/div).

shows the resultant gain/phase plots of Tpp (s) under a maximum duty cycle of 0.756 and a minimum of 0.460. A theoretical Tpp (s) is shown by the solid lines, and an approximated one by the dotted lines. The design of c satises Tss = 500 ms. The design results are summarized in Table II. VI. E XPERIMENTAL R ESULTS To verify the MPPT dynamics and its design, a hardware setup using the circuit parameters in Table I was constructed. Fig. 6 shows the test setup and photograph. The TEG consists

of three TEMs of 36-W rating at 12 V, DC-CL021, stacked together. The two outer units control the temperature of the center unit. The center unit generates TEG electrical characteristics according to the loading condition and temperature difference. The interactive power amplier amplies the electrical characteristic and controls the loading condition with a variation of the converter load. A boost converter, which employs three interleaved phases to permit a large current and a low ripple, is implemented with a TMS320F2808 digital signal processor. Fig. 7 shows the three-phase inductor current iL , TEG current iteg , and battery voltage bat . A 60-A MOSFET, Inneon IPW60R045CP, was used. The gate driver detects the drainto-source voltage during turn-on and shuts down if the current becomes too large. A 12-V leadacid battery and a resistor bank are used as an electrical load. Fig. 8 shows the MPPT performance while the TEG power is increased, where iR and ibat are the current of the load bank

3714

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 9, SEPTEMBER 2009

Fig. 10. Transient responses of adaptive MPPT under the electrical load step from 160 to 330 W (RL = 1 0.5 ).

Fig. 8. Experimental results for MPPT performance under increasing TEG power (2 s/div). (a) Input and (b) output waveforms.

Fig. 11. Dynamic responses of adaptive MPPT under an abruptly changed electromotive force eteg from 6 to 15 V (50 ms/div).

Fig. 9. Experimental results for MPPT performance under decreasing TEG power (2 s/div). (a) Input and (b) output waveforms.

power consequently returns to its previous level. The battery is discharged again due to lack of output power. Fig. 10 shows the transient waveforms during the load step from 160 to 330 W. Increased load requires the battery to be discharged to supply more iR , and thus, ibat is negative. Note that teg and iteg are stable during this variation, and the power converter maintains the maximum power point well. Fig. 11 shows the dynamic response, where, instead of TEG, a dc power supply with 0.187 series resistor is used for easy control. The power supply is commanded from 6 to 15 V. The output power begins at 46 W, which is close to the analytical point of 48 W. The teg and iteg are at 3.2 V (optimal 3 V) and 14.5 A (optimal 16.5 A). Increasing eteg to 15 V, iteg also increases. The new operating point settles down close to the theoretical maximum of 300.8 W at 40.3 A. The settling time is measured to 300 ms, which is well within the design goal of 500 ms. Fig. 12 shows the experimental measurements, where the upper two curves show the extracted power Pg and the theoretical maximum power Pmax , and lower markers are the errors given by Error(%) = Pg 100 %. Pmax (30)

and battery, respectively. Refer to Fig. 6(a). Shown in Fig. 8(a), when the maximum TEG power is increased from 70 to 296 W, the output power is increased accordingly, where the teg and iteg are regulated at 8 V and 37 A, respectively. Shown in Fig. 8(b), the battery begins charging, due to the increasing output power. Fig. 9 shows the experimental results when the TEG power returns to 70 W. It is seen that the iteg is reduced, and the output

The measurement is done at every 1-V increment of eteg and shows acceptable precision of less than 5% error, even under a resistance-temperature variation. The experimental results agree with the transient performance and stability expected by the associated design procedure, verifying the tracking accuracy of the adaptive MPPT.

KIM et al.: ANALYSIS AND DESIGN OF POWER POINT TRACKING SCHEME FOR ENERGY STORAGE SYSTEM

3715

Fig. 12. Extracted power Pg , theoretical maximum power Pmax , and relative error versus electromotive force eteg .

VII. C ONCLUSION In this paper, the theoretical analysis of the maximum power point scheme was presented, considering the dynamics of TEG and the dcdc converter. It was shown that the large scaling factor or small sampling interval increases the dc loop gain of MPPT, which causes less stability due to lack of phase margin. The four factors affecting the dynamics of MPPT were addressed, including the current loop bandwidth, output capacitor, digital sampling effect, and switching frequency. Based on this analysis, the design methodology was derived. Experimental results support the effectiveness of the analysis and design methodology. A PPENDIX Full-order current-to-duty transfer function Gid (s) = iteg = z1 s2 + z2 s + z3 p1 s3 + p2 s2 + p3 s + p4

d z = C C 1 s o RL Rb Rs Vbat z2 = [Co RL Rb + Cs Rs (RL + Rb )] Vbat Cs RL Rb Rs IL (1 D) z3 = RL Rb IL (1 D) + (RL + Rb ) Vbat p1 = L Cs Co RL Rb Rs p2 = L [Cs Rs (RL + Rb ) + Co RL Rb ] p3 = RL Rb Rs Cs (1 D)2 + Co + L (RL + Rb ) p4 = RL Rb (1 D)2 + Rs (RL + Rb ). R EFERENCES

[6] Y.-C. Kuo, T.-J. Liang, and J.-F. Chen, Novel maximum-power-pointtracking controller for photovoltaic energy conversion system, IEEE Trans. Ind. Electron., vol. 48, no. 3, pp. 594600, Jun. 2001. [7] K. Harada and G. Zhao, Controlled power interface between solar cells and AC source, IEEE Trans. Power Electron., vol. 8, no. 4, pp. 654662, Oct. 1993. [8] S. Chiang, K. Chang, and C. Yen, Residential photovoltaic energy storage system, IEEE Trans. Ind. Electron., vol. 45, no. 3, pp. 385394, Jun. 1998. [9] W. Xiao and W. G. Dunford, A modied adaptive hill climbing MPPT method for photovoltaic power systems, in Proc. IEEE Power Electron. Spec. Conf., 2004, pp. 19571963. [10] J.-M. Kwon, K.-H. Nam, and B.-H. Kwon, Photovoltaic power conditioning system with line connection, IEEE Trans. Ind. Electron., vol. 53, no. 4, pp. 10481054, Aug. 2006. [11] F. Liu, S. Duan, F. Liu, B. Liu, and Y. Kang, A variable step size INC MPPT method for PV systems, IEEE Trans. Ind. Electron., vol. 55, no. 7, pp. 26222628, Jul. 2008. [12] J. A. M. Bleijs and A. Gow, Fast maximum power point control of current-fed DCDC converter for photovoltaic arrays, Electron. Lett., vol. 37, no. 1, pp. 56, Jan. 2001. [13] W. Xiao, W. G. Dunford, P. R. Palmer, and A. Capel, Application of centered differentiation and steepest descent to maximum power point tracking, IEEE Trans. Ind. Electron., vol. 54, no. 5, pp. 25392549, Oct. 2007. [14] T.-F. Wu, C.-H. Chang, and Y.-K. Chen, A fuzzy-logic-controlled singlestage converter for PV-powered lighting system applications, IEEE Trans. Ind. Electron., vol. 47, no. 2, pp. 287296, Apr. 2000. [15] M. Veerachacy, T. Senjyu, and K. Uezato, Neural-network-based maximum-power-point tracking of coupled-inductor interleaved-boostconverter-supplied PV system using fuzzy controller, IEEE Trans. Ind. Electron., vol. 50, no. 4, pp. 749758, Apr. 2003. [16] N. Femia, G. Petrone, G. Spagnuolo, and M. Vitelli, Optimization of perturb and observe maximum power point tracking method, IEEE Trans. Power Electron., vol. 20, no. 4, pp. 963973, Jul. 2005. [17] S. Lineykin and S. Ben-yaakov, Modeling and analysis of thermoelectric modules, IEEE Trans. Ind. Appl., vol. 43, no. 2, pp. 505512, Mar./Apr. 2007. [18] Y. H. Kim and H.-D. Ha, Design of interface circuits with electrical battery models, IEEE Trans. Ind. Electron., vol. 44, no. 1, pp. 8186, Feb. 1997. [19] R. W. Erickson and D. Maksimovic, Fundamentals of Power Electronics, 2nd ed. Berlin, Germany: Springer-Verlag, 2001. [20] N. S. Nise, Control System Engineering, 4th ed. New York: Wiley, 2003. [21] S. B. Leeb, A. Oritiz, R. F. Lepard, S. R. Shaw, and J. L. Kirtley, Jr., Applications of real-time median ltering with fast digital and analog sorters, IEEE/ASME Trans. Mechatronics, vol. 2, no. 2, pp. 136142, Jun. 1997.

[1] T. Kyono, R. O. Suzuki, and K. Ono, Conversion of unused heat energy to electricity by means of thermoelectric generation in condenser, IEEE Trans. Energy Convers., vol. 18, no. 2, pp. 330334, Jun. 2003. [2] D. W. Auckland, R. Shuttleworth, A. C. Luff, B. P. Axcell, and M. Rahman, Design of a semiconductor thermoelectric generator for remote subsea wellheads, Proc. Inst. Elect. Eng.Elect. Power Appl., vol. 142, no. 2, pp. 6570, Mar. 1995. [3] R.-Y. Kim and J.-S. Lai, A seamless mode transfer maximum power point tracking controller for thermoelectric generator applications, IEEE Trans. Power Electron., vol. 23, no. 5, pp. 23102318, Sep. 2008. [4] N. Kasa, T. Iida, and L. Chen, Flyback inverter controlled by sensorless current MPPT for photovoltaic power system, IEEE Trans. Ind. Electron., vol. 52, no. 4, pp. 11451152, Aug. 2005. [5] H. Koizumi, T. Mizuno, T. Kaito, Y. Noda, N. Goshima, M. Kawasaki, K. Nagasaka, and K. Kurokawa, A novel microcontroller for gridconnected photovoltaic systems, IEEE Trans. Ind. Electron., vol. 53, no. 6, pp. 18891897, Dec. 2006.

Rae-Young Kim (S06) received the B.S. and M.S. degrees in electrical engineering from the Hanyang University, Seoul, Korea, in 1997 and 1999, respectively. He is currently working toward the Ph.D. degree in electrical engineering at the Virginia Polytechnic Institute and State University, Blacksburg. From 1999 to 2004, he was a Senior Researcher with the Hyosung Heavy Industry R&D Center, Seoul. He is currently a Graduate Research Assistant with the Future Energy Electronics Center, Virginia Polytechnic Institute and State University. His research interests include power converter modeling and control, and highefciency conversions for renewable energy applications.

3716

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 9, SEPTEMBER 2009

Jih-Sheng (Jason) Lai (S85M89SM93F07) received M.S. and Ph.D. degrees in electrical engineering from the University of Tennessee, Knoxville, in 1985 and 1989, respectively. From 1980 to 1983, he was the Head of the Electrical Engineering Department, Ming-Chi Institute of Technology, Taipei, Taiwan, where he initiated a power electronics program and received a Grant from his college and a fellowship from the National Science Council to study abroad. In 1986, he became a Staff Member at the University of Tennessee, where he taught control systems and energy-conversion courses. In 1989, he was with the Electric Power Research Institute (EPRI) Power Electronics Applications Center, where he managed EPRI-sponsored power electronics research projects. In 1993, he was with Oak Ridge National Laboratory as the Power Electronics Lead Scientist, where he initiated a high-power electronics program and developed several novel high-power converters, including multilevel converters and soft-switching inverters. In 1996, he joined the Virginia Polytechnic Institute and State University, Blacksburg, where he is currently a Professor and the Director of the Future Energy Electronics Center. He has published more than 195 technical papers and two books. He is the holder of 17 U.S. patents. His main research areas are in high-efciency power electronics conversions for high power and energy applications. Dr. Lai was the Chairman of the 2000 IEEE Workshop on Computers in Power Electronics, the 2001 IEEE/U.S. Department of Energy Future Energy Challenge, and the 2005 IEEE Applied Power Electronics Conference and Exposition. He is the recipient of several distinctive awards, including a Technical Achievement Award at Lockheed Martin Award Night, two IEEE Industry Applications Society Conference Paper Awards, and Best Paper Awards from the 1997 International Conference on Industrial Electronics, Control, and Instrumentation, the 2005 International Power Electronics Conference, and the 2007 Power Conversion Conference.

Ben York (S05) received the B.S.E.E. degree from The University of Alabama, Tuscaloosa, in 2008. He is currently working toward the Ph.D. degree in electrical engineering at the Virginia Polytechnic Institute and State University, Blacksburg. He is currently a Graduate Research Assistant with the Future Energy Electronics Center, Virginia Polytechnic Institute and State University. His research interests include power converter design and control for renewable energy applications.

Ahmed Koran (S09) received the B.S. degree in electrical power engineering from Yarmouk University, Irbid, Jordan, and the M.S. degree in electrical engineering from the Illinois Institute of Technology, Chicago, in 2004 and 2007, respectively. He is currently working toward the Ph.D. degree in electrical engineering at the Virginia Polytechnic Institute and State University, Blacksburg. From 2004 to 2006, he was a Laboratory Engineer at Yarmouk University, where he was an Instructor from 2007 to 2008. He is currently a Graduate Research Assistant with the Future Energy Electronics Center, Virginia Polytechnic Institute and State University. His research interest includes the development of power electronics systems for renewable energy applications.

S-ar putea să vă placă și