View
213
Download
0
Embed Size (px)
Citation preview
7/22/2019 dcdcpvarchitectures.pdf
1/13
674 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 29, NO. 2, FEBRUARY 2014
An Efficient Partial Power Processing DC/DC
Converter for Distributed PV ArchitecturesMohammed S. Agamy, Senior Member, IEEE, Maja Harfman-Todorovic, Member, IEEE,
Ahmed Elasser, Senior Member, IEEE, Song Chi, Member, IEEE, Robert L. Steigerwald, Fellow, IEEE,
Juan A. Sabate, Member, IEEE, Adam J. McCann, Li Zhang, and Frank J. Mueller
AbstractIn this paper, a dc/dc power converter for distributedphotovoltaic (PV) plant architectures is presented. The proposedconverter has the advantages of simplicity, high efficiency, andlow cost. High efficiency is achieved by having a portion of theinput PV power directly fed forward to the output without be-ing processed by the converter. The operation of this converterallows for a simplified maximum power point tracker design usingfewer measurements. The stability analysis of the distributed PVsystem comprised of the proposed dc/dc converters confirms thestable operation even with a large number of deployed converters.
The experimental results show a composite weighted efficiency of98.22% with very high maximum power point tracking efficiency.
Index TermsBuckboost converter, distributed photovoltaic(PV) architectures, maximum power point tracking (MPPT), par-tial power processing.
I. INTRODUCTION
PHOTOVOLTAIC (PV) power plants with distributed power
electronic converters provide several advantages over the
standard central inverter systems, including higher energy yield,
more flexibility in plant design, and improved monitoring and
diagnostics capabilities [1][18]. Studies show that the distribu-tion of dc/dc converters along with the maximum power point
tracking (MPPT) controllers associated with them, as shown in
Fig. 1, provides a significant increase in the annual energy yield
of the system. A tradeoff study including detailed energy yield,
reliability, and cost analysis showed that a selection of a string-
level MPPT architecture [see Fig. 1(c)] is the best approach
for large commercial installations and utility scale PV systems
(200 kW up to 2 MW) as it provides an annual energy yield gain
of 68% [4], which is more than enough to compensate for the
cost of the additional power electronics.
Manuscript received September 28, 2012; revised December 13, 2012 andFebruary 11, 2013; accepted March 11, 2013. Date of current version August20, 2013. This work was supported in part by the U.S. Depart-ment of Energyunder Grant DE-EE0000572. Recommended for publication by Associate Edi-tor M. Ponce-Silva.
M. S. Agamy is with the School of Engineering, University ofBritish Columbia, Kelowna, BC V1Y 9W9 Canada (e-mail: [email protected]).
M. Harfman-Todorovic, A. Elasser, S. Chi, R. L. Steigerwald, J. A. Sabate,A. J. McCann, L. Zhang, and F. J. Mueller are with the GE Global Re-search Center, Niskayuna, NY 12309 USA (e-mail: [email protected];
[email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]).
Color versions of one or more of the figures in this paper are available onlineat http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TPEL.2013.2255315
Fig. 1. PV plant architectures: (a) centralized, (b) multistring distribution, (c)string distribution, and (d) module distribution.
One of the key factors affecting the distributed PV system
design is the proper selection and design of the dc/dc converters
used in these architectures. Using converters of smaller power
ratings leads to a decrease in conversion efficiency as well as
an increase in cost per unit power as compared to large cen-
tralized converters. This has to be taken into account to ensure
that the benefits of distributing the dc/dc converters in a PVplant are not offset by the drop in conversion efficiency. DC/DC
converter efficiencies on the order of 98% or higher are needed
such that the gains in energy yield obtained by distributing the
dc/dc power conversion stage do not get canceled out by the
drop in the power converter efficiency. These converters replace
the input dc/dc stage in two-stage central inverters, whereas in
the case of single-stage inverters, their design is significantly
simplified as they can now be designed to operate with a con-
stant input dc voltage rather than a wide input voltage range
and the MPPT function is shifted to the dc/dc converters. Even
though the added stage adds to the losses, if this added con-
version stage is designed with high enough efficiency (98% or
higher) and with the added energy yield due to the distributionof MPPT, the impact of power converter distribution on effi-
ciency is reduced [4], [26]. This slight reduction in efficiency is
offset by the increase in energy yield in distributed systems, and
thus, the overall energy harvested from the solar array is im-
proved compared to central inverter systems. Furthermore, the
simplification of the inverter design improves the reliability of
the inverter stage. In this paper, a high-efficiency partial power
buckboost dc/dc converter is presented for use in distributed
PV systems based on its performance, reliability, and estimated
cost.
Based on the comparative system study in [4], power con-
verters at the string or multistring level (1.56 kW rated power)
0885-8993 2013 IEEE
7/22/2019 dcdcpvarchitectures.pdf
2/13
AGAMYet al.: AN EF FICIENT PARTIAL POWE R P ROC ESS IN G DC /D C C ONVE RTER F OR D IS TRIBUTE D P V AR CHITEC TUR ES 6 75
Fig. 2. (a) Full power versus (b) partial power processing structures.
show the best performance/cost tradeoff point. Therefore, a
3.5-kW partial power processing dc/dc converter design is pro-
posed in this paper. The input to the converter can be either one
multicrystalline silicon (mc-Si) string, multiple cadmium tel-
luride (CdTe) or copper indium gallium selenide (CIGS) strings.
The converter is composed of two interleaved 1.75 kW chan-
nels to reduce input current ripple. Efficiency is maximized by
using a partial power-processing scheme as well as operatingonly one of the two interleaved channels at light loads. Further-
more, in very light load conditions the converter is operated in
discontinuous conduction mode to reduce device turn on losses.
This paper is organized as follows. Section II presents a com-
parative study of some key candidate topologies for distributed
PV architectures. In Section III, the design and operation of
the proposed partial power converter are given. Experimental
results are presented in Section IV for the proposed converter
tested in a laboratory setup as well as an actual field test of a
set of parallel converters connected in a distributed architecture.
Further, measurements are also presented to show the stability
of the parallel-operated converters feeding a common grid-tiedinverter stage. Finally, Section V contains concluding remarks.
II. COMPARISON OFCANDIDATE DC/DC
CONVERTER TOPOLOGIES
One way to improve efficiency of the dc/dc converters is by
means of using partial power processing, as shown in Fig. 2
[19][26]. In these converters, part of the input power is directly
fed forward to the output, thus achieving close to 100% effi-
ciency, and the remaining part of the power processed by the
dc/dc converter is determined by the voltage regulation require-
ments, i.e., the percentage of power processed by the converter
depends on the voltage difference between the PV side and thedc-link voltage. Fig. 3 shows the relation between the required
voltage gain and the percentage of input power being processed
by the converter. With a proper design, the power converter can
be designed to handle around 3040% of the input power under
nominal operating conditions, thus improving its cost, size, and
efficiency. Therefore, the dc/dc converter block does not need
to have excessively high efficiency over its operating range to
achieve overall high conversion efficiency. An example of this
is shown in Fig. 4, where a dc/dc converter with an assumed
efficiency of 95% leads to an overall efficiency above 98% for
input voltages that are equal to 60% or higher of the output
dc-link voltage when used in a partial power conversion mode.
Fig. 3. Fraction of total power processed versus voltage gain for a partialpower converter.
Fig. 4. Example of overall efficiency of a partial power conversion topologyassuming a 95% dc/dc converter efficiency.
Energy yield and cost analysis of utility and large commercial
PV power plants led to the conclusion that a string-level dc/dc
power converters provide the optimum cost/benefit point for
plants constructed using high-power low-voltage modules (e.g.,mc-Si), whereas for plants made of higher voltage, lower power
thin film modules a string combiner dc/dc converter distribution
represents the optimal design point. Therefore, the target is to
identify the best converter: its design, operation, and control for
the string/multistring inputs mentioned in the previous section.
Converter size, cost, and ease of system integration are also key
factors in the selection process. For string converters rated at
(1.56 kW), the estimated gain in energy yield is in the range of
39% over the standard central inverter system [4]; therefore, a
target weighted efficiency of 98% is needed in order not to have
a significant negative impact on the annual yield. This weighted
efficiency is based on the California Energy Commission (CEC)weighted efficiency formula for solar inverters
CE C = 0.0410 % + 0.0520 % + 0.1230 %
+ 0.2150 % + 0.5375 % + 0.05100% . (1)
Key ratings targeted for the converter design are as follows:
1) rated power: 1.56 kW;
2) input dc-voltage range: 200600 V;
3) output dc-voltage: 600 V;
4) efficiency: 98% or above;
5) suitable for parallel operationwith several other converters
to feed a grid tied inverter stage.
7/22/2019 dcdcpvarchitectures.pdf
3/13
676 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 29, NO. 2, FEBRUARY 2014
Fig. 5. Three-level boost converter. (a) Circuit topology. (b) Efficiency fordifferent input power and input voltages.
The dc/dc converter topology to be used for such application
should provide the voltage step-up capabilities, MPPT, high
efficiency, high reliability, and design simplicity. Input output
isolation is not a necessary feature; however,it facilitates the par-
allel connection of multiple dcac inverter stages to the same
transformer, since nonisolated topologies require the inverters to
be isolated in order for them to be operated in parallel, which re-
quires the use of large power line medium voltage transformers.Providing the isolation in the dc/dc converter allows the invert-
ers to be tied to a multiwinding power transformer, which leads
to size reduction and cost savings. On the other hand, the ad-
dition of the high-frequency isolation transformer complicates
the design and adds to the converter cost.
In the following sections, some basic converter topologies are
assessed for their suitability for application in a distributed PV
architecture.
A. Transformerless Topologies
The boost converter is the simplest solution for this applica-
tion since it has a low part count and a simple design. However,the boost converter in this operational range has the following
disadvantages: a switching voltage in the (8001200 V) range
is required, which leads to using a relatively low switching fre-
quency and consequently a large input inductor. Furthermore,
the boost converter cannot meet the efficiency requirement with-
out adding auxiliary circuits for soft switching.
To solve the problem of low-frequency operation, a three-
level boost topology can be used, as shown in Fig. 5(a) [27].
Since the voltage stress on the switches is half the output volt-
age, high-frequency operation and compact size can be achieved
by using MOSFETs of lower voltage rating (400600 V). Ef-
ficiency can also be improved by operating the inductor in a
Fig. 6. Partial power boost converter. (a) Circuit topology. (b) Efficiency fordifferent input power and input voltages.
critical conduction mode. The efficiency curves of such a con-
verter are shown in Fig. 5(b).
Another solution is to use a modified boost converter with
partial power processing [23], [24] as shown in Fig. 6(a). In
this proposed topology, the output voltage is the sum of the
PV string voltage and the voltage of the output capacitor. Since
this converter does not need to process all the input power, theoverall conversion efficiency is very high for a very wide range
of the converter input power as shown in Fig. 6(b). The converter
has a very simple topology but it still needs to use devices of
high-voltage rating.
B. Topologies With High-Frequency Transformers
The full-bridge converter is one of the most used isolated
topologies in this power range whether it is a current-fed or
voltage-fed converter, as shown in Fig. 7. However, the converter
is built of four switches, an isolation transformer, and an out-
put rectifier; the added components lead to reduced reliability,
lower power density, and higher cost. Furthermore, achievingefficiencies higher than 98% at low cost becomes much more
challenging for these converters.
To improve the efficiency, a partial power processing full
bridge can be used [28], where only a fraction of the generated
power in the PV string flows through the full-bridge converter
and the remaining power is directly fed forward to the output of
the converter. However, this mode of operation leads to the loss
of the benefit of the converter being an isolated topology. Reso-
nant converters can also be used for building high-power-density
converters since they can operate at high switching frequencies
with natural zero-voltage switching. On the other hand, they
suffer from low partial load efficiency and complex design and
7/22/2019 dcdcpvarchitectures.pdf
4/13
AGAMYet al.: AN EF FICIENT PARTIAL POWE R P ROC ESS IN G DC /D C C ONVE RTER F OR D IS TRIBUTE D P V AR CHITEC TUR ES 6 77
Fig. 7. Full-bridge converter-based topologies: (a) current-fed, (b) voltage-fed, and (c) with partial power.
Fig. 8. Resonant converter with partial power. (a) Topology. (b) Efficiencycurves at different input power and input voltage.
control [29], [30]. The efficiency can be improved with connec-
tions similar to the full bridge with partial power processing as
shown in Fig. 8; however, the resulting converter still remains
complex and expensive in terms of topology design and control
implementation.
TABLE IMTBF COMPARISONS FORTWOCONVERTERSWITHELECTROLYTIC
ANDFILMCAPACITORS
C. Reliability Evaluation
With the higher number of converters and components in
distributed PV plants, the general perception is that the system
reliability will be reduced, leading to the requirement of a very
high reliability for the distributed dc/dc converters. However, a
more practical evaluation metric is system availability. Since in
a distributed system there are multiple converters operating in
parallel, the failure of one or a few converters will have a lowimpact on the plant power delivery capability. Therefore, dc/dc
converters with similar reliability to a central converter provide
high plant availability.
Theconverter reliability is dependent on the number andtypes
of components and their respective stresses, which gives an ad-
vantage to single-switch topologies and to partial power con-
verters, where components are exposed to lower overall stress.
The use of film capacitors to replace electrolytics is also a key
reliability factor [31][33]. As an example, since the three-level
boost converter, in Fig. 5, and the partial power boost con-
verter, in Fig. 6, appeared to best meet the performance/cost
requirements for distributed solar applications, these two con-verters were analyzed in more detail for reliability assessment.
Voltage and/or current stresses of the different converter com-
ponents were evaluated under different power and system volt-
age conditions in order to estimate the converter reliability us-
ing the Telcordia reliability prediction procedure for electronic
equipment [34]. Table I shows the mean time between failures
(MTBF) for the two converter options with film and electrolytic
capacitors. Two channel-interleaved variants of both converters
were analyzed at the rated converter power of 3.5 kW and an
input dc voltage of 480 V. For both cases, each 800 V rated film
capacitor was replaced with two electrolytic capacitors rated at
450 V. The voltage stresses were halved for each electrolytic
capacitor. The results indicate that reliabilities are significantlyimpacted for both topologies. The impact is more severe on
the three-level boost topology due to a larger number of ca-
pacitors being used. According to the values shown in Table I,
high-frequency converter designs provide a significant benefit
in terms of enabling the use of more reliable film capacitors
to replace electrolytic capacitors for this case. The reliability
assessment of semiconductor devices was made based on the
maximum voltage stresses they are required to withstand and
the power loss per device. The overall converter reliability for
both the partial power converter and the three-level boost con-
verter was found to be in excess of 290 000 h, which exceeds
the renewable energy market reliability requirements.
7/22/2019 dcdcpvarchitectures.pdf
5/13
678 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 29, NO. 2, FEBRUARY 2014
TABLE I ICOMPARISONBETWEENKEYTOPOLOGIESUNDERSTUDY
D. Comparison of Topologies
The major considerations when designing a solar power con-
verter are efficiency, design complexity, reliability, power den-
sity, and cost. The number of switching devices is a key factoraffecting efficiency, reliability, and cost; therefore, it is essential
to minimize the number of switches, which gives an advantage
to single-switch topologies. Replacing electrolytic capacitors
with film or ceramic capacitors provides clear reliability gains,
but for this to be practical, the switching frequency has to be in-
creased, which impacts efficiency. Since efficiency has a direct
impact on energy yield, there is a tradeoff between achieving
soft switching in fast switching converters and the complex-
ity of the converter and/or controller design. This leads to the
preference of methods such as discontinuous conduction mode
or critical conduction mode operation, or using new emerging
wide band gap devices with better switching characteristics, toreduce switching losses and improve efficiency compared to
adding auxiliary circuits or using complex resonant converter
control methods. The challenges with new devices are their cost
and unproven reliability, but initial designs and tests show them
as promising solutions [35]. Table II shows a comparison of
the different converter topologies with respect to the key fac-
tors influencing the dc/dc converter selection for distributed PV
architectures. Based on this comparison, it follows that simple
single-switch topologies with partial power processing, such as
the proposed partial power boost topology, to improve efficiency
are the strongest candidates for this type of application. Three-
level boost converters also present a good solution; however, the
need for twice the number of semiconductor devices and passive
components leads to higher cost and lower reliability. The use
of interleaved configurations of partial power boost converters
and three-level boost converters helps improve system perfor-
mance; however, the implementation of the interleaved topolo-
gies is much simpler from both control and topology aspects in
the case of the proposed, new, partial power boost converter.
III. PROPOSED CONVERTER OPERATION ANDDESIGN
Based on the comparison in Section II, the dc/dc converter
chosen for the distributed PV system is the partial power pro-
cessing boost converter as shown in Fig. 9(a), or a multichannel
Fig. 9. Partial power boost dc/dc converterwith a Si IGBT and a SiC Schottkydiode. One channel is rated at 1.75 kW. Input Voltage: 200 to 600 V, OutputVoltage: 600 V regulated. (a) one channel; (b) two channels.
interleaved converter structure as shown in Fig. 9(b). In this
topology, the output voltage is the sum of the PV string voltage
and the voltage of the output capacitor. Since this converter doesnot need to process all the input power, the overall conversion
efficiency is very high. The converter has a very simple topol-
ogy composed of only one switching device and one diode per
channel. The switches and diodes have to withstand the total
output voltage. Classical partial power processing converters
include a high-frequency transformer in their design. The pro-
posed topology does not require a high-frequency transformer,
which simplifies the design and reduces the required converter
cost.
The voltage gain of the regulated voltageVs under mediumto heavy loading conditions [in a continuous inductor current
conduction mode (CCM)]is given in (2), which is theconversionratio of a noninverting buckboost converter
Vs = d
1 d Vin . (2)
In a discontinuous inductor current conduction mode (DCM),
the capacitor voltage is given by
Vs = 1
2
1 +
2dRloadLin fsw
1
Vin (3)
whered is the duty ratio,fsw is the switching frequency,Lin isthe input inductance, and Rload is the equivalent load resistance.
And the fraction of input power processed by the converter is
7/22/2019 dcdcpvarchitectures.pdf
6/13
AGAMYet al.: AN EF FICIENT PARTIAL POWE R P ROC ESS IN G DC /D C C ONVE RTER F OR D IS TRIBUTE D P V AR CHITEC TUR ES 6 79
Fig. 10. Percentage of input power processed by the converter versus inputvoltage with a 600-V output voltage.
Fig. 11. (a)Stages of operationof thepartial power processing dc/dc converterand (b) the corresponding timing diagram.
given by
Fraction of Power Processed =
VsVin
VsVin
+ 1 . (4)
The target converter design for this application is an input
voltage range of (200600 V) with an output dc-link voltage of
600 V. For this range of operation, Fig. 10 shows the percentage
of input power to be processed by the converter over the entire
input voltage range, with a 600 V output.
The operation of the converter is similar to that of a simple
boost circuit. The stages of operation over a switching period
Ts are shown in Fig. 11 and can be summarized as follows.Stage 1 (0 < t < dTs ): In this stage, insulated gate bipolar
transistor (IGBT) (S) is turned ON and the inductor current
builds up
LindiL in
dt =Vin (5)
CsdVs
dt =
Vs+ Vin
Rload. (6)
Fig. 12. Critical inductor value versus duty ratio and fraction of rated powerat 30-kHz switching frequency.
Stage 2 (dTs < t < T2 ): The IGBT is turned OFF and the
inductor current is diverted to the diode (D) where the energy isdischarged into the capacitor Cs . For a continuous conductionmodeT2 =Ts , and the cycle ends at this stage
LindiL in
dt = Vs (7)
CsdVsdt
=iL inVs + Vin
Rload. (8)
Stage 3 (T2 < t < Ts ): This stage occurs in the case of DCM.In this mode, power is transferred from the input and output
capacitors to the output
Lin
diL in
dt = 0 (9)
CsdVsdt
= Vs+ VinRload
. (10)
It is also worth noting that during this mode of operation,
resonances can occur between the input inductor and device
capacitance. The DCM operation leads to zero-current turn-on
of the IGBT, thus reducing the turn-on losses at light load.
Under light-load conditions, the converter is designed to tran-
sition to a critical conduction mode and then DCM to reduce
power losses at turn-on. The input inductor value can be chosen
as a function of converter power and operational duty ratio as
shown in Fig. 12. Fig. 13 shows the loss breakdown of the dc/dc
converter under rated design conditions (3.5 kW, 400 V input),indicating the dominance of the IGBT switching losses over all
the other loss components. Therefore, a switching frequency of
30 kHz was chosen in order to meet the required efficiency tar-
get. Based on this chosen switching frequency, and in order to
reduce turn-on losses at light load, the input inductor is designed
to transition to a critical conduction mode when the input power
drops below 40% of the rated power of any converter channel.
Diode losses are primarily conduction losses since SiC Schottky
diodes are used to eliminate reverse recovery losses.
IGBTs with voltage rating of 1200 V were chosen for this ap-
plication in order to build converters suitable for operation with
a dc-link voltage up to 750 V. In lower voltage applications,
7/22/2019 dcdcpvarchitectures.pdf
7/13
680 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 29, NO. 2, FEBRUARY 2014
Fig. 13. Converter loss breakdown as a percentage of the total losses at nom-inal operating conditions.
the IGBTs can be replaced by MOSFETs, while in this voltage
range only SiC MOSFETs can be used and since they repre-
sent an emerging technology, they still do not meet cost targets
for PV applications. However, preliminary studies of these de-vices show that they can significantly improve the power density
of the propose converters while maintaining the same superior
performance [35]. Since the dc-link voltage is held constant by
the dc-link capacitors of the inverter, the converter capacitors
Cin andCs are designed to filter out the high-frequency ripplegenerated at the switching frequency or twice the switching fre-
quency depending on the number of operational channels. This
eliminates the need for electrolytic capacitors, which can now be
replaced by small-size, high-reliability thin-film capacitors. A
3.5-kW two-channel interleaved converter is shown in Fig. 9(b).
When the input PV power is less than 50% of the rating, the
switching is stopped in one of the channels, thus eliminating allthe associated switching losses and improving light-load effi-
ciency. Operation can be alternated between the two channels
in order to improve the reliability of the converter.
Converter control is based on having an output constant dc
voltage. The dc-link voltage is controlled by the dcac inverter
modulation. The duty cycle(d) is calculated to adjust the PVstring voltage Vin such that the maximum power of the PVstring/array can be tracked. In this case, Vs is determined inorder to compensate the difference between the PV voltage Vinand the output dc-link voltage. The input voltage command
Vi n C m d is calculated by the MPPT controller. A perturb andobserve MPPT algorithm is used for locating local maximum
power points. This algorithm is associated with a global PVvoltage sweep in order to locate the global maximum power
point [34][38]. The voltage sweep is performed periodically
whenever a substantial change in the PV array output power is
observed. A flowchart showing the implemented MPPT algo-
rithm is shown in the Appendix. With a constant output dc-link
voltageVou t , the MPPT control can be simplified to maximizethe output currentIou t through perturbing the input PV voltageVin . There is no power calculation required in this algorithm,thus reducing computation complexity and time. Fig. 14 shows
a simplified block diagram of the converter controller. Since the
system under study involved a large number of inverters run-
ning in parallel feeding a common dc link, the overall system
Fig. 14. Simplified block diagram of the controller including the MPPT con-trol block.
Fig. 15. Frequency response of the closed loop system with multiple paralleloperating converters with (a) input voltage Vin = 500 V and (b) input voltageVin = 200 V.
stability should be evaluated. Based on (5)(10) and with the
representation of the inverter side by its input capacitance, the
output to control transfer function is derived to be
G= Vin (1 + s/z )
(1 d)(1 + sQ n o n + s2
2o n)
(11)
where z = dV2in
L in (1d)2 R l o a d , on = (1d)
L in Cn, Qn =
(1d)R l o a dCn/ L in
,
Cn =Cdc + nCs , Cdc represents the inverter inputcapacitance,andn is the number of parallel-operating converters.
The system closed-loop frequency response at different in-
put voltage levels and different number of parallel converters is
shown in Fig. 15, indicating the overall system stability. The sys-
tem stability is verified experimentally including the frequency
response of the inverter in the following section.In addition to the MPPT control, the converter can also be
used for monitoring and diagnostics of the PV strings and the
dc-network in the plant, fault detection, as well as clamping the
PV string voltage.
IV. EXPERIMENTAL RESULTS
A 3.5-kW, 30-kHz, two-channel dc/dc converter prototype as
shown in Fig. 16 was built and tested. Converter component
values are listed in Table III. The converter power density is
19.5 W/in3 . SiC Schottky diodes are used for D1 and D2 inorder to avoid reverse recovery losses at high frequency. The
input PV voltage ranges from 200 to 600 V, while the output
7/22/2019 dcdcpvarchitectures.pdf
8/13
AGAMYet al.: AN EF FICIENT PARTIAL POWE R P ROC ESS IN G DC /D C C ONVE RTER F OR D IS TRIBUTE D P V AR CHITEC TUR ES 6 81
Fig. 16. Prototype of the 3.5 kW (two channels).
TABLE I II
CONVERTER COMPONENTS
Fig. 17. Measured efficiency for different input power, input voltages, andconverter configurations.
voltage is fixed at 600 V, which is regulated by the grid-tied
dcac inverter stage.
Converter efficiency evaluation was performed using a
variable-input dc voltage (200480 V) and an electronic load
operating in a fixed voltage mode with a set point of 600 V.
Efficiency measurements were performed across the full-load
range (10100%) in order to generate the weighted efficiency
number for solar converters as given in (1). As seen in Fig. 17,
Fig. 18. Efficiencyimprovementdue tocoordinatedswitching of thetwo chan-nels. (Solid lines two channels switching, Dotted lines: one channel switched
OFF at light load.)
Fig. 19. Block diagram of converter test setup.
the weighted efficiency of each individual channel exceeds 98%
with a peak efficiency reaching 98.9%. The number of channels
operated is decided by the input power; therefore, for power
levels less than 50% of the converter rating only one channel is
switched.This improves the efficiency profile, giving a weightedefficiency of 98.22%, as shown in Fig. 18. The converter was
then tested with a PV emulator input and its output connected
to the DC link of a grid tied inverter as shown in Fig. 19. In-
terleaved inductor currents and the resulting low ripple input
current are shown in Fig. 20(a). CCM operation at high input
power and DCM operation at low input power are shown in
Fig. 20(b) and Fig. 20(c), respectively.
Global MPPT sweep is performed, as shown in Fig. 21 to
locate the absolute maximum output power by sweeping the
PV voltage from its open-circuit value, to a set minimum value
(200 V in this test). Since the output voltage is constant, the
output current profile during the sweep matches the PV (power
voltage profile). Finally, the MPPT performance was studied byusing the PV emulator to apply a transient irradiance profile, as
shown in Fig. 22(a) to the PV string characteristic. Fig. 22(b)
shows the power extracted from the PV string and the tracking
efficiency of the MPPT controller.
At low power levels, only one channel is switched and as
the input PV power increases both channels are operated. With
a constant output voltage, a low-power condition is indicated
by the drop of the output current below a predefined threshold
(ideally 50% of the rated output current). The MPPT efficiency
achieved is above 99.78% under the static condition and during
fast transients it remains above 96%, which guarantees high PV
energy extraction.
7/22/2019 dcdcpvarchitectures.pdf
9/13
682 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 29, NO. 2, FEBRUARY 2014
Fig. 20. Converterwaveforms for different conditions:(a) interleaved inductorcurrents (yellow & green, LEM output 1 V/div1.6 A/div), total input current(magenta, 2 A/div), and gate of S1 (blue, 5 V/div), at input of 2 kW; (b) IGBTvoltage (yellow, 200 V/div), diode voltage (magenta, 200 V/div), and inductorcurrent (green, 3 A/div) of one channel, at total input of 1.75 kW and gate ofS1 (blue, 10 V/div); and (c) IGBT voltage (yellow, 200 V/div), diode voltage(magenta, 200 V/div), and inductor current (green, 2A/div) and gate of S1 (blue,10V/div) of one channel, at total input of 700 W.
Fig. 21. Converter waveforms during a global MPPT sweep: Ch1 (yellow):PV input voltage (100 V/div), Ch2 (blue): Output voltage (100 V/div), Ch3(magenta): Converter output current (1 A/div), Ch4 (green): Input PV current(1 A/div).
Distributed PV power plants consist of a large number of
dc/dc converters connected in parallel to the central dc/ac in-
verter as shown in Fig. 23.Usually, dc/dc converters aredesigned
for predefined stand-alone input and load conditions, e.g., for
an ideal input voltage source and a resistive load. However,
the predefined conditions are hardly met in this system. There-
fore, the dynamic performance of the dc/dc converter within thesystem can be very different from that in the stand-alone op-
eration. Specifically, each individual converter may have well
designed control loops in a stand-alone operation, but exhibit
deteriorated performance within a system. Generally, converter
input and output impedances are used for the verification of
the power system stability and dynamic performance [39][42].
Fig. 24 shows the schematic of the experimental setup used to
verify the system stability. The dc/dc converters under test were
connected to a single-phase dc/ac inverter. In order to ensure
the small signal stability of the interconnected system, formed
by integrating subsystems that are individually stable, the minor
loop gain TM =Zo/Zin must be checked for stability [43], [44].
Fig. 22. (a) Irradiance profile (200 W/m2/div) and (b) Extracted PV power(blue, 200 W/div) and MPPT efficiency (red, %).
Fig. 23. Parallel boardsconnectedto a three-phase inverterin a large PV plant.
To determine the system stability, the construction ofTM frommeasured impedancesZo andZin at the interface between thedc/dc converter and the dc/ac inverter is shown in Fig. 25. The
blue trace in Fig. 25(a) represents the measured loop gain of
the simple system shown in Fig. 24(a) with one dc/dc con-
verter connected to the inverter and the red trace represents the
measured loop gain with three dc/dc converters, the outputs of
which are connected in parallel to the dc link of the inverter
7/22/2019 dcdcpvarchitectures.pdf
10/13
AGAMYet al.: AN EF FICIENT PARTIAL POWE R P ROC ESS IN G DC /D C C ONVE RTER F OR D IS TRIBUTE D P V AR CHITEC TUR ES 6 83
Fig. 24. Interconnection system under test formed by integrating two subsys-tems: (a) one converter connected to one single-phase inverter and (b) threeparallel converters connected to one single-phase inverter.
for the 10.5 kW system under test. To determine if the system
is stable, the phase margin test is used [45]. From Fig. 25, itcan be seen that the phase margin for the case of one dc/dc
converter (blue trace) is measured to be 61 at 25.1 kHz and84 at 42.1 kHz, while for the case of three parallel dc/dc con-verters (red) trace the phase margin is 62 at 18 kHz and 100
at 50 kHz. This method can be expanded to systems with a
much larger number of parallel-connected converters feeding a
high-power dc/ac inverter stage to determine the overall system
stability.
Stability of the power system was also tested in the time
domain by applying global MPPT sweeps, as shown in Figs. 26
and 27. In this case, three dc/dc converters were connected in
parallel to the dc/ac inverter. The input of each converter is
connected to three parallel PV strings, each consisting of fourseries modules. The measured transient waveforms of the dc-
link voltage, input currents of the three dc/dc converters are
shown in Fig. 26. In Fig. 27, the dc-link voltage and the three
output currents from each converter are shown during a similar
global MPPT sweep. The measurements in Figs. 26 and 27
confirm the stability of the parallel-connected set of converters
even under dynamically varying conditions.
Further, the two figures show another benefit of distributed dc
converter architectures for solar applications since each of the
converters can be properly timed to locate the maximum power
point of the string or array it is connected to without causing
a significant dip in the overall system power as would be the
Fig. 25. Measured loop gain magnitude and phase in case of one dc/dc con-verter connected to one dc/ac inverter (blue) and three parallel converters con-nected to the same inverter (red).
Fig. 26. Input current of three parallel converters connected to a PV array andfeeding an inverter during shifted global MPPT sweeps. Ch 1 (yellow): dc-linkvoltage (100 V/div), Ch 2 (blue): input current of dcdc converter 1 (1 A/div),Ch 3 (magenta): input current of dcdc converter 2 (1 A/div), Ch 4 (green):input current of dcdc converter 3 (1 A/div).
Fig. 27. Output current of three parallel converters connected to a PV ar-
ray and feeding an inverter during shifted global MPPT sweeps. Ch1 (yel-low): dc-link voltage (100 V/div), Ch2 (blue): input current of dcdc converter1 (1 A/div), Ch 3 (magenta): input current of dcdc converter 2 (1 A/div), Ch 4
(green): input current of dcdc converter 3 (1 A/div).
7/22/2019 dcdcpvarchitectures.pdf
11/13
684 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 29, NO. 2, FEBRUARY 2014
case if a central inverter were to perform a global maximum
power point sweep. The output current waveform during the
voltage sweep, in Fig. 27, takes the typical form of a PV power
curve verifying the simplified MPPT control described earlier
that uses only the output current, rather than calculating power
in each computation cycle.
V. CONCLUSION
A simple high-efficiency dc/dc converter suitable for
medium- to large-scale distributed PV applications is proposed.
High efficiency is achieved by means of partial powerprocessing
as well as by coordinating the operation of the interleaved chan-
nels of the converter. The output of the converter being a fixed
dc bus also simplifies the MPPT implementation. Furthermore,
feedback signals can be used as monitoring and diagnostics
tools for assessing the condition of the PV plant. The proposed
converter was fully tested for performance parameters includ-
ing the efficiency, switching operation, MPPT performance, and
parallel operation under static and dynamic conditions. Exper-
imental results show excellent performance and fulfillment of
the design objectives. A set of three parallel converters was also
tested in a solar field and showed uninterrupted performance
over a long period of time under varying environmental and
weather conditions.
APPENDIX
Flowchart of the MPPT algorithm.
ACKNOWLEDGMENT
Disclaimer:This report was prepared as an account of work
sponsored by an agency of the U.S. Government. Neither the
U.S. Government nor any agency thereof, nor any of their em-
ployees, makes any warranty, express or implied, or assumes
any legal liability or responsibility for the accuracy, complete-
ness, or usefulness of any information, apparatus, product, or
process disclosed, or represents that its use would not infringe
privately owned rights. References herein to any specific com-
mercial product, process, or service by trade name, trademark,
manufacturer or otherwise does not necessarily constitute or im-
ply its endorsement, recommendation, or favoring by the U.S.
Government or any agency thereof. The views and opinions of
authors expressed herein do not necessarily state or reflect those
of the U.S. government or any agency thereof.
REFERENCES
[1] N. Kaushika and N. Gautam, Energy yield simulations of interconnectedPV arrays, IEEE Trans. Energy Convers., vol. 18, no. 1, pp. 127134,May 2003.
[2] N. Femia, G. Lisi, G. Petrone, G. Spagnuolo, and M. Vitteli, Distributedmaximum power point tracking of photovoltaic arrays: Novel approachand system analysis,IEEE Trans. Ind. Electron., vol. 55,no. 7, pp. 26102621, Jul. 2008.
[3] S. Poshtkouhi, V. Palaniappan, M. Fard, and O. Trescases, A generalapproach for quantifying the benefit of distributed power electronics forfine grained MPPT in photovoltaic applications using 3-D modeling,
IEEE Trans. Power Electron., vol. 27, no. 11, pp. 46564666, Nov. 2012.[4] A. Elasser, M. Agamy, J. Sabate, R. Steigerwald, R. Fisher, and
M. Harfman-Todorovic, A comparative study of central and distributedMPPT architectures for megawatt utility and large scale commercial pho-tovoltaic plants, in Proc. Ind. Electron. Conf., 2010, pp. 27532758.
[5] J. Popovic-Gerber, J. A. Oliver, N. Cordero, T. Harder, J. A. Cobos,M. Hayes, S. C. OMathuna, and E. Prem, Power electronics enablingefficient energy usage: Energy savings potential and technological chal-lenges,IEEE Trans. Power Electron., vol. 27, no. 5, pp. 23382353, May2012.
[6] G. Lijun, R. Dougal, L. Shengyi, and A. Lotova, Parallel-connected solarPV system to address partial and rapidly fluctuating shadow conditions,
IEEE Trans. Ind. Electron., vol. 56, no. 5, pp. 15481556, May 2009.[7] H.Pateland V. Agarwal,MATLAB-basedmodelingto study theeffectsof
partial shading on PV array characteristics,IEEE Trans.Energy Convers.,
vol. 23, no. 1, pp. 302310, Mar. 2008.[8] A. Chouder and S. Silvestre, Analysis model of mismatch power losses
in PV systems, J. Solar Eng., vol. 131, pp. 024504-1024504-5, May
2009.[9] N. Femia, G. Lisi, G. Petrone, G. Spagnuolo, and M. Vitteli, Distributed
maximum power point tracking of photovoltaic arrays: Novel approachand system analysis,IEEE Trans. Ind. Electron., vol. 55, no. 7, pp. 26102621, Jul. 2008.
[10] Y. Xue, L. Chang, S. Kjaer, J. Bordonau, and T. Shimizu, Topolo-gies of single-phase inverters for small distributed power generators: Anoverview,IEEE Trans. Power Electron., vol. 19, no. 5, pp. 13051314,Sep. 2004.
[11] J. Young-Hyok, K. Jun-Gu, P. Sang-Hoon, K. Jae-Hyung, and W. Chung-Yuen, C-language based PV array simulation technique considering ef-fects of partial shading, in Proc. Int. Conf., Ind. Technol., 2009, pp. 16.
[12] R. Bruendlinger, B. Bletterie, M. Milde, and H. Oldenkamp, Maxi-
mum power point tracking performance under partially shaded PV arrayconditions, in Proc. 21st Eur. Photovolt. Solar Energy Conf., Dresden,Germany, 2006, pp. 21572160.
[13] R. Ramabadran and B. Mathur, Effect of shading on series and parallelconnected solarPV modules,Modern Appl. Sci., vol.3, no.10,pp. 3241,Oct. 2009.
[14] M. Garcia, J. Maruri, L. Marroyo, E. Lorenzo, and M. Perez, Partialshadowing, MPPT performance and inverter configurations: Observationsat tracking PV plants,Prog. Photovolt.: Res. Appl., vol. 16, pp. 529536,Apr. 2008.
[15] U. Schwabe and P. Jansson, Performance measurement of amorphousandmonocrystallinesilicon PV modulesin eastern U.S. energyproductionversus ambient and module temperature, in Proc. IEEE Instrum. Meas.Technol. Conf., Singapore, May 57, 2009, pp. 16361641.
[16] S. Krauter, A. Preiss, N. Ferretti, and P. Grunow, PV yield prediction forthin film technologies and the effect of input parameters inaccuracies, inProc. 23rd Eur. Photovoltaic Solar Energy Conf. Exhib. , Valencia, Spain,
Sep. 15, 2008, pp. 740743.
7/22/2019 dcdcpvarchitectures.pdf
12/13
AGAMYet al.: AN EF FICIENT PARTIAL POWE R P ROC ESS IN G DC /D C C ONVE RTER F OR D IS TRIBUTE D P V AR CHITEC TUR ES 6 85
[17] (2010). [Online]. Available: http://www.steveransome.com/pubs.html[18] (2012). [Online]. Available: http://enphaseenergy.com/support/learning
center.cfm[19] R. Adler, A New DC/DC switching regulator topology enhances effi-
ciency and power density, in Proc. PowerCon, 1984, pp. F1-1F1-4.[20] Min, J. Lee, J. Kim, T. Kim, D. Yoo, and E. Song, A new topology with
high efficiency throughout all load range for photovoltaic PCS, IEEETrans. Ind. Electron., vol. 56, no. 11, pp. 44274435, Nov. 2009.
[21] P. S. Shenoy, K. A. Kim, B. B. Johnson, and P. T. Krein, Differential
power processing for increased energy production and reliability of pho-tovoltaic systems,IEEE Trans. Power Electron., vol. 28, no. 6, pp. 29682979, Jun. 2013.
[22] H. Li, H. Chen, and L. Chang, Analysis and design of a single-stage par-allel AC-to-DC converter,IEEE Trans. Power Electron., vol. 24, no. 12,pp. 29893002, Dec. 2009.
[23] M. de Rooij, J. Glaser, and R. Steigerwald, High efficiency photovoltaicinverter, U.S. Patent US2009/0323379 A1, Jun. 27, 2008.
[24] R. Steigerwald, M. Agamy, M. Harfman-Todorovic, A. Elasser, andJ. Sabate, Dc to Dc power converters and methods for controlling thesame, U.S. Patent US2012/0051095 A1, June 29, 2011.
[25] A. Min, J. Lee, J. Kim, T. Kim, D. Yoo, and E. Song, A new topologywith high efficiencythroughout allload range forphotovoltaic PCS,IEEETrans. Ind. Electron., vol. 56, no. 11, pp. 44274435, Nov. 2009.
[26] M. Agamy, M. Harfman-Todorovic, A. Elasser, J. Sabate, R. Steigerwald,Y. Jiang, and S. Essakiappan, DC/DC converter topology assessment for
large scale distributed photovoltaic plant architectures, in Proc. EnergyConvers. Conf. Expo., 2011, pp. 764769.
[27] G. Yao, L. Hu, Y. Liu, L. Chen, and X. He, Interleaved three-level boostconverter with zero diode reverse recovery loss, in Proc. Appl. Power
Electron. Conf., 2004, pp. 10901095.[28] J. Lee, B. Min, T. Kim, D. Yoo, and J. Yoo, High efficient interleaved
input-series-output-parallel-connected DC/DC converter for photovoltaicpower conditioning system, inProc. Energy Convers. Conf. Expo., 2009,pp. 327329.
[29] R. Steigerwald, A comparison of half-bridge resonant converter topolo-gies,IEEE Trans. Power Electron., vol. 3, no. 2, pp. 174182, Apr. 1988.
[30] M. Kazimierczuk andD. Czarkowski,Resonant Power Converters. NewYork, NY, USA: Wiley, 1995.
[31] RELIABILITY: Guidelines to understanding reliability prediction, Eur.Power Supply Manufacturers Assoc. Rep., Wellingborough, U.K., Edition24, Jun. 2005.
[32] E. Koutroulis and F. Blaabjerg, Design optimization of transformerless
grid-connected PV inverters including reliability, IEEE Trans. PowerElectron., vol. 28, no. 1, pp. 325335, Jan. 2013.
[33] Y. Song and B. Wang, Survey on reliability of power electronic systems,IEEE Trans. Power Electron., vol. 28, no. 1, pp. 591604, Jan. 2013.
[34] Reliability Prediction Procedure for Electronic Equipment, Telcordia,Piscataway, NJ, USA, Telcordia Rep. (Method 1 Case 3) SR-332, Issue 2,Sep. 2006.
[35] M. S. Agamy, S. Chi, A. Elasser, M. Harfman-Todorovic, Y. Jiang,F. Mueller, and F. Tao, A high power density Dc-Dc converter for dis-tributed PV architectures,IEEE J. Photovolt., vol. 3, no. 2, pp. 791798,Mar. 2013.
[36] T. Esram and P. Chapman, Comparison of photovoltaic array maximum
power tracking techniques,IEEE Trans. Energy Convers., vol. 22, no. 2,pp. 439448, Jun. 2007.
[37] Y.-H. Ji, D.-Y. Jung, J.-G. Kim, J.-H. Kim, T.-W. Lee, and C .-Y. Won, Areal maximum power point tracking method for mismatching compensa-
tion in PV array under partially shaded conditions, IEEE Trans. PowerElectron., vol. 26, no. 4, pp. 10011009, Apr. 2011.[38] A. Abdelsalam, A. Massoud, S. Ahmed, and P. Enjeti, High performance
adaptive perturb and observe MPPT technique for photovoltaic- basedmicrogrids,IEEE Trans. Power Electron., vol. 26, no. 4, pp. 10101021,Apr. 2011.
[39] D. C. Jones and R. W. Erickson, Probabilistic analysis of a generalizedperturb andobserve algorithmfeaturingrobust operationin thepresenceofpowercurve traps,IEEE Trans. Power Electron., vol.28,no.6, pp. 2912
2926, Jun. 2013.[40] Y. Panov and M. Jovanovic, Practical issues of input/ output impedance
measurements in switching power supplies and application of measureddata to stability analysis, in Proc. Appl. Power Electron. Conf., 2005,pp. 13391345.
[41] G. S. Thandi, R. Zhang, K. Xing, F. C. Lee, and D. Boroyevich, Model-ing, controland stability analysis of a PEBB based DC DPS,IEEE Trans.Power Del., vol. 14, no. 2, pp. 497505, Apr. 1999.
[42] J. He and Y. Wei Li, Generalized closed-loop control schemes withembedded virtual impedances for voltage source converters with LC or
LCL filters,IEEE Trans. Power Electron., vol. 27, no. 4, pp. 18501861,Apr. 2012.
[43] C. M. Wildrick, Stabiliy of distributed power supply systems, M.S.thesis, Dept. Electr. Comput. Eng., Virginia Polytech. Inst. State Univ.,Blacksburg, VA, USA, Feb. 1993.
[44] R. Ridley and B. Cho, Distributed power systems, in Proc. High Fre-quency Power Convers. Conf., 1992, pp. 184.
[45] R.W. Erickson and D. Maksimovic,Fundamental of Power Electronics.Norwell, MA, USA: Kluwer, 2001.
Mohammed S. Agamy (S01M08SM11) re-ceived the B.Sc. (Hons.) and M.Sc. degrees fromAlexandria University, Alexandria, Egypt, in 2000and 2003, respectively, and the Ph.D. degree fromQueens University, Kingston, ON, Canada, in 2008,
all in electrical engineering.Heis currentlyan AssistantProfessor atthe School
of Engineering, University of British Columbia,Kelowna, BC, Canada. From 2008 to July 2012, hewas a Lead Power Electronics Engineer at the Gen-
eral Electric Global Research Center, Niskayuna, NY,USA, where his research was focused on power supply technologies for renew-able energy sources and medical equipment. From May 2003 to October 2008,he was with the Energy and Power Electronics Applied Research Laboratory,Queens University as a Research Assistant and then a Postdoctoral Fellow.From September 2000 to April 2003, he was an Assistant Lecturer at Alexan-dria University. He holds two U.S. patents with six others pending and hasmore than 35 published technical papers in refereed journals and conferences.His research interests include resonant converters, power factor correction, soft-switchingtechniques, andmodeling andcontrol of power converters and electricmachines.
Dr. Agamy serves as a Reviewer for the IEEE TRANSACTIONS ON P OWER
ELECTRONICS, the IEEE TRANSACTIONS ON EDUCATION, the IEEE TRANSAC-TIONS ONINDUSTRYAPPLICATIONS, the IEEE TRANSACTIONS ON INDUSTRIAL
ELECTRONICS, IEEE JOURNAL OF PHOTOVOLTAICS , International Journal ofElectronics, and several IEEE conferences.
Maja Harfman-Todorovic (S02M08) receivedthe Dipl.Ing. degree from the Faculty of Electrical
Engineering, University of Belgrade, Belgrade, Ser-bia, in 2001, and the M.S. and Ph.D. degrees fromTexas A&M University, College Station, TX, USA,in 2004 and 2008, respectively.
Since March 2008, she has been a Lead Engineerin the Utility Power Electronics Laboratory, GeneralElectric Research Center, Niskayuna, NY, USA. Sheholds one U.S. patent with eight others pending andhas more than 30 published technical papers in refer-
eed journals and conferences. Her research interests include converters for PVapplications, subsea oil and gas applications, switching-mode power supply de-sign, uninterruptible power systems, energy storage devices, and digital controlof power converters.
Dr. Harfman-Todorovic serves as a Reviewer for the IEEE T RANSACTIONSON POWER ELECTRONICS, the IEEE TRANSACTIONS ONEDUCATION, the IEEETRANSACTIONS ONINDUSTRY APPLICATIONS, the IEEE TRANSACTIONS ONIN-DUSTRIALELECTRONICS, and several IEEE conferences.
7/22/2019 dcdcpvarchitectures.pdf
13/13
686 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 29, NO. 2, FEBRUARY 2014
Ahmed Elasser (S92M96SM12) was born inDemnate, Morocco, in 1963. He received the Inge-
nieur DEtat diploma in electric power engineeringfrom the Mohammadia Schoolof Engineering,Rabat,Morocco, in 1985, and the M.S. and Ph.D. degreesin electric power engineering and power electron-ics from Rensselaer Polytechnic Institute, Troy, NY,USA, in 1993 and 1996, respectively.
He worked as an Electrical Maintenance Engineer
andthen as a Laboratory Engineer from 1986to 1992,in Morocco. In 1996, he joined the General Electric
(GE) Global Research Center (GRC), Niskayuna, NY, as a Senior Professional.His previous research interests include the study, modeling, and application ofpowersemiconductor devices, systems modeling and simulation,silicon carbidedevices, six-sigmaquality, ande-engineering. From 2002to 2007, he ledthe new
ideas and innovation within the GRC Micro and Nano Structures Technology(MNST) organization and he was heading the MNST Disruptive TechnologyCouncil. He worked across GRC with various technology councils to create aculture of innovation and growth. His current research interests include siliconcarbide power devices fabrication, design, modeling, testing, characterization,and applications; he hasalso been working on photovoltaicsystems,focusing onplant architectures, distributed MPPT, and balance of systems work. He has pre-sented at many IEEE conferences on power electronics, power semiconductordevices, and PV systems. He has authored or coauthored more than 25 papers,holds 13 patents, and has several patents pending.
Dr. Elasser is a Regular Reviewer for various IEEE publications and confer-ences and has recently served as a Topic Chair for the ECCE 2011 sustainableenergy track as well as a Session Chair. He received the GE Dushman TeamAward, in 1996 and a GE Whitney team award, in 2012, he also earned numer-
ous awards from the GE Research Center for his technical and organizationalcontributions. GE Industrial Systems also recognized him for his numerouscontributions to circuit breaker modeling and design.
SongChi (S04M07) received the B.S. degreefromNortheastern University, Shenyang, China, in 1993,the M.S. degree from Tsinghua University, Beijing,China, in 2000, and the Ph.D. degree from The OhioState University, Columbus, OH, USA, in 2007, allin electrical engineering.
He is currently with the GE Global Research Cen-
ter, Niskayuna, NY, USA. Prior to joining GE, hewas a Senior Engineer with the Research and En-gineering Center of Whirlpool. He has authored orcoauthored 15 technical papers in IEEE conferences
and journals. He has two patents pending. His research interests include sensor-less control of ac drives, flux-weakening control of Surface mounted permanentmagnet/interior permanent magnet machines, controls of power conversion sys-tems such as distributed solar systems and high-fidelity gradient amplifiers ofmagnetic resonance imager.
Robert L. Steigerwald (S66M79SM85F94)received the B.S.E.E. degree from Clarkson College,Omaha, NE, USA, in 1967,and the M.E.E. and Ph.D.degrees from Rensselaer Polytechnic Institute, Troy,NY, USA, in 1968 and 1978, respectively.
He joined GE Global Research, in 1968, wherehe did research and development in many areas ofPower Electronics until his retirement in 2005. Heformed his consulting company (Adirondack PowerProcessing, LLC), in 2006. His inventions and de-signs have been applied to variable speed DC and AC
motor drives, induction heating, high-frequency lamp ballasts, uninterruptablepower supplies, utility interactive photovoltaics, satellite power supplies, radarand sonar power supplies, CT X-ray generators, and MRI gradient amplifiers.These applications employ both low- and high-voltage power supplies includ-ing high power factor converters and resonant, soft-switched converters. He hasreceived 120 patents and published more than 50 technical papers.
Dr. Steigerwald received the William E. Newell Outstanding Achievement
Award by the Power Electronics Society in 1993, and the IEEE 3rd Millenniummedal in 2000 for outstanding achievements. He is currently on the ScientificAdvisory Board of the Center for Power Electronic Systems at Virginia Tech,Blacksburg, USA.
Juan A. Sabate(M94) received the M.S. and Ph.D.degrees in electrical engineering from Virginia Tech,
Blacksburg, VA, USA, in 1988 and 1994, respec-tively.
Since 2000, he has been a Power Electronic Re-searcher in GE Global Research Center, Niskayuna,NY, USA. He has led multidisciplinary teams of sci-entists and engineers to develop power supplies andhigh-power switching amplifiers for GE energy and
medical applications. From 1997 to 2000, he waswithPhilips ElectronicsResearchin BriarcliffManor,
New York, where he conducted research and advanced development of highpower density power supplies for commercial applications, and ballasts for flu-orescent and HID lamps. From 1994 to 1997, he was with Hewlett-Packard intheir R&D center in Barcelona, Spain, where he designed high power density
dcdc converters andspecial purposesensors.In addition,from 1994to 1997, hewasa Lecturer and Adjunct Professor in the Ramon Llull University, Barcelona,where he was responsible for the Power Electronics Curriculum. His researchinterests include high-performance power conversion, distributed power gener-ation, lighting electronics, and modeling and control of PWM converters. Hehas published more than 50 journal and conference papers, and holds 14 U.S.patents.
Adam J. McCann was born in Salt Lake City, UT,
USA, in 1987. He received the B.S. and M.Eng. de-grees from Cornell University, Ithaca, NY, USA, inapplied engineering physics and electrical and com-puter engineering, in 2010 and 2011, respectively.
He joined General Electric Global Research Cen-ter (GRC) in Niskayuna, NY, USA, as part of theEdison Engineering Development program, in 2011.While at GRC, he has coauthored papers in signalprocessing and remote monitoring applications alongwith patents pending in the area of OCR and pattern
recognition. His current research interests include embedded device architec-tures and communication platforms in the Internet of Things.
Li Zhang received the Ph.D. degree from the Uni-versity of Massachusetts Amherst, Amherst, MA,
USA, in 2004, and the M.S. degree from the Univer-sity of Electronic Science and Technology of China,Chengdu, China, in 2000.
He is currently a Lead Scientist in the DistributedIntelligent Systems Lab at GE Global Research Cen-ter, Niskayuna, NY, USA. He has over 15 years ex-
perience in thefieldsof Internet of Things, electrome-chanical systems design, smart sensing systems, au-tomatic identification, wireless networking, remote
monitoring, and diagnostics. He is an industrys expert in the design of CyberPhysical Systems and is actively leading GEs research efforts in designinghardware, software, algorithms, networks and web-based interaction systemsfor digitizing millions of assets and objects found in aviation systems, powersystems, healthcare, pharmaceutics, etc.
Frank J. Mueller attended Rensselaer PolytechnicInstitute (RPI), Troy, NY, in 1977, pursuing a majorin physics. In 1979, he enlisted in the U.S. Air Forcewhere he completed a Navy ET-A school at GreatLakes Naval Base and then specialized in Meteoro-logical Equipmenttrainingat ChanuteAir Force Base(AFB), IL, USA (he graduated atthe top of his class).
Afterward, he went to Blytheville Air Force Base,Blytheville, AR, USA, where he maintained theweather radar and other meteorological equipmentfor the base. In 1983, he joined the microprocessor
laboratory at RPI assisting in lab preparations and later also supported a PowerElectronics Lab which included laying out gate drive circuits. In 1996, he joined
General Electric, Niskayuna, NY, USA, and has been active in Power Electron-ics testing and layout. He has experience in optics, ultrahigh vacuum systems,and high-voltage work up to 160 kV and has been leading his organizationssafety program since 1999. He holds two U.S. patents and has two pending.