Modeling and Simulation of Power Electronic

Converters
DRAGAN MAKSIMOVIĆ, MEMBER, IEEE, ALEKSANDAR M. STANKOVIĆ, MEMBER, IEEE,
V. JOSEPH THOTTUVELIL, MEMBER, IEEE, AND GEORGE C. VERGHESE, FELLOW, IEEE

Invited Paper

This paper reviews some of the major approaches to modeling The analysis and design of such systems presents significant
and simulation in power electronics, and provides references that challenges.
can serve as a starting point for the extensive literature on the sub- Modeling and simulation are essential ingredients of the
ject. The major focus of the paper is on averaged models of var-
ious kinds, but sampled-data models are also introduced. The im- analysis and design process in power electronics. They help
portance of hierarchical modeling and simulation is emphasized. a design engineer gain an increased understanding of circuit
operation. With this knowledge the designer can, for a given
Keywords—Averaged models, boost converter, circuit aver-
aging, dynamic phasors, hierarchical methods, modeling, power set of specifications, choose a topology, select appropriate
electronics, power factor correction, sampled-data models, simu- circuit component types and values, estimate circuit perfor-
lation, state-space averaging, switched models. mance, and complete the design by ensuring—using Monte
Carlo simulation, worst case analysis, and other reliability
and production yield analyses—that the circuit performance
I. INTRODUCTION
will meet specifications even with the anticipated variations
A. Modeling and Simulation in operating conditions and circuit component values.
Power electronic systems are widely used today to pro- The increased availability of powerful computing has
vide power processing for applications ranging from com- made direct simulation widely accessible [1]–[12] and
puting and communications to medical electronics, appliance has enlarged the set of tractable modeling and analysis
control, transportation, and high-power transmission. The as- approaches. Simulation of a full production schematic still
sociated power levels range from milliwatts to megawatts. remains an elusive goal; the obstacles include the need
These systems typically involve switching circuits composed for extensive model building, excessively long simulation
of semiconductor switches such as thyristors, MOSFETs, times, the challenges of automatically recognizing and
and diodes, along with passive elements such as inductors, exploiting modular or hierarchical or time-scale structure
capacitors, and resistors, and integrated circuits for control. [13], the difficulties of coupling diverse modeling and
simulation modalities, and the effects of layout, packaging,
and parasitics. Even if it were possible to simulate a full
schematic with sufficient accuracy and efficiency, it is
Manuscript received November 28, 2000; revised February 1, 2001. The doubtful whether this capability alone would provide the
work of D. Maksimović was supported by the National Science Foundation basis for good design. Typically, crucial insight and under-
under Grant ECS-9703449. The work of A. M. Stanković was supported standing are provided by hierarchical modeling, analysis,
by the National Science Foundation under Grants ECS-9502636 and ECS-
9820977, and by the Office of Naval Research under Grant N14-97-1-0704. and simulation, rather than working directly with a detailed
D. Maksimović is with the University of Colorado, Boulder, CO 80302 schematic. The combination of these insights with hardware
USA. prototyping and experiments constitutes a powerful and
A. M. Stanković is with Northeastern University, Boston, MA 02205
USA. effective approach to design.
V. J. Thottuvelil is with Tyco Electronics Power Systems, Mesquite, TX Issues of modeling, simulation, and, more generally, com-
75149 USA. puter-aided design in power electronics have been addressed
G. C. Verghese is with the Massachusetts Institute of Technology, Cam-
bridge, MA 02139 USA (e-mail: [email protected]). in this and other journals in past years. The papers [7], [14]
Publisher Item Identifier S 0018-9219(01)04106-8. provide particularly valuable perspectives on these issues.

0018–9219/01$10.00 ©2001 IEEE

898 PROCEEDINGS OF THE IEEE, VOL. 89, NO. 6, JUNE 2001

Fig. 1. Boost PFC converter circuit.

Our emphasis on modeling in this paper complements the a diode bridge that rectifies the input voltage (into pul-
more detailed treatment of simulation in [7]. sating dc).
• The rectified line voltage is applied to the boost con-
verter, the basic elements of which are , , ,
B. Example: The Boost PFC Rectifier and . The capacitor serves to filter switching-fre-
quency ripple current at the input of the boost converter
To illustrate the challenges in modeling and simulating
from the ac input. The circuit elements , , , ,
power electronic circuits, we build around the example
, and form a turnoff snubber for diode , to re-
of a boost dc–dc converter in this paper. The basic boost
duce the reverse recovery current.
converter, intended to provide a voltage step-up function, is
The control loops described below are aimed at: 1)
embedded in applications ranging from power-factor-cor-
making the input current of the boost converter closely
rected (PFC) rectifier circuits (whose input current is made
track the shape of its input voltage—i.e., track a mul-
to follow the waveshape of the input voltage) to circuits
tiple of the input voltage—over the duration of each
that power the RF amplifiers in cell phones. This range of
rectified ac cycle and 2) regulating the output voltage to
applications illustrates an interesting characteristic of power
the desired value, by slowly adjusting this multiple over
electronic circuits: depending on the application, the same
several rectified ac cycles. The close current tracking at
basic circuit can be embellished with additional elements or
the input causes the boost converter to appear resistive
used with different control methods that provide additional
to the ac line input, thereby resulting in a power factor
functionality or work better at the power levels demanded
of (essentially) unity.
by the application. The particular details that are important
• The voltage-regulation loop is formed by , , ,
from a modeling and simulation perspective will vary
, , , and , which derive an error voltage
correspondingly, so this versatility presents a challenge.
from the fed back output voltage and a reference
The boost PFC rectifier circuit illustrates many of the chal- voltage. The output of serves as the “multiple”
lenges and opportunities to combine simulation with smart referred to above, and is fed to a module in which it
analysis, as well as the need for a hierarchical approach [15]. multiplies (a signal proportional to) the rectified input
Consider the particular version shown in Fig. 1. This circuit voltage, thereby creating the desired current reference
is intended to provide a nominal 400 Vdc regulated output signal at the output of the multiplier block.
from an ac voltage in the range of 85–264 Vac at the rectifier • The current-regulation loop is formed by current-sense
input. The circuit functions as follows. resistor , and , , , , , and , cre-
• The ac line input is fed through an electromagnetic in- ating the current error signal at the positive input of
terference (EMI) filter composed of , , and , to and deriving from it a modulating signal to be fed

MAKSIMOVIĆ et al.: MODELING AND SIMULATION OF POWER ELECTRONIC CONVERTERS 899

 to a pulsewidth modulator (PWM) circuit. The PWM
circuit compares the modulating signal with a clocked
ramp voltage to create a duty cycle signal that is used
to drive the switch .
The basic analysis tasks for power electronic circuits can
be outlined as follows in the context of the boost PFC rectifier
of Fig. 1.
• Provide the steady-state relationships between input
and output voltages and currents, as a function of the
circuit and control parameters. From this operating
point information, one obtains peak and rms voltage
and current stresses and calculates element dissipation
due to losses in the key circuit elements (switch
, diode , inductor , and capacitor ). This
information can typically be obtained with sufficient
accuracy for a first-pass design through analytical
approaches employing linear circuit models for each
of the circuit configurations that arises through action Fig. 2. Simplified circuit with parameters V = 160 V, L = 0.32
mH, C = 22 F, R = 200 , f = 50 kHz, V = 400 V.
of the switches.
• Provide switching waveforms to design/select the com-
ponents. This often requires a detailed determination these can be successively refined and extended to include
of the voltage and current waveforms associated with a additional circuit details and gain more insight into circuit
device. For example, the relevant stress levels used to operation. The methods used to systematically develop dy-
select a switch include peak switch current, peak switch namic models (switched, averaged, or sampled) for power
voltage, and power dissipation (which involves the av- converters are outlined in Section III, using the boost con-
erage value of the product of the voltage and current verter throughout as an example. Section IV describes how
waveforms). While simplified analysis can predict ide- to fold models such as those of Section III into simulation
alized waveforms, real circuits typically have parasitic tools that yield practically useful results in both the time
elements and nonideal behavior that require detailed and frequency domains. Section V contains a concluding
circuit simulation, not only to predict stress levels but discussion.
also to take into account the sometimes complicated re-
lationships among circuit elements.
• Obtain the dynamic characteristics of the circuit for two II. INITIAL ANALYSIS AND SIMULATION
key purposes: 1) to enable robust design of the control
loops that regulate the output voltage and shape the The first step in the analysis of the circuit in Fig. 1 is to
waveform of the input current and 2) to confirm that the obtain the voltage and current waveforms that describe basic
circuit has acceptable transient response, with output power-stage circuit operation. The input voltage to the boost
voltage and input current staying within specified limits portion of the PFC rectifier is continuously varying at a fre-
under changes of input voltage and load current as well quency equal to twice the line frequency, since it is derived
as during start-up and shutdown conditions. as the rectified ac line voltage. However, we can make the
• After the components have been selected and the assumption that within a switching cycle (which is typically
designer has gone through the physical realization 25 s or smaller), and indeed over several switching cycles,
process, usually via a board layout, predict circuit the input voltage is essentially constant.
performance under abnormal conditions. Usually, With this assumption, the basic steady-state analysis can
auxiliary circuits such as over-voltage, over-current, be obtained using the simplified circuit shown in Fig. 2,
and over-temperature detection and shutdown are where denotes from Fig. 1, denotes from the
incorporated into the circuit. In addition, conducted earlier figure, and denotes the load, modeled as being
and radiated EMI performance also must be assessed purely resistive. Note that many of the circuit elements in
for conformance to regulatory requirements; this is Fig. 1, such as those associated with the EMI filter, bridge
typically determined today by exhaustive experimental rectifier, snubber, and voltage- and current-feedback loops,
testing. have been eliminated in Fig. 2, so as to focus on the basic
power processing function of the circuit. Note also that
the switch has been replaced by a simple switch model
C. Outline
which is now controlled by a duty cycle pulse derived from
Section II of the paper describes the process by which a modulating signal ; this signal is compared with the
simple analysis and simulations are performed to understand sawtooth waveform at the input to the comparator IC in
the basic steady-state operation of a given circuit (in our case, order to establish the duty ratio of the converter (as discussed
the boost converter embedded in a PFC rectifier), and how in Section III).

900 PROCEEDINGS OF THE IEEE, VOL. 89, NO. 6, JUNE 2001

 core losses, and losses in the capacitor’s equivalent series re-
sistance (ESR). Switching losses in switch and diode
can also be estimated by constructing analytical approxima-
tions for these quantities over a single switching cycle, and
accumulating them to compute the loss over a rectified line
cycle.
The next level of circuit elaboration would be to add a
snubber circuit. To analyze the more elaborate circuit, we
need a model that can replicate effects such as diode re-
verse recovery, as well as a circuit simulator that can deal
with such refinements. Although some models and simula-
tors exist, in general, substantial investments in model con-
struction and simulation time are needed to get useful re-
sults. In particular, the effects of temperature on device phe-
nomena such as diode reverse recovery (and core losses in
the case of magnetic components) can be substantial and re-
quire a sophisticated simulation, with coupling between elec-
trical and thermal simulators and an accurate representation
of the packaging of the devices, to yield even partially useful
results. Due to these constraints, such detailed circuit simu-
lation is not commonly used in actual design (except when
undertaking failure analysis, particularly if experimental in-
vestigations have yielded little insight, or when ultrahigh re-
liability is needed, such as in space or military applications).
Rather, the practice is to simulate or analyze the effects of
diode turnoff with much simpler enhancements of the basic
diode model, e.g., having a switch in parallel with the basic
diode model, and closing it for a very short time in synchro-
nism with the turning off of the diode, to allow a reverse cur-
rent for a short duration. Using the results obtained with this
simple model, one can design the snubber circuit with sig-
nificant margin, and finally carry out additional tests on a
prototype (especially at elevated temperatures), with further
Fig. 3. Waveforms of simplified circuit. From top to bottom: adjustment of the snubber as needed.
voltage that drives the switch, voltage across the switch, and
currents through the inductor, switch, and diode.

III. LARGE-SIGNAL AVERAGED AND SAMPLED-DATA

Fig. 3 shows key steady-state circuit waveforms obtained
MODELS
from simulating the circuit in Fig. 2 using a simple switched-
circuit simulator of the general sort described in [1]–[12], for Elementary circuit modeling of a power converter
example. The waveforms of the simplified power-stage cir- typically produces detailed continuous-time nonlinear
cuit can also be obtained using straightforward circuit anal- time-varying models in state-space form. These models
ysis techniques. Simulations or computations (such as de- have rather low order, provided one makes approximations
termination of component stresses) at this modeling level that are reasonable from the viewpoint of control-oriented
can frequently be implemented simply and conveniently in modeling (as seen in the transition from Figs. 1 and 2):
a spreadsheet program such as Excel or using a general com- neglecting dynamics that occur at much higher frequencies
putational tool such as Mathcad or MATLAB. than the switching frequency (for instance, dynamics due
The quasi-static analysis that describes boost circuit opera- to parasitics or snubber elements, whose time scales are
tion at any one input-voltage level can—with sufficient accu- typically much shorter than the switching period), and
racy for most purposes—be extended to describe operation of focusing instead on components that are central to the power
the full PFC rectifier over a rectified line cycle simply by set- processing and control functions of the converter.
ting up the analysis with the boost input voltage as a variable. Such models capture essentially all the effects that are
Key quantities such as rms currents through the switch and likely to be significant for analysis of the basic power con-
diode , and ripple currents in and over a complete version function, but they are generally still too detailed and
rectified line cycle, can be accurately estimated using such an awkward to work with. The first challenge, therefore, is to
analysis. These results can be used to estimate dissipations extract from such a detailed model a simplified approximate
by calculating the conduction losses (which are products of model, preferably time-invariant, that is well matched to
the rms currents and on-state resistances or forward drops), the particular analysis or control task for the converter

MAKSIMOVIĆ et al.: MODELING AND SIMULATION OF POWER ELECTRONIC CONVERTERS 901

being considered. There are systematic ways to obtain such corresponds to the familiar aliasing effect associated with
simplifications, notably through averaging, which blurs out sampling. Our assumption for the averaged models below
the detailed switching artifacts, and sampled-data modeling, will be that is not allowed to change significantly
again to suppress the details internal to a switching cycle, within a single cycle, i.e., that is restricted to vary
focusing instead on cycle-to-cycle behavior. Both methods considerably more slowly than half the switching frequency.
can produce time-invariant but still nonlinear models. In the As a result, in the th cycle, so at
remainder of this section, and following the development any time yields the prevailing duty ratio [provided also that
in [16], we illustrate the preceding comments through a , of course—outside this range, the duty ratio
more detailed examination of the boost converter that was is 0 or 1]. Generalizations to rapid small-signal variations in
introduced in the previous section. Extensions to other can be found in [17]–[19], and a discussion of issues
converters can be made along similar lines. of aliasing under such rapid variations may be found in [20]
Boost Converter Operation: In typical operation of the and [21]. The modulating signal is usually generated
boost converter under what may be called constant-frequency by a feedback scheme, for instance, of the form shown by
PWM control, the switch in Fig. 2 is closed (or turned on) the inputs to the PWM in Fig. 1.
every seconds, and opened (or turned off) seconds
later in the th cycle, , so represents the A. Switched State-Space Models
duty ratio in the th cycle. If we maintain a positive inductor Choosing the inductor current and capacitor voltage
current, , then when the transistor is on, the diode as natural state variables, picking the resistor voltage
is off, and vice versa. This is referred to as the continuous as the output , and using the notation in Fig. 2, it is
conduction mode, and the waveforms in Fig. 3 correspond to easy to see that the following state-space model describes the
steady-state operation in this mode. In the discontinuous con- idealized boost converter in that figure:
duction mode, on the other hand, the inductor current drops
all the way to zero some time after the transistor is turned off,
and then remains at zero, with the transistor and diode both
off, until the transistor is turned on again. We focus on the
case of continuous conduction.
Let us mark the position of the switch using a switching (1)
function . When , the switch is closed; when
, the switch is open. The switching function Denoting the state vector by (where
may be thought of as (proportional to) the signal that has to be the prime indicates the transpose), we can rewrite the above
applied to the gate of the MOSFET in Fig. 1 to turn it on and equations as
off as desired. Under the constant-frequency PWM switching
discipline described above, jumps to 1 at the start of each
cycle, every seconds, and falls to 0 an interval later (2)
in its th cycle, as reflected in the top waveform in Fig. 3.
The average value of over the th cycle is therefore ; where the definitions of the various (boldfaced) matrices and
if the duty ratio is constant at the value , then is vectors are obvious from (1). We refer to this model as the
periodic, with average value . switched or instantaneous model, to distinguish it from the
In Fig. 2, corresponds to the signal at the output of the averaged and sampled-data models developed in later para-
comparator. The input to the “ ” terminal of the comparator graphs. (Similar state-space models are not hard to obtain for
is a sawtooth waveform of period that starts from 0 at the more elaborate, less idealized circuit models, for instance, in-
beginning of every cycle, and ramps up linearly to by the cluding capacitor ESR.)
end of the cycle. At some instant in the th cycle, this ramp If our compensator were to directly determine itself,
crosses the level of the modulating signal at the “ ” ter- rather than determining the modulating signal , then the
minal of the comparator. Hence, the output of the comparator above model would be the one of interest. It is indeed pos-
is set to 1 every seconds when the ramp restarts, and it re- sible to develop control schemes directly in the setting of the
sets to 0 later in the cycle, at time , when the ramp crosses switched model (2); see, for instance, [22]–[25], and refer-
. (In practice, the output of the comparator would ac- ences in those papers.
tually be used to trigger a latch, so that the switch does not For the design of more conventional feedback control
operate more than once each cycle.) The duty ratio of the compensation, we require a model describing the con-
signal thus ends up being in the cor- verter’s response to the modulating signal or the duty
responding switching cycle. By varying from cycle to ratio , rather than the response to the switching
cycle, the duty ratio can be varied. function . Augmenting the model (2) to represent the
Note that the samples of are what determine relation between and would introduce additional
the duty ratios. We would therefore obtain the same se- nonlinearity and time-varying behavior, leading to a model
quence of duty ratios even if we added to any signal that is hard to work with. The averaged and sampled-data
that stayed negative in the first part of each cycle and crossed models considered below are developed in response to this
up through 0 in the th cycle at the instant . This fact difficulty.

902 PROCEEDINGS OF THE IEEE, VOL. 89, NO. 6, JUNE 2001

B. State-Space Averaged Models switching converter operating with low ripple in the state
To design an analog feedback control scheme, we seek a variables. There are alternative sets of assumptions that lead
tractable model that relates the modulating signal or the to the same approximations.
duty ratio to the output voltage. In fact, since the With the approximations in (6), the description (5) be-
ripple in the instantaneous output voltage is made small by comes
design, and since the details of this small output ripple are not
of interest anyway in designing the feedback compensation,
what we really seek is a continuous-time dynamic model that
relates or to the local average of the output
voltage (where this average is computed over the switching
(7)
period). Also, recall that , the duty ratio, is the local
average value of in the corresponding switching cycle. What has happened, in effect, is that all the variables in the
These facts suggest that we should look for a dynamic model switched state-space model (1) have been replaced by their
that relates the local average of the switching function average values. In terms of the matrix notation in (2), and
to that of the output voltage . with defined as the local average of , we have
Specifically, let us define the local average of to be
the lagged running average

(8)
(3)
This continuous-time state-space model is referred to as
and call the continuous duty ratio. Note that the state-space averaged model, [26], [27]. The model is
, the actual duty ratio in the th cycle (defined as extending driven by the continuous-time control input —with the
from to ). If is periodic with period , then constraint , and assuming continuous conduc-
, the steady-state duty ratio. Our objective is to tion—and by the exogenous input . It is time-invariant
relate in (3) to the local average of the output voltage, with respect to this pair of inputs, linear with respect to
defined similarly by , and bilinear with respect to . [If is fixed
at a constant value , then the model is LTI.] Note that,
under our assumption of a slowly varying , we can
(4)
take ; with this substitution, (8) becomes
an averaged model whose control input is the modulating
A natural approach to obtaining a model relating these aver- signal , as desired. The use and interpretation of this
ages is to take the local average of the state-space descrip- model should be restricted to frequencies significantly
tion in (1). The local average of the derivative of a signal below half the switching frequency; converter dynamics up
equals the derivative of its local average, because of the linear to around one-tenth the switching frequency are generally
time-invariant (LTI) nature of the local averaging operation well captured by the averaged model.
we have defined. The result of averaging the model (1) is The averaged model (8) can be used to solve for
therefore the following set of equations: steady-state or operating point relations obtained with con-
stant and (by setting the derivatives
to zero). It also leads to much more efficient simulations of
converter dynamic behavior than those obtained using the
switched model (2), provided only local averages of vari-
ables are of interest; the simulation can take larger time steps
(5) because it no longer needs to track the switching-frequency
ripple. This averaged model also forms a convenient starting
where the overbars again denote local averages. point for various nonlinear control design approaches; see,
The terms that prevent the above description from being for instance, [28]–[30], and references in those papers. More
a state-space model are and ; the average of a traditional small-signal control design to regulate operation
product is generally not the product of the averages. Under in the neighborhood of a fixed operating point can be based
reasonable assumptions, however, we can write on the corresponding LTI linearization of the averaged
model. Section IV has further discussion of these issues.
Current-Mode Control: The preceding averaged model
(6) can also be easily modified to approximately represent the
dynamics of a high-frequency PWM converter operated
One set of assumptions leading to the above simplification under so-called current-mode control [31]. The name comes
requires and over the averaging interval from the fact that a fast inner loop regulates the inductor
to not deviate significantly from and , re- current to a reference value, while the slower outer loop
spectively. This condition is reasonable for a high-frequency adjusts the current reference to correct for deviations of the

MAKSIMOVIĆ et al.: MODELING AND SIMULATION OF POWER ELECTRONIC CONVERTERS 903

Fig. 4. Nonlinear averaged circuit model of the boost converter.

output voltage from its desired value. Control of a PFC rec- number of different types of converters and switch elements,
tifier can be implemented on this basis as well. The current including phase-controlled rectifiers, pulsewidth modulated
monitoring and limiting that are intrinsic to current-mode converters in continuous or discontinuous conduction mode,
control are among its attractive features. resonant-switch converters, and so on. Because of its gen-
In constant-frequency peak-current-mode control, the erality and the ease with which the resulting models are
transistor is turned on every seconds, as before, but is simulated in standard circuit simulators such as SPICE or
turned off when the inductor current (or equivalently, the SABER, there has been a recent resurgence of interest in
transistor current) reaches a specified reference or peak circuit averaging of switched networks; see [38]–[50] and
level, denoted by . The duty ratio, rather than being further references in Sections III-D and IV.
explicitly commanded via a modulating signal such as The first step in circuit averaging is to replace all voltages
in Fig. 2, is now implicitly determined by the inductor and currents by their (running) averages. The resulting quan-
current’s relation to . (Instead of constant-frequency tities still respect Kirchhoff’s laws, and therefore constitute
control, one could use hysteretic or other schemes to confine valid circuit variables. All LTI components of the original cir-
the inductor current to the vicinity of the reference current.) cuit impose the same constraints on the averaged quantities
A tractable and reasonably accurate continuous-time as they do on the original instantaneous variables, and there-
model for the dynamics of the outer loop is obtained by fore remain the same in the averaged circuit. The switching
assuming that the average inductor current is approximately elements of the original circuit need to be handled differ-
equal to the reference current ently, however. If we represent the switching elements in the
original circuit as appropriately controlled voltage or current
(9) sources, then these can be circuit-averaged as well, but with
some approximations to convert the control relationships to
and then making the substitution (9) in (7) to eliminate . ones involving only averaged quantities.
The result is the following first-order nonlinear time-in- As an example, consider the ideal boost dc–dc converter
variant model: of Fig. 2. The diode can be replaced by a controlled cur-
rent source of value and the switch by a controlled
(10) voltage source of value , where .
By averaging these relationships and making the same ap-
This model is simple enough that one can use it for simu- proximations as in (6), we obtain
lations and to explore various nonlinear control possibilities (11)
for adjusting to control or ; a linearized
version of this equation can be used to design small-signal (12)
controllers for perturbations around a fixed operating point. where . These are the terminal relations that
More continues to be written on averaged models in power characterize the averaged switching elements. The averaged
electronics (see, e.g., [32]–[36]) as well as further references circuit model of Fig. 4 is the result of the above process. This
in Sections III-C, III-D, and IV. is a large-signal nonlinear, but time-invariant circuit model.
Not surprisingly, it is governed by the state-space averaged
C. Circuit Averaging and the Averaged Switch model in (7), but we have not had to derive a state-space
Instead of averaging the converter state equations, we model for each configuration in order to obtain the model.
could directly average the characteristics or waveforms Linearization of this circuit (as discussed in Section IV)
associated with each of the components in the converter yields small-signal models that are suited to conventional
[37]. This circuit averaging approach is widely used (al- feedback control design.
though sometimes in implicit rather than explicit ways). It should be noted that the definition of the switch net-
Because manipulations are performed on the circuit diagram work and its dependent variables is not unique. Different def-
instead of on its equations, the circuit averaging technique initions lead to equivalent but not identical averaged circuit
often gives a more physical interpretation to the model. models; some choices may be better suited than others to any
The circuit averaging technique can be applied directly to a particular analysis task.

904 PROCEEDINGS OF THE IEEE, VOL. 89, NO. 6, JUNE 2001

D. Generalized Averaging and Dynamic Phasors
The voltages and currents in power electronic converters
and electrical drives are typically periodic in steady state,
and often nonsinusoidal. The dynamics of interest for anal-
ysis and control are often those of deviations from periodic
behavior, for instance as manifested in deviations of the en-
velope of a quasi-sinusoidal waveform from its steady-state
value. For analysis of the steady state, one has familiar phasor
or harmonic or describing function methods [18], [51]–[53].
The analytical approach reviewed here is aimed at system-
atic derivation of phasor dynamics, from which the dynamic
behavior of the original waveform or its envelope can effi-
ciently be deduced. (A distinct approach to the notion of en-
velope following, directly implemented in a simulation set-
ting, may be found in [54].) The idea of deriving dynamical Fig. 5. Circuit schematic of a series resonant converter (typical
models for Fourier coefficients goes back to classical aver- parameters: v = 3.3 V, I = 1 A, R = 5 , switching frequency
aging (see [32], [34] and references therein). The recent in- above 38 kHz).
terest in these approaches for power electronics was sparked
by [49] and [55], which applied the approach to series res- tion is time-invariant, the standard definition of phasors
onant and switched mode dc–dc converters (also see [56] from circuit theory is recovered. A key property is that the
for series resonant converters); the approach taken in [49] derivative of the th Fourier coefficient is given by the fol-
involved direct circuit averaging. Some extensions may be lowing expression:
found in [57].
The generalized averaging that we perform to obtain our (15)
models is based on the observation [55] that a (possibly com-
plex) time-domain waveform can be represented on the This formula is easily verified using (13) and (14), and inte-
interval using a Fourier series of the form gration by parts.
The definitions given in (13) and (14) can also be gener-
alized for the analysis of polyphase systems, with the defi-
(13)
nition of dynamic positive-sequence, negative-sequence and
zero-sequence symmetric components at frequency ; see
where and are the complex Fourier co- [59].
efficients, which we shall also refer to as phasors. These The application of the above phasor calculus to obtaining
Fourier coefficients are functions of time since the interval an averaged model proceeds just as with state-space aver-
under consideration slides as a function of time. We are in- aging (and limiting attention to the zeroth-order phasor
terested in cases where a few coefficients suffice to provide actually recovers traditional state-space averaging). One be-
a good approximation of the original waveform, and where gins with a standard state-space description of the instan-
those coefficients vary slowly with time. taneous (switched) variables, then averages both sides, in-
The th coefficient (or -phasor) at time is determined voking the properties of dynamic phasors as needed. The next
by the following averaging operation: step is to make approximations that allow the averaged model
itself to be written in state-space form, using the dynamic
phasors as state variables. The slow variation of the phasors
(14)
is usually one of the critical assumptions in making reason-
able approximations.
The notation will be used to denote the averaging 1) Example: Resonant Converter: As an example of the
operation in (14). Our analysis aims to provide a dynamic application of generalized averaging, consider the series res-
model for the dominant Fourier series coefficients as the onant dc–dc converter shown in Fig. 5. Using the notation
window of length slides over the waveforms of interest. given in the figure, a state-space model can be written as
More specifically, we aim to obtain a state-space model in
which the coefficients in (14) are the state variables.
When the original waveforms are complex-valued, the
phasor equals (where is the complex
conjugate of ). Complex-valued waveforms arise, for in-
stance, when using complex space vectors [58] in dynamical (16)
descriptions of electrical drives. In the case of real-valued
time-domain quantities, and , so where denotes the switching frequency in rad/s, and
(13) can be rewritten as a one-sided summation involving are the instantaneous resonant tank voltage and current re-
twice the real parts of for positive . If in addi- spectively, is the instantaneous output voltage, and the

MAKSIMOVIĆ et al.: MODELING AND SIMULATION OF POWER ELECTRONIC CONVERTERS 905

load comprises a resistor in parallel with a current sink We illustrate how a sampled-data model may be obtained
(we have dropped the time argument from the variables , for our boost converter example. The state evolution of (1),
, and to avoid notational clutter). The “ ” in the above (2) for each of the two possible values of can be de-
equations is 1, the sign being that of its argument. scribed very easily using the standard matrix exponential ex-
To derive a dynamical phasor model corresponding to pressions for LTI systems, and the trajectories in each seg-
(16), it is assumed that both and are described with ment can then be pieced together by invoking the continuity
sufficient accuracy by their respective fundamental ( ) of the state variables. Recall that the matrix exponential can
components (with corresponding phasors , taken to have be defined, just as in the scalar case, by the (very well be-
angle 0, and , respectively), while is assumed haved) infinite matrix series
to be slowly varying, hence well described by its
component, or local average. These assumptions are rea- (18)
sonable in well-designed dc–dc series resonant converters.
from which it is evident that
Then the following dynamic phasor model can be derived
from (16) using (15): (19)

Under the switching discipline of constant-frequency PWM,

where for the initial fraction of the th switching
cycle, and for the rest of the cycle, and assuming
the input voltage is constant at , we find
(17)

(20)
(We have again dropped the time argument from ,
and .) This model can be written in the form of a fifth- where
order model involving real-valued quantities, for example,
by taking real and imaginary parts of the first two equations.
It turns out that the dynamic phasor model approximates the
switched model very closely, as shown in [55]. Control ex-
plorations using this model can be found in, e.g., [60] and (21)
[61].
Dynamic phasors can be used to obtain models with For a well-designed high-frequency PWM dc–dc con-
varying degrees of detail; for example, both the dc compo- verter in continuous conduction, the state trajectories in
nent and the fundamental switching-frequency component each switch configuration are close to linear, because the
were used to describe a boost converter in [57]. Dynamic switching frequency is much higher than the filter cutoff
phasors have been used very naturally and effectively frequency. What this implies is that the matrix exponentials
for a variety of power electronic converters of interest in (20) are well approximated by just the first two terms in
in high-power transmission systems. These flexible ac their Taylor series expansions
transmission system (FACTS) applications include the
thyristor-controlled series capacitor (TCSC) described in
[62], [63], and an unbalanced unified power flow controller (22)
(UPFC) that utilizes polyphase dynamic phasors, treated in
If we use these approximations in (20) and neglect terms
[64]. Application to unbalanced three-phase machines can
in , the result is the following approximate sampled-data
be found in [59]. The notion of a dynamic phasor can be
model:
of use in power systems even when no power electronics is
involved; see [65] and [66].
(23)
E. Sampled-Data Models
Sampled-data models are naturally matched to power elec- This model is easily recognized as the usual forward-Euler
tronic converters, firstly because of the cyclic way in which approximation of the continuous-time model in (8), obtained
power converters are operated and controlled, and secondly by replacing the derivative there by a forward difference.
because such models are well suited to the design of dig- (Retaining the terms in leads to more refined, but still very
ital controllers, which are increasingly used in power elec- simple, sampled-data models.) For an example of the use
tronics. Like averaged models, sampled-data models allow us in simulation of sampled-data and continuous-time models
to focus on cycle-to-cycle behavior, ignoring details of the in- based on this sort of approximation, see [67] and [68].
tracycle behavior. This makes them effective in studying and The sampled-data models in (20) and (23) were derived
controlling ripple instabilities (i.e., instabilities at half the from (1), (2), and therefore used samples of the natural state
switching frequency), and also in general simulation, anal- variables, and , as state variables. However, other
ysis, and design. choices are certainly possible, and may be more appropriate

906 PROCEEDINGS OF THE IEEE, VOL. 89, NO. 6, JUNE 2001

for a particular implementation. For instance, we could re- and small-signal frequency-response characteristics.
place by , i.e., the sampled local average of With simple device models, and ignoring details of
the capacitor voltage. switching transitions, simulations over many switching
An early reference on sampled-data models in power elec- cycles can be completed efficiently, using general-pur-
tronics is [69]. For more on sampled-data models, see [45] pose circuit simulators or specialized simulators that
and references there, and also, e.g., [70]–[72]. In particular, are developed to support fast transient simulation
[45] derives a sampled-data model for the boost PFC (but based on idealized, piecewise-linear device models,
sampling at the period of the rectified ac voltage rather than or based on a combination of piecewise-linear and
the switching period of the boost converter), and uses it to de- nonlinear models (see [1]–[12]).
sign a discrete-time feedback controller (with time constant • Averaged models are well suited for prediction of
on the order of the period of the ac input). converter steady-state and dynamic responses. These
models are essential design tools because they provide
physical insight and lead to analytical results that can
IV. SIMULATION OF SWITCHED AND AVERAGED DYNAMIC
be used in the design process to select component and
MODELS
controller parameter values for a given set of specifi-
In the design verification of power electronic systems by cations. A large-signal averaged circuit model, such as
simulation, it is often necessary to use component and system the model in Fig. 4, is very convenient for application
models of various levels of complexity. This section, which with general-purpose circuit simulators such as SPICE
elaborates on some of the issues raised in Sections I and II, or SABER. Simulations of averaged circuit models
is focused on switched and averaged models, although sam- can be performed to test for losses (apart from those
pled-data models have their particular role as well, especially due to switching) and efficiency, steady-state voltages
in careful stability studies and in control design for digital and currents, stability, and large-signal transient re-
controllers. sponses. Since switching transitions and ripples are
• Detailed, complex models that attempt to accurately removed by averaging, simulations over long time
represent the physical behavior of devices are neces- intervals and over many sets of parameter values can
sary for tasks that involve finding switching times, be completed efficiently. Therefore, averaged models
details of switching transitions and switching loss are also well suited for simulations of large electronic
mechanisms, or instantaneous voltage and current systems that include multiple switching converters
stresses. Component vendors often provide libraries [73]. Furthermore, although large-signal averaged
of such device models for use with general-purpose models are nonlinear, they are time-invariant and can
circuit simulators such as SPICE or SABER. To com- be linearized about any constant operating condition
plete a detailed circuit model, one must also carefully to produce LTI small-signal models, from which one
examine effects of packaging and board interconnects. can generate various frequency responses of interest
With fast-switching power semiconductors, simulation (see Section IV-B). References on averaged converter
time steps corresponding to a few nanoseconds or less modeling for simulation include [74]–[81].
may be required, especially during ON–OFF switching
transitions. Because of the complexity of detailed A. Transient Response Analysis
device models and the fine time resolution, the simu- In the design of control loops around converters, it is
lation tasks can be very time consuming. In practice, often necessary to perform transient simulations over many
time-domain simulations using detailed device models switching cycles. For example, in dc voltage regulator
are usually performed only on selected parts of the designs, it is necessary to verify whether the output voltage
system, and over short time intervals involving a few remains within specified limits when the load current takes
switching cycles. a step change. In the boost PFC rectifier of Fig. 1, transient
• Since an ON–OFF switching transition usually takes simulations can be used to determine current harmonic
only a small fraction of a switching cycle, the basic distortion, component stresses during start-up or load tran-
operation of switching power converters can be ex- sients, and so on. Such simulations can be performed on a
plained using simplified, idealized device models. For switching circuit model using a switched-circuit simulator
example, a MOSFET can be modeled as a switch with or a general-purpose simulator, or on the converter averaged
a small (ideally zero) resistance when on, and a very model, or using a sampled-data model.
large resistance (ideally an open circuit) when off. As an example, let us apply the first two approaches
Such simplified models yield physical insight into the to investigate a transient response of the boost converter
basic operation of switching power converters, and shown in Fig. 2 due to a step change in the switch duty
provide the starting point for the development of the cycle. Fig. 6 shows the inductor current and the capacitor
analytical models described earlier. Simplified device voltage waveforms during the transient. The waveforms
models are also useful for time-domain simulations obtained by switched-circuit transient simulation are shown
aimed at determining or verifying converter and together with the waveforms obtained by simulation of the
controller operation, switching ripples, current and averaged circuit model in Fig. 4. The converter transient
voltage stresses, responses to load or input transients, response is governed by the natural time constants of the

MAKSIMOVIĆ et al.: MODELING AND SIMULATION OF POWER ELECTRONIC CONVERTERS 907

 and numerical approaches for such problems may be found
in [87], which also examines the modeling and simulation of
more exotic phenomena such as chaos in power electronics.
Circuit averaging leads to a nonlinear, time-invariant
circuit model, as illustrated by the example shown in Fig. 4.
Both steady-state computations and the construction of
small-signal models are easily carried out with averaged
circuits. As an example, Fig. 7 shows the steady-state dc and
small-signal ac circuit model obtained by standard lineariza-
tion of the nonlinear controlled sources in Fig. 4 around a
steady-state operating point. This circuit model includes
an ideal transformer that explicitly illustrates the major
(a)
features of the boost dc–dc converter, namely a dc
conversion ratio (where ), small-signal natural
time constants determined by energy storage components, as
well as effects of duty cycle variations through the sources
and , where , the duty-ratio deviation from
steady state. This circuit model can be easily solved for
transfer functions of interest for classical controller design
based on the LTI model, including control-to-output and
line-to-output responses, as well as the output impedance.
Linearized averaged models are also the starting point for
the modeling and stability analysis of paralleled converters
(see [46], [88], and [89]).
Frequency responses of interest can alternatively be ob-
(b)
tained by appropriate time-domain simulations of switched-
Fig. 6. Transient waveforms in the boost converter example, for circuit models (see [90]–[93]), or by ac simulations of non-
i (t) and v (t). The duty cycle is increased from d = 0.55 to
d = 0.6 at t = 0.5 ms. linear averaged circuit models (see [50], [74]–[80]). As an
example, Fig. 8 shows magnitude and phase responses of
the boost control-to-output transfer function
converter. Since these time constants are much longer than (where is the perturbation in output voltage), obtained by
the switching period, the converter transient responses take ac simulation of the model in Fig. 4.
many switching cycles to reach a new steady state. In the
results obtained by simulation of the averaged circuit model,
the switching ripples are removed, but the low-frequency V. CONCLUDING DISCUSSION
portions of the converter transient responses match very First, an important disclaimer. Although we have cited sev-
closely the responses obtained by switched-circuit simula-
eral relevant references, there are at least as many other ones
tion. (Note that the converter goes through an interval in the
that we have not. The references listed here are intended to
discontinuous conduction mode, from around 1.2 to 2 ms. serve as pointers for the interested reader, and will quickly
An appropriate averaged-switch model can be derived to lead to much more that is likely to be useful.
handle this transition to discontinuous conduction and back;
Hierarchical approaches, using a variety of layered models
see [50].)
and simulations, form the basic strategy used today to an-
alyze and design power electronic circuits. Proceeding up
B. Steady-State and Small-Signal Analysis the hierarchy typically involves modeling individual mod-
There are many numerical/simulation approaches to deter- ules or portions of the circuit in a more aggregated or ab-
mining the steady state of a switched model (see [82]–[86] stracted form, allowing larger portions of the circuit or of a
and references therein, for example). Small-signal models, system with multiple circuits to be simulated in reasonable
and particularly sampled-data small-signal models, can now times with adequate accuracy.
be constructed to represent small deviations from this steady Switched-circuit and averaged simulators have also
state. proven to be very valuable in the synthesis of new power
A designer is also often interested in determining the electronic circuits. Generally, there are large numbers of pos-
boundaries in the space of parameters (such as input voltage sible combinations of switches and passive elements that can
amplitude, frequency, load resistance or current, and so on) be combined to create new circuit topologies. Simulation of
that mark transitions from one steady-state operating mode these topologies remains a key tool in comparing topologies
to another in a switched circuit. Of complementary interest is for an application, discovering problems in a new circuit
the determination of stability domains in the state space for or control approach, trying out variations to overcome each
particular operating modes, i.e., the sets of initial conditions successively discovered hurdle, and then refining the circuit
that respectively converge to these operating modes. Models or controller to meet performance requirements. The ability

908 PROCEEDINGS OF THE IEEE, VOL. 89, NO. 6, JUNE 2001

Fig. 7. DC and small-signal ac averaged circuit model of the boost converter.

istics. Although a few papers and simulators have re-

ported some work in this area [96]–[98], they still re-
main highly specialized activities requiring major in-
vestments in time to set up the models and simulations,
leading to practical use only in a few cases.
• Difficulties in getting reliable convergence of simula-
tors also remains an ongoing problem and a source of
frustration for power electronic circuit designers.
In addition to circuit simulators and analytical methods
such as averaging, control-system-oriented tools such as
MATLAB are also used today. They are used in much the
same way as for other control-system analysis problems,
with the exception that an appropriate circuit-averaged or
Fig. 8. Magnitude and phase responses of the control-to-output sampled-data model is used for the switching power stage
transfer function G (s) = v^=d^ of the boost dc–dc converter. and portions of the control circuitry that are not continuous.
Another area with increased activity recently is in de-
in a simulator to build models of increasing complexity, riving models that are related to the physical geometry for
starting from very idealized models, provides a strong tool such components as magnetics and printed wiring boards,
for the power electronics design engineer exploring a new [99]–[101]. Results from these analyses may be used to
design concept, in essence by using the computer tool to help derive circuit models that can then be combined with other
understand how a new circuit works (or does not work). The elements in a circuit simulator.
development of approaches to rapidly determining the cyclic Many of the methods described above can be folded into
steady-state sequence of a switching circuit has meant that a framework that assesses circuit performance with variation
long simulations to reach steady state are avoided, leading of circuit parameters or operating conditions. These include
to significant increases in design engineer productivity. worst case analysis (see [102] and references therein), and
Although general-purpose circuit simulators such as yield analysis using Monte Carlo or similar techniques. Most
SPICE are increasingly being used for power electronics commercially available simulators used in power electronics
simulation, they are still beset with problems when used include such capabilities.
to simulated detailed device behavior. These include the In summary, analysis of power electronic systems requires
following. multiple methods and tools to understand circuit operation
• Paucity of accurate device models for generally and obtain enough information to achieve a robust design.
This is no different than in other fields of electronics such
available commercial devices, especially for power
as digital systems where hierarchical approaches and mul-
switching devices such as diodes, MOSFETs, thyris-
tiple tools are routinely used. A key difference, valid even
tors, and IGBTs. Although there have been numerous
papers on modeling power semiconductor devices [94], today, is that experimental methods are still practical, effec-
tive, and heavily relied upon in power electronics. Neverthe-
[95], the models remain difficult to develop and involve
less, the widespread availability of inexpensive computing
somewhat large investments in time and equipment
and the refinement of simulation tools and techniques over
to build and validate. Also, power electronic design
engineers, who are generally focused on circuit-level the last decade have allowed us to come closer to the day
when a complete power electronic circuit can be simulated
design, rarely have a deep enough understanding of
and studied in software, then built with high confidence that
device behavior to understand the structure of the
it will work right the first time.
models or how to change model parameters to reflect
different devices. These barriers remain as significant
ACKNOWLEDGMENT
obstacles to the use of detailed circuit simulations in
power electronics for the future. The authors are grateful to an anonymous reviewer, and to
• Inadequate understanding and modeling of the role that V. Caliskan, G. Escobar, S. Leeb, and D. Perreault for com-
thermal effects play in changing electrical character- ments that helped improve the paper.

MAKSIMOVIĆ et al.: MODELING AND SIMULATION OF POWER ELECTRONIC CONVERTERS 909

REFERENCES [25] L. Malesani, L. Rossetto, G. Spiazzi, and P. Tenti, “Performance op-
timization of Ćuk converters by sliding-mode control,” IEEE Trans.
Power Electron., vol. 10, pp. 302–309, 1995.
[1] R. J. Dirkman, “The simulation of general circuits containing ideal [26] R. D. Middlebrook and S. Ćuk, “A general unified approach to mod-
switches,” in IEEE Power Electronics Specialists Conf. (PESC), eling switching converter power stages,” in IEEE Power Electronics
1987, pp. 185–194. Specialists Conf. (PESC), 1976, pp. 18–34.
[2] C. J. Hsiao, R. B. Ridley, H. Naitoh, and F. C. Lee, “Circuit-oriented [27] , “A general unified approach to modeling switching-converter
discrete-time modeling and simulation of switching converters,” power stages,” Int. J. Electron., vol. 42, pp. 521–550, June 1977.
in IEEE Power Electronics Specialists Conf. (PESC), 1987, pp. [28] S. R. Sanders and G. C. Verghese, “Lyapunov-based control for
167–176. switched power converters,” IEEE Trans. Power Electron., vol. 7,
[3] R. C. Wong, H. A. Owen, and T. G. Wilson, “An efficient algorithm pp. 17–24, 1992.
for the time-domain simulation of regulated energy-storage dc-to-dc [29] H. Sira-Ramírez and M. T. Prada-Rizzo, “Nonlinear feedback reg-
converters,” IEEE Trans. Power Electron., vol. 2, pp. 154–168, Apr. ulator design for the Ćuk converter,” IEEE Trans. Automat. Contr.,
1987. vol. 37, pp. 1173–1180, 1992.
[4] V. Rajagopalan, Computer-Aided Analysis of Power Electronic Sys- [30] G. Escobar, R. Ortega, H. Sira-Ramirez, J. P. Vilain, and I. Zein, “An
tems. New York: Marcel Dekker, 1987. experimental comparison of several nonlinear controllers for power
[5] A. M. Luciano and A. G. M. Strollo, “A fast time-domain algorithm converters,” IEEE Control Syst. Mag., vol. 19, pp. 66–82, 1999.
for simulation of switching power converters,” IEEE Trans. Power [31] S. P. Hsu, A. Brown, L. Resnick, and R. D. Middlebrook, “Modeling
Electron., vol. 2, pp. 363–370, July 1990. and analysis of switching dc-to-dc converters in constant-frequency
[6] D. Bedrosian and J. Vlach, “Time-domain analysis of networks with current-programmed mode,” in IEEE Power Electronics Specialists
internally controlled switches,” IEEE Trans. Circuits Syst. I, vol. 39, Conf. (PESC), 1979, pp. 284–301.
pp. 199–212, Mar. 1992. [32] P. T. Krein, J. Bentsman, R. M. Bass, and B. C. Lesieutre, “On the
[7] N. Mohan, W. P. Robbins, T. M. Undeland, R. Nilssen, and O. Mo, use of averaging for the analysis of power electronic systems,” IEEE
“Simulation of power electronics and motion control systems—An Trans. Power Electron., vol. 5, pp. 182–190, Apr. 1990.
overview,” Proc. IEEE, vol. 82, pp. 1287–1302, Aug. 1994. [33] S. Ben-Yaakov, D. Wulich, and W. M. Polivka, “Resolution of an av-
[8] N. Mohan, T. M. Undeland, and W. P. Robbins, Power Electronics: eraging paradox in the analysis of switched-mode dc–dc converters,”
Converters, Applications, and Design, 2nd ed. New York: Wiley, IEEE Trans. Aerosp. Electron. Syst., vol. 30, pp. 626–632, Apr. 1994.
1995. [34] B. Lehman and R. M. Bass, “Switching frequency dependent aver-
[9] P. Pejović and D. Maksimović, “A new algorithm for simulation aged models for PWM dc–dc converters,” IEEE Trans. Power Elec-
of power electronic systems using piecewise-linear device models,” tron., vol. 11, pp. 89–98, Jan. 1996.
IEEE Trans. Power Electron., vol. 10, pp. 340–348, May 1995. [35] J. Sun and H. Grotstollen, “Symbolic analysis methods for averaged
[10] P. Pejović, “A method for simulation of power electronic systems modeling of switching power converters,” IEEE Trans. Power Elec-
using piecewise-linear device models,” Ph.D. dissertation, Univ. tron., vol. 12, May 1997.
Colorado, Boulder, Apr. 1995. [36] J. Sun, D. M. Mitchell, M. Greuel, P. T. Krein, and R. M. Bass,
[11] D. Li, R. Tymerski, and T. Ninomiya, “PECS: An efficacious so- “Averaged modeling of PWM converters in discontinuous conduc-
lution for simulating switched networks with nonlinear elements,” tion mode: A reexamination,” in IEEE Power Electronics Specialists
in IEEE Power Electronics Specialists Conf. (PESC), 2000, pp. Conf. (PESC), 1998, pp. 615–622.
274–279. [37] G. W. Wester and R. D. Middlebrook, “Low-frequency characteriza-
[12] S. M. Sandler, SMPS Simulation with SPICE3. New York: Mc- tion of switched dc–dc converters,” IEEE Trans. Aerosp. Electron.
Graw-Hill, 1997. Syst., vol. AES-9, pp. 376–385, May 1973.
[13] S. B. Leeb, J. L. Kirtley, and G. C. Verghese, “Recognition of dy- [38] Y. S. Lee, “A systematic and unified approach to modeling switches
namic patterns in dc–dc switching converters,” IEEE Trans. Power in switch-mode power supplies,” IEEE Trans. Ind. Electron., vol.
Electron., vol. 6, pp. 296–302, Apr. 1991. IE-32, pp. 445–448, Nov. 1985.
[14] T. G. Wilson, Jr., “Life after the schematic: The impact of circuit [39] R. Tymerski and V. Vorperian, “Generation, classification and anal-
operation on the physical realization of electronic power supplies,” ysis of switched-mode dc-to-dc converters by the use of converter
Proc. IEEE, vol. 76, pp. 325–334, Apr. 1988. cells,” in Proc. Int. Telecommunications Energy Conf. (INTELEC),
[15] V. J. Thottuvelil, D. Chin, and G. C. Verghese, “Hierarchical ap- Oct. 1986, pp. 181–195.
proaches to modeling high-power-factor ac–dc converters,” IEEE [40] S. Freeland and R. D. Middlebrook, “A unified analysis of con-
Trans. Power Electron., vol. 6, pp. 179–187, Mar. 1991. verters with resonant switches,” in IEEE Power Electronics Special-
[16] G. C. Verghese, “Dynamic modeling and control in power elec- ists Conf. (PESC), 1987, pp. 20–30.
tronics,” in The Control Handbook, W. S. Levine, Ed. Boca Raton, [41] R. Tymerski, V. Vorperian, F. C. Lee, and W. T. Baumann, “Non-
FL: CRC Press—IEEE Press, 1996, pp. 1413–1423. linear modeling of the PWM switch,” in IEEE Power Electronics
[17] B. Y. Lau and R. D. Middlebrook, “Small-signal frequency response Specialists Conf. (PESC), 1988, pp. 968–976.
theory for piecewise-constant two-switched-network dc-to-dc [42] V. Vorperian, R. Tymerski, and F. C. Lee, “Equivalent circuit models
converter systems,” in IEEE Power Electronics Specialists Conf. for resonant and PWM switches,” IEEE Trans. Power Electron., vol.
(PESC), 1986, pp. 186–200. 4, pp. 205–214, Apr. 1989.
[18] J. Groves, “Small-signal analysis using harmonic balance methods,” [43] A. Witulski and R. Erickson, “Extension of state-space averaging to
in IEEE Power Electronics Specialists Conf. (PESC), 1991, pp. resonant switches and beyond,” IEEE Trans. Power Electron., vol.
74–79. 5, pp. 98–109, Jan. 1990.
[19] R. Tymerski, “Application of time varying transfer function for exact [44] V. Vorperian, “Simplified analysis of PWM converters using the
small-signal analysis,” in IEEE Power Electronics Specialists Conf. model of the PWM switch: Parts I and II,” IEEE Trans. Aerosp.
(PESC), 1991, pp. 80–87. Electron. Syst., vol. 26, pp. 490–505, May 1990.
[20] D. Perreault and G. C. Verghese, “Time-varying effects in models [45] J. G. Kassakian, M. F. Schlecht, and G. C. Verghese, Principles of
for current-mode control,” IEEE Trans. Power Electron., vol. 12, pp. Power Electronics. Reading, MA: Addison-Wesley, 1991.
453–461, May 1997. [46] A. Kislovski, R. Redl, and N. Sokal, Dynamic Analysis of
[21] G. C. Verghese and V. J. Thottuvelil, “Aliasing effects in PWM Switching-Mode DC/DC Converters. New York: Van Nos-
power converters,” in IEEE Power Electronics Specialists Conf. trand-Reinhold, 1991.
(PESC), 1999, pp. 1043–1049. [47] D. Maksimović and S. Ćuk, “A unified analysis of PWM converters
[22] W. W. Burns and T. G. Wilson, “Analytic derivation and evaluation in discontinuous modes,” IEEE Trans. Power Electron., vol. 6, pp.
of state trajectory control law for dc–dc converters,” in IEEE Power 476–490, July 1991.
Electronics Specialists Conf. (PESC), 1977, pp. 70–85. [48] S. R. Sanders and G. C. Verghese, “Synthesis of averaged circuit
[23] H. Sira-Ramírez, “Sliding motions in bilinear switched networks,” models for switched power converters,” IEEE Trans. Circuits Syst.,
IEEE Trans. Circuits Syst., vol. 34, pp. 919–933, 1987. vol. 38, pp. 905–915, Aug. 1991.
[24] S. R. Sanders, G. C. Verghese, and D. E. Cameron, “Nonlinear con- [49] J. M. Noworolski and S. R. Sanders, “Generalized in-place aver-
trol of switching power converters,” Control Theory Adv. Technol., aging,” in IEEE Applied Power Electronics Conf. (APEC), 1991, pp.
vol. 5, pp. 601–627, 1989. 445–451.

910 PROCEEDINGS OF THE IEEE, VOL. 89, NO. 6, JUNE 2001

 [50] R. W. Erickson and D. Maksimovic, Fundamentals of Power Elec- [75] , “Using the SPICE2 CAD package for easy simulation of
tronics, 2nd ed. Dordrecht, The Netherlands: Kluwer, 2001. switching regulators in both continuous and discontinuous conduc-
[51] A. Gelb and W. E. van der Velde, Multiple Input Describing Func- tion modes,” in Proc. 8th Nat. Solid-State Power Conversion Conf.
tions and Nonlinear Systems Design. New York: McGraw-Hill, (Powercon 8), Apr. 1981.
1968. [76] , “Using the SPICE2 CAD package to simulate and design the
[52] A. I. Mees and A. R. Bergen, “Describing function revisited,” IEEE current mode converter,” in Proc. 11th Nat. Solid-State Power Con-
Trans. Automat. Contr., vol. 20, pp. 473–478, Sept. 1975. version Conf. (Powercon 11), Apr. 1984.
[53] S. R. Sanders, “On limit cycles and the describing function method [77] D. Kimhi and S. Ben-Yaakov, “A SPICE model for current mode
in periodically switched circuits,” IEEE Trans. Circuits Syst. I, vol. PWM converters operating under continuous inductor current condi-
40, pp. 564–572, Sept. 1993. tions,” IEEE Trans. Power Electron., vol. 6, pp. 281–286, Apr. 1991.
[54] J. White and S. Leeb, “An envelope-following approach to switching [78] Y. Amran, F. Huliehel, and S. Ben-Yaakov, “A unified SPICE
power converter simulation,” IEEE Trans. Power Electron., vol. 6, compatible average model of PWM converters,” IEEE Trans. Power
pp. 303–307, Apr. 1991. Electron., vol. 6, pp. 585–594, Oct. 1991.
[55] S. R. Sanders, J. M. Noworolski, X. Z. Liu, and G. C. Verghese, [79] S. Ben-Yaakov, “Average simulation of PWM converters by direct
“Generalized averaging method for power conversion circuits,” implementation of behavioral relationships,” in IEEE Applied Power
IEEE Trans. Power Electron., vol. 6, pp. 251–259, Apr. 1991. Electronics Conf. (APEC), Feb. 1993, pp. 510–516.
[56] C. T. Rim and G. H. Cho, “Phasor transformation and its applica- [80] S. Ben-Yaakov and D. Adar, “Average models as tools for studying
tion to the dc/ac analyses of frequency phase-controlled series res- the dynamics of switch mode dc–dc converters,” in IEEE Power
onant converters (SRC),” IEEE Trans. Power Electron., vol. 5, pp. Electronics Specialists Conf. (PESC), 1994, pp. 1369–1376.
201–211, Apr. 1990. [81] V. M. Canalli, J. A. Cobos, J. A. Oliver, and J. Uceda, “Behav-
[57] V. A. Caliskan, G. C. Verghese, and A. M. Stanković, “Multifre- ioral large signal averaged model for dc/dc switching power con-
quency averaging of dc/dc converters,” IEEE Trans. Power Elec- verters,” in IEEE Power Electronics Specialists Conf. (PESC), 1996,
tron., vol. 14, pp. 124–133, Jan. 1999. pp. 1675–1681.
[58] D. W. Novotny and T. A. Lipo, Vector Control and Dynamics of AC [82] Y. Kuroe, T. Maruhashi, and T. Kanayama, “Computation of sen-
Drives. London, U.K.: Oxford Univ. Press, 1996. sitivities with respect to conduction time of power semiconductors
[59] A. M. Stanković, S. R. Sanders, and T. Aydin, “Analysis of unbal- and quick determination of steady state for closed loop power
anced AC machines with dynamic phasors,” in Naval Symp. Electric electronics systems,” in IEEE Power Electronics Specialists Conf.
Machines, Oct. 1998, pp. 219–224. (PESC), 1988, pp. 756–764.
[60] A. M. Stanković, D. J. Perreault, and K. Sato, “Synthesis of dissipa- [83] D. Maksimović, “Automated steady-state analysis of switching
tive nonlinear controllers for series resonant dc/dc converters,” IEEE power converters using a general-purpose simulation tool,” in IEEE
Trans. Power Electron., vol. 14, pp. 673–682, July 1999. Power Electronics Specialists Conf. (PESC), 1997.
[61] J. M. Carrasco, E. Galvan, G. Escobar, R. Ortega, and A. M. [84] T. Kato and W. Tachibana, “Periodic steady-state analysis of an au-
Stanković, “Analysis and experimentation with adaptive controllers tonomous power electronic system by a modified shooting method,”
for the series resonant converter,” IEEE Trans. Power Electron., IEEE Trans. Power Electron., vol. 13, May 1998.
vol. 15, pp. 536–544, May 2000. [85] D. Li and R. Tymerski, “Comparison of simulation algorithms for
[62] P. Mattavelli, G. C. Verghese, and A. M. Stanković, “Phasor accelerated determination of periodic steady state of switched net-
dynamics of thyristor-controlled series capacitor systems,” IEEE works,” IEEE Trans. Ind. Electron., vol. 47, Dec. 2000.
Trans. Power Syst., vol. 12, pp. 1259–1267, Aug. 1997. [86] B. K. H. Wong, H. S. H. Chung, and S. T. S. Lee, “Computations
[63] P. Mattavelli, A. M. Stanković, and G. C. Verghese, “SSR analysis of the cycle state-variable sensitivity matrix of PWM dc/dc con-
with dynamic phasor model of thyristor-controlled series capacitor,” verters and its applications,” IEEE Trans. Circuits Syst. I, vol. 47,
IEEE Trans. Power Syst., vol. 14, pp. 200–208, Feb. 1999. pp. 1542–1548, Oct. 2000.
[64] P. C. Stefanov and A. M. Stanković, “Dynamic phasors in modeling [87] S. Banerjee and G. C. Verghese, Eds., Nonlinear Phenomena in
of UPFC under unbalanced conditions,” in IEEE POWERCON, Dec. Power Electronics: Attractors, Bifurcations, Chaos and Nonlinear
2000. Control. Piscataway, NJ: IEEE Press, 2001.
[65] C. L. DeMarco and G. C. Verghese, “Bringing phasor dynamics into [88] V. J. Thottuvelil and G. C. Verghese, “Analysis and control design
the power system load flow,” in North American Power Symp., Oct. for paralleled dc/dc converters with current sharing,” IEEE Trans.
1993. Power Electron., vol. 13, pp. 635–644, July 1998.
[66] E. H. Allen and M. D. Ilić, “Interaction of transmission network and [89] A. Garg, D. J. Perreault, and G. C. Verghese, “Feedback control
load phasor dynamics in electric power systems,” IEEE Trans. Cir- of paralleled symmetric systems, with applications to nonlinear dy-
cuits Syst. I, vol. 47, pp. 1613–1620, Nov. 2000. namics of paralleled power converters,” in IEEE Int. Symp. Circuits
[67] L. García de Vicuña, A. Poveda, L. Martínez, F. Guinjoan, and and Systems (ISCAS), May 1999.
J. Majo, “Computer-aided discrete-time large-signal analysis of [90] P. Maranesi, “Small-signal circuit modeling in the frequency-domain
switching regulators,” IEEE Trans. Power Electron., vol. 7, pp. by computer-aided time-domain simulation,” IEEE Trans. Power
75–82, Jan. 1992. Electron., vol. 7, pp. 83–88, Jan. 1992.
[68] F. Guinjoan, J. Calvente, A. Poveda, and L. Martínez, “Large-signal [91] R. C. Wong and J. Groves, “An automated small-signal frequency-
modeling and simulation of switching dc–dc converters,” IEEE domain analyzer for general periodic-operating systems as obtained
Trans. Power Electron., vol. 12, pp. 485–494, May 1997. via time-domain simulation,” in IEEE Power Electronics Specialists
[69] H. A. Owen, A. Capel, and J. G. Ferrante, “Simulation and anal- Conf. (PESC), 1995, pp. 801–808.
ysis methods for sampled power electronic systems,” in IEEE Power [92] P. Huynh and B. H. Cho, “Empirical small-signal modeling of
Electronics Specialists Conf. (PESC), 1976, pp. 44–55. switching converters using PSpice,” in IEEE Power Electronics
[70] A. R. Brown and R. D. Middlebrook, “Sampled-data modeling of Specialists Conf. (PESC), 1995, pp. 809–815.
switching regulators,” in IEEE Power Electronics Specialists Conf. [93] D. Maksimović, “Automated small-signal analysis of switching con-
(PESC), 1981, pp. 349–369. verters using a general-purpose time-domain simulator,” in IEEE
[71] D. J. Shortt and F. C. Lee, “Extensions of the discrete-average Applied Power Electronics Conf. (APEC), 1998.
models for converter power stages,” in IEEE Power Electronics [94] H. A. Mantooth and P. O. Lauritzen, “Compact models of power
Specialists Conf. (PESC), June 1983, pp. 23–37. devices and power ICs for circuit simulation,” in IEEE APEC
[72] J. M. Burdío and A. Martínez, “A unified discrete-time state-space Tutorial, Feb. 2000, Available: http://www.ee.washington.edu/re-
model for switching converters,” IEEE Trans. Power Electron., vol. search/pemodels/ for a description of many publicly available
10, pp. 694–707, Nov. 1995. device models.
[73] H. N. Chandra and V. J. Thottuvelil, “Modeling and analysis of com- [95] A. N. Githiari, B. M. Gordon, R. A. McMahon, Z.-M Li, and P. A.
puter power systems,” in IEEE Power Electronics Specialists Conf. Mawby, “A comparison of IGBT models for use in circuit design,”
(PESC), 1989, pp. 144–151. IEEE Trans. Power Electron., vol. 14, pp. 607–615, July 1999.
[74] V. Bello, “Computer aided analysis of switching regulators using [96] A. R. Hefner and D. L. Blackburn, “Simulating the dynamic electro-
SPICE2,” in IEEE Power Electronics Specialists Conf. (PESC), thermal behavior of power electronic circuits and systems,” IEEE
1980, pp. 3–11. Trans. Power Electron., vol. 8, pp. 376–385, Oct. 1993.

MAKSIMOVIĆ et al.: MODELING AND SIMULATION OF POWER ELECTRONIC CONVERTERS 911

 [97] A. Ammous, K. Ammous, H. Morel, B. Allard, D. Bergogne, F. Sel-
lami, and J. P. Chante, “Electrothermal modeling of IGBTs: Applica- Aleksandar M. Stanković (Member, IEEE) received the Dipl. Ing. degree
tion to short-circuit conditions,” IEEE Trans. Power Electron., vol. in 1982 and the M.S. degree in 1986, both from the University of Belgrade,
15, pp. 778–790, July 2000. Yugoslavia, and the Ph.D. degree from the Massachusetts Institute of Tech-
[98] P. Mawby, “Compact electrothermal models for power electronics,” nology, Cambridge, in 1993, all in electrical engineering.
in Inst. Elect. Eng. Colloq. Power Electronic Systems Simulation, He has been with the Department of Electrical and Computer Engineering
1998, Ref. 1998/486, pp. 3/1–3/4. at Northeastern University, Boston, since 1993, presently as an Associate
[99] S. Cristina, F. Antonini, and A. Orlandi, “Switched mode power sup- Professor.
plies EMC analysis: near field modeling and experimental valida- Dr. Stanković is a member of IEEE Power Engineering, Power Elec-
tion,” in IEEE Int. Symp. Electromagnetic Compatibility, 1995, pp. tronics, Control Systems, Circuits and Systems, Industrial Electronics, and
453–458. Industry Applications Societies. He serves as an Associate Editor for the
[100] R. Prieto, J. A. Cobos, O. Garcia, P. Alou, and J. Uceda, “Model IEEE TRANSACTIONS ON CONTROL SYSTEM TECHNOLOGY, and as chair of
of integrated magnetics by means of finite element analysis tech- the technical committee on Power Electronics and Power Systems of the
niques,” in IEEE Power Electronics Specialists Conf. (PESC), 1999. IEEE Circuits and Systems Society.
[101] J. Pleite, R. Prieto, R. Asensi, J. A. Cobos, and E. Olias, “Obtaining
a frequency-dependent and distributed-effects model of magnetic
components from actual measurements,” IEEE Trans. Magn., vol.
35, pp. 4490–4502, Nov. 1999.
[102] N. Femia and G. Spagnuolo, “True worst-case circuit tolerance anal- V. Joseph Thottuvelil (Member, IEEE) received the B.S. degree from the
ysis using genetic algorithms and affine arithmetic,” IEEE Trans. Indian Institute of Technology, Madras, and the M.S. and Ph.D. degrees from
Circuits Syst. I, vol. 47, pp. 1285–1296, Sept. 2000. Duke University, Durham, NC, all in electrical engineering, in 1978, 1980,
and 1984, respectively.
He was with Digital Equipment Corporation from 1984 to 1993.
Between 1993 and 2000, he was with Bell Laboratories, Lucent Technolo-
gies, working on telecommunications and computer power systems and
components. He is now with Tyco Electronics Power Systems as Technical
Manager of the Applications Engineering Group.
Dr. Thottuvelil was Secretary of the IEEE Power Electronics Society from
1995–1998, and is currently Associate Editor for the IEEE TRANSACTIONS
ON POWER ELECTRONICS, covering the area of telecommunications.

George C. Verghese (Fellow, IEEE) received the

B.Tech. degree from the Indian Institute of Tech-
nology at Madras in 1974, the M.S. degree from
the State University of New York at Stony Brook
in 1975, and the Ph.D. degree from Stanford Uni-
versity, Stanford, CA, in 1979, all in electrical en-
gineering.
Since 1979, he has been at the Massachusetts
Dragan Maksimović (Member, IEEE) was born Institute of Technology, Cambridge, where
in Belgrade, Yugoslavia, on July 15, 1961. He he is Professor of Electrical Engineering in
received the B.S. and M.S. degrees in electrical the Department of Electrical Engineering and
engineering from the University of Belgrade and Computer Science and a member of the Laboratory for Electromagnetic
the Ph.D. degree from the California Institute of and Electronic Systems. His research interests and publications are in the
Technology, Pasadena, in 1984, 1986, and 1989, areas of systems, control, estimation, and signal processing, with a focus on
respectively. applications in power electronics, power systems, and electrical machines.
From 1989 to 1992, he was with the University He is co-author (with J. G. Kassakian and M. F. Schlecht) of Principles
of Belgrade. Since 1992, he has been with the De- of Power Electronics (Addison-Wesley, 1991), and co-editor (with S.
partment of Electrical and Computer Engineering Banerjee) of Nonlinear Phenomena in Power Electronics: Attractors,
at the University of Colorado, Boulder, where he Bifurcations, Chaos, and Nonlinear Control (IEEE Press, 2001). He has
is currently an Associate Professor and Co-Director of the Colorado Power served on the AdCom and other committees of the IEEE Power Electronics
Electronics Center (CoPEC). His current research interests include simula- Society, and also as founding co-chair (with V. J. Thottuvelil) of its
tion and control techniques, low-harmonic rectifiers, and power electronics technical committee and workshop on Computers in Power Electronics.
for low-power, portable systems. Dr. Verghese has served as an Associate Editor of Automatica, the IEEE
In 1997 he received the NSF CAREER Award, and a Power Electronics TRANSACTIONS ON AUTOMATIC CONTROL, and the IEEE TRANSACTIONS ON
Society Transactions Prize Paper Award. CONTROL SYSTEMS TECHNOLOGY.

912 PROCEEDINGS OF THE IEEE, VOL. 89, NO. 6, JUNE 2001