Hysteresis Control of the Pseudo Boost PFC

Converter
Aleksandra Lekić-Vervoort Milovan Majstorović, Leposava Ristić Dušan Stipanović
Intelligent Electrical Power Grids (IEPG), The Department of Power Converters Coordinated Science Laboratory
Electrical Sustainable Energy (ESE), and Drive Systems, University of Illinois
Faculty of Electrical Engineering, School of Electrical Engineering, Urbana, IL 61801, USA
Mathematics and Computer Science, TU Delft University of Belgrade [email protected]
Delft, Netherlands Belgrade, Serbia
[email protected] [email protected], [email protected]

Abstract—In this paper, a control design for the the Pseudo

Boost PFC converter with the application in the charger for
electric vehicles, is formulated and verified. The rectifier ends up
being energy efficient because due to containing smaller number
of switches. The converter is controlled using a method involving
the minimum Lyapunov function derivative and a hysteresis.
This control provides an averaged value of the converter sine
input current, and thus, decreasing the THD value. Theoretical
results are complemented with the corresponding simulations and
experimental waveforms.
Index Terms—AC-DC Converter, Electric Vehicles, Lyapunov,
Hysteresis. Fig. 1: AC-DC Pseudo Boost converter.
I. I NTRODUCTION
Nowadays, our society is facing the enormous CO2 pol- where we have used a common quadratic Lyapunov function
lution. One of the solutions to reduce pollution is to search that fits the converter’s dynamics. Now we take a step further
for the alternative transportation, which is not powered by and design a controller based on this Lyapunov function and
gasoline and does not produce exhaust gases. The electric a hysteresis. This control method guarantees stability of the
vehicles present this alternative and they are powered with a converter with the increased efficiency and reduced value for
battery or with some renewable energy source. However, there THD.
are still unsolved questions about construction of the high The paper consists of five sections. In the section I are given
speed chargers and on-side vehicle batteries, which require the basics about pseudo boost converter and its possible operat-
persistent and thorough research. Battery powered vehicles ing modes. For the determined converter’s model the quadratic
are the easiest for maintenance and thus the most common. Lyapunov function is proposed used for constructing the so-
Batteries should be periodically charged, and chargers should called hysteresis controller. In the section IV is demonstrated
satisfy specific standards. The standard SAE J1772, the level operation of the hysteresis controlled pseudo boost AC-DC
II residential charging applications [1] require 240 V DC volt- converter on the number of representative simulations, done
age. This DC voltage is obtained from the line voltage using in PSIM c . Finally, section V is the conclusion.
AC-DC converters, which should be carefully constructed to
meet criteria of power density, efficiency and power correction Notation
[2]. It should be noted that throughout the text vectors are
In order to construct energy efficient AC-DC converter, the denoted with lowercase bold identifiers and matrices with
new converter topologies are derived without the bridge stage. uppercase bold. In what follows, all state variables are time
These converters are known as bridgeless PFC converters [3]– dependent, but somewhere time dependence is not explicitly
[5]. One bridgeless AC-DC converter is AC-DC PFC pseudo noted in order to simplify the notation and avoid clutter. In the
boost converter depicted in Fig. 1 (also known as Ćuk), which cases when the time dependence is not clear from the context,
can have up to 98% power factor [6], [7] and can be controlled it is depicted as such.
to obtain low THD. Besides the standard controlling methods,
there have been proposed works related to the construction of II. AC-DC P SEUDO B OOST C ONVERTER
the changing duty ratio to fit better current waveform [8]. In The AC-DC Pseudo Boost converter depicted in Fig. 1 has
this paper, we will present the controller design for the AC- three state variables being x = [x1 x2 x3 ]T = [iL1 vC1 vC0 ]T .
DC Pseudo Boost PFC converter, which relies on our work [9] The AC input voltage is represented as vIN (t) = Vm sin(ω0 t),

978-1-7281-5635-4/20/$31.00 ©2020 IEEE 731

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PIAUI. Downloaded on March 19,2024 at 14:22:48 UTC from IEEE Xplore. Restrictions apply.
where Vm is its amplitude and ω0 is its frequency. This be designed to operate at the end of the switching period in
converter can operate in 6 different subsystems [9] described subsystem 4 [10]. Thus, for the positive input voltage vIN > 0
in the state space matrix form as: ẋ = Ai x+Bi , for i denoting the converter produces switching sequence e 1 → 2 → 3 → 4,
i-th operating subsystem. where with e 1 is considered any of the subsystems 1, 1∗ and
Representing conducting transistors and diodes as a series 1 ∗ ∗. For the negative input voltage, the switching sequence
combination of an ideal voltage source and a resistance, becomes e 1 → 3 → 2 → 4.
annotated as rq and VQ for the transistor and rd and VD for the The equilibrium of the system over the whole time period
x3
diode, and assuming that the conducting bidirectional switch q x1 = iL1 = 0, x2 = − 2 and x3 = M Vm , where
T0 is then
S is represented as VS = VQ − VD ≈ 0 and rs = rq + rd , the R 2L1
M = 2Re and Re = d2 TS , for d ∈ [0, 1] being assumed
following matrices can be derived [9]: constant duty-ratio in the steady-state.
1) Subsystem 1: If diodes D1 and D2 are turned off, then
only switch S is conducting and the system matrices are: III. H YSTERESIS LYAPUNOV C ONTROL
 rs   vIN  For the switching system, such as AC-DC Pseudo Boost
− L1 0 0 L1 converter, it is desired to find the common quadratic Lyapunov
A1 =  0 0 0 , B1 =  0  . (1)
1 function of the form:
0 0 − RC 0
0
V (y) = yT P y, (5)
2) Subsystem 1∗: If at moment t = t0 , vC1 (t0 ) −
rS iL1 (t0 ) ≥ VD1 , then diode D1 turns on and the system where y = x − x for the equilibrium point x. A common
matrices are given by quadratic Lyapunov function is desirable due to its continuous
  differentiability [11]. For the quadratic Lyapunov function (5)
rs rd1
− L1 (r s +rd1 )
− L1 (rrss+rd1 ) 0 one can compute its derivative to obtain
A1∗ =  C1 (rrss+rd1 ) − C1 (rs1+rd1 ) 0  ,

1
V̇i (y) = yT Qi y + yT By1 , (6)
0 0 − RC
h 0
iT where Qi = ATi P + P Ai , Byi = 2 P (Ai x + Bi ) and i ∈
vIN rs VD1 VD1
B1∗ = L1 + L1 (rs +rd1 ) , C1 (rs +rd1 ) , 0 . {1, 2, 3, 4}.
Using the calculations for the equilibrium from the previous
3) Subsystem 1∗∗: If at moment t = t0 , −vC1 (t0 ) − section x = [0 − x23 x3 ]T , where x3 = M Vm and defining
vC0 (t0 ) + rS iL1 (t0 ) ≥ VD2 , then diode D2 turns on and the matrix P = p diag{L1 , C1 , C0 } for any p > 0, the time
system matrices become derivatives of the Lyapunov function are as follows:
− L1 r(rssr+r − L1 (rrss+rd2 ) − L1 (rrss+rd2 )
 d2

y32
 
d2 ) 2 x3 y3
A1∗∗ = 
 rs
− C1 (rs1+rd2 ) − C1 (rs1+rd2 )  , V̇1 (y) = 2p −rs y1 − + (vIN − rs x1 ) y1 − ,
C1 (rs +rd2 ) R R
rs
− C0 (rs1+rd2 ) − R R+r s +rd2
y22 y2
C0 (rs +rd2 ) (rs +rd2 ) C0

rs rd1 2
− 3+
iT
V̇1∗ (y) = 2p − y1 −
h
vIN rs VD2
B1∗∗ = L1
− L1 (rs +rd2 )
, − C1 (rVsD2
+rd2 )
, − C0 (rVsD2
+rd2 ) . rs + rd1 rs + rd1 R
 
4) For subsystem 2, S is turned off, D1 is turned off, D2 rs (VD1 − x2 ) VD1 − x2
+ vIN + y1 + y2 −
is turned on, and the corresponding system matrices are as rs + rd1 rs + rd1

follows: x3 y3
− ,
 rd2
− L1 − L11 − L11
  vIN −VD2  R
L1
y22

A2 =  C11 0 0  , B2 =  0  . (2) rs rd2 2
1 1
V̇1∗∗ (y) = 2p − y1 − −
C0 0 − RC0 0 rs + rd2 rs + rd2
y 2 (R + rs + rd2 ) 2y2 y3
5) Subsystem 3 represents switch S being off, D1 is on, and − 3 − + vIN y1 −
D2 is off, resulting in R (rs + rd2 ) rs + rd2

(VD2 + x2 + x3 ) (−rs y1 + y2 + y3 ) x3 y3
 rd1
− L1 − L11 0
  vIN +VD1  − − ,
L1 rs + rd2 R
A3 =  C11 0 0  , B3 =  0  . (3)
y2

0 0 − RC1
0 V̇2 (y) = 2p −rd2 y12 − 3 + (vIN − VD2 − x2 − x3 ) y1 −
0 R
6) Subsystem 4, which is a discontinuous current inductor

x3 y3
mode, is characterized with all switches turned of and: − ,
R
y2
    
0 0 0 0
V̇3 (y) = 2p −rd1 y12 − 3 + (vIN + VD1 − x2 ) y1 −
A4 = 0 0 0  , B4 = 0 . (4) R
1
0 0 − RC 0
0 x3 y3

− ,
In order to minimize THD, the averaged value of the input R
 2 
inductor current over one switching period TS = 1/fS has the y x3 y3
2 V̇4 (y) = 2p − 3 − . (7)
sinusoidal form, hiL1 iTS = vIN2Ld1 TS , if the converter should R R

732

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PIAUI. Downloaded on March 19,2024 at 14:22:48 UTC from IEEE Xplore. Restrictions apply.
 Similar result is obtained for vIN < 0, when the system has
switching sequence e 1 → 3 → 2 → 4. We can introduce the
limit for switching between subsystems e 1 and 3 like:
V̇e1 (y) − V̇3 (y) =
 

= 2p y12 (rd1 − rs ) +y1 (−rs x1 − VD1 + x2 )

| {z }
 ≈0 
x3
= −2p y1 VD + = −ρ(t), (10)
2
and the minimum value of the negative inductor current
becomes
ρ(t)
y1,min (t) =
. (11)
Fig. 2: Limiting of the maximum inductor current using 2p VD + x23
hysteresis Lyapunov switching law. Minimum value of the inductor current when the input voltage
is negative has the same absolute value as y1,max for the
positive input voltage.
The hysteresis Lyapunov control is described in [12], [13], Knowing that the peak value of the inductor current has the
where it is designed controller for the DC-DC Ćuk converter. expression [9]:
It relies on the application of the minimal derivative of the 2vIN (t)
y1,max (t) = , (12)
Lyapunov function and the hysteresis for the construction dRe
of the switching law. The main idea of this design is to for some assumed duty-ratio d ∈ (0, 1) in the steady-state and
obtain constant switching period during which the on time using the expression for the y1,max (t) in the equation (9), it
of the switch S is determined using the minimum value of the can be written that
Lyapunov derivative. 2vIN (t) ρ(t)
According to the switching sequence when the input voltage = ,
dRe px3
1 → 2 → 3 → 4, on time of the switch S can
is positive being e
be limited by introducing the parameter ρ, such that switch where Re = d2L 2 T . After substitution of the equilibrium
1
S

S turns off when it becomes V̇e1 (y) − V̇2 (y) = 0. With this parameter ρ can be estimated as:
approach we are limiting the duration of the operation in the 2px3 vIN (t)
subsystem e 1 and the time instance of the switching between ρ(t) = . (13)
dRe
subsystems e 1 and 2. As it is shown in our previous work [9],
Using previously determined hysteresis formula (13), the
when the switch S is turned off, system stabilizes itself and
complete switching law is defined as:
switches always in the subsystem for which is Lyapunov func-
tion derivative minimal. Thus, switching between subsystems ρ
if S = 1 and iL1 > 0, iL1 > switch to S = 0,
in the off state of the switch S does not need to be further px3
ρ
controlled. if S = 1 and iL1 < 0, iL1 < switch to S = 0,
px3
AC-DC Ćuk converter switches from subsystem e 1 to the
if S = 0 switch to S = 1 if t = n TS , n ∈ Z. (14)
subsystem 2 when:
One realisation of the proposed hysteresis controller is
V̇e1 (y) − V̇2 (y) = depicted in Fig. 3, where it can be seen that the switches
  denoted as S are turned on with the constant frequency fS
= 2p y12 (rd2 − rs ) +y1 (−rs x1 + VD2 + x2 + x3 ) using short pulse on the set input of the SR flip-flop, triggered
| {z } on the edge. Control logic from equation (14) is used to reset
≈0
  the flip-flop and that way adjust the duty-ratio to obtain desired
x3 hysteresis.
= 2p y1 VD + = ρ(t), (8)
2
IV. N UMERICAL S IMULATIONS
which gives maximum value of the inductor current being: For the pseudo boost PFC depicted in Fig. 1, whose circuit
parameters are given in Table I, the simulations were per-
ρ(t) formed when it was controlled by the PID controller and by
y1,max (t) = x3

. (9)
2p VD + 2
the proposed hysteresis-based switching controller, and then
the results are compared and discussed.
This result is presented in the Fig. 2, where is visible how the In Fig. 5 are depicted waveforms for the PID controlled
equation (8) influences behavior of the inductor current. pseudo boost PFC constructed for the desired duty-ratio

733

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PIAUI. Downloaded on March 19,2024 at 14:22:48 UTC from IEEE Xplore. Restrictions apply.
 Fig. 3: Hysteresis controller implementation.

TABLE I: Pseudo boost PFC converter’s circuit parameters

L1 = 1 mH C1 = 68 nF C0 = 390 µF
R = 1 kΩ Vm = 80 V f0 = 50 Hz

d = 0.2 and switching frequency fS = 10 kHz. For these

values, it can be determined that Re = 500 Ω and expected
x3 = Vm = 80 V, which is the case. PID controller is de-
signed like in Fig. 4 according to [4].

Fig. 5: Startup transient of the pseudo boost PFC converter

controlled using PID controller from Fig. 4.

seen that in the steady-state THD of the hysteresis controlled

PFC is more stable (almost constant), while PID controlled
PFC changes over time. The second advantage is that the
THD value is slightly smaller for the hysteresis controlled
PFC. These simulations are also done in the PSIM c software
using THD measurement tool to verify the statement of the
paper.

V. C ONCLUSION
Fig. 4: PID controller.
In this paper, a construction and a validation of the hys-
In order to obtain the same steady-state operation of teresis Lyapunov controller for the special pseudo boost PFC,
the pseudo boost PFC, it is necessary to determine the are provided. Application of the Lyapunov control stems from
hysteresis waveform from equation (13). With the same the proven fact that the internal converter’s dynamics shows
circuit’s parameters as before, the required hysteresis is great correspondence with the switching according to the
2
ρ(t) =
2pVm sin(ωt)
= 128 sin(ωt) W, for some p = 1. Sim- minimum among the set of Lyapunov derivatives. Thus, this
dRe
ulations are done for the case of hysteresis controlled pseudo hysteresis controller is implemented such that the switching
boost PFC and the obtained startup transient waveforms are of the external control, applied to the controlled switches,
depicted in Fig. 6. As it can be seen from the diagrams in employs the same principle. In this work not only that the
Figs. 5 and 6, the hysteresis controlled converter shows smaller technical derivations are provided but the hysteresis control is
overshoot on the inductor current iL1 and capacitor voltage verified by the relevant simulations using PSIM c software.
vC1 , which decreases the stress on the used switches (diodes Additionally, simulation results for the pseudo boost PFC
and transistors) and also the price of the setup, since it requires that compare THD values for the PID controlled and for the
components for the smaller currents/voltages. hysteresis controlled converter are provided where they show
The same simulation results are analyzed with respect to a clear benefit of using the proposed control design.
THD in both cases: with PID and with hysteresis controller. As
ACKNOWLEDGEMENT
it can be seen from Fig. 7 in the steady-state is visible the full
impact of the use of hysteresis control. Namely, from the THD This work was partially supported by the Ministry of
diagrams depicted in Figs. 7, it can be seen that there are two Education, Science and Technological Development of the
benefits when using the hysteresis controller. First, it can be Republic of Serbia through project TR33020.

734

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PIAUI. Downloaded on March 19,2024 at 14:22:48 UTC from IEEE Xplore. Restrictions apply.
 R EFERENCES
[1] K. Morrow, D. Karner, and J. Francfort, “Plug-in hybrid electric vehi-
cle charging infrastructure review,” US Department of Energy-Vehicle
Technologies Program, vol. 34, 2008.
[2] M. N. Z. Abidin, “IEC 61000-3-2 harmonics standards overview,”
Schaffner EMC Inc., Edsion, NJ, USA, 2006.
[3] F. Musavi, M. Edington, W. Eberle, and W. G. Dunford, “Evaluation
and efficiency comparison of front end ac-dc plug-in hybrid charger
topologies,” IEEE Transactions on Smart grid, vol. 3, no. 1, pp. 413–
421, 2012.
[4] A. A. Fardoun, E. H. Ismail, M. A. Al-Saffar, and A. J. Sabzali, “New
“real” bridgeless high efficiency AC-DC converter,” in Applied Power
Electronics Conference and Exposition (APEC), 2012 Twenty-Seventh
Annual IEEE. IEEE, 2012, pp. 317–323.
[5] M. Mahdavi and H. Farzanehfard, “Bridgeless SEPIC PFC rectifier
with reduced components and conduction losses,” IEEE Transactions
on Industrial Electronics, vol. 58, no. 9, pp. 4153–4160, 2011.
[6] S. Ćuk, “Bridgeless PFC converter,” Oct. 14
2010, uS Patent App. 12/798,682. [Online]. Available:
https://www.google.com/patents/US20100259240
[7] ——, “Single-stage AC-to-DC converter with isolation and power
factor correction,” Dec. 1 2011, uS Patent App. 12/802,122. [Online].
Available: https://www.google.com/patents/US20110292703
[8] A. El Aroudi, M. Bodetto, A. Cid-Pastor, and M. Al-Numay, “Power
quality issues in single-stage ac-dc hbled drivers at low power levels:
Problems and solutions,” in 2017 International Conference on Green
Energy Conversion Systems (GECS). IEEE, 2017, pp. 1–6.
[9] N. Petrović, A. Lekić, and D. Stipanović, “Lyapunov characterization
and analysis of the operating modes of the AC-DC ćuk converter,” IEEE
Journal of Emerging and Selected Topics in Power Electronics, vol. 7,
Fig. 6: Startup transient of the pseudo boost PFC converter no. 2, pp. 1318–1328, 2019.
controlled using hysteresis controller. [10] A. A. Fardoun, E. H. Ismail, M. A. Al-Saffar, and A. J. Sabzali, “A
bridgeless resonant pseudoboost PFC rectifier,” IEEE Transactions on
Power Electronics, vol. 29, no. 11, pp. 5949–5960, 2014.
[11] A. F. Filippov, Differential Equations with Discontinuous Righthand
Sides. Dordrecht, The Netherlands: Kluwer Academic Publishers, 1988.
[12] A. Lekić and D. Stipanović, “Hysteresis switching control of the Ćuk
converter,” IEEE Transactions on Circuits and Systems I: Regular
Papers, vol. 63, no. 11, 2016.
[13] A. Lekić and D. Stipanović, “Hysteresis switching control of the
Ćuk converter operating in discontinuous conduction modes,” IEEE
Transactions on Circuits and Systems II: Express Briefs, vol. 64, no. 9,
2016.

Fig. 7: THD factor change during transient of pseudo boost

PFC converter controlled using PID controller and hysteresis
controller.

735

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PIAUI. Downloaded on March 19,2024 at 14:22:48 UTC from IEEE Xplore. Restrictions apply.