This paper proposes a pulse width modulation (PWM)based sliding mode controller (SMC) for a fullbridge DCDC converter that can eliminate static output voltage error. Hysteretic SMC in DCDC converter does not have a fixed switching frequency, and applying hysteretic SMC to fullbridge converters is difficult. Fixedfrequency SMC, which is also called PWMbased SMC, based on equivalent control overcomes these shortcomings. However, the controller order reduction in equivalent control in PWMbased SMC causes static output voltage error. To resolve this issue, an integral item is added to the PWMbased SMC. Sliding mode coefficients are designed by applying a standard secondorder system to the sliding mode surface. The effect of adding an integral item on the controller is analyzed, and an integral coefficient design method is proposed. Experiment results on a threelevel fullbridge DCDC converter verify the control scheme and design method proposed in this paper.
I. INTRODUCTION
Traditional linear PI controller is based on a small signal model, which is only effective in a very narrow range. Although the PI control scheme is widely used, it shows limited performance, especially poor robustness against parameter uncertainties. Thus, this scheme is unsuitable for variable structure systems
[1]
,
[2]
. Sliding mode controller (SMC) is a kind of nonlinear controller suitable for variable structure systems. This controller can guarantee stability and robustness against parameter uncertainties. Infinite frequency is the ideal operating frequency of SMC. However, switching frequencies of power converters must be restricted within a practical range
[3]

[5]
.
Hysteretic SMC requires a bangbang type of controller and has variable switching frequencies. Switching frequency is affected by input voltage and load. Variable switching frequencies challenge the feasibility of hysteretic SMC because the filter must suit a wide frequency range. Moreover, hysteretic SMC can only be used in simple DCDC converters such as buck and boost. Hysteretic SMC cannot be directly used in fullbridge converters
[6]
,
[7]
. The converters proposed in
[8]
,
[9]
require control variables, such as duty cycle and phaseshift angle. If SMC is used in the converters in
[8]
and
[9]
, then the equivalent control variable can be derived.
Equivalent control concept is introduced into the SMC to obtain the fixedfrequency control scheme. Fixedfrequency SMC, also called pulse width modulation (PWM)based SMC, can resolve the problems of hysteretic SMC. The PWM technique can be used in SMC. The duty cycle control variable is independent from the switching frequency. Thus, the application of the PWM technique to SMC does not contradict the concept of classical PWM controllers used in power electronics. SMC is based on SMC law, whereas classical PWM control is based on linear control law
[10]
. PWM technique gives SMC the ability to be used in a wide range of DCDC converters
[10]

[12]
. In
[11]
,
[12]
, the sling mode coefficients are analyzed and the design method is introduced.
An SMC with double integral sliding surface for buck converter is designed in
[13]
. In
[13]
, the SMC is analyzed and designed as a thirdorder system. Meanwhile, in the present paper, the secondorder sliding surface and the additional integral are analyzed and designed, respectively. In
[14]
a sampleddata reducedorder observer is adopted to deal with the unknown parameters, and a robust sampleddata output feedback control law developed to achieve global tracking control performance in high precision for buck converter.
In
[15]
, SMC scheme via multiresonant sliding surface is used for singlephase gridconnected voltage source inverter. In
[16]
, active damping based on SMC is used in threephase voltage inverters. Sliding mode observer is used in sensorless vector control of induction machine in
[17]
. The SMC strategy is applied to a permanent magnet synchronous machine vector control system in
[18]
to improve system robustness amid parameter changes and disturbances.
SMC with a fast output sampling strategy is used in
[19]
for an uncertain system to avoid switching of control and chattering. A new development in the time optimal control theory in SMC systems for multiquadrant buck converters with a variable load is presented in
[20]
. In
[21]
and
[22]
, the analysis and design of parallelconnected converter system using SMC techniques are proposed. In
[23]
, a complete SMC scheme is designed for the voltage tracking control of a conventional boost converter. A robust adaptive SMC is designed for a boost converter in
[24]
. An active antidisturbance control approach called disturbance observerbased control is analyzed in
[25]
. This controller provides an active and effective way to handle disturbances and improve robustness of the closedloop system. An SMC approach for systems with mismatched uncertainties via a nonlinear disturbance observer is developed in
[26]
.
However, equivalent control scheme reduces the order of the SMC. Traditional PWMbased SMC has no ability to eliminate static output voltage error, which is influenced by the load. In the present paper, the integral of output voltage error is added to the SMC as a new control item. Sliding mode coefficients are designed by applying a standard secondorder system to the sliding mode surface. The effect of adding an integral item on the controller is analyzed, and an integral coefficient design method is proposed.
This paper is organized as follows. The threelevel fullbridge converter and its phaseshift control scheme are introduced in Section II. Traditional PWMbased SMC is introduced in Section III. PWMbased SMC with an additional integral to eliminate static output voltage error is proposed in Section IV. Sliding mode coefficients and an additional integral coefficient are designed in Section V. Experiment results are given in Section VI, and the conclusion is presented in Section VII.
II. THREELEVEL FULLBRIDGE CONVERTER AND ITS PHASESHIFT CONTROL SCHEME
The structure of neutralpointclamped (NPC) threelevel fullbridge converter is shown in
Fig. 1
. The converter includes two NPC threelevel bridges. The center tap of the two capacitors provides neutral voltage for the converter. Each phase leg of the threelevel converter has two pairs of switching devices in series. The center of each device pair is clamped to the neutral through a clamping diode. In this kind of converter, the voltage stress of main switches is half of the input voltage by using the NPC technique. This structure is suitable for highvoltage and highpower applications
[27]
.
Threelevel fullbridge DCDC converter.
The converter has a highfrequency transformer and an uncontrolled rectifier bridge (
Fig. 1
). Inductor
L
and capacitor
C
comprise the lowpass
LC
filter. Resistor
R
_{L}
is the load of the converter. Compared with a twolevel fullbridge converter, this converter has more switches and its phaseshift control scheme is more complex.
One of the most useful phaseshift control schemes is shown in
Fig. 2
. Switches
Q
1 to
Q
4 are all switched out of phase. The right leg is similar to the left leg. Both voltages
u
_{ao}
and
u
_{bo}
have three levels. The primary side voltage of the transformer
u
_{ab}
has five levels. In this converter, the main two control variables are conduction angle
θ
and phaseshift angle
α
. In
Fig. 2
,
γ
he clamp angle and is shown in Equation (1). Hysteretic SMC cannot be directly used in a threelevel fullbridge converter (
Figs. 1
and
2
).
Phaseshift control of threelevel fullbridge converter.
THD of the primary side voltage of the transformer should be kept low because voltage and current harmonics will result in high power loss and excessive temperature rise, which will eventually diminish the useful life of the transformer.
Figure 3
shows the primary side voltage of the transformer in one switching period. The converter switching frequency must be selected as the fundamental frequency, and the THD of the waveforms in
Fig. 3
must be analyzed.
Primary side voltage of the transformer in one switching period.
THD of the waveforms is shown in Equation (2) and determined by conduction angle and phaseshift angle. These two control variables should be selected to keep the THD low, and thus find the optimal operation point. The minimum value of the THD can be searched using a mathematical tool. On the optimal operation point, the conduction angle value is 2.636 rad and the phaseshift angle value is 2.187 rad.
Phaseshift angle and conduction angle influence the fundamental output voltage. After the optimal operation point is selected, the transformer ratio will be designed according to the operation point and the desired output voltage. The phaseshift angle will then be close to the optimal operation point when the converter is working.
Conduction angle should be initialized when the converter is started. The phaseshift angle is generated by the SMC. Duty cycle control value is obtained from the PWMbased SMC. The relationship between duty cycle value and phaseshift angle is shown in Equation (3).
From this section, the hysteretic SMC cannot be used directly in the threelevel fullbridge converter. In a buck converter, the only switch can be driven with a bangbang type controller. Meanwhile, in fullbridge converter, basic SMC based on a mathematical model cannot be directly used. Thus, the equivalent SMC should be derived.
III. PWMBASED SMC
As described in Section I, hysteretic SMC cannot be directly used in a fullbridge converter. However, PWMbased SMC should be obtained from the hysteretic SMC model. In this section, hysteretic SMC will be initially derived, and then the PWMbased SMC will be obtained by applying the equivalent control concept.
 A. Hysteretic SMC
Fig. 4
shows the hysteretic SMC. The control function
u
is 0 or 1 (
Fig. 4
). This control function can be used in a buck or boost converter. Equivalent control function should be obtained from control function
u
. Three variables are selected as the state variables for the hysteretic SMC (
Fig. 4
), namely, output voltage error
x
_{1}
, differential of output voltage error
x
_{2}
, and integral of output voltage error
x
_{3}
. These variables are defined as Equation (5).
Hysteretic SMC for DCDC converter.
A new variable
v_{i}
is introduced to simplify the analysis, as shown in Equation (4). This variable is defined through input voltage and ratio of isolated transformer from primary side to secondary side.
State space variable vector is defined as Equation (5), and the system model of threelevel fullbridge DCDC converter is applied.
From Equation (5), the state space function of hysteretic SMC can be obtained, as shown in Equation (6). The three matrixes are coefficients of state space function.
where
In the SMC, variable
S
is defined as an instantaneous trajectory of state variable, as shown in Equation (7). The aim of SMC is to design a certain sliding surface in its control law that will direct the trajectory of state variables toward a desired region. By enforcing
S
= 0, a sliding surface can be obtained for the SMC.
where
In the sliding surface
S
= 0,
α
_{1}
,
α
_{2}
, and
α
_{3}
are the sliding mode coefficients to determine the trajectory of state variables. These three coefficients are the main parameters we need to design. A new matrix that represents sliding mode coefficients is defined as Equation (8).
Switching function shown in Equation (9) is adopted as the control law of the SMC. This control law is a bangbang type control function, and its output is determined by the trajectory of the state variable.
The entire sliding mode movement process can be divided into two stages. In the first stage, the controller forces the trajectory of the state variable to reach the sliding surface. When the trajectory is within a small vicinity of the sliding surface, the trajectory is in sliding mode operation, which is the second stage of the control process. The SMC will give the converter continuous control actions to maintain the trajectory of the state variable within a small vicinity of sliding surface.
 B. Existence Condition of SMC
In all of the SMCs, sliding coefficients should satisfy the existence condition of SMC. Such satisfaction will ensure that the trajectory of the state variable will always be directed toward the sliding surface. The direct method of Lyapunov is adopted in the existence condition analysis of SMC
[13]
. Time differential of Equation (7) is shown in Equation (10). To achieve the local reachability of SMC, Equation (11) should be satisfied.
Equation (11) can be rewritten as Equation (12).
Existence condition of SMC can be obtained from Equations (6)(8), and (12). The detailed condition including sliding coefficients is shown in Equation (13).
From Equation (13), the input voltage, load resistor, filter inductor, filter capacitor, and output voltage will all influence the existence condition. When designing the sliding coefficients, all these items will be considered.
 C. Equivalent Control of SMC
If PWM technique can be used in SMC, the switching frequency of the converter will be fixed. The control law will not be a bangbang type, and the control variable for the converter will become the duty cycle. Notably, PWM technique should not destroy the SMC property. In the SMC, the discrete control input
u
can be theoretically replaced by a smooth function known as the equivalent control signal
u_{eq}
. This equivalent control signal can be obtained by force time differentiation of Equation (7) equals zero, as shown in Equation (14).
Equivalent control signal is derived as Equations (15) and (16). At high switching frequency, the equivalent control is an effective duty cycle control
[4]
.
PWMbased SMC for threelevel fullbridge DCDC converter is shown in
Fig. 5
. In this controller, output voltage and capacitor current should be sampled. The SMC generates the equivalent duty cycle and uses this variable for the fullbridge DCDC converter.
PWMbased SMC for threelevel fullbridge DCDC converter.
 D. Static Output Voltage Error
In the PWMbased SMC, the equivalent control variable does not include the integral item of output voltage error. The secondorder hysteretic SMC becomes the firstorder PWMbased SMC after equivalent control conversion. The firstorder system has the ability to achieve good dynamic performance. However, it has no ability to resist the static output voltage error.
Fig. 6
shows the static error of PWMbased SMC under load step up and load step down obtained by simulation. The input voltage of the converter is 1000 Vdc. The ratio of the isolated transformer from primary side to secondary side is 2.0. The desired output DC voltage is 330 Vdc. In
Fig. 6
, the load before 0.1 s and after 0.16 s is 8W, and the load between 0.1 and 0.16 s is 4W. The output current causes voltage drop on the wire, transformer, and inductor. A large output current causes a large voltage drop and a large static output voltage error. The output voltage error cannot be eliminated because there is no integral item in the first order PWMbased SMC.
Static error of PWMbased SMC under load step up and load step down obtained by simulation.
Adding an integral item to the SMC is an effective and simple method to avoid and/or eliminate static output voltage error. This integral item will be analyzed in the next section.
IV. PWMBASED SMC WITH ADDITIONAL INTEGRAL
 A. Additional Integral Item
An integral item of output voltage error is added to the PWMbased SMC. The new control diagram is shown in
Fig 7
. The red part is the integral item to eliminate static output voltage error.
PWMbased SMC with additional integral for threelevel fullbridge DCDC converter.
With the additional integral item, a new duty cycle is shown in (17). From Equation (17), the duty cycle will always be regulated until the static error is eliminated. Meanwhile, integral coefficient
K_{i}
determines the speed of eliminating static error.
Fig. 8
shows the static error of PWMbased SMC with an additional integral under load step up and load step down obtained by simulation. The desired output DC voltage is 330 Vdc. In
Fig. 8
, the load before 0.1 s and after 0.16 s is 8 W, and the load between 0.1 s and 0.16 s is 4W. Moreover, no static output voltage error is observed all the time under different loads. The integral coefficient
K_{i}
is also equal to 100.
Static error of PWMbased SMC with integral under load step up and load step down obtained by simulation.
The dynamic performance of PWMbased SMC with an additional integral under load step up obtained by simulation is shown in
Fig. 9
. The three waveforms are obtained through three different integral coefficients. The integral eliminates the static error and simultaneously accelerates the dynamic response speed. When the integral coefficient is too large, the output voltage waveform overshoots and chattering may occur for a short time.
Dynamic performance of PWMbased SMC with additional integral under load step up obtained by simulation.
The whole system diagram of PWMbased SMC with an additional integral for threelevel fullbridge DCDC converter is shown in
Fig. 10
.
Entire system diagram of PWMbased SMC with additional integral for threelevel fullbridge dcdc converter.
The double loop PI controller can also eliminate static output error. However, PWMbased SMC in this paper is different from the PI controller. First, a double loop PI controller is designed based on a small signal model at a specific operation point. PI parameters are usually designed through a bode plot of the system. Double loop PI controller is a linear controller. PWMbased SMC is designed based on the trajectory of the state variable at a full range of load. SMC parameters are designed through a sliding surface. PWMbased SMC is a nonlinear controller. Second, load resistor and input voltage are reflected in the SMC. With this controller, the converter can achieve better robustness than the PI controller. Third and last, the outer loop generates current reference for the inner loop in a doubleloop PI controller. However, no current reference and no inner loop controller are found in PWMbased SMC proposed in this paper.
 B. Regulated Sliding Surface
The equivalent control function shown in Equation (16) is obtained from the trajectory of state variable, as shown in Equation (7). Sliding surface of the hysteretic SMC is
S
= 0. Adding an item in an equivalent control function means that the original trajectory of state variable should be regulated to adapt to the new control function. When the trajectory of the state variable is defined as Equation (18), the control function can be obtained as Equation (17). Thus, the sliding surface for PWMbased SMC with integral is
S
' = 0 .
where
A comparison of the old trajectory in Equation (7) with the new trajectory in Equation (18) shows that a double integral item is found added in the new trajectory. The original hysteretic SMC becomes a thirdorder system. Given that an equivalent control conversion will decrease the order of the SMC, the PWMbased SMC with an integral is a secondorder controller. This design is the fundamental reason this controller can eliminate static error.
 C. Regulated Existence Condition
A new state variable
x
_{4}
is introduced to eliminate static error, and the new state space function is shown in Equation (23). Given that the definition of the trajectory of state variable changes to Equation (18), the existence condition of SMC will be regulated. The direct method of Lyapunov is used to derive the new existence condition. The local reachability condition of SMC is shown in Equation (24).
where
The local reachability condition of Equation (24) can be written as Equation (25). The detailed condition, including sliding coefficients and integral coefficient, is shown in Equation (26).
V. DESIGN OF SLIDING COEFFICIENTS AND INTEGRAL COEFFICIENT
Sliding mode coefficients determine the dynamic response performance of SMC. Dynamic response performance is reflected by function
S
' = 0 , as shown in Equation (27). This thirdorder system needs to be analyzed. The coefficient
α
_{4}
represents the additional integral coefficient, as shown in Equation (21). The three sliding mode coefficients and integral coefficient are designed according to dynamic response performance of Equation (27). This design method is analyzed and used in
[11]
.
Analysis on a thirdorder system is relatively complex compared with that on a secondorder system. The function of integral coefficient is to eliminate the static error; its influence on the dynamic response of SMC is clearly shown in
Fig. 9
. Thus, the three sliding mode coefficients and the integral coefficient can be designed for PWMbased SMC with an additional integral. The three sliding mode coefficients mainly determine the dynamic response, and the integral coefficient mainly eliminates the static error.
 A. Design of the Three Sliding Mode Coefficients
Dynamic response of the SMC is reflected by Equation (28). Through time differential of Equation (28), a standard secondorder system can be obtained, as shown in Equation (29). In Equation (29),
ω_{n}
represents the undamped natural frequency and z represents the damping ratio. This standard secondorder system has three types of responses: under damped (0 ≤
ζ
＜ 1), critically damped (
ζ
= 1), and over damped (
ζ
> 1).
where
First, time constant
τ
for this converter should be determined. Time constant represents the speed of controller dynamic response. This time constant should be selected according to power rating and switching frequency. This phenomenon is an inherent characteristic of power converter.
When the damping ratio 0 ≤
ζ
≤ 1, the time constant is shown as Equation (30), expressed by damping ratio, and undamped natural frequency. The sliding mode coefficients can be obtained through Equations (29) and (30) and shown as Equations (31) and (32).
When the damping ratio
ζ
> 1, the time constant is shown as Equation (33), expressed by damping ratio and undamped natural frequency. The sliding mode coefficients can be obtained through Equations (29) and (33) and are shown as Equations (34) and (35).
 B. Design of the Integral Coefficient
If response speed is disregarded, then a very small integral coefficient will be able to eliminate static error. Thus, selecting a small integral coefficient is a simple, direct, and advisable design method. This coefficient also accelerates the response speed.
Fig. 11
shows a typical waveform of SMC under load step up. In the figure, the area
S
_{d}
represents the integral of output voltage error.
Typical waveform of SMC under load step up.
A comparison of two equivalent duty cycle functions, Equations (16) and (17), shows that the latter owns an extra item Δ
d
. This extra item is shown in Equation (36).
Equation (37) represents the desired duty cycle difference between PWWbased SMC and PWMbased SMC with integral. The integral coefficient can be obtained from Equation (37).
VI. EXPERIMENTAL RESULTS
Fig. 12
shows the threelevel fullbridge converter device. This device is designed for possible use on a ship’s power supply system. It is used to connect a high DC voltage bus and a low DC voltage bus in a ship. The load is uncertain in a wide range. The converter should have precise output voltage and highspeed dynamic response with the load changing.
Threelevel fullbridge converter device.
Input voltage is 1000 Vdc and output voltage is 330 Vdc. The load resistor may change from 2 W to 12 W. The PWMbased SMC proposed in this paper is used in this device to achieve highspeed dynamic response and precise output voltage. The parameters of this device and SMC are shown in
Table I
.
PARAMETERS OF THE THREELEVEL FULLBRIDGE CONVERTER
PARAMETERS OF THE THREELEVEL FULLBRIDGE CONVERTER
The maximum power of this device is about 50 kW. At this power rating, IGBT is selected as the main switch instead of MOSFET. To avoid high power loss, switching frequency of the converter is selected as 3.6 kHz. The parameter value of hardware is designed according to the inductor current ripple requirement and the output voltage ripple requirement.
Fig. 13
shows the softstart waveforms of output voltage and inductor current under ramp reference condition using PWMbased SMC with integral. This converter has a good softstart performance (
Fig. 13
). Therefore, the SMC can adapt for a large range of output voltage.
Softstart waveforms of output voltage under ramp reference condition using PWMbased SMC with integral.
Fig. 14
shows the steadystate experiment waveforms of threelevel fullbridge DCDC converter using PWMbased SMC with integral. The four waveforms are primary side current of transformer, drive signal of
Q
1, primary side voltage of transformer, and output voltage. This converter owns good steadystate performance (
Fig. 14
).
Steady state experiment waveforms of threelevel fullbridge converter using PWMbased SMC with integral.
Softstart waveforms and steady state waveforms obtained by PWMbased SMC without integral are similar with those obtained by PWMbased SMC with integral, except for the static output error. The static output error is difficult to observe in
Figs. 13
and
14
.
Static output voltage error is measured under a series of load. Relationship curves of DC output voltage and load resistor are shown in
Fig. 15
. The figure indicates that a heavy load causes a large static output error using PWMbased SMC. In addition, the PWMbased SMC with integral will not cause static output voltage error.
Relationship curves of DC output voltage and load resistor.
Fig. 16
shows the detailed experimental waveforms of static output voltage error using the two controllers under load step up and load step down. In load step up condition, the load changes from 8 W to 4 W. In load step down condition, the load change is contrary. The static output error is significant and influenced by the load [
Figs. 16
(a) and
16
(b)]. No static error is found and static error is not influenced by the load resistor [
Figs. 16
(c) and
16
(d)]. The advantage of PWMbased SMC with integral is significant.
Experimental waveforms of static output voltage error. (a) PWMbased SMC under load step up. (b) PWMbased SMC under load step down. (c) PWMbased SMC with integral under load step up. (d) PWMbased SMC with integral under load step down.
Fig. 17
shows the dynamic response performance of PWMbased SMC with additional integral under load step up using different integral coefficients. The load resistor changes from 8 W to 4 W. The integral eliminates the static error and accelerates the response speed. When
K
_{i}
> 0, no static output voltage error is observed. Meanwhile, a large integral coefficient will cause the output voltage waveform to overshoot and chatter for a short time. The time scale in
Fig. 16
is 200 ms/div, while the time scale in
Fig. 17
is 20 ms/div. Thus, in the time axis, the range in
Fig. 17
is only onetenth of the range in
Fig. 16
.
Figure 16
shows the significant static output voltage error in a large time range, and
Fig. 17
shows the significant specific dynamic response in a small time range.
Dynamic response performance of PWMbased SMC with additional integral under load step up using different integral coefficients. (a) K_{i} = 0; (b) K_{i} = 100. (c) K_{i} = 300. (d) K_{i} = 500. The load resistor changes from 8 W to 4 W.
VII. CONCLUSION
This paper proposes a PWMbased SMC for fullbridge DCDC converter that can eliminate the static output voltage error. A PWMbased SMC is derived through equivalent control concept. The secondorder hysteretic SMC becomes the firstorder PWMbased SMC after equivalent control conversion. The firstorder controller has the ability to achieve good dynamic performance. However, it has no ability to resist the static output voltage error. Thus, an integral item is added to the PWMbased SMC. The trajectory of the state variable of SMC is redefined, and the regulated existence condition is analyzed. The integral item eliminates the static error and accelerates the response speed. Experiment results on a threelevel fullbridge DCDC converter verified the control scheme and the design method proposed in this paper.
Acknowledgements
This project is supported by the Program for New Century Excellent Talents in the University of China (NCET110871) and supported by the National Key Basic Research Program 973 Project of China (2013CB035601)
BIO
Jilong Liu was born in 1988 in Hebei Province, China. He received his B.S. degree in 2010 from the School of Electrical Engineering, Xi’an Jiaotong University, Xi’an, China. From 2010 to 2012, he worked toward the M.S. degree in Xi’an Jiaotong University. He is currently working toward a Ph.D. degree in Xi’an Jiaotong University. His current research interests include distributed control scheme of power electronics system and highvoltage largecapacity power electronics equipment.
Fei Xiao was born in 1977 in Hubei Province, China. He received his B.S and M.S degrees in Electrical Engineering in 1999 and 2001, respectively, from the Naval University of Engineering, Wuhan, China. From 2003 to 2009, he was a lecturer at the Naval University of Engineering. He was promoted to an Associate Professor in 2009 and then a Full Professor in 2012. His current research interests include renewable energy generation, modeling and control of power electronics system, and highvoltage largecapacity power electronics equipment.
Weiming Ma was born in April 1960 in Jiangsu Province, China. He received his M.S. degree in Electrical Engineering in 1982 from the Naval University of Engineering, Wuhan, China, and his PhD degree in Electrical Engineering in 1996 from Tsinghua University, Beijing, China. He was elected to the Chinese Academy of Engineering and became an academician in 2001. He is currently a professor at the Naval University of Engineering and works as the Director of the National Key Laboratory for Vessel Integrated Power System Technology. He has published 2 monographs and more than 300 papers in several authoritative academic journals. He also achieved 38 national granted invention patents. Prof. Weiming Ma has received the following prizes: two firstclass prizes in the National Science and Technology Progress Award, two thirdclass prizes in the National Award for Invention, and four firstclass prizes in the Army Award for Science and Technology Progress.
Xuexin Fan was born in 1977 in Zhejiang Province, China. He received his B.S and M.S. degrees in Electrical Engineering in 2000 and 2003, respectively, from the Naval University of Engineering, Wuhan, China. From 2003 to 2013, he was a lecturer in the Naval University of Engineering. He was promoted to an Associate Professor in 2013. His research interests include independent power system and highvoltage multilevel converters.
Wei Chen was born in 1989 in Jilin Province, China. He received his B.S. degree from the Department of Automation from Tsinghua University, Beijing, China, in 2012. He is currently working toward his M.S. degree in the Naval University of Engineering, Wuhan, China. His current research interests include modeling and control of parallel connected threephase inverters system and largecapacity power electronics equipment.
Utkin V. I.
1993
“Sliding mode control design principles and applications to electric drives,”
IEEE Trans. Ind. Electron.
40
(1)
23 
36
DOI : 10.1109/41.184818
Lee S. H.
,
Joo Y.
,
Back J.
,
Seo J. H.
,
Choy I.
2011
“Sliding mode controller for torque and pitch control of PMSG wind power systems,”
Journal of Power Electronics
11
(3)
342 
349
DOI : 10.6113/JPE.2011.11.3.342
Venkataramanan R.
,
Sabanovic A.
,
Cuk S.
“Sliding mode control of dctodc converter,”
in Proc. 11th Annual Conf. of the IEEE Industrial Electronics Society (IECON)
1985
251 
258
Tan S. C.
,
Lai Y. M.
,
Tse C. K.
2008
“General design issues of sliding mode controllers in DCDC converters,”
IEEE Trans. Ind. Electron.
55
(3)
1160 
1173
DOI : 10.1109/TIE.2007.909058
Cardim R.
,
Teixeira M. C. M.
,
Assuno E.
,
Covacic M. R.
2009
“Variable structure control design of switched systems with an application to a dc–dc power converter,”
IEEE Trans. Ind. Electron.
56
(9)
3505 
3513
DOI : 10.1109/TIE.2009.2026381
Tan S. C.
,
Lai Y. M.
,
Cheung M. K. H.
,
Tse C. K.
2005
“On the practical design of a sliding mode voltage controlled buck converter,”
IEEE Trans. Power Electron.
20
(2)
405 
437
DOI : 10.1109/TPEL.2004.842977
Chen Y.
,
Kang Y.
2011
“The variablebandwidth hysteresis modulation slidingmode control for the PWMPFM converters,”
IEEE Trans. Power Electron.
26
(10)
2727 
2734
DOI : 10.1109/TPEL.2011.2158852
Lin B. R.
,
Chen C. C.
2014
“New threelevel PWM DC/DC converter – analysis, design and experiments,”
Journal of Power Electronics
14
(1)
30 
39
DOI : 10.6113/JPE.2014.14.1.30
Shi Y.
,
Yang X.
2013
“Zerovoltage switching PWM threelevel fullbridge DC–DC converter with wide ZVS load range,”
IEEE Trans. Power Electron.
28
(10)
4511 
4524
DOI : 10.1109/TPEL.2012.2232940
Tan S. C.
,
Lai Y. M.
,
Tse C. K.
,
Cheung M. K. H.
2005
“A fixedfrequency pulsewidth modulation based quasi sliding mode controller for buck converters,”
IEEE Trans. Power Electron.
20
(6)
1379 
1392
DOI : 10.1109/TPEL.2005.857556
Ackermann J.
,
Utkin V.
1998
“Sliding mode control design based on Ackermann’s formula,”
IEEE Trans. Autom. Control
43
(2)
234 
237
DOI : 10.1109/9.661072
SiraRamirez H.
1989
“A geometric approach to pulsewidth modulated control in nonlinear dynamical systems,”
IEEE Trans. Autom. Control
34
(2)
184 
187
DOI : 10.1109/9.21094
Tan S. C.
,
Lai Y. M.
,
Tse C. K.
2008
“Indirect sliding mode control of power converters via double integral sliding surface,”
IEEE Trans. Power Electron.
23
(2)
600 
611
DOI : 10.1109/TPEL.2007.915624
Zhang C.
,
Wang J.
,
Li S.
,
Wu B.
,
Qian C.
2015
“Robust control for PWMbased DCDC buck power converters with uncertainty via sampleddata output feedback,”
IEEE Trans. Power Electron.
30
(1)
504 
515
DOI : 10.1109/TPEL.2014.2299759
Hao X.
,
Yang X.
,
Liu T.
,
Huang L.
,
Chen W.
2013
“A slidingmode controller with multi resonant sliding surface for singlephase gridconnected VSI with an LCL filter,”
IEEE Trans. Power Electron.
28
(5)
2259 
2268
DOI : 10.1109/TPEL.2012.2218133
Guzman R.
,
Vicuna L. G.
,
Camacho A.
,
Matas J.
,
Castilla M.
,
Miret J.
“Active damping control for a three phase gridconnected inverter using sliding mode control,”
in Proc. 39th Annual Conference of the IEEE Industrial Electronics Society (IECON)
2013
382 
387
Comanescu M.
,
Xu L.
2006
“Slidingmode MRAS speed estimators for sensorless vector control of induction machine,”
IEEE Trans. Ind. Electron.
53
(1)
146 
153
DOI : 10.1109/TIE.2005.862303
Jin N.
,
Wang X.
,
Wu X.
2014
“Current sliding mode control with a load sliding mode observer for permanent magnet synchronous machines,”
Journal of Power Electronics
14
(1)
105 
114
DOI : 10.6113/JPE.2014.14.1.105
Janardhanan S.
,
Bandyopadhyay B.
2006
“Output feedback slidingmode control for uncertain systems using fast output sampling technique,”
IEEE Trans. Ind. Electron.
53
(5)
1677 
1682
DOI : 10.1109/TIE.2006.881962
Jafarian M. J.
,
Nazarzadeh J.
2012
“Minimum time regulation of DCDC converters in damping mode with an optimal adjusted sliding mode controller,”
Journal of Power Electronics
12
(5)
769 
777
DOI : 10.6113/JPE.2012.12.5.769
Shtessel Y. B.
,
Raznopolov O. A.
,
Ozerov L. A.
1998
“Control of multiple modular DCtoDC power converters in conventional and dynamic sliding surfaces,”
IEEE Trans. Circuits Syst. I, Fundam. Theory Appl.
45
(10)
1091 
1100
DOI : 10.1109/81.728863
Mazumder S. K.
,
Nayfeh A. H.
,
Borojevic A.
,
Gaya P.
,
López O.
2002
“Robust control of parallel DC–DC buck converters by combining integral variablestructure and multipleslidingsurface control schemes,”
IEEE Trans. Power Electron.
17
(3)
428 
437
DOI : 10.1109/TPEL.2002.1004251
Wai R.J.
,
Shih L.C.
2011
“Design of voltage tracking control for dc–dc boost converter via total slidingmode technique,”
IEEE Trans. Ind. Electron.
58
(6)
2502 
2511
DOI : 10.1109/TIE.2010.2066539
Oucheriah S.
,
Guo L.
2013
“PWMbased adaptive sliding mode control for Boost DCDC converters,”
IEEE Trans. Ind. Electron.
60
(8)
3291 
3294
DOI : 10.1109/TIE.2012.2203769
Li S.
,
Yang J.
,
Chen W.
,
Chen X.
2014
Disturbance ObserverBased Control: Methods and Applications
CRC Press
3 
17
Yang J.
,
Li S.
,
Yu X.
2013
“Slidingmode control for systems with mismatched uncertainties via a disturbance observer,”
IEEE Trans. Ind. Electron.
60
(1)
160 
169
DOI : 10.1109/TIE.2012.2183841
Liu J.
,
Xiao F.
,
Chen W.
,
Yang G.
“A novel control scheme for threelevel fullbridge converter achieving low THD output voltage,”
in Proc. 7th International Power Electronics Conference (IPEC)
2014
66 
71