The 6th International Conference on Renewable Power Generation (RPG)

19–20 October 2017

Based on STATCOM direct current full-order terminal sliding mode control

Ruo-bo Chen, Xue-mei Zheng, Song-nan Pang, Xin Li
Electrical Engineering Department, Harbin Institute of Technology, Harbin, People’s Republic of China
E-mail: xmzheng@hit.edu.cn

Published in The Journal of Engineering; Received on 11th October 2017; Accepted on 3rd November 2017

Abstract: This study proposes a static synchronous compensator (STATCOM), which cannot only detect and compensate for reactive current
according to the real-time change of grid reactive current, but also maintain the current phase and voltage phase of grid consistent. Aimed at the
shortages existing in the control of STATCOM at present, such as the parameters are difficult to adjust and the calculation is very complex, a
sliding mode controller for STATCOM reactive current inner-loop control system is designed. Finally, the control logic is evaluated in Matlab/
Simulink, and the simulation results are compared with the traditional proportional integral (PI) control design. It is shown that the proposed
method is able to achieve the response performance of the conventional PI method, and has stronger robustness against external disturbance.

1 Introduction operation. This operation mode is applicable to the grid voltage

waveform without distortion, the power system voltage distortion
The static synchronous compensator (STATCOM) is one of the flex-
is relative. When the grid voltage distortion, the application of in-
ible AC transmission system equipments controller by parallel connec-
stantaneous reactive power theory cannot accurately detect the re-
tion which can motion compensate the electric network’s reactive
active current. Also based on the generalised instantaneous
power; at the same time, it can compensate inductive and capacitive
reactive power theory, ip–iq operation even in the grid voltage dis-
reactive power, so STATCOM becomes one of the most reactive-load
tortion can accurately detect the reactive current. For the two above
compensation equipment in AC transmission system. Therefore, in the
methods, we can give the briefly introduction:
power system, STATCOM is widely used in recent years.
The control strategy of STATCOM is divided into direct current
control method and indirect current control method [1]. Simply speak- p–q algorithm: the application of instantaneous reactive power
ing, direct current control of the basic idea is to use the appropriate theory to decouple the load current into the fundamental compo-
pulse-width modulation (PWM) strategy for the system to deal with nents and harmonic components. As only reverse transform back
the instantaneous reactive current PWM pulse signal to drive the in- to the fundamental reactive power, so disconnect the channel, the
verter controllable power electronic device insulated gate bipolar tran- command current contains the fundamental reactive and harmonic.
sistor gate switch, and then control the inverter output [2, 3]. The The phase-locked loop (PLL) and the sine and cosine signal gener-
instantaneous value of the current and the instantaneous reactive ating circuit generate sinusoidal and cosine signals in phase with the
current of the system can be gotten if they are within the allowable de- a-phase voltage, and the converted low-pass filtered LPF obtains
viation, as is shown in Fig. 1. From the theoretical analysis, all kinds of the fundamental component.
PWM technology can control the current generated by the AC side of ip–iq operation: ip–iq operation to the standard sinusoidal signal
STATCOM. Also because of the direct control of the reactive current instead of voltage signal, through the linear look-up table interpol-
of STATCOM output, this method has the advantages of high control ation method, you can quickly and accurately get the corresponding
accuracy, good steady-state performance, fast transient response etc. angle corresponding to the ideal sine and cosine. The calculation
Additionally, the direct control can control the reactive power output steps are as follows:
of STATCOM more accurately than the indirect current control [4].
Therefore, various control strategies occupy a very important position (1) A phase voltage is used to obtain the phase angle of the A-phase
in the direct current control strategy with STATCOM [5]. voltage Ea through A-PLL and produce a sinusoidal signal sinωt and
Among various control strategies, sliding mode control (SMC) a corresponding cosine signal cosωt in phase with Ea.
has attracted significant amount of interest due to its fast global con- (2) The three-phase current on the load side is used to convert the abc
vergence, simplicity of implementation, order reduction, high ro- three-phase coordinate transformation into αβ two-phase coordinates,
bustness to external disturbances, and insensitivity to model and the corresponding value in the αβ coordinate is obtained.
errors and system parameter variations. Thanks to these advantages, (3) Use the following formula to find the active and reactive com-
SMC has been widely used in many applications, including electric- ponents of the current in α–β two-phase coordinates, as shown in
al, mechanical, chemical, industrial, civil, military, aeronautical, the below equation
and aerospace engineering [6, 7].



ip sin vt − cos vt ia
= (1)
2 Mathematical model of the STATCOM iq − cos vt − sin vt ib
2.1 Detection of reactive current
Based on the theory of instantaneous reactive power [8], the method (4) Pass the active component of the current ip through a digital
of detecting reactive power of three-phase circuit can be obtained by low-pass filter to obtain the fundamental active component of the
calculating p and q as the starting point, and it is called p–q current ip′.

J. Eng., 2017, Vol. 2017, Iss. 13, pp. 2067–2071 This is an open access article published by the IET under the Creative Commons
doi: 10.1049/joe.2017.0693 Attribution -NonCommercial License (http://creativecommons.org/licenses/by-nc/
3.0/)
 The model of the switching function on the two-phase αβ coord-
inate system is shown below

⎧ dV 3 
⎪
⎪ C dc = id sd + iq sq − iL
⎪
⎪
⎪
⎨ di
dt 2
L d
− vLiq + Rid = Ed − ud (5)
⎪
⎪ dt
⎪
⎪
⎪
⎩ L diq + vLi + Ri = E − u
d q q q
Fig. 1 STATCOM current decoupling control structure dt

where Ed, Eq are the three-phase power supply voltage converted to

(5) The fundamental active component iαf and iβf under αβ coordi- the dq coordinate system on the voltage value; sd, sq the unipolar
nates can be obtained by transforming ip′. binary logic switching function in the dq coordinate system; and
(6) Use the formula C23 to obtain the corresponding αβ three-phase id, iq the current value of the three-phase input current converted
coordinates of the fundamental active part of the current iaf, ibf, icf to the dq coordinate system.
Select the d-axis and the grid electromotive force vector coinci-
⎡ ⎤
 1 1 1 dence, the d-axis for the active component reference axis, q-axis
− − ⎥
2⎢ for the reactive component reference axis, can achieve active
T
C23 = C32 = ⎢ √2 √2 ⎥ (2)
3⎣ 3 3⎦ current and reactive current independent control, which can
0 − achieve the control of STATCOM.
2 2
For the state (5), regardless of the state variables, the original
state equation can be expressed as:
(7) Also then with the sampling current ia, ib, ic subtraction, to get ⎡ ⎤ ⎡E −u ⎤
the use of instantaneous power theory reactive power compensation
 R

− v d d
i̇d ⎢ L ⎥ id ⎢ L ⎥
required for reactive power detection current. =⎣ +
R ⎦ iq ⎣E −u ⎦ (6)
i̇q
−v − q q
As the p–q operation to collect the voltage and current of the grid, L L
and by the power grid voltage distortion and three-phase imbalance,
so the final choice is the second method. It can be seen from the above formula: in the d–q coordinate
system, the active current and reactive current are coupled with
each other, and any change in current in the direction of one
2.2 Current decoupling control based on dq coordinate system
axis causes a change in the current in the other direction. So
STATCOM main circuit is shown in Fig. 2. According to the basic the general proportional integral (PI) adjustment is difficult to
law of circuit science, STATCOM can be established based on achieve the desired control effect, for which the actual current
three-phase static coordinate system mathematical model, as will be detected to compensate for the cohesion between the two-
shown in the below equation phase current, thus eliminating the interaction between the
⎧ current to achieve the current decoupling control. We can intro-
⎪ dV  duce the new variable a, making the state equation into the fol-
⎪
⎪ C dc = ik sk − iL
⎪
⎪ dt lowing form:
⎪
⎪
⎪
⎨
k=a,b,c

dik 1  (3) ⎡ ⎤
⎪ L + Rik = Ek − Vdc sk − s
 R


⎪
⎪ dt 3 j=a,b,c j − 0
⎪
⎪  
id ⎢ L ⎥ id p1
⎪ =⎣ +
R ⎦ iq
⎪ (7)
⎪
⎩ Ek = ik = 0 iq p2
k=a,b,c k=a,b,c 0 −
L

where Ek (k = a, b, c) is the three-phase power supply voltage; ik

According to (6) and (7), we can get:
(k = a, b, c) the current value of the three-phase input; Vdc the cap-
acitance voltage; sk (k = a, b, c) is defined as:
ud = vLiq − Lp1 + Ed
 (8)
1 Upper arm is on, the lower arm switch is off .
sk = (4) uq = −vLid − Lp2 + Eq
0 Upper arm is off , the lower arm switch is on.

Also because the inverter side of the AC side of the voltage d,

q-axis components are:

ud = Vdc sd
(9)
uq = Vdc sq

The feedback control rate is:

vLiq − Lp1 + Ed
sd =
Vdc
(10)
−vLid − Lp2 + Eq
sq =
Fig. 2 STATCOM main circuit Vdc

This is an open access article published by the IET under the Creative Commons J. Eng., 2017, Vol. 2017, Iss. 13, pp. 2067–2071
Attribution -NonCommercial License (http://creativecommons.org/licenses/by-nc/ doi: 10.1049/joe.2017.0693
3.0/)
 For the active and reactive current loop, we adopt the following The q-axis current controller is designed to make the iq track the iq*.
PI control strategy: The error of the q-axis current is designed as follows:
 
K   eq = i∗q − iq (13)
ud = Ed − KiP + iI i∗d − id + vLiq
s
  (11) ∗ ∗ Eq − uq − Riq − vLid
K   ėq = i̇q − i̇q = i̇q − (14)
uq = Eq − KiP + iI i∗q − iq − vLid L
s
A full-order sliding mode manifold is designed in the following
The active current command value id* is proportional to the load form:
current and its value is the output of the voltage regulator:
sq = ėq + c1 |eq |a1 sgn(eq ) (15)
 
K  ∗ 
i∗d = KuP + uI Vdc − Vdc (12) where c1 > 0, 1 > α1 > 0 are constants.
s

Theorem 1: : The q-axis current error eq will converge to zero

within finite time if the full-order sliding mode surface sq is selected
as (15), and the control is designed as follows:
3 SMC current control design
The traditional PI method for controller design is the most common uq = uqeq + uqn (16)
method of engineering application, but because of the non-linearity
∗
of the STATCOM device itself, the inaccuracy of parameters and uqeq = Eq − Riq − vLid − Li̇q − c1 L|eq |a1 sgn(eq ) (17)
measurement, and the uncertainty introduced by the system
network side disturbance, it is not easy; and linear and robust u̇qn + Tuqn = vq (18)
methods because of the need for real-time identification of the  
state, the parameters of high precision, large amount of calculation, vq = − kd1 + kT 1 + h1 sgn(sq ) (19)
and other unfavorable factors limit the engineering application.
SMC includes conventional linear sliding mode control and ter- where η is a positive constant; kd1 a constant defined in kd1 . |u̇qn |,
minal sliding mode (TSM) control. The former is asymptotically T ≥ 0 and kT1 are selected to satisfy the following condition:
stable, while the latter is finite-time stable. The design of the kT1 . |Tuqn |.
SMC systems mainly consists of two steps: the choice of the
sliding mode surface, and the design of the SMC. The sliding Proof: The following Lyapunov function is considered:
mode surface is chosen such that an SMC system can behave in a
desirable fashion. The controller is designed to guarantee the exist- V = 0.5s2q (20)
ence condition of the sliding mode, so the system can be driven to
reach the sliding mode surface in finite time and remain on it there- the sliding mode manifold (13) can be rewritten as follows:
after [9, 10].
The control aim is to make id and iq better track their reference
V̇ = sq ṡq
values, so that compensation for reactive power is more accurate  
and the power factor is closer to 1. = sq vq − Tuqn
Fig. 3 is a block diagram of double closed-loop control using
   
full-order sliding mode algorithm. The following design of the
= sq − kd1 + kT 1 + h1 sgn(sq ) − Tuqn (21)
inner loop is based on the control block diagram.
  
= − kd1 + kT1 + h1 sq  − Tuqn sq
3.1 q-Axis current controller design
       
     
= −kd1 sq  − kT 1 sq  + Tuqn sq − h1 sq  ≤ 0

It means that the system will reach to sq = 0 in finite time and then eq
will converge to zero within finite time along sq = 0.

Remark: Through the design of control law (16)–(19), the negative

index terms of eq are prevented in the control laws. So singularity
can be avoided. Meanwhile, (18) is equal to a low-pass filter,
where vq is the input and uqnthe output of the filter. Although
there is switch function in vq, uqn is a continuous control signal.

3.2 d-Axis current controller design

The d-axis current controller is designed to make the id track the id*.
The error of the d-axis current is designed as follows:

ed = i∗d − id (22)

∗ ∗ Ed − ud − Rid + vLiq
ėd = i̇d − i̇d = i̇d − (23)
Fig. 3 Block diagram of full-order TSM control L

J. Eng., 2017, Vol. 2017, Iss. 13, pp. 2067–2071 This is an open access article published by the IET under the Creative Commons
doi: 10.1049/joe.2017.0693 Attribution -NonCommercial License (http://creativecommons.org/licenses/by-nc/
3.0/)
A full-order sliding mode manifold is designed in the following
form:
 a  
sd = ėd + c2 ed  2 sgn ed (24)

where c2 > 0, 1 > α2 > 0 are constants.

Theorem 2: The d-axis current error ed will converge to zero within

finite time if the full-order sliding mode surface sd is selected as
(24), and the control is designed as follows:

ud = udeq + udn (25)

∗  a   Fig. 4 A-phase voltage and current waveform
udeq = Ed − Rid + vLiq − Li̇d − c2 Led  2 sgn ed (26)

u̇dn + Tudn = vd (27)

 
vd = − kd2 + kT 2 + h2 sgn(sd ) (28)

where η is a positive constant; kd2 a constant defined in kd2 . |u̇dn |,

T ≥ 0 and kT2 are selected to satisfy the following condition:
kT 2 . |Tudn |.

Proof: The following Lyapunov function is considered:

V = 0.5s2d (29)

V̇ = sd ṡd
  Fig. 5 STATCOM absorption of active and reactive power
= sd vd − Tudn
   
= sd − kd2 + kT 2 + h2 sgn(sd ) − Tudn (30)
   
= − kd2 + kT2 + h2 sd  − Tudn sd
       
= −kd2 sd  − kT 2 sd  + Tudn sd − h2 sd  ≤ 0

It means that the system will reach to sd = 0 in finite time and then ed
will converge to zero within finite time along sd = 0.

4 Case study
STATCOM’s ability to track the inductive, capacitive, and reactive
power compensation and reactive load is verified by this simulation.
The simulation parameters of the STATCOM as follows: AC side
R = 0.01 Ω, AC side L = 4.3 mH, three-phase symmetrical input
power phase voltage U = 40 V, Vdc of the given value is 110 V,
AC-side voltage standard unitary value of 56.6 V. Fig. 6 Load side active current and reactive current waveform
Simulation results are shown in Fig. 4.
As is shown in Fig. 4, the system in the 0.1 s moment load sen-
of reactive load. Simulation results show that the proposed method is
sitive switch to the capacitive, at 0.2 s time and then by the capaci-
able to achieve the response performance of the conventional PI
tive switch to emotional. Before 0.1 s time, since the load is
method, and has stronger robustness against external disturbance.
inductive, the system needs to provide inductive reactive power,
since the work has been put into STATCOM, the compensation re-
active current required by the load, the system voltage, and system 6 Acknowledgment
current phase. Fig. 5 reflects the above situation occurred when
STATCOM absorbed active and reactive. The system provides cap- This work was supported by the National Natural Science
acitive and reactive power between 0.1 and 0.2 s, providing induct- Foundation of China (51577039), and HeiLongJiang Province
ive reactivity at other times. Natural Science (E201426).
As is shown in Fig. 6, the d-axis current and the q-axis current
vary with their reference values. In this way, such sliding mode 7 References
control algorithm to control the steady-state error is small.
[1] Encarnação L., Fernando Silva J.: ‘Stability condition based sliding
5 Conclusions mode modulators for multilevel power converters’. 2009 35th
Annual Conf. IEEE Industrial Electronics, 2009, pp. 877–882
In this paper, an SMC for STATCOM reactive current inner-loop
[2] Pati S., Kar S.K., Mohanty K.B., ET AL.: ‘A sliding mode controller
control system is proposed. STATCOM load conditions in a based STATCOM with battery storage for voltage and frequency sta-
variety of inductive, capacitive, mutation nature of the load is still bilization in a micro-grid’. 2016 Int. Conf. Circuit, Power and
a good tracking ability and capacity reactive power compensation Computing Technologies (ICCPCT), 2009, pp. 1–7

This is an open access article published by the IET under the Creative Commons J. Eng., 2017, Vol. 2017, Iss. 13, pp. 2067–2071
Attribution -NonCommercial License (http://creativecommons.org/licenses/by-nc/ doi: 10.1049/joe.2017.0693
3.0/)
 [3] Mohanty K.B., Pati S.: ‘Voltage profile improvement of micro-grids [7] Tang Y., He H., Ni Z., ET AL.: ‘Adaptive modulation for DFIG and
using SMC based STATCOM’. 2016 Annual IEEE Systems Conf. STATCOM with high-voltage direct current transmission’. IEEE
(SysCon), 2016, pp. 1–8 Trans. Neural Netw. Learn. Syst., 2013, pp. 1762–1772
[4] Hou R., Cao T., Qian B.: ‘Design of sliding mode controller for [8] Mohaddes M., Suriyaarachchi D.H.R., Karawita C.: ‘Performance en-
STATCOM with LCL filter’. 2010 Int. Conf. Computer, hancement of modular multi-level converters using a direct phase
Mechatronics, Control and Electronic Engineering, 2010, pp. 80–83 current control technique’. 13th IET Int. Conf. AC and DC Power
[5] Li S., Li W., Xu W., ET AL.: ‘The direct current control method of Transmission (ACDC 2017), 2017, pp. 1–4
STATCOM and it’s simulation’. 2013 Third Int. Conf. Intelligent [9] Pati S., Mohanty K.B., Dash S.S., ET AL.: ‘Sliding mode controller
System Design and Engineering Applications, 2013, pp. 1426–1429 based decoupled control of STATCOM for voltage profile improve-
[6] Sun Y., Li S.: ‘Comparison of conventional and a novel direct- ment in a micro grid system’. 2015 Int. Conf. Energy, Power and
current vector control approaches for a LCL-filter based Environment: Towards Sustainable Growth (ICEPE), 2015, pp. 1–6
STATCOM’. 2015 IEEE Power & Energy Society General [10] Bouzid A.M., Cheriti A., Doumbia M.L., ET AL.: ‘Comparative study
Meeting, 2015, pp. 1–5 between decoupled control with sliding mode & feedback lineariza-
tion control applied to STATCOM’. 2014 IEEE 27th Canadian
Conf. Electrical and Computer Engineering (CCECE), 2014, pp. 1–6

J. Eng., 2017, Vol. 2017, Iss. 13, pp. 2067–2071 This is an open access article published by the IET under the Creative Commons
doi: 10.1049/joe.2017.0693 Attribution -NonCommercial License (http://creativecommons.org/licenses/by-nc/
3.0/)