Académique Documents
Professionnel Documents
Culture Documents
Tine Vandoorn, Bert Renders, Frederik De Belie, Bart Meersman and Lieven Vandevelde
vg (t)
In the current paper, the control is performed in the time δff (t) = . (4)
domain without transformation of reference frame and by vdc (t)
using conventional PI-regulators. A single-phase grid is Using the following transfer function of duty-ratio to in-
studied and, in further research, this will be extended to verter current:
a three-phase grid. The voltage of the grid is controlled by îL (s) Vdc
an inner current control loop and an outer voltage control = (5)
δ̂(s) sL
loop. To constrain the inverter current within its safety lim-
its, a fast current controller is used in the inner loop, having the inner PI-regulator can be tuned.
a reference current obtained by the outer-loop voltage reg-
ulation. The input of the inner PI-regulator is the measured current
An advantage of the inner current control loop is its easy iL compared to its reference value i∗L . The output of this
current limit function. More advantages of the inner cur- regulator is the desired duty-ratio δ of the PWM module.
rent control loop are described in [8]. To obtain i∗L , the reference grid voltage vg∗ is compared
to its measured value vg and controlled by a second PI-
2. Control Strategy regulator. The PI-regulator to control the grid voltage vg is
tuned by using the transfer function
In this paper, a control strategy for inverters in island mode vˆg 1
is described, the topology of the VSI is shown in Fig. 1. In = (6)
ˆ
ic sC
this figure the grid is represented as a load. The aim is to
control both the amplitude and the frequency of the grid and a Padé approximation for delay time as a result of the
voltage vg (t). A schematic overview of the control strategy sample and hold procedure. The output of the outer PI-
is shown in Fig. 2. regulator is ∆i∗c , with ∆i∗c a small-signal deviation of i∗c .
The input of the inner PI-regulator is
In the fast inner current control loop, the measured in-
verter current iL (t) is compared with the set value i∗L (t) ∆iL = i∗L − iL (7)
of this current. The obtained current error is presented to a
discrete proportional-integral controller. The output of the and ∆iL consists of two parts:
current controller is the set value of the switching voltage
∗
vsw (t) or, equivalently, the duty-ratio δ(t). To obtain bet- ∆iL = ∆iL,1 + ∆iL,2 . (8)
ter disturbance rejection, a duty-ratio feed-forward branch
is added to the output of the current controller [10]. The In the previous equation ∆iL,1 = ∆i∗c using eq. (6) as ∆iL
sum of the duty-ratio and the duty-ratio feed-forward is is changed in order to decrease the difference between vg
the input of the PWM-unit, which calculates the switching and vg∗ . ∆iL,2 is an open-loop feed-forward of ∆iL or
signals for the inverter.
∆iL,2 = ig + i∗c − iL . (9)
The design of the current controller is based on:
The inner PI-regulator forces iL to its reference value.
diL (t)
L = vsw (t) − vg (t), (1) Another method to derive the transfer functions (5) and (6)
dt is by using the state space model:
and vsw the switching voltage averaged over a PWM pe-
dx
riod, given by = Ax + Bu (10a)
dt
vsw (t) = δ(t).vdc , (2)
y = Cx + Du (10b)
with δ ∈ [−1, 1]. Further in this paper, the time depen-
dence of the following functions will be taken implicitly. which gives:
Transformation to a small signal model in the Laplace do-
main results in 1 1
d iL 0 − iL 0 vsw
= L + L .
Vdc δ̂(s) δ0 v̂dc (s) v̂g (s) dt 1 1
îL (s) = + − , (3) vg 0 vg 0 − ig
sL sL sL C C
(11)
with δ0 the average duty-ratio and where hatted values x̂
denote small deviations from the steady state value of x. The PI-regulator must be robust for disturbances. The
This equation shows that the current of the inverter iL is bandwidths of the two PI-regulators are different and the
200 Obtained
Desired
Voltage Vg (V)
RL L S1 S4 100
Vdc 0
R Cf -100
-200
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18
Time (s)
S2 S3 (a) Overview
235
Obtained
Voltage Vg (V)
230 Desired
Figure 3: Resistive load: topology
225
220
Voltage Vg (V)
6 Desired
4
iˆL sC 2
= Vdc (12) 0
δ̂ 1 + s2 LC -2
-4
-6
and -8
vˆg 1 0.0548 0.0549 0.055 0.0551 0.0552 0.0553
= . (13) Time (s)
ˆ
ic sC
(c) Detail: phase error (zero-crossing of vg )
As the switching period of the PWM is at least ten times
shorter than the time constant of transfer function (12), the
Figure 4: Resistive load: grid voltage vg with its reference
response to a step input 1s in δ, viz an exponential function,
di value
can be approximated as linear with gradient dt :
230
Grid Power
1500 Active power (W)
1000 Reactive power (VA)
225
500
220 0
-500
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18
215
0.058 0.059 0.06 0.061 time (s)
Time (s)
Figure 8: Switching load: grid active and reactive power
Figure 5: Switching load: detail amplitude error in grid (vg and ig )
voltage vg
30
20
RL L S1 S4
Current IL (A)
10
0 Obtained
R2 R Cf
-10 Desired
Vdc
Ih
-20
-30
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time (s)
S2 S3
Figure 6: Switching load: sampled inverter current iL with
its reference value
Figure 9: Harmonic load: topology
20
10 235
Cirrent Ig (A)
Obtained
Voltage Vg (V)
0 230 Desired
-10 225
-20 220
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
T ime (s)
215
0.058 0.0585 0.059 0.0595 0.06 0.0605 0.061 0.0615
Time (s)
Figure 7: Switching load: grid current ig
Figure 10: Harmonic load: detail amplitude error in grid
voltage vg
transient behaviour can be concluded from this figure. The
inverter current iL equals ic + ig and the switching ripple
in iL is almost completely absorbed in the capacitor cur-
rent ic . The current ig is shown in Fig. 7, its peak current C. Harmonic load
before the switching instant is approximately 9.2 A and, as
expected, this is doubled (18.4 A) by halving the load re- In the next simulation, the load consists of the previous
sistance. Also, when comparing ig to iL a phase-difference load R of 25 Ω. This resistance is connected in parallel
is obtained as the capacitor injects a reactive current into with a current source which is placed in series with a sec-
the grid. ond resistance R2 of 25 Ω as shown in Fig. 9. The current
source has an amplitude of 5 A and a frequency of 250 Hz.
Before the transient the active power exported to the grid The simulation results of the obtained grid voltage vg and
equals the reference grid voltage vg∗ is shown in detail in Fig. 10
2
and the overview is analogous with Fig. 4(a). The error
vg,rms of the obtained voltage compared to the desired voltage is
= 1.058 kW, (18)
R small.
after 0.06 s the active power equals The sampled inverter current iL contains a fifth harmonic
component next to the fundamental component, as shown
2 in Fig. 11.
vg,rms
= 2.116 kW. (19)
0.5R
D. Robustness
The grid active and reactive power, calculated with the ac-
tive & reactive power block of MatLab SimPowerSystems, In this paragraph, the robustness to measurement inaccu-
are shown in Fig. 8. racy and parameter faults is studied.
20
20 Desired
Current IL (A)
Current I L (A)
10
10 Obtained
0
Desired
-105 Obtained
0
-20 -10
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 -20
T ime (s)
0.045 0.05 0.055 0.06 0.065
Time (s)
Figure 11: Harmonic load: sampled inverter current iL
with its reference value Figure 13: White noise: sampled inverter current iL with
its reference value
200
Voltage Vg (V)
100 235
Obtained
Voltage Vg (V)
0
230 Desired
-100-
Desired 225
-200 Obtained
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 220
T ime (s)
220
215 40 Desired
0.058 0.059 0.06 0.061 0.062
Current IL (A)
Obtained
Time (s) 20
-40
Figure 12: White noise: grid voltage vg with its reference 0 0.03 0.05 0.06
0.01 0.02 0.04 0.07
value T ime (s)
In the sampled inverter current iL a fifth harmonic caused In the sampled inverter current iL a fifth harmonic caused
by the load is obtained next to the ground wave, as shown by the load is obtained next to the ground wave, as shown
in Fig. 13, where one fundamental period of 20 ms is in Fig. 15, it is shown that the parameter sensitivity is suf-
shown. The error of the obtained current compared to the ficiently low.
desired current increases under increasing measurement
noise. The disturbance rejection of the inner loop is suf- 4. Conclusions
ficient. The inner loop is fast in comparison with the outer
voltage loop which results in an even better disturbance The control of the voltage of a single-phase microgrid with
rejection of the outer loop. one VSI is obtained. This control has two separate con-
trol loops: a voltage control loop and a fast current control
2) Parameter faults loop. The output of the voltage control loop is the input of
the current control loop, using separation of variables. The
In this simulation, the load consist of the previous har- control is studied under different loads, transient effects
monic load of Fig. 9. The real filter capacitor C equals and other disturbances resulting in a robust control strategy
with sufficiently low parameter sensitivity. An advantage 2000.
of this approach is that it can be adopted to control both [5] M. Prodanović and T. C. Green, “High-quality power gen-
single- and three-phase microgrids. In future work micro- eration through distributed control of a power park micro-
grids with multiple VSI’s will be considered. grid,” IEEE Trans. Ind. Electron., vol. 53, no. 5, pp. 1471–
1482, Oct. 2006.
Acknowledgement [6] N. Pogaku and T. Green, “Harmonic mitigation through-
out a distribution system: a distributed-generator-based so-
This research of Tine Vandoorn is funded by the Special lution,” IEE Proc. Gener. Transm. Distrib., vol. 153, no. 3,
pp. 350–358, May 2006.
Research Fund (BOF) of Ghent University. The research
was carried out in the frame of the inter-university At- [7] R. Teodorescu and F. Blaabjerg, “Flexible control of small
wind turbines with grid failure detection operating stand-
traction poles IAP-VI-021, funded by the Belgian Govern-
alone and grid-connected mode,” IEEE Trans. Power Elec-
ment. tron., vol. 19, no. 5, pp. 1323–1332, Sept. 2004.
[8] M. Prodanović, “Power quality and control aspects of par-
References allel connected inverters in distributed generation,” Ph.D.
dissertation, University of London, Imperial College, 2004.
[1] T. Green and N. Pogaku, “A model citizen approach to inte-
[9] K. De Brabandere, T. Loix, K. Engelen, B. Bolsen, J. V.
grating inverter-based distributed generation.” World Re-
Keybus, J. Driesen, and R. Belmans, “Design and opera-
newable Energy Congress (WREC), 2005, pp. 1101–1108.
tion of a phase-locked loop with kalman estimator-based
[2] E. Barklund, N. Pogaku, M. Prodanović, C. Hernandez- filter for single-phase applications,” in Conf. of the IEEE
Aramburo, and T. C. Green, “Energy management in au- Ind. Electron. Society (IEEE IECON’06), Paris, France,
tonomous microgrid using stability-constrained droop con- pp. 525–530, Nov. 7-10, 2006.
trol of inverters,” IEEE Trans. Power Electron., vol. 23,
[10] D. M. Van de Sype, K. De Gussemé, A. P. M. Van den Boss-
no. 5, pp. 2346–2352, Sept. 2008.
che, and J. A. Melkebeek, “Duty-ratio feed-forward for dig-
[3] R. H. Lasseter, A. Akhil, C. Marnay, J. Stephens, J. Da- itally controlled boost PFC converters,” IEEE Trans. Ind.
gle, R. Guttromson, A. Meliopoulous, R. Yinger, and J. Eto, Electron., vol. 52, no. 1, pp. 108–115, Feb. 2005.
“The CERTS microgrid concept, white paper on integration
[11] B. Renders, K. De Gussemé, W. R. Ryckaert, and L. Vande-
of distributed energy resources.” California Energy Com-
velde, “Input impedance of grid-connected converters with
mission, Office of Power Technologies - U.S. Department
programmable harmonic resistance,” IET Electr. Power
of Energy, Apr. 2002.
Appl., vol. 1, no. 3, pp. 355–361, May 2007.
[4] D. Moskovitz, “Profits and progress through distributed re-
sources.” Regulatory Assistance Project, Tech. Rep., Feb.