Beruflich Dokumente
Kultur Dokumente
The low voltage DC microgrid consider for this study-case Fig. 3: Control diagram for ESS.
is composed by two PV-based active generators (PV+ESS) and
a critical load (see Fig. 1). The microgrid is formed around a IESSi is the output current of each ESS. In this case, a typical
standard 48Vdc common bus, which is a kind of low voltage double-loop VCM controller is implemented for a bidirectional
DC power distribution system that is widely used, since it non-isolated buck converter, as can be seen in Fig.3. When the
allows working on a live conductor with minimum risk for same virtual resistance (Rd) is applied to each ESS control
personal injury and without special safety requirements [21]. loop, the current is equally shared between ESS (see Fig. 4(a)).
Moreover, non-isolated buck DC/DC converters operating in
continuous mode are used for the conversion stage of each Under the discharge of the battery, for balancing the SoC
RES and ESS as proposed in [15]. between ESS, the ESS with the highest SoC should supply
more power to the microgrid than the other. On the contrary,
In isladed operation, it is expected that the RES operate when the batteries are being charged the ESS with the smallest
by using a maximum power point tracking (MPPT) algorithm, SoC should get more energy from the microgrid than the
in order to obtain from the RES the maximum amount of other. This behavior can be achieved by weighting the virtual
available energy. Fort that reason, RES operate as constant resistance (Rd) by a factor i as is shown in Fig. 4(b), where
power sources following the current reference given by the the largest SoC have been assumed for ESS2. Therefore, (1)
MPPT algorithms. This current reference can be obtained from can now be rewritten as follows:
MPPT methods as the one proposed in [22]. However MPPT
VDC = VDC Rd i IESSi (2)
strategies are out of the scope of this paper, interested readers
may also refer to [23]. To get back to the point, the inductor
VDC (V) V (V)
current of the converter is controlled by typical inner current V*
DC
V*DC *
VDC=VDC-RdIESS2=VDC-RdIESS1
* V*
loops. Fig. 2 shows the scheme of the inner current control for DC DC
*
VDC=VDC-RdD2IESS2
each RES. VDC
VDC VDC
Meanwhile, the ESS operate in voltage control mode
*
VDC=VDC-RdD1IESS1
(VCM) being responsible of regulating the bus voltage. At
this mode, the batteries will be charged or discharged in order
to compensate the unbalance between the energy generated IESS1 = IESS2 I (A) IESS1 IESS2 I (A)
by RES and load consumption [20], [24]. Commonly, The
power unbalance is equally shared between ESS by means (a) (b)
of conventional droop control loop [5]. Therefore, the voltage Fig. 4: Droop characteristics: (a) Equal virtual resistance Rd
at the common bus (VDC ) is established by the following (b) Weighted virtual resistance Rd.
equation:
Finally, the microgrid system is complemented with an en-
VDC = VDC Rd IESSi (1)
ergy management system (EMS) that execute the equalization
where Rd is the virtual resistance of the droop control loop, algorithm in order to obtain the values for 1 and 2 as can
VDC is the voltage reference of the common DC bus, and be seen in Fig. 5.
294
I RES1
(PV1+ESS1)
DC/DC
DC/
C/DC DC/DC
DC/
C/DC
I ESS1
PV Array
a Converter
C
Co nverte
t r Converter
C
Co nverte
t r Battery
Batt
ttery
ry Array
ay 1
I RES 1 I ESS 1 I Bat1
VPV
P I PV VDC SoC Bat 1
DC
PWM
W W
PWM *
Currentt
Current I ref + + V DC
C
I ref Current
C rrent
Cu contr
t ol
control PI
P
- -
MPPT
MP
MPPT contr
t ol
control loop
l op
lo o
VDC
D D1
l op
loo
loop
Rd EMS
Primary
P im
Pr i ary
r Co
C
Control
ntr
t ol RES
RES 1
RE Primary
P im
Pr i ary Control
r Co
C t ol Battery
ntr Batt
ttery
r 1
I Load
Load I Load
Equalization
DC Common Bus
Algorithm
PV Array
ay DC/DC
C/DC Converter
DC/ C nverte
Co t r DC/DC
C/DC Converter
DC/ C nverte
Co t r Batt
tterryy Array
Battery ay 2
I RES 2 I ESS 2 I Bat 2
VPV
P IWTG SoC Bat 2
VDC
D
Primary
P im
Pr i ary Control
r Co
C t ol RES
ntr RES 2
RE Primary
P im
Pr i ary Control
r Co
C t ol Battery
ntr Batt
ttery
r 2 D2
I ESS 2
(PV2+ESS2)
I RES 2
Fig. 5: Diagram of the DC microgrid with conventional inner control loops.
88
It is important to say that this paper will only consider the
86
operation of the microgrid when the ESS are being charged I Bat1
84
or discharged. Anyhow, the operation of the microgrid should mSoC1
72
68
The algorithm is based on the fact that the rate of change of 0 1 2 3 4
Time (s)
5 6 7
295
where SoCBati = SoC(t)Bati SoC(0)Bati . Then (5) Meassure
External Data
takes the form:
IESSi = mSoCi KBati (6) Solve
K Bat1mSoC1 K Bat 2 mSoC 2 I RES1 I RES 2 I Load 0
being, mSoCi the rate of the SoC, and KBati a proportionality SoC (0) Bat1 mSoC1
't SoC (0) Bat 2 mSoC 2
't
constant that depends on the main parameters of the ESS.
In a general case, where n distributed active generators No If Ibat<0 Yes
(PV+ESS) are integrated into the microgrid, it is easy to derive
the Kirchhoff current law equation for the common node as:
n
n
If SoC (0) Bat1 ! SoC (0) Bat 2 If SoC (0) Bat1 ! SoC (0) Bat 2
IESSi + IRESi ILoad = 0 (7)
Yes No
i=1 i=1
where, (ILoad ) is the load current, (IRESi ) is the power D1 Cmin Cmax
supplied for each RES, and (IESSi ) is the current at each ESS. D2 D1 k Bat1mSoC1 k Bat 2 mSoC 2
Wait until
In particular, for the proposed DC microgrid shown in Fig. Time 't
No Yes
5, by combining (6) and (7), we have
D2 Cmin Cmax
2
2
D1 D 2 k Bat 2 mSoC 2 k Bat1mSoC1
mSoCi KBati + IRESi ILoad = 0 (8)
END
i=1 i=1
Fig. 7: Equalization algorithm.
Moreover, in order to perform the SoC equalization, it is
required that the straight-line equations, that represents the TABLE I: Main Parameters of the microgrid
behavior of the SoC within a dened period of time (t),
Parameter Symbol Value
are equalized between then as: Nominal Bus Voltage
VDC 48 (V)
10 ()
SoC(0)Bat1 + mSoC1 t = SoC(0)Bat2 + mSoC2 t (9) Nominal Load
Maximum (RES) Power Rating
RLoad
PRESmax 300 W
Nominal Voltage V bat 48V
to be more precise SoC(t)Bat1 = SoC(t)Bat2 . Nominal Battery Capacity Cbat 0.02 (Ah)
Period of the Equalization Algorithm t 5 (s)
To get back to the point, the main task of the equalization Duty-cycle D 0.93 (Ah)
algorithm is to solve the equation system composed by (8) Charging/discharging efciency Bati 100
and (9), in order to obtain the adequate values for mSoCi Virtual Resistance Rd 0.5 ()
296
51.0 0 51.0 0
Equalization
hjhkhjlkhjkhjjjj Equalization 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Equalization
Equalization 17 18 19 20 21 22 23 24
Algorithm Algorithm Algorithm Time (Sec.) Algorithm
{
{
{
100 100
90 SoCBat2 90 SoCBat2
SoC (%)
SoC (%)
#1:2 #1:2
80 80 #1:1
#1:1
70 70
60 0
SoCBat1 SoCBat1
60 0
0 1 2 3 4 5 6 7 8 9 10 11 2 22 23 24
1 12 13 14 15 16 17 18 19 20 21 0 1 2 3 4 5 6 7 8 6 17 18 19 20 21
1 12 13 14 15 16
9 10 11 2 22 23 24
Time (s) Time (s)
(a) (a)
-0.2
IBat (A)
-0.5 #1:2
-0.4
-1.0 -0.6
IBat1 IBat2
-1.5 0 -0.8 0
0 1 2 3 4 5 6 7 8 1 17 18 19 20 2
11 12 13 14 15 16
9 10 1 21 22 23 24 0 1 2 3 4 5 6 7 8 9 10 1
11 12 13 14 15 16 21 22 23 24
6 17 18 19 20 2
Time (s) Time (s)
(b) (b)
0
0
Diff(SoC)
Diff(SoC)
-5
-5
#1:1 #1:1
-10 0
-10 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (s)
Time (s)
(C)
(c)
Fig. 8: ESS Charge when CBat1 = CBat2 : (a) SoC, (b) IBat , Fig. 10: ESS Charge when CBat1 < CBat2 : (a) SoC, (b) IBat ,
(c) Diff(SoC). (c) Diff(SoC).
49.0 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Equalization
Equalization 17 18 19 20 21 22 23 24 25
48.0 0
Algorithm Time (Sec.) Algorithm 0 1 2 3 4 5 6 Equalization
7 8 9 10 11 12 13 14 15 16 Equalization
17 18 19 20 21 22 23 24 25
SoCBat2
{
#1:2
{
80 #1:1
Algorithm Time (Sec.) Algorithm
SoCBat2
{
{
#1:2
SoC (%)
60 80 #1:1
SoCBat1
SoC (%)
40 60
SoCBat1
20 0 40
0 1 2 3 4 5 6 7 8 9 10 11 2 22 23 24 25
1 12 13 14 15 16 17 18 19 20 21
20 0
Time (s) 0 1 2 3 4 5 6 7 8 9 10 11
1 12 13 14 15 16 17 18 19 20 21
2 22 23 24 25
(a)
Time (s)
(a)
2.0 IBat2
1.5 #1:2
IBat2
IBat (A)
2.0 #1:2
1.0 1.5
IBat (A)
0.5 1.0
#1:1
IBat1
0.0 0 0.5 #1:1
0 1 2 3 4 5 6 7 8 9 10 11 2 22 23 24 25
1 12 13 14 15 16 17 18 19 20 21 IBat1
0.0 0
Time (s) 0 1 2 3 4 5 6 7 8 11 12 13 14 15 16 17 18 19 20 2
9 10 1 21 22 23 24 25
(b)
Time (s)
(b)
5
Diff(SoC)
0 0
-5 #1:1
Diff(SoC)
-5
-10 #1:1
-15 0 -10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
-15 0
Time (s)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
(C)
Time (s)
Fig. 9: ESS Discharge when CBat1 = CBat2 : (a) SoC, (b) (c)
IBat , (c) Diff(SoC). Fig. 11: ESS Discharge when CBat1 < CBat2 : (a) SoC, (b)
IBat , (c) Diff(SoC).
Fig. 8(c) shows the difference between SoC Dif f (SoC) =
SoCBat1 SoCBat2 , where it is possible to see that the see clearly from Fig. 12 (from 10s to 11s), that the ESS is able
difference is practically zero after two iterations. Similarly, to smoothly transfer from charge to discharge modes during
Fig. 9 shows the response of the microgrid when the ESS are the equalization process in order to achieve the goal.
being discharged. Comparing Fig. 8 and Fig. 9, we can see
that for the discharging |IBat2 | > |IBat1 |, and for charging
|IBat1 | > |IBat2 | SoC equalization is achieved.
V. C ONCLUSION
B. Case CBat2 > CBat1
Fig. 10 and 11 show the response of the microgrid when The proposed been effective for SoC equalization in dis-
CBat1 = 0.01(A/h). We can see that when the equalization tributed ESS. Nevertheless, at least two iterations are required
is not applied, we have |IBat2 | > |IBat1 |. The reason of this in order to equalize completely the SoC, This is because the
is that ESS2 requires much more current in order to achieve transient and dynamic responses have not been considered
mSoC1 = mSoC2 . However when the algorithm is applied, the by the algorithm. Despite this, and by assuming a linear
current is adjusted in order to equalizing SoCs. behavior, the algorithm is able to equalize the SoC under
few iterations. This algorithm can be complemented by an
optimization process in order to minimize the period of time
C. Case CBat2 < CBat1
t, and by taking into account the power constrains of the
Fig. 12 and 13 show the response of the microgrid when system. Additionally, the algorithm can be easily adapted for
CBat2 = 0.01(A/h). Compared to the previous case |IBat2 | < AC microgrid and grid connected microgrids whit a larger
|IBat1 |, when the equalization is not applied. It is possible to number of interconnected active generators.
297
51.0 0
0 1 2 3 4 5 6 Equalization Equalization
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Algorithm Algorithm
active generator for smart grid applications, IEEE Transactions on
Time (Sec.)
Industrial Electronics, vol. 58, pp. 45834592, Oct 2011.
{
{
100
90 SoCBat2 [7] J. de Matos, F. e Silva, and L. Ribeiro, Power control in ac isolated
SoC (%)
#1:2
80 microgrids with renewable energy sources and energy storage systems,
#1:1
70
SoCBat1 IEEE Transactions on Industrial Electronics,, vol. PP, no. 99, pp. 11,
60 0
0 1 2 3 4 5 6 7 8 9 10 11 2 22 23 24 25
1 12 13 14 15 16 17 18 19 20 21 2014.
Time (s)
(a)
[8] F. Marra, G. Yang, C. Traeholt, J. Ostergaard, and E. Larsen, A
decentralized storage strategy for residential feeders with photovoltaics,
0.5
#1:1
IBat2
IEEE Transactions on Smart Grid, vol. 5, pp. 974981, March 2014.
IBat (A)
0.0 #1:2
[9] . I. S. C. Committee, Ieee guide for optimizing the performance and
-0.5
IBat1 life of lead-acid batteries in remote hybrid power systems, 2008.
-1.0 0
0 1 2 3 4 5 6 7 8 9 10 11 2 22 23 24 25
1 12 13 14 15 16 17 18 19 20 21 [10] D. Linden and T. Reddy, Handbook of batteries. McGraw-Hill hand-
Time (s) books, McGraw-Hill, 2002.
(b)
[11] Y.-K. Chen, Y.-C. Wu, C.-C. Song, and Y.-S. Chen, Design and
0
implementation of energy management system with fuzzy control for dc
Diff(SoC)
80 #1:1
SoCBat2 [14] Y. Zhang, H. J. Jia, and L. Guo, Energy management strategy of
60
SoC (%)
1.0 #1:2
[16] C. Li, T. Dragicevic, N. Diaz, J. Vasquez, and J. Guerrero, Voltage
0.5 #1:1
scheduling droop control for state-of-charge balance of distributed en-
IBat2
0.0 0 ergy storage in dc microgrids, in Energy Conference (ENERGYCON),
0 1 2 3 4 5 6 7 8 1 17 18 19 20 21
1 12 13 14 15 16
9 10 11 2 22 23 24 25
2014 IEEE International, pp. 13101314, May 2014.
Time (s)
(b) [17] C. Li, T. Dragicevic, M. G. Plaza, F. Andrade, J. C. Vasquez, and
0
J. M. Guerrero, Multiagent based distributed control for state-of-charge
balance of distributed energy storage in dc microgrids, in Industrial
Diff(SoC)
-5
#1:1
Electronics Society, IECON 2014 - 40th Annual Conference of the IEEE,
-10
pp. 21802184, Oct 2014.
-15 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 [18] X. Lu, K. Sun, J. Guerrero, J. Vasquez, and L. Huang, State-of-charge
Time (s) balance using adaptive droop control for distributed energy storage
(c)
systems in dc microgrid applications, IEEE Transactions on Industrial
Fig. 13: ESS Discharge when CBat1 > CBat2 : (a) SoC, (b) Electronics, vol. 61, pp. 28042815, June 2014.
IBat , (c) Diff(SoC). [19] Q. Shaee, T. Dragicevic, J. Vasquez, and J. Guerrero, Hierarchical
control for multiple dc-microgrids clusters, Energy Conversion, IEEE
R EFERENCES Transactions on, vol. 29, pp. 922933, Dec 2014.
[20] N. Diaz, T. Dragicevic, J. Vasquez, and J. Guerrero, Intelligent dis-
[1] W. Li, X. Mou, Y. Zhou, and C. Marnay, On voltage standards for dc tributed generation and storage units for dc microgrids - a new concept
home microgrids energized by distributed sources, in Power Electronics on cooperative control without communications beyond droop control,
and Motion Control Conference (IPEMC), 2012 7th International, IEEE Transactions on Smart Grid, vol. 5, pp. 24762485, Sept 2014.
vol. 3, pp. 22822286, 2012. [21] J. Akerlund, -48 v dc computer equipment topology-an emerging tech-
[2] D. McMenamin, Case studies supporting -48 vdc as the power input nology, in Telecommunications Energy Conference, 1998. INTELEC.
of choice for computer equipment deployed in the telecom network, Twentieth International, pp. 1521, 1998.
in Telecommunications Energy Conference, 1998. INTELEC. Twentieth [22] N. Diaz, A. Luna, and O. Duarte, Improved mppt short-circuit current
International, pp. 261265, 1998. method by a fuzzy short-circuit current estimator, in Energy Conversion
[3] C. Wang and P. Jain, A quantitative comparison and evaluation of 48v Congress and Exposition (ECCE), 2011 IEEE, pp. 211218, Sept 2011.
dc and 380v dc distribution systems for datacenters, in Telecommuni- [23] V. Salas, E. Olas, A. Barrado, and A. Lzaro, Review of the maximum
cations Energy Conference (INTELEC), 2014 IEEE 36th International, power point tracking algorithms for stand-alone photovoltaic systems,
pp. 17, Sept 2014. Solar Energy Materials and Solar Cells, vol. 90, no. 11, pp. 1555
[4] H. Kakigano, Y. Miura, and T. Ise, Distribution voltage control for dc 1578, 2006.
microgrids using fuzzy control and gain-scheduling technique, Power [24] Y. Zhang, H. J. Jia, and L. Guo, Energy management strategy of
Electronics, IEEE Transactions on, vol. 28, no. 5, pp. 22462258, 2013. islanded microgrid based on power ow control, in 2012 IEEE PES
[5] J. Guerrero, J. Vasquez, J. Matas, L. de Vicua, and M. Castilla, Innovative Smart Grid Technologies (ISGT), pp. 18, Jan 2012.
Hierarchical control of droop-controlled ac and dc microgrids a general [25] D. Wu, F. Tang, T. Dragicevic, J. Vasquez, and J. Guerrero, Au-
approach toward standardization, IEEE Transactions on Industrial tonomous active power control for islanded ac microgrids with pho-
Electronics, vol. 58, no. 1, pp. 158172, 2011. tovoltaic generation and energy storage system, Energy Conversion,
[6] H. Kanchev, D. Lu, F. Colas, V. Lazarov, and B. Francois, Energy IEEE Transactions on, vol. 29, pp. 882892, Dec 2014.
management and operational planning of a microgrid with a pv-based
298