Response Coordination of Distributed Generation and Tap Changers For Voltage Support
Response Coordination of Distributed Generation and Tap Changers For Voltage Support
Response Coordination of Distributed Generation and Tap Changers For Voltage Support
the DG in their simultaneous operation during peak load hours. Verr1 = V ref − Vt1 (1)
The objective of the coordination controller is to reduce, or
It should be noted that Vref is normally set at 1.0 p.u. The
eliminate the number of counter-acting operations between the
target bus of the OLTC is referred to as the furthest bus from
two devices.
the substation which belongs to the control zone of the OLTC.
In other words, the OLTC is meant to regulate the weakest
II. VOLTAGE REGULATION METHOD voltage point within their zone of responsibility. Definition of
Conventionally, the voltage on any distribution feeder is “control zone” will be discussed later on in this paper. Step 1
controlled by the OLTC placed at the substation. The main produces an output called Si, which takes a value of 0, or -1, or
purpose of this voltage regulation is to keep the voltages of all +1 depending on how difference the Verr1 compared to the
the customers within the acceptable limits, which is ± 5% for deadband a is, as shown in (2).
Australia distribution system. However, with the rapidly 0 for - a ≤ Verr1 ≤ a
increase of load, the OLTC can not always guarantee that it
will be able to keep the regulation requirement satisfied. The S i = + 1 for Verr1 > a (2)
- 1 for V < −a
recent trend of introducing DG into the distribution networks err1
has become a potential solution for solving voltage problem, Value of the deadband defines the total voltage range,
especially during peak hour. In this section, we introduce a around the voltage setting Vref, which the OLTC controller will
coordinated voltage control technique using parallel operation consider as a satisfied condition. Choice of a small deadband
of the OLTC and the DG with minimum level of will result in more tap changes to occur, but will provide a
communication system. better regulation. On the contrary, a larger deadband causes
A. Simple System Model fewer tap changes, but at the expense of a closely regulated
line.
To analyze the operation of the OLTC and the DG, as well Step 2 – Comparison of current and previous voltage status: in
their effects on the distribution feeder, a simplified model has this step, the current status Si will be compared to the previous
been used as shown in Fig.1. status Si-1. This is done to keep track of the changes of voltage
DG status at the target bus.
Step 3 – Counter setting: A counter is employed to determine
the time duration of the voltage error exceeding the deadband.
~ The counter basically starts accumulating the time when
OLTC Load voltage error is outside of the deadband. It will be reset in
d three cases: the voltage error fluctuates below and above the
L deadband, the voltage error changes sign, or the voltage error
is currently within the deadband. Output of the counter is
Fig. 1 Simplified model of distribution feeder determined based on the results of the comparison in Step 2:
0 for S i ≠ S i −1
The distribution feeder is connected to the substation,
which is modelled as a Thevenin source, at the point of Ci = 0 for S i = 0 (3)
common coupling. For simplicity, distribution line of length L C i −1 + ∆t for S i = S i −1
is represented as series resistance and reactance with load Where ∆t is the time step.
represented by equivalent lumped values at N number of Step 4 – Time delay setting: In this step, value of the counter
nodes. The network is equipped with an OLTC at the will be compared to the time delay setting DT, and the control
substation. The OLTC has provision for correcting the voltage action will take place only if the time delay is elapsed. The
from -5% to +15%, each step represents a 1.25% change in time delay varies depend on the control coordination
voltage. A DG connected to the distribution line through a algorithm, which has a great impact on the performance of the
load bus d km from the substation. It is modelled as a current system. Selection of the DT value will be discussed in more
source and able to inject both real and reactive power to the details in the section of “Coordinated voltage control”.
system. Step 5 – Action of tap changer: if there is a requirement of a
tap change and the counter value is equal or greater than the
B. Voltage Regulation by OLTC time delay setting (Ci ≥ DT), a signal will be sent to the tap-
The operation of the OLTC includes five major steps, which changer motor to move the tap up or down. The actual tap
are described in details below: ratio is:
Step 1 - Status determination: The OLTC recognizes the ni = ni −1 + u (∆n )
current voltage condition, which is represented by the value of (4)
voltage error. The voltage error of the OLTC controller, Verr1 if n min ≤ n i ≤ n max and V min ≤ Vlocal ≤ Vmax
depends on two inputs, the actual voltage at the target bus of Where ni and ni-1 are the current and previous actual tap ratio,
the OLTC, Vt1, and the reference voltage Vref. It can be respectively, ∆n is the tap step, and u is the desired tap ratio
calculated using (1): which is +1 for requirement of a tap up, -1 for requirement of a
tap down, and 0 for no requirement of tap action. nmin and nmax
3
represent the working band of the OLTC, Vmin and Vmax are the In this work, we propose a method for determination of
lower and upper voltage limits, respectively. voltage control zones according to the sensitivity analysis.
From the network equation YBUSVBUS = IBUS, for a N+1-load
C. Voltage Regulation by DG
bus system with bus 1, bus n, and bus n+1 represent the
The DG controller operates in a much similar manner with a substation, the remote load bus, and the internal bus which
sequence of 5 steps. The steps and their outputs are as connects the DG into the system, respectively, we obtain:
followings:
Step 1 - Status determination: The DG status SDGi can be (
V3− N = −Y3−1 Y2TVS − 2 + Y5VDG ) (9)
determined by (5);
V3
0 for - b ≤ Verr 2 ≤ b y1,3 K y1, n
= M , Y2 =
y 2,3 K y 2, n
Where, V3 − N
S DGi = + 1 for Verr 2 > b (5)
V n
- 1 for V
err 2 < −b
V
and Verr 2 = V ref − Vt 2 (6) V S − 2 = S , and VDG is voltage at the DG connection point.
Where, Verr2 is voltage error of the DG controller, Vt2 is actual V 2
voltage at the target bus of the DG, b is the voltage deadband The relative changes of V3-N due to the DG injection and the
of Vt2. tap operation can be obtained with substituting VS = 0 into (9)
Step 2 – Comparison of current and previous voltage status: and examining individually the response of the system to 1.0
SDGi is compared to SDGi-1. p.u. voltage at the DG connection point and 1.0 p.u. voltage at
Step 3 – Counter setting: Counter value is calculated as: V2, which is the secondary side of the OLTC, using the
0 for S DGi ≠ S DGi −1 superposition principle. The results are:
s DG = −Y3−1Y5
C DGi = 0 for S DGi = 0 (7) (10)
C DGi −1 + ∆t for S DGi = S DGi −1 0
s TX = −Y3−1Y2T (11)
Step 4 – Time delay setting: Time delay of the DG is denoted 1
as TDDG. Setting for time delay of the DG controller will be
If values of (10) & (11) are graphically plotted with respect
described later in section 2.4.
to the load bus from 3 to N, their intersection point could be
Step 5 – Action of the DG: The desired output of the DG, IDG,
is calculated using the proportional controller as in (8): defined as the boundary for control zones of the OLTC and the
DG. Let the ordinate of the intersection is Cb. The
I DG = K P V err 2 (8) responsibility of each device is defined by following rules:
Where KP is the proportional constant. - If the voltage sensitivity of bus i with respect to DG voltage
In this work, however, the DG is set to step-by-step update
is greater than Cb ( s DG ≥ C b ), bus i is belong to the control
i
its output towards to desired value, instead of making a
substantial change. This is done to allow time for the other zone of DG.
regulating devices (OLTC in this case) to take part into the - If the voltage sensitivity of bus i with respect to the V2 is
regulating process. Also, it helps to avoid unnecessary extra i
greater than Cb ( s tx ≥ C b ), bus i is belong to the control zone
energy delivered by the DG in case when the later change of of the OLTC.
the load demand would possibly eliminate the voltage The main purpose of voltage control zone determination is
violation. By controlling DG output in this way, we would be that we will be able to assign higher opportunity to a more
able to reduce the operation cost of the DG. However, it efficient device to operate. Thus, the voltage drop will be
causes increasing time of the DG response to voltage violation compensated more quickly with less number of control actions.
if the current load condition maintains sufficiently long. Let us assume that the boundary bus is bus m. In other
D. Coordinated Voltage Control words, the control zone of the OLTC is from bus 2 to bus m,
In order to avoid hunting between the units, proper while DG control zone is from bus (m+1) to bus n, which is
coordination is needed. The coordination makes sure that all the remote bus. Delay time values of the OLTC and DG, which
regulating devices in the system will response to any voltage are denoted as DT and DTDG, respectively, are not constant,
violation in a correct order. As we have mentioned earlier, the but changing from time to time depending on the network
task of coordinating responses of the OLTC and DG in this conditions. The delay time could be one of the following
paper is done using a definition of “control zone”. types:
Idea of the control zone is that the area subject to voltage - Transient time (TT for OLTC and TTDG for DG): the time
control is divided into two zones. Each device is mainly that has to elapse before any voltage correction is permitted to
responsible for controlling voltage in one zone, where its take place. This setting is to avoid voltage correction to occur
control action has higher influence to the voltage. The other due to temporary voltage excursions.
device will act as a supporter in case the action of the main is - Grading time (GT for OLTC and GTDG for DG): the time in
not satisfactory or fails. which a regulating equipment has to delay its action to enable
4
the equipment with higher priority to complete their voltage if VD post − VD pre ≤ w
correction. In this paper, since the operation cost of the OLTC
is less than that of the DG, we assume that the default Grading if TD ≤ TT then TD = TT
Time of the DG is two times higher than the OLTC’s Grading else then TD = ST
Time. end
- Second-action time (ST for OLTC and STDG for DG): the else
time delay before any further action can take place, if required. if the DG fails to operate
The next action has to be in the same direction with the if TD ≤ TT then TD = TT
previous one for the Second-action time to apply. Otherwise, else then TD = ST
the time delay with be either Transient Time or Grading Time.
end
It should be noted that Second-action Time is always shorter
else
than Transient Time.
The algorithm for the control coordination has been if TD ≤ TT DG + TT then TD = TTDG + TT
proposed with a set up of communication system as follows: else then TD = ST
- One way communication from the target load buses to its end
corresponding controller, which is depending on the control end
zones. end
- Two way communication betweens controllers of regulating end (12)
devices in the system.
The first part of the communication system is built to provide Time delay for the DG:
highly accurate feedback signal to the controller, which is the
if VD post − VD pre ≤ w
voltage status at target buses. The later part of the
communication system, on the other hand, is used to inform if the OLTC fails to operate
status of one device to another. This would be very helpful in if TD DG ≤ TT DG then TD DG = TT DG
the case that when one device fails to operate, their
responsibility could then be passed on to other equipment else then TD DG = ST DG
within the shortest time, and thus reduce the duration of end
voltage rise/drop. end
The priority of controller in this coordination scheme has else
been determined based on the idea of control duty. This means if TD DG ≤ TTDG then TD DG = TTDG
that if more significant load change occurs in the control zone
of one regulating device, it then has more priority to operate else if TTDG < TDDG ≤ GTDG then TDDG = GTDG
than the other. However, taking into account that operation of else then TD DG = STDG
the DG is more expensive than the OLTC, the DG, in case of
end
higher priority, will only issue one step change in its output
end (13)
then wait until the Grading Time has passed before doing any
Where, VDpre and VDpost are the simulated voltage drops
further voltage correction. This will allow time for the OLTC
before and after voltage excursion. ω is a threshold value to
to operate if necessary. Moreover, by using the communication
determine the priority of the regulating equipment.
system, a signal could be sent from one controller to another to
In (12) & (13), the three variables VDpre, VDpost, and ω are
inform their status. If one device is exceeding its capacity or
used to determine the knowledge of where the causes of
fails to carry out the desired action, the other one will operate
voltage violation are. This knowledge is used as a criterion to
with minimum delay time, regardless its priority.
assign the time delay values for the OLTC as well as the DG.
The time delay of the OLTC and the DG for this control
When the remote voltage is outside the deadband, this is due to
coordination is calculated as in (12) & (13) using “IF”
one or both of the following reasons:
statements. It should be noted that the equations applied only if
- A voltage change upstream the DG
Si ≠ 0 and SDGi ≠ 0, respectively. Vr and Vm are used to denote
- A load change downstream the DG
the actual remote voltage and the actual voltage at boundary
The distinction between those two can be found by monitoring
bus m, respectively.
the local voltage and current at DG point, as well as the
Time delay for the OLTC:
upstream feeder impedance. This information could provide us
if Si ≠ 0
a simulated voltage drop, from which we are able to define if
if TD ≤ TT then TD = TT there is any load change has occurred in the voltage control
else then TD = ST zone of the DG [8]. If the difference between simulated
end voltage drop before and after voltage excursion is large, then
end the second cause is more likely to occur.
if SDGi ≠ 0
5
Sensitivity values
69 nodes, in which node 1 is represented for the substation and
0.6
node 69 is denoted for the remote load bus. There are nine
laterals attached to the main, which is shown in Fig.2. 0.5
0.3
H F 0.2
E
0.1
D C 0
G B 12 22 32 42 52 62 72 82 92
Load bus number
A Fig. 3: Determination of the control zones
SMITHTON
The load variation, as shown in Fig.4, has been used to test
Fig. 2: Smithton - Woolnorth test feeder the proposed method of voltage control coordination. Total
In order to properly illustrate how the coordination methods simulation time is 400 seconds. During this period of time, the
work, modified form of the system has been used. The feeder total load demand increases from 1.02 MVA up to 4.02 MVA.
has an automatic OLTC at the substation and a DG 36 km At each second, there are a number of random load buses
from the substation. changing their load demand with a random magnitude. This
has been done to imitate the practical situation of load
IV. SIMULATION RESULTS fluctuations, which are random and non-uniform. Since under-
voltage situations are more likely to occur in the distribution
Firstly, we demonstrate how to determine the control zones
systems, an increasing load profile has been used in the
for each regulating unit by the approach of zone determination
simulation. By doing this test, we would be able to know how
with the system offline load data. The sensitivity values of load
much the maximum loading capacity is before the system
bus voltages with respect to voltage at the secondary side of
voltage fails to meet the safety margins.
the OLTC (sTX) and with respect to voltage of the DG
connection point (sDG) have been plotted on the same graph, as Real and reactive power of load demand
4.5
shown in Fig.3. The intersection of sTX and sDG divides the
feeder into two parts. The upstream part from bus 2 to bus 52 4
Real power
is belong to the OLTC control zone, while the downstream Reactive power
3.5
part from bus 53 to the remote end belongs to the DG control
P (MW) and Q(MVars)
3
zone. We should note that the sensitivity analysis here is
carried out for those load buses located on the main feeder 2.5
only. For the laterals, those attach to the backbone at points
2
which belongs to the OLTC control zone, will also be the
major regulating duty of the OLTC. Otherwise, the DG will be 1.5
the main responsible regulating unit.
1
0.5
0
0 50 100 150 200 250 300 350 400
Time (sec)
Fig. 4: Dynamic load profile
represent the variation limits of bus 52 and bus 69, not sufficient enough to activate the DG controller. When t =
respectively. Since bus 52 is at around two-third of the feeder 352 seconds, the DG starts responding to the excursion at the
from the substation, we would like to keep its voltage at a remote bus. After that, from t = 360 seconds, the DG output
reasonable level (±0.025 variation is chosen here) in order that increases rapidly to compensate the voltage drop as it detects
further voltage drops along the line will not cause the remote the saturation condition of the OLTC.
end to fall below 0.95 p.u. The variable b is chosen to be 0.045 Voltage profile of system shows that the voltage is well
instead of 0.05 to give some levels of tolerance to the control regulated most of the simulation time. Actions of the OLTC
system. has maintained the voltage within the safety margins, and thus
Figs.5, 6 and 7 show the results of the coordination method kept the DG offline for more than 85% of time. The DG only
using communication system, including the tap response, the participates into the control process when the voltage drops
DG response, and the voltage at the remote end, respectively. below the limit, and the OLTC fails to operate further. This
In Fig.5, we can see that the OLTC alters its ratio according to can be considered as a desired solution for solving voltage
the voltage variation at bus 52. At first, it increases with the problems in a distribution system, which maximises the
increase of load. After that, at t = 220 seconds, tap ratio stops capacity of the OLTC and utilises the DG only in real need.
rising at the value of 1.063. Even though the capacity of the Voltage at the remote end
0.99
OLTC has not exceeded yet, further increase of tap ratio is not
0.985
allowed because of upper voltage constraint at the secondary
side of the OLTC. 0.98
0.97
Tap ratio / Voltage at target bus (p.u.)
1.06
0.965
1.04
0.96
1.02 0.955
Tap ratio
Actual voltage at bus 52 0.95
1
0.945
0 50 100 150 200 250 300 350 400
0.98 Time (sec)
Fig. 7: Remote end voltage profile
0.96
V. CONCLUSION
0.94
0 50 100 150 200 250 300 350 400 In this paper, an algorithm for the control action
Time (sec)
coordination between the OLTC and the DG has been
Fig. 5: Tap response and voltage at the target bus – bus 52
developed with the utilization of communication system. The
Real and reactive power of DG
0.18 method works by incorporating several concepts, including the
control zones, settings of reference voltage, voltage deadband,
0.16
as well as time delay. Selection of these settings has been
0.14 discussed in details in the paper, which makes it easier for the
control engineers to design their systems. The coordination
P (MW) and Q(MVars)
VIII. BIOGRAPHIES
An D.T. Le received the B.E. (Hons.) from the University of Tasmania,
Australia, in 2004. She is currently pursuing the PhD degree at the University
of Tasmania. Her special fields of interests are power system analysis,
renewable energy, distributed generation, power system control and
protection.