Optimal Voltage Profile Control and Losses

Minimization of Radial Distribution Feeders

H. A. Attia,
Electrical Power and Machines Department, Faculty of Engineering, Cairo university, Giza, Egypt.
.

reducing losses and improving voltage, capacitors release

Abstract— The primary function of the power system is to capacity. Improving the power factor increases the amount of
satisfy load and energy requirements for customers with a real-power load the circuit can supply. Using capacitors to
reasonable assurance of continuity and quality of supply. The supply reactive power reduces the amount of current in the line,
electricity sector in all countries is interested in improving system
so a line of a given ampacity can carry more loads. This paper
quality not only to achieve legal obligations and standard
regulation but also to increase the revenue. Improving system
introduces a new method to find out the optimal location and
quality leads to increasing electricity sales, which represent a size of capacitor banks for reactive power compensation. The
profit to the utility, and contract damage. The appliances and compensation problem under study is an optimization problem
equipments will operate more efficiency during its lifetime which solved using nonlinear programming technique.
is a return to the customer. Two main items identifying the One of a utility’s core responsibilities is to deliver voltage to
quality of service were considered that directly affect the system customers within a suitable range, so utilities must regulate the
operation besides the utility revenue and also customer voltage. On distribution circuits, voltage drops due to current
satisfaction. Voltage profile control and losses minimization were flowing through the line impedances. Voltage regulator on
the two items considered. Owing to the load growth, the utility
feeder can adjust primary voltage. The voltage regulator has to
lies in a criticality due to the increasing voltage drop and losses in
addition to the capabilities of the distribution network. This
be assigned and allocated in an optimal issue such that
paper introduces a new method to find out the optimal location providing the best voltage profile on the feeder leading to
and quantity of capacitor banks for reactive power compensation. efficient performance of the customer loads besides reducing
There before a step voltage regulator for voltage profile control is system losses and improving system operation. The present
assigned and allocated to achieve the highest benefits for customer paper highlights these issues utilizing related mathematical
load patterns and utility loss reduction. The compensation formulations and computer programs as applied to study cases
problem under study is an optimization problem solved using in order to substantiate the validity of the proposed method. .
nonlinear programming technique. Appropriate Voltage profile
control and regulator allocation depends on the sight of the
II. VOLTAGE DROP PROBLEM
operation engineer and the load pattern and configuration of the
distribution feeder. Poor voltage regulation is a direct loss of revenue. Proper
Index Terms— Reactive Power Compensation, system voltage control will improve a company’s revenue.
Objective Function, Constraints, Load Flow, Losses, During the low voltage conditions, many types of equipment
Voltage Profile Control, Load characteristics. will operate at a lower power factor and draw more than rated
current without any increase in the power consumption, which
I. INTRODUCTION means more power losses for the utility and low efficiency of
customer equipment and a decrease in life time.
Capacitors provide tremendous benefits to distribution system
performance. Most noticeably, capacitors reduce losses, free up 1) Load models:
capacity, and reduce voltage drop. If applied properly and Electric load categories are affected differently by voltage
controlled, capacitors can significantly improve the variations. Loads may be classified as :
performance of distribution circuits. But if not properly • Constant power load
applied, the reactive power from capacitor banks can create The real and reactive power stays constant as the voltage
losses and high voltages. Good planning helps to ensure that changes. As voltage decreases, this load draws more current,
capacitors are sited properly. Feeder capacitors are closer to which increases the voltage drop. A constant power model is
the loads. Capacitors closer to loads more effectively good for induction motors.
release capacity, improve voltage profiles, and reduce line
losses. This is especially true on long feeders that have
considerable line losses and voltage drop. In addition to

978-1-4244-1933-3/08/$25.00 ©2008 IEEE 453

• Constant current load • Line .flow in each line section specified in kW and kVAr,
The current stays constant as the voltage changes, and the amps and degrees or amps and power factor.
power increases with voltage. As voltage decreases, the current • Power loss in each line section.
draw stays the same, so the voltage drop does not change. • Total feeder input kW and kVAr.
• Constant impedance load • Total feeder power losses.
The impedance is constant as the voltage changes, and the
power increases as the square of the voltage. As voltage Because the feeder is radial, iterative techniques commonly
decreases, the current draw drops off linearly; so the voltage used in transmission network power flow studies are not used
drop decreases. The constant impedance model is good for because of poor convergence characteristics. Instead, an
incandescent lights and other resistive loads. iterative technique “The ladder iterative technique” specially
Monitors at various load sides can determine the load active designed for a radial system is used. The distribution feeder is
and reactive power sensitivity to voltage variations to be used non-linear because most loads are assumed to be constant kW
directly or to identify parameters for more detailed load and kVAR.. Therefore, the non-linear characteristics of the
models. distribution feeder loads have to be taken into account.
EPRI had performed the power consumption measurement of
each category of electric appliances by varying the service The solution approach utilizes the “forward sweep” and the
voltage within the range 0.9 to 1.1 p.u. in the laboratory. A “backward sweep” methods to perform power flow solution.
sample of the results is shown in Appendix-1. The “forward sweep” will determine a computed source
voltage. This first “iteration” will produce a voltage that is not
2) The economical impact to the utility due to improving equal to the specified source voltage. The most direct
voltage profile of a distribution feeder: modification to the ladder network theory is to perform a
In order to highlight the economical revenue gained by the “backward sweep.” The backward sweep commences by using
utility when improving the voltage profile, consider the the specified source voltage and the line currents from the
following case: For a radial feeder feeding a distribution area “forward sweep. ” Kirchhoff’s Voltage Law is used to compute
having X Kw peak load. Consider only Z division of this load the voltage at all nodes of the line segments during backward
as a voltage sensitive load (lighting load). It is known that, the sweep. The forward and backward sweep process is continued
lighting load power is related to the voltage violations as Pα until the difference between the computed and specified voltage
V1.6. Then, the annual saving to the utility as a result of at the source is within a given tolerance.
regulating the voltage due to the considered division is given
by the following equation: 2) Feeder flow equations:
Annual Egyptian Pounds = X*Y*8760*(HRS/Year)* Take for example a balanced three-phase system with single
(EPS/Kwh)*(Z /total Kwh)*([Vn/Vo]1.6- 1) (1) feeder as shown in Fig. 1. Loads and line sections data are as
shown in figure. After assuming the voltage of the remote
Where: X : Annual feeder peak, Y: Annual load factor.
point 5, the current at this point is calculates as:
EPS: Cost of Kwh in L.E., Z: Voltage sensitive load Kwh.
*
Vn: Regulated voltage, . Vo : Unregulated voltage. S 
I 5 =  5  (2)
This equation makes the increased revenue of proper voltage  V5 
evident. Since the regulated voltage Vn, would normally be
held higher than unregulated voltage Vo, then the annual
Egyptian pounds will have to be increased. Actually, voltage
regulation is a good business.

II. POWER FLOW ANALYSIS OF RADIAL FEEDERS

Fig. 1 Radial feeder of a distribution system
1) Ladder iterative technique:
The power flow analysis of a distribution feeder is similar to For this “end node” case, the line current I45 is equal to the load
that of an interconnected transmission system. Typically what current I5. The voltage at node 4 (V4) can be determined using
will be known prior to the analysis will be the three-phase Kirchhoff’s Voltage Law:
voltages at the substation and the complex power of all of the V = V + Z .I
4 5 45 45
(3)
loads and the load model (constant complex power, constant
impedance, constant current or a combination). Shunt The load current I4 can be determined and then Kirchhoff’s
Capacitor Models were presented for the capacitor banks. Line Current Law applied to determine the line current I34.
segment models and step-voltage regulators were presented for I =I +I
34 45 4
(4)
the series components of a distribution feeder.
Kirchhoff’s Voltage Law is applied to determine the node
A power flow analysis of a feeder can determine the following voltage V3. This procedure is continued until a voltage (V1) has
• Voltage magnitudes and angles at all nodes of the feeder. been computed at the source. The computed voltage V1 is
compared to the specified voltage Vs. There will be a difference

454
between these two voltages. This is the first "forward sweep" Voltage level — The desired voltage (on 120-volt base) to be
iteration. The "backward sweep" iteration is then performed held at the “load center.” The load center is considered to be
using the specified source voltage and the line currents from the the remote node on the feeder.
“forward sweep.” Kirchhoff’s Voltage Law is used to compute Bandwidth — The allowed variance of the load center
the voltage at node 2 by: voltage from the set voltage level. .
V2 = V5 − Z12 .I12 (5) Time delay — Length of time that a raise or lower operation is
called for before the actual execution of the command.
This procedure is repeated for each line segment until a “new” Line drop compensator — Set to compensate for the voltage
voltage is determined at node 5. Using the “new” voltage at drop between the regulator and the load center. The settings
node 5, a second “forward sweep” is started that will lead to a consist of R and X settings in volts corresponding to the
“new” computed voltage at the source. The forward and equivalent impedance between the regulator and load center.
backward sweep process is continued until the difference ICT * x resistance of the line
between the computed and specified voltage at the source is R set = Potential transformer ratio
(6)
within a given tolerance.
ICT * x reactance of the line
X set = (7)
III. APPLYING VOLTAGE REGULATOR Potential transform er ratio
Where: ICT Current transformer primary current rating.
TO DISTRIBUTION FEEDER The rating of a regulator is based on the kVA transformed, not
the kVA rating of the line. In general this will be 10% of the
1) Step-voltage regulators theory and operation line rating since rated current flows through the series winding
A step voltage regulator consists of an autotransformer and a which represents the ±10% voltage change.
load tap changing mechanism. The voltage change is obtained
by changing the taps of the series winding of the 2) Voltage regulator allocation to control voltage profile
autotransformer. The position of the tap is determined by a A 11 kV radial feeder in the distribution company in Egypt is
control circuit (line drop compensator). Standard step considered. A simplified model of such feeder is shown in Fig.
regulators contain a reversing switch enabling a ±10% regulator 4 Global feeder data are given in appendix-1. The load flow
range, usually in 32 steps. This amounts to a 5/8% change per program for radial feeders utilizing forward and backward
step or 0.75 volt change per step on a 120-volt base. sweep is applied. Load data (P,Q), line sections data, input
A typical step voltage regulator is shown in Fig. 2. The tap voltage and peak current measured at the beginning of the
changing is controlled by a control circuit shown in the block feeder are fed to the program. The program output is the
diagram of Fig. 3. voltage at each feeder point and the flow in each line sections.
The losses in each section and allover the feeder are also given.

Fig. 4 A 11 kV radial feeder representation.

The voltage profiles in cases of off peak and peak loads are
Fig. 2 A 32-Step voltage regulator. shown in Fig. 5 without voltage regulator.

Fig. 3 Regulator control circuit.

The control circuit requires the following settings: Fig. 5 Voltage profile of the feeder at light and peak loads.

455
The input voltage was assumed 1 pu (11 kV), while the Table 1 Summary of regulator calculations and results
measured current was 170 amp. at peak load and 46 amp. at Voltage Regulator Site Just Besides Load
light load. The voltage profile shows that most loads suffers Branching no. 6
from voltage drop problem at on peak hours. The voltage drop Voltage Regulator Rating 62.33 kVA
exceeds the permissible tolerance allowed by standard Rset 16 volt
regulation (± 5%). The regulated point at the end of the feeder Xset 13 volt
is assumed to have a voltage drop 1% irrespective of load
No. of Steps Bucking or 15
changes. This necessitates using a voltage regulator boosting
Boosting
9% voltage at peak load located as shown in Fig. 6 to control
Voltage Variation Before 0.8629 – 0.9537 pu
the voltage profile of all the feeder load points to be inside the
Volt. Reg.
permissible range.
Voltage Variation After 0.9529 – 1.043 pu
Volt. Reg.

IV. BENEFITS QUANTIFICATION OF THE FEEDER

OPERATION AS A RESULT OF USING VOLTAGE
REGULATOR
Voltage profile control is not only performed to accommodate
the standards regulations, but also to achieve technical and
economical aspects. Besides to the increase in the revenue for
the utility as a result of the increased energy sales when
improving the feeder voltage profile, (Equation (1)), the
customer's apparatus will operate more efficiently. Other
Fig. 6. Voltage profile of the feeder with voltage regulator advantages are gained such as deferred capacity in the active
and reactive power in addition to decreasing feeder energy
3) Voltage regulator connection and calculations losses. This can be evaluated as follows.
When a three-phase, three-wire circuit (the case of the 11kV
distribution systems in Egypt) incorporates three single-phase 1) Before using voltage regulation:
voltage regulators in a closed-delta connected bank as shown in The application of the load flow program for the considered
Fig. 7, the overall range of regulation of each phase is 11kv feeder at peak load gives the power flow and currents in
dependent on the range of regulation of each regulator. This all line sections and the voltage of each node. Before installing
type of connection will give approximately 50% more the prospective voltage regulator, the feeder will be like that
regulation than is obtained with two regulators in an open-delta one shown in Fig. 8, viewed as two parts, from the source to the
configuration.. regulator location and from regulator location up to regulated
point at the end of the feeder.

n Fig. 8. The two parts of the feeder without regulator.

Fig. 7 Connection of three voltage regulators ia three-phase, .

three-wire Delta circuit The size of the regulator unit is given
2) After using voltage regulation:
as: After introducing a voltage regulator in the feeder, besides to
Rated load in amps= (three-phase kVA x 1000/ the voltage profile improvement, there is also a substantial
line-to-line volts x 1.732) saving in the feeding power and reduction in energy losses.
Regulator rated kVA = load amps x range in kV The current has to be decreased in both sections in Fig. 9 while
Range in kV = range x line-to-neutral kV using regulator as compared with Fig. 8 without regulator.
after using regulator. Table 2 illustrates the values of the active
Table 1 shows a summary of the regulator required and its power loss kW, reactive power loss kVAR and energy losses
setting to accommodate the assigned voltage regulation of the for the considered feeder without and with voltage regulator.
considered feeder. The percentage reduction in losses is about 15% which is
considered a good contribution for voltage regulator.

456
 Where Vmax : Maximum limitation of bus voltage
Vmin : Minimum limitation of bus voltage
m : Total number of buses on the feeder

2) Limitation of the capacitor

The change of bus voltage before and after being equipped with
bus relates to the capacity of the compensated capacitor of the
feeder; to assure power supply quality of the distribution
system. Fixed type of capacitors are selected. A limitation
Fig. 9 The two parts of the feeder with regulator. function should be added to accommodate the nature of the
capacitor to be non-continuous as in equation (8).
Table 2 Power and energy losses in each part of the feeder m

without and with voltage regulator. ∑Q −SQ

i =1
i i o =0 (8)

Item/Feeder Part Part 1 Part 2 All the Where Qo : Smallest capacity selected for the capacitor.
Feeder Si : Positive integer assigns the number of capacitors.
Active Without V.R. 92 186 278
3) The mathematical model of the problem and constraints
Power Loss With V.R. 79 157 236 Based on the above illustration, the mathematical model and
kW % age saving 15 % 15% 15 % constraints of the capacitor allocation problem is given as :
Reactive Without V.R. 78 119 197 Minimize :
Power Loss With V.R 67 100 167 m
kVAR % age saving 14 % 16% 15% F = K p Ploss + K c ∑ Qi (9)
i =1
Energy Without V.R 287243 578937 866180 Subject to :
Losses
kWH
With V.R
% age saving
245370
15%
488724
15%
734094
15%
Vmin ≤ Vi ≤ Vmax ; i = 1,2,3,...., m
m

V. OPTIMAL CAPACITOR ALLOCATION

∑Q −SQ
i =1
i i o =0
The compensating capacitor in the distribution system is used
to provide reactive power, in order to reduce power loss of the Qi ≥ 0
feeder and improve the stability of bus voltage. The installation
aims to define the most appropriate position and size of the 4) Solution Approach
compensating capacitor to minimize power loss and investment LINGO optimization program is utilized to solve the capacitor
cost at the same time. By doing so, the system will work to allocation optimization problem. LINGO program is a simple
minimize the total cost goal, and the objective function is tool for utilizing the power of linear and nonlinear optimization
demonstrated in the following equation (6): to formulate large problems briefly, solve them, and analyze
m the solution. The objective function and constraints including
F = K p Ploss + K c ∑ Qi (6) the problem variables are fed to the program. The goal of
i =1
optimization is to find the values of a model's variables that
Where Kp : Cost factor of real power loss (dollar/kW-year) generate the best value for the objective function, subject to
Kc : Annual cost of capacitor per kVAr (dollar/kVAr- any limiting conditions placed on the variables.
year)
Ploss : Total power loss of the system during peak load Applying such program for the feeder of Fig. 4 reveals the
(kW) following.
Qi : Capacitance placed of bus / (kVAr); we assume it i. The capacitor size is 1500 kVAR located at point
is an integral of the unit Qo no. 8 on the feeder.
ii. The losses has reduced by 5%
i=l,2,3,...,m refers to the position of compensated bus iii. The voltage of the regulated point has improved
that is selected. from 0.91 pu to 0. 935 pu
Therefore, besides loss minimization an additional advantage
1) Limitation of the bus voltage was gained as viewed by the regulator. The regulator is then
This study minimizes the total cost F and limits the voltage of relieved. A load bonus is gained, the assigned regulator may be
each bus to the acceptable range shown as equation (7) : then replaced by a smaller size or keeping the larger size for
Vmin ≤ Vi ≤ Vmax ; i = 1,2,3,...., m (7)
future load growth.

457
 V. CONCLUSIONS [7] Carlisle, J. C. and El-Keib, A. A., “A Graph Search
Algorithm for Optimal Placement of Fixed and Switched
Due to expansions and load growth in the distribution systems, Capacitors on Radial Distribution Systems,” IEEE Transactions
the voltage drop problem has arised. The utility has to provide on Power Delivery, vol. 15, no. 1, January 2000.
end users with energy and power as economically as possible [8] Grainger, J. J. and Lee, S. H., “Optimum Size and Location
while maintaining good quality service. The utility has to of Shunt Capacitors for Reduction of Losses on Distribution
provide service to customers with a voltage level meeting the Feeders,” IEEE Transactions on Power Apparatus and Systems,
standard regulations. Several advantages are gained from vol. PAS-100, no. 3,March 1981
improving voltage profile of the distribution feeders. For the
utility, an increase in the energy sales is gained, in addition to
improving the system operation. The currents in the feeder Appendix-1 The power consumption measurement
section, energy losses, active and reactive power losses are
compressed. Deferred capacity for transformers and lines are Type of Load Model
also gained. For customers, their apparatus will operate more Appliance
efficiently besides to the loss reduction in final distribution Refrigerator P = 1.0 + 1.3958 ∆V + 9.881 ∆V2
circuits. + 84.72 ∆V3 + 293 ∆V4
For solving the voltage drop problem, the voltage regulator was Washing machine P = 1.0 + 1.2786 ∆V + 3.099 ∆V2
found to be the master. The size and location of the regulator + 5.939 ∆V3
to control the feeder voltage profile was given. The settings Clothes dryer P = 1.0 – 0.1968 ∆V – 3.637 ∆V2
and the operation of the regulator are, also, illustrated. The - 28.32 ∆V3
impact of the regulator in improving system operation is TV P = 1.0 + 1.2471 ∆V + 0.562 ∆V2
quantified. Electric pot P = 1.0 + 0.3769 ∆V + 2.003 ∆V2
Additional improvement of system operation can be gained by Electronic pot P = 1.0 + 0.3758 ∆V + 2.326 ∆V2
allocating capacitor optimally on the feeder. A substantial Dehumidifier P = 1.0 + 0.5179 ∆V + 3.721 ∆V2
decrease in losses, improvement in voltage profile and relieve Fan P = 1.0 + 0.7101 ∆V + 2.133 ∆V2
to the regulator are gained. The system can meet future Lamp P = 1.0 + 0.6534 ∆V – 1.650 ∆V2
extensions and load growth while maintaining satisfactory ∆V : Normalized voltage deviation from rated voltage
service quality.

REFERENCES
DISCUSSION FOR COMMENTS
[1] Electric Power Distribution Engineering, Turan Gonen,
Voltage drop problem in 11kV distribution feeders in Egypt is
McGraw Series In Electrical Engineering, 1986.
an annoying phenomena and affects both utility and customers.
[2] C.G. Renato, “New method for the analysis of distribution
The appraisal of such effect necessitates clarifying how do
networks, IEEE Trans. Power Delivery, Vol.5, No.1, Jan 1990.
electric appliances power are influenced by voltage deviations
[3] Bishop, M. T., Foster, J. D., and Down, D. A., “The from rated values. Therefore, it was important to introduce
application of Single-Phase Voltage Regulators on Three-Phase
load models and mathematical forms identifying the power
Distribution Systems,” IEEE Industry Applications Magazine,
consumption of different appliances as a function of voltage
pp. 38–44, July/August 199
deviation. This is also beneficial for utility for financial
[4] Power Distribution Planning Reference Book, H. Lee Willis
justification and surplus revenue quantification when deciding
ABB Power T&D Company Inc. Cary, North Corolina, Marcel
to improve feeders voltage profile, in addition to meet standard
Dekker. Inc. 1997.
regulations.
[5] Foundations of Optimization, Charles S. Beightler, 1979
The radial feeder load flow was solved using a software tool,
by Prentice Hall, Inc.,Englewood Cliffs,N.J
however, an explanation of the ladder iterative solution
[6] M.M.A. Salama, A.Y. Chikhani, “A Simplified Network
method – that the software is built on – was mentioned to
Approach to the VAR Control Problem for Radial Distribution
explain the peculiarity of the radial feeder and for the seek of
Systems”, IEEE Trans. Power Delivery, Vol. 8, No.3, Jul 1993.
complementing radial feeder load flow solution procedure.

458