HIGH-EFFICIENCY VOLTAGE REGULATOR FOR

RURAL NETWORKS
ABSTRACT
This paper presents a high-efficiency voltage regulator, which combines robustness, low
costs and easy maintenance without power electronics components. Power quality is the
combination of voltage quality and current quality. Quality of supply is a combination of voltage
quality and the non-technical aspects of the interaction from the power network to its customers.
These characteristics make it suitable for rural networks, where investments and operational cost
in power quality improvement are limited. The regulator consists of a multi winding reduced-
power transformer, and provides serial voltage compensation.
This paper presents a new voltage regulator that fulfills the rural networks needs: high efficiency,
robustness, easy maintenance and low cost. ection !! presents the design of the voltage
regulator, describing its power circuit and control system. ome practical considerations
regarding the design of the voltage regulator are presented in ection !!!. "nd finally, ection !#
presents the operation e$perience data of voltage regulators installed in the distribution network.
%ifferent voltage compensation steps are obtained by modifying the connection and the
polarity between the primary and secondary windings. The transformer design has been
optimi&ed to obtain a high-efficiency and low-cost regulator. "n automatic controller monitors
the output voltage and sets the optimal compensation step. "t present more than '(( units of the
voltage regulator are in operation.
)$perimental records for the operation of installed voltage regulators have shown their
reliability, high efficiency, and their capacity to improve power quality in rural networks.
I. INTRODUCTION
L*+,-duration voltage variation -undervoltage and overvoltage. is a central issue in
distribution network power quality. upply voltage and power quality are regulated by certain
standards, such as the )uropean )+ /(01( 203 or the "merican "+! 45'-0 263. These standards
are complemented in each country or state by specific codes and rules 273. The )uropean )+
/(01( stipulates that the ma$imum voltage amplitude variation accepted is 0(8, while the
"merican "+! 45'-0 defines a normal operating range of 06( # /8. +ational rules usually
define more restrictive voltage ranges9 for instance, the panish rule for voltage quality 2'3 sets
the ma$imum variation of the voltage at the load connection point at 67( # :8. The value of
voltage amplitude is an important quality issue, because loads are designed to work correctly
within a specific voltage range. everal problems in domestic and industrial equipment are
associated with long duration undervoltages, such as malfunctioning in relays and contactors,
incandescent lighting dim, switch-off of discharge lighting, failure of nonlinear loads -e.g.,
computer power supplies., and torque reduction in induction machines. *n the other hand, long
duration overvoltages usually result in the overheating of loads -motors and transformers., and
hence a reduction in their e$pected durability. ;ow voltage rural distribution networks compared
with urban networks are more susceptible to long-term voltage variations, due to the dispersed
configuration of customers. #oltage variations in rural areas are usually associated with long
distances between the loads and the distribution transformer. +owadays, the integration of non-
controllable dispersed generation in these networks is a new potential source of voltage variation
problems. To minimi&e long-term voltage variations in rural networks, distribution companies
have traditionally performed different actions: 0. tap change control in the main distribution
transformer9 6. installation of compensation equipment, such as capacitor banks, voltage
regulators, boosters, or auto-boosters9 and 7. as a last resort, because it is the most e$pensive
alternative, the distribution company upgrades the low voltage network -increasing the line
capability, or changing the network rated voltage. 2/3. !n rural areas, the ratio of contracted
power per connection point is much smaller than for urban areas9 therefore, investments to solve
specific voltage problems are limited. !n this situation, the use of compensation equipment such
as voltage regulators becomes an interesting alternative. <oreover, the distribution company also
takes into account the operation and maintenance costs and the energy losses resulting from the
different options for solving voltage problems. 4onsequently the cost-efficiency of voltage
regulators is also a key issue. 4urrently, there are different technologies for voltage regulators
both in commercial devices and in the literature: tap-switching, ferroresonant, and electronic 213.
The most advanced commercial voltage regulators are based on power electronics and provide
accurate voltage output. +owadays there are few approaches to voltage regulators in rural
networks 2:3=20(39 moreover these approaches still do not completely cover the needs of rural
distribution networks. This paper presents a new voltage regulator that fulfills the rural networks
needs: high efficiency, robustness, easy maintenance and low cost. ection !! presents the design
of the voltage regulator, describing its power circuit and control system. ome practical
considerations regarding the design of the voltage regulator are presented in ection !!!. "nd
finally, ection !# presents the operation e$perience data of voltage regulators installed in the
distribution network.
Power disri!"io# $o#ro%
Distribution System
Electrical power is transmitted by high voltage transmission lines from
sending end substation to receiving end substation. At the receiving end
substation, the voltage is stepped down to a lower value (say 66kV or 33kV
or kV!. "he secondary transmission system transfers power from this
receiving end substation to secondary sub#station. A secondary substation
consists of two or more power transformers together with voltage regulating
e$uipments, buses and switchgear. At the secondary substation voltage is
stepped down to kV. "he portion of the power network between a
secondary substation and consumers is known as distribution system. "he
distribution
system can be classified into primary and secondary system. %ome large
consumers are given high voltage supply from the receiving end substations
or secondary substation.
"he area served by a secondary substation can be subdivided into a number
of sub# areas. Each sub area has its primary and secondary distribution
system. "he primary distribution system consists of main feeders and
laterals. "he main feeder runs from the low voltage bus of the secondary
substation and acts as the main source of supply to sub# feeders, laterals or
direct connected distribution transformers. "he lateral is supplied by the
main feeder and e&tends through the load area with connection to
distribution transformers. "he distribution transformers are located at
convenient places in the load area. "hey may be located in specially
constructed enclosures or may be pole mounted. "he distribution
transformers for a large multi storied building may be located within the
building itself. At the distribution transformer, the voltage is stepped down to
'((V and power is fed into the secondary distribution systems. "he
secondary ' distribution system consists of distributors which are laid along
the road sides. "he service connections to consumers are tapped off from
the distributors. "he main feeders, laterals and distributors may consist of
overhead lines or cables or both. "he distributors are 3# phase, ' wire
circuits, the neutral wire being necessary to supply the single phase loads.
)ost of the residential
and commercial consumers are given single phase supply. %ome large
residential and commercial consumer uses 3#phase power supply. "he
service connections of consumer are known as service mains.
"he consumer receives power from the distribution system. "he main part of
distribution system
includes*#
. +eceiving substation.
,. %ub# transmission lines.
3. -istribution substation located nearer to the load centre.
'. %econdary circuits on the .V side of the distribution transformer.
/. %ervice mains.
Power Flow
0or distribution system the power flow analysis is a very important and
fundamental tool. 1ts results play the ma2or role during the operational
stages of any system for its control and economic schedule, as well as during
e&pansion and design stages. "he purpose of any load flow analysis is to
compute precise steady#state voltages and voltage angles of all buses in the
network, the real and reactive power flows into every line and transformer,
under the assumption
of known generation and load.
-uring the second half of the twentieth century, and after the large
technological developments in the fields of digital computers and high#level
programming languages, many methods for solving the load flow problem
have been developed, such as 3auss#%iedel (bus impedance matri&!,
4ewton#+aphson5s (4+! and its decoupled versions. 4owadays, many
improvements have been added to all these methods involving assumptions
and appro&imations of the transmission lines and bus data, based on real
systems conditions.
"he 0ast -ecoupled 6ower 0low )ethod (0-60)! is one of these improved
methods, which was based on a simplification of the 4ewton#+aphson5s
method and reported by %tott and Alsac in 78'. "his method due to its
calculations simplifications, fast convergence and reliable results became the
most widely used method in load flow analysis. 9owever, 0-60) for some
cases, where high +:; ratios or heavy loading (.ow Voltage! at some buses
are present, does not converge well. 0or these cases, many efforts and
developments have been made to overcome these convergence obstacles.
%ome of them targeted the convergence of systems with high +:; ratios,
others those with low voltage buses. "hough many efforts and elaborations
have been achieved in order to improve the 0-60), this method can still
attract many researchers, especially when computers and simulations are
becoming more developed and are now able to handle and analy<e large si<e
system.
Objectives of Radial Distribution System:-
. 6lanning, moderni<ation and automation.
,. "o provide service connection to various urban, rural and industrial
consumer in the allocated area.
3. )a&imum security of supply and minimum duration of interruption.
'. %afety of consumers, utility personnel.
/. "o provide electricity of accepted $uality in terms of *#
(a! =alanced three phase supply.
(b! 3ood power factor.
(c! Voltage flicker within permissible limits.
(d! .ess voltage dips.
(e! )inimum interruption in power supply.
Advantages of Radial Distribution System:-
(a! +adial distribution system is easiest and cheapest to build.
(b! "he maintenance is easy.
(c! 1t is widely used in sparsely populated areas.
Drawback of Radial Distribution System:-
(a! "he end of the distributor nearest to the feeding point will be heavily
loaded.
(b! "he consumers are dependent on a single feeder and single distributor.
"herefore, any fault on the feeder or distributor cuts off supply to the
consumers who are on the side of the fault away from the sub#station.
(c! "he consumers at the distant end of the distributor would be sub2ected to
serious voltage fluctuations when the load on the distributor
!e single line diagram of a ty"ical low tension distribution system#
$istory of Distribution System
1n the early days of electricity distribution, direct current -> generators were
connected to loads
at the same voltage. "he generation, transmission and loads had to be of the
same voltage because there was no way of changing -> voltage levels,
other than inefficient motor#generator sets. .ow -> voltages were used (on
the order of (( volts! since that was a practical voltage for
incandescent lamps, which were then the primary electrical load. "he low
voltage also re$uired less insulation to be safely distributed within buildings.
"he losses in a cable are proportional to the s$uare of the current, the length
of the cable, and the
resistivity of the material, and are inversely proportional to cross#sectional
area. Early transmission networks were already using copper, which is one of
the best economically feasible
conductors for this application. "o reduce the current and copper re$uired
for a given $uantity of
power transmitted would re$uire a higher transmission voltage, but no
convenient efficient method e&isted to change the voltage level of -> power
circuits. "o keep losses to an economically practical level the Edison ->
system needed thick cables and local generators.
%odern Distribution System
"he modern distribution system begins as the primary circuit leaves the sub#
station and ends as the secondary service enters the customer?s meter
socket. A variety of methods, materials, and e$uipment are used among the
various utility companies, but the end result is similar. 0irst, the energy
leaves the sub#station in a primary circuit, usually with all three phases. "he
most common type of primary is known as a &ye configuration (so named
because of the shape of a @A@.! "he Bye configuration includes 3 phases
(represented by the three outer parts of the @A@! and a neutral (represented
by the centre of the @A@.! "he neutral is grounded both at the substation and
at every power pole. "he other type of primary configuration is known as
delta. "his method is older and less common. -elta is so named because of
the shape of the 3reek letter delta, a triangle. -elta has only 3 phases and
no neutral. 1n delta there is only a single voltage, between two phases
(phase to phase!, while in Bye there are two voltages, between two phases
and between a phase and ,8 neutral (phase to neutral!. Bye primary is safer
because if one phase becomes grounded, that is, makes connection to the
ground through a person, tree, or other ob2ect, it should trip out the circuit
breaker tripping similar to a household fused cut#out system. 1n delta, if a
phase makes connection to ground it will continue to function normally. 1t
takes two or three phases to make connection to ground before the fused
cut#outs will open the circuit. "he voltage for this configuration is usually
'C(( volts.
Re'uirement of Distribution system
A considerable amount of effort is necessary to maintain an electric power
supply within the
re$uirements of various types of consumers. %ome of the re$uirements of a
good distribution
system are* proper voltage, availability of power on demand, and reliabilit
Pro"er (oltage:
Dne important re$uirement of a distribution system is that voltage
variations at consumers5 terminals should be as low as possible. "he changes
in voltage are generally caused due to the variation of load on the system.
.ow voltage causes loss of revenue, inefficient lighting and possible burning
out of motors. 9igh voltage causes lamps to burn out permanently and may
cause failure of other appliances. "herefore, a good distribution system
should ensure that the voltage variations at consumers5 terminals are within
permissible limits. "he statutory limit of voltage variations is E(F of the
rated value at the consumers5 terminals. "hus, if the declared voltage is ,3(
V, then the highest voltage of the consumer should not e&ceed ,'' V while
the lowest voltage of the consumer should not be less than ,6 V.
Availability of Power Demand:
6ower must be available to the consumers in any amount that they may
re$uire from time to time. 0or e&ample, motors may be started or shut down,
lights may be turned on or off, without advance warning to the electric
supply company. As electrical energy cannot be stored, therefore, the
distribution system must be capable of supplying load demands of the
consumers. "his necessitates that operating staff must continuously study
load patterns to predict in advance those ma2or load changes that follow the
known schedules.
Reliability:
<odern industry is almost dependent on electric power for its operation. >omes and office
buildings are lighted, heated, cooled and ventilated by electric power. This calls for reliable
service. ?nfortunately electric power, like everything else that is man-made, can never be
absolutely reliable. >owever, the reliability can be improved to a considerable e$tent by -a.
inter-connected system, -b. reliable automatic control system and -c. providing additional
reserve facilities.
C%&ssi'i$&io# o' Disri!"io# S(se)
" distribution system may be classified according to:
*i+ N&"re o' $"rre#,
"ccording to nature of current, distribution system may be classified as -a. d.c. distribution
system and -b. a.c. distribution system. +ow-a-days a.c. system is universally adopted for
distribution of electric power as it is simpler and more economical than direct current method.
*ii+ T(-e o' $o#sr"$io#,
"ccording to type of construction, distribution system may be classified as -a. overhead system
and -b. underground system. The overhead system is generally employed for distribution as it is
/ to 0( times cheaper than the equivalent underground system. !n general, the underground
system is used at places where overhead construction is impracticable or prohibited by the local
laws.
*iii+ S$.e)e o' $o##e$io#,
"ccording to scheme of connection, the distribution system may be classified as -a. radial
system, -b. ring main system and -c. inter-connected system. )ach scheme has its own
advantages and disadvantages.
R&di&% Disri!"io# S(se)
A radial system has only one power source for a group of customers. A power
failure, shortcircuit, or a downed power line would interrupt power in the
entire line which must be fi&ed before power can be restored. "he figure of
+adial -istribution %ystem is shown as *#
Radial Distribution System
1n this system, separate feeders radiate from a single sub#station and feed
the distributors at one end only. 0igure (a! shows a single line diagram of a
radial system for d.c. -istribution where a feeder D> supplies a distributor
A= at point A. Dbviously, the distributors are fed at one point only i.e. point A
in this case. 0igure (b! shows a single line diagram of radial system for a.c.
distribution. "he radial system is employed only when power is generated at
low voltage and the sub#station is located at the centre of load. "his is the
simplest distribution circuit and has the lowest initial cost.
Single )ine Diagram of Radial Distribution System
*ode Radial Distribution *etwork:-
Objectives of Radial Distribution System:-
. 6lanning, moderni<ation and automation.
,. "o provide service connection to various urban, rural and industrial
consumer in the allocated area.
3. )a&imum security of supply and minimum duration of interruption.
'. %afety of consumers, utility personnel.
/. "o provide electricity of accepted $uality in terms of *#
(a! =alanced three phase supply.
(b! 3ood power factor.
(c! Voltage flicker within permissible limits.
(d! .ess voltage dips.
(e! )inimum interruption in power supply.
Advantages of Radial Distribution System:-
(a! +adial distribution system is easiest and cheapest to build.
(b! "he maintenance is easy.
(c! 1t is widely used in sparsely populated areas.
Drawback of Radial Distribution System:-
(a! "he end of the distributor nearest to the feeding point will be heavily
loaded.
(b! "he consumers are dependent on a single feeder and single distributor.
"herefore, any fault on the feeder or distributor cuts off supply to the
consumers who are on the side of the fault away from the sub#station.
(c! "he consumers at the distant end of the distributor would be sub2ected to
serious voltage fluctuations when the load on the distributor
/ODELINOGF DISTRIBUTIOSYNS TE/
CO/PONENTS
The individual components of & distribution system are modeled by their mathematical
equivalents. The three-phase modeling of distribution system components is given . The series
impedance matri$ of a three-phase line section is given by equation
This equation is obtained after @ronAs reduction. !t takes care of the effects of the neutral or
ground. "t each bus i, the comple$ power S, is given by,
where P:pe' and Q;," are the specified real and reactive powers respectively of bus i. The
equivalent current inBection at bus i for the kfh iteration is given as,
THREE-PHASE DISTRIBUTION LO AD FLOW ANALY.S.I S
<ost of the distributionAsystems are. radial in nature with a single voltage source. This special
property of the distribution system is used to derive various formulations. %ifferent iterative
methods similar to ,auss-eidel method are discussed in this paper. !n this section, the
algorithms for these methods are given.
A. Implicit Z-bus Method
The implicit C-bus method is the most commonly used method 010. The method works on the
principle of superposition as applied to the system bus voltages. "ccording to the principle of
superposition, only one type of source is considered at a time for the calculation of bus voltages.
"n iterative procedure is used in this method. !nitially, all the bus voltages are assumed to be
equal to the swing bus voltage -only swing bus is considered as the source in the system with all
the current inBections at load buses taken as &ero.. !n the ne$t step, since the current inBections
and bus voltages are .dependent on each other, these quantities are required to be determined
iteratively. The swing bus is short-circuited while calculating the component of bus voltages due
to the current inBections.
The following steps are involved in this algorithm:
0. The bus voltages are assumed to have some initia0,value. The D-bus (Y,) is formed.A
6. The current inBections are computed by using equation 7 for which the recent values of bus
voltages are taken.
2. The voltage deviations (VD) due to current inBections are computed by the factori&ation of D-
bus,
f = [Y! [ VD]' -'.
4. The voltage deviations calculated in step 2 are superimposed on the no load bus voltage
(VNL)>. ence, the bus voltages are updatedAas. 3" = VNL + [ VD]' (5)
/. The convergence is checked. !f the method has not converged. then steps from " to ' are
repeated.
. Modi#ied Gauss-Se2.d Method
The implicit C-bus method described earlier requires the factori&ation of the full D-bus matri$,
adversely affecting the performance i# terms of speed. .>ence, a new method has been suggested
in 213 by blending the implicit C-bus method and the ,auss-eidel method to improve the
computational efficiency. Eor a distribution system with n buses, where P:peAand Qtpec are the
specified powemat bus i, the bus voltage for k'" iteration can be calculated by using the ,auss-
eidel method
as.
The values of voltages used in the modified ,auss-eidel method are the most recently
computed values, whereas thevalues of voltages used in the implicit C-bus method are the
Two matrices are developed, vi&. the bus inBection to branch current -F!F4. matri$ and branch
current to bus voltage -F4F#. matri$. Fy . . using simple matri$ multiplication of these two
matrices, the Two developed matrices, F!F4 and F4F# are used to obtain the load flow
solution. The development of these two matrices is e$plained with reference to Eig. . The figure
shows a simple distribution system. !t has sub-station at its bus number !, and bus numbers " to 1
are the load buses loadflowGsolution is obtained
$. %o&'(&d-(c)'(&d *ubstitutio+
!n all the previous methods, the voltages at all the buses in the system are calculated in one step,
by using the matrices. !n forward-backward substitution, the @4; and @#; are applied at each.
node and branch respectively. Fy solving these equations iteratively, the solution is obtained 2!.
The following steps involved in this method:
,ptim(l o&de&i+- o. +odes:
+odes are renumbered according to source node - load node relationship to facilitate the
forward and backward substitution. Thus, a forward path is created from the source node to the
load node and a backward path is traced from the load node to the source node. The branch node
nearer to the source is called as the parent node and the other node is called as the child node.
!nitially, the flat voltage start is assumed.
(c)'(&d substitutio+:
This is used to calculate the current in each branch. The current in the last branch is equal to the
current inBection at the corresponding end node. The voltage values are kept constant. The
network is traced in the backward direction. The currents in all the other branches can be found
out by using @4; as given by the equation.
where I,, ("1). Ib (U,) and I, (ut) are the branch currents of line section m, and ib,, i" and
iLc are the equivalent current inBections at the child node (i) of branch m. M is the set of line
sections connected to mrh branch at its child node (p is the number of a line section which is an
element of <..
respectively. These values of the voltages are used for calculating the currents by backward
substitution in the ne$t iteration.
/hec) for convergence:
The forward and backward substitutions are performed in each iteration of the load flow. The
voltage magnitudes at each bus in an iteration are compared with their values in the previous
iteration. !f the error is within the tolerance limit, the procedure is stopped. *therwise, the steps
of backward substitution, forward substitution and check for convergence are repeated
0. 1(dde& 2et'o&) 3heo&4
The ladder network theory given in [9] is very much similar to the forward-backward
substitution method. Though the basic principle of both the methods is same, there are
differences in the steps of implementation. !n the ladder network theory, the optimal ordering of
nodes is done 'irs. !n the backward substitution, the node voltages are assumed to be equal o
some initial value in .e 'irs iteration. T.e currents in each branch are computed by @4; using
equation 66. !n addition to the branch currents, the node voltages are also computed by using
equation. Thus, the value of the swing bus voltage is &%so determined. This calculated value of
the swing bus voltage is compared with its specified value. !f the error is within the limit, then
the load flow converges9 otherwise the forward substitution is performed &s e$plained in the case
of forward-backward substitution method. Thus, in the ladder network theory, the !"s voltages
are calculated twice in the same iteration as compared to only once for the forward-backward
substitution method. .The convergence is checked in the ladder network theory by comparison
between the specified and calculated voltage values of the swing bus, whereas the difference
between the values of bus voltages at the present and previous iterations is considered for
convergence in the forward-backward substitution method.
Eorward sirbstitufion
This is used to calculate the voltage at each node -starting from the child node of the first branch.
by using @#;. The swing bus voltage is set to its specified value. The current in each branch is
held constant at the value obtained in the backward substitution. Thus, using the branch currents
calculated in the backward substitution, the values of voltages are calculated by using the
equation,
Fast Decou"led Power Flow for Radial Distribution System
1n +adial -istribution %ystem, the large +:; ratio causes problems in
convergence of conventional load flow algorithm. "herefore for the better
convergence some modified load flow
methods are used. 0or the purposes of power flow studies, we model a radial
distribution system as a network of buses connected by distribution lines,
switches, or transformers to a voltage specified source bus. Each bus may
also have a corresponding load, shunt capacitor, and:or co#generator
connected to it. "he model can be represented by a radial interconnection of
copies of the basic building block shown in 0igure '., the dotted lines from
the co#generator, shunt capacitor, and load to ground are to indicate that
these elements may be connected in an ungrounded delta#configuration.
%ince a given branch may be single#phase, two#phase, or three#phase. "he
basic building block of radial distribution systemis shown on the ne&t page
as*#
Figure
!e +asic +uilding +lock of Radial Distribution System#
Dne of the key concepts behind our formulation is that the voltage and
current at one bus can be
e&pressed as a function of the voltage and current at the ne&t bus. 1f we let
the e$uations GH as
"he branch update function GH is given below as*
Bhere Wk is a vector containing the real and imaginary parts of the voltages
and currents at bus
k# "he function gk is determined by the sub#laterals attached at bus k as
well as the models for
distribution lines, switches, transformers, loads, shunt capacitors, and co#
generators. 0rom Vk we
can compute the currents in2ected by the loads, shunt capacitors, and co#
generators. 3iven Ik , -
and the currents Ij in2ected into sub#laterals branching off from bus k, we
apply I>. at bus k to
caculate current GH given as*#
Bhere Ak is the set of buses ad2acent to bus k on sub#laterals.
0rom the following e$uation (,C!, we can solve for the voltage and current at
the primary given
the voltage and current at the secondary GH as*#
"herefore by solving e$uation (,8! , we get
%o that by using this method we get the converged value easily and fast
than the other ordinary
methods.
1n mathematics, an incidence matri. is a matri& that shows the relationship
between two classes of ob2ects. 1f the first class is X and the second is Y, the matri&
has one row for each element of X and one column for each element of Y. "he entry
in row x and column y is if x and y are related (called incident in this conte&t!
and ( if they are not. "here are variationsJ see below.

Gr&-. .eor(
U#dire$ed &#d dire$ed 3r&-.s
"n undirected graph
!n graph theory an undirected graph G has two kinds of incidence matri$: unoriented and
oriented. The i#$ide#$e )&ri4 -or "#orie#ed i#$ide#$e )&ri4. of G is a p H q matri$ -b
ij
.,
where p and q are the numbers of vertices and edges respectively, such that b
ij
I 0 if the verte$ v
i

and edge x
j
are incident and ( otherwise.
Eor e$ample the incidence matri$ of the undirected graph shown on the right is a matri$
consisting of ' rows -corresponding to the four vertices. and ' columns -corresponding to the
four edges.:
The i#$ide#$e )&ri4 of a directed graph D is a p H q matri$ 2b
ij
3 where p and q are the number
of vertices and edges respectively, such that b
ij
I J 0 if the edge x
j
leaves verte$ v
i
, 0 if it enters
verte$ v
i
and ( otherwise. -+ote that many authors use the opposite sign convention..
"n orie#ed i#$ide#$e )&ri4 of an undirected graph G is the incidence matri$, in the sense of
directed graphs, of any orientation of G. That is, in the column of edge e, there is one K0 in the
row corresponding to one verte$ of e and one J0 in the row corresponding to the other verte$ of
e, and all other rows have (. "ll oriented incidence matrices of G differ only by negating some
set of columns. !n many uses, this is an insignificant difference, so one can speak of !e oriented
incidence matri$, even though that is technically incorrect.
The oriented or unoriented incidence matri$ of a graph G is related to the adBacency matri$ of its
line graph L-G. by the following theorem:
where "-L-G.. is the adBacency matri$ of the line graph of G, #-G. is the incidence matri$, and
$
q
is the identity matri$ of dimension q.
The @irchhoff matri$ is obtained from the oriented incidence matri$ M-G. by the formula
M-G.M-G.
%
.
The integral cycle space of a graph is equal to the null space of its oriented incidence matri$,
viewed as a matri$ over the integers or real or comple$ numbers. The binary cycle space is the
null space of its oriented or unoriented incidence matri$, viewed as a matri$ over the two-
element field.
Si3#ed &#d !idire$ed 3r&-.s
The incidence matri$ of a signed graph is a generali&ation of the oriented incidence matri$. !t is
the incidence matri$ of any bidirected graph that orients the given signed graph. The column of a
positive edge has a K0 in the row corresponding to one endpoint and a J0 in the row
corresponding to the other endpoint, Bust like an edge in an ordinary -unsigned. graph. The
column of a negative edge has either a K0 or a J0 in both rows. The line graph and @irchhoff
matri$ properties generali&e to signed graphs.
/"%i3r&-.s
The definitions of incidence matri$ apply to graphs with loops and multiple edges. The column
of an oriented incidence matri$ that corresponds to a loop is all &ero, unless the graph is signed
and the loop is negative9 then the column is all &ero e$cept for L6 in the row of its incident
verte$.
H(-er3r&-.s
Fecause the edges of ordinary graphs can only have two vertices -one at each end., the row of an
incidence matri$ for graphs can only have two non-&ero entries. Fy contrast, a hypergraph can
have multiple vertices assigned to one edge9 thus, the general case describes a hypergraph.
I#$ide#$e sr"$"res
The i#$ide#$e )&ri4 of an incidence structure & is a p H q matri$ 2b
ij
3, where p and q are the
number of -oi#s and %i#es respectively, such that b
ij
I 0 if the point p
i
and line L
j
are incident
and ( otherwise. !n this case the incidence matri$ is also a biadBacency matri$ of the ;evi graph
of the structure. "s there is a hypergraph for every ;evi graph, and vice'vers(, the incidence
matri$ of an incidence structure describes a hypergraph.
Fi#ie 3eo)eries
"n important e$ample is a finite geometry. Eor instance, in a finite plane, ) is the set of points
and * is the set of lines. !n a finite geometry of higher dimension, ) could be the set of points
and * could be the set of subspaces of dimension one less than the dimension of *9 or ) could be
the set of all subspaces of one dimension + and * the set of all subspaces of another dimension e.
B%o$5 desi3#s
"nother e$ample is a block design. >ere ) is a finite set of MpointsM and * is a class of subsets of
), called MblocksM, subBect to rules that depend on the type of design. The incidence matri$ is an
important tool in the theory of block designs. Eor instance, it is used to prove the fundamental
theorem of symmetric 6-designs, that the number of blocks equals the number of points.
Cre&e 3r&-.s 'ro) &# i#$ide#$e )&ri4
Des$ri-io#
graph.incidence creates a bipartite igraph graph from an incidence matri$.
Us&3e
graph.incidence(incidence, directed = FALSE, mode = c("all",
"out",
"in", "total"), multiple = FALSE, weighted = NULL, add.names
= NULL)
Ar3")e#s
incidence
"he input incidence matri&. 1t can also be a sparse matri& from
the atri! package.
directed
.ogical scalar, whether to create a directed graph.
mode
A character constant, defines the direction of the edges in
directed graphs, ignored for undirected graphs. 1f Kout5, then edges
go from vertices of the first kind (corresponding to rows in the
incidence matri&! to vertices of the second kind (columns in the
incidence matri&!. 1f Kin5, then the opposite direction is used. 1f
Kall5 or Ktotal5, then mutual edges are created.
multiple
.ogical scalar, specifies how to interpret the matri& elements. %ee
details below.
weighted
"his argument specifies whether to create a weighted graph from
the incidence matri&. 1f it is NULL then an unweighted graph is
created and the multiple argument is used to determine the edges
of the graph. 1f it is a character constant then for every non#<ero
matri& entry an edge is created and the value of the entry is
added as an edge attribute named by the weighted argument. 1f it
is "#UE then a weighted graph is created and the name of the edge
attribute will be Kweight5.
add.names
A character constant, NA or NULL. graph.incidence can add the row
and column names of the incidence matri& as verte& attributes. 1f
this argument is NULL (the default! and the incidence matri& has
both row and column names, then these are added as the Kname5
verte& attribute. 1f you want a different verte& attribute for this,
then give the name of the attributes as a character string. 1f this
argument is NA, then no verte& attributes (other than type! will be
added.
Incidence Matrices
by Dr Richard Klitzing
(reproduced with permission)
Taken purely abstract, plytpes are described by their surtpial ele!ents plus the
relati"e incidences. The !st basic #ay t gi"e thse incidences are $-%-!atrices #ith
% !eaning &incident& and $ &nt&. 'ut already the easiest plytpes #uld ask (r huge
!atrices. This is the entrance (r sy!!etry, the sy!!etry ( the plytpe itsel(. )like
surtpial ele!ents n# can be classed tgether "ia sy!!etrical e*ui"alence, and the
incidence relatin #ill be gi"en (r the classes instead. This reduces the size ( the
!atri+ cnsiderably. The diagnal ele!ents ( these reduced !atrices #ill gi"e the ttal
cunt ( ele!ents ( each ( the respecti"e e*ui"alence classes. The nn-diagnal
ele!ents ,-(n,!) #ill pr"ide the nu!bers ( incident surtpes ( class ! #ith any (
the ele!ents ( class n. The subdiagnal parts ( the r#s thus still describe the
surtpial ele!ent classes. The superdiagnal parts ( the r#s describe their
en"irn!ental aspacts, i.e. "erte+ (igures, edge (igures etc.
Regular plytpes are bund t pr"ide a single class ( surtpes per di!ensin, but in
general there #ill be !re sy!!etry-ine*ui"alent ele!ents ( the sa!e di!ensinality.
There(re it is cn"eniant t display the di!ensinal brders as #ell as a superi!psed
guiding grid. .erte+-transiti"ity (r instance can be read (( (r! an incidence !atri+
directly, as thse plytpes sh# up nly a single "erte+ class.
/ere as an e+a!ple the incidence !atri+ ( the truncated cube is gi"en. The Dynkin
diagra!s ( the relati"e classes are pr"ided in additin in (rnt ( the r#s.
0+1+
. . . $$ %& $ % ' $ ' %
(((((())(((()((((((()((((
. ! . $$ % $ %& * $ ' '
. . ! $$ % $ * '% $ + %
(((((())(((()((((((()((((
o,! . $$ , $ , + $ - *
. !&! $$ - $ & & $ * .
This matri$ shows that there are 6' vertices, all having the same symmetry -upper-left element..
The lowest two rows show that the 6-dimensional elements have 7 or 5 vertices -lower-left
block. and therefore are triangles or octagons. The rightmost entries of the first row show further
that at each verte$ 0 such triangle and 6 octagons are incident. Eurther there are 6 types of edges,
the upper one is incident to 0 triangle and 0 octagon, the other one is incident to 6 octagons only.
The middle block of the bottom two rows shows that the triangle will have all edges of the first
type clearly, but the octagons do use edges of both types -alternatingly.. "ltogether there are 5
triangles and 1 octagons -lower-right block..
Two relations on these numbers are generally valid. The one is the equation
!N-n,n.O!N-n,m.I!N-m,n.O!N-m,m.. This is true whether incident representants of those classes of
subpolytopes do e$ist or not, as in the latter case the corresponding non-diagonal elements are
both &ero. The other observation is, and this derives right from the diagrammatic representation
of the (-0-matri$, the so called >asse diagram, that this diagram read top-down instead of
bottom-up would describe the dual abstract polytope. The same is even true for the reduced
matrices, where the matri$ of the dual polytope can be read off by Bust rotating the matri$ half
way around an a$is orthogonal to the writing plane, thereby interchanging counts of vertices and
facets, or dualising the numbers of the verte$ figures into those of facets and vice versa. Eurther-
on to each of the subdiagonal parts of the rows, the superdiagonal parts of the rows, and the
diagonal itself, the )uler formula might be applied9 but appropriate e$tensions like genus,
density etc. would have to be considered.
+ote that the same polytope might be a fi$-element under different symmetry groups. Thus there
could be different -reduced. incidence matrices, all describing the same polytope. )specially the
identity map, taken as reducing symmetry, would reproduce the (-0-matri$. *n the other hand
incidence matrices Bust like >asse diagrams only depend on the structure of the abstract
polytope. That is, different isomorph realisations of it would have the same incidence matri$. Eor
instance a conve$ polygon PnQ abstractly can not be distinguished from the polygram PnRdQ as
long there are no incidences of different types.
POWER 6UALITY
The contemporary container crane industry, like many other industry segments, is often
enamored by the bells and whistles, colorful diagnostic displays, high speed performance, and
levels of automation that can be achieved. "lthough these features and their indirectly related
computer based enhancements are key issues to an efficient terminal operation, we must not
forget the foundation upon which we are building. Power quality is the mortar which bonds the
foundation blocks. Power quality also affects terminal operating economics, crane reliability, our
environment, and initial investment in power distribution systems to support new crane
installations. To quote the utility company newsletter which accompanied the last monthly issue
of my home utility billing: G?sing electricity wisely is a good environmental and business
practice which saves you money, reduces emissions from generating plants, and conserves our
natural resources.S "s we are all aware, container crane performance requirements continue to
increase at an astounding rate. +e$t generation container cranes, already in the bidding process,
will require average power demands of 0/(( to 6((( kT = almost double the total average
demand three years ago. The rapid increase in power demand levels, an increase in container
crane population, 4U converter crane drive retrofits and the large "4 and %4 drives needed to
power and control these cranes will increase awareness of the power quality issue in the very
near future.
P*T)U Q?";!TD PU*F;)<
Eor the purpose of this article, we shall define power quality problems as:
G"ny power problem that results in failure or mis operation of customer equipment, manifests
itself as an economic burden to the user, or produces negative impacts on the environment.S
Then applied to the container crane industry, the power issues which degrade power quality
include:
V Power Eactor
V >armonic %istortion
V #oltage Transients
V #oltage ags or %ips
V #oltage wells
The "4 and %4 variable speed drives utili&ed on board container cranes are significant
contributors to total harmonic current and voltage distortion. Thereas 4U phase control creates
the desirable average power factor, %4 4U drives operate at less than this. !n addition, line
notching occurs when 4USs commutate, creating transient peak recovery voltages that can be 7
to ' times the nominal line voltage depending upon the system impedance and the si&e of the
drives. The frequency and severity of these power system disturbances varies with the speed of
the drive. >armonic current inBection by "4 and %4 drives will be highest when the drives are
operating at slow speeds. Power factor will be lowest when %4 drives are operating at slow
speeds or during initial acceleration and deceleration periods, increasing to its ma$imum value
when the 4USs are phased on to produce rated or base speed. "bove base speed, the power
factor essentially remains constant. ?nfortunately, container cranes can spend considerable time
at low speeds as the operator attempts to spot and land containers. Poor power factor places a
greater k#" demand burden on the utility or engine-alternator power source. ;ow power factor
loads can also affect the voltage stability which can ultimately result in detrimental effects on the
life of sensitive electronic equipment or even intermittent malfunction. #oltage transients created
by %4 drive 4U line notching, "4 drive voltage chopping, and high frequency harmonic
voltages and currents are all significant sources of noise and disturbance to sensitive electronic
equipment
!t has been our e$perience that end users often do not associate power quality problems
with 4ontainer cranes, either because they are totally unaware of such issues or there was no
economic 4onsequence if power quality was not addressed. Fefore the advent of solid-state
power supplies, Power factor was reasonable, and harmonic current inBection was minimal. +ot
until the crane Population multiplied, power demands per crane increased, and static power
conversion became the way of life, did power quality issues begin to emerge. )ven as harmonic
distortion and power Eactor issues surfaced, no one was really prepared. )ven today, crane
builders and electrical drive ystem vendors avoid the issue during competitive bidding for new
cranes. Uather than focus on "wareness and understanding of the potential issues, the power
quality issue is intentionally or unintentionally ignored. Power quality problem solutions are
available. "lthough the solutions are not free, in most cases, they do represent a good return on
investment. >owever, if power quality is not specified, it most likely will not be delivered.
Power quality can be improved through:
V Power factor correction,
V >armonic filtering,
V pecial line notch filtering,
V Transient voltage surge suppression,
V Proper earthing systems.
!n most cases, the person specifying andRor buying a container crane may not be fully
aware of the potential power quality issues. !f this article accomplishes nothing else, we would
hope to provide that awareness.
!n many cases, those involved with specification and procurement of container cranes
may not be cogni&ant of such issues, do not pay the utility billings, or consider it someone elseSs
concern. "s a result, container crane specifications may not include definitive power quality
criteria such as power factor correction andRor harmonic filtering. "lso, many of those
specifications which do require power quality equipment do not properly define the criteria.
)arly in the process of preparing the crane specification:
V 4onsult with the utility company to determine regulatory or contract requirements that must be
satisfied, if any.
V 4onsult with the electrical drive suppliers and determine the power quality profiles that can be
e$pected based on the drive si&es and technologies proposed for the specific proBect.
V )valuate the economics of power quality correction not only on the present situation, but
consider the impact of future utility deregulation and the future development plans for the
terminal.
T>) F)+)E!T *E P*T)U Q?";!TD
Power quality in the container terminal environment impacts the economics of the terminal
operation, affects reliability of the terminal equipment, and affects other consumers served by the
same utility service. )ach of these concerns is e$plored in the following paragraphs.
0. )conomic !mpact
The economic impact of power quality is the foremost incentive to container terminal operators.
)conomic impact can be significant and manifest itself in several ways:
a. Power Eactor Penalties
<any utility companies invoke penalties for low power factor on monthly billings. There
is no industry standard followed by utility companies. <ethods of metering and calculating
power factor penalties vary from one utility company to the ne$t. ome utility companies
actually meter k#"U usage and establish a fi$ed rate times the number of k#"U-hours
consumed. *ther utility companies monitor k#"U demands and calculate power factor. !f the
power factor falls below a fi$ed limit value over a demand period, a penalty is billed in the form
of an adBustment to the peak demand charges. " number of utility companies servicing container
terminal equipment do not yet invoke power factor penalties. >owever, their service contract
with the Port may still require that a minimum power factor over a defined demand period be
met. The utility company may not continuously monitor power factor or k#"U usage and reflect
them in the monthly utility billings9 however, they do reserve the right to monitor the Port
service at any time. !f the power factor criteria set forth in the service contract are not met, the
user may be penali&ed, or required to take corrective actions at the userSs e$pense. *ne utility
company, which supplies power service to several east coast container terminals in the ?",
does not reflect power factor penalties in their monthly billings, however, their service contract
with the terminal reads as follows:
GThe average power factor under operating conditions of customerSs load at the point
where service is metered shall be not less than 5/8. !f below 5/8, the customer may be required
to furnish, install and maintain at its e$pense corrective apparatus which will increase the Power
factor of the entire installation to not less than 5/8. The customer shall ensure that no e$cessive
harmonics or transients are introduced on to the 2utility3 system. This may require special power
conditioning equipment or filters. The !))) td. /0W-0WW6 is used as a guide in %etermining
appropriate design requirements.S
The Port or terminal operations personnel, who are responsible for maintaining container
cranes, or specifying new container crane equipment, should be aware of these requirements.
?tility deregulation will most likely force utilities to enforce requirements such as the e$ample
above. Terminal operators who do not deal with penalty issues today may be faced with some
rather severe penalties in the future. " sound, future terminal growth plan should include
contingencies for addressing the possible economic impact of utility deregulation.
b. ystem ;osses
>armonic currents and low power factor created by nonlinear loads, not only result in
possible power factor penalties, but also increase the power losses in the distribution system.
These losses are not visible as a separate item on your monthly utility billing, but you pay for
them each month. 4ontainer cranes are significant contributors to harmonic currents and low
power factor. Fased on the typical demands of todaySs high speed container cranes, correction of
power factor
alone on a typical state of the art quay crane can result in a reduction of system losses that
converts to a 1 to 0(8 reduction in the monthly utility billing. Eor most of the larger terminals,
this is a significant annual saving in the cost of operation.
c. Power ervice !nitial 4apital !nvestments
The power distribution system design and installation for new terminals, as well as
modification of systems for terminal capacity upgrades, involves high cost, speciali&ed, high and
medium voltage equipment. Transformers, switchgear, feeder cables, cable reel trailing cables,
collector bars, etc. must be si&ed based on the k#" demand. Thus cost of the equipment is
directly related to the total k#" demand. "s the relationship above indicates, k#" demand is
inversely proportional to the overall power factor, i.e. a lower power factor demands higher k#"
for the same kT load. 4ontainer cranes are one of the most significant users of power in the
terminal. ince container cranes with %4, 1 pulse, 4U drives operate at relatively low power
factor, the total k#" demand is significantly larger than would be the case if power factor
correction equipment were supplied on board each crane or at some common bus location in the
terminal. !n the absence of power quality corrective equipment, transformers are larger,
switchgear current ratings must be higher, feeder cable copper si&es are larger, collector system
and cable reel cables must be larger, etc. 4onsequently, the cost of the initial power distribution
system equipment for a system which does not address power quality will most likely be higher
than the same system which includes power quality equipment.
6. )quipment Ueliability
Poor power quality can affect machine or equipment reliability and reduce the life of
components. >armonics, voltage transients, and voltage system sags and swells are all power
quality problems and are all interdependent. >armonics affect power factor, voltage transients
can induce harmonics, the same phenomena which create harmonic current inBection in %4 4U
variable speed drives are responsible for poor power factor, and dynamically varying power
factor of the same drives can create voltage sags and swells. The effects of harmonic distortion,
harmonic currents, and line notch ringing can be mitigated using specially designed filters.
7. Power ystem "dequacy
Then considering the installation of additional cranes to an e$isting power distribution system, a
power system analysis should be completed to determine the adequacy of the system to support
additional crane loads. Power quality corrective actions may be dictated due to inadequacy of
e$isting power distribution systems to which new or relocated cranes are to be connected. !n
other words, addition of power quality equipment may render a workable scenario on an e$isting
power distribution system, which would otherwise be inadequate to support additional cranes
without high risk of problems.
'. )nvironment
+o issue might be as important as the effect of power quality on our environment.
Ueduction in system losses and lower demands equate to a reduction in the consumption of our
natural nm resources and reduction in power plant emissions. !t is our responsibility as occupants
of this planet to encourage conservation of our natural resources and support measures which
improve our air quality
R"r&% &re&s
Uural areas are large and isolated areas of an open country with low population density. The
terms McountrysideM and Mrural areasM are not synonyms: a McountrysideM refers to rural areas that
are open. Eorest, wetlands, and other areas with a low population density are not a countryside.
"bout W0 percent of the rural population now earn salaried incomes, often in urban areas. The 0(
percent who still produce resources generate 6( percent of the worldSs coal, copper, and oil9 0(
percent of its wheat, 6( percent of its meat, and /( percent of its corn. The efficiency of these
farms is due in large part to the commerciali&ation of the farming industry, and not single family
operations
Vo%&3e $o#ro%
!n those pre-digital <!%! days, synths used a different system to control themselves and each
other. !nstead of digital bits and bytes, information was passed between modules through wires
that carried a voltage.
" voltage is Bust a measure of how much electrical ApushA a circuit has. Plug a voltage source -
like a battery, or synth module - into a circuit and it will push the electricity around so it starts
flowing. The amount of this push - you can think of it as electrical pressure - is measured in units
called #olts.
!n an analogue synth, voltages are used to control how much each module does what itXs
designed to do. Turn up the voltages to an amplifier, for e$ample, and the sound gets louder. %o
the same to an oscillator and its pitch pitch goes up. Try it with a filter and the filter opens.
<odulation sources - low-frequency oscillators, "%Us, and so on - are cunningly designed
bo$es that ramp voltages up and down automatically in predicted ways. Tithout them, Dou and
Dour friends would have to turn knobs and dials by hand every time Dou hit a note. <ost synths
use a standard voltage-control system. This defines a common one-volt-per-octave scale - in
other words, every time the control voltage goes up by one volt, the frequency of an oscillator
doubles. Turn up the voltage by 0R06th of a volt, and the pitch goes up by a semitone.
4onsequently, a '# signal would cause an oscillator to produce a pitch one octave higher than a
7# signal, and thatXs the theory of one-volt-per-octave.
*ne fo the clever -and strange- things about analogue synthesis is that you can interchange
control voltage and audio lines, because thereXs no practical difference between the two. "n
audio signal is Bust a voltage thatXs wobbling up and down particularly fast, but itXs still basically
a voltage, Bust like youXll get from any module.
This means you can use the output of an oscillator to change the pitch of another, or of a filter, or
the gain of a amplifier -vca.. This gives you acces to strange and unusual effects that you canXt
create any other way.
G&e &#d Tri33er
The gate signal told the synth that you had pressed a key. This voltage, usually in the /#-0(#
range, would remain constant as long as the key was held. "s soon as the key was released, the
gate signal would stop -drop to *#., and the synthAs envelope would immediately Bump to its
release stage.
The trigger signal also told the synth you were playing a note, but unlike a gate, it was a
momentary -about /ms. rather than continuous signal, and could not tell the synth to produce a
sound9 it worked in conBunction with the gate. The trigger signalAs purpose was to start the synthAs
envelope generators, thus articulating the attack of the note. Thenever the synth received a new
trigger, the envelope generators would be restarted, and the attack of that new note would be
articulated. Tithout a trigger signal, a new note would sound using the current state of the
envelope generator - much the same as what hoppens in the AlegatoA mode found on modern
synths.
Triggers came in two varieties. The first, used by "UP instruments, was a momentary spike
where the voltage Bumped from *# to 0(# then back down to *#. The other type, called an -
trigger -or switch trigger., was used on <oog synthesi&ers. !t consisted of a continuous gate-type
voltage that dropped to *# when a key was pressed. This voltage was used to control a switch
that produced a trigger when it closed in response to the voltage drop.
VOLTAGE REGULATOR PRINCIPLE OF OPERATION
The proposed voltage regulator provides discrete serial voltage compensation. #oltage is
compensated using tapswitching, which combines two regulation principles: voltage
ratio regulation and polarity selection. The voltage regulator has a transformer with two
independent primary windings, which is fed from the network, and a secondary
compensationwinding, which is serial connected. %ifferent compensation values can be obtained
by changing the connections between the primary and secondary windings, using three power
contactors, with four poles each. "n automatic controller measures the output voltage and selects
the optimal voltage compensation connection. The characteristics of the proposed design make it
suitable for the needs of rural distribution networks:
Y,ep vo-(ge reg.-(ion: voltage is adBusted within the required quality range, and for
industrial or commercial applications in rural networks there is usually no need for an accurate
voltage regulation based on small steps or continuous regulation.
Y/ob.sness: the voltage regulator will be usually placed outdoors, in dispersed locations, some
of them with difficult access. Eor greater reliability and easy maintenance, electromechanical
contactors are preferred to power electronics.
YLo0 cos: using serial voltage compensation instead of a full power converter reduces the
device si&e and cost, increasing the efficiency notably. <oreover, the higher reliability
reduces the maintenance costs. ection !!! in this paper presents a cost analysis of the voltage
regulator. The regulator is intended for the use in distribution networks, both indoors or
outdoors9 for instance, pole mounted in overhead lines -Eig. 0.. The design proposed in this
paper can be used for one-phase and three-phase voltage regulators. This paper is focused on the
one-phase voltage regulator, including a description of its power circuit and the control system.
A. Po'e& /i&cuit
The power circuit of a one-phase voltage regulator consists of a multiwinding transformer -see
Eig. 6., with two primary windings and a secondary serial compensation winding. !n addition,
three power contactors are used to connect the windings and the network. Tith this design, five
different voltage ratios can be achieved in the transformer by changing the connection between
the windings -see Table !.. The selection of the adequate compensation step is performed by a
microprocessor-based control unit. The primary windings of the transformer P0 and P6 have the
same number of turns , and they are connected to the input voltage with contactor 40. The
primary windings can be parallel or series connected using contactor 46, so the effective number
of turns can be or 6 respectively. )ach connection determines a certain voltage ratio, as indicated
in Table !. The secondary winding of the transformer is the serial compensation winding, and
can be series connected with the distribution network or bypassed using contactor40. Einally, the
polarity of the magnetic coupling is set by contactor 47. The output voltage is then increased or
decreased depending on contactor 47. The output voltage can be formulated as the input voltage
plus the compensation voltage set by the regulator -0.. The compensation voltage is given by the
ratio of the secondary winding turns and the primary winding turns , times
the voltage at the input of the device
where:
Y is the number of turns of windings P0 or P69
Y is the number of turns of winding 9
Y is the connection constant, which can be 0 or 6 for parallel
and series connection respectively of P0 and P6. !n addition this value will be positive for direct
winding coupling, and negative for inverse coupling of windings. The proposed design of the
voltage regulator has five different compensation steps can be achieved by changing the position
of the three contactors, as is shown in Table !. The standby mode of the voltage regulator is set
with the three contactors opened, and guarantees that the device is disconnected from the
distribution network. >ence, the secondary winding is short-circuited and input and output
voltages are the same, as . This stand-by mode protects loads connected to the voltage regulator
from any failure of the device. ,iven the design of the voltage regulator, the rated power of the
transformer is lower than the power that the voltage regulator can supply -6.. The power
difference depends on the ratio between the compensation and the rated voltages. Eor instance,
the transformer of a voltage regulator with '(-# compensation and 67(-# rated voltage, will
have a rated power of 0:8 the ma$imum load that the voltage regulator can supply
. /o+t&ol *4stem
The obBective of the voltage regulator is to improve the line voltage whenever it can be achieved,
and to guarantee system security in the event of a failure of the voltage regulator or severe
contingency in the distribution network. !n this situation, the device will be automatically
disconnected from the network. Eor this purpose, the voltage regulator includes a control system
that consists of three modules -Eig. 7..
Y%!e Me(s.re Mo+.-e registers voltage and current at the voltage regulator output. )very 76
cycles, average U< voltage values are calculated, and these are used to decide if voltage
compensation is needed. Eor the sake of security, if the voltage e$ceeds the security range - , .
the voltage regulator will be disconnected from the network.
Y%!e &omp(rison Mo+.-e compares the average value of a set of output voltage values with the
reference voltage range - , ., and decides if a new compensation step is required.
Y%!e &ompens(ion Mo+.-e receives a step-change order from the 4omparison <odule and
opens or closes the three contactors to achieve a new compensation step in the voltage regulator.
" voltage compensation e$ample is presented in Eig. ', using a voltage regulator with a '(-#
secondary, and two 67(-# primary windings. The reference voltage values, and , are in dash
lines, and the solid lines represent the theoretical relationship between the input and output
voltages for different compensation steps. !n the e$ample the initial voltage at the output is 66(
# at point ", which is within the reference voltage range. #oltage suddenly decreases to 0W( #
-point F.. Then, the
voltage regulator corrects the voltage with two maneuvers, using steps ' and / -path F-4-%..
II. DESIGN AND CONSTRUCTION, PRACTICAL CONSIDERATIONS
This section discusses various aspects of the construction of the voltage regulator, given the need
to guarantee its correct and safe operation.
". %r(ns1ormer Design
" shell-type transformer has been selected for its hardness and its lower operation temperature,
which makes it more appropriate for achieving greater efficiency in the functioning of the
voltage regulator. " software tool has been developed to analy&e the costs of different
transformer column designs. This tool analy&es the active material cost and the energy losses
cost of the transformer 2003. The cost analysis for a 67(R'(-# transformer is shown in Eig. /,
where different designs for the column length of the magnetic core are presented. The cost
calculation assumes the following prices: the steel lamination cost is )uros 7Rkg, the copper cost
is )uros 5./Rkg, and the energy losses for 1( ((( hours at )uros (.0RkTh. ,iven the results in
Eig. /, the optimal transformer design has a column length of W7 mm. Eor this design, total
transformer costs are )uros /55./6, which can be split into )uros 'W6.16 for energy losses and
)uros W/.W for active material cost. " detailed electro-magnetic analysis for the selected length
of the 67(R'(-# transformer has been performed with "+*ETSs
<a$well oftware 2063. This analysis corresponds to the operation at point % of the e$ample
presented in Eig. ', and voltages in the windings are shown in Eig. 1. The ma$imum flu$ density
distribution in the transformer for the voltage regulator is shown in Eig. :. Te observe that the
transformer has been designed with a low flu$ density in order to reduce the energy losses in the
steel laminations.
#. Vo-(ge /e1erence /(nge De1iniion
The reference voltage range will be defined usually in accordance with the quality
standards in each country. Eor instance, if voltage requirements at the load are 67( #
:8, the proposed settings for the voltage regulator will be
, and . >owever, for certain voltage-sensitive applications, a reduced reference voltage range
may be required. !n this case, some adaptations in the design are required to avoid oscillations in
the compensation maneuvers. !f the control system step-ups because output voltage is below
the reference , the new compensated voltage
should always be lower than the reference voltage , to avoid oscillations. This constraint can be
formulated for each step c as follows:
The proposed design of the voltage regulator can be then adapted to a small-voltage reference
range. The values can be achieved by selecting the adequate winding turns of the primary
and secondary windings.
#. &ompens(ion M(ne.vers ,eq.ence
"s defined in the previous section, a compensation maneuver changes the position of the
contactors of the device. The compensation maneuver starts with the disconnection of the
contactor 40. Then, contactors 46 and 47 are opened or closed depending on the compensation
step needed. This operation is performed off-load, which minimi&es possible transients and
e$tends the service life of contactors 46 and 47. "nd finally, 40 is closed, and a new
compensation step is obtained. )$perimental results for the compensation maneuver are shown
in Eig. 5, which shows voltage in the coil of contactor 40, and voltage in the primary winding,.
&. ,0ic!ing &onro-
The reliability and e$pected life of the voltage regulator is mainly determined by the power
contactor 40, which operates on-load. ?nfavorable switching conditions will shorten the life
of the contactor, and some malfunctioning can occur when the contactor fails to open -because
contactor contacts are welded. or fails to close -because contacts have lost their conducting
surface.. Three improvements have been implemented in the design of the voltage regulator to
enhance the switching maneuver of contactor 40.
YTwo poles of contactor 40 are parallel connected to open the load current -see Eig. 0.. The
other two poles are series connected to open the rated voltage.
Y" capacitor is connected in the primary winding P0.
YThe 4ompensation <odule guarantees the switching of contactors to &ero current crossing
2073.
D. 211icienc3 %es
The efficiency of a voltage regulator, whose technical data are enclosed in the Table !!, has been
tested. The tests have been carried out with a variable voltage source, in order to emulate the real
operation in a non-constant voltage network. "dBustable loads at power factor 0 and power factor
(.5 have been used to
charge the voltage regulator. The results of the efficiency tests are shown in Table !!! and Table
!#. "lthough the voltage regulator has '.78 total energy losses, its efficiency at any load is
remarkable. imilarly noteworthy is the fact that, due to its 7 ((( #" transformer, this voltage
regulator can manage apparent power at the output of up to 01 /(( #". !n short, the voltage
regulator design presented in this paper offers low cost and high efficiency.
REFERENCES
203 Z#oltage characteristics of electricity supplied by public distribution networks,[ in 4roc.
&2N2L2&, Frussels, Felgium, 6((0.
263 "+! 45'.0 "merican +ational tandard for )lectric Power ystems and )quipmentY
#oltage Uatings -1( >&. "merican +ational tandards !nstitute, 6((1.
273 U. <alaman, \. "fonso, ;. ;o chiavo, ". Uomero, 4. epulveda, U. #roliBk, and F.
Tharmby, Quality of )lectricity upply: !nitial Fenchmarking on "ctual ;evels, tandards and
Uegulatory trategies 4ouncil of )uropean )nergy Uegulators -4))U., 6((0.
2'3 U% 0W//R6((( de 0 de diciembre por el que se regulan las actividades de transporte,
distribuci]n, comerciali&aci]n, suministro y procedimientos de autori&aci]n de instalaciones de
energ^a el_ctrica F.*.)., %ec. 6:, 6(((, pain 2*nline3. "vailable: http:RRwww.boe.es
2/3 ,. alis and ". afiagianni, Z)conomically optimi&ation of a radial power network based on
voltage criterion,[ "+v. 2ng. ,o10., vol. 66, no. 0, pp. 0=6(, 0WW/.
213 T. ). @a&ibwe and <. >. endaula, 2-ecric 4o0er 5.(-i3 &onro- %ec!niq.es. +ew Dork:
#an +ostrand Ueinhold, 0WW7.
2:3 %. >. \ang and ,. >. 4hoe, Ztep-upRdown "4 voltage regulator using transformer with tap
changer andPT<"4 chopper,[ $222 %r(ns. $n+. 2-ecron., vol. '/, no. 1, pp. W(/=W00, %ec.
0WW5.
253 . ,. Peschel, ". <olden, and *. Tonello, Z!n-;ine FuckRFoost#oltage Uegulation ystems
and "pparatus,[ ?.. Patent /5'''(6, %ec. 0WW5.
2W3 F. <eyer, \. erdyn, >. \. Feukes, and U. ,. tephen, Ztretching low-voltage electrification
networks by means of electronic voltage regulators,[ in 4roc. 4o0er 2ngineering ,ocie3
,.mmer Meeing, \ul. 0/=0W, 6((0, vol. 0, pp. :6=::.
20(3 *. <. U. ,arcia and ". ,. )$posito, Zolid-state voltage regulator for dispersed rural
distribution systems,[ in 4roc. $222 4o0er %ec! &on1. , Fologna, !taly, \un. 6((7, vol. 7, pp.
67=61.
2003 \. >. >arlow, 2-ecric 4o0er %r(ns1ormer 2ngineering. Foca Uaton,
E;: 4U4 Press, 6(('.
2063 M(x0e-- ,o10(re. )lmwood Park, +\: "nsoft 4orporation, 0WW5.
2073 4. >. Elurscheim, 4o0er &irc.i #re(6er %!eor3 (n+ Design, ser. !)) Power )ngineering
eries. ;ondon, ?.@.: !)), 0W56.
20'3 <. T. Fishop, \. %. Eoster, and %. ". %own, ZThe application of singlephase voltage
regulators on three-phase distribution systems,[ in 4roc.
37! "nn.. /.r(- 2-ecric 4o0er &on1., 4olorado prings, 4*, "pr. 0WW'.
20/3 ,. ". <anos, %. <akridou, and 4. %. #ournas, Z;ocal and global bifurcations in a power
system with two cascaded regulators,[ in 4roc. 7! $222 Me+ierr(ne(n &on1. &onro- (n+
".om(ion, Patra, ,reece, \ul. 6(((.