Transformers

Electrical Machines I
Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao
Transformers
1 Introduction
Michael Faraday propounded the principle of electro-magnetic induction in 1831. It states that a voltage appears across the terminals of an electric coil when the ux linked with the same changes. The magnitude of the induced voltage is proportional to the rate of change of the ux linkages. This nding forms the basis for many magneto electric machines. The earliest use of this phenomenon was in the development of induction coils. These coils were used to generate high voltage pulses to ignite the explosive charges in the mines. As the d.c. power system was in use at that time, very little of transformer principle was made use of. In the d.c. supply system the generating station and the load center have to be necessarily close to each other due to the requirement of economic transmission of power. Also the d.c. generators cannot be scaled up due to the limitations of the commutator. This made the world look for other ecient methods for bulk power generation and transmission. During the second half of the 19th century the alternators, transformers and induction motors were invented. These machines work on alternating power supply. The role of the transformers became obvious. The transformer which consisted of two electric circuits linked by a common magnetic circuit helped the voltage and current levels to be changed keeping the power invariant. The eciency of such conversion was extremely high. Thus one could choose a moderate voltage for the generation of a.c. power, a high voltage for the transmission of this power over long distances and nally use a small and safe operating voltage at the user end. All these are made possible by transformers. The a.c. power systems thus got well established.
Indian Institute of Technology Madras
Transformers can link two or more electric circuits. In its simple form two electric circuits can be linked by a magnetic circuit, one of the electric coils is used for the creation of a time varying magnetic led. The second coil which is made to link this eld has an induced voltage in the same. The magnitude of the induced emf is decided by the number of turns used in each coil. Thus the voltage level can be increased or decreased by changing the number of turns. This excitation winding is called a primary and the output winding is called a secondary. As a magnetic medium forms the link between the primary and the secondary windings there is no conductive connection between the two electric circuits. The transformer thus provides an electric isolation between the two circuits. The frequency on the two sides will be the same. As there is no change in the nature of the power, the resulting machine is called a transformer and not a converter. The electric power at one voltage/current level is only transformed into electric power, at the same frequency, to another voltage/current level.
Even though most of the large-power transformers can be found in the power systems, the use of the transformers is not limited to the power systems. The use of the principle of transformers is universal. Transformers can be found operating in the frequency range starting from a few hertz going up to several mega hertz. Power ratings vary from a few milliwatts to several hundreds of megawatts. The use of the transformers is so wide spread that it is virtually impossible to think of a large power system without transformers. Demand on electric power generation doubles every decade in a developing country. For every MVA of generation the installed capacity of transformers grows by about 7MVA. These gures show the indispensable nature of power transformers.
Basic Principles
As mentioned earlier the transformer is a static device working on the principle of
Faradays law of induction. Faradays law states that a voltage appears across the terminals of an electric coil when the ux linkages associated with the same changes. This emf is proportional to the rate of change of ux linkages. Putting mathematically, e= d dt (1)
Where, e is the induced emf in volt and is the ux linkages in Weber turn. Fig. 1 shows a
Figure 1: Flux linkages of a coil
coil of N turns. All these N turns link ux lines of Weber resulting in the N ux linkages. In such a case, = N and e=N d dt volt (3) (2)
The change in the ux linkage can be brought about in a variety of ways coil may be static and unmoving but the ux linking the same may change with time. 3
ux lines may be constant and not changing in time but the coil may move in space linking dierent value of ux with time. both 1 and 2 above may take place. The ux lines may change in time with coil moving in space. These three cases are now elaborated in sequence below, with the help of a coil with a simple geometry.
L B
Figure 2: Static coil
Fig. 2 shows a region of length L m, of uniform ux density B Tesla, the ux lines being normal to the plane of the paper. A loop of one turn links part of this ux. The ux linked by the turn is L B X Weber. Here X is the length of overlap in meters as shown in the gure. If now B does not change with time and the loop is unmoving then no emf is induced in the coil as the ux linkages do not change. Such a condition does not yield any useful machine. On the other hand if the value of B varies with time a voltage is induced in the coil linking the same coil even if the coil does not move. The magnitude of B 4
is assumed to be varying sinusoidally, and can be expressed as, B = Bm sin t (4)
where Bm is the peak amplitude of the ux density. is the angular rate of change with time. Then, the instantaneous value of the ux linkage is given by, = N = N LXBm sin t The instantaneous value of the induced emf is given by, e= d = N m . cos t = N m .. sin(t + ) dt 2 (6) (5)
Here m = Bm .L.X. The peak value of the induced emf is em = N m . and the rms value is given by E= N m . 2 volt. (7)
Further, this induced emf has a phase dierence of /2 radian with respect to the ux linked by the turn. This emf is termed as transformer emf and this principle is used in a transformer. Polarity of the emf is obtained by the application of Lenzs law. Lenzs law states that the reaction to the change in the ux linkages would be such as to oppose the cause. The emf if permitted to drive a current would produce a counter mmf to oppose this changing ux linkage. In the present case, presented in Fig. 2 the ux linkages are assumed to be increasing. The polarity of the emf is as indicated. The loop also experiences a compressive force.
Fig. 2(b) shows the same example as above but with a small dierence. The ux density is held constant at B Tesla. The ux linked by the coil at the current position is 5
= B.L.X Weber. The conductor is moved with a velocity v = dx/dt normal to the ux, cutting the ux lines and changing the ux linkages. The induced emf as per the application of Faradays law of induction is e = N.B.L.dx/dt = B.L.v volt.(Here N=1)
Please note,the actual ux linked by the coil is immaterial. Only the change in the ux linkages is needed to be known for the calculation of the voltage. The induced emf is in step with the change in and there is no phase shift. If the ux density B is distributed sinusoidally over the region in the horizontal direction, the emf induced also becomes sinusoidal. This type of induced emf is termed as speed emf or rotational emf, as it arises out of the motion of the conductor. The polarity of the induced emf is obtained by the application of the Lenzs law as before. Here the changes in ux linkages is produced by motion of the conductor. The current in the conductor, when the coil ends are closed, makes the conductor experience a force urging the same to the left. This is how the polarity of the emf shown in g.2b is arrived at. Also the mmf of the loop aids the eld mmf to oppose change in ux linkages. This principle is used in d.c machines and alternators.
The third case under the application of the Faradays law arises when the ux changes and also the conductor moves. This is shown in Fig. 2(c).
The uniform ux density in space is assumed to be varying in magnitude in time as B = Bm sin t. The conductor is moved with a uniform velocity of change in the ux linkages and hence induced emf is given by e = N. d(Bm . sin t.L.X ) dx = N.L.X.Bm .. cos t. + N.Bm . sin t.L. V olt. dt dt (8)
dx dt
= v m/sec. The
The rst term is due to the changing ux and hence is a transformer emf. The second term is due to moving conductor or is a speed emf. When the terminals are closed such as to permit a current the conductor experiences a force and also the mmf of the coil opposes the change in ux linkages. This principle is used in a.c. machines where the eld is time varying and conductors are moving under the same.
The rst case where there is a time varying eld and a stationary coil resulting in a transformer emf is the subject matter in the present section. The case two will be revisited under the study of the d.c machines and synchronous machines. Case three will be extensively used under the study of a.c machines such as induction machines and also in a.c. commutator machines.
Next in the study of the transformers comes the question of creating a time varying led. This is easily achieved by passing a time varying current through a coil. The winding which establishes the eld is called the primary. The other winding, which is kept in that eld and has a voltage induced in it, is called a secondary. It should not be forgotten that the primary also sees the same time varying eld set up by it linking its turns and has an induced emf in the same. These aspects will be examined in the later sections. At rst the common constructional features of a transformer used in electric power supply system operating at 50 Hz are examined.
Constructional features
Transformers used in practice are of extremely large variety depending upon the
end use. In addition to the transformers used in power systems, in power transmission and distribution, a large number of special transformers are in use in applications like electronic supplies, rectication, furnaces, traction etc. Here the focus is on power transformers only. The principle of operation of these transformers also is the same but the user requirements dier. Power transformers of smaller sizes could be air cooled while the larger ones are oil cooled. These machines are highly material intensive equipments and are designed to match the applications for best operating conditions. Hence they are tailor made to a job. This brings in a very large variety in their constructional features. Here more common constructional aspects alone are discussed. These can be broadly divided into 1. Core construction 2. Winding arrangements 3. Cooling aspects
3.1
Core construction
Transformer core for the power frequency application is made of highly permeable
material. The high value of permeability helps to give a low reluctance for the path of the ux and the ux lines mostly conne themselves to the iron. Relative permeability r well over 1000 are achieved by the present day materials. Silicon steel in the form of thin laminations is used for the core material. Over the years progressively better magnetic properties are obtained by going in for Hot rolled non-oriented to Hot rolled grain oriented steel. 8
Later better laminations in the form of cold Rolled Grain Oriented (CRGO), -High B (HiB) grades became available. The thickness of the laminations progressively got reduced from over 0.5mm to the present 0.25mm per lamination. These laminations are coated with a thin layer of insulating varnish, oxide or phosphate. The magnetic material is required to have a high permeability and a high saturation ux density, a very low remanence Br and a small area under the B-H loop-to permit high ux density of operation with low magnetizing current and low hysteresis loss. The resistivity of the iron sheet itself is required to be high to reduce the eddy current losses. The eddy current itself is highly reduced by making the laminations very thin. If the lamination is made too thin then the production cost of steel laminations increases. The steel should not have residual mechanical stresses which reduce their magnetic properties and hence must be annealed after cutting and stacking. In the case of very small transformers (from a few volt-amperes to a few kilo voltamperes) hot rolled silicon steel laminations in the form of E & I, C & I or O as shown in Fig. 3 are used and the core cross section would be a square or a rectangle. The percentage of silicon in the steel is about 3.5. Above this value the steel becomes very brittle and also very hard to cut. The saturation ux density of the present day steel lamination is about 2 Tesla. Broadly classifying, the core construction can be separated into core type and shell type. In a core type construction the winding surrounds the core. A few examples of single phase and three phase core type constructions are shown in Fig. 4. In a shell type on the other hand the iron surrounds the winding. In the case of very small transformers the conductors are very thin and round. These can be easily wound on a former with rectangular or square cross section. Thus no special care is needed for the construction of the core. The cross section of the core also would be square or rectangular. As the rating of the transformer increases the conductor size 9
(a )
(b)
(c) Figure 3: E and I,C and I and O Type Laminations
10
1.phase 3.phase
HV LV core
Single phase
LV HV
Three phase
LV HV
(a)Core type
(b) Shell type
Figure 4: Core and Shell Type Construction also increases. Flat conductors are preferred to round ones. To wind such conductor on a rectangular former is not only dicult but introduces stresses in the conductor, at the bends. From the short circuit force with stand capability point of view also this is not desirable. Also, for a given area enclosed the length of the conductor becomes more. Hence it results in more load losses. In order to avoid all these problems the coils are made cylindrical and are wound on formers on heavy duty lathes. Thus the core construction is required to be such as to ll the circular space inside the coil with steel laminations. Stepped core construction thus becomes mandatory for the core of large transformers. Fig. 5 shows a few typical stepped core constructions. When the core size increases it becomes extremely dicult to cool the same (Even though the core losses are relatively very small). Cooling ducts have to be provided in the core. The steel laminations are grain oriented exploiting the simple geometry of the transformer to reduce the excitation losses. The iron losses in the lamination, when the ux
11
d
duct
duct
Figure 5: Stepped Core Construction
12
is oriented in the direction of grain orientation, is about 30% of that in the normal direction. Another important aspect to be carefully checked and monitored is the air gaps in
Path of flux HV
LV
Windings
Core
(a)
(b)
Figure 6: Typical stacked Core and wound core Construction series in the path of the main ux. As the reluctance of air path is about 1000 times more than that of the steel, an air path of 1mm will require a mmf needed by a 1 meter path in iron.
Hence butt joints between laminations must be avoided. Lap joints are used to provide alternate paths for ux lines thus reducing the reluctance of the ux paths. Some typical constructional details are shown in Fig. 6. In some power transformers the core is built up by threading a long strip of steel through the coil in the form of a toroid. This construction is normally followed in instrument transformers to reduce the magnetizing current and hence the errors.
13
Large cores made up of laminations must be rendered adequately sti by the provision of stiening plates usually called as itch plates. Punched through holes and bolts are progressively being avoided to reduce heating and melting of the through bolts. The whole stack is wrapped up by strong epoxy tapes to give mechanical strength to the core which can stand in upright position. Channels and angles are used for the frame and they hold the bottom yoke rigidly.
3.2
Windings
Windings form another important part of transformers. In a two winding trans-
former two windings would be present. The one which is connected to a voltage source and creates the ux is called as a primary winding. The second winding where the voltage is induced by induction is called a secondary. If the secondary voltage is less than that of the primary the transformer is called a step down transformer. If the secondary voltage is more then it is a step up transformer. A step down transformer can be made a step up transformer by making the low voltage winding its primary. Hence it may be more appropriate to designate the windings as High Voltage (HV) and Low Voltage (LV) windings. The winding with more number of turns will be a HV winding. The current on the HV side will be lower as V-I product is a constant and given as the VA rating of the machines. Also the HV winding needs to be insulated more to withstand the higher voltage across it. HV also needs more clearance to the core, yoke or the body. These aspects inuence the type of the winding used for the HV or LV windings. Transformer coils can be broadly classied in to concentric coils and sandwiched coils Fig. 7. The former are very common with core type transformers while the latter one 14
HV
LV
Core LV HV
(a)Concentric coil
LV
HV
Core
(b) Sandwich coil
Figure 7: Concentric and Sandwich Coils
15
are common with shell type transformers. In the gure the letters L and H indicate the low voltage and high voltage windings. In concentric arrangement, in view of the lower insulation and clearance requirements, the LV winding is placed close to the core which is at ground potential. The HV winding is placed around the LV winding. Also taps are provided on HV winding when voltage change is required. This is also facilitated by having the HV winding as the outer winding. Three most common types of coils viz. helical, cross over and disc coils are shown in Fig. 8.
Helical coils
Disc coils
cross over coils
Figure 8: Disc, Crossover and Helical Coil Construction
Helical Windings One very common cylindrical coil arrangement is the helical winding. This is made up of large cross section rectangular conductor wound on its at side. The coil progresses as a helix. This is commonly used for LV windings. The insulation 16
requirement also is not too high. Between layers no insulation (other than conductor insulation) is needed as the voltage between layers is low. The complexity of this type of winding rapidly increases as the current to be handled becomes more. The conductor cross section becomes too large and dicult to handle. The eddy current losses in the conductor rapidly increases. Hence two or more conductors have to be wound and connected in parallel. The parallel circuits bring in problems of current sharing between the circuits. Transpositions of the parallel paths have to be adopted to reduce unequal current distribution. The modern practice is to use continuously transposed and bunched conductors. Cross over coils The second popular winding type is the cross over coil. These are made of circular conductors not exceeding 5 to 6 sq mm in cross section. These are used for HV windings of relatively small transformers. These turns are wound in several layers. The length and thickness of each block is made in line with cooling requirements. A number of such blocks can be connected in series, leaving cooling ducts in between the blocks, as required by total voltage requirement. Disc coils Disc coils consist of at conductors wound in a spiral form at the same place spiralling outwards. Alternate discs are made to spiral from outside towards the center. Sectional discs or continuous discs may be used. These have excellent thermal properties and the behavior of the winding is highly predictable. Winding of a continuous disc winding needs specialized skills. Sandwich coils Sandwich windings are more common with shell type core construction. They permit easy control over the short circuit impedance of the transformer. By bringing HV and LV coils close on the same magnetic axis the leakage is reduced and the mutual ux is increased. By increasing the number of sandwiched coils the 17
reactance can be substantially reduced.
3.3
Insulation
The insulation used in the case of electrical conductors in a transformer is varnish
or enamel in dry type of transformers. In larger transformers to improve the heat transfer characteristics the conductors are insulated using un-impregnated paper or cloth and the whole core-winding assembly is immersed in a tank containing transformer oil. The transformer oil thus has dual role. It is an insulator and also a coolant. The porous insulation around the conductor helps the oil to reach the conductor surface and extract the heat. The conductor insulation may be called the minor insulation as the voltage required to be withstood is not high. The major insulation is between the windings. Annular bakelite cylinders serve this purpose. Oil ducts are also used as part of insulation between windings. The oil used in the transformer tank should be free from moisture or other contamination to be of any use as an insulator.
3.4
Cooling of transformers
Scaling advantages make the design of larger and larger unit sizes of transformers
economically attractive. This can be explained as below. Consider a transformer of certain rating designed with certain ux density and current density. If now the linear dimensions are made larger by a factor of K keeping the current and ux densities the same the core and conductor areas increase by a factor of K 2 . The losses in the machine, which are proportional to the volume of the materials used, increase by a factor of K 3 .The rating of the machine increases by a factor of K 4 .
18
The surface area however increases by a factor of K 2 only. Thus the ratio of loss per surface area goes on increasing by a factor of K . The substantial increase in the output is the major attraction in going in for larger units. However cooling of the transformer becomes more and more dicult. As the rating increases better cooling techniques are needed.
Simple air cooling of the transformers is adopted in dry type transformers. The limit for this is reached by the time the rating is a few kVA. Hence air cooling is used in low voltage machines. This method of cooling is termed as AN(Air Natural). Air Blast(AB) method improves on the above by directing the blast of air at the core and windings. This permits some improvement in the unit sizes.
Substantial improvement is obtained when the transformer is immersed in an oil tank. The oil reaches the conductor surface and extracts the heat and transports the same to the surface of the tank by convection. This is termed as ON (Oil Natural) type of cooling. This method permits the increase in the surface available for the cooling further by the use of ducts, radiators etc.
OB(Oil Blast) method is an improvement over the ON-type and it directs a blast of air on the cooling surface. In the above two cases the ow of oil is by natural convective forces. The rate of circulation of oil can be increased with the help of a pump, with the cooling at the surface remaining natural cooling to air. This is termed as OFN (Oil Forced Natural). If now a forced blast of air is also employed, the cooling method become OFB( Oil Forced Blast). A forced circulation of oil through a radiator is done with a blast of air 19
Main tank
Radiator
Tubes
(a)
Bushing
Conservator & Breather

water outlet
oil pump
Radiator
water inlet
(b)
Conservator& Breather
Bushing
Radiator
Oil pump for O.F.B
Fan motor
(c) 20 Figure 9: Some Typical Cooling Arrangements
over the radiator surface. Substantial amount of heat can be removed by employing a water cooling. Here the hot oil going into the radiator is cooled by a water circuit. Due to the high specic heat of water, heat can be evacuated eectively. Next in hierarchy comes OFW which is similar to OFB except that instead of blast of air a forced circulation of cool water in the radiator is used in this. Some cooling arrangements are shown in Fig. 9.
In many large sized transformers the cooling method is matched with the amount of heat that is required to be removed. As the load on the transformer changes the heat generated within also changes. Suitable cooling method can be pressed into service at that time. This gives rise to the concept of mixed cooling technique. ON/OB Works as ON but with increased load additional air blast is adopted. This gives the ratings to be in the ratio of 1:1.5 ON/OB/OFB Similarly gives the ratings in the ratio of 1:1.5:2
3.4.1
Properties of the transformer coil Even though the basic functions of the oil used in transformers are a) heat conduc-
tion and b) electrical insulation, there are many other properties which make a particular oil eminently suitable. Organic oils of vegetative or animal origin are good insulators but tend to decompose giving rise to acidic by-products which attack the paper or cloth insulation around the conductors.
Mineral oils are suitable from the point of electrical properties but tend to form sludge. The properties that are required to be looked into before selecting an oil for transformer 21
application are as follows: Insulting property This is a very important property. However most of the oils naturally full this. Therefore deterioration in insulating property due to moisture or contamination may be more relevant. Viscosity It is important as it determines the rate of ow of the uid. Highly viscous uids need much bigger clearances for adequate heat removal. Purity The oil must not contain impurities which are corrosive. Sulphur or its compounds as impurities cause formation of sludge and also attack metal parts. Sludge formation Thickening of oil into a semisolid form is called a sludge. Sludge formation properties have to be considered while choosing the oil as the oil slowly forms semi-solid hydrocarbons. These impede ows and due to the acidic nature, corrode metal parts. Heat in the presence of oxygen is seen to accelerate sludge formation. If the hot oil is prevented from coming into contact with atmospheric air sludge formation can be greatly reduced. Acidity Oxidized oil normally produces CO2 and acids. The cellulose which is in the paper insulation contains good amount of moisture. These form corrosive vapors. A good breather can reduce the problems due to the formation of acids. Flash point And Fire point Flash point of an oil is the temperature at which the oil ignites spontaneously. This must be as high as possible (not less than 160 C from the point of safety). Fire point is the temperature at which the oil ashes and continuously burns. This must be very high for the chosen oil (not less than 200 C ).
22
Inhibited oils and synthetic oils are therefore used in the transformers. Inhibited oils contain additives which slow down the deterioration of properties under heat and moisture and hence the degradation of oil. Synthetic transformer oil like chlorinated diphenyl has excellent properties like chemical stability, non-oxidizing, good dielectric strength, moisture repellant, reduced risk due re and explosion.
It is therefore necessary to check the quality of the oil periodically and take corrective steps to avoid major break downs in the transformer.
There are several other structural and insulating parts in a large transformer. These are considered to be outside the scope here.
Ideal Transformer
Earlier it is seen that a voltage is induced in a coil when the ux linkage associated
with the same changed. If one can generate a time varying magnetic eld any coil placed in the eld of inuence linking the same experiences an induced emf. A time varying eld can be created by passing an alternating current through an electric coil. This is called mutual induction. The medium can even be air. Such an arrangement is called air cored transformer. Indeed such arrangements are used in very high frequency transformers. Even though the principle of transformer action is not changed, the medium has considerable inuence on the working of such devices. These eects can be summarized as the followings. 1. The magnetizing current required to establish the eld is very large, as the reluctance
23
of the medium is very high. 2. There is linear relationship between the mmf created and the ux produced. 3. The medium is non-lossy and hence no power is wasted in the medium. 4. Substantial amount of leakage ux exists. 5. It is very hard to direct the ux lines as we desire, as the whole medium is homogeneous. If the secondary is not loaded the energy stored in the magnetic eld nds its way back to the source as the ux collapses. If the secondary winding is connected to a load then part of the power from the source is delivered to the load through the magnetic eld as a link. The medium does not absorb and lose any energy. Power is required to create the eld and not to maintain the same. As the winding losses can be made very small by proper choice of material, the ideal eciency of a transformer approaches 100%. The large magnetizing current requirement is a major deterrent. However if now a piece of magnetic material is introduced to form the magnetic circuit Fig. 10(b) the situation changes dramatically. These can be enumerated as below. 1. Due to the large value for the permeance ( r of the order of 1000 as compared to air) the magnetizing current requirement decreases dramatically. This can also be visualized as a dramatic increase in the ux produced for a given value of magnetizing current. 2. The magnetic medium is linear for low values of induction and exhibits saturation type of non-linearity at higher ux densities. 3. The iron also has hysteresis type of non-linearity due to which certain amount of power is lost in the iron (in the form of hysteresis loss), as the B-H characteristic is traversed. 24
Primary
x
Leakage flux
Secondary Mutual flux
(a)
Leakage flux
X
Primary Mutual flux
Secondary Iron core
(b)
Figure 10: Mutual Induction a) air core b) iron core
25
4. Most of the ux lines are conned to iron path and hence the mutual ux is increased very much and leakage ux is greatly reduced. 5. The ux can be easily directed as it takes the path through steel which gives great freedom for the designer in physical arrangement of the excitation and output windings. 6. As the medium is made of a conducting material eddy currents are induced in the same and produce losses. These are called eddy current losses. To minimize the eddy current losses the steel core is required to be in the form of a stack of insulated laminations. From the above it is seen that the introduction of magnetic core to carry the ux introduced two more losses. Fortunately the losses due to hysteresis and eddy current for the available grades of steel is very small at power frequencies. Also the copper losses in the winding due to magnetization current is reduced to an almost insignicant fraction of the full load losses. Hence steel core is used in power transformers.
In order to have better understanding of the behavior of the transformer, initially certain idealizations are made and the resulting ideal transformer is studied. These idealizations are as follows: 1. Magnetic circuit is linear and has innite permeability. The consequence is that a vanishingly small current is enough to establish the given ux. Hysteresis loss is negligible. As all the ux generated connes itself to the iron, there is no leakage ux. 2. Windings do not have resistance. This means that there are no copper losses, nor there is any ohmic drop in the electric circuit. 26
v1=V1mcost
i o
0
T1
+ e1 + e2 -
i1 + + e2 -
i2
T2
v1=V1sint e1 +
(a)Unloaded machine form

v1=V1cost i1 N + i2 ZL e2 + e1 + T1
(b) Circuit
T2
(c)Loaded machine
Figure 11: Two winding Ideal Transformer unloaded and loaded
27
In fact the practical transformers are very close to this model and hence no major departure is made in making these assumptions. Fig. 11 shows a two winding ideal transformer. The primary winding has T1 turns and is connected to a voltage source of V1 volts. The secondary has T2 turns. Secondary can be connected to a load impedance for loading the transformer. The primary and secondary are shown on the same limb and separately for clarity.
As a current I0 amps is passed through the primary winding of T1 turns it sets up an mmf of I0 T1 ampere which is in turn sets up a ux through the core. Since the reluctance of the iron path given by R = l/Ais zero as , a vanishingly small value of current I0 is enough to setup a ux which is nite. As I0 establishes the eld inside the transformer it is called the magnetizing current of the transformer. F lux = mmf I 0 T1 I0 T1 A = l = . Reluctance l A (9)
This current is the result of a sinusoidal voltage V applied to the primary. As the current through the loop is zero (or vanishingly small), at every instant of time, the sum of the voltages must be zero inside the same. Writing this in terms of instantaneous values we have, v1 e 1 = 0 (10)
where v1 is the instantaneous value of the applied voltage and e1 is the induced emf due to Faradays principle. The negative sign is due to the application of the Lenzs law and shows that it is in the form of a voltage drop. Kirchos law application to the loop will result in the same thing.
28
This equation results in v1 = e1 or the induced emf must be same in magnitude to the applied voltage at every instant of time. Let v1 = V1peak cos t where V1peak is the peak value and = 2f t. f is the frequency of the supply. As v1 = e1 ; e1 = d1 /dt but e1 = E1peak cos t E1 = V1 . It can be easily seen that the variation of ux linkages can be obtained as 1 = 1peak sin t. Here 1peak is the peak value of the ux linkages of the primary. Thus the RMS primary induced emf is e1 = d1 d(1peak sin t) = dt dt 1peak . 2f T1 m = = 4.44f m T1 2 2 (11) (12) volts
= 1peak .. cos t or the rms value E1 =
Here 1peak is the peak value of the ux linkages of the primary. The same mutual ux links the secondary winding. However the magnitude of the ux linkages will be 2peak = T2 .m . The induced emf in the secondary can be similarly obtained as , e2 = d2 d(2peak sin t) = dt dt = 2peak .. cos t or the rms value 2f T2 m = 4.44f m T2 2 E1 T1 = E2 T2 volt (13) (14)
E2 = which yields the voltage ratio as
(15)
The voltages E1 and E2 are obtained by the same mutual ux and hence they are in phase. If the winding sense is opposite i.e., if the primary is wound in clockwise sense and the secondary counter clockwise sense then if the top terminal of the rst winding is 29
I1 + V1 + E1 E2 -
I2
V2
Figure 12: Dot Convention
at maximum potential the bottom terminal of the second winding would be at the peak potential. Similar problem arises even when the sense of winding is kept the same, but the two windings are on opposite limbs (due to the change in the direction of ux). Hence in the circuit representation of transformers a dot convention is adopted to indicate the terminals of the windings that go high (or low) together. (Fig. 12). This can be established experimentally by means of a polarity test on the transformers. At a particular instant of time if the current enters the terminal marked with a dot it magnetizes the core. Similarly a current leaving the terminal with a dot demagnetizes the core.
So far, an unloaded ideal transformer is considered. If now a load impedance ZL is connected across the terminals of the secondary winding a load current ows as marked in Fig. 11(c). This load current produces a demagnetizing mmf and the ux tends to collapse. However this is detected by the primary immediately as both E2 and E1 tend to collapse. The current drawn from supply increases up to a point the ux in the core is restored back to its original value. The demagnetizing mmf produced by the secondary is neutralized by
30
additional magnetizing mmf produces by the primary leaving the mmf and ux in the core as in the case of no-load. Thus the transformer operates under constant induced emf mode. Thus, i 1 T1 i 2 T2 = i 0 T1 i 2 T2 = i 1 T1 but i0 0 and the rms value I2 T2 = I1 T1 . (16) (17)
If the reference directions for the two currents are chosen as in the Fig. 12, then the above equation can be written in phasor form as, I 1 T1 = I 2 T2 Also E1 E2 T2 or I .I 2 1 = T1 T1 I2 = = E1 I1 = E 2 I2 T2 I1 (18) (19)
Thus voltage and current transformation ratio are inverse of one another. If an impedance of ZL is connected across the secondary, 2 E I 2 = ZL L = E2 or Z I 2 (20)
The input impedance under such conditions is i = E1 = ( T1 )2 . E2 = ( T1 )2 .Z L Z T2 I T2 I 1 2
(21)
1 An impedance of ZL when viewed through a transformer of turns ratio ( T ) is seen T2
T1 2 as ( T ) .ZL . Transformer thus acts as an impedance converter. The transformer can be 2
interposed in between a source and a load to match the impedance.
Finally, the phasor diagram for the operation of the ideal transformer is shown in Fig. 13 in which 1 and 2 are power factor angles on the primary and secondary sides. As 31
V1
E1 I2 E2 I1 1 2 V2
Figure 13: Phasor diagram of Operation of an Ideal Transformer
the transformer itself does not absorb any active or reactive power it is easy to see that 1 = 2 .
Thus, from the study of the ideal transformer it is seen that the transformer provides electrical isolation between two coupled electric circuits while maintaining power invariance at its two ends. This can be used to step up or step down the voltage /current at constant volt-ampere. Also, the transformer can be used for impedance matching. In the case of an ideal transformer the eciency is 100% as there are no losses inside the device.
32
Practical Transformer
An ideal transformer is useful in understanding the working of a transformer. But it
cannot be used for the computation of the performance of a practical transformer due to the non-ideal nature of the practical transformer. In a working transformer the performance aspects like magnetizing current, losses, voltage regulation, eciency etc are important. Hence the eects of the non-idealization like nite permeability, saturation, hysteresis and winding resistances have to be added to an ideal transformer to make it a practical transformer. Conversely, if these eects are removed from a working transformer what is left behind is an ideal transformer.
Finite permeability of the magnetic circuit necessitates a nite value of the current to be drawn from the mains to produce the mmf required to establish the necessary ux. The current and mmf required is proportional to the ux density B that is required to be established in the core. B = H ; B= A (22)
where A is the area of cross section of the iron core m2 . H is the magnetizing force which is given by, H = i. T1 l (23)
where l is the length of the magnetic path, m. or = B.A = A(iT1 ) l = permeance mmf (here that of primary ) (24)
The magnetizing force and the current vary linearly with the applied voltage as long as the magnetic circuit is not saturated. Once saturation sets in, the current has to vary in 33
a nonlinear manner to establish the ux of sinusoidal shape. This non-linear current can be resolved into fundamental and harmonic currents. This is discussed to some extent under harmonics. At present the eect of this non-linear behavior is neglected as a secondary eect. Hence the current drawn from the mains is assumed to be purely sinusoidal and directly proportional to the ux density of operation. This current can be represented by a current drawn by an inductive reactance in the circuit as the net energy associated with the same over a cycle is zero. The energy absorbed when the current increases is returned to the electric circuit when the current collapses to zero. This current is called the magnetizing current of the transformer. The magnetizing current Im is given by Im = E1 /Xm where Xm is called the magnetizing reactance. The magnetic circuit being lossy absorbs and dissipates the power depending upon the ux density of operation. These losses arise out of hysteresis, eddy current inside the magnetic core. These are given by the following expressions: Ph B 1.6 f Pe B 2 f 2 t2 Ph -Hysteresis loss, Watts B - Flux density of operation Tesla. f - Frequency of operation, Hz t - Thickness of the laminations of the core, m. (25) (26)
For a constant voltage, constant frequency operation B is constant and so are these losses. An active power consumption by the no-load current can be represented in the input circuit as a resistance Rc connected in parallel to the magnetizing reactance Xm . Thus the no-load current I0 may be made up of Ic (loss component) and Im (magnetizing component
34
as ) I 0 = Ic j Im
2 Ic Rc gives the total core losses (i.e. hysteresis + eddy current loss) 2 Xm - Reactive volt amperes consumed for establishing the mutual ux. Im
(27)
Finite of the magnetic core makes a few lines of ux take to a path through the air. Thus these ux lines do not link the secondary winding. It is called as leakage ux. As the path of the leakage ux is mainly through the air the ux produced varies linearly with the primary current I1 . Even a large value of the current produces a small value of ux. This ux produces a voltage drop opposing its cause, which is the current I1 . Thus this eect of the nite permeability of the magnetic core can be represented as a series inductive element jxl1 . This is termed as the reactance due to the primary leakage ux. As this leakage ux varies linearly with I1 , the ux linkages per ampere and the primary leakage inductance are constant (This is normally represented by ll1 Henry). The primary leakage reactance therefore becomes xl1 = 2f ll1 ohm (28)
A similar eect takes place on the secondary side when the transformer is loaded. The secondary leakage reactance jxl2 arising out of the secondary leakage inductance ll2 is given by
xl2 = 2f ll2
(29)
Finally, the primary and secondary windings are wound with copper (sometimes aluminium in small transformers) conductors; thus the windings have a nite resistance (though 35
I1 V1
+
r1
jxl1 Io Rc
I2
+
E1
-
~
-
jXm
T1
r2
jxl2
I2
+
V2
ZL
E2
-
T2
I1
r1
(a)Physical arrangement
jXl1
I2
V1
Rc
Ic
Im
Io
E1
r2 jXl2 I2
jXm
E2
ZL V2
(b)Equivalent circuit
Figure 14: A Practical Transformer
36
small). This is represented as a series circuit element, as the power lost and the drop produced in the primary and secondary are proportional to the respective currents. These are represented by r1 and r2 respectively on primary and secondary side. A practical transformer sans these imperfections (taken out and represented explicitly in the electric circuits) is an ideal transformer of turns ratio T1 : T2 (voltage ratio E1 : E2 ). This is seen in Fig. 14. I2 in the circuit represents the primary current component that is required to ow from the mains in the primary T1 turns to neutralize the demagnetizing secondary current I2 due to the load in the secondary turns. The total primary current vectorially is I 1 = I2 + I0 Here I2 T1 = I2 T2 or I2 = I2 T2 T1 (30) (31) (32)
T2 + I Thus I 1 = I2 0 T1
By solving this circuit for any load impedance ZL one can nd out the performance of the loaded transformer.
The circuit shown in Fig. 14(b). However, it is not very convenient for use due to the presence of the ideal transformer of turns ratio T1 : T2 . If the turns ratio could be made unity by some transformation the circuit becomes very simple to use. This is done here by replacing the secondary by a hypothetical secondary having T1 turns which is equivalent to the physical secondary. The equivalence implies that the ampere turns, active and reactive power associated with both the circuits must be the same. Then there is no change as far as their eect on the primary is considered. Thus V2 = aV2 , where a -turns ratio
T1 T2
I2 =
I2 , a
r2 = a 2 r2 ,
xl 2 = a 2 xl 2
ZL = a 2 ZL .
37
This equivalent circuit is as shown in Fig. 16(a). As the ideal transformer in this case has a turns ratio of unity the potentials on either side are the same and hence they may be conductively connected dispensing away with the ideal transformer. This particular equivalent circuit is as seen from the primary side. It is also possible to refer all the primary parameters to secondary by making the hypothetical equivalent primary winding on the input side having the number of turns to be T2 . Such an equivalent circuit having all the parameters referred to the secondary side is shown in g. 15.
The equivalent circuit can be derived, with equal ease, analytically using the Kirchos equations applied to the primary and secondary. Referring to g. 14(a), we have (by neglecting the shunt branch) V1 = E1 + I1 (r1 + jxl1 ) E2 = V2 + I2 (r2 + jxl2 ) T1 I 0 = T 1 I 1 + T 2 I 2 = a = or I1 = I2 + I0 a (33) (34) (35)
I2 + Ic + Im a
T1 . T2
Multiply both sides of Eqn.34 by a [This makes the turns ratio unity and retains the power invariance]. aE2 = aV2 + aI2 (r2 + jxl2 ) Substituting in Eqn.33 we have but aE2 = E1 (36)
38
V1 = aV2 + aI2 (r2 + jxl2 ) + I1 (r1 + jxl1 ) = V2 + I1 (a2 r2 + ja2 xl2 ) + I1 (r1 + jxl1 ) = V2 + I1 (r1 + r2 + jxl1 + xl2 ) (37)
A similar procedure can be used to refer all parameters to secondary side. (Shown in g. 15.)
r1
jxl1 Ic
I1
Io
Im
r2
jxl2
I2
V1
Rc
jXm
ZL
V2
Figure 15: Equivalent Circuit Referred to the Secondary Side
Phasor diagrams
The resulting equivalent circuit as shown in Fig. 16 is known as the exact equivalent
circuit. This circuit can be used for the analysis of the behavior of the transformers. As the no-load current is less than 1% of the load current a simplied circuit known as approximate equivalent circuit (see Fig. 16(b)) is usually used, which may be further simplied to the one 39
I1
r1
jxl1
r2
jxl2
Ic
V1
Rc
Io
Im
jXm
V2
ZL
I1 Ic V1 Rc
I2
r1 Im
jxm
(a)
jxl1
r2
jxl2
Io
R=r1+r2
I1
jX
I2
ZL
V2 x=xl1+xl2
I1=I2
V1
V2
(b)
(c)
Figure 16: Exact,approximate and simplied equivalent circuits
40
shown in Fig. 16(c).
On similar lines to the ideal transformer the phasor diagram of operation can be drawn for a practical transformer also. The positions of the current and induced emf phasor are not known uniquely if we start from the phasor V1 . Hence it is assumed that the phasor is known. The E1 and E2 phasor are then uniquely known. Now, the magnetizing and loss components of the currents can be easily represented. Once I0 is known, the drop that takes place in the primary resistance and series reactance can be obtained which when added to E1 gives uniquely the position of V1 which satises all other parameters. This is represented in Fig. 17(a) as phasor diagram on no-load.
Next we proceed to draw the phasor diagram corresponding to a loaded transformer. The position of the E2 vector is known from the ux phasor. Magnitude of I2 and the load power factor angle 2 are assumed to be known. But the angle 2 is dened with respect to the terminal voltage V2 and not E2 . By trial and error the position of I2 and V2 are determined. V2 should also satisfy the Kirchos equation for the secondary. Rest of the construction of the phasor diagram then becomes routine. The equivalent primary current I2 is added vectorially to I0 to yield I1 . I1 (r1 + jxl1 )is added to E1 to yield V1 . This is shown in g. 17(b) as phasor diagram for a loaded transformer.
Testing of Transformers
The structure of the circuit equivalent of a practical transformer is developed earlier.
The performance parameters of interest can be obtained by solving that circuit for any load 41
V1
IoX l1 Ior1 E2
E1
Io Im
V1
Il
(a)No-load
I1X l1 I1r1 E2 Il Io
(b)On-load Figure 17: Phasor Diagram of a Practical Transformer
E1 I2
I2x2 I r 2 2
V2
I2
42
conditions. The equivalent circuit parameters are available to the designer of the transformers from the various expressions that he uses for designing the transformers. But for a user these are not available most of the times. Also when a transformer is rewound with dierent primary and secondary windings the equivalent circuit also changes. In order to get the equivalent circuit parameters test methods are heavily depended upon. From the analysis of the equivalent circuit one can determine the electrical parameters. But if the temperature rise of the transformer is required, then test method is the most dependable one. There are several tests that can be done on the transformer; however a few common ones are discussed here.
7.1
Winding resistance test

This is nothing but the resistance measurement of the windings by applying a small
d.c voltage to the winding and measuring the current through the same. The ratio gives the winding resistance, more commonly feasible with high voltage windings. For low voltage windings a resistance-bridge method can be used. From the d.c resistance one can get the a.c. resistance by applying skin eect corrections.
7.2
Polarity Test
This is needed for identifying the primary and secondary phasor polarities. It is
a must for poly phase connections. Both a.c. and d.c methods can be used for detecting the polarities of the induced emfs. The dot method discussed earlier is used to indicate the polarities. The transformer is connected to a low voltage a.c. source with the connections made as shown in the g. 18(a). A supply voltage Vs is applied to the primary and the
43
V3 Vs
A2
S
V1
a2
V2
A1 a1
+ -
A2
a2
+
A1
a1
(b)D.C.test
(a)A.C.test Figure 18: Polarity Test
readings of the voltmeters V1 , V2 and V3 are noted. V1 : V2 gives the turns ratio. If V3 reads V1 V2 then assumed dot locations are correct (for the connection shown). The beginning and end of the primary and secondary may then be marked by A1 A2 and a1 a2 respectively. If the voltage rises from A1 to A2 in the primary, at any instant it does so from a1 to a2 in the secondary. If more secondary terminals are present due to taps taken from the windings they can be labeled as a3 , a4 , a5 , a6 . It is the voltage rising from smaller number towards larger ones in each winding. The same thing holds good if more secondaries are present. Fig. 18(b) shows the d.c. method of testing the polarity. When the switch S is closed if the secondary voltage shows a positive reading, with a moving coil meter, the assumed polarity is correct. If the meter kicks back the assumed polarity is wrong.
44
W A V1 V V2
V1
Io
Im jXm
Ic Rc
(a)Physical Arrangement (b)Equivalent Circuit Figure 19: No Load Test
7.3
Open Circuit Test

As the name suggests, the secondary is kept open circuited and nominal value of
the input voltage is applied to the primary winding and the input current and power are measured. In Fig. 19(a) V, A, W are the voltmeter, ammeter and wattmeter respectively. Let these meters read V1 , I0 and W0 respectively.Fig. 19(b) shows the equivalent circuit of the transformer under this test. The no load current at rated voltage is less than 1 percent of nominal current and hence the loss and drop that take place in primary impedance r1 + jxl1 due to the no load current I0 is negligible. The active component Ic of the no load current I0 represents the core losses and reactive current Im is the current needed for the magnetization.
45
Thus the wattmeter reading W0 = V1 Ic = Pcore Ic = Im = Rc = W0 V1 V1 Ic

2 2 I0 Ic
(38) (39)
or V1 Im
(40) (41)
andXm =
V1
Io
Figure 20: Open Circuit Characteristics
The parameters measured already are in terms of the primary. Sometimes the primary voltage required may be in kilo-Volts and it may not be feasible to apply nominal voltage to primary from the point of safety to personnel and equipment. If the secondary voltage is low, one can perform the test with LV side energized keeping the HV side open circuited. In this case the parameters that are obtained are in terms of LV . These have to be referred to HV side if we need the equivalent circuit referred to HV side.
46
Sometimes the nominal value of high voltage itself may not be known, or in doubt, especially in a rewound transformer. In such cases an open circuit characteristics is rst obtained, which is a graph showing the applied voltage as a function of the no load current. This is a non linear curve as shown in Fig. 20. This graph is obtained by noting the current drawn by transformer at dierent applied voltage, keeping the secondary open circuited. The usual operating point selected for operation lies at some standard voltage around the knee point of the characteristic. After this value is chosen as the nominal value the parameters are calculated as mentioned above.
7.4
Short Circuit Test

The purpose of this test is to determine the series branch parameters of the equiv-
alent circuit of Fig. 21(b). As the name suggests, in this test primary applied voltage, the current and power input are measured keeping the secondary terminals short circuited. Let these values be Vsc , Isc and Wsc respectively. The supply voltage required to circulate rated current through the transformer is usually very small and is of the order of a few percent of the nominal voltage. The excitation current which is only 1 percent or less even at rated voltage becomes negligibly small during this test and hence is neglected. The shunt branch is thus assumed to be absent. Also I1 = I2 as I0 0. Therefore Wsc is the sum of the
copper losses in primary and secondary put together. The reactive power consumed is that absorbed by the leakage reactance of the two windings.
2 Wsc = Isc (r1 + r2 )
(42) (43) (44)
Zsc = (xl1 + xl2 ) =
Vsc Isc
2 (r + r )2 Zsc 1 2
47
A Vsc V
(a)Physical Arrangement
Isc
Vsc
r1
jxl1
r2
jxl2
(b)Equivalent Circuit Figure 21: Short Circuit Test
48
If the approximate equivalent circuit is required then there is no need to separate r1 and r2 or xl1 and xl2 . However if the exact equivalent circuit is needed then either r1 or r2 is determined from the resistance measurement and the other separated from the total. As for the separation of xl1 and xl2 is concerned, they are assumed to be equal. This is a fairly valid assumption for many types of transformer windings as the leakage ux paths are through air and are similar.
7.5
Load Test
Load Test helps to determine the total loss that takes place, when the transformer
is loaded. Unlike the tests described previously, in the present case nominal voltage is applied across the primary and rated current is drown from the secondary. Load test is used mainly 1. to determine the rated load of the machine and the temperature rise 2. to determine the voltage regulation and eciency of the transformer. Rated load is determined by loading the transformer on a continuous basis and observing the steady state temperature rise. The losses that are generated inside the transformer on load appear as heat. This heats the transformer and the temperature of the transformer increases. The insulation of the transformer is the one to get aected by this rise in the temperature. Both paper and oil which are used for insulation in the transformer start getting degenerated and get decomposed. If the ash point of the oil is reached the transformer goes up in ames. Hence to have a reasonable life expectancy the loading of the transformer must be limited to that value which gives the maximum temperature rise tolerated by the insulation. This aspect of temperature rise cannot be guessed from the electrical equivalent circuit. Further, the losses like dielectric losses and stray load losses are not modeled in the 49
equivalent circuit and the actual loss under load condition will be in error to that extent. Many external means of removal of heat from the transformer in the form of dierent cooling methods give rise to dierent values for temperature rise of insulation. Hence these permit dierent levels of loading for the same transformer. Hence the only sure way of ascertaining the rating is by conducting a load test.
It is rather easy to load a transformer of small ratings. As the rating increases it becomes dicult to nd a load that can absorb the requisite power and a source to feed the necessary current. As the transformers come in varied transformation ratios, in many cases it becomes extremely dicult to get suitable load impedance.
Further, the temperature rise of the transformer is due to the losses that take place inside the transformer. The eciency of the transformer is above 99% even in modest sizes which means 1 percent of power handled by the transformer actually goes to heat up the machine. The remaining 99% of the power has to be dissipated in a load impedance external to the machine. This is very wasteful in terms of energy also. ( If the load is of unity power factor) Thus the actual loading of the transformer is seldom resorted to. Equivalent loss methods of loading and Phantom loading are commonly used in the case of transformers. The load is applied and held constant till the temperature rise of transformer reaches a steady value. If the nal steady temperature rise is lower than the maximum permissible value, then load can be increased else it is decreased. That load current which gives the maximum permissible temperature rise is declared as the nominal or rated load current and the volt amperes are computed using the same.
50
In the equivalent loss method a short circuit test is done on the transformer. The short circuit current is so chosen that the resulting loss taking place inside the transformer is equivalent to the sum of the iron losses, full load copper losses and assumed stray load losses. By this method even though one can pump in equivalent loss inside the transformer, the actual distribution of this loss vastly diers from that taking place in reality. Therefore this test comes close to a load test but does not replace one.
W A 2Io V1 V I 2
Io
Io
I2
I2
I 2
A Vs V
Figure 22: Back to Back Test - Phantom Loading
In Phantom loading method two identical transformers are needed. The windings are connected back to back as shown in Fig. 22. Suitable voltage is injected into the loop formed by the two secondaries such that full load current passes through them. An equiv51
alent current then passes through the primary also. The voltage source V1 supplies the magnetizing current and core losses for the two transformers. The second source supplies the load component of the current and losses due to the same. There is no power wasted in a load ( as a matter of fact there is no real load at all) and hence the name Phantom or virtual loading. The power absorbed by the second transformer which acts as a load is pushed back in to the mains. The two sources put together meet the core and copper losses of the two transformers. The transformers work with full ux drawing full load currents and hence are closest to the actual loading condition with a physical load.
Per Unit Calculations

As stated earlier, transformers of various sizes, ratings, voltage ratios can be seen
being used in a power system. The parameters of the equivalent circuits of these machines also vary over a large range. Also the comparison of these machines are made simple if all the parameters are normalized. If simple scaling of the parameters is done then one has to carry forward the scaling factors in the calculations. Expressing in percent basis is one example of scaling. However if the scaling is done on a logical basis one can have a simple representation of the parameters without the bother of the scaling factors. Also dierent units of measurement are in use in the dierent countries (FPS, CGS, MKS, etc;). These units also underwent several revisions over the years. If the transformer parameter can be freed from the units then the system becomes very simple. The per unit system is developed keeping these aspects in mind. The parameters of the transformer are referred to some base values and thus get scaled. In the case of power system a common base value is adopted in view of dierent ratings of the equipments used. In the case of individual equipments, its own nominal parameters are used as base values. Some base parameters can be chosen 52
as independent base values while some others become derived base parameters. Once the base values are identied the per unit values are calculated for any parameter by dividing the same by its base value. The units must be the same for both the parameters and their bases. Thus the per unit value is a unit-less dimensionless number. Let us choose nominal voltage and nominal current on the primary side of a transformer as the base values Vbase and Ibase . Other base values like volt ampere Sbase , short circuit impedance Zbase can be calculated from those values. Pbase , Qbase , Sbase = Vbase Ibase Rbase , Xbase , Zbase = Gbase , Bbase , Ybase Vbase Ibase Ibase = Vbase (45) (46) (47)
Normally Sbase and Vbase are known from name plate details. Other base values can be derived from them. Vp.u = Ip.u Zp.u V (volt) , Vbase (volt) I (Amps) I (amps) = = Sbase Ibase (amps) V
base
(48) (49)
Z (ohm) Ibase Sbase = = Z (ohm) = Z (ohm). 2 Zbase (ohm) Vbase Vbase
Many times, when more transformers are involved in a circuit one is required to choose a common base value for all of them. Parameters of all the machines are expressed on this common base. This is a common problem encountered in the case of parallel operation of two or more transformers. The conversion of the base values naturally lead to change in the per unit values of their parameters. An impedance Zp.u.old on the old base of Sbaseold and Vbaseold shall get modied on new base Sbasenew ,Vbasenew as Zp.u.new = (Zp.u.old .
2 Vbase Sbase old old
Sbase 2 Vbase
new new
(50)
53
The term inside the bracket is nothing but the ohmic value of the impedance and this gets converted into the new per unit value by the new Sbase and Vbase .
If all the equivalent circuit parameters are referred to the secondary side and per unit values of the new equivalent circuit parameters are computed with secondary voltage and current as the base values, there is no change in the per unit values. This can be easily seen by, Sbase 2 Vbase 1 .Zohm a2
Zp.u. = Zohm .
but Zohm =
(51)
Where a - is the turns ratio of primary to secondary Z - impedance as seen by primary, Z - impedance as seen by secondary. Sbase = Sbase - as the transformer rating is unaltered.
1 Vbase = Vbase . a
From the above relationships it can be seen that Zp.u. = Zp.u..
This becomes obvious if we realize that the mmf of the core for establishing a given ux is the same whether it is supplied through primary or the secondary. Also the active power and reactive power absorbed inside the transformer are not dependant on the winding connected to supply. This is further illustrated by taking the equivalent circuit of a transformer derived earlier and expressing the same in per unit form.
Thus the per unit values help in dispensing away the scaling constants. The veracity of the parameters can be readily checked. Comparison of the parameters of the machines with those of similar ones throw in useful information about the machines. Comparing the 54
eciencies of two transformers at any load one can say that the transformer with a higher p.u.resistance has higher copper losses without actually computing the same.
Application of per unit values for the calculation of voltage regulation, eciency and load sharing of parallel connected transformers will be discussed later at appropriate places.
Voltage Regulation
Modern power systems operate at some standard voltages. The equipments work-
ing on these systems are therefore given input voltages at these standard values, within certain agreed tolerance limits. In many applications this voltage itself may not be good enough for obtaining the best operating condition for the loads. A transformer is interposed in between the load and the supply terminals in such cases. There are additional drops inside the transformer due to the load currents. While input voltage is the responsibility of the supply provider, the voltage at the load is the one which the user has to worry about. If undue voltage drop is permitted to occur inside the transformer the load voltage becomes too low and aects its performance. It is therefore necessary to quantify the drop that takes place inside a transformer when certain load current, at any power factor, is drawn from its output leads. This drop is termed as the voltage regulation and is expressed as a ratio of the terminal voltage (the absolute value per se is not too important).
The voltage regulation can be dened in two ways - Regulation Down and Regulation up. These two denitions dier only in the reference voltage as can be seen below.
55
Regulation down: This is dened as the change in terminal voltage when a load current at any power factor is applied, expressed as a fraction of the no-load terminal voltage. Expressed in symbolic form we have, Regulation = |Vnl | |Vl | |Vnl | (52)
Vnl and Vl are no-load and load terminal voltages. This is the denition normally used in the case of the transformers, the no-load voltage being the one given by the power supply provider on which the user has no say. Hence no-load voltage is taken as the reference. Regulation up: Here again the regulation is expressed as the ratio of the change in the terminal voltage when a load at a given power factor is thrown o, and the on load voltage. This denition if expressed in symbolic form results in Regulation = Vnl is the no-load terminal voltage. Vl is load voltage. Normally full load regulation is of interest as the part load regulation is going to be lower. |Vnl | |Vl | |Vl | (53)
This denition is more commonly used in the case of alternators and power systems as the user-end voltage is guaranteed by the power supply provider. He has to generate proper no-load voltage at the generating station to provide the user the voltage he has asked for. In the expressions for the regulation, only the numerical dierences of the voltages are taken and not vector dierences.
56
(a) Equivalent Circuit

V1
A V2
B I2Xe
I2Re
C E
I2
(b)Phasor Diagram Figure 23: Regulation of Transformer
57
In the case of transformers both denitions result in more or less the same value for the regulation as the transformer impedance is very low and the power factor of operation is quite high. The power factor of the load is dened with respect to the terminal voltage on load. Hence a convenient starting point is the load voltage. Also the full load output voltage is taken from the name plate. Hence regulation up has some advantage when it comes to its application. Fig. 23 shows the phasor diagram of operation of the transformer under loaded condition. The no-load current I0 is neglected in view of the large magnitude of I2 . Then I1 = I 2 . V1 = I2 (Re + jXe ) + V2 OD = V1 = = - power factor angle, - internal impedance angle=tan1 Also, V1 = V2 + I2 .(Re + jXe ) = V2 + I2 (cos j sin )(Re + jXe ) RegulationR = (1 + v1 ) +
2 2 v2 2 Xe Re
(54)
[OA + AB + BC ]2 + [CD ]2 (55)
[V2 + I2 Re cos + I2 Xe sin ]2 + [I2 Xe cos I2 Re sin ]2
(56)
|V1 | |V2 | = |V2 |
2 (1 + v1 )2 + v2 1
(57)
2 2 + v1 ) v2 v2 2 (1 + v1 ) + +[ ] = (1 + v1 + )2 (58) 2(1 + v1 ) 2(1 + v1 ) 2(1 + v1 ) Taking the square root (59) 2 2(1 v2 .
(1 + v1
)2
2 v2
2 v2 = 1 + v1 + 2(1 + v1 )
(60)
where v1 = er cos + ex sin and v2 = ex cos er sin er =

I2 R e =per V2
unit resistance drop
58
ex =
I 2 Xe =per V2
unit reactance drop
as v1 and v2 are small.

2 v2 v2 1 v1 + 2 2(1 + e1 ) 2 (ex sin er cos )2 regulation R = er cos ex sin + 2
1 + v1 +
(61) (62)
2 v2 2(1 + v1 )
2 v2 (1 v1 ) . 2 2 (1 v1 )
2 v2 .(1 v1 ) 2
2 v2 2
(63)
Powers higher than 2 for v1 and v2 are negligible as v1 and v2 are already small. As v2 is small its second power may be neglected as a further approximation and the expression for the regulation of the transform boils down to regulation R = er cos ex sin
The negative sign is applicable when the power factor is leading. It can be seen from the above expression, the full load regulation becomes zero when the power factor is leading and er cos = ex sin or tan = er /ex or the power factor angle = tan1 (er /ex ) = tan1 (Re /Xe ) leading.
Similarly, the value of the regulation is maximum at a power factor angle = tan1 (ex /er ) = tan1 (Xe /Re ) lagging. An alternative expression for the regulation of a transformer can be derived by the method shown in Fig. 24. Here the phasor are resolved along the current axis and normal to it.
59
V1
V2
I2
I2Xe
I2Re
A B
Figure 24: An Alternate Method for the Calculation of Regulation
We have, OD 2 = (OA + AB )2 + (BC + CD )2 (64)
= (V2 cos + I2 Re )2 + (V2 sin + I2 Xe )2 (65) RegulationR = (V sin + I2 Xe ) (V2 cos + I2 Re ) + 2 1 V2 V2 =

2 (cos + Rp.u )2 + (sin + Xp.u )1 2 2
OD V2 OD = 1 V2 V2
(66) (67) (68)
Thus this expression may not be as convenient as the earlier one due to the square root involved.
Fig. 25 shows the variation of full load regulation of a typical transformer as the 60
power factor is varied from zero power factor leading, through unity power factor, to zero power factor lagging.
5 4 3 %Regulation 2 1 leading 0 0. 5 -1 -2 -3 -4 -5 1. 0 0. 5 0 lagging
power factor
Figure 25: Variation of Full Load Regulation with Power Factor
It is seen from Fig. 25 that the full load regulation at unity power factor is nothing but the percentage resistance of the transformer. It is therefore very small and negligible. Only with low power factor loads the drop in the series impedance of the transformer contributes substantially to the regulation. In small transformers the designer tends to keep the Xe very low (less than 5%) so that the regulation performance of the transformer is satisfactory.
61
A low value of the short circuit impedance /reactance results in a large short circuit current in case of a short circuit. This in turn results in large mechanical forces on the winding. So, in large transformers the short circuit impedance is made high to give better short circuit protection to the transformer which results in poorer regulation performance. In the case of transformers provided with taps on windings, so that the turns ratio can be changed, the voltage regulation is not a serious issue. In other cases care has to be exercised in the selection of the short circuit impedance as it aects the voltage regulation.
10
Eciency
Eciency of a power equipment is dened at any load as the ratio of the power
output to the power input. Putting in the form of an expression, Ef f iciency = output power Input power losses inside the machine = (69) input power Input power losses inside the machine = 1 = 1 ef f iciency inputpower output power = output + losses inside the machine
output input power power
More conveniently the eciency is expressed in percentage. % =
100
A typical curve for the variation of eciency as a function of output is given in Fig. 26 The losses that take place inside the machine expressed as a fraction of the input is some times termed as deciency. Except in the case of an ideal machine, a certain fraction of the input power gets lost inside the machine while handling the power. Thus the value for the eciency is always less than one. In the case of a.c. machines the rating is expressed 62
100
75
Efficiency%
50
25
0. 5
Figure 26: Eciency
in terms of apparent power. It is nothing but the product of the applied voltage and the current drawn. The actual power delivered is a function of the power factor at which this current is drawn. As the reactive power shuttles between the source and the load and has a zero average value over a cycle of the supply wave it does not have any direct eect on the eciency. The reactive power however increases the current handled by the machine and the losses resulting from it. Therefore the losses that take place inside a transformer at any given load play a vital role in determining the eciency. The losses taking place inside a transformer can be enumerated as below: 1. Primary copper loss 2. Secondary copper loss 3. Iron loss 63
4. Dielectric loss 5. Stray load loss These are explained in sequence below.
Primary and secondary copper losses take place in the respective winding resistances due to the ow of the current in them.
2 2 Pc = I 1 r1 + I 2 r2 = I22 Re
(70)
The primary and secondary resistances dier from their d.c. values due to skin eect and the temperature rise of the windings. While the average temperature rise can be approximately used, the skin eect is harder to get analytically. The short circuit test gives the value of R e taking into account the skin eect.
The iron losses contain two components - Hysteresis loss and Eddy current loss. The Hysteresis loss is a function of the material used for the core. Ph = Kh B 1.6 f For constant voltage and constant frequency operation this can be taken to be constant. The eddy current loss in the core arises because of the induced emf in the steel lamination sheets and the eddies of current formed due to it. This again produces a power loss Pe in the lamination. Pe = K e B 2 f 2 t2 where t is the thickness of the steel lamination used. As the lamination thickness is much smaller than the depth of penetration of the eld, the eddy current loss can be reduced by 64
reducing the thickness of the lamination. Present day laminations are of 0.25 mm thickness and are capable of operation at 2 Tesla. These reduce the eddy current losses in the core. This loss also remains constant due to constant voltage and frequency of operation. The sum of hysteresis and eddy current losses can be obtained by the open circuit test.
The dielectric losses take place in the insulation of the transformer due to the large electric stress. In the case of low voltage transformers this can be neglected. For constant voltage operation this can be assumed to be a constant.
The stray load losses arise out of the leakage uxes of the transformer. These leakage uxes link the metallic structural parts, tank etc. and produce eddy current losses in them. Thus they take place all round the transformer instead of a denite place , hence the name stray. Also the leakage ux is directly proportional to the load current unlike the mutual ux which is proportional to the applied voltage. Hence this loss is called stray load loss. This can also be estimated experimentally. It can be modeled by another resistance in the series branch in the equivalent circuit. The stray load losses are very low in air-cored transformers due to the absence of the metallic tank.
Thus, the dierent losses fall in to two categories Constant losses (mainly voltage dependant) and Variable losses (current dependant). The expression for the eciency of the transformer operating at a fractional load x of its rating, at a load power factor of 2 , can be written as = xS cos 2 xS cos 2 + Pconst + x2 Pvar (71)
Here S in the volt ampere rating of the transformer (V2 I2 at full load), Pconst being constant 65
losses and Pvar the variable losses at full load.
For a given power factor an expression for in terms of the variable x is thus obtained. By dierentiating with respect to x and equating the same to zero, the condition for maximum eciency is obtained. In the present case that condition comes out to be Pconst = x2 Pvar or x = Pconst Pvar (72)
That is, when constant losses equal the variable losses at any fractional load x the eciency reaches a maximum value. The maximum value of that eciency at any given power factor is given by, xS cos 2 xS cos 2 = xS cos 2 + 2Pconst xS cos 2 + 2x2 Pvar
max =
(73)
From the expression for the maximum eciency it can be easily deduced that this maximum value increases with increase in power factor and is zero at zero power factor of the load. It may be considered a good practice to select the operating load point to be at the maximum eciency point. Thus if a transformer is on full load, for most part of the time then the max can be made to occur at full load by proper selection of constant and variable losses. However, in the modern transformers the iron losses are so low that it is practically impossible to reduce the full load copper losses to that value. Such a design wastes lot of copper. This point is illustrated with the help of an example below.
Two 100 kVA transformers A nd B are taken. Both transformers have total full load losses to be 2 kW. The break up of this loss is chosen to be dierent for the two transformers. Transformer A: iron loss 1 kW, and copper loss is 1 kW. The maximum eciency of 98.04% 66
occurs at full load at unity power factor. Transformer B: Iron loss =0.3 kW and full load copper loss =1.7 kW. This also has a full load of 98.04%. Its maximum occurs at a fractional load of
42 42+0.6 0.3 1.7
= 0.42. The maximum eciency at unity power factor being
100 = 98.59%. At the corresponding point the transformer A has an eciency of 100 = 97.28%. Transformer A uses iron of more loss per kg at a given ux
42 42+1.0+0.1764
density, but transformer B uses lesser quantity of copper and works at higher current density.
10.1
100
All day eciency

100
50
Load % of full load
50
Power Loss %
s P
Time,hrs
12
18
24
12
24
(a)Load factor Figure 27: Calculation of Load Factor and Loss Factor
(b) Loss factor
Large capacity transformers used in power systems are classied broadly into Power transformers and Distribution transformers. The former variety is seen in generating stations and large substations. Distribution transformers are seen at the distribution substations. 67
The basic dierence between the two types arise from the fact that the power transformers are switched in or out of the circuit depending upon the load to be handled by them. Thus at 50% load on the station only 50% of the transformers need to be connected in the circuit. On the other hand a distribution transformer is never switched o. It has to remain in the circuit irrespective of the load connected. In such cases the constant loss of the transformer continues to be dissipated. Hence the concept of energy based eciency is dened for such transformers. It is called all day eciency. The all day eciency is thus the ratio of the energy output of the transformer over a day to the corresponding energy input. One day is taken as a duration of time over which the load pattern repeats itself. This assumption, however, is far from being true. The power output varies from zero to full load depending on the requirement of the user and the load losses vary as the square of the fractional loads. The no-load losses or constant losses occur throughout the 24 hours. Thus, the comparison of loads on dierent days becomes dicult. Even the load factor, which is given by the ratio of the average load to rated load, does not give satisfactory results. The calculation of the all day eciency is illustrated below with an example. The graph of load on the transformer, expressed as a fraction of the full load is plotted against time in Fig. 27. In an actual situation the load on the transformer continuously changes. This has been presented by a stepped curve for convenience. The average load can be calculated by Average load over a day =
n i=1
Pi
24
Sn
n i=1
xi ti cos i 24
(74)
where Pi is the load during an interval i. n intervals are assumed. xi is the fractional load. Si = xi Sn where Sn is nominal load. The average loss during the day is given by Pc
n i=1
Average loss = Pi +
x2 i ti
24
(75)
This is a non-linear function. For the same load factor dierent average loss can be 68
there depending upon the values of xi and ti . Hence a better option would be to keep the constant losses very low to keep the all day eciency high. Variable losses are related to load and are associated with revenue earned. The constant losses on the other hand has to be incurred to make the service available. The concept of all day eciency may therefore be more useful for comparing two transformers subjected to the same load cycle.
The concept of minimizing the lost energy comes into eect right from the time of procurement of the transformer. The constant losses and variable losses are capitalized and added to the material cost of the transformer in order to select the most competitive one, which gives minimum cost taking initial cost and running cost put together. Obviously the iron losses are capitalized more in the process to give an eect to the maximization of energy eciency. If the load cycle is known at this stage, it can also be incorporated in computation of the best transformer.
11
Auto Transformer
The primary and secondary windings of a two winding transformer have induced
emf in them due to a common mutual ux and hence are in phase. The currents drawn by these two windings are out of phase by 180 . This prompted the use of a part of the primary as secondary. This is equivalent to fusing the secondary turns into primary turns. The fused section need to have a cross sectional area of the conductor to carry (I2 I1 ) ampere! This ingenious thought led to the invention of an auto transformer. Fig. 28 shows the physical arrangement of an auto transformer. Total number of turns between A and C are T1 . At
69
I1
C
T1 V1
I2
T2 I2
A
V2
ZL
I1
Figure 28: Autotransformer - Physical Arrangement
point B a connection is taken. Section AB has T2 turns. As the volts per turn, which is proportional to the ux in the machine, is the same for the whole winding, V1 : V2 = T 1 : T2 (76)
For simplifying analysis, the magnetizing current of the transformer is neglected. When the secondary winding delivers a load current of I2 ampere the demagnetizing ampere turns is I2 T2 . This will be countered by a current I1 owing from the source through the T1 turns such that, I 1 T1 = I 2 T2 (77)
A current of I1 ampere ows through the winding between B and C . The current in the winding between A and B is (I2 I1 ) ampere. The cross section of the wire to be selected for AB is proportional to this current assuming a constant current density for the whole winding. Thus some amount of material saving can be achieved compared to a two 70
winding transformer. The magnetic circuit is assumed to be identical and hence there is no saving in the same. To quantify the saving the total quantity of copper used in an auto transformer is expressed as a fraction of that used in a two winding transformer as, copper in autotransf ormer (T1 T2 )I1 + T2 (I2 I1 ) = copper in two winding transf ormer T1 I 1 + T 2 I 2 2T2 I1 = 1 T1 I 1 + T 2 I 2 ButT1 I1 = T2 I2 The Ratio = 1 2T2 I1 T2 =1 2T1 I1 T1 (78)
(79) (80)
I1+I2
I1
I2
V1
I2
V2
V1+V2
ZL
I1+I2
I1
I2
Figure 29: Two Winding Transformer used as auto transformer
This means that an auto transformer requires the use of lesser quantity of copper given by the ratio of turns. This ratio therefore denotes the savings in copper. As the 71
space for the second winding need not be there, the window space can be less for an auto transformer, giving some saving in the lamination weight also. The larger the ratio of the voltages, smaller is the savings. As T2 approaches T1 the savings become signicant. Thus auto transformers become ideal choice for close ratio transformations. The savings in material is obtained, however, at a price. The electrical isolation between primary and secondary has to be sacriced.
If we are not looking at the savings in the material, even then going in for the auto transformer type of connection can be used with advantage, to obtain higher output. This can be illustrated as follows. Fig. 29 shows a regular two winding transformer of a voltage ratio V1 : V2 , the volt ampere rating being V1 I1 = V2 I2 = S . If now the primary is connected across a supply of V1 volt and the secondary is connected in series addition manner with the primary winding, the output voltage becomes (V1 + V2 ) volt. The new output of this auto transformer will now be I2 (V1 + V2 ) = I2 V2 (1 + V1 V1 ) = S (1 + ) V2 V2 I2 = V1 (I1 + I2 ) = S (1 + ) I1 (81) (82)
Thus an increased rating can be obtained compared to a two winding transformer with the same material content. The windings can be connected in series opposition fashion also. Then the new output rating will be
I2 (V1 V2 ) = I2 V2 (
V1 V1 1) = S ( 1) V2 V2
(83)
The dierential connection is not used as it is not advantageous as the cumulative connection. 72
11.1
Equivalent circuit
I1
r1,xl1
V1
I1
I2
r2,xl2
(I2 -I1)
V2
I1
I2
Figure 30: Kirchos Law Application to auto transformer
As mentioned earlier the magnetizing current can be neglected, for simplicity. Writing the Kirchos equation to the primary and secondary of Fig. 30 we have V1 = E1 + I1 (r1 + jxl1 ) (I2 I1 )(r2 + jxl2 ) (84)
Note that the resistance r1 and leakage reactance xl1 refer to that part of the winding where only the primary current ows. Similarly on the load side we have, E2 = V2 + (I2 I1 )(r2 + jxl2 ) The voltage ratio V1 : V2 of the primary. 73 = E 1 : E2 = T 1 : T2 = (85) a where T1 is the total turns
Then E1 = aE2 and I2 = aI1 multiplying equation(84) by a and substituting in (83) we have V1 = aV2 + a(I2 I1 )(r2 + jxl2 ) + I1 (r1 + jxl1 ) (I2 I1 )(r2 + jxl2 ) = aV2 + I1 (r1 + jxl1 + r2 + jxl2 ar2 ajxl 2) + I2 (ar2 + jaxl2 r2 jxl2 ) = aV2 + I1 (r1 + jxl1 + r2 + jxl2 + a2 r2 + ja2 xl2 ar2 ajxl2 ar2 jaxl2 ) = aV2 + I1 (r1 + r2 (1 + a2 2a) + jxl1 + xl2 (1 + a2 2a)) = aV2 + I1 (r1 + (a 1)2 r2 + jxl1 + (a 1)2 xl2 ) (86)
Equation (85) yields the equivalent circuit of Fig. 31 where Re = r1 + (a 1)2 r2 and
Re
jXe
Re=r1+(a-1)2r2 Xl=xl1+(a-1)2xl2
V1 Rc
Ic
Io
Im
jXm V2=aV1
Figure 31: Equivalent Circuit of auto transformers
Xe = xl1 + (a 1)2 xl2 . The magnetization branch can now be hung across the mains for completeness. The above equivalent circuit can now be compared with the approximate equivalent circuit of a two winding case Re = r1 + a2 r2 and Xe = xl1 + a2 xl2 . Thus in the case of an auto 74
I1x1 (I2-I1)x2 V1
(I2-I1)r2
I1r1
I2 E1 I1 (I2-I1)r2 E2 (I2-I1)x2
V2
1
I2
Figure 32: Phasor Diagram of Operation of an autotransformer
75
transformer total value of the short circuit impedance is lower and so also the percentage resistance and reactance. Thus the full load regulation is lower. Having a smaller value of short circuit impedance is sometimes considered to be a disadvantage. That is because the short circuit currents become very large in those cases. The eciency is higher in auto transformers compared to their two winding counter part at the same load. The phasor diagram of operation for the auto transformer drawing a load current at a lagging power factor angle of 2 is shown in Fig. 32. The magnetizing current is omitted here again for simplicity.
From the foregoing study it is seen that there are several advantages in going in for the autotransformer type of arrangement. The voltage/current transformation and impedance conversion aspects of a two winding transformer are retained but with lesser material (and hence lesser weight) used. The losses are reduced increasing the eciency. Reactance is reduced resulting in better regulation characteristics. All these benets are enhanced as the voltage ratio approaches unity. The price that is required to be paid is loss of electrical isolation and a larger short circuit current (and larger short circuit forces on the winding).
Auto transformers are used in applications where electrical isolation is not a critical requirement. When the ratio V2 : V1 is 0.3 or more they are used with advantage. The normal applications are motor starters, boosters or static balancers. Another wide spread application of auto transformer type of arrangement is in obtaining a variable a.c. voltage from a xed a.c. voltage supply. Here only one winding is used as in the auto transformer. The secondary voltage is tapped by a brush whose position and hence the output voltage is variable. The primary conductor is bared to facilitate electrical 76
Vari abl e a.c output
V in
M ovi ng contact
Figure 33: Variable Secondary Voltage Arrangement
contact Fig. 33. Such arrangement cannot exploit the savings in the copper as the output voltage is required right from zero volts upwards.
The conductor is selected based on the maximum secondary current that could be drawn as the output voltage varies in practically continuous manner. These are used in voltage stabilizers, variable d.c. arrangements (with a diode bridge) in laboratories, motor starters, dimmers etc.
77
12
Harmonics
In addition to the operation of transformers on the sinusoidal supplies, the harmonic
behavior becomes important as the size and rating of the transformer increases. The eects of the harmonic currents are 1. Additional copper losses due to harmonic currents 2. Increased core losses 3. Increased electro magnetic interference with communication circuits. On the other hand the harmonic voltages of the transformer cause 1. Increased dielectric stress on insulation 2. Electro static interference with communication circuits. 3. Resonance between winding reactance and feeder capacitance. In the present times a greater awareness is generated by the problems of harmonic voltages and currents produced by non-linear loads like the power electronic converters. These combine with non-linear nature of transformer core and produce severe distortions in voltages and currents and increase the power loss. Thus the study of harmonics is of great practical signicance in the operation of transformers. The discussion here is conned to the harmonics generated by transformers only.
78
e i i t i t i i
Figure 34: Harmonics Generated by Transformers
12.1
Single phase transformers

Modern transformers operate at increasing levels of saturation in order to reduce
the weight and cost of the core used in the same. Because of this and due to the hysteresis, the transformer core behaves as a highly non-linear element and generates harmonic voltages and currents. This is explained below. Fig. 34 shows the manner in which the shape of the magnetizing current can be obtained and plotted. At any instant of the ux density wave the ampere turns required to establish the same is read out and plotted, traversing the hysteresis loop once per cycle. The sinusoidal ux density curve represents the sinusoidal applied voltage to some other scale. The plot of the magnetizing current which is peaky is analyzed using Fourier analysis. The harmonic current components are obtained from this analysis. These harmonic currents produce harmonic elds in the core and harmonic voltages in the windings. Relatively small value of harmonic elds generate considerable magnitude of harmonic voltages. For example a 10% magnitude of 3rd harmonic ux produces 30% 79
magnitude of 3rd harmonic voltage. These eects get even more pronounced for higher order harmonics. As these harmonic voltages get short circuited through the low impedance of the supply they produce harmonic currents. These currents produce eects according to Lenzs law and tend to neutralize the harmonic ux and bring the ux wave to a sinusoid. Normally third harmonic is the largest in its magnitude and hence the discussion is based on it. The same can be told of other harmonics also. In the case of a single phase transformer the harmonics are conned mostly to the primary side as the source impedance is much smaller compared to the load impedance. The understanding of the phenomenon becomes more clear if the transformer is supplied with a sinusoidal current source. In this case current has to be sinusoidal and the harmonic currents cannot be supplied by the source and hence the induced emf will be peaky containing harmonic voltages. When the load is connected on the secondary side the harmonic currents ow through the load and voltage tends to become sinusoidal. The harmonic voltages induce electric stress on dielectrics and increased electro static interference. The harmonic currents produce losses and electro magnetic interference as already noted above.
12.2
Three phase banks of single phase transformers

In the case of single phase transformers connected to form three phase bank, each
transformer is magnetically decoupled from the other. The ow of harmonic currents are decided by the type of the electrical connection used on the primary and secondary sides. Also, there are three fundamental voltages in the present case each displaced from the other by 120 electrical degrees. Because of the symmetry of the a.c. wave about the time axis only odd harmonics need to be considered. The harmonics which are triplen (multiples of three) behave in a similar manner as they are co-phasal or in phase in the three phases. The 80
non-triplen harmonics behave in a similar manner to the fundamental and have 120 phase displacement between them. The harmonic behavior of poly-phase banks can be discussed now. Dd connection In three phase banks with mesh connection on both primary side and secondary side a closed path is available for the triplen harmonics to circulate currents. Thus the supply current is nearly sinusoidal (but for the non-triplen harmonic currents). The triplen harmonic currents inside the closed mesh winding correct the ux density wave to be nearly sinusoidal. The secondary voltages will be nearly sinusoidal. Third harmonics currents ow both in the primary and the secondary and hence the magnitudes of these currents, so also the drops due to them will be lower. Dy and Yd connection (without neutral connection) Behavior of the bank with mesh connection on one side is similar to the one discussed under Dd connection. The harmonic currents and drops and the departure of the ux density from sinusoidal are larger in the present case compared to Dd banks. Yy connection without neutral wires With both primary and secondary connected in star no closed path exists. As the triplen harmonics are always in phase, by virtue of the Y connection they get canceled in the line voltages. Non-triplen harmonics like fundamental, become 3 times phase value and appear in the line voltages. Line currents remain sinusoidal except for non-triplen harmonic currents. Flux wave in each transformer will be at topped and the phase voltages remain peaked. The potential of the neutral is no longer steady. The star point oscillates due to the third harmonic voltages. This is termed as oscillating neutral. Yy connection with neutral wires When a neutral wire is provided the triplen har81
monic current can ow and the condition is similar to the single phase case (with a star connected 4 wire source or with the system earth). The neutral wire carries three times the triplen harmonic current of one transformer as these currents are co-phasal. Unloaded secondary neutral will not be operative. Other polyphase connections not discussed above explicitly will fall under one type or the other of the cases discussed. In a Yy connection, to obtain third harmonic suppression one may provide a third winding which is connected in mesh, which can be an unloaded winding. It is called a tertiary. This winding improves the single phase to earth fault detection also. Further, this winding can be used to feed some permanent station loads also. Such transformers are designated as Yyd transformers. If the neutral wires are provided and also mesh connected winding is present, then triplen harmonics are shared between them depending upon their impedances.
12.3
Three phase transformers units

As against a bank of three single phase transformers connected to three phase
mains, a three phase transformer generally has the three magnetic circuits that are interacting. The exception to this rule is a 3-phase shell type transformer. In the shell type of construction, even though the three cores are together, they are non-interacting. Three limb core type 3-phase transformer is the one in which the phases are magnetically also linked. Flux of each limb uses the other two limbs for its return path. This is true for fundamental and non-triplen harmonics. The triplen harmonics being co-phasal cannot use other limbs for the return path (this holds good for zero sequence, unbalanced fundamental mmf also). The ux path is completed through the air. So substantially large value of the mmf produces a low value of third harmonic ux as the path of the ux is through the air and has a very 82
high reluctance. Thus the ux in the core remains nearly sinusoidal, so also the induced emf. This happens irrespective of the type of connection used. The triplen order ux, sometimes links the tank and produces loss in the same.
Other harmonics can be suppressed by connecting tuned lters at the terminals. Harmonic current compensation using special magnetic circuit design is considered to be outside the scope here.
13
Poly Phase connections and Poly phase Transformers

The individual transformers are connected in a variety of ways in a power system.
Due to the advantages of polyphase power during generation, transmission and utilization polyphase power handling is very important. As an engineering application is driven by techno-economic considerations, no single connection or setup is satisfactory for all applications. Thus transformers are deployed in many forms and connections. Star and mesh connections are very commonly used. Apart from these, vee or open delta connections, zig zag connections , T connections, auto transformer connections, multi winding transformers etc. are a few of the many possibilities. A few of the common connections and the technical and economic considerations that govern their usage are discussed here. Literature abounds in the description of many other. Apart from the characteristics and advantages of these, one must also know their limitations and problems, to facilitate proper selection of a conguration for an application. Many polyphase connections can be formed using single phase transformers. In some cases it may be preferable to design, develop and deploy a 83
polyphase transformer itself. In a balanced two phase system we encounter two voltages that are equal in magnitude diering in phase by 90 . Similarly, in a three phase system there are three equal voltages diering in phase 120 electrical degrees. Further there is an order in which they reach a particular voltage magnitude. This is called the phase sequence. In some applications like a.c. to d.c. conversion, six phases or more may be encountered. Transformers used in all these applications must be connected properly for proper functioning. The basic relationship between the primary and secondary voltages (brought about by a common mutual ux and the number of turns), the polarity of the induced emf (decided by polarity test and used with dot convention) and some understanding of the magnetic circuit are all necessary for the same. To facilitate the manufacturer and users, international standards are also available. Each winding has two ends designated as 1 and 2. The HV winding is indicated by capital letters and the LV winding by small letters. If more terminals are brought out from a winding by way of taps there are numbered in the increasing numbers in accordance to their distance from 1 (eg A1 , A2 , A3 ...). If the induced emf at an instant is from A1 to A2 on the HV winding it will rise from a1 to a2 on the LV winding.
Out of the dierent polyphase connections three phase connections are mostly encountered due to the wide spread use of three phase systems for generation, transmission and utilization. Three balanced 3-phase voltages can be connected in star or mesh fashion to yield a balanced 3-phase 3-wire system. The transformers that work on the 3-phase supply have star, mesh or zig-zag connected windings on either primary secondary or both. In addition to giving dierent voltage ratios, they introduce phase shifts between input and output sides. These connections are broadly classied into 4 popular vector groups.
84
1. Group I: zero phase displacement between the primary and the secondary. 2. Group II: 180 phase displacement. 3. Group III: 30 lag phase displacement of the secondary with respect to the primary. 4. Group IV: 30 lead phase displacement of the secondary with respect to the primary. A few examples of the physical connections and phasor diagrams are shown in Fig. 35 and Fig. 36 corresponding to each group. The capital letters indicates primary and the small letters the secondary. D/d stand for mesh, Y/y - for star, Z/z for zig-zag. The angular displacement of secondary with respect to the primary are shown as clock position, 0 referring to 12 oclock position. These vector groups are especially important when two or more transformers are to be connected in parallel.
Star connection is normally cheaper as there are fewer turns and lesser cost of insulation. The advantage becomes more with increase in voltage above 11kv. In a star connected
1 winding with earthed-neutral the maximum voltage to the earth is ( )of the line voltage. 3
Also star connection permits mixed loading due to the presence of the neutral. Mesh connections are advantageous in low voltage transformers as insulation costs are insignicant
1 and the conductor size becomes ( ) of that of star connection and permits ease of winding. 3
The common polyphase connections are briey discussed now. Star/star (Yy0, Yy6)connection This is the most economical one for small high voltage transformers. Insulation cost is highly reduced. Neutral wire can permit mixed loading. Triplen harmonics are absent in the lines. These triplen harmonic currents cannot ow, unless there is a neutral wire. This connection produces oscillating neutral. Three 85
G roup1 0 0 Phase shi f t Wi ndi ngs & Term i nal s

A1 B1 C1
E. M. F Vector di agram s
a1 b1 c1 a1 b1 c1
A 2 a2 B 2 b2 C 2 c2 A 2 a2 B 2 b2 C 2 c2
A2
a2
C2 A2
B2 a2
A1 B1 C1
C2 A2
B2
n
A1 B1 C1 A 2 a4 B 2 b4 C 2 c4 a1 b1 c1
C2
B2
(a)

A1 B1 C1
A1 B1 C1
A1
a1 b1 c1
N n
A 2 a1 B 2 b1 C 2 c1
A 2a1 B 2b1 C 2c1 a2 b2 c2
A2
C2
A2
B2
a1 b1 c1
n
A 2 a3 a1 b1
C2
A2
B2
B1
B 2b3 C 2 c3
C1
c1
C2
B2
(b) Figure 35: Vector Groups for 3-phase Transformer Connections 86

n
A1 B1 C1 A 2 a2 B 2 b2 C 2 c2 a1 b1 c1 a1 b1 c1
A2
a2
N
A1 B1 C1
A 2 a2 B 2 b2 C 2 c2
C2
A2 B2
a2
N n
A1 B1 C1 A 2 a4 B 2 b4 C 2 c4 a1 b1 c1
C2 A2
B2
C2
B2
(a)
G roup4 + 30 0 Phase shi f t Wi ndi ngs & Term i nal s

n
A1 B1 C1 A 2 a2 B 2 b2 c2 a1 b1 c1
a2
A1 B1 C1
N
A 2 a1 B 2 b1 C 2 c1
C2
a2 b2 c2 a1 b1 c1
a3 b3 c3
A2
B2
C2 B2 A2
C2 B2
A1 B1 C1
N
A2 B2
a4 b4
a1 b1 c1
C2 B2
C 2 c4
(b) Figure 36: Vector Groups for 3-phase Transformer Connections 87
phase shell type units have large triplen harmonic phase voltage. However three phase core type transformers work satisfactorily. A tertiary mesh connected winding may be required to stabilize the oscillating neutral due to third harmonics in three phase banks.
Mesh/mesh (Dd0, Dd6) This is an economical conguration for large low voltage transformers. Large amount of unbalanced load can be met with ease. Mesh permits a circulating path for triplen harmonics thus attenuates the same. It is possible to operate with one transformer removed in open delta or Vee connection meeting 58 percent of the balanced load. Three phase units cannot have this facility. Mixed single phase loading is not possible due to the absence of neutral.
Star/mesh(Dy or Yd ) This arrangement is very common for power supply transformers. The delta winding permits triplen harmonic currents to circulate in the closed path and attenuates them. Zig zag/ star (ZY1 or Zy11) Zigzag connection is obtained by inter connection of phases. 4-wire system is possible on both sides. Unbalanced loading is also possible. Oscillating neutral problem is absent in this connection. This connection requires 15% more turns for the same voltage on the zigzag side and hence costs more.
Generally speaking a bank of three single phase transformers cost about 15% more than their 3-phase counter part. Also, they occupy more space. But the spare capacity cost will be less and single phase units are easier to transport.
88
Mesh connected three phase transformers resemble 3- single phase units but kept in a common tank. In view of this single tank, the space occupied is less. Other than that there is no big dierence. The 3-phase core type transformer on the other hand has a simple core arrangement. The three limbs are equal in cross section. Primary and secondary of each phase are housed on the same limb. The ux setup in any limb will return through the other two limbs as the mmf of those limbs are in the proper directions so as to aid the same. Even though magnetically this is not a symmetrical arrangement, as the reluctance to the ux setup by side limbs is dierent from that of the central limb, it does not adversely aect the performance. This is due to the fact that the magnetizing current itself forms a small fraction of the total phase current drawn on load. The added advantage of 3-phase core is that it can tolerate substantially large value of 3rd harmonic mmf without aecting the performance. The 3rd harmonic mmf of the three phases will be in phase and hence rise in all the limbs together. The 3rd harmonic ux must therefore nd its path through the air. Due to the high reluctance of the air path even a substantially large value of third harmonic mmf produces negligible value of third harmonic ux. Similarly unbalanced operation of the transformer with large zero sequence fundamental mmf content also does not aect its performance. Even with Yy type of poly phase connection without neutral connection the oscillating neutral does not occur with these cores. Finally, three phase cores themselves cost less than three single phase units due to compactness.
89
14
Parallel operation of one phase and two phase transformers

By parallel operation we mean two or more transformers are connected to the same
supply bus bars on the primary side and to a common bus bar/load on the secondary side. Such requirement is frequently encountered in practice. The reasons that necessitate parallel operation are as follows. 1. Non-availability of a single large transformer to meet the total load requirement. 2. The power demand might have increased over a time necessitating augmentation of the capacity. More transformers connected in parallel will then be pressed into service. 3. To ensure improved reliability. Even if one of the transformers gets into a fault or is taken out for maintenance/repair the load can continued to be serviced. 4. To reduce the spare capacity. If many smaller size transformers are used one machine can be used as spare. If only one large machine is feeding the load, a spare of similar rating has to be available. The problem of spares becomes more acute with fewer machines in service at a location. 5. When transportation problems limit installation of large transformers at site, it may be easier to transport smaller ones to site and work them in parallel. Fig. 37 shows the physical arrangement of two single phase transformers working in parallel on the primary side. Transformer A and Transformer B are connected to input voltage bus bars. After ascertaining the polarities they are connected to output/load bus
90
V1
E1
A
E2
IA
V2
load
IB
E1 E2
supply bus
Load bus
Figure 37: Parallel Operation of Two Single Phase Transformers - Physical
91
bars. Certain conditions have to be met before two or more transformers are connected in parallel and share a common load satisfactorily. They are, 1. The voltage ratio must be the same. 2. The per unit impedance of each machine on its own base must be the same. 3. The polarity must be the same, so that there is no circulating current between the transformers. 4. The phase sequence must be the same and no phase dierence must exist between the voltages of the two transformers. These conditions are examined rst with reference to single phase transformers and then the three phase cases are discussed. Same voltage ratio Generally the turns ratio and voltage ratio are taken to be the same. If the ratio is large there can be considerable error in the voltages even if the turns ratios are the same. When the primaries are connected to same bus bars, if the secondaries do not show the same voltage, paralleling them would result in a circulating current between the secondaries. Reected circulating current will be there on the primary side also. Thus even without connecting a load considerable current can be drawn by the transformers and they produce copper losses. In two identical transformers with percentage impedance of 5 percent, a no-load voltage dierence of one percent will result in a circulating current of 10 percent of full load current. This circulating current gets added to the load current when the load is connected resulting in unequal sharing of the load. In such cases the combined full load of the two transformers can never be met without one transformer getting overloaded. 92
Per unit impedance Transformers of dierent ratings may be required to operate in parallel. If they have to share the total load in proportion to their ratings the larger machine has to draw more current. The voltage drop across each machine has to be the same by virtue of their connection at the input and the output ends. Thus the larger machines have smaller impedance and smaller machines must have larger ohmic impedance. Thus the impedances must be in the inverse ratios of the ratings. As the voltage drops must be the same the per unit impedance of each transformer on its own base, must be equal. In addition if active and reactive power are required to be shared in proportion to the ratings the impedance angles also must be the same. Thus we have the requirement that per unit resistance and per unit reactance of both the transformers must be the same for proper load sharing. Polarity of connection The polarity of connection in the case of single phase transformers can be either same or opposite. Inside the loop formed by the two secondaries the resulting voltage must be zero. If wrong polarity is chosen the two voltages get added and short circuit results. In the case of polyphase banks it is possible to have permanent phase error between the phases with substantial circulating current. Such transformer banks must not be connected in parallel. The turns ratios in such groups can be adjusted to give very close voltage ratios but phase errors cannot be compensated. Phase error of 0.6 degree gives rise to one percent dierence in voltage. Hence poly phase transformers belonging to the same vector group alone must be taken for paralleling. Transformers having 30 angle can be paralleled to that having +30 angle by reversing the phase sequence of both primary and secondary terminals of one of the transformers. This way one can overcome the problem of the phase angle error.
93
Phase sequence The phase sequence of operation becomes relevant only in the case of poly phase systems. The poly phase banks belonging to same vector group can be connected in parallel. A transformer with +30 phase angle however can be paralleled with the one with 30 phase angle, the phase sequence is reversed for one of them both at primary and secondary terminals. If the phase sequences are not the same then the two transformers cannot be connected in parallel even if they belong to same vector group. The phase sequence can be found out by the use of a phase sequence indicator. Performance of two or more single phase transformers working in parallel can be computed using their equivalent circuit. In the case of poly phase banks also the approach is identical and the single phase equivalent circuit of the same can be used. Basically two cases arise in these problems. Case A: when the voltage ratio of the two transformers is the same and Case B: when the voltage ratios are not the same. These are discussed now in sequence.
14.1
CASE A: Equal voltage ratios

Always two transformers of equal voltage ratios are selected for working in parallel.
This way one can avoid a circulating current between the transformers. Load can be switched on subsequently to these bus bars. Neglecting the parallel branch of the equivalent circuit the above connection can be shown as in Fig. 38(a),(b). The equivalent circuit is drawn in terms of the secondary parameters. This may be further simplied as shown under Fig. 38(c). The voltage drop across the two transformers must be the same by virtue of common connection at input as well as output ends. By inspection the voltage equation for the drop can be
94
RA
j XA
IA
V 2
RA
jX A
V1
IA
ZA
V1
RB j XB IB
RB
jX B
IB
V 2
ZB
(a)
ZA IA I IB
(b)
ZB VL V
Load
V L
ZL
(c) Figure 38: Equivalent Circuit for Transformers working in Parallel -Simplied circuit and Further simplication for identical voltage ratio
95
written as IA ZA = IB ZB = IZ = v (say ) (87) (88)
HereI = IA + IB And Z is the equivalent impedance of the two transformers given by, Z= ZA ZB ZA + Z B Thus IA = IB v IZ ZB = = I. and ZA ZA ZA + Z B v IZ ZA = = = I. ZB ZB ZA + Z B
(89) (90) (91)
If the terminal voltage is V = IZL then the active and reactive power supplied by each of the two transformers is given by
PA = Real(V IA )andQA = Imag (V IA )and PB = Real(V IB )andQB = Imag (V IB )
(92) (93) (94)
From the above it is seen that the transformer with higher impedance supplies lesser load current and vice versa. If transformers of dissimilar ratings are paralleled the transformer with larger rating shall have smaller impedance as it has to produce the same drop as the other transformer, at a larger current. Thus the ohmic values of the impedances must be in the inverse ratio of the ratings of the transformers. IA ZA = IB ZB
IA IB
ZB . ZA
Expressing the voltage drops in p.u basis, we aim at the same per unit drops at any load for the transformers. The per unit impedances must therefore be the same on their respective bases. Fig. 39 shows the phasor diagram of operation for these conditions. The drops are magnied and shown to improve clarity. It is seen that the total voltage drop inside the 96
V V2 IAXA IX IBXB
B
IA
A
IB IL
A B
IARA
IR
IBRB
Figure 39: Phasor Diagram of Operation for two Transformers working in Parallel
transformers is v but the currents IA and IB are forced to have a dierent phase angle due to the dierence in the internal power factor angles A and B . This forces the active and reactive components of the currents drawn by each transformer to be dierent ( even in the case when current in each transformer is the same). If we want them to share the load current in proportion to their ratings, their percentage ( or p.u) impedances must be the same. In order to avoid any divergence and to share active and reactive powers also properly, A = B . Thus the condition for satisfactory parallel operation is that the p.u resistances and p.u reactance must be the same on their respective bases for the two transformers. To determine the sharing of currents and power either p.u parameters or ohmic values can be used.
97
14.2
Case B :unequal voltage ratios

RA j XA
IA I
RB
j XB
IB
EA
EB
ZL
VL
Figure 40: Equivalent Circuit for unequal Voltage Ratio
One may not be able to get two transformers of identical voltage ratio in spite of ones best eorts. Due to manufacturing dierences, even in transformers built as per the same design, the voltage ratios may not be the same. In such cases the circuit representation for parallel operation will be dierent as shown in Fig. 40. In this case the two input voltages cannot be merged to one, as they are dierent. The load brings about a common connection at the output side. EA and EB are the no-load secondary emf. ZL is the load impedance at the secondary terminals. By inspection the voltage equation can be written as below: EA = IA ZA + (IA + IB )ZL = V + IA ZA EB = IB ZB + (IA + IB )ZL = V + IB ZB Solving the two equations the expression for IA and IB can be obtained as 98 (95)
IA = IB
EA ZB + (EA EB )ZL ZA ZB + ZL (ZA + ZB ) EB ZA + (EB EA )ZL = ZA ZB + ZL (ZA + ZB )
and
(96)
ZA and ZB are phasors and hence there can be angular dierence also in addition to the dierence in magnitude. When load is not connected there will be a circulating current between the transformers. The currents in that case can be obtained by putting ZL = ( after dividing the numerator and the denominator by ZL ). Then, IA = IB = (EA EB ) (ZA + ZB ) (97)
If the load impedance becomes zero as in the case of a short circuit, we have, IA = EA ZA and IB = EB ZB (98)
Instead of the value of ZL if the value of V is known , the currents can be easily determined ( from Eqns. 95 ) as IA = EA V ZA and IB = EB V ZB (99)
If more than two transformers are connected across a load then the calculation of load currents following the method suggested above involves considerable amount of computational labor. A simpler and more elegant method for the case depicted in Fig. 41 is given below. It is known by the name parallel generator theorem. IL = IA + IB + IC + ...... But IA = V EA V , ZA = IL .ZL IB = EB V , ZB IC = EC V ZC (100) 99
RA EA EB RC EC RB
jXA jXB jXC
IA IB IC ZL
V
Figure 41: Parallel Generator Theorem
Combining these equations V EA V EB V EC V = + + + ... ZL ZA ZB ZC Grouping the terms together V( 1 1 1 1 EA EB EC + + + + ...) = + + + ... ZL ZA ZB ZC ZA ZB ZC = ISCA + ISCB + ISCC + .... 1 1 1 1 1 + + + + ...) = ZL ZA ZB ZC Z V = Z (ISCA + ISCB + ISCC + ....) ( (101)
(102) (103) (104)
From this V can be obtained. Substituting V in Eqn. 100, IA , IB etc can be obtained. Knowing the individual current phasor, the load shared by each transformer can be computed.
100
15
Transformer voltage control and Tap changing

Regulating the voltage of a transformer is a requirement that often arises in a power
application or power system. In an application it may be needed 1. To supply a desired voltage to the load. 2. To counter the voltage drops due to loads. 3. To counter the input supply voltage changes on load. On a power system the transformers are additionally required to perform the task of regulation of active and reactive power ows.
Booster transformer Regulation transformer B
1 2
Booster transformer B
1
Reverser
Main transformer
tertiary2
Reverser
Figure 42: Tap changing and Buck Boost arrangement
The voltage control is performed by changing the turns ratio. This is done by provision of taps in the winding. The volts per turn available in large transformers is quite high and hence a change of even one turn on the LV side represents a large percentage change in the voltage. Also the LV currents are normally too large to take out the tapping from the windings. LV winding being the inner winding in a core type transformer adds to the 101
diculty of taking out of the taps. Hence irrespective of the end use for which tapping is put to, taps are provided on the HV winding. Provision of taps to control voltage is called tap changing. In the case of power systems, voltage levels are some times changed by injecting a suitable voltage in series with the line. This may be called buck-boost arrangement. In addition to the magnitude, phase of the injected voltage may be varied in power systems. The tap changing arrangement and buck boost arrangement with phase shift are shown in Fig. 42.
Tap changing can be eected when a) the transformers is on no- load and b) the load is still remains connected to the transformer. These are called o load tap changing and on load tap changing. The O load tap changing relatively costs less. The tap positions are changed when the transformer is taken out of the circuit and reconnected. The on-load tap changer on the other hand tries to change the taps without the interruption of the load current. In view of this requirement it normally costs more. A few schemes of on-load tap changing are now discussed. Tap 1 2 3 4 5 6 7 8 9 switches closed 1,s 1,2 2,s 2,3 3,s 3,4 4,s 4,5 5,s
Reactor method The diagram of connections is shown in Fig. 43. This method employs 102
1 s 2 3 Reactor 4 5
HV
LV
Figure 43: Reactor Method of Tap Changer ( with table of switching)
103
an auxiliary reactor to assist tap changing. The switches for the taps and that across the reactor(S) are connected as shown. The reactor has a center tapped winding on a magnetic core. The two ends of the reactor are connected to the two bus bars to which tapping switches of odd/even numbered taps are connected. When only one tap is connected to the reactor the shorting switch S is closed minimizing the drop in the reactor. The reactor can also be worked with both ends connected to two successive taps. In that case the switch S must be kept open. The reactor limits the circulating current between the taps in such a situation. Thus a four step tapped winding can be used for getting seven step voltage on the secondary(see the table of switching). The advantage of this type of tap changer are 1. Load need not be switched o. 2. More steps than taps are obtained. 3. Switches need not interrupt load current as a alternate path is always provided. The major objection to this scheme seems to be that the reactor is in the circuit always generating extra loss. Parallel winding, transformer method In order to maintain the continuity of supply the primary winding is split into two parallel circuits each circuit having the taps as shown in Fig. 44. Two circuit breakers A and B are used in the two circuits. Initially tap 1a and 1b are closed and the transformer is energized with full primary voltage. To change the tap the circuit breaker A is opened momentarily and tap is moved from 1a to 2a. Then circuit breaker A is closed. When the circuit A is opened whole of the primary current of the transformer ows through the circuit B. A small dierence in the number of turns between the two circuit exists. This produces a circulating current between them. Next, circuit breaker B is opened momentarily, the tap is changed from 104
1b to 2b and the breaker is closed. In this position the two circuits are similar and there is no circulating current. The circulating current is controlled by careful selection of the leakage reactance. Generally, parallel circuits are needed in primary and secondary to carry the large current in a big transformer. Provision of taps switches and circuit breakers are to be additionally provided to achieve tap changing in these machines.
A B
a2 a4 a3
a1
b1
b2 b3
b4
HV
LV
Figure 44: Parallel Primary Winding Tap Changing
Series booster method In this case a separate transformer is used to buck/boost the voltage of the main transformer. The main transformer need not be having a tapped arrangement. This arrangement can be added to an existing system also. Fig. 42shows the booster arrangement for a single phase supply. The reverser switch reverses the polarity of the injected voltage and hence a boost is converted into a buck and vice 105
versa. The power rating of this transformer need be a small fraction of the main transformer as it is required to handle only the power associated with the injected voltage. One precaution to be taken with this arrangement is that the winding must not be open circuited. If it gets open circuited the core (B in g) gets highly saturated. In spite of the small ratings and low voltages and exibility, this method of voltage control costs more mainly due to the additional oor space it needs. The methods of voltage regulation discussed so far basically use the principle of tap changing and hence the voltage change takes place in steps. Applications like a.c. and d.c. motor speed control, illumination control by dimmers, electro-chemistry and voltage stabilizers need continuous control of voltage. This can be obtained with the help of moving coil voltage regulators. moving coil voltage regulators Fig. 45 shows the physical arrangement of one such transformer. a, b are the two primary windings wound on a long core, wound in the opposite sense. Thus the ux produced by each winding takes a path through the air to link the winding. These uxes link their secondaries a2 and b2 . A short circuited moving coil s is wound on the same limb and is capable of being held at any desired position. This moving coil alters the inductances of the two primaries. The sharing of the total applied voltage thus becomes dierent and also the induced emf in the secondaries a2 and b2 . The total secondary voltage in the present case varies from 10 percent to 20 percent of the input in a continuous manner. The turns ratios of a1 : a2 and b1 : b2 are 4.86 and 10.6 respectively. the bottom position it becomes
5 4.86 95 4.86
+ +
95 10.6 5 10.6
= 10% when s is in the top position. In = 20%. By selecting proper ratios for the
secondaries a2 and b2 one can get the desired voltage variation. Sliding contact regulators These have two winding or auto transformer like construction.
106
5%
a1
a2
V out Vi n
S
95%
b1 b2
Figure 45: Moving Coil Voltage Regulator
Variable secondary a.c voltage
V1 V2
V1
sliding contact
a) without electrical isolation
b) with electrical isolation
Figure 46: Sliding Contact Regulator
107
The winding from which the output is taken is bared and a sliding contact taps the voltage. The minimum step size of voltage change obtainable is the voltage across a single turn. The conductor is chosen on the basis of the maximum load current on the output side. In smaller ratings this is highly cost eective. Two winding arrangements are also possible. The two winding arrangement provides electrical isolation also.These are shown in Fig. 46.
108

Transformers

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Transformers

Caricato da

Copyright:

Formati disponibili

Electrical Machines I

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Figure 1: Flux linkages of a coil

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Figure 2: Static coil

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

is assumed to be varying sinusoidally, and can be expressed as, B = Bm sin t (4)

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

(c) Figure 3: E and I,C and I and O Type Laminations

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

(b) Shell type

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Figure 5: Stepped Core Construction

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

(b) Sandwich coil

Figure 7: Concentric and Sandwich Coils

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

cross over coils

Figure 8: Disc, Crossover and Helical Coil Construction

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

reactance can be substantially reduced.

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Conservator & Breather

Oil pump for O.F.B

(c) 20 Figure 9: Some Typical Cooling Arrangements

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao

Indian Institute of Technology Madras

Prof. Krishna Vasudevan, Prof. G. Sridhara Rao, Prof. P. Sasidhara Rao