Transformer Protection

The main protection applied to the power Transformers are the Transformer differential protection , REF protection , and Back up over current protection apart from the winding temperature and oil temperature trips and pressure relief , Bucholz etc . Strictly speaking the Winding temperature and oil temperature , bucholz etc are not operating in instatanious way but in a gradual way
It may be noted that REF protection is used to cover the unprotected part of the winding by the differential protection .


10.High impedance REF protection

Before discussing the details of the REF protection it is better to have a good understanding between high impedance and low impedeance protection . It is a well known fact that the REF protection and the differential protections are basically working on the krichoffs current low ( there are volage differential relay also !)in which CTs are placed to measure the incoming current to the protected area and the out going current from the protected area . If this currents are equal then we conclude that there is no internal fault in side the area where the protection is supposed to be protected . obviously if there is any difference beteween the incoming and out going current then , theoritically we will conclude that there is a fault inside the protected zone
It is intersting to note that eventhough we are aware of the krichoffs current law applicable to a point or node , it is equally applicable to a surface. To understand the concept cleraly we can draw a complete closed path encircling the protected object and so that the drawn closed path appears like a surface . Now try to add all the incoming currents and outgoing currents algebrically ,then in all cases of normal system ( no fault ) this current adds to Zero ( of course it applies to only lenior networks and there shall be proper ratio factor applied if the surface includes a Transformer ).

Now coming back to the discussion on the REF protection, it is now clear that all the CT currents are added and they add up to zero under normal system condition . Now let us discuss about the causes of errors in this type of system under practical condition .
If all the CTs involved are of same characteristics then their error also become same and obviously there is no error current flowing through the relay and the relay will be stable . However if there is any mismatch among the charecteristics of the CTs involve inthe REF protection it will add to error current and and if this error current is above the relay setting the relay will maloperate under normal operating condition . The most important cause of error in a Ct is the magnetising current taken by the CT . ie the secondry current will not be propotional to the primary current devided by the transformation ratio but it will be lesser than this expecetd current by the magnetising current taken by the CT . Hence if the CTs involved having different magnetising current , then it will appear as a differential current in the relay and it will maloperate if the setting is below the maximum error current so obtained . Hence the setting of the relay is done based on the maximum expected error current in the CTs under consideration . In order to reduce the error current , the prefered CTs for this type of protection should have low magnetising current as possible.
Now it is understood that , if the CTs are not designed for the full expected through fault current , then there is every possibility for the relay to operate in case one of the CTs involved are operating in different operating point .
Now let us discuss about the high impedance resistor connected in series with the relay and its use in case of REF protection
During a through earth fault condition in which one of the CT is driven to saturation due to various factors like inadequate kneepoint voltage or a possible practical error in the CT material, construction etc , then only current from the othr CT is only present to the relay and hence the relay will trip as there is no current from the other CT to balance the healthy CT current . To overcoeme such a condition and to make the protection stable for such condition , it can be considered that the saturated CT is no longer working as a current source but it is acting as a simple winding with RCT ( resistance of CT ) . Hence the current from the other CT will also be flowing thru this resistance a making a voltage drop acros it . The maximum possible voltage across this resistance is found to be during a condition under which all the healthy phase CT current is flowing thru this saturated CT . The voltage across the saturated CT due to the current due to the healthy phase CT is found out and as this voltage is directly acting across the relay circuit , a stabilising resistance is added in series with the Relay circuit so that the current flowing thru the relay circuit is kept below the set current of the relay .
From the forgoing discussion , it can be concluded that there is error current in the REF relay during the following condition which requires special attention of the designer
1. The error current introduced due to difference in magentising charecteristics o the CTs involved - This can be over come by the basic setting of the relay
2. The error current introduced under fully saturated condition of one of the CT - This can be overcome by the proper designing of the stabilising resistor
3. The error current introduced due to partial CT saturation - this is a state of the CT in between the 1 and 2 above and appears .We have two cases to be analysed in this case
a, A through earth fault so that only one of the phase CT and nuertal CT is practically involved in the fault . As both the currents are opposing each other in Relay circuit there will not be any current in the Relay circuit > But in case of one of the CT is saturated partially ,the healthy CT will drive the current through the saturated CT partially and the voltage is developed across the Rct of the saturated CT in part of the cycle .But still no operation of the relay is possible .
b. In case of through phase faults . In this case if there is no CT saturation , there will not be any spill current from the there phase CTs and practically no current from the neutral CT . But if one of the phase Ct is saturated , the other two CTs will try to complete the path thru the saturated CT . But as the primary current is balanced , the current which will pass thru the saturated CT can not be more than the actual current that would flow if the CT is not saturated . But the moment the CT is saturated , the current will be revered which would develope voltage across the Rct . But as the stabilising resistance is designed for this current also , the circuit will be stable .
c. The case of charging of transformer - During this operation due the different point on switching and remenance in the power transformer as well as the CT , the current could be unbalanced one . Ideally if the primary current is unbalanced and no saturation of the CTs , the spill current from the phase CT will exactly match with the current from the neutral CT . But if there a phase CT saturation, then the current in the ( to be continued )
10.1 A comparison of low impedance and high impedance protection based on the stability

In case of the high impedance REF or differential protection , the stabilising resistor is used to prevent the unwanted operation durring a through fault condition when one of the CT is completely saturated . Hence the circuit is stable even if the CT is completly sturated . However if the CT is not designed to meet the transient and steady state condition , there is a chance for a spill current if one of the CT is partially saturated . ie the CT is working in the upper non lenior portion of the magnestising current , Under this condition there can be considerable difference between the mag current of the CTs involved which may lead to tripping the relay if the error current is more than the set value
However in case of low impedance protection using the bias technique which is used to overcome the CT errors and mismatch of the CTS and not a measure for the CT full saturation . (If the bias characteristics of the differential relay is adjusted so that the tripping is prevented for one of the CT complete saturation, then it will not trip for an internal fault at all!)

10.2 Current setting of the REF relay
What is the minimum setting that can be applied in an REF relay is intersting to discuss. Theoritically you can set any current above zero so that , whenever there is some differential current in the relay , the relay can operate. However in practice the CTs even though it is manufatured with same caharecteristics , there are erros in the CT . This errors could be due the slight difference in the magentising current and due to the turns ratio error . For exmple a class X CT , the turns ratio error can be a maximum of +- 0.25% and this will give a error current of 2.5mA at 1A rating . If one of the CT involved is having an error of +2.5 and the other -2.5 , then both will add up in the REF circuit and give rise to spill current of 5mA under normal load condition . But this error will be much more under through fault condition for which the relay is not supposed to operate . Hence the this error shall be considerd before the minimum setting is decided on the REF relay . .

10.3Stabilising Voltage of REF relay

As discussed in the previous paragraph , when one of the CT in a REF circut is saturated , it will no longer work as CT and such a saturated CT can be replaced by its resistance . Hence the healthy CT current will be flowing through resistance of the saturated CT circuit . Hence as the relay circuit is in parallel with this CT circuit , the voltage that appears acros the saturated CT also appears acros the Relay circuit .This volatge is called stabilising Voltage . In order to prevent the operation of the Relay under this condition , the relay circuit burden is increased to a value so that the current in the Relay circuit is kept below the set current of the Relay . This is achieved by adding a resistance in series with the Relay circuit and this resistance is called stabilising resistance.

10.4 Sensitivity and magnetising current

The sensitivity of the REF protection depends on the magentising current of the connecetd CTs . The current that is available to the relay is the balance current after supplying the magnetising current of all the CTs. ie the current that is available in the secondary circuit of the CT is not equal to the step down current based on the transformation ratio of the CT but lesser than it be the magentising current required by the CT core . Hence if the fault current available in the primary of the CT is very low such that it is just sufficient for the CT magnetising curent ,there will be no current available in the secondray circuit for the relay to operate . Hence for such a low primary faullt current , the relay may not operate at all and in otherwise the relay is not sensitive to such currents .
Ideally for a internal fault and suppose the fault current supplied by nuetral CT and there is no current from the phase CTS . under such a fault condition the nuetral CT,apart from its own maganetising current has to supply the parallel conncetd CTs ie another 4CTs . Hence the relay will not operate for 4 times the magnetising current plus the set current of the relay ( converted to primary ). In otherwords the relay is not sensitive up to the above current.

10.5 Stabiliy and magnetising current

As in the case of sensitivity , the magnetising current also effects the stability of the REF protection . The CT sizing calculation is done for the through fault condition of the Transformer . ie the CT should be stable for a through fault condition . Under a through earth fault condition , the faulted phase CT has to supply the magnetising current of the unfaulted phase the other two CTs will be clamped to the same pottential as that of the faulted phase . Hence this current ie the magnetisng current supplied to the healthy phases will be acting as the error current in the REF and the relay mimimum operating current will be above this current.

10.6 Effect of value of stabilising resistance on stability and sensitivity

The value of stabilising resistor is foumd out based on the value of the stabilising voltage obatined when one of the CT is completly saturated . Now if the stabilising resistance is set at a higher and higher value the stability will be more and more for CT saturation . However for an internal fault as the burden seen by the CT is higher , there is a possibility for the CT to saturate and threby preventing the relay from operating

11.Transformer Differential relay

Transformer differential relay also working on the principle of current differential as in REF relay and they can also be high impedance and low impedance based . The high impedance principle is same as that of the REF principle as discussed above and the discussion in this topic is restricted to the low impedance principle

In order to make the relay stable for the differential current due to the slight mismatch in the primary and secondary current due to the effect of tap position during the normal operating condition and to account for the mismatch due to relatively large mismatch that would appear due to CT error and the slight saturation for through faults the relay uses the so called bias or retrain technique
The low impedance bias differential relay has basically 3 setting points and slopes . Before we proceed further on the matter , it is to be understood that , for every setting , the bottom line is that the Relay shall not trip under normal operation . Hence the limiting conditions are no load , load and the through fault of the transformer under which the relay shall not trip or in other words it shall refrain from operating . The more the settings are close to the above limiting condition , the more sensitive the relay operation is obtained . Hence we always have some fault condition under which the relay may not operate because the fault current is below the setting selected .
To understand the point more clearly , we can discuss the basic setting of the transformer differential relay . This setting is usually selected to avoid the tripping of the transformer under no load condition and to avoid any error introduced by the CTs involved . Generally 0.2 to 0.3 of the Full load current is sufficient to take care of the above . Now there is fault which is less than the above setting , then the relay may not operate for such low value of fault current .
11.1 Setting criteria

There are basically the following settings required to be done in a transformer differential protection
1. Basic setting
2. Slope of the first portion of the bias characteristics
3. The end of section of the above slope.
4. The slope of the second portion of the bias characteristic
5. The setting required for the unrestrained operation ( if available in the relay )
6. The setting required to block for the magnetic inrush
7. The vector compensation setting
8. The zero sequence current filtering setting
11.2 Basic setting

This setting should cover the primary magnetizing current of the transformer and the minor expected error of the Current transformer . Normally the magnetizing current is in the range of 5% of the full load current . Suppose a 40 MVA 66/11 KV transformer with full load current of 2100 A at the 11KV side then a 105 A of current is usually sufficient to set the basic setting . Suppose a CT ratio of 2500/1 A , then the relay need to be set at a minimum of 105/2500 = 0.042 A . But as a general practice it is set at 0.2 A to avoid any other error which would otherwise be introduced .
More accurately , the no load magnetizing current can be obtained from the FAT report of the manufacturer and the setting can be decided accordingly

11.3 Bias characteristics

Before discussing the the slope of the bias characteristics, it would be interesting to know basic principle behind the bias characteristics and its requirement in the transformer differential relay . The question here is what happens to the relay if it is set like a instantaneous over current relay which operates when the differential current crosses the set level .
Suppose one of the involved CT is getting saturated partially for a through fault and because of this saturation a differential current is measured by the Relay and if it is above the setting , the relay may trip unwontedly. To avoid such an unwanted tripping , a term called biasing is introduced in the relay so that the differential setting current is raised to a new value from the basic setting during such through faults . As the effective setting of the differential current is made a variable value by the application of the bias or restraining quantity in such a way that , the maloperation of the Relay is avoided during the through fault condition
Here the differential current which tend to operate the relay is determined by the phasor difference between the primary and secondary currents ( I1‾ +‾I2) and the restraining current which tends to hold the relay from operation is determined by averaging the primary and secondary currents algebraically ( Normally I1+I2/2) . Hence in the graph is drawn with differential current in the Y axis and the restraining current( bias current ) in the X axis . As the bias current I1+I2 /2 is nothing but a measure of the actual load current , the X axis of the graph can be treated as load current itself and the Y axis the differential current .

11.4 The slope setting of the first portion of the bias characteristics

The first portion of the slope characteristics is provided to get the stable operation during the actual loading of the transformer . During an actual loading , Normally the CT correction factor is applied in the primary CT and the Secondary CT when the tap is under normal operating tap . In other words it is done when the rated primary voltage is applied at the primary resulting in to the rated secondary current . But in actual condition there are chances for the primary voltage to change from the actual value . It may be lesser than the rated value requiring the tap position to raise from the normal or it may be higher requiring the tap position to lower . In both the cases if the load is assumed to be constant , the primary current is different from that of the normal tap and the secondary current is constant as there is no change in the secondary Voltage . Hence a relay which is designed to be balanced at the normal tap will find a spill current( differential current ) under tha above condition tend to operate . As the load increases , the difference also tries to increase . Hence the differential setting shall also increase to avoid the unwanted tripping . This is done by introducing a slope characteristics in the differential setting with respect to the load . The limiting criteria for this slope can be found out by various operating conditions under maximum and minimum tap position so that the tripping is avoided during the various operating condition of the Transformer . generally 30% of the setting will cover the stability in this region . If the slope of this setting is increased more than that is required , then the sensitivity of the relay to the fault condition reduces.

11.5 End of section of the first portion of the bias characteristics

As discussed in the previous section , the slope of the first characteristics determines the stability of the relay during the normal operating range of the Transformer . Hence end of section of the first slope characteristics is determined by the normal operating range of the Transformer with some safer margin. Generally 150 to 175 % of the normal load current is taken as the end of section of the first portion of the load characteristics. Any current beyond this point is generally taken as actual internal fault or through fault .

11.6 Slope of the second portion of the bias characteristics

One of the important setting requirement for the Transformer differential protection is to avoid the tripping during a through fault condition due to the CT saturation . Hence the second portion of the bias charecteristics determines the through fault condition . theoretically , the limiting condition for this slope setting can be found out by assuming that one of the CT is completely saturated .In this case the slope will be found out to be 200% ( that is the slope calculated with only one current is present and the other CT current is not available due to complete saturation of the CT ) But in practice if the slope is set based on the above criteria, then the relay will not operate for a internal fault fed from only one side . Hence the assumption made in this case is that severe saturation in the CT is not assumed but partial saturation so that the slope can be set a more realistic value for tripping on actual faults

11.7 The setting required for the unrestrained operation .

This setting is required for the relay to operate for without the effect of the restraining function for a heavy internal fault . The limiting condition for this setting shall be normally more than the through fault current of the transformer . If the relay see a current that is more than the through fault current , then it must naturally be an in zone fault current and the tripping shall be done more quickly with out the restraing force . Generally 120% of the through fault current is a good approximation for this setting.
It shall also be noted that , generally this setting shall be above the inrush current of the transformer during energisation . The inrush current may appear to the transformer as fault current as it is available only on the winding where the no load energistaion is performed

11.8 The setting requirement to block the tripping due to magnetic inrush
The magnetic inrush current is the initial current flows to the transformer during the energizing of the transformer . This current will be normally very high and is available in the primary side ( or one side ) only and the relay will see it as differential current and trips the CB if proper care is taken to block the same . The magnitude of this inrush current depends on lot factors including the remenant flux available in the transformer core and the point of switching of the voltage . The important charecteristics of this current is that it contains second harmonic . Hence relays are provided facility to block the operation if the second harmonic component of the current is crossing a preset value . Now the modern power transformers are available with very low value as low as 5% of second harmonic component . Hence to avoid the tripping during the energisation , the second harmonic to the fundamental component shall be set accordingly
The FAT report of the particular transformer gives the available second harmonic component in the Transformer and the relay may be set accordingly

11.9 The vector compensation setting.

In a transformer there is always a phase angle difference between the primary and the secondary due to the connection of the winding .i.e. , the three phases of the primary winding can be connected in star and the three windings of the secondary can be connected in delta and similar combinations . Hence there is a phase angle difference between the primary and secondary windings current due to the above . As the relay is set for the differential current which is set to operate for the magnitude and the phase difference between Primary and secondary currents , it will mal operate and trip the breaker even if the primary and secondary currents equal in magnitude if the phases angle of the primary and secondary currents are not made same . Interposing CTs are used to introduce this vector compensation in Electromechanical Relays . However with the arrival of various numerical relays , there are software interposing CTs available inside the Relays which can be set to compensate for the vector difference between the primary and secondary based on the actual phase difference .

11.10.Zero sequence filtering

The basic requirement of the stability of the relay during the through fault is achieved based on the applied CT correction factors in the relays ( of course with proper vector compensation). By applying the proper correction factors the Current from the primary and secondary are made equal to rated current of the relay under full load condition .In other words , the primary current is the reflection of the secondary current and the secondary current is transferred to primary based on the transformation ratio of the Transformer . By selecting proper CT ratio and CT ratio correction factor the currents in the primary and secondary are made equal and applied to the Relay Hence they cancel each other at the differential element and the relay is made stable . However this stability is valid only for the three phase loading or three phase fault condition. If there is a single phase loading or single phase to earth fault in the star side of a delta – star transformer with star earthed , then we can see that , this earth fault current is not reflected to the primary based on the actual transformation ratio of the Transformer . Obviously the relay will not be stable . Hence we go for a method to make this current equal under this condition also and is called zero sequence filtering.
The word zero sequence filtering is based on the fact that , the earth fault current is zero sequence current and hence if proper methods are taken to equate the primary and secondary current fed to the relay called zero sequence filatering
Traditionally , the zero sequence filtering and vector compensation and amplitude matching is achieved by using a Auxilary CT ( matching CT )conncetd star star in delta side of the main transdormer and star delta Aux CT in Star side of the transformer with Auxilary CT delta side connected to the relay
It is important to note that , there is no phase shift between primary and secondary current during an earth fault ( except the 180 degree normal shift )in the star side of the power transformer with star earthed . ( even though there is a phase shift when there is 3 phase current present ).

In numerical relays the matching factor , the vector compensation and the zero sequence filtering is done by software methods . The method of zero sequence filteration and vector compensation differs based on the phase shift between primary and secondary.