Differential protection is a very sensitive protection scheme for large electrical machines. Ordinary current transformers (CT) are not enough for this service and we employ Class 'PS' type of CT for this. The following is a very nice excerpt on what is PS and why is it used for differential protection. This is taken from GlobalSpec forum. The full thread may be accessed here.
Normally protection CTs like 5P, 10P or 15P are used in almost all protection schemes. But, for Unit Protection Schemes like Differential, REF, etc., these CTs are not preferred. Why?
In unit protection schemes, it is very very important that the scheme operates only and only for the internal faults and must remain stable for all external faults. That is, when the unit protection scheme operates, one can be pretty sure that something is wrong within the protected equipment.
Also, unit protection schemes are employed for very critical equipment in the network. As such, whenever any unit protection scheme operates, all hell breaks loose. And one cannot put back the equipment into service, without conducting an array of tests and ensuring that the equipment is fit to be put back to service. But, this will take time and effort. And until such time, the plant will be shut down.
So, it is all the more imperative that the unit protection scheme operates only for genuine internal faults and NOT for any external faults.
Now, if we employ conventional protection class CTs like 5P or 10P for this application, let us see what happens. Lets us assume that one has selected 5P10 Class CTs for a Unit Protection Scheme. Let us say, the relay setting is 10%; this means that any differential current of 10% will operate the relay. Now, a 5P10 CT means that the CT will maintain its accuracy at least up to 10 times the rated current. This means that the CT will not saturate at least up to 10 times the rated current.
This also means that the CT may saturate anywhere after 10 times its rated current. This level will differ for different CTs. Among the same two 5P10 Class CTs, one may saturate at 12 times and the other may saturate at 13.5 times. In such a condition, during a through fault condition, there will be differential current and the relay will operate for external faults too. Even when both CTs are identically manufactured, the deterioration of its core properties over time may differ and yet they may behave differently over time.
Also, even when the CTs may be supplying to unit protection scheme of the same equipment, it is highly impossible that all the CTs of the scheme will be located at the same place. The incoming side CTs or the outgoing side CTs may have to be located far away from the relay location, thereby incurring extended lead lengths, thus imposing additional burden on the CTs. This increased burden will also shift the saturation level, as we have already seen.
Thus again, during a through fault condition, there will be differential current and the relay will operate for external faults too. There are many other similar factors contributing to the maloperation of unit protection schemes, when conventional protection class CTs are employed. Thus, it has called for a special class of CTs for such applications. That Special Class is called Class PS. (PS is the abbreviation of the French Word "Protection Speciale")
Here, instead of generalising on the minimum saturation level of the CT, the users have to exactly specify the saturation level of the CT. This is called the Knee Point Voltage (VKP), as it appears as a human-knee in the CT Magnetisation Characteristics. This specification will take into account the maximum through fault current, the actual lead burden, the relay burden & the resistance of the CT secondary winding, as also a factor of safety.
The minimum Knee Point Voltage for a given PS Class CT is calculated by:
VKP = K * I(f)s (RCT + RB), where,
If(s) = Maximum thro fault current as reflected at the CTsecondary terminals ( = If(P) / CT Ratio)
RCT = CT Secondary Winding Resistance
RB = Connected Burden, includes the relay burden & the burden of the connecting leads
K = Factor of Safety, normally taken as 2
VKP = Knee Point Voltage of the CT
As can be seen from the above formula, here the customer is specifying the level of saturation, duly taking into account the maximum possible fault current in his network, the actual burden connected to the CT, etc. If the factor of safety is taken as two, this means that at least up to two times the maximum possible fault current the CTs will not saturate. Which also means that at the maximum possible fault current, both the incoming and outgoing side CT characteristics would exactly coincide. That is, their secondary currents would match exactly and the scheme would not operate for any external fault.