Replacing 1 4000:5 CT with 2 2000:2.5 CTs in Parallel. Are different Relay Classes a consern?

[TWeber]

This question pertains to the usage of relay class CTs for protection and metering in low voltage switchgear.

In the case of using a double wide breaker (meaning 2 connections per phase) typically the two split legs of a phase are bussed together in order to run through a single CT. For instance, for a double wide 4000A breaker the two legs (each carrying 2000A) are bussed together to run through a 4000:5 C200 CT before feeding a circuit.

I would like to know what are the ramifications of using a two 2000:2.5 C100 CT and wiring them in parallel instead to running bus between the spit phases. Continuous thermal rating at 30C is 1.0 and for B0.1 and B0.2 accuracy is 0.3.

There is concern from the customer and some engineers that since a 4000:5 C200 is typically used, lowering the class to C100 will reduce the saturation point of the CT, reducing protection. Theoretically the current should be split (with minor resistance differences), so class can be divided by 2.

Can someone enlighten me? Is there any literature or standard that addresses this? Is 1 4000:5 C200 CT equal to 2 2000:2.5 CT wired in parallel on a split branch circuit?

 

[collies99]

On old 25kv GE KSO oil CB, we series (liked ratio) to effectively double the vsat capability of the CTs to drive high burden CO relays, one is not doable nybitself during faults.

Without knowing for certain on your configuration, paralleling CTs should just give you the total current for the phase. Since voltage sources are in parallel, the connected burden at fault levels must be within the lowest CT vsat rating or total RMS current will be inaccurate.

 

[TWeber]

Clarification on the configuration.
The breaker has 6 connections on the line side and 6 connections on the load side. The breaker is rated for 4kA. The phase currents are split in half, so A1-B1-C1 on left side of breaker and A2-B2-C2 on right side.

Let's say it's a main/feeder. Each leg is fed 2000A for a total of 4000 incoming on the line side per phase. On the load side for meetering/protection we bus A1 and A2 together and run it through a CT before cabling out 4000, and the same with B and C.

What I am proposing is instead of busing 2000 from A1 and 2000 from A2 together with copper, use a 2x 2000:2.5 on their respective sides and wire them in parellel back to the relay. This eliminates the need for the copper connection.

The question arises from the Class difference. Would a 2 C100 used on a split current acceptably replace one C200 on the whole current.

 

[collies99]

That is the doubtfull answer.

If you treat each CT as a voltage source with vsat as max voltage, your total voltage in parallel is basically vsat.

If you place 2 CTs is series, with ratio being equal, your total voktage in series is vsat1 + vsat2.

Therefore in your CT parallel configuration you still need to generate Isec fault x burden deliverable by each CTs since produce 1/2 of isec.

There have been discussion here on lower vsat CT being used as as acceptable for 50 elements if detectable by some form of peak current algorithm, which can possibly work in your case, however if time base 51 needs to operate, the proper RMS and for that matter the proper vsat rating would be required.

 

[rcw retired EE]

You will have better luck getting an accurate C200 2000:5 CT than a non-standard 2000:2.5 C100 CT. If I was asked for a 2000:2.5 CT, I would supply a 4000:5 and change the label. Since you only have one primary turn (the busbar) the CT will need 4000/5= 800 turns of secondary wire whether it is 4000:5 or 2000:2.5. You won't save any costs.

You could assume that the current division between the two busses was perfect, take two 2000:5 C100 CT's and wire the secondaries in series to get close to the performance of the 4000:5 C200. The CT's would share the voltage requirements under fault conditions. BUT, if the currents in the two busses ever got unbalanced the CT's would probably fail as they fought to make the secondary current match their primary bus current. If they do survive, the current to the relaying would be inaccurate.

If the price of six 4000:5 C200 CT's is cheaper than your bus work, then using two in parallel may be a viable option. But don't try to use two C100's unless you have full information on the customer's circuit and are willing to take the liability if the 4,000 A breaker doesn't trip on a 50,000A fault, that destroys equipment and possibly injures or kills someone.

 

[davidbeach]

Two 2000:5 in parallel, not series, and not 2000:2.5. You can only put CTs in parallel if the primaries are also in parallel (measuring the exact same current). These might look to be in parallel, but they're measuring different, though similar, currents.

 

[TWeber]

See attached picture for arrangement clarification. The picture is a top view of the breaker, there would be another set of stabs below.

I have a source for the 2000:2.5 C100 CTs. The costs I will save is in copper bus bar.

 

[rcw retired EE]

If the original design or specification requires three 4000:5, C200 CT's, your parallel secondary design will require six 2000:2.5 C200 CT's. C100's will not work if C200 is needed.

As a mechnical analogy, the original design uses a pump that can deliver 100 gpm at 200 psi and you want to replace it with two pumps that deliver 50 gpm each at only 100 psi into the same piping system. The parallel pumps will not be able to push enough flow if the piping system needs the 200 psi to get to 100 gpm.

Also, how are you saving any costs of bus work? You still have to connect the two A phase busses together to connect at the load terminals or the switchgear bus. Are you planning on having your client connect 1/2 his cables to each bus?

 

[TWeber]

Thank you rcwilson. That makes a lot of sense now. You are correct on the first paragraph, that was my intention.

Cable in 2000A to 2 seperate line side lugs landings of the breaker. Load side to horizonal cross bus which distribues to small 800-2000 feeders. With a 4000:5 we need to bring the 2 side Phases together before spitting again into the cross bus.

 

[jghrist]

I think you should use a 4000:5 CT on each bushing. Let's say you have total of 40000A fault current and a burden of 4 ohms. With a single 4000:5 (CTR=800) CT, you would get 40000/800 = 50A secondary current and a voltage of 50*4 = 200V. With a 4000:5 CT on each bushing, you would get 20000/800 = 25A secondary on each which will add together to get 50A. You would get 50*4 = 200V through the burden, which would be impressed across both CTs, so you would still need C200 class.

This ignores the fact that some of the burden is the CT winding resistance and that each parallel CT winding will only see 25A. You would actually need somewhat less than a C200.