Cable Insulation Level on a Delta System

Status
Not open for further replies.

FaradayFF

Senior Member
Location
California
Hello,
Is there any requirement in the Code to provide 173% insulation level for 3 phase, 3 wire feeder? If there is a single-phase ground fault on this ungrounded delta system, the phase to neutral voltage would for the two upfaulted conductors would increase by factor of 1.73. If ground fault protection is not provided and the feeder is not disconnected from service in a timely manner, the system would be subjected to prolonged overvoltage condition which would likely damage the insulation and has the potential to become a L-L fault. Besides a good practice and requirement in the Code to provide conductor protection, are there any specific callouts on this in the code book?
Thank you,
EE
 
see below excerpted from anixter webpage.

"Low-Voltage Cables
For cables rated 0 through 2,000 volts, the in-service voltage stress on the insulation is so low that the concept of insulation levels is largely unnecessary. For example, the operating voltage stress on a typical 600-volt cable is about 5 volts per mil of insulation thickness, i.e., each mil (0.001 inch) of insulation must withstand only 5 volts of electrical stress. On the other hand, the in-service voltage stress on a typical 15 kV cable is about 50 volts per mil or 10 times more than that for a 600-volt cable. 600-Volt rated cables have much thicker insulation per volt of applied electrical stress and are thus over insulated from a voltage stress point of view. As a result, the cable insulation thickness specified by industry standards for use in a grounded (100 percent) electrical system is also acceptable for use on an ungrounded (133 percent) electrical system.



There may be some applications that require 173 percent insulation levels for low-voltage systems such as high-resistant grounding systems. 173 percent insulation levels are not addressed by most industry standards. However, NEMA and ICEA standards such as ICEA S-95-658-1999 Standard for Nonshielded Power Cables Rated 2000 V or Less for the Distribution of Electrical Energy recommends the use of a cable rated at least 1.73 times the phase-to-phase system voltage. For example, a 480-volt system would require the use of a cable rated at least 830 volts. Because of commercial availability, a 1 kV rated cable is typically used for this type of application in Canada and a 2 kV cable would be used in the United States."
XHH

You should be able to find 1kV rated XHHW-2 insulated cable.
 
if you are doing an industrial plant design, its not uncommon for engineer's to specify XHHW for low voltage systems, as its a more robust cable relative to THHN/THWN cable, so it wouldn't be that unusual for you to specify XHHW if you are still concerned about the voltage issue during a fault.
 
if you are doing an industrial plant design, its not uncommon for engineer's to specify XHHW for low voltage systems, as its a more robust cable relative to THHN/THWN cable, so it wouldn't be that unusual for you to specify XHHW if you are still concerned about the voltage issue during a fault.

Nice Avi. 345-115kv 400MVA?
 
Hello,
Is there any requirement in the Code to provide 173% insulation level for 3 phase, 3 wire feeder? If there is a single-phase ground fault on this ungrounded delta system, the phase to neutral voltage would for the two upfaulted conductors would increase by factor of 1.73. If ground fault protection is not provided and the feeder is not disconnected from service in a timely manner, the system would be subjected to prolonged overvoltage condition which would likely damage the insulation and has the potential to become a L-L fault. Besides a good practice and requirement in the Code to provide conductor protection, are there any specific callouts on this in the code book?
Thank you,
EE

No in general. For several reasons.

First it only exists that way at the fault and it’s short term. Don’t forget insulation is tested at 200% of voltage class plus 1 kV. So a 600 V rated cable is tested at 2200 V. A 5 kV cable is tested at 11 kV. 300 V insulation if you can find it is tested at 1600 V. So even though it would be illegal for ungrounded delta it can still work.

Second at low voltage (under 1 kV) for delta or resistance grounded you must run line-line rated cable but on solidly grounded wye lower voltage is allowed. So theoretically that’s 600 and 300 V cables for most installations. But the cost difference with 300 V cable is so small that it is usually more expensive due to lack of demand. So most installations are 600 V cable for common voltages.

In medium voltage shielded cables are where insulation level applies. Cable manufacturers only make shielded cable for 5 kV and up. This ignores “temporary” hookup wire and motor lead wire for instance that are not Listed for NEC purposes, only as components of assemblies. The thing is that a 5 kV shielded cable actually only sees 2.4 kV line to ground so the manufacturers “cheat” knowing this fact. This is fine with a solidly grounded wye. So this is 100% insulation level. But with delta or resistance grounding you have to add extra. The insulation stress is radial so you don’t actually need 173%. Only 133% is needed. There’s a chart that says that if the fault can remain indefinitely use 133%. Intermediate levels can be used if you clear the fault quickly. But again we hit a demand issue so in practice really only the 100% and 133% levels are used.

Getting back to assemblies inside panels you see crazy stuff. The most popular unshielded MV cable for years was Exlar brand although there are competitors. You can get voltage stress and tracking damage if Exlar cable gets too close to another phase or a grounded metal surface. So extensive use of insulators and GP3 fiberglass (Glastic brand) and lots of tie wraps are used. And since air is insulation you can mix the two. So since anything above 15 kV is made to order often you see 15 kV cable on 25 KV or 35 KV insulators in 25 or 35 KV class switchgear. It’s not a big deal if you know what’s going on and take steps to ensure cables don’t touch each other or grounded surfaces. You have to do it anyway. In practice what I usually see is the air gaps are equal to worst case phase to phase distances and the cable may as well
Be bare so the insulation just makes it look pretty. This is at 15 kV and up. At 5 and 8 KV quite often anything goes. But after ten years or so it becomes very obvious who knows what they are doing and who doesn’t. This applies to both installers and manufacturers. Some of the biggest names (GE) are in my example photos.

Finally let’s be clear about why you do NOT want to use ungrounded deltas. You cannot insulate to a high enough level. You can insulate at 1000% and it’s not enough. Ungrounded delta just shouldn’t be used except at high voltages (69 kv+) where arcing faults are rare.


Instead for 250-8 kV use a high resistance ground. It is the cheapest ground to install and maintain. It controls transients and can still maintain power during faults like ungrounded, plus it is easy to detect and remove ground faults unlike ungrounded.,above 10 kV the system charging current gets excessive. In that case a low resistance (400 A) grounding system has most of the advantages of high resistance grounding but a smaller resistor. The resistor is usually 10 seconds rated so we must trip quickly.
 
In medium voltage shielded cables are where insulation level applies. Cable manufacturers only make shielded cable for 5 kV and up. This ignores “temporary” hookup wire and motor lead wire for instance that are not Listed for NEC purposes, only as components of assemblies. The thing is that a 5 kV shielded cable actually only sees 2.4 kV line to ground so the manufacturers “cheat” knowing this fact. This is fine with a solidly grounded wye. So this is 100% insulation level. But with delta or resistance grounding you have to add extra. The insulation stress is radial so you don’t actually need 173%. Only 133% is needed. There’s a chart that says that if the fault can remain indefinitely use 133%. Intermediate levels can be used if you clear the fault quickly. But again we hit a demand issue so in practice really only the 100% and 133% levels are used.

I believe now it's if the ground fault can be cleared within 1 minute you use 100% rated cable, if the ground fault will be cleared within 1 hour you use 133% rated cable, and if the ground fault could remain indefinitely you would use 173% rated cable. Voltage rating is based on phase-to-phase voltage. Looking at my Southwire book, my Okonite book, and IEEE 525-2016.
 
Finally let’s be clear about why you do NOT want to use ungrounded deltas. You cannot insulate to a high enough level. You can insulate at 1000% and it’s not enough. Ungrounded delta just shouldn’t be used except at high voltages (69 kv+) where arcing faults are rare.


This is an interesting white paper, and not something I'd thought about before. I may have overlooked it in the paper, but under what conditions does an arcing fault develop with a single grounded conductor on a ungrounded system? I see the example is a system with a 1 A charging current (10 uf capacitance), but I'm wondering how realistic that is.
 
This is an interesting white paper, and not something I'd thought about before. I may have overlooked it in the paper, but under what conditions does an arcing fault develop with a single grounded conductor on a ungrounded system? I see the example is a system with a 1 A charging current (10 uf capacitance), but I'm wondering how realistic that is.

Very realistic That’s not much in terms of transformers, motors, and shielded cable. In MV applications typically 1-2 A is on the low end. At 480 you’re typically close to that. GE has a nice estimating chart that I’ve used for years that is reasonably accurate. See page 4.


It doesn’t really matter what the inputs are. It occurs over a wide range of inputs. They just gave an example. The thing is it’s hard to detect because arcing faults tend to come and go and the equipment damage it causes can happen anywhere on the system but typically afflicts motors because impulse resistance is so much lower. A standard low voltage (under 600 V) NEMA motor must pass 1000-1200 V hipot. An inverter duty is 1400-1800 V. In contrast #14 THHN with a mere 15 mils of PVC tests up to 2800+ V according to NEMA tests. US Navy reports roughly 4 times motor failure rates. They require ungrounded deltas by Code due to corrosion concerns. So if an arcing fault develops it quickly builds up an arcing voltage then discharges itself somewhere else via an innocent motor.

Keep in mind also there are statistics among the safety folks claiming somewhere around 90%+ of faults start as single line to ground arcing faults. I kind of doubt these numbers but calculating it from IEEE 493 I get around 40-80% depending on equipment type. And I think it is grossly under reported because most people will just see a three phase fault which is true after it starts arcing. So although I don’t have better numbers and I found nothing backing up the 90% claims it makes intuitive sense.

This effect does not happen with bolted faults but generally speaking bolted faults are typically from actions such as incorrect wiring. Arcing faults are generally due to equipment failures so we expect them more often. And the research on arcing faults shows regardless of how they start, they normally evolve into three phase arcing faults within 1-2 cycles for enclosed gear.

These numbers have been proven out to the point that FM Global, a very popular business loss insurance carrier, all but mandates converting ungrounded deltas to high resistance grounds in their standards.

HRG is simple, cheap, and gives you all the advantages of both the ungrounded and solidly grounded systems with none of the disadvantages. It is mandatory in many EU countries, underground and coal mines in North America, and used extensively in many high reliability plants since the 1970s. It is virtually a de facto standard for 4160 and 7200 V in modern installations where grounded systems pretty much destroy a cubicle if a fault occurs. With the exception of some instrumentation such as PTs for switchgear I don’t see any reason for ungrounded deltas today except ignorance.
 
That's good info. I see ungrounded delta systems in some of the facilities I work for, which were installed to support critical systems that need to operate to failure. I doubt they will transition any of the existing equipment, but this gives me more ammunition for moving to grounded systems where it makes sense going forward.
 
Status
Not open for further replies.
Top