Surge Arrestor

i thought that the surge arrestors have to be rated substantially higher than the peak voltage of the system? In your case >47kV ? i am no expert in MV and HV systems, and my experience is limited to using MOVs (Metal-oxide varistors) in LV systems. (What is MCOV?) The clamping voltage is usually >2xRMS system voltage. Here for example the system voltage of 115V, one uses the varistor which starts conducting at 180V and clamps below 300V.

I'll try and answer both questions. MCOV is Maximum Continuous Operating Voltage that defines the maximum continuous rms phase to ground voltage applied across the arrester terminals. For effectively grounded systems this is normally the maximum system phase (line) to ground voltage. For ungrounded or impedance-grounded systems the MCOV should be at least 90% of the maximum phase-to-phase operating voltage. Insulation coordination is important to ensure that the arrester provides adequate protective margins for the equipment it is intended to protect. That equipment will be subjected to the arrester discharge voltage plus the voltage drop of line and ground leads during arrester discharge to ground of impulse overvoltages. Another consideration is the arrester TOV or Temporary Overvoltage capability, which defines the time duration in seconds that the arrester can withstand a temporary power frequency overvoltage above its MCOV. It's important to realize that surge arresters are designed to safely discharge impulse overvoltages to ground; not power frequency.

Thank you. You have certainly improved my knowledge. So, just to be sure, the surge arrestor in this system should be ≥90% x 33 kV? Or since the phase-to-ground is 33/√3 = 19.05 and peak of that is x √2 = 27kV ? Much obliged if you could clarify, thanks in advance.

That's fundamentally correct but the rating is based on RMS voltage; not peak or crest but again, it depends upon how the system is grounded and whether the 33 kV is nominal or maximum. That relationship is dictated by system voltage regulation. For distribution systems in the USA that is normally 5%. In other words, nominal x 1.05 = maximum.

Other considerations in the selection of gapless polymer housed distribution class surge arresters include the duty class (Normal, Heavy, Riser Pole), which defines the fault current withstand and duty cycle characteristics.

You might take a look at the DynaVar® surge arrester catalogs and other information available at www.hubbellpowersystems.com. I think you'll find it useful.

Why would you choose an arrestor at below MCOV?

A "first cut" from the comments is that :

1. You need to tell the VCB supplier how your system is earthed and its power rating (supply voltage limits and impedance). It is then his responsibility to ensure the arrestors protect his VCB panels adequately and safely (you are paying him for the arrestors and their proper installation "for lightning surge protection").
2. You should look at surge voltage limits of your equipment connected to VCB. Ask the panel supplier for the peak let-through voltage surge of his proposed arrestors, telling him that you wish to check if his arrestors protect your own equipment.
3. For your information. The arrestor peak breakdown voltage ( not the same as its peak continuous service voltage rating) matters for the protection of your equipment, but the service rms voltage matters for the recovery of the arrestor from a surge, as does the arrestor current. Once "fired", the arrestor must be able to "choke-off" power-frequency currents and stop the discharge in its gaps.

Your advice is sound except for what happens when the arrester "fires". Since the arrester, in this case, is polymer-housed, it is most likely gapless MOV (metal-oxide varistor). In that case when the arrester is subjected to an overvoltage it goes into conduction (from capacitive to totally resistive), discharges the surge to ground and returns to normal. Older silicon carbide distribution arresters have gaps in series with the blocks such that the blocks are not subjected to full line to ground voltage except during discharge. If there is no current-limiting capability in the gap structure, the arrester must rely primarily on a zero crossing power frequency voltage in order to reseal against system voltage. That's the only "choke-off" feature available to the arrester.

The total duration of a typical lightning stroke is less than 1000 microseconds, including restrikes. The time to crest of 1/4 cycle of 60 Hz power frequency is 4,167 microseconds. Thus, the arrester normally discharges lightning surges to ground in less than 1/4 cycle. If the arrester fails or is unable to reseal, a circuit breaker or recloser operation is required to clear the power frequency fault.

Thanks for information, Bluestone, my knowledge on HV kit is "old technology". I did not know MOVs had grown from LV to serious HV!