Safety of Using Adiabatic Method?

[Cookie]

I'm thinking, and while I could be wrong, is the adiabatic method a possible latent safety hazard? As well as undersized CPCs on circuits 32 amps and under?

Normally DNO sources are strong with relatively high PFC, and we have a known Ze value. However, should someone bring a portable generation unit on site after an outage, or a large yet short circuit limited inverter (Tesla Power Ball), Ze values can be such that MCB opening time increases. Such an increase can overheat the CPC, up to the point of melting it.

From a touch voltage perspective I humbly do not see a major concern however, in that 20 x 10 = 200 amps, 1.15 ohms or 46kw- a 46kw load on a 25kw generator would result in a significant drop in the output voltage whereby 230 volts dropping to 110 volts giving 55 volts to remote earth such that a longer disconnection time will not result in organic bodily injury.

However, the risk of fire and shock from an overheating CPC is relevant.

What do others make of this point of view?

 

[pc1966]

Many generators have RCD on the output for this sort of reason, as well as protection for socket outlets against shock. Typically don't have the PFC in reality to trip things (even if they measure as "stiff" due to electronic regulation on they usual style of impedance meter).

(We are looking at a generator in due course, probably 10-20kVA 3-phase and probably will ask the supplier for a delay 100mA RCD and hard-wired terminals instead of socket and 30mA "instant" as it is for a system with a lot of electronic leakage.)

 

[Lucien Nunes]

In theory, before the supply characteristics are changed, the suitability of the protective measures in the installation should be confirmed as suitable for the new supply. I expect this is rarely done for small installations, not least because the impedance of a small generator in the tens of kVA is usually difficult to define, over the ADS tripping time under consideration.

A rule of thumb sometimes applied to synchronous generators in this size range is PFC = 7*FLC but there are too many variables to allow this simple figure to be used to assess the suitability of edge-case ADS design. For example, where in the cycle the fault occurs (DC component contribution); what load is on the machine at the time (prime mover instantaneous power availability); load angle (demagnetisation effect of fault current); AVR characteristics etc. Due to the relatively high R/Z of a machine of this size, the pf of the fault current is unfavourable and it will both demagnetise the machine and mechanically arrest it, resulting in a very broad range of current/time curves.

For all the above reasons, it is normal to employ overall RCD protection in this scenario.

 

[Cookie]

Many generators have RCD on the output for this sort of reason, as well as protection for socket outlets against shock. Typically don't have the PFC in reality to trip things (even if they measure as "stiff" due to electronic regulation on they usual style of impedance meter).

(We are looking at a generator in due course, probably 10-20kVA 3-phase and probably will ask the supplier for a delay 100mA RCD and hard-wired terminals instead of socket and 30mA "instant" as it is for a system with a lot of electronic leakage.)

Is it required by BS7671?

Me personally I really do not want leakage current to trip an RCD supplying power to fire alarms, emergency lighting or even medical equipment.

 

[Cookie]

In theory, before the supply characteristics are changed, the suitability of the protective measures in the installation should be confirmed as suitable for the new supply. I expect this is rarely done for small installations, not least because the impedance of a small generator in the tens of kVA is usually difficult to define, over the ADS tripping time under consideration.

A rule of thumb sometimes applied to synchronous generators in this size range is PFC = 7*FLC but there are too many variables to allow this simple figure to be used to assess the suitability of edge-case ADS design. For example, where in the cycle the fault occurs (DC component contribution); what load is on the machine at the time (prime mover instantaneous power availability); load angle (demagnetisation effect of fault current); AVR characteristics etc. Due to the relatively high R/Z of a machine of this size, the pf of the fault current is unfavourable and it will both demagnetise the machine and mechanically arrest it, resulting in a very broad range of current/time curves.

For all the above reasons, it is normal to employ overall RCD protection in this scenario.

Right, but who does this in an emergency? Supermarket, data center, school turned into a shelter- during a storm the owner asks that a gen be wheeled and hooked up to the mains supply

Would you know of any typical values for units in the 10 to 100kw range? What about 500-1000kw?

From what I've read a single phase to ground fault is capable of producing 10x the FLC for one second, then that value drops to 3x for 9 seconds thereafter. But this is really thin on my part.

 

[Strima]

Most back up supplies that are brought in are normally only for essential circuits, such as Supermarkets should only have support for fridge's and freezers, basic lighting circuits etc. These curcuits should be designed to accommodate this.

Data centres tend to use a UPS to support essential circuits until the generator comes on full load from a dark start, roughly 2 minutes for most places. The loads are fed through the UPS which normally limits any fault currents as part of it design.

 

[Cookie]

Most back up supplies that are brought in are normally only for essential circuits, such as Supermarkets should only have support for fridge's and freezers, basic lighting circuits etc. These curcuits should be designed to accommodate this.

Data centres tend to use a UPS to support essential circuits until the generator comes on full load from a dark start, roughly 2 minutes for most places. The loads are fed through the UPS which normally limits any fault currents as part of it design.

Right, but you've never seen a mobile gen come in or a permanently installed one put in place latter?

 

[Strima]

Right, but you've never seen a mobile gen come in or a permanently installed one put in place latter?

Normally the on site generators are short term, I know data centres work on a four hour window. If the outage exceeds this then external generators are brought in, these plug into the site generator point which would then be running the essential circuits as I mentioned earlier.

As Lucien pointed out generators can be problematic and the installation should be assessed before they're connected.

 

[Cookie]

Normally the on site generators are short term, I know data centres work on a four hour window. If the outage exceeds this then external generators are brought in, these plug into the site generator point which would then be running the essential circuits as I mentioned earlier.

As Lucien pointed out generators can be problematic and the installation should be assessed before they're connected.

Understood, but where I'm from no one really assesses. Certainly not a home owner who gets a 10 or 15kw portable unit hooked up to his home after a storm.

Commercial places will call a gen rental company and they will hook up to the main switch board without going through every circuit.

 

[DefyG]

but where I'm from no one really assesses.

They do but you may be right in that home owners don't!

 

[Cookie]

They do but you may be right in that home owners don't!

Which is a latent hazard that BS7671 should address IMO. I can understand the adiabatic method being a necessity back in the day when materials were more costly, but today I think it largely has more cons than pros. But I could be wrong here.

 

[Lucien Nunes]

Without detracting from the valid theoretical point Cookie makes, the real-life risk intuitively seems very small. Very few systems have their CPCs designed down to the minimum, very few systems are ever hooked up to a generator, and faults cleared by ADS are rare. I wonder how often the three coincide, and indeed whether genuine thermal damage will typically result even if they do?

FWIW I have never seen a CPC of any size thermally damaged other than by stray welding current and by a failed CNE, both of which are recognised causes.

 

[pc1966]

Is it required by BS7671?

RCD on sockets less then 63A is.

General RCD use on generators is not.

Me personally I really do not want leakage current to trip an RCD supplying power to fire alarms, emergency lighting or even medical equipment.

For bigger systems where that is a possible problem then you may have the MCCB style RCD addition with adjustable trip time and current. It could be as high as 5A!

That is not "stray leakage", that is a real fault! Of course, it also requires your down-steam circuits to mostly have RCDs with shorter trip times and lower thresholds, but if you are sizing backup for critical systems that ought to be part of the consideration.

As for adiabatic risk, I suspect in the real world many smaller generators would simply stall if shorted (and no fast clearing RCD) so you have some energy-limiting that way.

 

[Cookie]

Without detracting from the valid theoretical point Cookie makes, the real-life risk intuitively seems very small. Very few systems have their CPCs designed down to the minimum, very few systems are ever hooked up to a generator, and faults cleared by ADS are rare. I wonder how often the three coincide, and indeed whether genuine thermal damage will typically result even if they do?

FWIW I have never seen a CPC of any size thermally damaged other than by stray welding current and by a failed CNE, both of which are recognised causes.

You make a point which can not be ignored, not many cases are documented in the real world.

However Table 54.7 in BS7671 still stands out at me. While the CPC values are enormous vs the NEC, I can not help but think of the electrical theory behind it.

 

[pc1966]

However Table 54.7 in BS7671 still stands out at me. While the CPC values are enormous vs the NEC, I can not help but think of the electrical theory behind it.

Table 54.7 is the generic "half phase size, if that is above 16mm, or phase size" rule for cases then you can't use the adiabatic approach (for whatever reason, OCPD unknown or just too much trouble).

In itself is really just a relaxation of the fairly obvious case that is a cable has suitable protection for L-N fault/overload, then the same CPC is fine, and going L-E is (I would hope!) always a fault condition and not overload sort of durations so it can be relaxed a little.

Of course the TN-C-S "open PEN" fault is a clear exception to these assumption hence the additional regs in Table 54.8 that set a lower limit to CPC size so a 6mm (or below) sub-main cable needs additional CPC if bonding to extraneous conductive parts under these conditions.

 

[Cookie]

Table 54.7 is the generic "half phase size, if that is above 16mm, or phase size" rule for cases then you can't use the adiabatic approach (for whatever reason, OCPD unknown or just too much trouble).

In itself is really just a relaxation of the fairly obvious case that is a cable has suitable protection for L-N fault/overload, then the same CPC is fine, and going L-E is (I would hope!) always a fault condition and not overload sort of durations so it can be relaxed a little.

Of course the TN-C-S "open PEN" fault is a clear exception to these assumption hence the additional regs in Table 54.8 that set a lower limit to CPC size so a 6mm (or below) sub-main cable needs additional CPC if bonding to extraneous conductive parts under these conditions.

Agree. The thing is, and its being argued in US forums, that Ze can change over time for several reasons and such device clearing time can change. With a full/half size CPC there is no worry of shock or fire should tripping time increase.