Make the right choice on protective devices

Choosing the correction protective device for a low-voltage distribution system requires system performance analysis.

By David Loucks, PE, Eaton Corporation October 14, 2011

Power system engineers are faced with a wide array of protective devices from which to choose. Circuit breakers or fuses? Current-limiting or not? Would ground-fault protection be an appropriate choice even if not required by code? Regardless of the situation, every decision affects some aspect of system performance.

So, what is a designer of a low-voltage distribution system to do? Modern protective devices do a pretty good job of protecting a branch circuit. Differences tend to be related to less frequently occurring "corner cases." Some examples might include:

1. Breaker manufacturers questioning the quality of circuit protection if the incorrect fuse is installed. How often is this really going to occur? I think we can all assume that supply chain buys the correct replacement and that the service professionals installing the replacement fuses know and use the correct type and size.

2. Fuse manufacturers claiming better arc flash performance than circuit breakers on bolted fault. An arc flash does, after all, require an arc. If a circuit has no arcing connections (as a system with bolted fault current would not have), then measuring "arc flash incident energy" at that current has little meaning. The equation used in IEEE-1584 describes testing where the arcing fault currents were as little as 10% of the available bolted fault current. In low-voltage systems, real fault currents are a fraction of bolted values.

Focus your efforts

So if these are the corner cases, where should power system engineers focus their efforts?

According to IEEE 493-2007 (Gold Book) Table 10-32, well over 90% of electrical faults are arcing faults that involve ground. This tells us at least two things:

  • Arcs include impedance. IEEE-1584 testing calculated that arcing impedance was related to many factors, but for low-voltage systems, arcing current was typically in the range of 80%down to as low as 10% of bolted fault current.
  • A ground detection system might provide improved fault clearing performance.

These two factors strongly suggest that applying ground-fault protection to circuit protective devices would provide improved selectivity. This is because ground-fault current pickup settings can be more sensitive than phase settings—so sensitive that a fault current below full load current could be cleared instantaneously. 

By themselves, fuses cannot differentiate between phase and ground faults and so cannot clear a fault if the current is below the fuse rating.

By themselves, fuses cannot provide ground-fault protection except for relatively high-level ground faults. To provide ground-fault discrimination in a fused system, a system must be installed that detects ground faults and sends a trip signal to the disconnecting element. 

The IEEE Gold Book states that you are 2.5 to 70 times more likely to have a ground fault in a system than a phase-to-phase fault. Thus, the protection against ground faults is a priority, yet it is impossible to predict the actual fault current of a ground fault because you cannot predict its path with absolute certainty. You have to prepare for the maximum load, although it is very rare that you would be faced with a 100,000-amp fault. 

Furthermore, because you are up to 70 times more likely to have ground faults, the current will most often be less than 10,000 amps, maybe much less. If you follow this toward an assumption that a system must operate quickly at 10,000 amps and less, fuses simply have a more difficult time clearing those faults quickly. 

Since the fuse curve is very steep, when the fault current increases, energy let through is decreased. The problem is that the process works in reverse. If you decrease the fault current, the increase of energy let through increases dramatically. 

While perhaps not intuitive, a fault of less than 10% of your maximum current level can cause higher arc flash incident energy than a fault at full bolted fault levels. This is because as current decreases, time increases, as does arc flash danger. Circuit breakers operate somewhat differently in that once current exceeds an “instantaneous” value, the clearing time does not vary much. 

Therefore, as current drops below bolted fault levels, the time remains constant. With decreasing current and constant time, the incident energy is decreasing. Fuses don’t have a zone where they switch from time delay to instantaneous, so for fuses the incident energy increases as the current decreases. 

Fuses perform best on high magnitude, bolted faults. However, there is, by definition, no arc flash with a bolted fault. It is important to measure mechanical (magnetic) withstand capability, but this highest incident energy on a fuse protected circuit occurs at lower fault current levels.

Ground-fault protection

Additional considerations for applying ground-fault protection include:

Zone selective interlocking is recommended. While conventional selective coordination techniques can insure that the protective device closest to the fault clears first, those techniques work by adding intentional delay to the clearing time of the upstream device. 

This goes counter to our goal of using ground fault to clear the fault as rapidly as possible. One solution is to interconnect up- and downstream protective devices in a zone selective interlocking scheme. Should a ground occur within a "zone" bounded by two ZSI-equipped protective devices, the upstream device will clear without any intentional delay.

Make sure the protective device is rated to interrupt the fault current or include an automatic override. Switches used in a fused disconnect are rated to interrupt full load current, but not fault current. If a ground is detected, the relay must be inhibited from opening the disconnect if the fault current exceeds the load-interrupting rating of the switch. 

Pay particular attention to GF relays that inhibit operation at high currents. These relays, called Class II GF relays, are typically supplied with motor control. Since motor starters have very low interrupting ratings, the inhibit levels will be very low. 

Currents exceeding these low levels will be inhibited from opening the switch, and the fuse will be required to clear that fault. In many cases the fuse clearing time of a Class II GF relay application can easily exceed 10 sec clearing time. 

To work around that problem, some GF relays, rather than measuring current and inhibiting opening on high current, will simply add a time delay. The purpose of the time delay (typically 0.1 sec) is to allow sufficient time for the fuse to clear the high current fault before opening the switch. The problem is that technical papers have shown that during the 0.1 sec window where tripping is inhibited, a single-phase arcing ground fault has enough time to propagate into a 3-phase fault. At that point the phase currents are more closely “balanced” and the GF relay may no longer issue a trip signal. In this case, the ground-fault relay provided no benefit to reduce arc flash incident energy.

So be careful when calculating arc flash incident energy on low-voltage systems. Many of the standards are focused on measuring incident energy from 3-phase bolted fault current levels when, according to IEEE, most faults are single-phase arcing ground faults where fuses in particular have problems. However, at the very highest current levels, fuses reduce peak current better than most circuit breakers. 

Do your homework, ask the tough questions of your vendors, and you’ll have a safer system. 

Loucks is PCM Solution Manager for the Eaton Corporation and a Senior Member of IEEE.