Overcurrent protects the capacitor bank from large faults, but it does not see arcing or small internal faults. When a grounded-wye bank is energized, the charging current can approach the ground trip level and cause false trips. IEEE C37.99-2000 also describes breaker failure for the bank, with tripping of neighboring breakers after a delay, and the role of surge arresters.
Short answer
Overcurrent covers the large faults of a capacitor bank: phase to phase or phase to ground. It does not cover well the small faults inside the bank or arcing, which are the job of unbalance protection. When the bank is energized, the charging current can look like a ground fault. If the bank breaker does not open, breaker failure protection (50BF) isolates the bank by tripping the neighboring breakers.
IEEE C37.99-2000 deals with these points in clauses 7.2.1, 7.2.3, 7.2.6 and 7.2.7.
What overcurrent covers and what it does not
The standard separates two types of fault:
- Major fault, such as phase to phase or phase to ground. It requires protection external to the bank: power fuses, breakers or circuit switchers with their relays. For the best protection, the relays are set as low and as fast as possible, with just enough delay not to trip on external system disturbances.
- External arcing within the bank. Overcurrent is usually not sensitive to an arc in the bank and, if it detects it, it is normally very slow. Unbalance protection governs there. In large banks, the standard suggests redundant unbalance protection and an unbalance current or negative-sequence function.
In large extra-high-voltage banks, the standard shows two sets of three-phase 50/51 overcurrent relays, with instantaneous and short-time functions, as redundant protection against short circuits inside the bank.
The arc that overcurrent does not see is covered in external arcing and rack faults; the voltage protection that goes with it, in bank voltage protection (59, 59B and 27B).
Overcurrent points according to the standard
| Topic | IEEE C37.99-2000 criterion |
|---|---|
| Minimum pickup of inverse time | Desirable 135 % of rated phase current in grounded-wye banks and 125 % in ungrounded banks |
| Instantaneous | If used, it must be set high to ride through inrush and outrush transients |
| Relays with fundamental filter | Modern relays with a bandpass filter at fundamental frequency are not sensitive to inrush or outrush |
| Neutral 51N relay with low burden | Reduces the high voltages at the current transformer secondary caused by high-frequency outrush current during an external fault |
| System not effectively grounded and ungrounded bank | The 51N is set to a low tap to detect and clear low-magnitude ground faults fast |
| Effectively grounded system and grounded-wye bank | The outrush current to an external fault does not usually operate the 51N, but the unbalanced bank load from that fault can make it pick up if set too low. Its trip is set above the capacitor phase current |
These are references from the standard, not the setting of your relay. The real setting depends on your bank, your current transformers and your network.
Energization: the current that looks like a fault
Each time a large grounded-wye bank is energized, the momentary charging currents of the capacitors, in one phase and in the neutral, can approach the ground fault trip level. When a bank is already energized in parallel, those currents can be in the order of thousands of amperes.
The standard lists the possible consequences:
- Spurious relay operations.
- Relay failures.
- Current transformer failures.
- Energized substation fences.
- Problems in the ground grid.
To limit the problem, the standard asks that the neutral connection between several banks not run through the ground grid. A single-point or peninsula ground should be used. The same transients also reach the current transformers and the control cables.
That current and its limits are in switching a capacitor bank, and its effect on CTs and relays, in switching transients in CTs, VTs and control circuits.
Bank breaker failure
If the bank switching device is a breaker or a circuit switcher, the standard contemplates a breaker failure scheme as local backup. If the device fails, the bank is isolated by tripping, after a delay, the adjacent breakers connected to it. In some cases a remote trip may be needed.
- Fault in the bank. The protective relay orders the bank breaker to trip
- The breaker does not open. Current continues and an overcurrent element confirms it
- Breaker failure starts. An auxiliary relay starts it and a delay runs
- Adjacent breakers trip. The bank is isolated from the system
- Remote trip if needed. Order to the remote end when local isolation is not enough
Two precautions from the standard:
- Breaker failure detection is initiated with auxiliary relays. The usual schemes supervise the timer start with an overcurrent element, or supervise the timer output.
- Those overcurrent elements must operate for all types of bank fault, including currents lower than rated. A phase level of 50 % of the normal rated bank current is generally considered adequate in grounded and ungrounded banks.
The standard also asks you to check the capacitive current switching capability of the adjacent breakers that will trip. That capability is explained in classes C1 and C2.
Surge arresters
Lightning and switching overvoltages are controlled with standard surge arresters. The bank absorbs overvoltages because, for sudden voltage changes, it behaves for an instant like a short circuit. Overvoltages around the bank are greatly reduced, but the protection is not complete. The overvoltage on the bank depends on the line length between the bank and the point where the transient is generated, and on the duration of the surge. More on the equipment in surge arrester.
Common mistakes
- Expecting overcurrent to detect an arc inside the bank. Unbalance is there for that.
- Setting the 51N too low in a grounded bank. The unbalanced load from an external fault makes it trip without a fault in the bank.
- Joining the neutrals of several banks through the grid. It increases transients in the grid and in the secondaries.
- Not checking the adjacent breakers. The breaker failure scheme depends on their ability to switch capacitive currents.
What we do at Tiza Group
In our remote relay configuration service we review the overcurrent, the 51N and the breaker failure protection of your bank and their coordination with the unbalance protection. We work remotely, worldwide. Contact us and we will quote the review.
Related guides
- IEEE C37.99: shunt capacitor bank protection guide
- Capacitor bank voltage protection: 59, 59B and 27B
- Bank transients in CTs, VTs and control circuits
- Switching a bank: breaker, inrush and surge arresters
- Breaker failure 50BF (RBRF): what it is and how it works
References
- IEEE Std C37.99-2000, IEEE Guide for the Protection of Shunt Capacitor Banks, clauses 7.2.1, 7.2.3, 7.2.6 and 7.2.7, Figure 6.
Frequently asked questions
Why does overcurrent not protect a capacitor bank from an internal arc?
Because it is not sensitive to an arc inside the bank and, if it detects anything, it does so late. IEEE C37.99-2000 assigns that task to unbalance protection and, in large banks, to redundant unbalance protection or to a negative-sequence current function.
Why does ground protection trip when a grounded-wye bank is energized?
Because the charging current of the capacitors, in one phase and in the neutral, can approach the ground fault trip level for an instant. If another bank is already energized in parallel, the current can be in the order of thousands of amperes.
What happens to the neighboring breakers when the bank breaker fails?
If the bank breaker does not open, breaker failure protection trips the adjacent breakers that feed the bank, after a delay. In some cases a remote trip is required. The standard reminds you to verify that those breakers can handle capacitive current switching.