Negative-sequence protection (ANSI 46) — why phase unbalance heats a motor faster than the current alone suggests

The guide on motor overload classes (10/20/30) covers thermal overload protection that reacts to the total, measured phase current. This article covers an additional protection function that does not look at the total current, but specifically at the unbalance between the three phase currents: negative-sequence protection, designated by ANSI code 46, also referred to as phase-unbalance protection.

What negative-sequence current is and why it is so damaging

Any set of three-phase currents can be mathematically decomposed into three symmetrical components: a positive-sequence component (the normal, correctly rotating part), a negative-sequence component (a part that rotates in the opposite direction), and a zero-sequence component. Under a fully balanced, healthy three-phase load, the negative-sequence current (I2) is negligibly small. If unbalance occurs — due to unequal supply voltage, a loose or broken phase, an unevenly loaded single-phase load on a three-phase supply, or an internal fault — a measurable I2 appears, even though the total, average phase current may still remain within normal limits.

For a motor, this negative-sequence current is particularly damaging: because I2 rotates in the direction opposite to the rotor, it induces currents in the rotor at a frequency of roughly twice the supply frequency. Due to the skin effect, these currents concentrate near the surface of the rotor bars and end rings, causing intense, localized heating there. As a result, a relatively small negative-sequence current — often just a few percent of rated current — causes a much larger temperature rise than the same percentage of additional symmetrical overload would: the thermal loading of the rotor increases roughly with the square of I2, not linearly.

Why an ordinary overcurrent or thermal protection can miss this

With a broken phase (for example a loose terminal connection or a blown fuse in one phase), the current in the remaining two phases may rise, but the total, three-phase-averaged current may still stay below the pickup threshold of an ordinary thermal overload protection, while the negative-sequence component — and hence the rotor heating — has in fact increased significantly. Negative-sequence protection is considerably more sensitive to this, because it responds specifically to the unbalance component rather than to the total current.

How the protection works in practice

A 46 relay continuously calculates the negative-sequence component I2 from the three measured phase currents and compares it against a set threshold, typically expressed as a percentage of rated current. Because the thermal loading of the rotor increases roughly with the square of I2, an inverse-time characteristic is often applied (similar to thermal overload protection): the higher the measured I2, the shorter the time to trip, so that the protection treats a prolonged, moderate unbalance differently from a short, extreme unbalance (for example due to an actually broken phase).

Note: the exact threshold and time characteristic of a 46 relay follow from the manufacturer-specified negative-sequence thermal limit of the specific motor or generator; this article covers the principle, not a ready-made setting for every machine.

Practical relevance

For a motor that repeatedly overheats without the measured, average phase current showing a clear overload, it is important to check the individual phase currents for unbalance — a relatively small, persistent current unbalance (for example caused by a slightly unequal supply voltage or a deteriorated terminal connection in one phase) can cause significant, localized rotor heating that an ordinary thermal overload protection alone does not detect in time.

Common mistakes

Assuming a normal, average phase current means there is no problem — significant phase unbalance can be present while the average current still stays within normal limits.
Relying on ordinary thermal overload protection (10/20/30) alone for negative-sequence protection — the latter is designed for symmetrical overload, not specifically for the much more damaging, squared heating caused by I2.
Not tuning the 46 setting to the manufacturer-specified negative-sequence thermal limit of the specific machine — a generic setting may be too insensitive for a smaller, but still vulnerable motor, or too sensitive for a more robust machine.
Ignoring a persistent unbalance alarm without investigating the cause — for example a deteriorated terminal connection, an asymmetric supply voltage, or a partially broken phase, which can lead to a complete failure if not addressed.