GE IAC and ABB HU Transformer Protective Relays in Nuclear Power Plants

Protective relay panel, photo courtesy North American Switchgear, whose website it located at, microprocessor-based protective relays aren’t yet a common staple in nuclear power plants because of the trepidation associated with installing sensitive solid-state circuitry in safety-critical applications. This trend, however, is slowly changing. Until then, though, it is useful to become aquainted with the older kinds of relays still in operation.

This article describes the basic kinds of transformer faults and the type of transformer protective relays that guard against them. It will then focus on two specific inductive relay models: the GE IAC and ABB HU.


Standard circuit breakers in lower voltage circuits (such as 120 VAC residential breaker panels or 480 VAC MCCs) detect the load current directly. Trip (instantaneous, short-time, long-time) and time-delay settings, if adjustable, are made by adjusting various dials on their front faces. These dials tend to adjust physical spacings to change the amount of electromagnetic force required to open the breaker.

Breakers whose operation is triggered by signals from protective relays are different. They aren’t really breakers at all, but more like power switches. Their contacts open when a control relay is actuated. Output contacts on protective relays actuate the breaker control relay, which then opens the breaker contacts and breaks the circuit.

The breaker trip settings, then, must be set on the relay, not the breaker. The settings are functions, typically, of current supplied by the secondary side of a current transformer. The CTs image the current flowing through the transformer, but on a smaller scale and with smaller current magnitudes that are proportional to the actual phase currents. Instead of adjusting dials on the breaker, then, you select various tap ratios on the relay to establish a relay’s sensitivity to fault currents and set instantaneous trips and time delays. Time delays on protective relays are adjusted by turning a time dial.


Two common protection schemes are differential relay protection and overcurrent protection. Differential relays compare the current entering the transformer to that leaving the transformer. Under normal conditions, they will be balanced. The overcurrent relays detect high instantaneous currents. The relays on the primary side and secondary side are separate, and they must be coordinated (time delay and magnitude). Their setpoints tend to be set higher than the transformer’s maximum inrush current to avoid misoperation, which means the overcurrent relays will not detect smaller currents caused by internal faults. They also usually have time delays to help avoid nuisance tripping on inrush.

Where time- and instantaneous-overcurrent relays protect the transformer from external faults like short-circuits upstream or downstream, differential relays protect against various internal faults: winding shorts (phase-to-phase or phase-to-ground), core failures, and tank faults.

There are several challenges to overcome when designing differential protection for transformers. One is inrush current, which resembles an internal fault to the differential relays.

Another issue is current transformer saturation. During a heavy external fault, the current magnitudes can be so high that the CTs saturate on the side of the transformer (primary or secondary) experiencing the fault. When they saturate, their output rises, peaks, then decreases. The consequence is that the differential relays appear to be detecting a differential fault because the secondary current produced by the opposite set of CTs does not track that of the side where the external fault occurs.

To summarize, the differential relay will detect a differential current under three main conditions:

  1. During transformer energization (inrush)
  2. During an external fault (CT saturation)
  3. During an internal fault (the fault of interest)

The differential relay must be able to correctly distinguish among these three different scenarios.

To detect inrush current, many differential relays employ what is called second-harmonic blocking. Because of the magnetic nature of inrush conditions, it turns out that inrush can be distinguished from an internal fault by examining the harmonic content. The magnetizing nature of inrush currents produces significant second harmonic content. The relays detect this second harmonic and compare it against the fundamental frequency. If the detected ratio is greater than the setpoint ratio (15% or so), then the relay is prevented from operating so that it doesn’t nuisance trip (misoperate) on inrush.

There are multiple ways in theory to overcome CT saturation, such as using a CT that won’t saturate under the worst-case expected conditions. But this kind of solution is “running away from the problem” by trying to design away the problem of CT saturation during an external fault. This form of solution can have other undesirable repercussions, such as increases in cost and size when these variables are usually constrained. But as a GE white paper has put it, “the elegant resolution of this problem would be to utilize the fact that the fault CT may saturate as a basis for the solution.”

That’s what the classical induction relay does. Applying the principle of induction, it uses sinusoidal currents to induce electromagnetic forces in a disc. These forces generate a torque that causes the disc to spin, much like a motor. A contact is attached to the disc, and the rotating disc pushes the movable contact into the fixed contact. This image from clearly illustrates the concept:

Induction relay illustration courtesy of Fikret at
Illustration of inductive relay operating principle

Because differential relays are normally very sensitive to differential currents, they employ a secondary coil to offset the tendency of the operating coil to actuate. This secondary coil is called a restraining coil because it decreases the relay’s inherent sensitivity to differential currents and restrains the relay from operating falsely. The differential currents produced by internal faults must be large enough to overcome the restraining effect of the restraining coil.


ABB HU-model differential relays employ this set of features to compensate for these false-positive conditions. The HU relays enforce a certain minimum threshold of differential current that must be flowing before the relays actuate.

The “HU” model differential relays were originally Westinghouse products, and they are frequently found in nuclear power plants protecting transformers. At some point in the past, the Westinghouse protective relay division was purchased by ABB, so now they are “ABB HU” relays.

The ABB relay setting process is relatively straight-forward as long as you are familiar with the variables involved: select the correct taps by following ABB’s 7-step procedure. The procedure contains a built-in CT performance check to ensure the initially determined tap settings are adequate. If they aren’t, then adjustments should be made. The manual contains the relay-setting process, and you can download the ABB HU manual by clicking this link. The calculation steps begin on page 10, and examples are presented on pages 14 and 15

IAC-model GE overcurrent relays have been used to protect transformers from instantaneous faults. They are also induction-type relays. Compared to the ABB HU differential relay, they are a bit simpler. Like the differential relay, the disc rotates a movable contact, mounted to the disc shaft, into a fixed contact upon operation. As the manual explains, “the disc shaft is restrained by a spiral spring to give the proper contact-closing current, and its motion is retarded by a permanent magnet acting on the disk to give the correct time delay.” The tap setting selection procedure is a bit simpler than the ABB HU process, too. You can read the manual by clicking this link. A relay setting example is given on page 20 of the PDF.


Large power transformers are critical pieces of equipment. They take a long time to manufacture, so they are not easily or quickly replaced. If they are damaged, they can take a while to repair. To a power plant whose primary purpose is power production and distribution to the power grid, losing a main transformer because of a fault that could have been mitigated is a costly failure.

Differential and overcurrent protective relays guard against a wide array of fault conditions. They act to break the fault as soon as possible in order to prevent it from severely damaging the transformer. Two common types of protective relays used in the nuclear power industry are those of the GE IAC family and the ABB HU family. Knowledge of these relays is helpful when performing transformer replacements. The new transformers usually come equipped with new CTs whose characteristics will probably differ from those presently installed. In these cases, it is useful to understand the protective relaying circuitry and know how the variations in CT tap ratios and burdens, not to mention changes in transformer full-load or fault let-through current, will potentially impact the existing relay settings.

What do you think?