Resistance earthing (NGR) — low-resistance versus high-resistance star point, and why it differs from a Petersen coil

The Petersen coil guide covers how a resonantly tuned reactor compensates the capacitive earth-fault current of a cable-rich network. This article covers the other main route for deliberately not earthing a network's star point solidly: a neutral grounding resistor (NGR), also called resistance earthing, connected between the star point (directly, or via a zigzag transformer, see the related guide) and earth.

Two variants: low-resistance (LRG) and high-resistance (HRG)

An NGR deliberately limits the earth-fault current to a predetermined, manageable value, instead of the very high current that would occur with a solidly (directly) earthed star point. There are two common implementations, with a fundamentally different purpose:

Low-resistance grounding (LRG): limits the earth-fault current to typically a few hundred amperes (for example 200–400 A, system-dependent) — high enough for a sensitive, fast and easily coordinated earth-fault protection (see the guide on REF protection), but considerably lower than the fault current with a solidly earthed star point, limiting damage at the fault location.
High-resistance grounding (HRG): limits the earth-fault current to typically only a few amperes. With a sufficiently low fault current, the installation can, much like a Petersen-compensated network or an IT system, continue operating through the first fault without immediate disconnection, because the fault current is too small to cause significant damage or a dangerous touch voltage.

Why the HRG resistor value is not chosen arbitrarily

In a high-resistance earthed network, the current flowing through the NGR resistor itself during a fault must generally be at least equal to (or greater than) the network's total capacitive charging current — the same capacitive current covered in the Petersen coil guide. If the resistive current through the NGR is too small relative to this capacitive current, a transient overvoltage can occur when the fault is cleared, because the capacitive energy is insufficiently damped. A Petersen coil solves this problem by actively compensating the capacitive current with an opposing inductive current (resonance); a high-resistance NGR solves the same underlying problem instead by resistively damping the current sufficiently — two different mechanisms for the same purpose, and therefore not interchangeable techniques.

The short-time thermal rating of the NGR resistor

Unlike a continuously loaded component, an NGR resistor is typically designed for a short-time thermal load — often specified as the ability to withstand the rated fault current for typically 10 seconds — because the associated earth-fault protection is expected to clear the fault well within that time. An NGR is therefore not designed to withstand a continuous earth fault; a protection relay set too slowly relative to the resistor's thermal rating can burn out the resistor itself before the fault is cleared.

Note: the exact choice between low-resistance and high-resistance earthing, the resistor value and the thermal rating follow from the network study (fault current level, transient overvoltage risk, desired protection selectivity) of the relevant installation or network operator; this article covers the principle, not a ready-made selection table for every application.

Why a high-resistance earthed network still needs insulation monitoring

Just as with an IT system or a Petersen-compensated network, being able to ride through a first fault does not give licence to leave that fault undetected: a second, independent fault on another phase could then cause a much higher, undamped fault current between the two fault locations. A high-resistance earthed network therefore requires a current relay that monitors the (small) current through the NGR and signals a fault as soon as one occurs, so that it can be located and cleared before a second fault develops.

Practical relevance

When assessing a medium-voltage installation with resistance earthing, it is important to recognise which of the two variants has been applied: a low-resistance earthed network is designed to trip an earth fault quickly, whereas a high-resistance earthed network is designed to ride through the first fault — with the accompanying obligation to detect and clear that first fault in good time via monitoring.

Common mistakes

Confusing low-resistance and high-resistance earthing — the first is intended for fast tripping, the second for riding through the first fault.
Confusing a high-resistance NGR with a Petersen coil — both limit the earth-fault current, but via a fundamentally different mechanism (resistive damping versus active inductive compensation).
Assuming a high-resistance earthed network needs no insulation monitoring because the fault current is already small — just as with an IT system, a first fault must still be signalled and cleared.
Ignoring the NGR resistor's thermal (time) rating when setting the associated earth-fault protection — a relay setting that is too slow can burn out the resistor before the fault is cleared.