Occurrence and detection of an earth fault in MV power systems with isolated neutral

Methods of neutral earthing

The mode of operation of an MV industrial power system that is galvanically isolated from the public power supply is determined by the method of neutral earthing on the secondary side of the transfer transformer.

The following methods of neutral earthing can be used to earth secondary-side neutral points of transfer transformers:

• Isolated neutral,
• Earth-fault compensation or resonant neutral earthing,
• Low-impedance neutral earthing.

Power systems with low-impedance neutral earthing also include those with isolated neutral or with resonant neutral earthing whose neutral point is earthed temporarily on each occurrence of an earth fault.

Solid neutral earthing, which is not mentioned, is not particulary important in MV industrial power systems due to of the high line-to-earth short-circuit currents and the resulting interference (EMC), earthing (permissible touch voltage) and dimensioning problems (required rated short-time withstand current of the cable screen).

Power system with isolated neutral

A power sytem with isolated neutral (Figure 1) is defined as a power system in which the neutral points of transformers and generators are either not connected to earth or only connected to earth through measurement and protective devices with a very high impedance or through an overvoltage protector.

Fault current: I_F = I_CE = j × 3 × ω × C_E × U_LE ; I_CE ≤ 30 A

Where:

• C_E – Line-to-earth capacitance
• I_CE – Capacitive earth-fault current
• U_LE – Line-to-earth voltage (U_LE = U_LL / √3)
• U_EN – Neutral-point displacement voltage
• ω – Angular frequency (2 × π × ƒ)

Operation with isolated neutral is the simplest form of earth-fault-oriented neutral-point connection. In the event of an earth fault, voltage displacement to earth occurs.

The line-to-line voltages of the earth-faulted power system, on the other hand, do not change. Because of this, no reaction arises that is disadvantageous for the loads connected to the power system and operation can be maintained even in the event of an earth fault.

Where:

• U_L1, U_L2, U_L3 – Star voltage of lines L1, L2, L3
• U_L2, U_L3 – Star voltage on the faultless phases L2 and L3 raised by the factor √3
  (power-frequency overvoltage)
• U_EN or U_en – Neutral-point displacement voltage
• I_C-L2 or I_C-L3 – Capacitive charging current of line L2 or L3
• I_CE – Capacitive earth-fault current at the fault location

However, the voltage displacement to earth causes the capacitive earth-fault current I_CE of the entire power system to flow through the earth fault location. This earth-fault current, which is largely determined by the earth capacitance of the lines C_E, has the magnitude I_CE ≈ 3 × ω × C_E × U_LE.

Due to the comparatively low thermal stress at I_CE < 30 A, the earth-fault arc can burn for longer without destroying the insulation of the intact conductors. Destruction of this insulation would entail a serious danger of the earth fault developing into a double earth fault or short circuit.

To detect the earth-fault location, SIPROTEC relays with sensitive earth-fault detection can be used. They measure the capacitive residual currents. The residual currents of the faultless and earth-faulted feeder differ in their magnitude and direction (Figure 3).

In faultless feeders, all residual currents flow in the same direction. Their magnitude depends on the magnitude of the relevant capacitive charging current.  The residual current of the earth-faulted feeder is the sum of the charging currents of all faultless feeders flowing in the opposite direction.

By means of the measurement with reference to the neutral-point displacement voltage U_en of this capacitive residual current (sinφ measurement), the SIPROTEC relay detects the feeder with the earth fault.

The neutral-point displacement voltage U_en required for earth-fault direction detection is obtained by connecting single-pole-earthed inductive voltage transformers (Figure 4).

If this is used in a system with isolated neutral, there is a risk of relaxation oscillations (ferroresonance). Relaxation oscillations are caused by interaction of the non-linear no-load inductance of the voltage transformers connected to earth with the earth capacitance of the network.

The relaxation oscillations, which mainly occur when an earth fault arc is extinguished or on an energizing operation, put the iron core in the saturated state and cause high core losses.

As a consequence of these high core losses, the voltage transformer can be thermally overloaded and finally destroyed.

The simplest and safest way of avoiding relaxation oscillations is to insert an ohmic damping resistor R_D into the earth-fault windings of the three voltage transformer units interconnected in an open delta (Figure 4). The damping resistor R_D is rated so that neither it nor the voltage transformer is thermally overloaded.

Table 1 contains standard values for damping resistors that have proven convenient in practice.

Table 1 – Standard values for damping resistors R_D

The choice of R_D is based on the thermal limiting output (rated long-duration current) of the e-n or da-dn winding of the voltage transformer. If these standard values cannot be used, other values can be calculated.

The calculation must be performed as follows (Equations 1, 2, 3 and 4):

Where:

• R_D-req required damping resistor (minimum value)
• R_D-select selected damping resistor
• P_V-req required thermal load capacity (minimum value)
• P_V-select selected thermal load capacity
• U_rsec secondary rated voltage of the earth-fault winding
• S_rsec thermal rated limiting output

Example

Example of calculation of the damping resistor dimensioning according to above formulas (R_D-req and P_V-req): see Table 2 below.

Table 2 – Calculation of the damping resistor dimensioning

A further negative phenomenon in isolated-neutral systems are overvoltage levels that result in an excessive voltage stress on the equipment insulation.

On occurrence of an earth fault, the capacitive charge reversal of the faultless lines is brought about by a transient. This transient occurs as a medium-frequency oscillation that briefly causes an overvoltage. The transient overvoltage on occurrence of an earth fault can be 3 to 3.5 times the star voltage.

This transient overvoltage is brought under control by the insulation level that is standardized and assigned to the specific rated voltage of the item of equipment (Table 3).

Table 3 – Standardized insulation levels in the range 1 kV < U_m ≤ 36 kV according to DIN EN 60071-1 or IEC 60071-1

1. Alternating voltage with a frequency between 48 Hz and 62 Hz and a duration of 60 sec.
2. Voltage pulse with a rise time of 1.2 μsec and a time to half-value of 50μsec
3. For the MV range 1kV < U_m ≤ 36 kV, no rated switching impulse level is defined in a standard. It is considered in the insulation coordination with a absolute value of 0.8×U_rp . Because of the longer time to half-value of the transient switching overvoltage, U_rSIL is less than  U_rp.

Compliance with these time limits must be ensured in the operation of cable networks with isolated neutral.

Despite its negative side-effects (ferroresonance, high transient and long-lasting power-frequency overvoltages), operation with isolated neutral is very important for industrial power supplies. This method of neutral earthing is preferred in small cable networks and those without (n–1) redundancy.

Source: Planning Guide for Power Distribution Plants by Dr.-Ing. Hartmut Kiank and Dipl.-Ing. Wolfgang Fruth (Siemens)