IT systems as reflected in the standards

A distinction must be made between the power source for a "normal" IT system and for a medical IT system. Disconnection in the form of basic insulation is required for the power source. In practice, this usually takes the form of an isolating transformer. This function can also be performed by a battery, by a standalone photovoltaic (PV) system or by a mobile power generator.  In a medical IT system, the leakage currents during intracardial interventions (e.g. open-heart surgery) must be extremely low given the potential risk to the patient. The required isolating transformer is described in IEC 61558-2-15:2011-11.

Figure 2: Layout of a medical IT system in acc. with IEC 60364-7-710

The fault current I_d following the occurrence of a first fault against an exposed-conductive-part or earth is very low, and automatic shutdown is not necessary (IEC 60364-4-41:2017-03, 411.6.1), presupposing that the earthing requirement defined in Section 411.6.2 has been satisfied. This means that the protective conductor resistance R_A is parallel to the body resistance and that the already very low fault current flows through this protective conductor and that the touch voltage remains significantly below the maximum permitted value of 50 V in AC systems. This is of particular benefit in the medical sector.

The value of fault current I_d at a first fault is determined by the nominal voltage, nominal frequency and the parallel circuit comprising the system leakage capacitance and insulation resistance of the electrical system to earth. The fault current, in the event of a first fault, flows with negligible impedance between a line conductor and an exposed-conductive-part. With a good level of insulation in an electrical system, I_d can be approximated by the system leakage capacitance and can be calculated as follows: 

For a 3-phase system

 I_Ce = U⁄√3 ×3ω × C_e = U ×√3×ω ×C_e 

For single-phase system

I_Ce = U × ω × C_e 

Figure 4: Fault current Id with a first insulation fault in the IT system (backup circuit diagram)

Figure 5: Example of touch voltage UT after a first fault

With an insulation fault locating system (IFLS), faulty outputs and/or devices can be located during operation, i.e. there is no need to shut down the system. For troubleshooting purposes, measuring pulses are superimposed on the IT system that are in turn picked up and evaluated by measuring current transformers. Based on the assignment of the measuring current transformer/output, it is then easy to identify the faulty output.

Figure 6: Principle diagram of IT system with an IMD and several IFLS

In acc. with IEC 60364-4-41:2017-03 subclause 411.3.3, additional protection must be provided in AC voltage systems in the form of fault current protective devices (RCDs) for < 32 A sockets that are intended for use by non-specialists and for general use. However, this requirement is not practicable in allowing RCDs to achieve the desired protective action in IT systems. For a start, the operating principle of an RCD after a first fault requires a fault current I_d that must be above the rated fault current I_Δn of an RCD (e.g. > 30 mA). However, in practice, this is not the case. Even two independent insulation faults at both live conductors or connected equipment items do not cause an RCD to trip because these two faults act like a load.

In addition, IEC 60364-4-41:2017-03, states that additional protection for socket-outlet circuits using RCDs with I_Δn < 30 mA is also required for IT systems if a fault current I_d > 15 mA flows in the event of a first fault. On closer inspection, however, it becomes apparent that this requirement is technically questionable:

• Apart from the fact that sockets are the exception rather than the norm in IT systems, how should the fault current I_d be even determined in the planning phase? The fault current I_d is determined mainly by the cable length and the number of loads as well as the leakage currents of the electrical system to earth. However, no planner can determine these variables with even the best of planning.
• RCDs currently available on the market do not have a response value of I_Δn < 15 mA. In addition, the German DIN VDE 0100-530 (VDE 0100-530):2018-06, 538.4 states that discriminating (direction selective) residual current devices are recommended in AC IT systems to avoid unwanted notifications/alarms due to leakage currents when high leakage capacitances are likely to occur downstream of where the residual current device is connected. The problem here is that such discriminating RCDs are also not available. 

Para. 8: Two insulation faults on different live conductors "behind" an RCD

Para. 9 Division of system leakage capacitances "before" and "after" an RCD

Para. 10 Use of RCDs in branched IT systems

Selection of insulation monitoring devices

An insulation monitoring device is selected on the basis of the following criteria:

• Maximum nominal voltage
• Network type, i.e. AC, DC or AC / DC
• Main circuit, control circuit or specialist application
• System leakage capacitance
• Response values
• Inclusion of an insulation fault location system
• Specific environmental conditions

To simplify the process of selecting insulation monitoring devices for the planner and user, the product standard for insulation monitoring devices, IEC 61557-8:2014/CORI:2016 includes further stipulations:

• Designation of the corresponding application
• A distinction is made between insulation monitoring device types based on possible components in the IT system:

   

  • type AC IMD for pure AC IT systems
  • type DC IMD for pure DC IT systems
  • type AC/DC IMD for AC IT systems with directly connected rectifiers and for pure DC IT systems and for DC IT systems with directly connected AC inverters

   

• A distinction is made between insulation monitoring device types for specific applications:

   

  • Medical
  • Photovoltaics

Figure 11: Examples for the identification of IMDs

The response value must be set appropriately for the affected system. In accordance with the German standard subclause 538.1.3, a value of 100 Ω/V and, for prewarning, a value of 300 Ohm/V is recommended. In the previous edition of the standard, a guide value of 50 Ω/V is recommended. Both parameters are correct in principle and are influenced by the number of consumers and by the quality of the installation (e.g., humidity, dust etc.). In practice, the value displayed on the screen of the IMD is used to set an alarm value that is less than this displayed value, and that therefore represents the desired minimum value, leaving sufficient latitude for service and maintenance work. Also bear in mind here that all important system outputs are also in operation.

However, another advantage is that any significant change to the insulation resistance by switching on or off a load or system component is displayed by the IMD, enabling potential weaknesses to be identified.

In accordance with IEC 60364-4-41:2017-03, subclause 411.6.4, after an initial fault occurs, the conditions for an automatic disconnection following a second fault occurring on a different live conductor must be satisfied. In practical terms, this means that a defined level of loop impedance must be achieved. For IT systems without a neutral conductor, the loop impedance is defined as:

Z_S ≤ U/(2 × I_a )

Where:

U = Nominal AC voltage between the line conductors
I_a = Current (in A) that causes a protective device to be triggered within a time period defined in subclauses 411.3.2.2 / 411.3.2.3.

If for reasons of availability, if the system is not intended to shut down when an initial fault occurs, ensure that the prevailing fault current I_d is < 0.4 I_Δn when using RCDs. For a 30 mA RCD this means that I_d is < 12 mA.

Additionally, it should be stated that symmetrical faults on different live conductors should never cause a fault current that causes a system to shut down.

An RCD can be used for each consumable if and only if the shutdown conditions for overcurrent protection cannot be satisfied, because for example:

• it is not possible to define the loop impedance precisely (cable lengths difficult to estimate, metallic materials close to lines)
• the fault current is so low that the maximum permitted disconnection time when using overcurrent protective devices cannot be achieved
• the loop resistance is too high to ensure an automatic shutdown
• and additional equipotential bonding is not possible

Figure 13: The second fault in the IT system