Uninterruptible static power supplies and the protection of persons

..........................................................................Collection Technique

Cahier technique no. 129

Uninterruptible static power suppliesand the protection of persons

J.-N. Fiorina

Building a New Electric World

"Cahiers Techniques" is a collection of documents intended for engineersand technicians, people in the industry who are looking for more in-depthinformation in order to complement that given in product catalogues.

Furthermore, these "Cahiers Techniques" are often considered as helpful"tools" for training courses.They provide knowledge on new technical and technological developmentsin the electrotechnical field and electronics. They also provide betterunderstanding of various phenomena observed in electrical installations,systems and equipments.Each "Cahier Technique" provides an in-depth study of a precise subject inthe fields of electrical networks, protection devices, monitoring and controland industrial automation systems.

The latest publications can be downloaded from the Schneider Electric internetweb site.Code: http://www.schneider-electric.comSection: Press

Please contact your Schneider Electric representative if you want either a"Cahier Technique" or the list of available titles.

The "Cahiers Techniques" collection is part of the Schneider Electric’s"Collection technique".

ForewordThe author disclaims all responsibility subsequent to incorrect use ofinformation or diagrams reproduced in this document, and cannot be heldresponsible for any errors or oversights, or for the consequences of usinginformation and diagrams contained in this document.

Reproduction of all or part of a "Cahier Technique" is authorised with thecompulsory mention:"Extracted from Schneider Electric "Cahier Technique" no. ....." (pleasespecify).

no. 129Uninterruptible static powersupplies and the protectionof persons

ECT 129 first issue, September 2004

Jean-Noël FIORINA

Joined Merlin Gerin in 1968 as a laboratory technician in the ACS(power supplies and static converters) department where he wasinvolved with debugging static converters.He then rejoined the ACS department in 1977 having graduated fromENSERG (Ecole Nationale Supérieure d’Electricité et deRadioélectricité de Grenoble, France).Working initially as a design engineer and then project manager,he subsequently became Innovations Manager at MGE UPS Systemsbefore taking up the post of Scientific Director. In some ways, he isthe father of medium and high-power inverters.

Cahier Technique Schneider Electric no. 129 / p.2


Uninterruptible static power suppliesand the protection of persons

The primary function of uninterruptible power supplies is to ensurecontinuity of electrical power. Even when the line supply fails, they are ableto provide the necessary power via their integrated batteries, a backuppower source, or even an emergency standby source in the event of a fault.

Static transfer switches can be used to supply power to a load via twoindependent power sources. In the event of a fault, they automaticallytransfer the loads from one source to the other.

However, the number of power sources and the multiplicity of possibleconfigurations, as well as the different types of earthing system increasethe complexity of setting up these devices in accordance with therequirements of installation standards IEC 60364 and NF C 15-100.

This “Cahier Technique” summarizes these difficulties and offersexplanations to assist with the selection of the most suitable solutions forthe various case scenarios.

Contents

1 Reminders: Low-voltage neutral earthing 1.1 Description p. 4systems 1.2 In practice p. 4

2 The main configurations or types of UPS 2.1 Double conversion UPS (known as online UPS) p. 6

2.2 UPS in line-interactive operation p. 6

2.3 UPS in passive stand-by operation p. 7

2.4 UPS connection configurations p. 7

2.5 Specific features associated with different types of UPS p. 8

2.6 Specific restrictions p. 8

3 Protection against direct contact p. 10

4 Protection against indirect contact 4.1 Selection of earthing systems upstream of the UPS p. 11

4.2 TN-C earthing system upstream of the UPS p. 11

4.3 Different earthing systems downstream and upstream of the UPS p. 11

4.4 Identical earthing systems upstream and downstream of the UPS p. 12

4.5 Detailed requirements regarding the extent of protectionby means of automatic power supply disconnection p. 13

4.6 Protection against input voltage feedback p. 13

5 Application 5.1 Application in relation to single UPSs p. 14

5.2 Application to UPSs connected in parallel p. 21

5.3 Application to STSs p. 26

6 Protection against indirect contact 6.1 Devices for monitoring DC circuits p. 31for DC circuits and the battery 6.2 Main applications p. 34

7 Conclusion p. 40

Bibliography p. 41


1 Reminders: Low-voltage neutral earthing systems(standards IEC 60364-1 and NF C 15-100 Section 312)

1.1 Description

These standards define three types of earthingsystem for low-voltage installations:c The TT system, known as directly earthedneutralc The TN system, known as connected to neutralc The IT system, known as unearthed orimpedance-earthed neutral

They are identified by 2 letters:c The first letter denotes the relationship of theneutral at machine zero to earth:T = Neutral connected directly to earthI = Either all live components isolated fromearth or neutral connected to earth via animpedance

c The second letter denotes the relationship ofthe frames to earth:T = Frames connected directly to earthN = Frames connected to neutral

Two other letters are also used with TN systems:c TN-S = When the protection function isprovided by a separate conductor from theneutral or the live earthed conductorc TN-C = When the neutral and protectionfunctions are combined in a single conductor(PEN conductor).

The table in Figure 1 next page summarizes therequirements of the standard governing theinstallation and operation of these systems.

1.2 In practice

c Although the TT system is the simplest interms of design and operation, it does need tobe protected by means of residual currentdevices (RCDs) on all circuits.

c The TN-S system is highly recommended forpower supplies for data processing and othersimilar equipment due to the presence ofnumerous filtering capacities, which create asignificant earth leakage current.

c The TN-C system is not recommended forcommunication equipment due to the inevitabledifferences in voltage between the variousenclosures. Moreover, its use is prohibited insome cases (for example for electrical powersupplies in hazardous areas where there is a riskof fire or explosion) due to the circulation ofsignificant currents in the PEN conductor and in

conductor components connected in parallel withthe PEN (shielding, metal structures in buildings,etc.)

c Due to its ability to tolerate the first fault, theIT system may be required for safety equipment.It requires a permanent insulation monitor (PIM)to rectify the first fault before a second appears.Note that a single installation may be designedwith one of these earthing systems or with anumber of earthing systems (see Fig. 2 ).

For more detailed information about earthingsystems, see the following “Cahier Techniques”:c “Earthing systems in LV”, no. 172c “Earthing systems worldwide and evolutions”,no. 173c “The IT earthing system (unearthed neutral) inLV”, no. 178.

PEN

TN-C TN-S

TN-C-S

TT

NN

PE PE

IT

PE

3

N

Fig. 2 : Example of coexistence between the various earthing systems


PE

123N

RB

123NPE

RB RA

123PEN

RB

123NPE

RB

123NPE

RB

a - Directly earthed neutral (TT) b - Connected to neutral (TN-C)

: Permanent insulation monitor

c - Connected to neutral (TN-S)

d - Connected to neutral (TN-C-S) e - Unearthed neutral (IT)

Neutral to earth -TT- (system a)Operating principle:Breaking on first insulation fault.Principle for the protection of persons:The use of residual current devices (RCDs) (at leastone at the supply end of the installation) isnecessarily associated with the earthing of frames.This is the simplest solution in terms of design andinstallation. It does not require permanent insulationmonitoring, although each fault will lead to thecomponent affected being disconnected from thepower supply.Note: If, for specific operational reasons, the earthconnection for the frames (applications) must beseparated from the earth connection for the neutral(inverter) downstream of a UPS, only this neutralto earth (TT) system may be used.

Connected to neutral (TN)In accordance with standards IEC 60364 andNF C 15-100, the TN system comprises a numberof subsystems:c TN-C (system b): If the neutral N and PEconductors are combined (PEN)c TN-S (system c): If the neutral N and PEconductors are separatec TN-C-S (system d): Use of a TN-S downstreamof a TN-C (vice versa is not permitted)Please note that the use of the TN-C system is notpermitted on lines with conductors with cross-sectional areas of less than 10 mm2.Operating principle:Breaking on first insulation fault.Principle for the protection of persons:c Frames and neutral MUST be interconnectedand earthed

c Immediate breaking on first fault is achieved bytripping the overcurrent protection devices (circuit-breakers or fuses) or by means of a differential device.Inexpensive in terms of installation, the TN schemerequires an installation study and skilled operatingpersonnel. It results in the circulation of high faultcurrents, which may damage some sensitiveequipment.

Isolated (IT) or high-impedance systemWith this system (e), the first insulation fault is notdangerous.Operating principle:c Signaling of first insulation faultc Fault must be located and eliminatedc Breaking in the event of two simultaneousinsulation faultsPrinciple for the protection of persons:c Interconnection (a) and earthing of frames inaccordance with the TT system if all frames are notinterconnected, in accordance with the TN systemotherwisec Monitoring of first insulation fault by means ofpermanent insulation monitorc Breaking on second fault is achieved by meansof overcurrent protection devices (circuit-breakersor fuses) or a differential deviceThe IT system is the best solution for ensuringcontinuity of service. The signaling of the first faultenables protection to be provided against all risksof electrocution. It requires skilled maintenancepersonnel (first fault troubleshooting).

Public distributionThe most common earthing systems are TT andTN. Some countries, for example Norway, use theIT system.

Fig. 1 : Summary of the three earthing systems defined by IEC and NF standards (boxed text)


2 The main configurations or types of UPS

As its name suggests, the primary function of anuninterruptible power supply or UPS is continuityof service. However, a UPS can also performother functions, in particular in terms ofimproving the quality of the voltage supplied to

the load. This explains the various configurationsused. The simplified systems below aredesigned to illustrate the main characteristics ofthese configurations. For more detailedinformation, see UPS standard IEC 62040-3.

2.1 Double conversion UPS (known as online UPS)

In normal operation, the load is supplied withpower permanently via the inverter.The DC voltage required by the inverter is suppliedby the rectifier if the voltage of the main powersupply lies within the tolerance limits permittedby the rectifier, or by the battery if it does not.

In this system (see Fig. 3 ), the battery ispermanently represented on the DC section ofthe inverter and the rectifier also functions as abattery charger. In some configurations, thebattery has a separate charger and is connectedto the inverter directly via a switch when themain power supply fails. In other configurations,in particular in the event of low battery voltage,

~= ~

=Main power supplyorpower supply 1

Backup power supplyorpower supply 2

Load

SwitchBy-pass channel

Battery

Rectifier/Charger Inverter

2.2 UPS in line-interactive operation

~=

Powersupply 1

Load

Battery

Inverter-charger

Power interface

Fig. 3 : Schematic diagram of a double conversion or online UPS

Fig. 4 : UPS in line-interactive operation

a DC/DC converter is inserted between thebattery and the DC section of the inverter.In this case, the battery may have a specificcharger or the charger function may be providedby the DC/DC converter. A channel known as abypass channel can be used to increase theavailability of the power supply in the event of anoverload or if the inverter shuts down following afault or for maintenance purposes.

This configuration is the most popular, inparticular for medium or high-power UPSs, as itoptimizes the quality of the voltage supplied tothe load. It is therefore suitable for all types ofload, even the most sensitive.

This time, in normal operation, the power supplyfor the load is provided via the line by means ofa power interface (see Fig. 4 ). This powerinterface, in conjunction with the inverter connectedin parallel with the load, enables the quality of thevoltage supplied to be improved on a configuration-specific basis by introducing voltage regulation,and also supports filtering of the harmonicsgenerated by the load.The battery is charged by the inverter, whichoperates in reversible mode.If the main power supply voltage strays outsidethe tolerance limits permitted by the UPS, theinverter will continue to supply power to the loadvia its battery.


Note that in order to avoid load disturbancewhen the main voltage strays outside theselimits, an ultra-fast breaking device must beinstalled on the power interface.

Moreover, the voltage supplied to the load has tobe of the same frequency as that of the mainpower supply. If this frequency deviates oroscillates rapidly when the main power supply isreplaced by a generating set, for example, you

2.3 UPS in passive stand-by operation

~= ~

=

Powersupply 1

Load

SwitchDirect channel

Battery

Rectifier/Charger Inverter

Filter

~= ~

=

Unit 1

~= ~

=

Unit 2

~= ~

=

Unit 3

Powersupply 2

Powersupply 1

Powersupply 1

Powersupply 1

Load

Switch

Automatic bypass

Manual bypass

Fig. 6 : Three UPSs connected in parallel with activeredundancy and one automatic bypass. In this case,two UPS units are sufficient to provide the total powerrequired by the load

Fig. 5 : UPS in passive stand-by operation or offline UPS

are left with a choice of transmitting thisdisturbance to the load or discharging thebattery. It is for this reason that this configurationis less common than the double conversion UPSone and is only used in applications withaverage or low sensitivity.

Note that it is possible to add a bypass channelwith a switch to this system in the same way asfor a double conversion system.

In normal operation, the load is supplied withpower directly via the line (see Fig. 5 ). If the linevoltage strays outside the permissible loadtolerance limits, the switch can be activated inorder to continue to supply the load with powervia the inverter.

The switch is often a relay on which the breakingtime is less than 10 ms on source transfer, asthis system is usually reserved for low powerratings.

This configuration is designed for low-sensitivityloads such as personal computers, for which thevoltage quality of the main power supply (linesupply or possible replacement sources) isconsidered sufficient to ensure the correctoperation of the devices.

The main function of this type of UPS istherefore to operate when there is no voltage orwhen the voltage strays outside the permissibletolerance limits.

2.4 UPS connection configurations

c UPS connected in parallel

As a general rule, the power available can beincreased by connecting the UPS in parallel.This configuration is obligatory when the powerrequired by the load exceeds the maximumpower available for a UPS.

However, in most cases, parallel connectionsare used to increase availability in order to beable to continue to supply the load with power inthe event of a UPS fault.

In order to avoid parallel connections beingassociated systematically with increasedavailability, standard IEC 62040-3 reserves theterm “paralleling” for increased power.Paralleling in order to increase availability isdesignated by “parallel redundant UPS”.

In this case, the total power of the UPSsconnected in parallel will exceed the powerrequired by the load by at least one UPS unit.

v Paralleling with active redundancy with onebypass (see Fig. 6 )


The single switch can be used to transfer theload to power supply 2 in the event of failure ofall UPS units.

This single switch is usually installed in a specificcell, which also supports paralleling of the UPSunits and connects the feeders to the load.

v Paralleling with active redundancy with onebypass channel per UPS unitThis time, in order to supply the load with powervia power supply 2, all switches, which inpractice are static switches, must be controlledsimultaneously (see Fig. 7 ).

~= ~

=

Unit 1

Powersupply 2Powersupply 1

Powersupply 2Powersupply 1 Load

C1

Bypass channel 1

~= ~

=

Unit 2

C2

Bypass channel 2

Unit n Load

Sequential redundancy

SS1

SS2

~= ~

=

Unit 1

~= ~

=

Unit 2

Powersupply 2

Powersupply 1

Powersupply 1

c UPS with passive redundancy

If one of the UPS units fails during operation, theUPS in passive stand-by operation unit is startedup in order to restore the power supply to the load.This configuration may or may not feature abypass.Figure 8 shows two identical double conversionUPSs. The emergency power supply for unit 1 isreplaced by unit 2, which provides much greateravailability than that of the line.

Fig. 7 : Three UPSs connected in parallel with activeredundancy and one automatic bypass channel perUPS unit. In this case, two UPS units are sufficient toprovide the total power required by the load

Fig. 8 : Example of two identical UPSs with passiveredundancy and bypass

2.5 Specific features associated with different types of UPS

This overview of the different UPS configurationshighlights the following essential points:

c The UPS acts as a load for the power supplyvoltage source(s).

c The UPS acts as a source from the point ofview of the loads.

c In principle, the UPS output voltage isavailable even in the absence of one or more ofthe power supply sources.

c UPSs are an arrangement of AC and DC circuits.

c In some operational sequences, the load maybe supplied with power directly via one of thepower supply sources.

Moreover, the protection of persons is based onneutral earthing systems, which requireappropriate protection devices.

Depending on the installation, the earthingsystems upstream and downstream of the UPSmay be identical or different.

2.6 Specific restrictions

Presence of electrical isolation in a UPS

Depending on configuration, UPSs may featuretransformers in various places, which electricallyisolate certain circuits from the power supplies orloads, requiring the following cases to be considered:

c With or without electrical isolation betweenupstream and downstream equipment, seeSection 5

c With or without electrical isolation between thebattery and DC circuits on the one hand and the

upstream and downstream equipment on theother, see Section 6

For instances of redundant double conversionUPS paralleling, Figure 9 next page shows thevarious possible locations for transformers.

This representation is as complete as it can bedue to the fact that some channels do not existin other configurations.

The switches have been replaced by the staticswitches (SS) used as standard.


Requirement for continuity of service

In order to optimize power supply continuity forloads connected to the UPS, devices forproviding protection against overcurrents mustbe discriminated in terms of whether theseovercurrents are due to a fault between liveconductors or an insulation fault.

Moreover, as the short-circuit current of aninverter is relatively low (between 2 and 3 timesthe nominal current), the protection devices mustbe defined with great care.

Also, the possible presence of EMC filters, inparticular for the power supply for computerequipment type loads, must be taken intoaccount when defining protection devices.Essentially, these filters include capacitorslocated between the live conductors and earth,which can interfere with the operation ofresidual current protection devices(see “Cahier Technique” no. 114, “Residualcurrent devices in LV”).

Unit 1

TR

TB

TI

Load

SS1

Unit 2

TR

TB

TI

SS2

Unit n

Powersupply 2

Powersupply 1

Powersupply 2

Powersupply 1

~= ~

=

~= ~

=

Fig. 9 : The various possible locations for transformersfor instances of redundant double conversion UPSparalleling


3 Protection against direct contact

The following standards govern protectionagainst electric shock:

c For installationv IEC 60364-4-41v IEC 61140v EN 61140v NF C 15-100 Part 4-41

c For type-tested and partially type-testedassemblies (referred to as equipment previouslyassembled in the factory)v IEC 60439-1v EN 60439-1

c For UPSsv IEC 62040-1-1v EN 62040-1-1v EN 62040-1-2

Persons can be protected against direct contactwith a component that is live under normalcircumstances by installing UPSs (rectifier,inverter and possibly other devices such asstatic switch) inside enclosures (see Fig. 10 ).The degree of protection of these enclosuresmust be at least IP 2xx or IP xxB (in accordancewith IEC and EN 60529).

As far as battery packs are concerned,observance of these standards and theassociated operating restrictions produce threeinstallation modes:

c Integration of batteries with other static powersupply components (rectifier, inverter, bypass,transfer switch, etc.) in a single cell (in this case,the batteries are compartmentalized)

c Installation of batteries in separate enclosures

Fig. 10 : UPS for protecting persons against risksof direct contact with degree of protection IP 215(Galaxy 3000 UPS, source MGE-UPS)

c Grouping of batteries in specialized locations(delimited by partitions in buildings or enclosuresinside a cabinet) reserved for the electricalpower supply

Moreover, the risks inherent in battery packs(dissipation of explosive gases, corrosivesubstances) impose specific installation conditions.


4 Protection against indirect contact

In accordance with standards IEC 60364-4-41and NF C 15-100 Part 4-41, indirect contact isunderstood to include contact by persons oranimals with frames energized accidentally as aresult of an insulation fault.Usually, this type of protection is provided by:c Interconnecting and earthing metal machineframes (equipotentiality)

c Eliminating a fault that may endangerpersons (or property) by means of a protectiondevice selected on the basis of neutral earthingsystems (see Section 1.1)

Safety can also be achieved using othermethods (class II, electrical isolation, etc.),which are not usually employed in UPSinstallations.

4.1 Selection of earthing systems upstream of the UPS

It ought to be possible to use all standardizedearthing systems due to their equivalence interms of the protection of persons.However, it is important to understand their mainoperating characteristics in order to be able tomake a definitive selection (see Fig. 1).

Moreover, depending on their relationship to theUPSs (upstream and downstream), earthingsystems must be installed in accordance withother requirements, which are outlined in thefollowing subsections.

4.2 TN-C earthing system upstream of the UPS

The PEN (protective earth and neutralconductor) must never be broken. The continuityof the neutral conductor is always assured.

4.3 Different earthing systems downstream and upstream of the UPS

Any change in earthing system, with the exceptionof the transition from an upstream TN-C to adownstream TN-S or TT, requires total electricalisolation of the circuits concerned.If a UPS is used, this isolation, which comprises

Upstream transformer

Transformer upstream ofUPS

Transformer downstreamof UPS

Downstreamtransformer

Transformer inbypass channel

Internal transformers

a -

IsolatedRBI network

b -

Non-isolatedRBI network

Non-isolatedRBI network

UPS withoutelectrical isolation

UPS withelectrical isolation

UPS withoutelectrical isolation

Direct or

bypass channel

Direct or

bypass channel

Fig. 11 : The various ways to electrically isolate line supplies upstream and downstream of a UPS[a] UPS without bypass channel or without direct channel and reversible UPS (direct interaction with the line supply)[b] UPS with bypass channel or direct channel

one or a number of transformers with separatewindings, can be achieved in various waysdepending on the number of channels andthe use of transformers in the UPS(see Fig. 11 ).

The downstream system may therefore be TN-C,TN-S or TT, without the need for any specificconfiguration.


If isolated RBI (Rectifier - Battery - Inverter)network UPSs are used, there may be one ora number of transformers. Therefore, theDC circuits and the battery may be connectedwith upstream, downstream or total isolation.Of the numerous different systems, the systems

outlined here demonstrate the principles used toset up earthing systems on an installationfeaturing UPSs. Knowledge of these principles isessential in order to be able to understand andanalyze the offers available from the variousmanufacturers to meet a specific requirement.

4.4 Identical earthing systems upstream and downstream of the UPS

There are two possible scenarios:

c The upstream system is a TN-C system.

c The upstream and downstream systems areTN-S, TT or IT.

The upstream and downstream systems areTN-C systems

As for the general scenario in Section 4-2 witha TN-C system upstream, there are no specificconfiguration requirements.

The upstream and downstream systems areTN-S, TT or IT

In such cases, when a protection device isoperating or maintenance is being carried out,the neutral conductor upstream of the UPS maybe broken or isolated.

c During this interruption, if electrical isolation isnot assured on all channels, the downstreamsystem is an IT system, regardless of the type ofupstream system.Therefore, for TN-S and TT systems:

v Equipment compatibility with the IT earthingsystem must be assured, in particular in terms ofthe withstand voltage of the capacitors designedto meet EMC (electromagnetic compatibility)requirements. Essentially, these capacitors,which accept the phase voltage in neutralearthing systems, run the risk of having toaccept a phase-to-phase voltage in the event ofa fault on the IT system. In the event of a phase-to-earth insulation fault, the potential of the otherphases in relation to earth is the same as thephase-to-phase voltages.However, for the purpose of maintenanceoperations, which are usually completedrelatively quickly, the probability of a faultoccurring is very low. It is for this reason that it isgenerally accepted that capacitors do not sufferthis phase-to-phase voltage.

v You must check that the conditions forprotection against indirect contact have beensatisfied, not only for the TN-S or TT system butalso for the IT system. For this purpose, makesure, in the IT system, that in the event of a

double fault, the current is sufficient to trip theprotection devices (NFC 15-100, Part 4Section 411-6-4 and IEC 60364, Part 4Section 413-1-5-5).

v You must make sure that the neutral conductoris protected by using an overcurrent detectiondevice that will cut off the power supply to all liveconductors, including the neutral (NFC 15-100and IEC 60364, Part 4 Section 431-2).

v During this operating period, insulationmonitoring is not carried out, although theprotection of persons remains assured.However, this state does not usually last for longas the tripping of a protection device is theconsequence of a fault, which can be rectifiedquickly. Moreover, if discrimination has beencalculated correctly, the tripping of the protectiondevice need only affect the faulty feeder and notthe other feeders, which are operating correctly.It is for this reason that standards do not requirePIMs on installations that usually operate asTN-S or TT systems with a UPS in accordancewith this configuration.

For the IT system, operation remains with thesame system, although insulation monitoringcannot be guaranteed during the period in whichthe neutral is broken upstream.

c If there is total electrical isolation upstream anddownstream of the UPS, the various scenariosillustrated before with different earthing systemsupstream and downstream apply:

The earthing system must be reconstructedupstream of the UPS:

v As a TN-S or TT system by means of directearthing upstream of this electrical isolation

v As an IT system by means of the installation ofa new permanent insulation monitor

Total electrical isolation is recommended forimproved protection against disturbance presenton upstream power supplies.


4.5 Detailed requirements regarding the extent of protection by means of automaticpower supply disconnection (NFC 15-100 Part 4 Section 411 and IEC 60364 Part 4 Section 413-1)

When load circuits downstream of the UPS canbe supplied with power directly via the mainsource or replacement source (UPS with bypasscircuit or offline UPS), you must check theconditions for protection against indirect contactin accordance with the requirements of thesestandards, taking into account the main source(power supply 1) and, if applicable, thereplacement source (power supply 2).

Moreover, in all cases, you must check theprotection conditions when the circuits aresupplied with power via the inverter, taking intoaccount the operating characteristics of the latterprovided by the manufacturer.

It is particularly important to be aware of themaximum current that can be supplied by theinverter for overloads that may lead to short-circuits and the period of time during which theinverter is able to provide this current.

Usually, the manufacturer will provideinformation about these devices along with theirrated data in order to safeguard this type ofprotection.

Two practical recommendations

c In order to avoid load disturbance whileprotection devices are in operation, the use ofhigh-speed current-limiting circuit-breakers orfast-acting fuses is recommended.

c In order for protection devices to operate, asthe short-circuit current of an inverter is limited tovalues between 2 and 3 times its nominalcurrent, maximum subdivision of the feeders isrecommended in order to maximize the ratiobetween the fault current for a feeder and the tripcurrent for the protection device.

Therefore, with 4 identical feeders and aninverter short-circuit current equal to 2.5 times itsnominal current, when applied to a feeder, thiscurrent is equal to 10 times its rated current.

Special case of low-power UPSs connectedto the main line supply via a power outletLow-power UPSs are power supplies with arating of less than 3 kVA, which are connectedvia a single-phase power outlet with a maximumrated current of 16 A. They must not be used tosupply power to a permanent installation.

When the cable is disconnected, the protectiveearth conductor PE is broken and the powersupply to the load circuits is provided via theUPS battery pack.

The installation can be considered small-scalewith floating potential. Protection is thenprovided by means of the equipotentialitybetween the various hardware componentsconnected. This requires that the power outletsdownstream of the UPSs have an earth contact.

4.6 Protection against input voltage feedback

UPS standards IEC and EN 62040-1-1 and62040-1-2 require the provision of protectionagainst input voltage feedback.

c For low-power UPSs connected via an outlet,this requirement is justified in particular by thefact that when the cable is disconnected, thepresence of voltage at the male connectorscould put persons or animals at risk.

c For permanent UPSs, this requirement isjustified for maintenance reasons, whenpersonnel have to intervene upstream of theUPS. This device can be integrated into the

equipment or installed externally. In the lattercase, the manufacturer must specify the type ofinsulation device to be used. A label reading“ISOLATE THE UNINTERRUPTIBLE POWERSUPPLY BEFORE WORKING ON THISCIRCUIT” must also be affixed by the user to allinsulation devices installed in a zone not in theimmediate vicinity of the UPS.


5 Application

5.1 Application in relation to single UPSs

Different earthing systems upstream anddownstream

c UPS with electrical isolation: the UPS featuresa transformer on each of its channels.

In the systems described below, the transformerhas been installed upstream of the RBI network andis marked TR (for rectifier transformer). Thischannel can also be electrically isolated by installinga transformer downstream of the inverter (TI).

v TT system: The load neutral is earthed at thedownstream UPS terminal (see Fig. 12 ).

A

B

Cb DTB

TR

SS

E

F

G

U3

U2

U1

RBIC

Protection of persons is achieved using an RCDconnected to a circuit-breaker either globally orindividually on each feeder.

v TN system: The load neutral is earthed via theUPS terminal.

Protection of persons is usually provided byovercurrent protection devices.

If the load earthing system is a TN-C system, thecombined protective earth and neutral conductor(PEN) is distributed via this terminal(see Fig. 13 ).

A

B

Cb DTB

TR

SS

E

F

G

U3

U2

U1

PEN

RBIC

Fig. 12 : TT earthing system downstream(any type of earthing system upstream - TT, TN or IT)

Fig. 13 : TNC earthing system downstream(any type of earthing system upstream - TT, TN or IT)


If, as is most common, the load earthing systemis a TN-S system, the protective earth andneutral conductors are connected to the sameearth connection on the UPS terminal (seeFig. 14 ). RCDs may then be installed, ifnecessary, on the feeders.

v IT system: A PIM, connected between one ofthe live conductors and earth, detects insulationfaults upstream of the UPS (load) as well as onthe UPS as far as the rectifier or invertertransformers (TR and TI respectively) (seeFig. 15 ). Protection of persons in the event of asecond fault is usually provided by overcurrentprotection devices.

c UPS without electrical isolation (no transformerson any of the channels)

Without isolation, the only possible combinationsare the TN-C earthing system upstream of theUPS with TN-S or TT earthing systemsdownstream (see Fig. 16 next page). In effect,

there is no need for an additional transformer,regardless of the number and position of anytransformers that might be present in the UPSdue to the continuity of the combined protectiveearth and neutral (PEN) conductor.

In this case, the combined PEN conductor splitsinto a separate neutral and protective earth atthe UPS output terminal.

The neutral is distributed with the PE for the partof the installation supplied with power in TN-Sconfiguration, but for the part supplied with powerin TT configuration, only the neutral is distributedas the PE is distributed via a local earth.

If other earthing system combinations arerequired upstream and downstream, theprevious scenario should be adopted with theaddition of one or a number of transformers,possibly externally to the electronic equipment ifthe UPS does not feature built-in transformerson each of the channels.

U3

U2

U1

A

B

Cb DTB

TR

SS

E

F

G

PE

RBIC

U3

U2

U1

A

B

Cb DTB

TR

SS

PIM

E

F

GRBI

C

Fig. 14 : TNS earthing system downstream (any type of earthing system upstream - TT, TN or IT)

Fig. 15 : IT earthing system downstream (any type of earthing system upstream - TT, TN or IT)


c Double conversion UPS with invertertransformer (TI)

This type of UPS with an inverter transformerand without isolation in the bypass channel is themost common for medium and high-powerUPSs.

A bypass transformer (TB) must therefore beinstalled in the bypass channel.v Any type of system upstream, TT systemdownstream (see Fig. 17 )The load neutral is earthed at the UPS terminal.Protection of persons is achieved using an RCD

connected to a circuit-breaker either globally orindividually on each feeder.v Any type of system upstream, TN-S systemdownstreamIn this case, which is the most frequent forTN systems, the protective earth conductor andthe neutral are connected to the same earthconnection on the UPS terminal (system identicalto Fig. 15). Protection of persons is usuallyprovided by overcurrent protection devices.RCDs can be installed on the various feeders ifthe fault current is not sufficient to reach the tripthreshold for the overcurrent protection devices.

~

= ~

==

=

SS

TNSsystem

TT system

TT load earth

PE

PEN

PE

~

= ~

=

SS

TI

TB

Fig. 16 : Without electrical isolation, the possible combinations are the TN-C earthing system upstream and TN-Sor TT earthing systems downstream

Fig. 17 : Any type of system upstream, TT earthing system downstream


v Any type of system upstream, TN-C systemdownstream (see Fig. 18 )The load neutral is earthed at the UPS terminal.Protection of persons is usually provided byovercurrent protection devices.The combined protective earth and neutral (PEN)conductor is distributed via this terminal.v Any type of system upstream, IT systemdownstream (see Fig. 19 )A PIM connected between one of the liveconductors (the neutral in this case) and earthdetects insulation faults on the load as well ason the UPS as far as the bypass and invertertransformers (TB and TI).

Protection of persons in the event of a secondfault is usually provided by overcurrentprotection devices.

c Double conversion UPS without transformerThis system, which is usually reserved for lowpower ratings, is becoming increasingly popularin medium-power applications and is starting tobe used in high-power applications.In the very general case, where there is onemain channel and one bypass channel, twotransformers must be used in order to ensureupstream electrical isolation. The system requiredfor the load is reconstructed upstream of the UPSas in previous scenarios.

~

= ~

=

SS

TI

PEN

TB

~

= ~

=

SS

TI

TB

PIMPE

Fig. 18 : Any type of earthing system upstream, TN-C earthing system downstream

Fig. 19 : Any type of earthing system upstream, IT earthing system downstream


Internal inverter faults are protected by means ofupstream circuit-breakers.This two-transformer configuration is used fortwo main reasons:- If the two AC voltage sources are not providedby the same busbar system- In order to improve availability, because a faulton a transformer or input will not affect the otherchannelFigure 20 illustrates an example of this with anytype of system upstream and a TT systemdownstream.

It is also possible to use just one transformer if itis located at the output. In this case, the requiredearthing system can be reconstructed downstreamof the output transformer (see Fig. 21a ).

The neutral upstream of the UPS is only used ifrequired for the UPS to operate. The transformerinduces a voltage drop in relation to the UPSoutput, which may be harmful if the load currenthas an increased distortion rate and in particularwith the third harmonic. In this case, a transformerwith a zigzag-connection secondary may be usedin order to reduce zero-sequence impedance.

~

= ~

===

SS

TB

TR

Transformer earth Load earth

PE

~

= ~

==

=

a -

b -

~

= ~

==

=

N

SS

TS

Upstream source earth

a -

b -

Load earth

Load earth

PE

SSTE

Transformer earthPE

Fig. 20 : If the two transformers (battery and rectifier) are located upstream (2 different sources), the upstream anddownstream earthing systems do not have to be identical

Fig. 21 : With a single transformer TS [a] or TE [b], different earthing systems can be used upstream and downstream


Another possibility is to regulate the transformersecondary voltage, thereby matching itscharacteristics to those of the initial device.If, as is most common, two AC power supplysources are provided by a single busbar system,the system illustrated in Figure 21b precedingpage, with a single transformer, is also possible.This case is the same as UPSs with bypass anda single incoming line for the AC power supply,with the only difference that bridging betweenthe two channels is implemented inside the UPS.

Identical earthing systems upstream anddownstream and no electrical isolationIf electrical isolation is applied, the scenario isthe same as before, with the same systemupstream as downstream. It is for this reasonthat only the scenario in which electrical isolationis not applied is described below.

c For the TN-C system

There are no particular requirements to be met dueto the continuity of the neutral conductor. Thedownstream system can even be a TN-S or TTsystem without any particular configuration beingrequired, as in the scenario described before.The example in Figure 22a illustrates a scenariowith a TN-C system both upstream anddownstream for a UPS without a transformer.

c For TN-S, TT or IT systems upstream anddownstream

The isolation and disconnection of the neutralupstream of the UPS when specific events occurisolates this conductor from the earthdownstream of the UPS. This configurationrequires careful consideration in order to ensurethe safety of personnel working on the backed-up section of the installation.

Figure 22b illustrates this scenario for the TN-Ssystem.

The solution often advocated is to provideelectrical isolation on all channels.Under these conditions, the possible systemsare those considered in the scenarios before andwhich support any type of upstream system.However, the problem does need to be analyzed:During this period, the operating conditions forthe load are those of the IT system.v For property or persons, this operating modeposes no risks, as long as the conditions forprotecting persons against indirect contact havebeen met for the IT system.v Devices supplied with power by the UPS mustbe compatible for operation with this system. Inparticular, filter capacitors for EMC must be ableto withstand the phase-to-phase voltage in theevent of a fault on the IT system (see Section 4.2).

~

= ~

==

=

SS

~

= ~

==

=

SS

PEN

PEN

a -

b -

PE

N

Fig. 22 : No electrical isolation, identical earthing systems upstream and downstream


v In the absence of a PIM, although theoperating conditions are the same as those of anIT system, the lack of a PIM will not affectoperation. This device simply detects the firstfault in order that it can be repaired before asecond fault cuts off the power supply.In order to optimize continuity of service in an ITsystem, a PIM can be installed downstream ofthe UPS and only put into operation when theupstream voltage is lost. Figure 23 illustratesthis option and therefore features PIM 1.PIM 2 is put into operation via relay R2 only inthe event of a loss of voltage upstream of theUPS.In spite of the absence of voltage upstream ofthe UPS, it is possible, even very probable, thatthe neutral conductor will not be broken.A circuit-breaker trips either in the event of afault or for maintenance purposes.

If there is a risk that PIM 2 may be subject todisturbance from PIM 1, a relay R1 can be usedto shut down the former when the transformersecondary voltage is zero.Moreover, in the case of the TN-S system,depending on the country of application, it maynot always be the case that the neutral is broken(see CT 173). If the neutral is not broken, thereis no particular problem and the scenario will beas for the TN-C system.

The only difference is the presence of a neutralconductor and a PE instead of a PEN(see Fig. 24 ).

Low-power UPSs connected via a socketoutletFor these devices, there is only one incomingline, the main and bypass power supplies (if theyexist) are combined.

Fig. 23 : IT earthing system upstream and downstream, PIM2 put into operation on loss of upstream voltage

Fig. 24 : TN-S earthing system upstream and downstream

~

= ~

==

=

SS

R1

PIM 1

PIM 2

R2

PE

N

~

= ~

==

=

SS

PE

N


Moreover, the link to earth is made via the green/yellow cable conductor, which is always at risk ofbecoming broken or disconnected from the socketoutlet.Note that inserting an isolating transformer inorder to create an earthing system downstreamof the UPS, which is different to the earthingsystem upstream, will not alter this risk.

In the very general case, where the upstreamand downstream systems are identical andwithout user knowledge of them, there are nospecific requirements to be met. The safety ofpersons is ensured via the equipotentiality of theframes. It is therefore necessary that all socketsdownstream of the UPS have an earth contact.

5.2 Application to UPSs connected in parallel

Different earthing systems upstream anddownstream of UPSs

c UPS with electrical isolationIf the UPS has transformers on each channel(UPS with electrical isolation), it is sufficient torecreate the required system downstream, as inthe case of single UPSs.

In the system illustrated in Figure 25 , transformershave been installed upstream of the RBI networksand are marked TR (for rectifier transformers). Thischannel could also have been electrically isolatedby installing a transformer in the inverter (TI).

c UPS without electrical isolation

If the UPSs have not been totally electricallyisolated, isolating transformers must be added tothe channels that do not have them.A single transformer can be used for all mainpower supplies and one for all secondary powersupplies.

However, when parallel connections are used inorder to improve reliability, it is not advisable touse a single transformer for all main powersupplies, as the transformer is a reliability node.

For the secondary power supply, a singletransformer is usually sufficient as it is morereliable than the line supply.

When there is only one standby systemimplemented by an NS static switch (the idealsolution for reliability), a single transformer isrequired for the secondary power supply.

Connecting both single devices and staticswitches in parallel allows a number of paths forthe neutral and phase currents on the powersupply circuits. There is therefore a risk ofcertain conductors becoming overloaded. As it isnot easy to control current distribution in each ofthe conductors, this risk is usually eliminated byusing a single cable as far as the vicinity of UPSs.

Fig. 25 : The electrical isolation achieved by installing upstream transformers permits different earthing systems tobe used upstream and downstream of UPSs

A

B

Cb DTB

TR

SS

E

F

G

PERBI

Unit n

TRG

RBI

Unit 2

Unit 1

U3

U2

U1

(*) Loads not needing to be put into operation

(*)


v Single static switch

- UPS with inverter transformer (TI)Figure 26 illustrates the example of a downstreamTN-S system and any type of system upstream.In this case, it is sufficient to add a transformer inthe bypass channel (TB) to achieve totalisolation. This configuration applies equally toincreasing power and to active redundancy.

- UPS without transformerIn this case, a transformer must be added in eachof the UPS units in addition to the transformer onthe bypass channel (TB) (see Fig. 27 ).The transformers should be located upstream ofthe rectifiers (TR) rather than downstream of the

inverters in order to preserve the quality of theinverter output voltage.

If paralleling is employed simply to increasepower, a single rectifier transformer may be usedupstream of the rectifiers.

v Parallel connections with one bypass channelper UPS unitGenerally speaking, transformers must beinstalled on channels that do not already havethem.

- UPS without isolating transformerIn this case and in the event of a redundantparallel connection, transformers must beinstalled on each channel.

A

B

Cb DTB

TO

SS

E

F

G

PERBI

Unit n

TOG

RBI

Unit 2


Unit 1

U3

U2

U1

(*)

A

B

Cb DTB

TR

SS

E

F

G

PE

RBI

Unit n

TRG

RBI

Unit 2

Unit 1

U3

U2

U1


(*)

Fig. 26 : Installation of a transformer in the bypass channel (TB) will achieve total isolation of a UPS installationwith inverter transformer (TI), permitting different earthing systems to be used upstream and downstream of UPSs

Fig. 27 : Installation of transformers upstream of rectifiers (TR) will achieve total isolation of a UPS installationwithout transformer, permitting different earthing systems to be used upstream and downstream of UPSs withoutaffecting the quality of the output voltage


As in the example before, transformers upstreamof rectifiers (TR) should be used instead oftransformers downstream of inverters (TI) inorder to preserve the quality of the voltagesupplied by the inverters (see Fig. 28 ).

For increased power, one transformer may beused for the rectifiers and one for the bypasschannels, or even a single transformer for both(see Fig. 29 ).

In this case, however, you must check that thecurrent distribution in the various bypass

channels is correct in the phase and neutralconductors.If, for operational reasons, the neutral is requiredfor rectifier operation, the problem becomes alittle more delicate due to the presence of doublethe number of neutrals.

c TN-C system upstream with TN-S or TTsystem downstream

As for single devices, no special requirementshave to be met due to the continuity of theneutral (see Fig. 30 next page).

Fig. 28 : Connection of a number of UPSs in parallel, each of which feeded by a transformer (TR) and each of whichhas a bypass channel with upstream transformer (TB), will achieve total electrical isolation of the installation, thuspermitting the use of different earthing systems upstream and downstream of the UPSs without affecting thequality of the output voltage

Fig. 29 : Installation of an upstream transformer will achieve total electrical isolation of the installation, which isessential if different earthing systems are to be used upstream and downstream of the UPSs connected in parallel,each of which has a bypass channel

A

B

Cb

C

C

DTB

TR

SS

E

F

PE

RBIUnit 1

U3

U2

U1

Cb DTB

TR

SS

E

F

RBIUnit 2


(*)

A

B

Cb

C

C

DSS

E

F

PE

RBIUnit 1

U3

U2

U1

Cb DSS

E

F

RBIUnit 2


(*)


Identical earthing systems upstream anddownstream and no electrical isolationAs for single UPSs, in the event of electricalisolation or if this electrical isolation is required inorder to limit high-frequency disturbancetransmitted via the power supplies or to avoidpossible problems with the operation of the ITsystem while the neutral is broken, the situationis identical to the scenario before with identicalsystems upstream and downstream. It is for thisreason that only the scenario without electricalisolation is described below.

c For the TN-C earthing system

There are no specific requirements to be metdue to the continuity of the neutral conductor.

The downstream earthing system can even bea TN-S or TT system without the need for aspecial configuration, as described above(chapter 2).

Figure 31 illustrates a scenario with anupstream TN-C and a downstream TN-C for aredundant parallel UPS connection without atransformer.


A

B

Cb

C

C

DSS

E

F

PE

PEN

RBIUnit 1

U3

U2

U1

Cb DSS

E

F

RBIUnit 2

(*)


A

B

Cb

C

C

DSS

E

F

PE

PEN

RBIUnit 1

U3

U2

U1

Cb DSS

E

F

RBIUnit 2

(*)

Fig. 30 : TNC earthing system downstream and TNS earthing system upstream

Fig. 31 : TNC earthing system downstream and upstream


c For TN-S, TT or IT earthing systems

As for single UPSs, the only difficulty is presentedby the possible disconnection of the neutralconductor.The comments made in previous Section remainvalid and the same configurations are applicable.v Example of parallel connection of UPS unitswith inverter transformer (TI) and single staticswitch (NS)Figure 32a illustrates this scenario with a TN-Ssystem upstream and downstream.In this case, if the neutral is broken (e.g. in theevent of an upstream circuit-breaker B or Cbtripping), the downstream earthing systembecomes an IT system, the protection of personsis assured and the PIM is not obligatory, asexplained in Section 4.3.

In some cases, precautions must be taken, inparticular in the case of the TT system, on whichunintentional tripping of the RCD may occurwhen cables are connected in parallel.v Example of devices connected in parallelwithout isolation with one bypass channel perUPS unitFigure 32b illustrates the scenario with a TTsystem upstream and downstream.Because the two bypass channels areconnected in parallel, the currents in the phaseand neutral conductors on both channels aredistributed on the basis of the impedances of theconductors and static switches. The sum of thecurrents for each of the channels is not zero, andthis risks tripping the RCDs if they are located onthese channels.


A

B

Cb D

TO

SS

E

F

G

PE

RBI

Unit n

TOG

RBI

Unit 2

Unit 1

U3

U2

U1

A

(*)

a -

b -

B

Cb

C

C

DSS

E

F

PE

RBI Unit 1

(*)

U3

U2

U1

Cb DSS

E

F

RBI Unit 2

Fig. 32 : Examples of several UPS units connected in parallel, [a] TN-S upstream and downstream with invertertransformer TI and a single static switch, [b] TT upstream and downstream without isolation with one bypasschannel per UPS unit


The solution is therefore to install a single commonRCD at the supply end of the two channels,which will act upon the common circuit-breaker B.Note: If UPS units are connected in parallel with asingle static switch, this problem does not occur.

However, as the number of possible configurationsis high, it is advisable to ask manufacturers tospecify the precautions to be taken as well asthe associated equipment.

5.3 Application to STSs

Principle and difficulty

STSs - Static Transfer Switches - are devicesfor switching from one source to another(see Fig. 33 ).They can be used to improve power availabilityfor specific loads, which may obtain their powersupply from 2 different sources.

c TT system downstream (see Fig. 34a on theopposite page)The neutral conductors on the two transformersare interconnected and connected to a local earth.This common neutral is the neutral for the load,and the sources are transferred using three-phase static switches (the neutral is not switched).The protective earth conductor PE for the load isconnected to the load local earth.

c TN-S system downstream (see Fig. 34b on theopposite page)The neutral conductors on the two transformersare connected to the load earth.This common neutral is the neutral for the load,and the sources are transferred using three-phase static switches (the neutral is not switched,as in the case of the TT system).The PE conductor is connected to this earth.

c TN-C system downstream (see Fig. 34c on theopposite page)The neutral conductors on the two transformersare connected to the load earth.The combined protective earth and neutralconductor is distributed via this earth connection.The sources are transferred using three-phasestatic switches (the neutral is not switched, speciallyin this case, where it is prohibited to disconnect itas it also serves as the protective earth).

c IT system downstream (see Fig. 34d on theopposite page)The neutral conductors on the two transformersare interconnected but isolated from earth.The PE conductor is drawn from the local earth.A PIM is connected between the combined neutraland earth to monitor the installation’s insulation.If the sources have the same earthing systembut the load does not, it is obviously possible touse the previous configuration, although it is alsopossible to use just a single transformer.However, the neutral conductors on the twosources must be connected to the same earth. Ifthe IT system is used, the neutral conductors willobviously be isolated from earth. They must beinterconnected if a single PIM is monitoring theinstallation.

Although offering fewer advantages in terms ofavailability (as the transformer is an availabilitynode), this configuration is more cost-effective.In this case, the downstream system isimplemented on the transformer secondary.The source is also transferred in three-phasemode (the neutral is not used).

Source 1 Source 2

Load

Fig. 33 : Using static switches to switch the powersupply for a load between two separate sources

Initially, one of the sources is assigned to theload. If this source fails, the static switchestransfer to the second source without interruptingthe supply of power to the load.

The two sources, and the load, do not necessarilyhave identical neutral earthing systems.

If the systems for the two sources and the loadare identical, the neutral conductors pose aproblem. If they are interconnected, they riskbeing crossed by significant currents. If they arenot, they must be switched in the event of sourcetransfers.

This switching can be performed without overlap- i.e. with disconnection of the power supply - orwith overlap in order to avoid this disconnection.However, a transient current may occur, causing,for example, the RCD (if there is one) to trip, oreven the circuit-breakers if the exchange currentis significant.

If the upstream and downstream earthing systemsare not identical, the situation is less complicated,as one or a number of transformers must be usedand the required system set up downstream ofthe device. Switching will therefore only affectthe phase conductors.

Different systems upstream and downstreamIn the very general case in which the two sourcesand the load have different earthing systems, thesolution is to use two isolating transformers. Thefigures below illustrate scenarios for the variouspossible downstream earthing systems in caseswhere any type of upstream system can be used.


a - TT system downstream

SS

PE

LoadLoad earth

SS

b - TNS system downstream

SS

PE

Load

SS

c - TNC system downstream

SS

PEN

Load

SS

d - IT system downstream

SS

PE

Load

SS

PIM

Fig. 34 : Any type of earthing system upstream, downstream earthing system [a] TT, [b] TN-S, [c] TN-C, [d] IT


Figure 35 shows how connections are made in aconfiguration with an upstream TT system and adownstream TN-S system.Figure 36 shows how connections are made in aconfiguration with an upstream IT system and adownstream TN-S system.Here, the two PIMs check insulation in relation tothe same earth to which the protective earth PEis connected.

c Special case of TN-C system upstream andTN-S downstream (see Fig. 37 next page)This time, the transformer is not required due tothe continuity of the neutral.

The PEN conductor is separated into N and PEat the STS output terminal.

Identical systems upstream and downstream

Although using transformers as described aboveis not problematic, it is possible to do withoutthem by taking the following precautions.

c For all earthing systems, with the exception ofthe IT system, the neutral conductors on thetransformers must be connected to the sameearth connection in order that the earthingsystem and the load share the same protectiveearth (PE) as the sources.

SS

PELoad

Transformerearth

STS earth

Load earth

SS

SS

PE

PE

PE

PE

Load

Sourceearth

Load earth

SS

PIM 1PIM 2

Fig. 35 : TT earthing system upstream and TN-S system downstream

Fig. 36 : IT earthing system upstream and TN-S system downstream


SS

PE

PEN

Load

Transformerearth

SS

SS

PEN

Load

Transformerearth

SS

If this connection is not made, the PE used forthe STS and its load will only be associated withone of the sources, rendering it impossible toensure the correct operation of the protectiondevices in the event of an insulation fault.If this configuration is not possible, the earths onthe two sources must at least be interconnectedon the upstream LV switchboards.

c For the same reason, on the IT system, thetwo sources must share the same earthconnection. Moreover, the neutral conductors onthe transformers should be interconnected if

possible in order to avoid sudden variations inthe potential of the load’s live conductors in theevent of source transfers. This implies that theentire installation is monitored by a single PIM.

c If the neutral conductor is not distributeddownstream, three-phase transfers can be madefor the previous scenarios.If the neutral conductor is distributed downstream,this conductor should also be switched, excepton the TN-C system, on which the neutralconductor and protective earth are combined.This configuration is illustrated in Figure 38 .

Fig. 37 : TN-C earthing system upstream and TN-S system downstream

Fig. 38 : With the TN-C earthing system, the neutral does not have to be switched even if it is distributeddownstream


For other scenarios, the switching of the neutralis justified by the significant circulating currents,which may exist in this conductor if it ispermanently connected to the two sources (seeFig. 39 ). The neutral conductors must therefore

SS SS

SS SS

In1-Ic1-Ic2 In2+Ic1+Ic2

In1-Ic1 In2+Ic1

Ic1+Ic2

Ic2

Ic1

In1 In2

Loads 1 notneeding to beput into operation

Loads A withcontinuity ofservice

Loads B withcontinuity ofservice

Main powersupply 1

Main powersupply 2

Loads 2 notneeding to beput into operation

be switched. There are two possible ways inwhich this can be achieved:

1- Switching without overlapThese conductors are not connected in parallel,even briefly, although the transfer time mayexceed 10 ms. Moreover, there is a brief instantduring which the potential of the STS loadneutral concerned is unclear, posing a possiblerisk to the balance of the phase voltages. It musttherefore be ensured that all devices to whichpower is being supplied can withstand this typeof disturbance.

2- Switching with overlapThe transfer is faster (a few ms) and thepotential of the load neutral remains fixed duringswitching. In order to ensure overlapping of theneutrals, one possible solution is to connect aswitch, which will ensure continuity, in parallelwith the thyristors on each SS. In effect, if thecurrent is cut off in one of the neutralSS thyristors, the thyristor will only be capable ofinverse conduction once the voltage at itsterminals reaches a sufficient value, indicatingthat the potential of the neutral is different to thatof the source.

Moreover, a transient exchange current appears,and it must be ensured that this will not pose aproblem for the installation, in particular on TTsystems with RCDs that “see” this current. Inorder to avoid unintentional tripping, these RCDsmust have an appropriate threshold or a timedelay (see Fig. 40 ).

On the above system, loads 1 and 2, which do notneed to be put into operation, cause currents In1and In2 to flow in the neutral conductor.

If the neutral conductors are permanentlyconnected, there will be circulating currentsbetween these two conductors due to the twopossible backfeed paths to the sources.

If currents In1 and In2 are very different, thevalues of the circulating currents may be significantand, in certain cases, even exceed the ratings ofthe STSs.

If the common system is a TN-S or IT system,protection against overloads will be effective, but tothe detriment of continuity of service.

In a TT system, it is the RCDs which “see” apermanent zero-sequence current, even on circuit-breakers at the supply end of the sources.

Fig. 39 : With the exception of TN-C earthing systems,the switching of the neutral is justified due to thesignificant circulating currents, which may flow throughthis conductor

Fig. 40 : siE super-immune RCD (from Merlin Gerin)


6 Protection against indirect contactfor DC circuits and the battery

The mandatory emergency supply backup batteryfor a static UPS must be isolated from earth, asthis operating mode supports continuity of service.In the absence of a transformer, and due to theuse of semiconductors in static UPSs, there iselectrical continuity between the upstreaminstallation, the UPS and the downstreaminstallation. Therefore:

c An insulation fault on the DC circuits can bedetected by the upstream and downstreaminstallation protection devices.

c These protection devices can be subject todisturbance due to the presence of an insulationfault on the DC circuits.

However, in some operational sequences, thebattery and the DC circuits can be totally isolatedfrom the upstream and downstream installations.According to IEC 60364 and NFC 15-100, thesecircuits do not have to be monitored specifically if:

c There is equipotentiality between theconductor components in the building and theframes, which is the case when the battery and

the DC circuits are in the same location, inelectrical cabinets or in the same cabinet asother UPS components (local equipotentiality ofthe UPS)

c The installation meets the requirements ofSection 413-2 of IEC 60364-4-41 or Section 412of NFC 15-100 (e.g. additional isolation bymeans of a class II link if the battery is locatedaway from the rest of the power supply)

Moreover, in other cases, the conditions to bemet by the installation (IEC 60364-5-55 andNF C 15-100 Section 554-2) of the battery andits connection as far as the “battery” circuit-breaker render the risk of faults highlyimprobable, and this is considered sufficient forprotecting persons against indirect contact onthis section of the installation.

However, the measures to be taken in order toensure personal protection on the DC section ofthe installation must be examined (this section isdelimited by the rectifier, the inverter and the“battery” circuit breaker).

6.1 Devices for monitoring DC circuits

As indicated before, the poles on a battery packfor a UPS are isolated. Therefore, permanentinsulation monitors (PIMs) are the devices mostfrequently used to monitor DC circuits, althoughdepending on the neutral earthing systems onAC installations, RCDs may also be used.

Permanent insulation monitors

c Permanent insulation monitor withlow-frequency current injection (2, 5 or 10 Hz)(see Fig. 41 )How it works: The PIM applies a low-frequencyAC voltage source between one of the poles onthe DC circuits and earth. An insulation fault onthe DC circuits will generate a current, which isdetected by the measurement circuits (seeFig. 42 ).

These monitors, which are also able to monitorAC, mixed and DC line supplies, also supporttroubleshooting for insulation faults. They aretherefore recommended if:v The line supply is a DC-only supply (severalloads), which is not the case with UPSsv There is no electrical isolation between thebattery and the equipment downstream of theUPS

Line supply to be monitored

50 Hz (or 60 Hz) rejection filterdepending on the supplyto be monitored

Measurement circuit

Voltage source(low frenquency or DC)

Fig. 41 : Schematic diagram of a permanent insulationmonitor (PIM) with current injection

Fig. 42 : Vigilohm XM200, a PIM with low-frequencycurrent injection and Vigilohm TR22A, a PIM withDC injection (from Merlin Gerin)


c Permanent insulation monitor with DC injection

The principle is the same as that used by thePIMs cited before although its source supplies adirect current. This type of PIM, designedspecifically for monitoring AC line supplies, is notsuitable for DC circuits, as the DC voltageappearing in the event of an insulation fault altersthe sensitivity of the monitor’s threshold device.It is generally not recommended on mixed(AC and DC) line supplies. However, someDC injection PIMs are capable of indicating thepresence of a fault on the “DC” section of a linesupply (see Fig. 42).

c Voltmetric-balance permanent insulationmonitor (see Fig. 43 )

This is a passive permanent insulation monitor.

Residual current devices (RCDs)

RCDs are designed to detect any abnormalearth fault currents.However, precautions must be taken whenselecting RCDs if the installation has one ACsection and one DC section without electricalisolation (mixed line supply).

Standard IEC 60755 defines 3 types of RCD:

c Class AC RCDs: Will only operate withAC-only line supplies

c Class B RCDs: Will operate with all faultcurrent shapes, including DC only

c Class A RCDs: Can operate withnon-alternating currents in accordance with thediagrams in Figure 45

With single-phase rectifiers, the waveforms offault currents on the DC circuit are most oftencompatible with type A RCDs. However, withsome systems, the fault current can be DC,requiring type B RCDs.

DC line supply to be monitored

Measurement circuit

+

–

Fig. 43 : Schematic diagram of a voltmetric-balancepermanent insulation monitor (PIM)

It comprises a resistive divider, which creates amid-point for the DC voltage. The fault detectioncircuit is located between this mid-point andearth. An insulation fault on the (+) or (-) willgenerate a current to earth via one of theresistors and the detection circuit, which,connected to a threshold device, will activate analarm or a trip (see Fig. 44 ).

Fig. 44 : Vigilohm TR5, a voltmetric-balance PIM (fromMerlin Gerin)

6 mA

α = 90°

α = 135°

Fig. 45 : Different non-alternating fault current shapesapplicable for class A RCDs

Finally, in most cases, DC fault currents forthree-phase rectifiers are DC. They thereforerequire the use of type B RCDs (see Fig. 46 onopposite page).

You must therefore contact the manufacturer tofind out which type of RCD can be connected tothe converter if this information has not beenprovided in the installation manual.


Disturbance due to filters

EMC filters based on capacitors connected toearth are frequently used upstream of computersand UPSs. They can cause disturbance onprotection devices (undesirable alarms and/ortrips), in particular:

c They often cause RCDs to trip unintentionally:

v During operation (earth leakage current due tocapacitor unbalance between phases and earth),the RCD threshold must therefore be increasedv On power-up (capacitor load), although anRCD time delay should rectify this

c When used with PIMs, filters can sometimesgenerate:

v A transient fault signal on power-up forDC injection PIMs (current flowing through thecapacitors)v Or a permanent signal with AC-injection PIMs

One possible way to avoid these phenomena isto make sure that the total capacity of thesefilters does not exceed:v 30 µF for a 2 or 5 Hz injection PIMv 6 µF for a 10 Hz injection PIM

00.020.01

Shape of fault currentRectifier type

0

Id

t

(s)

00

Id

t

00

Id

t

Fig. 46 : Different fault current shapes requiring class B RCDs

Disturbance due to high-frequency (HF)leakage currentsAlthough the high-frequency currents emitted byUPSs are considered weak in accordance withstandard IEC 62040-2, they may be harmful tosome residual devices and cause them to tripunintentionally. Moreover, HF currents from loadsconnected downstream of the UPS or even thesum of the UPS and load currents may cross theresidual devices during some operating sequences.The use of RCDs immune to this type of disturbanceis recommended in order to avoid these difficulties.Note: These leakage currents do not affect low-frequency current injection insulation monitors.

Interaction between monitoring devices onDC circuits and those on upstream anddownstream equipmentc This interaction is directly connected with theUPS system and depends in particular on:v The presence or absence of a static switchv The number of UPSs, one or more with activeor passive redundancyv The presence or absence of TR or TI electricalisolation transformers


c This interaction is directly dependent on theprotection device selected and the earthingsystems on the “upstream” and “downstream”equipment. Note:v For the most common system (independentUPS without transformer on bypass), it isessential that the downstream neutral earthingsystem is the same as the one upstream of theUPS (not excluding transition from TNC to TN-Sor TN to TT).v For continuity of service, a downstream (andupstream) IT system is the best solution.

c This interaction can be:v Positive, e.g. the upstream protection devicealso monitors the DC circuitsv Negative:- Between two PIMsAs on AC circuits, two devices of the same typeconnected on two sections of an installation

without electrical isolation will cause mutualdisturbance. A relay such as R1 on the systemillustrated in Figure 22 should be used to avoidthis eventuality.- Between an injection PIM and a voltmetric-balance PIMA DC injection or low-frequency PIM willmeasure the internal resistance (R/2) of avoltmetric-balance device (several tens of kW).Located either side of a power converter(rectifier or inverter) without electrical isolation,the extent to which each causes disturbance onthe other will be directly dependent on theconduction rate of the converter semiconductors.v Or zero:- If there is electrical isolation between thebattery and the upstream and downstreamequipment (AC)- Between PIM and RCD or circuit-breaker

6.2 Main applications

The solutions advocated in this sectionsupplement the measures to be taken for theprotection of persons on equipment upstreamof the UPS already described in Section 4.Without exception, they must be employed oneach UPS link.

If a UPS includes DC circuits, which have notbeen isolated from the upstream and/ordownstream equipment, and if the protection ofpersons requires the use of RCDs, these RCDsmust be type A or B devices (see previoussections).

In general, when DC circuits are isolated fromupstream and downstream equipments, one PIMmust be used: the PIM 3.

If the DC circuits are not isolated, protection forpersons should be considered on the basis ofthe electrical links between these circuits and theupstream and downstream equipment, resultingin the various scenarios outlined below.

DC circuits isolated from upstream anddownstream equipment (inclusion of rectifierand inverter transformers - see Fig. 47 )

There is no interaction with the monitoring devices.The PIM 3 designed for monitoring insulation onthe DC circuits may be a voltmetric-balance orlow-frequency current injection monitor.

Protection in the event of a second fault is thenensured:c By means of circuit-breaker H if the UPS isstandalone

c By means of circuit-breaker H and theupstream (or internal) protection on the rectifierif the UPS is connected to the line supply

DC circuits not isolated from the upstreamequipment (Inclusion of an inverter transformer(TI), but no rectifier transformer)

For the following scenarios, as a general rule, ifthe power supply voltage is present, protection isprovided by the upstream devices.

If this voltage is lost, the DC circuits are isolatedfrom upstream (rectifier blocked) anddownstream (via the inverter transformer). Thesecircuits then temporarily form part of an IT system.

C1 G1

Cn

TR

H

TI

PIM 3

R O

Gn

Fig. 47 : Insulation monitoring on a UPS link withbattery pack isolated from the upstream anddownstream equipment


As illustrated in the previous section, thisoperating mode does not put safety at risk as itis able to withstand a first fault.

In the event of a second fault, the battery circuit-breaker and/or internal inverter protectiondevices will provide protection.

However, if monitoring for these circuits and inparticular for the battery is required, in the eventof prolonged independent operation or if thebattery is located away from the UPS, loss ofvoltage results in the rectifier being blocked andthe DC circuits being isolated. A PIM can beinstalled to indicate the first insulation fault onthese DC circuits. PIM 3 features in the followingfigures and is connected automatically via avoltage relay (R3).

c Upstream equipment operating as a TT system(see Fig. 48 )

If there is voltage upstream of the rectifier, theRCD installed on the UPS power supply willprovide protection by tripping circuit-breaker C.Note: Put into operation by the R3 relay, PIM 3controls the insulation of DC circuits when theyare isolated.

C1

R3

G1

Cn

H

TI

PIM 3

R O

Gn

Fig. 48 : Insulation monitoring on a UPS link withbattery pack not isolated from the upstream equipmentoperating as a TT system and isolated from thedownstream equipment

C1

R3

G1

Cn

H

TI

PIM 3

R O

Gn

Fig. 49 : Insulation monitoring on a UPS link withbattery pack not isolated from the upstream equipmentoperating as a TN or IT system and isolated from thedownstream equipment

c Upstream equipment operating as an IT system(see Fig. 49 )

If there is voltage upstream of the rectifier, PIM 1on the upstream equipment monitors theinsulation of the DC circuit (positive interaction).

c Upstream equipment operating as a TNsystem (see Fig. 49 )

If there is voltage upstream of the rectifier, theprotection of persons can be ensured by:v The circuit-breaker C installed on the UPS, ifthe impedance calculation for the fault loop haspermitted confirmation of the selectionv An RCD or additional interconnection if not,with the TN-S systemNote: Put into operation by the R3 relay, PIM 3controls the insulation of DC circuits when theyare isolated.

Note: Put into operation by the R3 relay, PIM 3controls the insulation of DC circuits when theyare isolated.Protection in the event of a second fault is thenensured:v If the UPS power supply is supplied with powervia the circuit-breaker (C) installed on the UPSpower supply, if the impedance calculation forthe fault loop has permitted confirmation of theselection, or via an RCD if notv During independent operation, via the batterycircuit-breaker

Special cases of UPSs supplied with power viaan intermediate power outlet, with a maximumrated current equal to 32 A.

These power outlets must be protected bya high-sensitivity RCD, I∆n < 30 mA(NF C 15-100 § 532). Also, if these UPSs do notfeature a rectifier transformer and specificmeasures (class II, double isolation) have notbeen taken, this protection must be provided bya high-sensitivity type A RCD.

DC circuits not isolated from the downstreamequipment

(Inclusion of a rectifier transformer (TR) but noinverter transformer)

In this case, if the neutral systems upstream anddownstream of the UPS are not identical, anisolating transformer (TB) must be installed.


c Downstream equipment operating as aTT system

v Without bypass transformerIf there is voltage upstream of the rectifier, theRCD installed on the bypass will provide protectionin the event of a fault on the DC circuits bytripping the battery circuit-breaker H and thecircuit-breakers C and Cb installed on theUPS power supply (see Fig. 50 ).If this voltage is lost, the DC circuits and theupstream equipment will continue to be protectedby the same RCD if the neutral conductor remainsconnected. This requires that the auxiliary power

supply for this RCD is connected downstream ofthe UPS.If the neutral conductor is broken, the DC circuitsand the downstream equipment will switch toIT system operating conditions.v With bypass transformer (TB)This same protection is provided permanently byan RCD installed on the neutral earth connectionat the UPS output (see Fig. 51 ).With or without a bypass transformer, theseconfigurations do however present a problemas they totally disconnect the power supply toall loads. Also, if continuity of service is essential

A

B

Cb D

TR

SS

E

F

G

U3

U2

U1C

H

R O


(*)

A

B

Cb D

TR

TB

SS

E

F

G

U3

U2

U1C

H

R O

(*)


Fig. 50 : Devices for protecting persons on a UPS link for DC circuits isolated from the upstream equipment butnot from the downstream equipment functioning as a TT system, without a bypass transformer

Fig. 51 : Devices for protecting persons on a UPS link for DC circuits isolated from the upstream equipment butnot from the downstream equipment functioning as a TT system, with a bypass transformer (TB)


and in particular in the case of redundant UPSs,the following measures can be taken. Theyconsist in isolating the fault link from the rest ofthe installation. For this purpose:- Every output in the link is permanentlymonitored by an RCD, which will provideprotection by tripping the circuit-breakers C andH and the switch or circuit-breaker J on the faultylink (see Fig. 52 ).- In the absence of a bypass transformer, thediscrimination must be checked between theupstream RCD and the downstream RCDs. If thetripping of output switch G can be controlled,device J does not have to be included as theRCD can act directly upon it.

c Downstream equipment operating as aTN system

To protect persons, usually, as soon as aninsulation fault occurs, circuit-breakers C and Htrip. If the conditions required for correctoperation are not met, in particular for the TNSsystem, the measures described in the previoussection (for the TT system) must be taken.

c Downstream equipment operating as anIT system

The DC circuits are monitored by:v PIM 1, if the bypass maintains electrical continuityv PIM 2, if the UPS is providing an autonomouspower supply to the loads (see Fig. 53 ).

A

BCb D

TR

SS

E

F

G1 J1

U3

U2

U1

C1

HR O

Cn Gn Jn

(*)


Fig. 52 : Measures to be taken to isolate a UPS link in the event of an insulation fault, for UPSs isolated from theupstream equipment but not from the downstream equipment operating as a TT system

A

B

Cb D

TR

SS

E

F

G

U3

U2

U1C

H

R O

R2 PIM 2

(*)

PIM 1

R1


Fig. 53 : Devices for protecting persons on a UPS link for DC circuits isolated from the upstream equipment butnot from the downstream equipment functioning as an IT system, without a bypass transformer


PIM 2 is started up via relay R2 only in the eventof the loss of the voltage upstream of the bypass.Despite the absence of the upstream voltage, itis possible, even very probable, that the neutralconductor will remain connected. A circuit-breaker will trip either in the event of a fault or formaintenance purposes.If PIM 2 is at risk from disturbance from PIM1, itcan be shut down by relay R1 if the transformersecondary voltage is zero.Protection in the event of a second fault is thenensured:v In a TN system if the frames are connectedv In a TT system if the frames are not connected

DC circuits not isolated from upstream anddownstream equipment

In this case, the neutral systemsupstream and downstream of the UPS must beidentical.

c Equipment operating as a TT system (seeFig. 54 )

If there is voltage upstream of the rectifier andbypass, the RCD installed on the common powersupply will provide protection by tripping circuit-breakers C and Cb and the battery circuit-breaker H.If the power supply upstream of the UPS is lostand if the neutral conductor remains connected,this RCD will continue to provide protection, aslong as its power supply is connected to theUPS output.If the neutral conductor is broken, operation willswitch to an IT system.This configuration does however present aproblem as it totally disconnects the powersupply to all loads.If continuity of service is essential and inparticular in the case of redundant UPSs, usethe configurations in Figure 55 .

A


BCb D

SS

E

F

G

U3

U2

U1C

H

R O

(*)

Fig. 54 : Protection of persons for DC circuits in a UPS link not isolated from the upstream and downstreamequipment operating as a TT system -Basic implementation-

A

B

Cb DSS

E

F

U3

U2

U1

H

R OG1 J1

Gn

C1

Cn Jn

(*)


Fig. 55 : Protection of persons for DC circuits in a UPS link not isolated from the upstream and downstreamequipment operating as a TT system -Implementation for optimizing continuity of service-


The RCDs installed upstream of the rectifier anddownstream of the inverter on each link will tripcircuit-breakers C and J on each of the links.Discrimination between these RCDs and theRCD upstream of the bypass channel must bechecked in order to ensure the availability of thischannel.

c Equipment operating as a TN system

To protect persons, usually, as soon as aninsulation fault occurs, circuit-breakers C and Htrip. If the conditions required for correctoperation are not met, for example for the TNSsystem, the configurations described in theprevious section (for the TT system) must beused.

c Downstream equipment operating as an ITsystem

The DC circuits are monitored by:

Fig. 56 : Protection of persons for DC circuits in a UPS link not isolated from the upstream and downstreamequipment operating as an IT system

A

B

Cb DSS

E

F

G

U3

U2

U1C

H

R O

R1

R2

(*)


PIM 1

PIM 2

v PIM 1, if there is voltage upstream of the UPSv PIM 2, if the UPS is providing an autonomouspower supply to the loads (see Fig. 56 ).PIM 2 is started up via relay R2 only in theevent of the loss of the upstream voltage.Despite the absence of the upstream voltage, itis possible, even very probable, that the neutralconductor will remain connected. A circuit-breaker will trip either in the event of a fault orfor maintenance purposes.If PIM 2 is at risk from disturbance from PIM 1,it will be shut down by relay R1 if thetransformer secondary voltage is zero.Protection in the event of a second fault is thenensured:v In a TN system if the frames are connectedv In a TT system if the frames are notconnected


7 Conclusion

This document demonstrates that setting upequipment to protect persons against indirectcontact on line supplies with DC circuits isactually more complex than it first seems.

Particular attention must therefore be paid to thevarious configurations that can be installed ormay occur during operation, for example in theevent of a fault.

This is why it is often advisable to use equipmentsupported by thorough documentation andassistance from specialists or professionalsystem builders to obtain a complete and reliablesolution for a specific requirement in a givenconfiguration.


Bibliography

Standards

c IEC 60364, NF C 15-100: Low-voltageelectrical installations

c IEC and EN 60439-1: Low voltage switchgearand controlgear assemblies Type-tested andpartially type-tested assemblies

c IEC 60529: Degrees of protection provided byenclosures (IP code)

c IEC 60755: General requirements for residualcurrent protection devices

c IEC and EN 62040: Uninterruptible powersystems (UPS)v Part 1-1: General and safety requirements forUPS used in operator access areasv Part 1-2: General and safety requirements forUPS used in restricted access locationsv Part 3: Method of specifying the performanceand test requirements

Guide

c UTE C 15402: Guide pratique. Alimentationssans interruption ( ASI ) de type statique. Règlesd’installation.

Schneider Electric Cahiers Techniquesc Residual current devices in LV,R. CALVAS, Cahier Technique no. 114

c Earthing systems in LV,B. LACROIX, R. CALVAS,Cahier Technique no. 172

c Earthing systems worldwide and evolutions,B. LACROIX, R. CALVAS,Cahier Technique no. 173

c The IT earthing system (unearthed neutral)in LV, F. JULLIEN, I. HERITIER,Cahier Technique no. 178

Schneider Electric Direction Scientifique et Technique,Service Communication TechniqueF-38050 Grenoble cedex 9Télécopie : 33 (0)4 76 57 98 60

E-mail : [email protected]

Transl: Lloyd International - Tarporley - Cheshire - GBDTP: AXESS - Valence (26)Editor: Schneider Electric

- 20 € - © 2

004

Sch

neid

erE

lect

r ic

09-04xxxxx

Uninterruptible static power supplies and the protection of persons

Documents

Transcript of Uninterruptible static power supplies and the protection of persons