LHC Project Note 153 Preliminary study Powering the Transfer

This is an internal CERN publication and does not necessarily reflect the views of the LHC project management.LHC Project Note 153

1998-07-22

Paolo.Burla@cern.ch

Preliminary study: Powering the Transfer Lines from SPS to LHC Paolo Burla / LHC-CRI, John Pedersen / SL-PO Keywords: Transfer Lines, Powering, Power Converters

Summary

After the decision to realise the transfer lines between SPS and LHC with room temperature magnets, a preliminary study of the powering scheme for these facilities has been performed.The purpose of this note is to present the method used for this preliminary study and the results obtained. This study proved the feasibility of a room temperature solution. The results of this note were presented and discussed in the group SL/PO responsible for the powering scheme completion. The LHC Technical Co-ordination Committee at the “Technical Review”of October 1997 [3], [4], approved these results.

1. Introduction and Background

Up to late 1994 different designs were considered for the new transfer lines between the SPS and the LHC. The use of superconducting magnets was at the base of all these schemes.

A cost estimate suggested that a substantial overall cost saving could result from the elimination of the cryogenic system. In addition a powering estimate showed that the existing power distribution system at SPS should be able to tolerate the added power request with only some reasonable adaptations.

Injection into P2 and P8 of LHC and the fact that the transfer lines will be required to operate no more than a few hours per day have now resulted in the choice of small aperture,room temperature magnets.

The LHC Technical Committee [2], after discussion of the above considerations, adopted the warm transfer line solution at its eighth meeting on June 6, 1995.

At the time of this preliminary study, the requirements for the Neutrino line had not been defined and therefore are not taken into consideration.

2. Power conversion

2.1Basic philosophy

The Transfer Lines from SPS to LHC are considered as extensions of the SPS machine.Thus substantial savings can be achieved by re-using, wherever possible, existing infrastructures such as buildings, galleries, power conversion systems, etc.

The application of this concept led to design the transfer lines arc lattice main magnets in such a way that existing LEP dipole and quadrupole power converters can be reused.

At the centre of this preliminary study is a set of small databases. These are interrelated and generated using FileMaker-Pro 3.0. Within each database, a set of calculation is performed. The results allow the selection of the adequate power converters, based on an input consisting of basic magnet circuit data. The D.C. cable characteristics are part of the input data.

The structure of this database [5] is shown in the figure below:

Fig. 1: Structure of the Transfer Lines Powering Database The transfer lines will be subpided in the usual manner into the following sections:

1.Extraction from SPS

2.Line matching

3.Arc lattice

4.Injection matching

5.Injection into LHC.

With the exception of the elements for injection into LHC, the above method has shown that existing LEP or SPS power converters having the required ratings can be found in almost all cases.

The list of existing power converters types, able to fulfil the transfer lines powering requirements, is shown in Fig. 2.

- 2 -

Fig. 2: List of existing PC types selected for powering the transfer lines from SPS to LHC

- 3 -

It should be noted that in a limited number of cases, the power converter type foreseen to be reused, might need such an extensive renovation that it might be better to install new equipment instead.

The elements for the extraction from SPS, for the line matching and for the arc lattice main magnets are powered from power converters located in SPS buildings (BA4 and BA7) and fed from the SPS pulsed network. In this way the existing reactive power compensation system, required for pulsing applications for the SPS, can also be reused for the injection tunnels.

On the other hand, the power converters for injection matching and injection into LHC elements are located in LHC buildings (SR’s) where the power grid should not be disturbed by pulsing equipment. Therefore these elements will be used in D.C. mode.

The transfer lines powering study described in the following paragraphs are based on data issued in [1] and arising from beam optics considerations.

When ref. [1] was edited, the correctors for the transfer lines were not known yet and therefore are not taken into account in this study.

Although some line optics parameters may have changed since then, the developed method for powering calculation remains valid and the order of magnitude of the overall electrical powering data should be only marginally affected.

However the cost of the cabling of the missing correctors may be significant.

2.2Power converters for TI2

Particles extracted from point 6 of SPS are transported to the point 2 of LHC in order to be injected in the clockwise direction in beam 1. This is done by means of a transfer line installed in the new tunnel TI2.

2.2.1Equipment fed from the SPS

The initial part of TI2, on the SPS side of the transfer tunnel, contains the elements for the extraction from SPS, for the line matching and for the arc lattice. This set of elements constitutes a system of 9 pulsed warm magnet circuits.

Each circuit is fed from an inpidual power converter, located in the SPS building BA7.

For the calculation of the power requirements, it has been assumed that the current cycle applied to the pulsed warm magnet circuits in TI2 will be the SPS injection and acceleration current cycle for filling the LHC as given in the "Pink book”. The flattop current data are taken from [1] and the power converters maximum D.C. ratings are resulting from the calculation achieved through the database mentioned in 2.1. The type of power converter is chosen from the list shown in table 1, according to these figures.

Fig. 3 below shows the power cycles for the whole set of pulsed circuits (addition of the 9 inpidual power converter cycles). This was generated using the Signal Analysis Software FAMOS 3.0.

- 4 -

- 5 -

-2

02468101214161820MW MVA -202468101214161820

MVar

2

4

6

8

10

12

14

16

18Sec

Fig. 3: Active, Reactive and Apparent Power for the pulsed part of TI2 injection line The connections between the power converters located in BA7 and the circuits of TI2are made with a partially existing D.C. cabling system, routed through the pit PA7, the access tunnel TA7 and the initial section of TT60.

The TI2 Main Dipole circuit will consist of 112 MBI type magnets connected in series.The current for this circuit will be supplied by a renovated LEP main dipole power converter rated for the peak figures of 1607 V / 5480 A.

The focusing quadrupole circuit QF will consist of a series of 40 MQI type magnets whereas the defocusing quadrupole circuit QD will contain 41 such magnets in series. A refurbished LEP quadrupole power converter will be connected to each of these circuits. It might be convenient to match the maximum D.C. output voltage (1300 V) of this type of power converter with the peak load voltage of the circuit (720 V to 900 V) so as to reduce the reactive power swing during the current cycle.

The power losses in the air due to the D.C. cable resistance must be kept as low as possible due to the small size of the transfer tunnel. The cabling in the underground facilities has been designed to reduce the heat dissipation to the air to an acceptable level (see Table 1and Table 2).

The Fig. 4 below contains detailed information about the magnet circuits powered from the building BA7, in particular data related to the D.C. cabling.

2.2.2

Equipment fed from LHC

The part of the TI2 Transfer Line, near the LHC tunnel in point 2, accommodates the elements for Injection Matching and for Injection into LHC. A total of 12 D.C. warm magnet circuits make up these 2 sections.

Each circuit is connected to an inpidual power converter located in the building SR2 by means of the partly reused LEP D.C. cabling system.

In the Fig. 5 below detailed information about magnet circuits powered from the building SR2 is given, in particular data related to the D.C. cabling.

Peak Apparent Power: 18.2 MVA Peak Reactive Power: 16.1 MVar

MVA

Peak Active Power: 12.2 MW

Fig. 4: TI2 pulsed circuits powered from BA7 with power summary and cabling data

- 6 -

Fig. 5: TI2 D.C. circuits powered from SR2 with power summary and cabling data

- 7 -

- 8 -

2.3Power converters for TI8

Particles extracted from point 4 of SPS are transported to the point 8 of LHC in order to be injected in beam 2 (anticlockwise). This is done by means of a transfer line installed in the new tunnel TI8.

2.3.1

Equipment fed from the SPS

The initial part of the TI8 on the SPS side of the transfer tunnel contains the elements for the extraction from SPS, for the line matching and for the arc lattice. All these elements are forming a system of 16 pulsed warm magnet circuits.

With the exception of the circuit MBB802, actually an SPS main dipole magnet in series with those of the SPS ring, each of the 15 magnet circuits will be connected to an inpidual power converter located in the SPS building BA4.

For the calculation of the power requirements of the 15 remaining pulsed circuits of TI8,it has been assumed that the SPS injection and acceleration current cycle for filling the LHC given in the "Pink book” is applied to the warm pulsed circuits of TI8. The flattop currents are taken from [1] and the maximum D.C. ratings are resulting from the calculation made through the database mentioned in 2.1. The type of power converter is chosen from the list shown in table 1, according to these figures.

In Fig. 6 below the power cycles for the whole set of TI8 pulsed circuits (except

MBB802) is shown. This was generated using the Signal Analysis Software FAMOS 3.0

-5

05101520253035MW MVA -5

05101520253035

MVar

2

4

6

8

10

12

14

16

18Sec

Fig. 6: Active, Reactive and Apparent Power for the pulsed part of TI8 injection line The connections between the power converters located in BA4 and the magnet circuits of TI8 will be realised by means of a partially existing D.C. cabling system routed through the pit PP4, the cavern ECX4 and the new tunnel TT40.

Peak Apparent Power: 32 MVA

Peak Reactive Power: 28.5 MVar Peak Active Power: 21.5 MW

The TI8 Main Dipole circuit will consist of 236 MBI type magnets connected in series. The current for this circuit will be supplied by a power converter system composed of 2 refurbished LEP main dipole power converters connected in series in order to get the required peak figures of 3200 V/5600 A.

The voltage drop due to the length of the D.C. cables between the Main Dipole circuit and the associated power converter must be minimised since the magnets themselves require a voltage very close to the maximum of the power converter. This requires a connection with a minimum total cross section of 1500 mm2 per polarity. The ohmic losses from such a heavy link can not be dissipated to the air. This link will have to be made up of water-cooled cables.

The focusing quadrupole circuit QF and the defocusing quadrupole circuit QD will each consist of a series of 35 MQI type magnets. A refurbished LEP quadrupole power converter will be connected to each of these circuits. It might be convenient to match the maximum D.C. output voltage (1300 V) of this type of power converter with the peak load voltage of the circuit (750 V to 785 V) so as to reduce the reactive power swing during the current cycle.

The cabling will be designed according to the same philosophy as for TI2.

Fig. 7 below contains detailed information about the magnet circuits powered from the building BA4, in particular data related to the D.C. cabling.

2.3.2Equipment fed from LHC

The extremity of the TI8 Transfer Line, near the LHC tunnel in point 8, accommodates the elements for Line Matching and for Injection into LHC which constitute a system of 11 D.C. warm magnet circuits.

These circuits will be connected to inpidual power converters located in the building SR8 by means of the partly recuperated LEP D.C. cabling system.

In Fig. 8 below detailed information about magnet circuits powered from the building SR8 are given, in particular data related to the D.C. cabling.

- 9 -

Fig. 7: TI8 pulsed circuits powered from BA4 with power summary and cabling data

- 10 -

Fig. 8: D.C. circuits powered from SR8 with power summary and cabling data

- 11 -

3.Power dissipation

The power dissipated by the various components of the transfer lines will be absorbed either by the water cooling system or in the air by the ventilation system. A good knowledge of the power dissipation in the various areas containing equipment for the transfer lines is required to specify both cooling systems.

A breakdown of the power dissipation in both cooling media, ordered by location, is shown in the following paragraphs.

3.1Power dissipated in TI2

The breakdown of the power dissipation in the various locations related to TI2 is given in Table 1 below.

Table 1: Breakdown of TI2 Power dissipation by location and cooling medium

3.2Power dissipated in TI8

The breakdown of the power dissipation in the various locations related to TI8 is given in Table 2 below.

Table 2: Breakdown of TI8 power dissipation by location and cooling medium

- 12 -

4. A.C. Power distribution

4.1General

This chapter describes the power distribution installations to be realised for the LHC injection power converters. The preliminary a.c. power distribution design, as presented in the following, is based on a.c. power requirements for the power converters, needed for the injection tunnels, as defined previously. The design guidelines for the general services system are not presented.

4.2Power requirements, reactive power compensation

The peak value of the total apparent power, needed for the power converters of the two injection tunnels, amount to 50 MVA. With the given pulse shape, this will correspond to about 22 MVA r.m.s (see Fig. 3 and 6).

As described earlier, an important number of magnet circuits of both tunnels will be operated in pulsed mode, and their power converters will therefore have to be supplied by the pulsed network of SPS. It is not feasible to supply them from the stable network of the LHC, as the pulsing may perturb the mains and have an adverse effect on the stability of the machine power converters.

The peak values for the required active and apparent power for the magnet system, operated in pulsed mode, are shown in table 3.

TI218.2 MVA12.2 MW16.1 Mvar.

TI832.0 MVA21.5 MW28.5 Mvar

Table 3: Power requirements for converters in pulsed mode The rest of the magnet circuits will be operated in D.C. mode. All of these will be fed from the LHC side.

On the LHC side, i.e. in access points 2 and 8, ample power is available. The demands for the injection tunnel installations amounts to 1.25 MVA for TI2 and 1.75 MVA for TI8, which is small compared to the reserves available.

On the SPS side, the situation requires a closer examination. The power converters would ideally be fed from the so-called SPS pulsed loop. This is an 18 kV cable loop, present in all BA buildings. It feeds all the auxiliary converters of the SPS, which will operate in synchronism with the pulsed converters for the injection tunnels. The 18 kV cable, constituting this loop, has a rated current corresponding to an apparent power of 20 MVA, each side. The SPS load is about 4 MVA, each side.

The r.m.s. value of the apparent power of TI2 is about 9 MVA. For TI8 the value is about 16 MVA. Adding the load of TI8, would bring the load on the side including BA4 to the limit of the capacity of the cable of the SPS pulsed loop.

Another critical parameter is the reactive power compensation. The compensation of the pulsed loop is provided by the Main Compensator in the Prévessin main substation. The limiting parameter is the dynamic range of the compensator, which is 110 Mvar. Operating at a beam energy of 450 GeV, the SPS already creates a reactive power swing of 85 Mvar, leaving a free capacity of 25 Mvar.

- 13 -

The peak value of the total reactive power of both tunnels is about 45 Mvar (see table 3). These 45 Mvar have to be compensated to avoid voltage variations in the power system. It is thus not possible to feed the converters for both TI2 and TI8 from the pulsed loop. Even TI8 alone would bring the compensator to the limit of its capacity. This solution should thus be avoided.

However, CERN has a second large compensator, the North Compensator, identical to the Main Compensator. It provides reactive power compensation for the installations in the North Area, and operates with the present load at a maximum reactive power swing of 72 Mvar. It offers thus a reserve of 38 Mvar. Unfortunately, this compensator does not have any links to either of the two proposed installation sites in the SPS: BA 4 and BA 7.

To obtain compensation of the reactive power of the two injection tunnels the following solution will be applied: The load of TI2 ( BA 7 ) will be connected to the Main Compensator via the pulsed loop, and the load of TI8 ( BA 4 ) will be connected to the North Compensator. To realise the latter connection, it is necessary to lay an 18 kV cable link from the Prévessin Main Sub-station to the BA 4 building. The distance from Prévessin to BA 4 is much shorter than from Prévessin to BA 7, and the cable routing is mainly across open land, along existing CERN buried links.

This solution is far cheaper than the installation of a dedicated reactive power compensation plant in BA 4.

5.Power distribution for the power converters

5.1Power distribution installations for the converters

The need to have the installations partly on the LHC side and partly on the SPS side means, for both injection tunnels, that there will be power distribution system extensions both in the LHC and in the SPS, i.e. installation work in SR 2, SR 8, BA 4 and BA 7.

The major power converter units, for the dipole magnet circuits and for some of the quadrupole magnet circuits, will have dedicated 18 kV feeders. The smaller converter units will be fed from a 400 V distribution switchboard

On the LHC side the work will be minor: A reorganisation of the low voltage switchboards, feeding small power converters.

On the SPS side the interventions will be much more important. However, the underlying principle in the chosen solutions is to avoid civil engineering. The only new construction being the necessary transformer pits for the converter transformers.

For TI2 the equipment on the SPS-side will be installed in the SPS access building BA 7, where space is available. In this building an 18 kV switchboard will be installed downstream of the existing pulsed 18 kV loop. New low voltage equipment will be installed to feed the smaller converter units (see Fig.9).

On the LHC side the equipment will be installed in building SR 2. Space will be available after dismantling of LEP, and the power distribution infrastructure is already present.

For TI8 the equipment on the SPS-side will be installed in building BA 4. The 18 kV supply will be brought to the site directly from the Prévessin Main Sub-station via a new cable link. A new 18 kV switchboard will be installed in BA 4 (see Fig. 10).

- 14 -

On the LHC side the equipment will be installed in building SR 8. Space will be available after dismantling of LEP, and the power distribution infrastructure is, in general, already present.

Outside BA 4 and BA 7 transformer pits for the power converter transformers will be made. At BA 4, the cabling from the pit to the building will require a short, connecting duct or gallery. Depending on the size of the oil-retention basin in place, and on the oil volume of the largest transformer to be installed, the capacity of the retention system may have to be augmented.

The remaining circuits, 12 in TI2 and 11 in TI8, will be operated in D.C. mode. These are the ones to be installed in LHC (see tables 4 and 5 for the distribution between SPS and LHC).

Table 4 recapitulates the circuits in the injection tunnel TI2, table 5 those in TI8. The tables indicate the power need of each circuit, the location where the power converter will be installed and the voltage level of the mains, from which the converter will be fed: 18 kV or 400 V.

Both the active and the reactive power consumption by the units to be installed in LHC can easily be coped with by the present LEP installations.

The power requirements are expressed in r.m.s. terms as well as in terms of peak power. The parameter used is the apparent power, as this determines the current. The r.m.s. value is used to determine the distribution cabling and switchgear.

This note is based on calculations made before the details of the eight LSS 4 extraction bumpers were available. The power supply and the converters for these circuits will be installed in BA 4. Their power requirements are limited, in total about 1.2 MVA. This additional load can be accepted within the framework of the proposed solution.

- 15 -

Fig. 9: New 18 kV installation in BA 7

- 16 -

Circuit no Circuit Location Power/kVA

r.m.s., 3 ph

Power/kVA

peak, 3 ph

Mains Feeder

rating/A

Cable

section

1MBI 218BA 765501310018 kV750 2MQI 205BA 742084718 kV630 3MQI 206BA 742084718 kV630 4MBAH 291SR 2215429400 V4001260 5MSIW 294SR 2140279400 V2501260 6MBAV 203BA 7800160718 kV1200630 7MBAV 261BA 78001590400 V12001260 8MBAV 284SR 290173400 V160630 9MQI 201BA 71326400 V32240 10MQI 202BA 73056400 V60240 11MQI 203BA 73058400 V60240 12MQI 204BA 74587400 V100240 13MQI 286SR 23366400 V60630 14MQI 287SR 23058400 V60630 15MQI 288SR 21428400 V40240 16MQI 289SR 22549400 V60240 17MQI 290SR 22549400 V60240 18MQI 291SR 21019400 V16240 19MQI 292SR 2612400 V16240 20MQI 293SR 21836400 V40240 21MQI 294SR 21836400 V40240

Table 4: Power needs for inpidual circuits in TI

- 17 -

Circuit no Circuit Location Power/kVA

r.m.s., 3 ph

Power/kVA

peak, 3 ph

Mains Feeder

rating/A

Cable

section

1MBI 822BA 42677518 kV1500 2MQI 808BA 481218 kV630 3MQI 809BA 484718 kV630 4MBSL 801BA 4150304400 V2501260 6MBAS 804BA 4613400 V10240 7MBAH 884BA 4200398400 V4001260 8MSIW 887BA 4200375400 V4001260 9MBAV 817BA 47001382400 V12001260 10MBLV 883SR 8150283400 V250630 11MQSL 801BA 4175357400 V4001260 12MQTL 802BA 44078400 V60240 13MQTL 803BA 43570400 V60240 14MQTL 804BA 41021400 V10240 15MQTL 805BA 41633400 V30240 16MQTL 806BA 43568400 V60240 17MQI 807BA 42754400 V60240 18MQI 878SR 84077400 V60630 19MQI 879SR 85097400 V100630 20MQI 880SR 84284400 V60630 21MQI 881SR 84080400 V60630 22MQI 882SR 85095400 V100630 23MQI 883SR 83056400 V60240 24MQI 884SR 83569400 V60240 25MQI 885SR 82038400 V40240 26MQI 886SR 83058400 V40240 27MQI 887SR 81836400 V40240

Table 5: Power needs for inpidual circuits in TI8

- 18 -

5.2High voltage installations

The largest power converters, each having its own rectifier transformer, will be supplied from dedicated 18 kV feeders.

For TI2 one 18 kV feeder will be used for the main bending magnet circuit, MBI 218, one for the vertical bending magnet circuit, MBAV 203, and two for the quadrupole magnet circuits, MQI 205 and MQI 206. To these must be added an incomer to the 18 kV switchboard. This new switchboard, to be called EMD3/A7, will be fed from the existing switchboard: EMD1/A7 (see Fig. 9).

For TI8 one 18 kV feeder is foreFn for the main bending magnet circuit, MBI 822, and two for the quadrupole magnet circuits MQI 808 and MQI 809. One additional circuit breaker is necessary as incomer for the switchboard.

The 18 kV installation for TI8 will be an antenna at the end of the future cable link foreseen from the main Prévessin sub-station to the BA 4 building. The new switchboard, to be called EMD3/A4, will be fed from the existing switchboard in the main substation: EMD1/BE. It will have a back-up supply from EMD1/A4 (see Fig. 10). The cable link will be laid in earth, and rated to the r.m.s. load of the pulse.

The 18 kV switchgear will be modern, standard equipment with standard protection. The upstream feeders of the EMD1/A7 and EMD1/A4 switchboards are existing equipment.

Table 6 shows the numbers of 18 kV feeders to be installed:

Tunnel Building Number of feeders

TI2BA75

TI8BA 46

Table 6:18 kV feeders for LHC injection tunnels

5.3Low voltage installations for small power converters

The smaller converter units will be fed from dedicated low voltage switchboards. In BA 7 a total of 5 feeders are needed. The total estimated r.m.s. current of the converters to be fed is 1400 A, the peak current about 2800A. The thermal load would require a switchboard of two columns, equipped with an incomer and 5 fused switch feeders.

In BA 4 a total of 10 feeders are needed. The total estimated r.m.s. current of the converters to be fed is 2000 A, the peak current about 4000A. The thermal load would require a switchboard of three columns, equipped with an incomer and 10 fused switch feeders.

The switchgear will be new, as the existing is obsolete and no longer up to the present operational and safety standards.

- 19 -

6.Cabling

6.1 D.C. Cabling

This chapter describes the D.C. cabling between the power converters and the magnet circuits.

The cables between the power converters and the magnets are selected from the normalised standard range of power cables. The cross sections mentioned in this paper are based on the results of the calculations using the database. The detailed study of the cabling of the inpidual circuits may lead to slight changes.

6.1.1Water cooled cables

Water-cooled cables will be used to connect the converters to the end of the dipole magnet chains. An effort is being made to maximise the recuperation of water cooled cables from decommissioned SPS installations. However, the moving of such cables, as well as the fittings used, makes it financially uninteresting to recuperate lengths shorter than about 100 m. Tables 4 and 5 show the chosen cross sections.

6.1.2Conventional D.C. cables

All other magnet circuits will be realised using conventional, air-cooled cables. The parameter that determined the cross sections of these cables was the amount of losses to the air. Tables 4 and 5 show the chosen cross sections.

6.2 A.C. Cabling

The A.C. cabling in the injection tunnels will be limited to a 400 V installation for general services: lighting, socket boxes etc. However, the length of the tunnels means that extreme care has to be taken in the design of the cabling, in particular concerning the cabling for the magnet installation vehicle. The normally used cross sections for a given power will lead to unacceptably high voltage drops. For reasons of cost and limitations on available space, one cannot just increase the cross sections. The detailed study will contain a proper segmentation of the load, with distribution on different circuits.

7.Planning

The overall LEP planning requires the installations of the two injection tunnels to be performed as follows.

Power distribution Magnet cabling

TI8:May - October 2002November 2002 - July 2003.

TI2:February - July 2004June 2004 - March 2005

Well before the work is put in hand, a precise planning of the cabling must be made. This is essential, because during the cable laying, a tunnel, or a tunnel segment, is not available for other works.

- 20 -

8.Safety aspects

8.1Emergency stop system

The general emergency stop in a tunnel will cut all feeders supplying circuits in the tunnel or a closed segment of it. Since the tunnels are fed both from SPS and LEP, the emergency stop system will be particular. The segmentation of the tunnel due to radiation protection may complicate matters further.

Equipment fed from UPS systems can be maintained even after an emergency stop, provided they are properly signalled. The magnet systems should be part of the machine interlock.

8.2Materials

Like in all CERN installations, and in particular underground installations, all materials in the power distribution and cabling system will respect the CERN safety instruction SI 23. This safety instruction allows only the use of halogen free plastics that are flame retarding and do not give of neither opaque, toxic nor corrosive fumes if subject to fire.

- 21 -

本文来源：https://www.bwwdw.com/article/bisq.html