4 October, 2000

 

TO: DISTRIBUTION

FROM: C NEUMEYER

SUBJECT: OPTIONS FOR NCSX POWER SUPPLY SYSTEM

 

This memo presents some thoughts on options for implementation of the NCSX power supply system. I think that we should call a meeting including all of the electric power engineers here at PPPL, along with interested NCSX parties, to discuss. Then after refinement, and maybe with the development of some cost information, the choices should be presented to lab management.

 

Background Information

 

1)      D-site Power Supply (PS) Units are 12-pulse rectifiers consisting of two 6-pulse Power Supply Sections (PSS) each pulse rated 1kV/24kA-6sec, once every 300 sec., and continuous rated at +/-1kV/3.25kA. In some cases we have operated up to 30kA with reduced pulse length. So each PS is pulse rated ≈ 2*24 = 48MVA, and continuous rated ≈ 2*3.25 = 6.5MVA. Dimensions of a PS unit are approx. 12’ wide x 12’ tall x 4’ deep per PS, not counting the cable throat in the rear in which the transformer-to-rectifier cables are routed, along with the ACCTs.

2)      Each PS is fed by a single three winding transformer (XFMR), oil filled and installed outdoors.

3)      The PSS are fully insulated from one another, but subject to a common firing generator/fault detector unit such that they operate in identical modes at identical firing angles. However, control modifications would permit independent control of the two PSS.

4)      The PSS are 2-quadrant, unipolar current, meaning +/- V but +I only. For bipolar current, need to connect pairs of PSS in anti-parallel.

5)      Equivalent internal impedance of a PSS is of order 13.3mΩ, meaning that a 400V drop will occur at 30kA.

6)      There are 37 fully rated PS, hence 74 fully rated PSS. There are two additional PS which were used for the TFTR HF and VC systems but these are partially rated and use smaller indoor dry type transformers, which are trapped in the building by the other equipment.

7)      Each PSS consists of six parallel 6-pulse rectifier modules, and two parallel bypass modules. For an NCSX pulse with t ≤ 3 seconds, only 1/2 of the modules would be required. Furthermore, the bypass modules are probably not required for NCSX.

8)      Based on certain assumed rates of inflation the cost of each XFMR in Y2K $ would be $266K, and each PS would be $344K, total $610K per PS/XFMR unit. This amounts to $12/KVA pulsed, $94/KVA continuous.

9)      NSTX requires a minimum of 21 of the 37 fully rated PS, based on the present configuration, with the further assumption that the presently available on-line spares would be sacrificed to NCSX. This leaves 16 PS units for NCSX, total 32 PSS (see list which follows).

 

 

10)  The PS units in the east-west wing of the FCPC building, formerly the TFTR TF power supplies, are presently used by the NSTX TF system and OH system. Because these units were connected in four parallel series strings, and because of the high door located at the end of the aisle, these 12 PS units could be disconnected with relative ease and removed.

11)  The PS units in the north-south wing of the FCPC building, formerly the TFTR PF power supplies, are presently used by the NSTX PF and CHI systems. These units are interconnected with each other and with other specialized components in such a way that significant re-cabling would be required to separate them from the NSTX circuits. Also, there is no simple way to physically remove them from the building. However, their power modules, Master Gate Drive (MGD) systems, and their XFMRs are modular and could be easily removed and relocated. In this case new frames, with suitable AC and DC bus systems, cooling water systems, and local control systems would have to be constructed to service the relocated modules.

12)  While the PS power components are still in good condition and good working order, the control systems are obsolete and showing serious signs of wear and tear. These facts, combined with the likely need for NCSX to exercise controls over the individual PSS rather than the pairs which make up each PS, suggests that control modernization will be required in any case. One option to consider in this regard would be to move toward the generation of firing pulses direct at the central control computer to be transmitted digitally and at the last step, optically, to the distributed MGD systems.  This would significantly reduce the parts count in the local controls, and could reduce the cost and enhance the reliability of the final system.

13)  The levels of power under discussion exceed the capability of the present PPPL connection to the 138kV grid. Therefore one MG system will be required. At present, only one of the two MG systems is in use.  Therefore either the MG now operating would be time shared between NCSX and NSTX, or the second MG set would be restored to service just for NCSX. However, the latter approach would require that the 13.8kV feeder configuration be re-arranged and segregated into two groups such that the NSTX feeders connected to one MG output bus and the NCSX feeders to the other one.

14)  To implement a direct connection of 13.8kV MG power from D-site to C-site would require that either iso-phase bus be extended from D-site to C-site, or pyrobreakers be installed to limit the very high fault power available from the MG. This approach is not practical. A better approach is to take the individual13.8kV feeder output of the variable frequency switchgear, which is current limited and protected, and extend it to C-site. In this case it may be desirable to add an additional switchgear line-up at C-site, in series with the D-site switchgear, for the purposes of providing an isolating safety break locally.

15)  Power could be transmitted in DC form from the rectifier outputs at D-site to the loads at C-site. This would not impose an intolerable voltage drop or power loss (e.g. the distance may not be that much greater than already is the case on NSTX).

16)  The C-site MG sets consist of three rotating shaft line-ups, each consisting of a motor, flywheel, and four DC generators (total 12 generators). The generators were built by Allis-Chalmers and first operated in 1960. The generators are separately excited DC generators. The motors are wound rotor induction motors with speed control by liquid rheostat. The exciters are cascade DC generators. Main specifications are given in the following table.

 C-site MG Specifications

(per shaft or per generator as applicable)

 

Parameter

Value

Units

Motor Power

7000

Hp

Flywheel Weight

96

tons

Maximum Speed

355

rpm

Minimum Speed

284

rpm

Deliverable Energy

137MJ

MJoule

Generator Maximum Voltage

850

V

Generator Maximum Voltage Derivative

1.0

kV/sec

Generator Continuous Ampacity

5075

amp

Generator Peak Pulse Current

22.3

kA

Generator Peak Pulse Power

16.67

MW

Generator Peak Pulse ESW

2.0

sec

Generator Peak Pulse Duty Cycle

5.15

%

Generator Armature Time Constant

0.0165

sec

Generator Field Time Constant

1.85

sec

Flat Top Current Regulation

+/-1.0

%

 

Comparing one of the twelve C-site generators to one of the D-site PSS, the power level is 78%, and (assuming 1/3 of the shaft energy per generator, and that the D-site PSS runs for 6 seconds at full rating and 80% p.f.) 40% of the energy. Controllability is relatively poor. Ripple is relatively good.

 

Implementation Options

 

Nine implementation options are presented on the attached spreadsheet. The following is a discussion of same.

 

Option 1

 

The 12 PS units and their transformers not required by NSTX in the east-west TF wing of FCPC would be relocated to C-site. Relatively simple modifications would be made to the DC cabling in FCPC to accommodate the removal.  However, the 24 PSS would not be sufficient for NCSX. To cover the shortfall, the C-site MG sets could be brought into service.

 

Option 2

 

The 12 units mentioned in option 1, along with 4 additional PS/XFMR units from the FCPC north-south PF wing, not required for NSTX, would be relocated to C-site. The DC cabling in the FCPC PF wing would be modified to accommodate the change. The PS units to be moved would have to be completely disassembled since the frames are trapped in place. Then, either new frames would be built at C-site, or the existing frames would be cut apart and re-welded at C-site.  This scheme would make available up to 32PSS, all identical to the existing PS configuration.

 

Option 3

 

The 12 units mentioned in option 1 would be relocated as per option 1. In addition the power modules, MGDs, and XFMRs would be relocated to C-site and new frames, AC/DC bus bar, water distribution, and local control systems would be built to form new PS units. The old frames, etc., would be left at D-site, and no DC cable changes would be required at D-site on the NSTX systems.  This scheme would make available up to 32 PSS, 12 of the existing configuration, and 4 with the new assembly and accessories. The D-site system could be restored in necessary in the future by returning the XFMRs, power modules, and MGDs.

 

Option 4

 

Taking from all 16 of the PS units not required by NSTX, the power modules, MGDs, and XFMRs would be relocated to C-site and new frames, AC/DC bus bar, water distribution, and local control systems would be built to form new PS units. The old frames, etc., would be left at D-site, and no DC cable changes would be required at D-site on the NSTX systems.  This scheme would make available up to 32 PSS, all of the new configuration. The D-site system could be restored in necessary in the future by returning the XFMRs, power modules, and MGDs.

 

Option 5

 

This is the same as option 4 except that new XFMRs would be used, and the D-site XFMRs would be left behind. Because NCSX requires pulse lengths of ≤ 3 seconds, it would be possible to realize up to 64PSS using the existing power modules.

 

Option 6

 

DC cabling at FCPC would be completely re-configured such that the outputs of the PSS were available for a variety of connections in a patch-panel type of scheme. This might be accomplished by adding penetrations between the 1st and 2nd floors of FCPC, removing the offices located there, and installing suitable cable/bus bar/bus link/disconnect switch systems. In this case all 74PSS would be available to NSTX or NCSX, but not at the same time. The system would be designed to allow switchover on the time scale of one week. NCSX would be located at C-site, and the power would be shipped from D-to-C-site in DC form. Current measurement and safety disconnect/grounding switches would probably be located at D-site. And shared by NSTX and NCSX.

 

 

Option 7

 

Same as Option 6 except NCSX installed at D-site to eliminate high current DC transmission from D-to-C-site.

 

Option 8

 

DC cabling at FCPC would be significantly re-configured such that the outputs of the PSS for NSTX and those for NCSX were fully segregated. This might require the addition of penetrations between the 1st and 2nd floors of FCPC, and the removal of the offices located there, to facilitate the installation of suitable cable/bus bar/bus link/disconnect switch systems. In this case the 32PSS not required by NSTX would be available to NCSX. The power would be shipped from D-to-C-site in DC form. Current measurement and safety disconnect/grounding switches would probably be located at C-site.

 

Option 9

 

Same as Option 6 except NCSX installed at D-site to eliminate high current DC transmission from D-to-C-site.

 

Cc:

 

R Hatcher

P Heitzenroeder

S Ramakrishnan

W Reiersen

J Schmidt

A Von Halle

M Williams