Alcator C-Mod Run 941123 Information

Back to runs

Miniproposals
Miniproposal:
Date Filed:
Title:
First Author:
Session Leader:Ian Hutchinson (shots 1-21)

Operators
Session leader(s):Ian Hutchinson
Physics operator(s):Earl Marmar,Steve Wolfe
Engineering operator(s):Vinny Bertolino,Joe Daigle

Engineering Operator Run Comment
restart cmod

Session Leader Plans

Physics Operators Plans

Session Leader Summaries
Entered: Jul 7 2004 03:49:11:320PM
Author: To Be Determined
Session Leader Summary for 941123:
SL:Hutchinson PO:Marmar/Wolfe EO:Bertolino/Daigle

MiniProposal 65:Implementing Orthogonal Shape Controllers.

Because of practical difficulties with the magnetics and PCS, we decided
to use the older standard shot for the controllers.

The standard call of the contihh controllers uses shot 930817004 for the reference.
This is nearly a marginally limited nearly double null shot. It gives rise to the
following controllers.

Voltage Controller*1.E3 before time-constant scaling.
1 constraint(s) 1 major constraint(s).
Major Parameters: PSI0 RCUR ZCUR
Current and Voltage limits:
15000. 15000. 10000. 10000. 22000. 5000. 50000. 50000. 50000.
400. 400. 350. 350. 3650. 1050. 900. 200. 200.
Time constant(s) in voltage optimization 100.000 ms.
PSI0 RCUR ZCUR CLEARIN RXL ZXL RXU ZXU Cnstr...
0.326 -0.198 0.393 -0.124 -0.011 0.072 0.104 -0.521 -1.245
0.326 -0.199 -0.393 -0.125 0.120 0.657 0.008 -0.015 -1.244
0.175 -0.046 0.284 1.449 0.308 -0.789 0.380 0.656 11.046
0.175 -0.046 -0.284 1.448 0.659 -1.440 0.056 0.118 11.046
1.871 2.850 0.000 -1.081 -0.195 0.549 -0.011 -0.056 -28.101
-0.357 0.500 0.000 -0.122 0.042 -0.021 0.057 0.061 87.877
1.017 -0.593 -0.002 0.057 -0.048 -0.011 -0.050 -0.009 0.859
0.275 -0.161 0.282 0.096 0.019 -0.046 -0.076 -0.017 0.269
0.275 -0.161 -0.283 0.096 -0.070 -0.018 0.003 0.002 0.269

Compared with the original miniproposal 65 table using 940623018:
\item{} PSI0 and ZCUR controllers are identical.
\item{} RCUR is larger and CLEARIN is smaller in magnitude by about 20%.
\item{} RXL, ZXL are rather similar.
\item{} RXU, ZXU are larger by a factor of 2 to 3, but this is mostly
going to be compensated by the different sensitivity of the predictor.
\item{} EF4 (the constraint) is identical.

One concludes that using the older reference shot will make little practical
difference.

Starting from shot 941122027 which was a nice diverted shot from
yesterday, very similar to the planned shot.

Tasks Accomplished:

The current rise was fixed apparently quite nicely by increasing the early (10ms)
RCUR proportional gain and bringing the derivative gain back to 10ms also.

All of the orthogonal controllers except EF4 were installed.
The current and position controllers were tuned somewhat although it looks as
if we still need more slow ZCUR gain. Plasma current gain appears to be somewhat
less also.
The shape controllers CLEARIN, RXL, ZXL are all operational but not necessarily
tuned up. The controllers on EF1,2L are backed down to zero once the other are on.
RXU and ZXU are not demonstrated to be in a satisfactory state by the end of the
run. They need to be completed.
No significant exercising of the controllers was done.


Shot Remarks
Load of 22027. Fill 1.0e-5. Some engineering problems
initially. Looked at the segment switch glitch. RCUR has two
programmed step effects: in the offset of 10000 in segment 1, 0 in
segment 2, and in the program p.
1 PSdud OH1 current too low. Bus rogowski signals lost. Switch not closed?
Fredian admits that a software change had broken the CAMAC. Fixed.
2 PSdud Got bus data. Confirmed that OH1 switches did not close.
Reset the switches etc.
3 Fiz 40kA bouncer.
Increased the RCUR gain on the 10ms point from 2 to 4 on proportional.
Raised fill to 1.4
4 Pla Looked good.
Bring RCUR D earlier to 10ms from 20 and doubled to 2.
Loaded wire1 RCUR controller. Increased pulse gas.
Turned on lots of stuff.
5 Pla Nice current rise with no bounce. Higher CII light. Low density.
Disruption at 1.1 sec.
Loaded wire 2 slow ZCUR fluxihh. Called predictor. Called controller.
6 Pla Disruption. We forgot to change the fast predictor to fluxihh. Hence there
is a discrepancy. EFC current ramps up till the rail then we lose vertical
position.
Installed fluxihh Ip controller. Set physics gain to 2s^-1 (remember
that there is an extra 1/0.8a gain). Installed fluxihh for fast zcur.
Tweaked the gas.
7 Pla Disuption early. .29s. Rather high early density. Vertical position OK.
leans out rather on the limiter.
No changes: repeat
8 Pla Long. This shot shows that although the controller is workig for ZCUR,
there is a gain (and possibly offset) difference between the old filament
predictor and the fluxihh. By inference, the fluxihh predictor has a
sensitivity about a factor of two lower.
Reloaded the filament predictor for slow and fast zcur.
9 Pla Long. Plasma vertical position back as before.
Loaded CLEARIN controller.
10 Pla Long. Got down to 200 kA. Clearin missing its demand. We are getting less
effective gain out of this controller. This is probably because OH2 is
removed from this controller while the EF2 is not too different.
Barely diverts and stays diverted during down ramp.
Stretched the CLEARIN gain to 12 s^-1 from about 4.
11 Pla Still disrupted on ramp down. Aout seems to indicate that the error is worse
but the analysis predictor is between 9 and 10. A funny MHD burst.
12 Pla Disrupts at .7s. Looks as if we are fighting some other control.
The disruption apears to be caused by a monstrous spike on the 2pi.
Zeroed the EF2U gain at .5s; the EF2L was already zeroed there.
Went back to PID=4 on the CLEARIN gain.
13 Pla Comes right down. Slightly better clearin.
Raise slow zcur (wire 2) PID by 50%.
Raise clearin gain back up to 8 (phys 23.6s^-1).
14 Pla Helped to prevent the chopper from having a big excursion on the down
motion. Clearin looks better but not quite as high a gain as the previous
controller. Leave it for now.
Load new RXL controller. Gain to 15 s^-1 (1.2 /V) (But this might not have
taken).
15 Pla Works fine. Lost about 40kA off the plasma current. Looks slightly better
controlled on RXL, nothing big.
RXL 36s^-1 (3).
Load new ZXL. Gain (36s^-1) (9).
Turn on Argon.
16 Pla ZXL error dropped. RXL error rose somewhat.
Loaded RXU, ZXU controllers. (Gain still zero).
Turn off EF1L control at 0.4s
17 Pla Disrupts at 0.55s EF1L starts to ramp down at 0.4s, presumably because
we drop that gain before the other comes in.
Move start of RXL,ZXL and CLEARIN control gains to 0.35s.
18 Pla Diverted somewhat earlier thanks to CLEARIN earlier.
Turn on RXU gain at 0.35s at 10s^-1. Turn of EF1 at 0.5s.
19 Pla Disrupts at 0.5s. Not obvious why. Antenna spot still present.
Reprogrammed early position to .66m to try to avoid the hot-spot on the
antenna by moving in. (In segment 1.)
20 Pla Disrupts at 0.65s. No bright spot on the antenna. RXU causes EF2U to ramp
down to -4000A and saturate. Seems too strong on EF2. Chopper is pushed
to its limit and then we get a VDE (UPWARDS).
RXU gain to 2s^-1. Turn on ZXU at 0.35s at 2^-1 paramter to +0.05m.
21 Fiz No obvious reason why. Perhaps the upward disruption stirred new dirt.

Scorecard etc.:

Shot,Duration(s),Ipmax(MA) Shot,Duration(s),Ipmax
1 0.000 0.011 2 0.000 0.013
3 0.000 0.047 4 1.362 0.820
5 1.124 0.825 6 0.618 0.827
7 0.361 0.559 8 1.066 0.822
9 1.357 0.822 10 1.251 0.820
11 1.055 0.830 12 0.684 0.825
13 1.340 0.820 14 1.362 0.822
15 1.329 0.802 16 1.277 0.812
17 0.562 0.882 18 1.329 0.817
19 0.514 0.884 20 0.653 0.884
21 0.000 0.026
Finished summary for 941123 Total of 21 shots.
21 shots. 17 plasmas. 2 fizzles. 2 duds.
Fizzle = gt 17 lt 50kA.

Physics Operator Summaries
Entered: Jul 7 2004 04:36:52:037PM
Author: To Be Determined
Run 941123

MP 065 Orthogonal Controllers

SL Hutch
PO Marmar
Asst Wolfe
EO Vinnie B., Joe Daigle

After a slow start with some OH1 supply problems (switch not closing) and a
camac problem, we made good progress on the mini-proposal. The start up was
quite reliable today. Following the fizzle on shot 3, we moved the RCUR
feedback earlier (8 msec) and increased its gain. We got 17 good current rises
in a row, with no changes in fill pressure or early puff (this was a wire-head
run). New controllers were installed on rcur, zcur, clearin, Ip, rxl, rxu,
zxl, and zxu. Considerable fiddling is still required; see the session
leader's summary for details.

The shot summary:

Shot 1 reload of 941123027 fill = 1.0 e-5 dud;
OH1 current deficit possible camac problems
2 retry; dud; OH1 still bad, switch did not close; camac OK
3 retry; OH1 seems OK; 50 kA fizzle
4 fill 1.4e-5; pulsed gas move 1 ms earlier; Good shot; .82 MA
RCUR prop gain increased from 2 to 4 at 8 ms
5 Put in new RCUR controller; tweak up pulsed gas; disrupt at 1.1 s
6 Put in new slow ZCUR controller; a little more pulsed gas; disrupt at
.62 s; vertical position problem (slow ZCUR and fast ZCUR using
different predictors
7 Match ZCUR predictors, put in new Ip controller; raise pulsed gas
again; disrupt at .275 s
8 retry; plasma was about 5 cm below midplane, disrupt at 1.06 s
9 put back old ZCUR predictors; still using new controllers; Pulsed gas
tweak; good shot
10 put in new clearin controller; tweak pulsed gas; change
vis_brem_fluct to AC coupled, 1 MHz, upped gain; good shot
11 upped gain on clearin prop gain from 1.5 to 4.; reduced gain on
vis_brem_fluct; plasma; disrupt at 1.05 s; interesting coherent
MHD at about .6 s, shows up on VB_fluct
12 upped clearin gain from 4 to 8; upped pulsed gas; disrupt at 0.68
13 lowered clearin gain from 8 to 4; turned off EF2 U&L feedback at .5 s
Good shot
14 clearin gain back up to 8; slow zcur gains up by factor 1.5; good
shot; interesting MHD in middle of shot
15 put in new rxl controller; good shot
16 put in new zxl controller; tweaked pulsed gas up; argon on; good shot
17 put in new rxu and zxu controllers, gains=0; turned off ef1L control
at 0.4 s; disrupts at .55 s, down
18 moved clearin, rxl, zxl feedback on time to .35 s, delayed ef1L
feedback turnoff to .55; good shot
19 turned on rxu feedback at .35 s; turned off ef1U at .55; disrupt at
0,51 s
20 rcur change after 50 msec to get off antenna; disrupt up at .65 s; ef2U
hit its current limit and we lost vertical control
21 turn on zxu controller; reduce gain on rxu to 2/s from 10/s; zxu also
on at .35 s, gain 2/s.; fizzle, very little current.

Session Leader Comments

Physics Operator Comments
Nov 23 1994 07:33:57:140AMSteve WolfeRun 941123
SL: Hutch
PO: Marmar
Asst: Wolfe
EO: Bertolino/Daigle

MP#065 - Orthogonal controllers

Loaded 9411022027.
New Hybrid configuration notes change of BP12_BC to BP12_DE, F14 off.
Recalculated all SEG01 and SEG02 predictors that depend on these.
At present the FLUXIHH predictors still expect F14 to be good. If it is
dead again today this will cause a small error, but the machine seemed
to run with it OK yesterday.

Plan:
Re-establish a shot and proceed to calculate new SEG02 controllers
with algorithm XCONTIHH, as called for in MP065.

If F14 turns out to be a problem, Ian will generate a new IAN tree which
down-weights the FLUX_14 and FLUX_15 locations.

Nov 23 1994 07:52:33:190AMSteve WolfeOn trying to do a LOAD_INTO_MODEL got the deadlock error (three times).
Saved the changes as PCS_941123101, then got out of PCS, restarted with
PCS,/nocopy and loaded successfully (I think).

DId a HYB_DIS_INIT; all matrices respond ok.

Engineering Operator Comments
ShotTimeTypeStatusComment
108:33:24:160AMPlasmaOkcamac errors on bus instrumentation and oh1
209:36:20:490AMPlasmaBadoh1 switch didn't fire
309:53:35:060AMPlasmaOkno faults
410:05:30:900AMPlasmaOkno faults
510:19:47:580AMPlasmaOkef1l comm fault
610:38:03:180AMPlasmaOkno faults
711:04:53:620AMPlasmaOkno faults
811:21:49:100AMPlasmaOkno faults
911:40:58:380AMPlasmaOkno faults
1011:56:49:730AMPlasmaOkno faults
1112:10:12:520PMPlasmaOkno faults
1212:29:03:660PMPlasmaOkNo FAults
1312:46:34:550PMPlasmaOkNo Faults.
1401:05:21:880PMPlasmaOkno faults
1501:25:34:100PMPlasmaOkNo faults
1601:44:03:810PMPlasmaOkNo faults
1702:00:02:470PMPlasmaOkNo Faults.
1802:20:47:210PMPlasmaOkno faults
1902:33:59:780PMPlasmaOkNo faults.
2002:45:44:710PMPlasmaBadOH2L comm fault
2103:03:17:980PMPlasmaOkNo faults