| Miniproposals | ||||||||||
| ||||||||||
|
| Operators | |
| Session leader(s): | Joseph Snipes,Marco Ferrara |
| Physics operator(s): | Steve Wolfe |
| Engineering operator(s): | Bill Parkin,Andy Pfeiffer,Ed Fitzgerald |
| Engineering Operator Run Comment |
| AM: MP#529 Bt dependence of L-H threshold (SL=Snipes) PM: MP#517 Kalman |
| Session Leader Plans |
| Entered: Apr 29 2008 01:54:51:347PM |
| Author: Joseph Snipes |
| Session Leader Run Plan for Wednesday 1080430 first half run day
MP 529: Toroidal Field dependence of the L-H transition at low density with ICRF Overnight ECDC in D2. We will start by copying 1080311019 and raise the TF to 3.3 T, Ip to 0.6 MA, and nl04 to 0.4 x 10^20 m^-2. This is a SNL standard shaped plasma. Turn on the ICRF at 50 MHz from 0.7 to 1.5 s with a slow ramp from 0.5 MW to the highest available power from J-port. We will look for the H-mode threshold, then reduce the density in small steps of 0.05 x 10^20 m^-2 looking for an increase in the threshold power as we reduce the density down to a density of nl04 = 0.2 x 10^20 m^-2 or until H-mode is not achieved at full power. Then, we will go back and increase the density in small steps up to nl04 = 6 or 7 x 10^19 m^-2 while again ramping the ICRF power to look for L-H thresholds. |
| Entered: Apr 30 2008 08:22:39:647AM |
| Author: Marco Ferrara |
| MP#517 (Kalman filter)
1) load shot 1071211009 as target equilibrium (k=1.7, control gains to their nominal values P=4 D=6, one of the shots from the experiment we ran in December). 2) Click on OBSERVERS, then on dpcs_kalmanfull2_zcur. A widget will open. Remove this procedure. 3) Click on OBSERVERS, then on New. Choose dpcs_kalman_zcur from the scroll-down menu. 4) Click on OBSERVERS, then on dpcs_kalman_zcur. A widget will open. Fill out the fields like follows: - write {ZCUR_phys_noise:16e6,Chopper_Gain:200.} in Parameters. Check the box for IDL Value. - write [0.8,2.0] in Switch Times. - write 10 in Wires. Then click Apply and Dismiss. Double-check that it took the inputs by re-opening the widget. 5) Run the shot. In this case the Kalman will output on wire 10 so that we can monitor that everything is fine before operating the filter in the loop (1-2 shots). 6) Change the number in Wires from 10 to 3. This will switch the output of the filter on the fast vertical control wire, so now the Kalman is in the loop. 7) Run the shot and verify the results (1-2 shots). 8) If everything looks fine, push the elongation, until disruption (1-2 shots). 9) Now increase the derivative gain from 6 to 8 or 10. The rationale here is that without noise we might be able to increase the derivative gain and maybe control higher elongations (2-3 shots). 10) Go back to the initial derivative gain (6) and lower the proportioanl gain (2.5) to excite the low frequency vertical resonance. The rationale here is that we want to see how the filter behaves when the real system follows a dynamic that is not included in the model of the filter (2-3 shots). 11) End of the half run. Total 7-12 shots. |
| Physics Operators Plans |
| Entered: Apr 29 2008 04:42:34:770PM |
| Author: Steve Wolfe |
| Physop plan for Wednesday, April 30, 2008
AM: MP#529 Bt dependence of L-H threshold (SL=Snipes) PM: MP#517 Kalman filter for vertical stability (SL=Ferrara) PO: wolfe ----------------- Engineering Setup ----------------- Run begins at 09:00 and ends at 17:00 Power systems as on: 1080311019 Acoil: +Dtop -Dbot -Jtop +Jbot (standard) Hybrid Enabled Gas setup: Fill B-Top with 6 psi D2 Hybrid enabled (PG4) fill B-side lower with 1 psi Ar Hybrid DISABLED (PG1) leave B-side upper as is Hybrid DISABLED (PG2) fill B-main (C-side) with 40 psi D2 Hybrid enabled (PG3) fill NINJA with 10 psi D2 DISABLED Enable gatevalves and shutters: ECE, VUV, Z-bolo (DNB if not on fire or full of water) Torvac gatevalve toggle (yes/no): NO Boronization(yes/no): no Overnight ECDC (yes/no): yes ICRF(yes/no): yes (J-port @50MHz) LH(yes/no): no DNB(yes/no): yes (if they want) Cryopump (yes/no): no Vessel temperature: 35/35/35 ------------------------------ ECDC Parameters (if requested) ------------------------------ gas and pressure: D2 at 2e-4 Torr sweep: 44/45/103 cm scan: 20/120 s Vessel temp for ECDC 60/60/60 -------------------------------------- Physop plan: ------------ MP#529: (L-H threshold) 1) Start from 1080311019 (LSN low density TF rampdown shot), and change parameters per SL plan: TF to 3.3T (92kA) Ip to 600kA nl04 to 0.4e20 Ramp ICRF power to find threshold 2) Decrease nl04 shot to shot by 0.05e20 per shot, to 0.2e20 or until no H-modes are obtained at max power. 3) Increase density to 6 or 7e19 and document threshold at higher density. MP#517 (Kalman filter) This run will be ohmic, so ICRF can shut down. Follow SL_plan, which may look something like this: 1) load shot 1071211009 as target equilibrium (k=1.7, control gains to their nominal values P=4 D=6, one of the shots from the experiment we ran in December). 2) Click on OBSERVERS, then on dpcs_kalmanfull2_zcur. A widget will open. Remove this procedure. 3) Click on OBSERVERS, then on New. Choose dpcs_kalman_zcur from the scroll-down menu. 4) Click on OBSERVERS, then on dpcs_kalman_zcur. A widget will open. Fill out the fields like follows: - write {ZCUR_phys_noise:16e6,Chopper_Gain:200.} in Parameters. Check the box for IDL Value. - write [0.8,2.0] in Switch Times. - write 10 in Wires. Then click Apply and Dismiss. Double-check that it took the inputs by re-opening the widget. 5) Run the shot. In this case the Kalman will output on wire 10 so that we can monitor that everything is fine before operating the filter in the loop (1-2 shots). 6) Change the number in Wires from 10 to 3. This will switch the output of the filter on the fast vertical control wire, so now the Kalman is in the loop. 7) Run the shot and verify the results (1-2 shots). 8) If everything looks fine, push the elongation, until disruption (1-2 shots). 9) Now increase the derivative gain from 6 to 8 or 10. The rationale here is that without noise we might be able to increase the derivative gain and maybe control higher elongations (2-3 shots). 10) Go back to the initial derivative gain (6) and lower the proportioanl gain (2.5) to excite the low frequency vertical resonance. The rationale here is that we want to see how the filter behaves when the real system follows a dynamic that is not included in the model of the filter (2-3 shots). 11) End of the half run. Total 7-12 shots. |
| Session Leader Summaries |
| Entered: Apr 30 2008 03:00:33:843PM |
| Author: Joseph Snipes |
| The H-mode threshold portion of today's run went reasonably well. There were 19 shots with 18 plasmas and one fizzle. We investigated the low density limit to the H-mode threshold at 3.3 T and 0.6 MA with 50 MHz ICRF H-minority heating. The ICRF worked quite well reaching 3 MW without difficulty. In contrast to the previous Ohmic H-mode threshold run at 2.2 T on 1080311, the low density limit at 3.3 T with ICRF was exactly where it was at 5.4 T with ICRF around nebar = 8 x 10^19 m^-3. This was even verified today by repeating a few shots at 5.4 T with D&E port ICRF up to 2.5 MW with H-minority heating. While the Ohmic H-mode results indicate that the low density limit decreases roughly linearly with TF, the ICRF H-mode results show no change in the low density limit between 5.4 and 3.3 T. Perhaps, increased sputtering with high power ICRF at low density results in more impurity radiation that increases the low density limit. |
| Entered: Apr 30 2008 05:47:02:787PM |
| Author: Marco Ferrara |
|
The second half of the run 1080430 was dedicated to the test of a Kalman filter for the vertical position of the plasma. Shot 20 was a high elongation k=1.87 without the Kalman, it did not disrupt but showed significant activity on the vertical trace at about 55Hz, probably driven by the pick-ups on the PF supplies. On shot 21 we introduced the Kalman filter at 0.8s but the shot disrupted with the nominal params CHOPPER_GAIN=200 and ZCUR_PHYS_NOISE=1.6e7. We tried different setups with CHOPPER_GAIN=250, ZCUR_PHYS_NOISE=1.6e7 (shot 022), CHOPPER_GAIN=200, ZCUR_PHYS_NOISE=0.4e7 (shot 023), CHOPPER_GAIN=150, ZCUR_PHYS_NOISE=0.4e7 (shot 024), but we only had disruptions. We took the Kalman off, on the assumption that the plasma was too elongated and unstable even without the Kalman. Shot 025 in fact disrupted, but 026 did not. The lesson from this first set of shots seemed to be that the filter was introducing a small delay which was not tolerable with very critical plasmas, a fairer test would need a more conservative target. On shot 027 we reduced the elongation to <=1.85 and ran the shot without the Kalman. Then on shot 028 we turned the Kalman on with the nominal paramas and it ran fine, only slightly larger oscillations (+11% RMSvalue, but also different features). On shots 29 adn 30 we reduced the proportional gain to P=3 and P=2.5 respectively, to try to excite the low frequency branch of the vertical instability, which is not included in the model of the filter. In these two cases the Kalman worsened things with respect to the shot when the Kalman was off, namely 031. Finally, on shot 32 we reloaded the aggressive target 021 but changed the derivative gain to D=8 (from 6). The shot disrupted, but oscillations looked smaller. Therefore we increased the derivative gain even further on shots 033 (D=10) and 034 (D=12.5) and these plasma didn't disrupt, in spite of the large elongations k>1.85. Oscillation amplitude was down and the plasma seemed to be able to recover from a large disturbance at t=1.3s in shot 034. On shot 035 we changed the derivative gain back to D=10 and turned the Kalman off and large high frequency oscillations showed up on the vertical trace. Gotta understand this behavior. |
| Physics Operator Summaries |
| Entered: Apr 30 2008 05:50:39:447PM |
| Author: Steve Wolfe |
Physics Operator summary for Wednesday April 30, 2008
MP#529 Toroidal field dependence of L-H threshold... (Snipes)
MP#517 Experimental test of kalman filter for vertical postiion (Ferrara)
PO: wolfe
EO: Parkin, Pfeiffer, Fitzgerald
A good run. The machine ran well and both experiments were completed
successfully. Marco actually got mostly the results he was looking for;
Joe got good threshold values at low and normal field, but the low density
limit was essentially the same at the two fields, which is contrary to
his ohmic result so there is something more subtle going on.
The low density shots in the morning were, as usual, prone to early
runaways. I tried increasing pg4 up to 31msec, setting rcur offset
to 1800, and increasing the pg3 in seg1 somewhat, but didn't find a
robust cure. Even when there were runaways they dumped early in the
discharge, and so didn't affect the results at 0.7sec and later, when
the RF was ramping. There wasn't much problem getting down to low density,
although the rf caused increases of ~1 to 2e19 when it turned on; it
seemed to me that this effect was stronger from J than from D&E, but
I'm not positive of that. Radiated power was bad at the low density,
especially at 3.3T, and that will complicate the analysis. The bottom
line is that we were unable to get H-mode with programmed nl04 of 3e19
at either 3.3T (50MHz) or 5.4T (80MHz), while we got transitions at 3.5e19 at
each field (much better looking H-mode at 5.4). The low density L-mode
shots at 3.3T had very high neutron rates, up to 1.7e13/sec, with low
stored energy ~40kJ. We have seen this before, and the tentative explanation
is second harmonic heating of a D+ tail. The 2nd harmonic damping is
FLR, (k_perp rho)^2 ,effectively proporional to beta_ion, so favors the
lower field. The 5.4 T cases weren't in Joe's original run plan, but
were done to verify that the unexpected results at 3.3T weren't due to
machine conditions.
The kalman filter experiments went pretty smoothly, although the
initial target was perhaps a bit too challenging in stability margin
at kappa=1.87. Plasmas were pseudo-DIII-D shape at 4T, 600kA, elongations ramping to above 1.8, mostly by increasing right gap. Everything ran well technically. The kalman filter
successfully reduced the response to noise with little or no compromise
in effective stability. Operation with reduced P gain resulted in
enhanced low frequency oscillation with the kalman in the loop, which
was not unexpected since the low frequency branch is not part of the
kalman design. Increased derivative gain (to 8, 10, 12.5) was well-tolerated
with the kalman in the loop, and seemed to suppress the high frequency oscillation that was observed without the kalman and D=10 [see shots 32-35]
There were a couple of delays for cooling, which seemed to be spurious.
The A-right top TF turn was showing high temperature (-159C) for almost
20minutes, then would suddenly drop to <-165. Once we just bypassed
cryo and took the shot, and the next shot this turn came right down.
Data acquisition was relatively trouble-free today. The big board display
stopped operating for a few shots and Joe Bosco had to reboot the PC to
get it back. The web-based logbook interface got lost, along with a
lot of the C-Mod website, between shots 7 and 9. Josh got it restored; I
haven't seen an explanation of what happened to it.
Startup was quite robust, although not as good as Joe got yesterday.
Scorecard:
Tests/Nopower : 0
Duds : 0
Fizzles : 1
Plasmas : 34
========== ==
Shots 35
|
| Session Leader Comments | |||
| Apr 30 2008 09:27:37:383AM | 1080430001 | Joseph Snipes | Shot 1: Got H-mode but the TF was at 3.5 T rather than 3.3 T. Reduce the TF a bit for the next shot. ICRF also tripped at 1.4 s, but that is not a problem for this shot. |
| Apr 30 2008 09:29:04:683AM | 1080430002 | Joseph Snipes | Shot 2: Got dithering H-mode throughout the RF ramp with the density also increasing with the RF power. There was an impurity injection at 0.89 s that may have caused some problems. Let's repeat without the impurity injection. |
| Apr 30 2008 09:48:35:603AM | 1080430003 | Joseph Snipes | Shot 3: Got dithering H-mode throughout the RF ramp with the density also increasing with the RF power again. There was no impurity injection but very little change in the dithering. |
| Apr 30 2008 10:00:49:927AM | 1080430004 | Joseph Snipes | Shot 4: Try reducing nl04 to 3.5 x 10^19 m^-2. Disrupted at 1.47 s. No H-mode even up to 3 MW of ICRF power. |
| Apr 30 2008 10:18:54:970AM | 1080430005 | Joseph Snipes | Shot 5: Try increasing nl04 to 4.5 x 10^19 m^-2. Got dithering again but no clear H-mode. |
| Apr 30 2008 10:29:24:807AM | 1080430006 | Joseph Snipes | Shot 6: Raise nl04 to 5 x 10^19 m^-2 and raise the RF power to start at 1 MW and ramp up to 3 MW at 1.2 s then stay at 3 MW until 1.5 s. Got H-mode at 0.77 s at nebar = 0.95 x 10^20 m^-3 and 3.3 T. |
| Apr 30 2008 10:46:39:820AM | 1080430007 | Joseph Snipes | Shot 7: Decrease nl04 to 4.5 x 10^19 m^-2. Got H-mode at 0.81 s at 1.5 MW of ICRF at nebar = 0.93 x 10^20 m^-3 . |
| Apr 30 2008 10:54:33:217AM | 1080430008 | Joseph Snipes | Shot 8: Decrease nl04 to 4.0 x 10^19 m^-2. Got H-mode at 0.77 s at 1.3 MW of ICRF at nebar = 0.83 x 10^20 m^-3. |
| Apr 30 2008 11:06:32:790AM | 1080430009 | Joseph Snipes | Shot 9: Decrease nl04 to 3.5 x 10^19 m^-2. Move the RF pulse to start at 0.85 s so that the TF is down to 3.3 T and keep the same ramp rate but turn off the RF pulse at 1.5 s as before. Weak dithering H-modes starting at 0.945 s with nebar = 0.7 x 10^20 m^-3 and PRF = 1.4 MW. |
| Apr 30 2008 11:18:27:207AM | 1080430010 | Joseph Snipes | Shot 10: Decrease nl04 to 3.0 x 10^19 m^-2. Dithering but no clear H-mode up to 2.5 MW. |
| Apr 30 2008 11:38:29:737AM | 1080430011 | Joseph Snipes | Shot 11: Repeat with the RF ramp starting at 1.5 MW and ramping up to 3 MW and staying constant at 3 MW. Disrupted at 1.29 s. Some dithering but no clear H-mode. |
| Apr 30 2008 12:05:55:130PM | 1080430013 | Joseph Snipes | Shot 13: Try ICRF at 3.0 MW with no ramp. Perhaps a bit of dithering but no clear H-mode. Disrupted at 1.52 s. |
| Apr 30 2008 12:24:18:740PM | 1080430014 | Joseph Snipes | Shot 14: Set nl04 to 6 x 10^19 m^-2 and return to an RF ramp from 1 MW to 2.5 MW from 0.85 s to 1.5 s. Clear H-modes starting at 0.965 s at PRF = 1.5 MW. |
| Apr 30 2008 12:33:57:967PM | 1080430015 | Joseph Snipes | Shot 15: Repeat with D&E instead of J at 5.4 T at the same q95 so that Ip = 0.98 MA. Got H-modes starting 1.02 s at nebar = 1.18 x 10^20 m^-3. |
| Apr 30 2008 12:44:12:177PM | 1080430016 | Joseph Snipes | Shot 16: Set nl04 to 4.5 x 10^19 m^-2. Got H-modes starting at 1.05 s with nebar = 0.91 x 10^20 m^-3 and PRF = 1.5 MW and Ptot =2.38 MW. |
| Apr 30 2008 12:56:32:450PM | 1080430017 | Joseph Snipes | Shot 17: Set nl04 to 3.5 x 10^19 m^-2. Got H-modes starting at 1.25 s with nebar = 0.78 x 10^20 m^-3 and PRF = 2.2 MW and Ptot = 2.85 MW. |
| Apr 30 2008 01:08:54:100PM | 1080430018 | Joseph Snipes | Shot 18: Set nl04 to 3.0 x 10^19 m^-2. Fizzle! |
| Apr 30 2008 01:19:58:563PM | 1080430019 | Joseph Snipes | Shot 19: Set nl04 to 3.0 x 10^19 m^-2. No H-mode up to 2.5 MW of RF. |
| Apr 30 2008 01:38:02:187PM | 1080430020 | Marco Ferrara |
kalman is on wire 10 is working. k=1.87 on this shot! next shot put kalman on wire 3 |
| Apr 30 2008 01:52:03:307PM | 1080430021 | Marco Ferrara | Disrupts at 1.4s.
Next shot try to increase CHOPPER_GAIN to 250., was 200. |
| Apr 30 2008 02:00:59:127PM | 1080430022 | Marco Ferrara | Disrupts at 1.18s.
Next shot try to decrease ZCUR_PHYS_NOISE = 0.4e7 CHOPPER_GAIN back to 200. |
| Apr 30 2008 02:25:29:473PM | 1080430024 | Marco Ferrara | Disrupts at 1.5s. Almost made it.
Next shot put the Kalman on wire 10. See if the plasma disrupts. |
| Apr 30 2008 02:42:49:517PM | 1080430025 | Marco Ferrara | Disrupts at 1.3s, without the Kalman, when k=1.86.
Next shot back off a little bit on the elongation. |
| Apr 30 2008 03:04:47:910PM | 1080430026 | Marco Ferrara | Makes it, even with a peak elongation k=1.87.
Next shot back off a little bit on the elongation. Kalman is still off the loop, on wire 10. |
| Apr 30 2008 03:15:47:980PM | 1080430027 | Marco Ferrara | Makes it, peak elongation k=1.83 at 1.36s.
Next shot put Kalman in the loop with nominal params CHOPPER_GAIN = 200. and ZCUR_PHYS_NOISE = 1.6e7 |
| Apr 30 2008 03:32:06:400PM | 1080430028 | Marco Ferrara | Stable and nice!!! peak elongation k=1.84 at 1.36s.
Next shot lower the proportional gain to 3 at 0.8s and see how the Kalman behaves when the low frequency branch of the instability is excited. |
| Apr 30 2008 03:45:45:973PM | 1080430029 | Marco Ferrara | Slow frequency seems to be enhanced, but shot is stable with the Kalman. Try P=2.5 with the Kalman next. |
| Apr 30 2008 04:03:40:640PM | 1080430030 | Marco Ferrara | Slow frequency branch is fairly big. Shot survives. Remove the Kalman in the next shot. |
| Apr 30 2008 04:16:58:857PM | 1080430031 | Marco Ferrara | Things are better without the Kalman. Gotta think about this. |
| Apr 30 2008 04:28:28:787PM | 1080430032 | Marco Ferrara | Load back shot 021 (very aggressive on the elongation).
Kalman in the loop with nominal params. D gain up to 8 from 6 at 0.8s. Oscillations seem now to be smaller, but shot disrupts at 1.4 (same as 021). Next shot try D=10. |
| Apr 30 2008 04:56:23:157PM | 1080430033 | Marco Ferrara | Shot survives with peak elongation 1.88. Oscillations are down with respect to shot 020.
Next shot increase D=12.5 from D=10. |
| Apr 30 2008 04:57:34:700PM | 1080430034 | Marco Ferrara | Shot survives with peak elongation 1.85. Oscillations are down with respect to shot 020, but a large perturbation at 1.3s. Shot survives and the perturbation is dumped.
Next shot decrease back to D=10 and switch the Kalman off. |
| Apr 30 2008 05:09:04:197PM | 1080430035 | Marco Ferrara | Shot survives with peak elongation 1.85. Oscillations are large and at high frequency, very different from case with the Kalman. |
| Physics Operator Comments | |||
| Apr 30 2008 07:22:57:633AM | Steve Wolfe |
Setting up for the morning experiment MP#529 (low density L-H threshold)
Load from 1080311019
Segments 1 and 2 active.
Segment 1 is titled "Startup 1080305"?
Looks normal, Rcur=1800,Br_0=0,EF4=-1340, pg4=26msec.
Notes from 1080305 indicate I was doing something with
the current rise for 400kA startup, "want to put back a normal startup
for part C". Bob was physop on 1080311; notes indicate he loaded 1080304025.
It looks like this was after I had turned off Ip gain in seg1 and changed
the OH1 voltage back (?) to 250V. However, I see lots of diff lines from
PCS_compare with yesterday's startup, including all the usual suspects
and pg3,pg2 and Ip (which is off anyway). Note the last shots yesterday
used He3 in seg1, so I don't want that.
I'm going to import seg 1 from 1080429018 (last shot without He3, and
adjust the pg4 for no cryopump.
Seg1: pg4 from 56 to 23msec
Seg2: TF from -56400 to -92000
Ip from -.55e6 to -.6e6
nl04 from 2.1e19 to 4.0e19
recalculate nl04 predictor
Load at 30-Apr-2008 07:21:48.00
Open tree /home/wolfe/pcs_scratch -1
Open tree done
| |
| Apr 30 2008 07:46:26:557AM | Steve Wolfe |
Continuing setup, import seg2 from 1071211009 into seg3 for afternoon MP.
Change title from "DIII-D Shape 6T" to "DIII-D shape 4T"
Note that not all of the predictors were updated (1st and 3rd row of
rcur are stale, no F11). Go through and call them all (wires 1-8,14)
Pick up dpcs_kalman_zcur under OBSERVERS | New
Bring up the setup widget and Turn it OFF
Now fill out the fields as in SL_PLAN:
Parameters: {ZCUR_phys_noise:16e6,Chopper_Gain:200.} click to IDL value
Switch_times: [0.8,2.0]
Wires: 10
Should now be ready to start Marco's run by turning on seg3 and turning on
the kalman procedure.
DON'T DO IT YET!
Load at 30-Apr-2008 07:46:05.00
Open tree /home/wolfe/pcs_scratch -1
Open tree done
| |
| Apr 30 2008 09:17:38:550AM | 1080430001 | Steve Wolfe | Shot#1: plasma full length Got 3.47T instead of 3.3, multiple H-modes starting at 0.95sec Next: reduce TF from 92kA to 88kA I have no signals from the pulse gas, but clearly we got gas. That is, I see signals but they are all zeros, which is clearly wrong. I have no signals from the pulse gas, but clearly we got gas. For example \ENGINEERING::TOP.TORVAC.GAS.PVALVE_2:WAVEFORM |
| Apr 30 2008 09:29:44:637AM | 1080430002 | Steve Wolfe | Shot#2: reduce TF from 92kA to 88kA raise pg4 from 23msec to 26msec Raise initial Ip point from -.55 to -.6MA at 0.1sec plasma full length, but had early runaways; may not have raised pg4 enough Lots of dithers and/or short repetitive h-modes, density rises along with ICRF ramp. Couple of injections, one at 0.86sec during rf Gas camac signals are ok this time, probably an INIT failure last time on the modules, but didn't see it on actmon. Next: repeat, raise pg4 again. |
| Apr 30 2008 09:38:52:067AM | 1080430003 | Steve Wolfe | Shot#3: repeat,pg4 from 26 to 28msec Plasma, still getting runaways. Still dithering. Next: drop nl04 to 3.5e19 |
| Apr 30 2008 10:02:24:723AM | 1080430004 | Steve Wolfe | Shot#4: nl04 from 4 to 3.5e19 pg4 from 28 to 30msec, rcur offset from 1000 to 1800 Disrupted at 1.5sec, lighting up inner wall opposite GH limiter. No h-modes go 4.5e19 |
| Apr 30 2008 10:17:01:690AM | 1080430005 | Steve Wolfe | Shot#5: nl04 3.5e19 to 4.5e19 plasma full length, dithering, but no clear transition. Next: repeat, start rf at higher power |
| Apr 30 2008 10:26:59:900AM | 1080430006 | Steve Wolfe | Shot#6: start rf at higher power, pg4 from 30 to 28msec nl04 from 4.5 to 5e19 (ignore previous claim we were going to repeat). Got some h-modes, also got early runaways again. First transition at 0.777sec, 1.27MW Next: nl04 back to 4.5 |
| Apr 30 2008 10:34:35:890AM | 1080430006 | Steve Wolfe | Shot#7: nl04 from 5 to 4.5e19; pg4 from 28 to 29msec Web based logbook has vanished, apparently along with all the C-Mod website. Josh is investigating. Meanwhile take shots. Big board also still not updated. |
| Apr 30 2008 10:39:41:457AM | 1080430007 | Steve Wolfe | Shot#7: nl04 from 5 to 4.5e19; pg4 from 28 to 29msec Web based logbook has vanished, apparently along with all the C-Mod website. Josh is investigating. Meanwhile take shots. Big board also still not updated. Plasma with h-modes (and some rf music) First transition at 0.817sec, 1.4MW Next: nl04 to 4e19 |
| Apr 30 2008 10:54:46:297AM | 1080430008 | Steve Wolfe | Shot#8: nl04 from 4.5e19 to 4e19 plasma with h-modes, first transition 0.78sec, Prf=1.3MW but density is up to 4.7, a little lower than 7 Next: down to 3.5e19 |
| Apr 30 2008 11:05:14:397AM | 1080430009 | Steve Wolfe | Shot#9: nl04 from 4e19 to 3.5e19, move rf pulse later to start at 0.85sec Web Logbook is back plasma with h-modes, first transition at .945? but very dithery. 1.5MW Next: down to 3e19 |
| Apr 30 2008 11:18:57:127AM | 1080430010 | Steve Wolfe | Shot#10: nl04 from 3.5 to 3e19 plasma dithers starting at .92sec but no clear transition Runaway dump early. Next: start rf at higher power, raise pg4 again. |
| Apr 30 2008 11:40:33:660AM | 1080430011 | Steve Wolfe | Shot#11: pg4 to 31msec, same density, rf waveform change. plasma, still early runaways, disrupts at 1.3sec Still dithering, no transition. goes vertically after an injection at 1.266sec, maybe a mode locking at 1.260 starts it all off. Next: raise pg3 in seg1 |
| Apr 30 2008 11:56:24:373AM | 1080430012 | Steve Wolfe | Shot#12:raise pg3 in seg1, rf to 2.5 MW flat same nl04 (3e19 program) plasma disrupts at 1.5sec, no h-mode, but neutron rate is up to 1.7e13/sec with Wmhd only 35kJ??? This seems to be the low density limit, and it's not as low as predicted. Next: same density, go for higher power |
| Apr 30 2008 12:04:07:973PM | 1080430013 | Steve Wolfe | Shot#13:raise pg3 in seg1a little more, rf to 3MW flat same nl04 (3e19 program) disrupts at 1.11sec, No h-mode, got 3MW, but prad is over 2MW. RF brings density from 2.5 to 4e19 Next: raise nl04 to 6e19, put ramp back on rf |
| Apr 30 2008 12:18:29:890PM | 1080430014 | Steve Wolfe | Shot#14: raise nl04 to 6e19, put ramp back on rf plasma with h-modes first transition at 0.965sec, 1.5MW Next: Go to 5.4T, 980kA, use D&E |
| Apr 30 2008 12:29:13:477PM | 1080430015 | Steve Wolfe | Shot#15:Go to 5.4T, 980kA (same q95), use D&E at 80MHz same density to start (6e19) plasma with H-modes. transition at 1.01sec, 1.4MW next: drop density |
| Apr 30 2008 12:41:46:957PM | 1080430016 | Steve Wolfe | Shot#16: drop density to 4.5e19 plasma with h-modes start at 1.05sec next: drop density |
| Apr 30 2008 12:55:34:920PM | 1080430017 | Steve Wolfe | Shot#17: drop density to 3.5e19 H-modes, then disrupted in rampdown H-mode at 1.27sec, nl04=4.4e19, 2.1MW rf Next: 3e19 program |
| Apr 30 2008 01:09:21:947PM | 1080430018 | Steve Wolfe | Shot#18: drop density to 3.e19 fizzle! Marco will give Joe one more try. Drop pg4 3msec |
| Apr 30 2008 01:18:47:483PM | 1080430019 | Steve Wolfe | Shot#19: pg4 from 31msec to 28msec density to 3.e19 plasma with runaways, no h-mode Next: On to the other MiniProposal |
| Apr 30 2008 01:34:46:370PM | 1080430020 | Steve Wolfe | Shot#20: Start on Marco's MP Seg2 off seg3 on Turn on the dpcs_kalman_zcur procedure Load at 30-Apr-2008 13:22:33.00 Open tree /home/wolfe/pcs_scratch -1 Open tree done plasma full length. Filter output shows up on wire10 starting at 0.8sec. Next: drop pg4 a bit, and do what Marco says about putting it in loop |
| Apr 30 2008 01:46:37:037PM | 1080430021 | Steve Wolfe | Shot#21: drop pg4 from 28 to 26msec Change filter setting from wire 10 to wire 3 plasma disrupts at 1.39sec, vertical instability. Next: Marco has many solutions |
| Apr 30 2008 01:59:56:080PM | 1080430022 | Steve Wolfe | Shot#22: Change chopper_gain parameter from 200 to 250. pg4 from 26 to 25msec disrupts at 1.186 (not a win) Next: noise param to 0.4e7 |
| Apr 30 2008 02:12:51:853PM | 1080430023 | Steve Wolfe | Shot#23: Change chopper_gain parameter from 250 to 200 zcur_phys_noise from 1.6e7 to 0.4e7 pg4 from 25 to 24msec disrupts at 1.5sec, nearly made it but has a growing oscillation. next: chopper_gain to 150 |
| Apr 30 2008 02:24:54:833PM | 1080430024 | Steve Wolfe | Shot#24: chopper_gain to 150, leave noise param alone pg4 from 24 to 23msec Load at 30-Apr-2008 14:14:54.00 Open tree /home/wolfe/pcs_scratch -1 Open tree done disrupts at 1.46sec Next run without kalman (back to wire10) |
| Apr 30 2008 02:40:32:313PM | 1080430025 | Steve Wolfe | Shot#25: run without kalman (back to wire10) No pg4 change this time Load at 30-Apr-2008 14:26:00.00 disrupts at 1.316sec, so this plasma is just too unstable. Next: back off on clearin |
| Apr 30 2008 03:02:29:017PM | 1080430026 | Steve Wolfe | Shot#26: clearin from -.005 to -.006 long delay for cooling, a-right top was hanging up at -158, then junped to -165 plasma full length, but still gets to kappa=1.87 back off another mm to be a bit safer |
| Apr 30 2008 03:14:03:370PM | 1080430027 | Steve Wolfe | Shot#27: clearin from -.006 to -.007 change endpoint of rcur ramp from .643 to .646 Still no kalman. plasma full length Next: kalman back in with original parameters |
| Apr 30 2008 03:29:53:773PM | 1080430028 | Steve Wolfe | Shot#28: kalman back in with original parameters noise_param=1.6e7, chopper_gain=200., wire=3 plasma full length. Clear decrease in noise-related demand. For this shot it works. Next: lower P gain (step 10 of plan) |
| Apr 30 2008 03:45:03:830PM | 1080430029 | Steve Wolfe | Shot#29: P gain on wire 3 from 4 to 3 pg4 from 23 to 22msec Load at 30-Apr-2008 15:32:24.00 plasma full length. Low frequency is bigger than on 28 Next: P gain from 3 to 2.5 (push low freq some more) |
| Apr 30 2008 04:02:43:737PM | 1080430030 | Steve Wolfe | Shot#30: P gain on wire 3 from 3.0 to 2.5 Load at 30-Apr-2008 15:45:36.00 Bypass cryo and go with A-right top at -159 after 19min plasma full length. Low frequency is stronger. Next: take Kalman out-of-loop |
| Apr 30 2008 04:16:45:793PM | 1080430031 | Steve Wolfe | Shot#31:take Kalman out-of-loop (back to wire 10) Load at 30-Apr-2008 16:03:15.00 plasma full length. Marco thinks the oscillation is worse with Kalman but it doesn't disrupt. Next: reload from 21 and increase D gain from 6 to 8 |
| Apr 30 2008 04:27:51:253PM | 1080430032 | Steve Wolfe | Shot#32: reload from 21, the more agressive shape that disrupted earlier, and increase D gain from 6 to 8 after 0.8sec (P gain is at 4 as usual) turn kalman wire back to 3 (back in the loop) Load at 30-Apr-2008 16:20:05.00 plasma disrupts at 1.42sec, similar to #21 Marco notes that oscillation amplitude is lower until it finally loses it. Next: D3 to 10 |
| Apr 30 2008 04:42:30:000PM | 1080430033 | Steve Wolfe | Shot#33: increase D3 gain from 8 to 10 after 0.8sec pg4 from 22 to 21msec since 32 disrupted Load at 30-Apr-2008 16:30:19.00 plasma full length, kappa gets to 1.87 Next: raise D3 gain from 10 to 12.5 |
| Apr 30 2008 04:53:53:430PM | 1080430034 | Steve Wolfe | Shot#34: increase D3 gain from 10 to 12.5 after 0.8sec Load at 30-Apr-2008 16:45:39.00 plasma full length, one big thrash near peak elongation 1.33sec Next: drop D gain back to 10 and turn off Kalman |
| Apr 30 2008 05:05:32:653PM | 1080430035 | Steve Wolfe | Shot#35: drop D3 gain from 12.5 to 10 and put Kalman back on wire 10 (out of loop) Load at 30-Apr-2008 16:56:17.00 Plasma full length. Didn't need kalman to survive at D=10. Chopper is railing on both sides. There is a high frequency oscillation now that wasn't there on 33 or 34. END OF RUN |
| Apr 30 2008 05:34:33:307PM | 1080430035 | Steve Wolfe | After the end of the run, remove the dpcs_kalman_full observer procedure and load into model, so this won't be around for tomorrow's run. |
| Engineering Operator Comments | ||||
| Shot | Time | Type | Status | Comment |
| 1 | 09:08:46:793AM | Plasma | Ok | |
| 2 | 09:22:36:723AM | Plasma | Ok | |
| 3 | 09:35:08:797AM | Plasma | Ok | |
| 4 | 09:54:32:267AM | Plasma | Ok | |
| 5 | 10:10:07:107AM | Plasma | Ok | |
| 6 | 10:23:02:117AM | Plasma | Ok | |
| 7 | 10:35:40:797AM | Plasma | Ok | |
| 8 | 10:48:57:180AM | Plasma | Ok | |
| 9 | 11:01:29:020AM | Plasma | Ok | |
| 10 | 11:14:57:843AM | Plasma | Ok | |
| 11 | 11:33:42:217AM | Plasma | Ok | |
| 12 | 11:46:02:053AM | Plasma | Ok | |
| 13 | 11:58:29:360AM | Plasma | Ok | |
| 14 | 12:10:54:323PM | Plasma | Ok | |
| 15 | 12:25:50:007PM | Plasma | Ok | |
| 16 | 12:38:39:780PM | Plasma | Ok | |
| 17 | 12:51:01:680PM | Plasma | Ok | |
| 18 | 01:03:44:097PM | Plasma | Ok | |
| 19 | 01:16:16:250PM | Plasma | Ok | |
| 20 | 01:30:39:493PM | Plasma | Ok | |
| 21 | 01:43:26:270PM | Plasma | Ok | |
| 22 | 01:56:46:640PM | Plasma | Ok | |
| 23 | 02:09:05:633PM | Plasma | Ok | |
| 24 | 02:22:25:737PM | Plasma | Ok | |
| 25 | 02:37:19:263PM | Plasma | Ok | |
| 26 | 02:58:30:250PM | Plasma | Ok | |
| 27 | 03:11:36:820PM | Plasma | Ok | |
| 28 | 03:24:11:097PM | Plasma | Ok | |
| 29 | 03:37:12:183PM | Plasma | Ok | |
| 30 | 03:59:10:153PM | Plasma | Ok | |
| 31 | 04:11:26:993PM | Plasma | Ok | |
| 32 | 04:23:46:063PM | Plasma | Ok | |
| 33 | 04:36:08:870PM | Plasma | Ok | |
| 34 | 04:48:34:100PM | Plasma | Ok | |
| 35 | 05:00:55:510PM | Plasma | Ok | |
| System Availability | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Apr 30 2008 09:08:38:120AM | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||