Princeton Plasma Physics Laboratory annual report, October 1, 1983-September 30, 1984 Page: 26 of 123

                                
simultaneously discharging capacitor banks in the
equilibrium-field system to augment the rectifier
power supplies. In full compression experiments, the
plasma is brought into contact with the Inconel
bellows cover plates, which form the small major-
radius limiter, and maintained for —1 sec at the final
position (R = 2.10 m, a => 0.48 m). In free-expansion
experiments, the plasma is rapidly moved part way
across the vessel and allowed to slowly expand until
it contacts the limiters.
For the standard compression scenario, the time
evolution of the Thomson scattering electron density
and temperature profiles during compression are
shown in Figs. 29 and 30. Calculated profiles based
2.522 sec
MAJOR RADIUS (m)
Fig. 29. Thomson scattering measurements of the electron
density protile evolution during a lull compression
discharge. The dotted curves indicate the predicted ideal,
adiabatic scaling of the density.
> .#64X1112
2.522 sec
MAJOR RADI US (m)
Fig. 30. Thomson scattering measurements of fhe electron
temperature profile evolution during a full compression
discharge. The dotted curves indicate the predicted ideal
adiabatic scaling of the temperature.
on adiabatic scaling of the orecompression profiles
are shown for comparison. The density profile follows
the expected scaling (ne a C2, a °= C-1/2) until the
plasma reaches the inner wall when a slight
broadening and loss of central density occurs,
although the total particle count is conserved within
the accuracy of measurement. However, the electron
temperature profiles exhibit significant deviations
from ideal scaling (Te <x C4'3). Measurements of the
central electron temperature using pulse-height
analysis of the X-ray spectrum (accumulated over
several shots) also show that the central electron
temperature does not scale adiabatically. Interpre-
tation of the electron temperature data is complicated
by the occurrence of a large sawtooth oscillation
towards the end of compression. This feature is
always present but the magnitude and timing of the
sudden drop in central electron temperature varies
from shot-lo-shot, depending (at least in part) on the
phase of the sawtooth prior to compression;
consequently, there is a large scatter (±200 eV) in
the Te(0) values at the end of compression. To reduce
sensitivity to sawtooth effects, the electron temper-
ature scaling has been examined in terms of the total
electron stored energy, which for a single Thomson
scattering profile is known to within 10%. At the end
of compression, the electron stored energy is —78%
of the value expected for ideal scaling, with only a
small scatter associated with the sawtooth and shot-
to-shot reproducibility.
The increase in the central ion temperature is more
nearly in accordance with the expected adiabatic
scaling (Fig. 31). A detailed analysis ol the strong
compression discharges was made using the
TFtANSP code with measured R(t), Te(R), ne(R), and
Zgtf data as input. Good agreement is obtained
between measured and calculated neutron fluxes by
assuming 1 xChang-Hinton neoclassical transport
for the ions. The defect in central ion temperature
Adiabatic Scaling
■ f i
8 0.5
2.6
TIME(sec)
Fig. 31. Time evolution of the charge-exchange and
neutron measurements of the central ion temperature for
a lull compression discharge.
26