High performance discharges are routinely obtained on JET with low or rever
sed magnetic shear (s = (r/q)dq/dr), and the potential for steady state ope
ration of such discharges is under investigation. With the use of the prope
r heating and fuelling, these 'optimized shear' (OS) discharges exhibit an
internal transport barrier (ITB), resulting in a strong peaking of the pres
sure profile, and thus in high fusion performance. These regimes have been
extensively studied during the last (DD and DT) JET campaigns in order to p
romote this type of scenario as the basis for 'advanced tokamak' operation.
A review is given of the highest performance achieved on JET OS discharges
during the last experimental campaigns, in both DD (up to 5.6 x 10(16) neu
trons/s) and DT operation (fusion power up to 8.2 MW, n(io)T(io)tau(E) up t
o 10(21) m(-3) keV s). The role of the plasma edge is pointed out, as the p
ower required to trigger an ITB is often higher than the H mode power thres
hold, leading to double barrier regimes. The presence of an H mode pedestal
both modifies the ITB and induces edge bootstrap and ELM activity, which s
hould be controlled to prolong such discharges. The operational procedure o
f optimization is then discussed, addressing the problems of ITB formation
(power threshold, timing of the main heating phase, i.e. optimization of th
e target q profile, influence of the heating scheme, electron versus ion IT
Bs), ITB evolution (expansion of the ITB footpoint, H mode formation) and I
TB termination (disruptive and/or 'soft' MHD limits). Finally, the crucial
problem of the route to steady state for such OS discharges is addressed, b
oth in terms of ITB sustainment and control within the stability domain and
in terms of edge pedestal control by means of impurity injection. The impu
rity behaviour is found, and examples of high performance discharges sustai
ned for several energy confinement times are given (beta(N) = 1.95, H-89 =
2.3, P-fusion(eq) similar to 10 MW, Q(DT)(eq) similar to 0.4 sustained for
similar to 3 s). Extrapolation towards fully non-inductive current drive is
discussed.