The first pioneering equilibrium reconstructions were carried out by Jim Luxon and the late Bruce Brown on an array processor for D-III using a conventional equilibrium code, which demonstrated that in a non-circular tokamak, external magnetic measurements outside the plasma can determine both the plasma stored energy and the current profile peakedness, in addition to the usual shape. A typical reconstruction took about 20-30 minutes on the old array processor.
These positive results led to the development of the filament current fitting code MFIT (Magnetic Fitting) and later EFIT during the last phase of D-III operation, with support and encouragement from Wayne Pfeiffer and Ron Stambaugh. Many useful numerical techniques were also provided by Holger St. John. MFIT uses filament currents to model the plasma current profile. It therefore is computationally very inexpensive but suffers from lack of accuracy. EFIT retains the computational efficiency of the filament code approach by interleaving the equilibrium and the fitting iterations to find the optimum solution, but improves on the accuracy by allowing a distributed plasma current source constrained by equilibrium.
With this approach, an equilibrium reconstruction now requires only the equivalent of 1-2 equilibrium calculations rather than many, as in the conventional approach. This, together with the advances in computer hardware over the last 10 years, substantially reduce the amount of time required to find the optimum solution. On a modern HP-735 UNIX workstation, a typical reconstruction now takes only a few seconds.
.