Neural network is considered as a parameter estimation tool in plasma contr
ols for next generation tokamak such as ITER. The neural network has been r
eported to be so accurate and fast for plasma equilibrium identification th
at it may be applied to the control of complex tokamak plasmas. For this ap
plication, the reliability of the conventional neural network needs to be i
mproved. In this study, a new idea of double neural network is developed to
achieve this. The new idea has been applied to simple plasma position iden
tification of KSTAR tokamak for feasibility test. Characteristics of the co
ncept show higher reliability and fault tolerance even in severe faulty con
ditions, which may make neural network applicable to plasma control reliabl
y and widely in future tokamaks. (C) 2001 American Institute of Physics.