A Neural Network Approach to Tokamak Equilibrium Control
We exploit the properties of the multilayer perceptron to develop a neural network approach to the feedback control of plasma position and shape in a tokamak experiment. The requirements of large bandwidth and high precision have led us to develop a custom hybrid analogue-digital hardware implementation of the neural network using conventional components. It is planned to demonstrate a complete system on the COMPASS tokamak at Culham Laboratory.
Unable to display preview. Download preview PDF.
- Bishop C M, Strachan I G D, O’Rourke J, Maddison G P and Thomas P R (1992) Reconstruction of Tokamak Density Profiles using Feedforward Networks Neural Computing and Applications 1 No 1.Google Scholar
- Osborne T H, Fukumoto H, Hosogane N et al (1986) Plasma Shape and Position Control on DIIID, Bull. Am. Phys. Soc. (1986) 31 No 9, 1502.Google Scholar