Modelling and Control of Plasma Position in the STOR-M Tokamak
Controlled nuclear fusion holds the promise of large supplies of commercial energy. Many significant engineering problems remain to be solved, however, before this energy can be economically provided. The tokamak is currently the most successful magnetic confinement scheme for controlled nuclear fusion, and control of the plasma motion is a fundamental requirement for its operation. The STOR-M Tokamak is equipped with turbulent heating (TH) capability. Turbulent heating is essentially ohmic heating (OH) with an artificially enhanced plasma resistivity. In addition to expected plasma heating effects, a significant density increase and consequent improvement in energy confinement time have been observed well after the turbulent heating pulse. Since a tokamak device is basically an electric system consisting of a transformer, plasma, eddy currents in a vacuum vessel, power supplies, and control windings, a new model in which these components are treated together as an electrical circuit has been developed and the effects of turbulent heating on the circuit elements included in the model. Since this model is nonlinear, and since the plasma motion needs to be controlled only in a region near equilibrium, the plasma dynamics has been approximated with linear time-invariant state equations. The simple model was used to study the behaviour of the system around an operating point of interest, and based on this study, a control system has been designed. Simulation results have shown that the plasma position is rather insensitive to variation in the time constant of the control current driver system, as long as it is less than 1.4 ms. It has also been shown that the presence of the iron-core results in unstable roots in the open loop plasma position Laplace transfer function. Although the plasma position model was very useful at the early stages of the control system design, due to practical constraints the parameters of this model cannot be determined very accurately. A Least Squares Error algorithm (LSE) has been used off-line to determine better estimates of the parameters of the model. A control strategy for plasma position using analog control during ohmic heating has been developed and successfully demonstrated. However rapid changes in plasma parameters during and after turbulent heating have been shown to require adaptive control. A digital controller based on a TMS320c25 microprocessor has been designed, built, and verified for control of the plasma position after a TH pulse. The final result of this work is a stably controlled plasma that can be maintained for the entire duration of a plasma discharge, for the both OH and TH modes.
Doctor of Philosophy (Ph.D.)
Electrical and Computer Engineering