Plasma Shape Control at JET

Part of the Advances in Industrial Control book series (AIC)


The controller design described in this chapter has been carried out in the framework of a project aimed at assessing the possibility of controlling accurately highly elongated plasmas at JET with the existing active circuits and control hardware [78]. One of the steps needed to achieve this objective has been the redesign of the Joint European Torus (JET) shape controller, since the previous controller did not guarantee satisfactory performance. This section describes the features of the new JET controller, which is called the eXtreme Shape Controller (XSC). This new controller is the first example of a multivariable tokamak controller that enables control with high accuracy of the overall plasma boundary, specified in terms of a certain number of gaps. The problem is formulated as an output regulation problem for a non-rightinvertible plant, that is a plant in which the number of independent control variables is less than the number of independent outputs to regulate. In this case it is not possible to guarantee that the difference between the reference and the controlled plant output (tracking error) is zero at steady-state. To tackle this problem, we essentially make use of singular value decomposition in order to isolate the part of the plant output that can be better regulated at steady-state. Moreover singular value decomposition gives us an insight into the steady-state control effort: since some of the singular values of the plant static gain are small, we truncate these singular values introducing a trade-off between the tracking error and the control effort.


Tracking Error Controller Design Plasma Current Feedback Scheme Plasma Boundary 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer London 2008

Personalised recommendations