Breakdownstudies

Fig 1: Probability of breakdown depending on two parameters Fig 2: Examples of Paschen’s Curves

Goals: Investigate probability of plasma breakdown (creation of plasma) using learning machines algorithms i.e. Support Vector Machine with different setups of tokamak parameters. Mainly the role of the working gas and its pressure, toroidal electric field orientation and breakdown electric field will be investigated. The physical background is rather known but application of learning machines can help to extract more interesting information.

Reports: Learning machines report

Example of use:

./predict_breakdown.py --train --plot
./predict_breakdown.py --test --Ub=400 --Ucd=500  --pressure=27.84  --Tcd=0.01
./predict_breakdown.py --outliers

Links to files :

predict_breakdown.py - the algorithm itself
data_object.py - class Data
data_breakdown.npz - data from shot 5000 to 10700 - see data_mining for more details