Published online by Cambridge University Press: 03 August 2017
We present here a new method for analysing stellar variability: a geometrical study of the attractor associated with a variable star. In contrary to Fourier analysis, which focuses on times scales, this approach attemps to define the shape of the attractor in its phase space. It presents the advantage of being almost unsensitive to gaps in the data as long as uninterrupted sequences contain several tens of characteristic times. But, its noise sensitivity calls for very accurate photometry (10−4), still extremely difficult to perform up to now.
A preliminary test has been done on the variable white dwarf GD 66 on data obtained by classical photometry from the ground.