Cascade and multibatch subspace system identification for multivariate vacuum-plasma response characterisation

A particular cascade structure system identification problem is formulated for the purpose of characterising the vacuum-plasma response for a magnetic confinement fusion experiment. A predictor-form closed-loop subspace system identification approach is advocated due to (i) plant instability (ii) sizes of input-output vectors and (iii) inherent multivariate eigenmodes of the physical system. Since experiment data come in relatively short batches, specialised means for data merging for subspace identification are developed. A batchwise delete-group jackknife procedure is utilised to estimate the standard error of the estimate of the dominant unstable empirical plasma response eigenvalue.