DIII-D RESEARCH OPPORTUNITIES FORUM FOR THE 2013 EXPERIMENTAL CAMPAIGN Review | Direct submission with log-in | Request submission without log-in
Print this page

Print this page
Title	272: Dither injection for closed-loop system identification of vacuum and plasma response
Name:	Lorenzo Frassinetti none	Affiliation:	ZZZ
Research Area:	Torkil Jensen Award	Presentation time:	Requested
Co-Author(s):	K.E.J. Olofsson (Columbia), F.A. Volpe (Columbia) P.R. Brunsell (KTH), J.R. Drake (KTH)	ITPA Joint Experiment :	No
Description:		ITER IO Urgent Research Task :	No
Experimental Approach/Plan:	The envisaged application of dither injection to RWM control at DIII-D is as follows: - the I-coils will be used to control the RWM with one of the algorithms commonly used at DIII-D - a small perturbation (random in time and space) will be applied to the voltages that drive the I-coils current. - the system response (plasma plus any conductive structure present in the wall) will be measured by the sensor coils - Post processing analysis will allow the RWM growth. The technique will initially be applied to vacuum shots, both as a preliminary test as well as to obtain information on the vacuum response. Dithering will be applied to both the I- and C-coils. In particular using the external C-coils will allow estimating the wall diffusion time [Olofsson et al., 2012, Fus. Eng. Design, in press]. The technique will then be tested on few plasma shots using a feedback algorithm normally deployed at DIII-D [for example as in Okabayashi et al., 2009 Nucl. Fusion 49, 125003]. If successful, we propose to then apply the dithering to several plasma shots with different beta in order to study in closed-loop the RWM growth-rate dependence. The technique inherently takes into consideration the 3D effects of the conductive structure. This allows a non-disruptive experimental cross-check of VALEN predictions [Okabayashi et al., 2005 Nucl. Fusion 45, 1715].
Background:	Dither injection is a well-known technique in automatic control and it is largely applied in industry [e.g. Y. Zhu, Multivariable System Identification for Process Control (Elsevier Science 2001)] where quantitative models are required for the improvement of plant operations. It consists in the application of random perturbations to the signals in input to a physical system. The perturbations are usually generated as the output of a (designed) low-pass filter fed with white noise. By analyzing the response of the system (i.e. the relation between the input signals and output signals) it may be possible to identify the physical properties of the system. The technique is in general useful for the assessment of linear model applicability and prospecting the time-domain prediction horizon of linear models (potentially useful for routine plasma control). Recently, dither injection was successfully applied to the system-identification of RWM growth rates in the EXTRAP T2R reversed field pinch [Olofsson et al., 2012, Fus. Eng. Design, in press]. Preliminary tests were also carried out on RFX-mod. In EXTRAP T2R, dither injection allowed suppressing a large range of unstable RWMs avoiding disruptions and prolonging the discharge while simultaneously identifying the RWM growth rate. The estimated growth rates are in very good agreement with theoretical expectations, proving the validity of the method [Olofsson et al., 2011 Plasma Phys. Controll. Fusion 53, 084003].
Resource Requirements:
Diagnostic Requirements:
Analysis Requirements:
Other Requirements: