Image Reconstruction for Pressure Filtration Using EIDORS 3D

T A York1, Ł Mazurkiewicz3, N Polydorides1, R Mann1 and B Grieve2

1 UMIST, PO Box 88, Manchester, M60 1QD, UK, tay@umist.ac.uk

2 Syngenta Ltd, Technology & Projects Department, PO Box A38, Huddersfield, HD2 1FF, UK

3 Computer Engineering Department, Technical University of Lodz, Poland, lmazurk@writeme.pl

ABSTRACT

A certified, intrinsically safe, electrical resistance tomography system has been applied to two production pressure filters at Syngenta in Huddersfield, UK and has been collecting data, more or less continuously, for two years. A simple approach has yielded cross-sectional images but this paper reports first efforts to produce 3D images using tomographic algorithms. The EIDORS software toolsuite, running under MATLAB, has been employed and has yielded results on simulated data. Real data presents challenges due to effects of the zener barrier diodes that are integral to the intrinsically safe instrumentation. Dramatic changes in the pervading conductivity within the filter result in degraded images due to the limitations of using a single sensitivity matrix. To overcome this a new approach is introduced that proposes the use of dynamic sensitivity maps that evolve according to the conductivity changes that occur in the filter during a batch.

Keywords Electrical tomography, pressure filtration, intrinsic safety, 3D image reconstruction