Choices and Implications in Practical EIT

D. R. Stephenson1, T. A. York2, R. Mann1

1 School of Chemical Engineering and Analytical Science, University of Manchester, M60 1QD, UK.

2 School of Electrical and Electronic Engineering, University of Manchester, M60 1QD, UK. Email: t.york@manchester.ac.uk

ABSTRACT

Recent developments have presented an increased number of choices to EIT practitioners. These choices include instrumentation, measurement strategy, reconstruction algorithms and reconstruction parameters. For example, three ERT instruments are readily available; the University of Cape Town bi-polar pulse, the University of Manchester / Syngenta LCT and the Industrial Tomography Systems P2000. Increasingly these instruments afford the choice of measurement strategy and the option to export measured voltage data for reconstruction in alternative software environments. For example, the Eidors toolkit is an open source SourceForge.net project that enables users to perform both the forward and inverse 2 and 3-dimensional EIT problems under Octave or Matlab. The combination of flexible instrumentation and bespoke image reconstruction opens exciting possibilities whilst, in the author’s view, simultaneously presenting a bewildering array of choices for inquisitive EIT practitioners. This paper aims to outline the choices available to EIT users, and examines the implications of those choices in terms of accuracy and confidence in reconstructed images. Several common measurement strategies are employed to interrogate a real-life process application and results are analysed. Various reconstruction algorithms are applied to process data with the results of both historical and state-of-the-art reconstruction algorithms and parameters presented. Finally, the use of data-driven a priori information is discussed.

Keywords EIT, Measurement Strategy, Reconstruction Algorithm, Data-Driven a priori information.