Determination of composition and motion of multi-component mixtures in process vessels using electrical impedance tomography (EIT) I: Principles and process engineering applications

F. J. Dickin, Richard A Williams, M. S. Beck

Research output: Contribution to journalArticlepeer-review

53 Citations (Scopus)

Abstract

A method of mapping composition profiles and phase boundaries within process reactors and pipelines using robust, low-cost and non-intrusive electrical sensors is described. The method involves utilising differences in electrical resistivity between phases in a disperse system to chart their momentary distribution in a cross-section of the vessel. Sensing electrodes are placed at known locations around the periphery of the vessel and the sensor signals are fed into a suitable algorithm to solve the boundary value problem. The algorithm solves Laplace's equation (⊇.ρ-1⊇V=0) inversely: i.e. a computed image of the resistivity profile of the two-dimensional cross-section is formed using electrical measurements obtained from the sensing electrodes. The technique is analogous to that used in medical tomography, but the sensing system employed in this method is solid-state, of much lower cost and can be used in a process plant environment. Composition maps can be obtained in real time. The principles, scope and limitations of electrical impedance tomography and some practical details of its implementation on process-scale vessels are discussed and illustrated by two cases studies (dynamic sensing of liquid-liquid mixing; quantitative reconstruction of concentration profiles for a process at steady state).
Original languageEnglish
Pages (from-to)1883-1897
Number of pages15
JournalChemical Engineering Science
Volume48
Issue number10
Publication statusPublished - 1993

Fingerprint

Dive into the research topics of 'Determination of composition and motion of multi-component mixtures in process vessels using electrical impedance tomography (EIT) I: Principles and process engineering applications'. Together they form a unique fingerprint.

Cite this