Optimised electrical resistance tomography performance for coarse-particle suspensions with a settled bed

Abstract

Electrical Resistance Tomography (ERT) offers a non-intrusive method to visualise the dynamic distribution of internal solids in suspension flows. While ERT can produce good qualitative images, quantitative analysis faces challenges due to low spatial resolution, sensitivity to electric noise, and the ill-posed nature of the inverse problem. This study aims to enhance ERT's performance for coarse-particle suspensions by considering system configuration, injection current, background conductivity and reconstruction techniques. Validation is performed on both static and dynamic suspensions with physically sensible flow scenarios. To ensure high-quality ERT data acquisition, the background conductivity must allow a range of injection current while maintaining a favourable signal-to-noise ratio. A more conductive carrier is needed for effective current injection in densely packed beds or larger-scale pipes. Reconstruction based on current and voltage measurements following Ohm's law is essential. The commonly adopted Linear Back Projection (LBP) algorithm in a commercial ERT system exhibits limited information in the bed-occupied area, leading to underestimations of integrated concentration. ERT predictions of chord-averaged concentration profiles in a dynamic suspension display similar trends as those from the one-dimensional Eskin model. When compared to experimental measurements, ERT results generally underestimate the integrated concentration by approximately 15 % v/v in a dynamic suspension.