Evaluation of methods for on-line image analysis of electrical impedance tomography images

This paper describes the manner in which EIT (Electrical Impedance Tomography) was used to online estimate separation efficiency in solid/liquid separators. Although off-line image analysis can be used to determine the characteristics of the internal flow, it is necessary to develop a technique for on-line control of the separation process. Three different methods were evaluated to determine a suitable strategy for on-line image analysis of electrical impedance tomography data: (i) Use of back-propagation artificial neural networks (ANN), (ii) air core scaling and (iii) use of radial basis functions and detection of abrupt changes. One case study is presented: A 44 mm internal-body diameter hydrocyclone pilot-scale plant.