Design of a sample-based Prior using a phenomenological model for Annular Flow
S. P. Pellegrini, J. L. Baliño, F. C. Trigo
Departamento de Engenharia Mecânica, Escola Politécnica da Universidade de São Paulo, (Av. Prof. Mello Moraes 2231, 05508-030 São Paulo-SP, Brasil)
The characterization of multi-phase flows through electrical impedance tomography (EIT) requires addressing the inherent ill-posed nature of this inverse estimation problem, in order to regularize it. One of the manners to achieve this goal is by using a priori information that suits the system under analysis. For such a purpose, a phenomenological model is created by combining five different correlations from the literature in order to provide a characterization of the three-dimensional phase distribution in horizontal two-phase annular and wet gas flow regimes.
First, the model is presented. Then, careful sampling is performed for the input variables of the model and the model is used, along with its expected deviations, to determine a sample of flows, described by their phase distribution in the continuum. Afterwards, the phase distribution is converted to representative values of electrical properties and mapped to a discrete mesh. Finally, this sample of discrete distributions of electrical properties is summarized as a Gaussian density function. This density function shall be used, in a future work, as a prior to solve the inverse EIT problem.
The model-based sampling approach for the design of priors presented in this study is especially useful for multiphase flow, given the extensive literature of semi-empirical correlations for different flow patterns.
Keywords annular flow, electrical impedance tomography, sample-based prior, wet gas
Copyright © International Society for Industrial Process Tomography, 2016. All rights reserved.