Application of Alternating - Minimization (AM) in Dual Source Gamma Ray Computer Tomography for Imaging Three Phase Systems

R. Varma1, J. A.O’Sullivan2 and M. H. Al-Dahhan1

1 Chemical Reaction Engineering Laboratory, Department of Energy, Environment and Chemical Engineering, Campus Box 1198, 1 Brookings Drive, Washington University, St. Louis, MO 63130 USA, Email: rv2@cec.wustl.edu

2 Electronic Systems and Signals Research Laboratory, Department of Electrical and Systems Engineering, 1 Brookings Drive, Washington University, St. Louis, MO 63130-4899 USA

ABSTRACT

Dynamic three phase reaction and process systems are used widely in the chemical, petrochemical and biochemical industries. Dual Source Gamma ray Tomography (DSCT) has the potential for use as a non-invasive tool for determining the phases’ holdup distribution. Most of the dual source (or energy) gamma ray and X-ray tomography methods reported in the literature use low energy gamma ray photons. These have poor penetration depths, thereby limiting their application to very small flow domains (~ 3in.) However, since most flow domains are large, the use of higher energy gamma ray

photons is inevitable. In this study, the 661 keV photon from 137Cs and the 1332 keV photon from 60Co, have been explored for imaging phase holdup distribution in a gas-liquid-solid (GLS) system.

Simulated gamma ray transmission data based on a synthetic GLS phantom was processed for image

reconstruction using the Alternating minimization (AM) algorithm developed by O’Sullivan and Benac (2007).

Traditionally phase holdup distribution images are obtained by determining the attenuation images of the flow system by a suitable reconstruction process for both the energies first. The attenuation values of the individual phases for both the energies are then used to determine the phase holdup distribution images. This approach is referred to as the mono energetic approach. However, the high magnitude of error in the results makes this approach unusable. In a new approach, called the polyenergetic approach, data from both energies were simultaneously used to iteratively reconstruct the holdup images directly. The results show very little error. The polyenergetic approach shows tremendous potential for its application for image reconstruction related to dual source or energy tomography. Both the monoenergetic and polyenergetic approaches used in this study were based on the AM algorithm.