Image Fusion of ECT/ERT for Oil-Gas-Water Three-Phase Flow

Image Fusion of ECT/ERT for Oil-Gas-Water Three-Phase Flow

Lifeng Zhang (North China Electric Power University, China)
Copyright: © 2013 |Pages: 6
DOI: 10.4018/978-1-4666-2645-4.ch011
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

The tomographic imaging of process parameters for oil-gas-water three-phase flow can be obtained through different sensing modalities, such as electrical resistance tomography (ERT) and electrical capacitance tomography (ECT), both of which are sensitive to specific properties of the objects to be imaged. However, it is hard to discriminate oil, gas and water phases merely from reconstructed images of ERT or ECT. In this paper, the feasibility of image fusion based on ERT and ECT reconstructed images was investigated for oil-gas-water three-phase flow. Two cases were discussed and pixel-based image fusion method was presented. Simulation results showed that the cross-sectional reconstruction images of oil-gas-water three-phase flow can be obtained using the presented methods.
Chapter Preview
Top

2. Ert And Ect Sensors

2.1. The Structure of Sensors

The 16-electrode ERT and ECT sensors are depicted in Figures 1(a) and (b).

Figure 1.

The structure of sensor: (a) ERT sensor and (b) ECT sensor

In ERT, the electrodes are mounted equally on the interior of pipe. The measured objects in pipe must be conductive and the electrodes must contact with them. While in ECT, the electrodes are mounted equally on the exterior of pipe. The measured objects in pipe must be nonconductive or most of the mixture of the multi-phase material is nonconductive.

2.2. Image Reconstruction Algorithms

There are many different image reconstruction methods for ERT and ECT, which can be mainly classified into two categories, direct methods (linear back projection, Tikhonov regularization and truncated singular value decomposition) and iterative methods (Conjugate Gradient and Landweber) (Yang & Peng, 2003; Marashdeh, Warsito, Fan, & Teixeira, 2006; Wang, Tang, & Cao, 2007). In our study, Landweber iterative algorithm with optimal step length was adopted which is defined as (Liu, Fu, & Yang, 1999):

(1)
(2)
(3) where , is the measurements, is the step length and is the normalized image grey in the kth iteration, respectively.

Complete Chapter List

Search this Book:
Reset