Publication: Pemodelan sistem pintar untuk menentukan nilai pecahan minyak
Loading...
Date
2005-03-01
Authors
Mohamood, Nadzrie
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The project title is modeling of an intelligent system for oil-fraction determination. The purpose of this project is to determine the fractions of oils in a pipeline containing two components, which are oil and gas. There are many types of flow regimes that could form in a pipeline such as stratified, core, annular, bubble, and homogenous flows. In this project all flow regimes need to be created by way of geometrical segments using program and will be fed into the Electrical Capacitance Tomography (ECT) simulator to get a set of ECT data which is represented in the form of independent capacitance measurements. This project uses an Artificial Neural Network (ANN) to solve the problem of oil-fraction determination. The entire work of this project realizes Matlab version 6.5.1 software. This
project does not go through the image reconstruction stage to determine the oil-fraction. All those data from the ECT simulator will be divided into 3 categories to be used in learning process of ANN. 40% of the data are used for training, 20% for validation, and the remaining 40% for testing data. Each data division is done randomly. The ANN architecture used is Radial Basis Function (RBF). The RBF will be trained with the training data chosen before regarding to the various probability of fraction of oil in a vessel. The output is a value between 0 and 1, corresponding to the fraction of oil in a pipe cross section.