Identification system of two-phase flow patterns for oil, gas and steam pipes

• Grant number: 07/01264-0
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: November 01, 2008
• End date: December 31, 2009
• Field of knowledge: Engineering - Mechanical Engineering - Transport Phenomena
• Principal Investigator: Grazieli Luiza Costa Carosio
• Grantee: Grazieli Luiza Costa Carosio
• Company: Mselli Engenharia e Consultoria Ltda
• Associated researchers: Oscar Mauricio Hernandez Rodriguez ; Paulo Seleghim Júnior
• Associated scholarship(s): 09/09690-4 - Identification system of two-phase flow patterns for oil, gas and steam pipes, BP.PIPE

Abstract

The fundamental objective of this technological innovation project is the development of a system to identify/detect liquid-liquid or liquid-gas two-phase flow patterns. Three immediate industrial applications for the proposed system can be appointed, among others: i) Oil slug detection in multiphase flows (from crude oil wells) at the three-phase separator input and other on-shore and off-shore equipment (a well known and not yet solved problem in the crude oil industry); ii) Heavy oil flow assisted by water monitoring and control (FAPESP project 04/13374-7), which is very important as it permits the exploration of previously impracticable wells and considerably decreases the energy consumption in heavy oil transportations; iii) Water carryover in steam pipes (these carryovers take place in the form of bubbles or small slugs of water), provoking corrosion, damages or even total destruction of steam turbines. Therefore, signal detection technologies will be implemented in an artificial intelligence platform. The first technology is based on the time-frequency analysis of the local pressure gradient, followed by the application of artificial intelligence techniques (artificial neural networks). The joined application of these technologies was found to be efficient in detecting these signals (SELLI, 2007) in laboratory environment, which makes the industrialization of this system very encouraging. The specialization of the system will be done during the neural network training and tuning phases of the operating parameters, allowing for the development of a totally innovative system with an improved performance, besides such a system is easily adaptable to complex practical situations, commonly found in oil and energy industries. It is important to stand out that this proposal is motivated and based on the results extremely encouraging obtained in one of our working team researcher's Ph.D. project. (AU)