Petrol flow pattern identification via data mining techniques

Abstract

Nowadays, petrol is an important resource for whole world, researchers are working on several mathematical models for flow pattern identification. One previous study is to find characterization of reservoir modeling in petrol flow data. Spatial data-mining can be used in reservoir geological research and ranking reservoir modeling. To find petrol flow patterns there is a study which aims to investigate and analyze the hole cleaning performance of gasified drilling fluids in horizontal, directional and vertical wells experimentally. Also, to identify the drilling parameters those have the major influence on cuttings transport, to define the flow pattern types and boundaries as well as to observe the behavior of cuttings in detail by using digital image processing techniques, and to develop a mechanistic model based on the fundamental principles of physics and mathematics with the help of the experimental observations. In this study we worked on petrol flow data with following features: mud flow rate, mud superficial velocity, pipe rotation per minute, rate of penetration, pressure transmitter and drill pipe. These features have been used in different classification and clustering algorithms to classify in nine class; Dispersed, Moving Bed, Stationary Bed, Dispersed Annular, Bubble, Elongated Bubble, Slug, Wavy Stratified, and Wavy Annular.We have received very promising results from 93% to 100% accuracy using different data mining algorithms.