OilGasArticles - Oil and Gas Industry Articles, News and Blogs - http://www.oilgasarticles.com
Decision Trees for new Installation in Transmission Pipelines
http://www.oilgasarticles.com/articles/497/1/Decision-Trees-for-new-Installation-in-Transmission-Pipelines/Page1.html
Muhammad Shaheen
Muhammad Shaheen Khan Tanoli has been enrolled in PhD (Computer Science) at University of Engineering & Technology Lahore, Pakistan. He has got M.Phil in Computer Science with a Gold medal in academics and research and M.Sc (Computer Science) – Gold Medalist. Mr. Shaheen has contributed at different research journals and conferences across the world. His research interests are Data Mining, Software metrics, Operations research and GIS 
By Muhammad Shaheen
Published on 11/27/2008
 
Decisions about installations of pipelines in gas industry determine the directions and course of billion rupees every year. The complexity of decision can range from simple to complex but analytical. These decisions determine cost of installation, operational optimization, timing of increasing facility capacity and long term utilization of pipeline. The complicated decisions within a company can take months of preparation. This paper addresses the mechanism to introduce decision trees (data mining) in pipeline industry to optimize the decisions.

Decision in Pipeline industry

 

D

ecisions about installations of pipelines in gas industry determine the directions and course of billion rupees every year. The complexity of decision can range from simple to complex but analytical. These decisions determine cost of installation, operational optimization, timing of increasing facility capacity and long term utilization of pipeline. The complicated decisions within a company can take months of preparation. Sui Northern Gas Pipelines Ltd has an asset of 2.92 million consumers connected through standard distribution piping network which is connected to gas wells through transmission pipeline network. The transmission system is equipped with pipelines, valve assemblies, Cathodic Protection Stations and Compressor Stations while the distribution network contains Town Border Stations (TBSs), District regulatory Stations (DRSs) and Consumer meter stations connected through pipelines. The company has a broader pipeline network including 6,625 Kilometers of transmission pipelines and 51,911 kilometers of distribution pipelines. The network has rapidly been growing for last twenty years and yet expected to grow a lot.

Extensions in the existing pipeline network or installation of a new pipeline causes pipeline planner to keep his eyes on wrist with a calculator on the table and the manpower diagram on the desktop. The pipelines specification codes should be on the finger tips to go through a progressive pilot project. Statistics and ASME pipeline code standards are the wisdom bibles for the decision makers to evaluate uncertainty, reduce risks and choose workable solutions. More involved cases require a clearly figured sketch of decision alternatives that combines information from multiple disciplines for which Decision tree is a selectable option. A number of methods are available to help decision makers evaluate decisions. The methods include net present value (NPV) calculations, discounted cash flow analysis, Monte Carlo simulations, Portfolio theory, preference theory etc which are good to compute basic expected values but these strategies are of no weight for graphical simulation of complex procedures and codes. Decision tree is a good option to frame complex situations.