Riskbedömning för nedgrävda gasledningar

Abstract

Submerged in the fields below us exists a network of pipelines for the distribution of necessary resources for our communities. It is of great importance that maintenance and repair of these wiring is performed satisfactorily. Swedegas is a company in Gothenburg that owns pipelines for natural gas. For several years, the company has collected data from the pipeline to document damage and properties of the damaged pipe sections. Based on this data, we have created models to predict where and why damage occurs. These models have been based on count data on the number of damage on a partial range of the entire pipeline. To model count data, we have used poisson regression and related methods such as negative binomial regression. Through the analyzes with the help of named methods, we have created a better picture over the mechanisms that affect submerged pipelines. In addition we have opened the door for future, in-depth analyzes of these factors.