An improved method of estimation for risk loss cost in natural gas network layout optimization

EasyChair Preprint 698

Abstract

We develop an improved layout optimization procedure, in which the BP neural network is applied and three different independent variables are analyzed. Herein, the procedures include two crucial steps. The first step is to forecast two risk loss costs by applying a neural network based on three different independent variables, and the second step is to verify the effective of the new estimation by using the predicted risk cost as the edge weight of the minimum spanning tree algorithm. The new method is applied in three different cases, leading to three distinct optimal layout. The results indicate that the economic benefit of using four independent variables is greater than the economic benefits of using three and two independent variables. Then, two optimal strategies for the pipeline network layouts are presented. These strategies realize a 1.54 to 13.23% greater economic benefit than that of the shortest layout.

Keyphrases: Natural gas network planning, layout optimization, minimum spanning tree, neural network, risk loss

@booklet{EasyChair:698,
  author    = {Jinyu An and Jianbo An},
  title     = {An improved method of estimation for risk loss cost in natural gas network layout optimization },
  howpublished = {EasyChair Preprint 698},
  year      = {EasyChair, 2018}}