Comparing energy consumption for rail transit routes through Symmetric Vertical Sinusoid Alignments (SVSA), and applying artificial neural networks. A case study of Metro Valencia (Spain)

Abstract

This paper presents the training of an artificial neural network using consumption data measured in the metropolitan network of Valencia, Spain, to estimate the energy consumption of a metro system. After calibration and validation of the neural network, the results obtained show that it can be used to predict energy consumption with high accuracy. Once fully trained, the neural network is used for testing hypothetical operational scenarios aimed to reduce the energy consumption of a metro system. These operational scenarios include different vertical alignments that prove that Symmetric Vertical Sinusoid Alignments (SVSA) can reduce energy consumption by 18.41% in contrast to a flat (0% gradient) alignment.

Keywords

• Symmetric Vertical Sinusoid Alignments (SVSA)
• gradient
• energy consumption
• artificial neural networks
• metro system