Sensitivity analysis of simulated railway energy consumption to elevation data processing and train length

As railroads strive to increase efficiency and reduce energy consumption, train performance simulation using detailed and reliable information on curvature and vertical gradient is critical for comparing technological alternatives on a given rail corridor. While major freight railroads have access to proprietary simulation tools, track charts and locomotive event recorder data, other researchers and firms interested in developing energy-saving technologies must rely on third-party train performance simulation tools and public data sources. Unfortunately, public digital elevation data does not differentiate between top-of-rail elevations and the surrounding ground, potentially leading to gradient discrepancies at tunnels, bridges and high cuts and fills. While longer freight trains may balance out localized grade discrepancies, they may be critical for analyzing emerging technology such as shorter trains of self-propelled autonomous railcars (SPARCs). To facilitate improved analysis of these technologies, this study aims to determine how geometric discrepancies impact the energy consumption of trains of varying length, and investigate the effectiveness of techniques to process and correct the data. Two primary elevation data sources are compared along several study corridors: data transcribed from Class 1 railroad track charts serving as “official reference”, and public digital elevation model (DEM) data processed by filtering and smoothing using open-source geospatial packages. At each processing stage, the elevation data is compared to the track chart data, and the Advanced Locomotive Technology and Rail Infrastructure Optimization System (ALTRIOS) is used to calculate the resistance and energy consumption of trains of differing lengths. Comparing total energy consumption and maximum resistance forces across different processing steps and freight train lengths demonstrates that longer trains naturally smooth out discrepancies, while short trains become sensitive to elevation spikes. The results underscore the importance of input data selection and offer practical guidance on data preparation for simulating the energy consumption of emerging railway technologies.