16 May 2025
AI and smarter analysis makes trains more reliable
Electric railways are key to sustainable transport, but their power systems are often affected by electrical disturbances that can cause wear, overheating, and operational disruptions. New research shows how these issues can be identified and addressed using advanced analysis and AI, paving the way for more reliable train operations.
Railway power systems are characterized by moving loads, fluctuating high demand, and multiple electromagnetic environments and voltage levels. This makes them susceptible to electrical disturbances, which can impact both operational reliability and the lifespan of technical equipment. One particularly challenging form of disturbance is known as waveform distortion. These occur when electrical signals deviate from their ideal shape and can result in overheating, equipment wear, and, in some cases, service interruptions. Since these distortions vary across time and location, they are difficult to predict and manage.
Addressing this issue requires a deeper understanding of how distortions emerge, how they propagate, and how they affect different parts of the railway infrastructure. This is the focus of a doctoral thesis by Rafael De Souza Salles, PhD student in Electric Power Engineering at Luleå University of Technology.
"Waveform distortion is not static. It changes over time and is influenced by traffic intensity, the position of the vehicles, and the type of technology in use," says Rafael De Souza Salles.
Rafael De Souza Salles, doctoral student in Electric Power Engineering at Luleå University of Technology.
AI and modeling improve analysis of power distortions
In his research, he has developed a framework for analyzing waveform distortion in AC railway power systems. While the focus is on the Swedish 15 kV 16 ⅔ Hz system, the methods have broader relevance. Combining traditional power system analysis with advanced statistics and unsupervised deep learning allows distortions to be identified and understood more effectively.
"We have developed methods that make it possible to pinpoint where distortions occur, how they move through the system, and which parts are most affected," says Rafael De Souza Salles.
The research is based on detailed measurements from Swedish traction converter stations and includes models of how disturbances spread through the railway power network. The results contribute to a better understanding of electromagnetic compatibility and power quality, key factors in preventing technical issues and maintaining reliable train service.
"Rail transport plays a vital role in the green transition, and the underlying power systems must be just as dependable. This research can help us build a more robust and sustainable transport infrastructure," says Rafael De Souza Salles.
Contact
Rafael De Souza Salles
- Associate Senior Lecturer
- 0910-585365
- rafael.de.souza.salles@ltu.se
- Rafael De Souza Salles
Published:
Updated: