Life length estimation of rolling stock

Publicerad: 14 augusti 2017

This research approaches the problem (MRL estimation of rolling stock) as a system because the vehicle and the track cannot be improved further in isolation of each other.

Sponsor: Trafikverket/JVTC

Researchers: Behzad Ghodrati

Duration: 2016-2018

The term rolling stock originally referred to any vehicles that move on a railway. It usually includes both powered and unpowered vehicles, for example locomotives, railroad cars, coaches, and wagons. The rolling stock health condition is important for both passenger and freight trains in terms of safety, availability, punctuality and efficiency. The estimation and knowing the Remaining Useful Life (RUL) of rolling stock equipment is important and helps/leads to design, plan and perform various inspection and maintenance methodologies on these systems to fulfill the planned performance measures.

In this project different methods for RUL estimation are studied and classified in groups (e.g. Knowledge-based models, Life expectancy models, Artificial neural networks (ANN), and Physical models). Thereafter the best and applicable methods are selected and used in estimation of RUL of rolling stock used in Sweden railway.

The output of this project will be useful and used in the maintenance planning of rolling stock

The best network architecture for RUL prediction of the wheel set subsystem
The best network architecture for RUL prediction of the wheel set subsystem