Train Based Differential Eddy Current Sensor System for Rail Fastener Detection
Development of a train based system for monitoring track defects and rail track components by anomaly detection in the modulated magnetic field, generated and measured by a differential eddy current sensor.
Researchers: Matti Rantatalo (PL), Praneeth Chandran (PhD candidate), Florian Thiery, Stephen M Famurewa, Johan Odelius, Uday Kumar, Jan Lundberg
Sponsor: Trafikverket/JVTC, INFRASWEDEN, IN2SMART
Goal : The main goal of this project is to develop an automated rail fastener inspection method that can be carried out using vehicle-mounted differential eddy current measuring equipment, operating in regular traffic.
Project status and results:
Rail transport has emanated as a significant mode of transportation, forming a major contributing factor in the economic growth and social development of modern society through mobilisation and transportation of people and commodities. Growth in overall transport demand has led to railways experiencing higher demand on operational capacity, service quality, and safety. A higher operational capacity can lead to an increase in traffic and load applied to the existing infrastructure. Increase in load and traffic leads to deterioration of the infrastructure quality and degradation to its components, resulting in a higher number of Maintenance and Renewal interventions. The downtime arising from these M&R of networks is responsible for nearly half of all the delays to passengers. Hence, the track and its components need to be periodically inspected to decrease interruptions of train operation, reduce cost and ensure safety. One of the crucial components in rail tracks is the rail fastening system, which acts as a mean to fix the rails onto the sleeper, upholding the track stability and track gauge. Failures of fasteners can increase wheel flange wear, reduce the safety of train operations, and may lead to derailment due to gage widening or wheel climb. In Sweden, the inspection of track fasteners is mainly carried out either manually by trained inspectors or by using measurement cars. Manual inspections are slow, cost-intensive, labor-intensive, pose safety issues for maintenance personal involved, and are prone to human errors. Inspections based on measurement car requires track possession and are cost intensive and thus cannot be utilised frequently without compromising the operational capacity. Further, the adverse weather condition, especially in the north of Sweden for the majority of the year, limit regular fastener inspection that depends on such traditional inspection methods. Over the past two decade machine vision has been gradually adopted by the railway industry as a track inspection technology, however these automated visual inspection techniques are relatively an expensive technique to carry out and becomes a challenge when, the rail and the fasteners is obscured due to dust coverage, surface erosion, rusting or covered under snow or other debris.
The purpose of this project is to facilitate the development of a train-based automated differential eddy current measurement system for inspecting railway fastening systems and detect and analyse anomalies from the fastener inspection measurement. Figure 1 depicts the field measurement carried out along the iron ore line in Sweden, using the sensor system mounted on an unloaded freight train. The sensor was succesful in capturing all the fastener signature from a distance of 110mm above the railhead, during an actual train measurement. Signal processing techniques and feature extraction methods were used to extract useful informations from the raw signal pertaining to the fastener signature. Unsupervised anomaly detection techniques based on machine learning algorithms were implemented to identify and segregate the anomalous data points from the healthy or normal fasteners. Figure 2 and Figure 3 depicts the output of the detection algorithm for a measurement carried out over a track section of approximately 2.5 km. The detection algorithm was able to detect all the anomalous points precisely and separate them from the healthy group of points. Further, the proposed clustering model was also able to detect missing clamps (both one and two) from fastening systems and weld joints and segregate them with distinct boundaries.
Read more about the reults in the PhD thesis by Praneeth Chandran.