Research to make tramway, metro and mainline operations more reliable and cost effective

Predictive Maintenance employing Non-intrusive Inspection & Data Analysis (PM ‘n’ IDEA1) is a jointly funded European research project focused on developing innovative inspection and maintenance technologies for railway track infrastructure. It is a three-year project that began in June 2009 with a budget of approximately €5 million. Sixteen partners are involved in the project and UNIFE2, the Association of the European Rail Industry, is the coordinator.

PM ‘n’ IDEA is focused on the development of non-intrusive track inspection systems to increase track availability, increase the life span and reduce life cycle costs of track com – ponents, and improve the safety of both workers and users of urban rail systems. The project has utilised and further developed existing innovative technologies to monitor the health status and the rate of degradation of track components to provide visibility of future maintenance and renewal requirements.

The project has delivered six ‘Key Innovations’ that are aimed at improving the integrity of rail transport networks through the deployment of intelligent design and sensor technologies into cost effective products and targeted non-intrusive monitoring processes.

Key Innovation 1: Use of intelligent image acquisition and analysis techniques for undertaking objective track inspection

The project team has risen to the challenge and designed a cost effective image acquisition system capable of very high image quality and resolution, which can easily be mounted on the bogie of a tramway, metro, or a mainline service or maintenance vehicle. This approach avoids the need for specialist and costly inspection units since ruggedised image acquisition units could be installed as part of a regular inspection service offering.

The system comprises a pair of high speed linescan cameras and a novel bank of very high intensity LED illumination systems to ensure consistently sharp images of each rail and the surrounding track infrastructure. The acquired images are stored in an on-board computer. The true novelty of the system lies in the software developed for the automated analysis for the detection of identified defect types or features. The system can be operated at vehicle speeds of 90mph with an image resolution of 0.1 mm/pixel.

Analysis is performed using mostly bespoke algorithms and has been shown to be capable of recognising a variety of defects and features including rolling contact fatigue cracks, squats, wheelburns, and the earliest signs of corrugation. Of special note is the development of the algorithms to recognise and characterise the running band. The monitoring of running band and deviations thereof are very early indications of the malfunction of the rail wheel contact condition and hence facilitate the preventative maintenance approach that is desired by all railways.

Key Innovation 2: Laser-sensor dimensional measuring system with on-board diagnostics

This is an evolutionary technique used to recon – struct the 3D shape from linear measurements.

This technology gives an additional dimension beyond the conventional 2D flat image used to inspect infrastructure components. The local structural information, particularly depth profile (Z level), can be measured and identified on an image. The depth information is superimposed by colour on to a 2D image where conventional image processing algorithms applying pattern matching detection can be used. Moreover, access to depth information is an alternative approach to the problem of low level of contrast in conventional images. The resolution and accuracy of the system varies according to the speed of the vehicle.

The system can be installed on a bogie or on the vehicle body, the nearer to the wheel the better to minimise the offset from the rail in curves.

The benefits of Laser Profile 3D analysis for end users lie in image processing to obtain further information about the track elements in addition to greyscale data from linescan cameras. The innovation allows inspection of the track bed simultaneously with visual inspection of the rail and fasteners.

Key Innovation 3: Assessment of internal integrity of embedded rails

The objective of this innovation is the nondestructive identification and localisation of electrically corroded rail foots and other defects in embedded tracks. The need for such an inspection system has been further highlighted by the increasingly popular practice of weld restoration of worn grooved (girder) rails in track. Existing technologies such as ultrasound inspection techniques and eddy current inspection techniques have been tested and proven not applicable for girder rails.

The technique developed in the project is based upon the injection of a high frequency large bandwidth excitation at the rail top using a small impact hammer. The vibration response is measured using a high-frequency accelero – meter or microphone and post-processed to identify corrosion of the rail foot. The detection relies on the analysis of the autopower spectrum of the measured vibration response and on the analysis of the cepstrum.

The finished system is a hand-driven mobile detection system that is able to localise corroded rails embedded in the street; measurement data is collected at several locations along the rail. Besides the measured response, the GPS coordinates as well as the distance from the starting point are recorded.

Key Innovation 4: Inspection technologies for the assessment of track quality

The objective of this innovation is to design a system for continuous monitoring of track quality in tramway and metro networks, through measurements taken by service vehicles during standard operations, as a new approach for reliable and cost-effective infrastructure monitoring and maintenance.

The basic idea of inspection technologies for the assessment of track quality is to extract valuable information from vibration measure – ments taken on-board the vehicle by a ‘track signature’ approach and/or a model-based technique for input reconstruction. Moreover, these approaches allow the integration of sensors in small boxes easily installed around the coach. The acquisition system can be easily installed outside and inside the commercial vehicle and coupled to an odometer.

For end users, the benefit of this approach is being able to obtain information about track quality using service vehicles rather than requiring dedicated equipment. Moreover, the two approaches to analyse the data supplement each other, as they have different advantages and drawbacks. The system could also be used for measuring track quality at more regular intervals than currently possible allowing greater understanding of track deterioration over a longer period.

Key Innovation 5: A scientifically validated methodology for establishing actionable boundary limits for the wear of rails

Running steel wheels on steel rails causes the two mating components to wear. The theoretical acceptable wear limit on a grooved or vignole rail is the minimum section dimension at which the rail is still capable of withstanding the stresses and strains imposed on it by the passing vehicle without breaking or deforming to endanger the safe running of vehicles. External constraints, such as the need to avoid fishplate strike or flange running, can sometimes become the rate determining factor ahead of the theoretical wear limit being reached. However, in practice the wear limits adopted by different networks are often based on past practice, which acts as a barrier to change because of the fear of introducing new risks. The wear limits adopted by different railway networks also reveal a wide spread even under very similar loading and operational conditions. This is indicative of either premature replacement or increased risk of rail breakage both of which are undesirable. The financial implications of the chosen wear limit have been assessed using measured wear rates on a partner network and the case study results are summarised in Figures 10 and 11.

Figure 10: Key Innovation 5 – average rail life as a function of acceptable vertical wear limit

Figure 11: Key Innovation 5 – financial savings as a function of wear limit

Thus, it is imperative that the wear limits used are not only appropriate to ensure safety but also to maximise life of the rail.

The work within the PM ‘n’ IDEA project was focused on developing a scientifically validated methodology of determining the acceptable levels of vertical, side wear of both vignole and grooved rails and also keeper wear for grooved rails. The methodology requires knowledge of the stresses and strains generated in a rail section as a function of the rail section, range of track forces experienced, and the track support characteristics.

Detailed FE models have been developed to establish the distribution of stresses through the rail cross section using track forces obtained from actual measurements on the Sheffield Supertram network. The FE model developed was validated using bespoke testing of machined rail sections under controlled laboratory conditions. The developed method – ology can now be applied to the full range of rail sections for the loading conditions on different networks.

Key Innovation 6: Automatic assessment of degradation and the integrity of intelligent track components

The contribution of the PM ‘n’ IDEA project towards reducing the cost of track inspection comes from the development and application of a range of sensors capable of identifying the reaction of key components to the forces imparted by passing vehicles. The below table shows what combinations have been studied.

The studies include use of Fibre Bragg Grating strain gauges, coupled with novel data processing to develop diagnostic routines for monitoring changes of condition of track components, particularly the condition of the nose of a switch blade, and to develop methods of advanced signal processing to classify data as a function of fault level. This developed system is capable of providing robust and sensitive condition monitoring with potential application in switches and crossings and for the measure – ment of temperature at locations susceptible to track buckling.

The work undertaken has considerably advanced the use of Axle Box Acceleration to enable remote monitoring of track components. The ability to detect changing conditions has been demonstrated e.g. clear evidence of fishplate deterioration was detected using axle box acceleration methods.

Compact self powered sensors with local intelligent monitoring have also been developed. These are small rail mounted low power units for condition monitoring of track components using miniaturised accelerometers, gyroscopes, and strain gauges, with analysis carried out locally before transmission. Work is concentrating on the application to stretcher bars and axle boxes. The final product is a flexible track component monitor, easily adapted for use on different vehicles and providing much greater levels of information than patrol based monitoring.

To summarise

The PM ‘n’ IDEA Key Innovations will make tramway, metro and mainline operations more reliable and cost effective. As a consequence, road traffic congestion and its impact on the environment will be reduced. For the public, a higher quality and more reliable service will be delivered.

References

1. www.pmnidea.eu

2. www.unife.org

About the author

Nicolas Furio has been Infrastructure Manager at UNIFE since November 2010. He is responsible for UNIFE activities in the field of rail infrastructure. Nicolas graduated from the French engineering school, INSA Lyon, as a civil works engineer and has a Master’s degree in industrial marketing and international strategy from the EM Lyon Business School. Prior to his role as Infrastructure Manager at UNIFE, he acted as Project Manager at Egis Rail which designs urban and railway transport projects.