Stanislav Jovanovic explains how software developed by ERRI in 1991-98 is increasingly being used in Europe, North America and Australia to plan maintenance and renewal over long periods, based on the observed and forecast condition of track elements

generic-tracks-cz_10.jpg

Stanislav Jovanovic is Project Engineer for Ecotrack at the European Rail Research Institute (ERRI) in Utrecht and a research engineer in the Railway Engineering Group at Delft University of Technology

THE COST of permanent way and its maintenance and renewal forms a significant proportion of total infrastructure spending, so any reduction has a major impact not only on the overall efficiency of infrastructure management but on total rail costs. However, the process of determining whether, when, where and how to intervene so as to minimise not just M&R costs but the total cost of operating the railway is a very complex problem.

In part, this is because sections of track tend to behave differently under the effects of loading. The decision processes for M&R works are also closely inter-related technically and economically, and in complex ways. Finally, the decision-making process requires a large quantity of technical and economic data, extensive knowledge and, above all, experience.

Ecotrack software enables managers and engineers responsible for making those decisions to plan the works so as to minimise life-cycle cost. It was developed by the European Rail Research Institute in Utrecht between 1991 and 1998 at the behest of UIC, and with the participation of 24 European railways. UIC was particularly concerned about the financial impact on the track of higher loads and speeds (RG 6.96 p365).

On completion of this task five years ago, it was soon appreciated that Ecotrack (ECOnomical TRACK) could only perform effectively once a large amount of data had been captured. This is normally contained in various databases and asset management systems, and it requires careful selection and migration into the Ecotrack database before it can be used for track condition analysis and work planning.

Before any railway can use Ecotrack, all of these issues must be faced. Recognising the complexity of the tasks necessary to put Ecotrack to work, ERRI/ UIC decided to help railways by offering feasibility studies (implementation trials)to facilitate identification and resolution of the most problematic issues related to implementation, such as data migration. This offer was taken up by SBB (Switzerland), Railtrack (UK, now Network Rail), FS (Italy, now RFI), RIB in the Netherlands, Serbian Railways and QR in Australia.

Today, after five years of practical experience, Ecotrack is clearly fulfilling the expectations of users and is provided long-awaited answers to questions that track maintenance managers face. In addition to providing solutions to the problems posed by maintaining high quality track at minimum cost, it is also resolving often delicate trade-offs between maintenance and renewal.

Locating assets accurately

Most of the world's railways were built between 170 and 90 years ago, since when there have been many alterations to track layouts and other features of the infrastructure. As a consequence, many track plans currently in use that define the location of assets with respect to kilometre posts are inaccurate, and hence unreliable.

Apart from track itself, the existence of several features such as switches and crossings, bridges, culverts and level crossings can also affect track behaviour - a concrete bridge deck normally provides much firmer support than adjacent embankments, for example. The locations of all these objects must be surveyed to ensure that their effects on track behaviour are properly interpreted.

So before existing location data is migrated into Ecotrack, it has to be corrected to the required accuracy. Given the vast extent of rail networks, this requires a very fast, accurate and economical survey. The only system available that met the criteria was FLI-MAP, produced by Fugro Impark bv, and it has been widely adopted. Among users to date are Spoornet in South Africa, Amtrak in the US Northeast Corridor, and in Europe Railtrack, DB, SNCF, Renfe and CFR Infrastructure.

Fast Laser Imaging & Mapping Airborne Platform locates points on the earth's surface with a scanning laser carried by a helicopter (Fig 1). The formation is located using GPS, and with two lasers measuring the distance from the formation to objects being scanned 22000 times a second, and the helicopter flying 50 to 150m above the ground at 50 to 70 km/h, the data density can vary from 5 to 25 points/m2.

By recognising the spatial relationship of these points, objects such as catenary wires and supports, rails, kilometre posts, signals and switches can be recognised and automatically identified. The accuracy of location is in the range 5 cm vertically and 10 cm horizontally relative to the national survey grid, while the relative local accuracy is even better (3 to 5 cm).

The purpose of deploying FLI-MAP is to define the position and attributes of all fixed assets, essential as a first step in building any asset register or maintenance management system. One result of the Ecotrack feasibility study conducted for Serbian Railways (ZTP Beograd) was that the Advanced Maintenance Management System will use FLI-MAP to reference all objects on the infrastructure by their geographic location.

After populating the track database with accurately-located assets, they can be linked to data obtained by inspection and measurement. The behaviour of track elements can then be modelled with the help of Ecotrack. In the case of Amtrak's Northeast Corridor, the FLI-MAP survey was already completed and the database populated as to location and condition, so the only task left would be to process all of that through Ecotrack in order to optimise M&R, and then balance track costs with quality.

Track condition monitoring

The second crucial prerequisite for successful track M&R management is condition monitoring. It is now widely recognised that the only efficient way to monitor the condition of track and other infrastructure assets is to use a train that can collect the data for all parameters simultaneously when running, where possible, at the normal line speed. Several such recording trains have been built, and more are being developed.

MerMec SpA in Italy produces equipment for track infrastructure condition monitoring. The measurement data can be successfully integrated with Ecotrack for in-depth condition analysis and M&R management. For example, RFI's Archimede diagnostic train recently brought into service in Italy (RG 9.02 p552) can monitor at 220 km/h all of the parameters listed in Table I.

A major advantage of Archimede is that all kinds of measurements concerning track quality such as geometry and ride quality can be correlated and analysed in an integrated way through Ecotrack, readily enabling truly optimised planning of M&R works.

These measurements can be analysed at two levels:

  • Level 1 analyses a single parameter such as gauge;
  • Level 2 analyses more than one parameter, such as track geometry with ride quality.

MerMec has already carried out four tests with Ecotrack at Level 1:

  1. Verified that condition parameters measured by MerMec are already handled by Ecotrack;
  2. Defined new tables/fields for new data;
  3. Visualised and analysed the results at both levels;
  4. Defined new thresholds and/or new rules.

Measurement data included in the current tests are components of track geometry, rail profile and rail corrugation. Tests still to be completed will cover wheel-rail interaction forces, vehicle accelerations and wheel-rail contact geometry. In both cases, the component parameters under each heading are as set out in Table I. Steps to follow include:

  • Performing import tests - track geometry standard deviation over 200m has already proved successful;
  • Analysing track geometry and ride quality in an integrated way;
  • Analysing other measuring systems such as catenary;
  • Defining new rules based on new data and combining them with existing rules.

Worldwide implementation

One of the most ambitious exercises carried out so far with Ecotrack involved complete processing and analysis in 2001-02 of the entire Railtrack network by FaberMaunsell.

Also in 2001-02, trial implementation by RIB (now ProRail) covered 87track-km in the Netherlands, 83 km in Queensland, and 93 km on Serbian Railways. These exercises threw up many interesting aspects and problems, but also revealed a great deal of similarity between these distant railways.

In practically all cases, track referencing and annotation was found to be inconsistent, and sometimes non-existent. Each railway had various information systems or databases in place, varying from comprehensive systems such as MIMS (Railtrack) and SAP (RIB and QR) to various custom-made databases and simple Access and Excel tables.

Migrating to and/or linking this data with the Ecotrack database proved rather straightforward. What did generate problems was the fact that the same piece of track would be identified inconsistently in systems produced for different purposes, having different end points for example. This caused rather serious problems in matching track geometry records, for example, with track inspection reports and track component inventories.

In these limited exercises, it was feasible to do the job manually, but automatic transfer of this data proved almost impossible without creating complex interfaces. This was not a problem caused by Ecotrack, however. It resulted from the absence of accurate mapping calling for global resolution.

Reconstructing work history

Most records of work carried out on the track were paper-based, so there can be no question of automated data capture. Many of these records are also of questionable accuracy. Just how important work history is to Ecotrack can be seen from Figs 2 and 3.

The left of Fig 2 shows how unrecorded work (mainly tamping) caused what appear to be illogical variations in track geometry. Recognising the importance of resolving this problem, a special tool called Work Tracer was developed for the application in RIB and QR trials. This uses the track geometry records to work backwards and identify the most likely time and place where tamping or other work was done.

Work Tracer software is based on the premise that track geometry does not improve by itself, so if there is an improvement between two measurements, tamping must have been performed. Based on this assumption, Fig 3 shows how Work Tracer has reconstructed a work history that can be used as the basis for planning future tamping or other work.

Fig 4 shows how the overall average standard deviation of the vertical track geometry of a 200 m segment changed over five years. There was a marked decrease in the SD in 2001 and 2002. Since there were no major works such as ballast cleaning or renewal on that track during the period, the only explanation could be that - after seeing the decrease in quality prior to 1999-2000 - the maintenance contractor decided to put more effort into tamping during 2000 and 2001.

This was confirmed when the statistics of the real (as measured) geometry behaviour (Fig 4)were compared with the statistics of the reconstructed tampings (Fig 5) showing a particularly high level of activity in 2001. This proved that Work Tracer was effective. The tool could be particularly useful in a situation where a railway has no access to tamping history, nor any ability to check the work that its maintenance contractors are actually doing.

Global planning and costing

All of these exercises were based on a collaborative application of the Ecotrack rules (slightly modified where necessary) and newly-created specific rules for the particular railway. For example, 25 new rules were created for RIB/ProRail in accordance with the local regulations and the valid maintenance policy of the Dutch infrastructure manager. This allowed the first shot at analysis and work planning over a long period (Fig 6).

It must be emphasised that fine-tuning of the rules and thresholds to match the needs of each infrastructure manager has to be done with great care, and with the direct involvement of senior personnel from the railway in question.

Fig 6 sets out the volume of work required on the 21·3track-km section 117A over the next 50 years. The work is expressed as track-km for each operation, and it can be displayed graphically or in tabular format to show costs (Fig 7).

As can be seen, there are peaks in ballast + sleeper renewal around 2019 to 2023, as well as ballast only, ballast + rail and rail only in the period 2025 to 2033, sleeper replacement from 2011 to 2017 and 2037 to 2041, and one more isolated peak of ballast renewal in 2044.

These peaks can be levelled up and made more uniform by altering the 'coherence settings' that apply various strategies for grouping work. It is also possible to delay or advance the work to match the costs to the available budget, or balance costs against variations in quality.

Ecotrack proves its worth

After five years of experience, Ecotrack has demonstrated its potential as a comprehensive and effective M&R decision-support system, aimed at solving difficult problems of maintaining track at the required quality for minimum cost. It also resolves those difficult trade-offs between maintenance and renewal as well as between costs and quality. The extensive database and powerful decision rules ensure thorough yet rapid track condition analysis, enabling engineers to investigate different scenarios to optimise M&R plans.

Ecotrack makes it possible to replace prescriptive M&R with a condition-based preventive policy, where work is scheduled only if, when and where it is really needed, and thus improving both the quality and efficiency of rail infrastructure.

Table I. Typical parameters addressed by measuring systems

TABLE: Track geometry

Left/right longitudinal levels

Left/right alignments

Curvature

Gauge

Cant

Twist over 2 and 9m or other base

Rail profile

Vertical wear

Horizontal wear

Gauge corner wear at 45í

Rail defects

Corrugation in 2 to 300 cm wavebands

Flaws and cracks

Shelling

Abrasion from wheel spin and slide

Welding anomalies

Rubble imprints

Sleepers

Cracks

Absence of fastener

Integrity of fastener

Ride quality

Dynamic Y and Q forces

Y/Q ratio

Bogie lateral acceleration correlation

Comfort and RMS value index calculations, (statistical analysis according to UIC518)

Contact geometry

Wheel/rail contact angles and equivalent conicity at several values of sigma

Overhead line

Static & dynamic stagger

Static & dynamic height

Contact wire gradient

Ambient temperature

Wire thickness (up to eight wires)

Pantograph

Pan/wire contact forces

Contact strip vertical speed and acceleration

Contact strip height

Contact strip temperature

Hard spots

Space domain frequency analysis

  • CAPTION: Fig 1. Helicopter survey using GPS and FLI-MAP on an electrified line
  • CAPTION: Fig 2. Fluctuations in vertical degradation of track geometry appear illogical without the tamping history
  • CAPTION: Fig 3. Tamping history as recreated by Work Tracer from track geometry records
  • CAPTION: Fig 4. Annual standard deviations for the average vertical track geometry over five years
  • CAPTION: Fig 5. The recreated tamping history shows the peak of activity in 2000-01 corresponding to the decrease in the real (as measured) SD in the same period (that is, the quality increase)
  • CAPTION: Fig 6. A 50-year work plan for Dutch track section 117A created by Ecotrack for RIB (ProRail)
  • CAPTION: Fig 7. Cost evaluation for the 50-year work plan in Fig 6

ECOTRACK supports condition-based decisions that minimise track costs

Stanislav Jovanovic explains how software developed by ERRI in 1991-98 is increasingly being used in Europe, North America and Australia to plan cost effective maintenance and renewal operations on track and other assets over periods extending to 50 years. The process starts with accurate mapping of all infrastructure assets, and the transfer of existing records on their condition into a geographically accurate database. In some cases where maintenance records are deficient, ECOTRACK can even be used to recreate recent tamping history from track geometry data.

Ecotrack conforte les décisions conditionnelles qui minimisent les coûts pour la voie

Stanislav Jovanovic explique comment des logiciels développés par l'ERRI de 1991 à 98 sont de plus en plus utilisés en Europe, en Amérique du Nord et en Australie afin de planifier les opérations coûteuses de maintenance et de renouvellement de voie et autres biens sur une période s'étalant sur 50 ans. Le processus commence par un inventaire précis de tous les équipements et le transfert des enregistrements existants sur leur condition dans une base de données géographique précise. Dans quelques cas, lorsqu'il y a un déficit de données, Ecotrack peut même être employé pour reconstituer l'historique récent des opérations de bourrage à partir de données telles que la géométrie de la voie

Ecotrack unterstützt zustandsabhängige Entscheidungen zum Minimieren von Gleiskosten

Stanislav Jovanovic erläutert wie vom ERRI von 1991 bis 1998 entwickelte Software in Europa, Nordamerika und Australien eingesetzt wird, um kostengünstige Unterhalts- und Ersatzarbeiten bei Gleisen und anderen Einrichtungen über einen Zeitraum von bis zu 50 Jahren hinaus zu planen. Der Prozess beginnt mit einer genauen Erfassung aller Infrastruktur-Einrichtungen und der Übertragung bestehender Unterlagen über deren Zustand in eine geografisch genaue Datenbank. In gewissen Fällen, in welchen Wartungsberichte ungenügend sind, kann Ecotrack sogar dazu benutzt werden, um die letzten Richt- und Stopf-Protokolle aus den Gleisgeometrie-Daten zu erzeugen

Ecotrack soporta decisiones con base en condiciones que minimizar? n los costes de la vía

Stanislav Jovanovic explica cómo el software desarrollado por ERRI en 1991-98 se est? utilizando cada día m? s por toda Europa, Norteamérica y Australia para planear operaciones de mantenimiento y de renovación económicas de la vía y de otros activos para un período de 50 años. El proceso se inicia mediante un mapeo preciso de todos los activos en infraestructuras y la transferencia de historiales existentes en su condición a una base de datos geogr? ficamente precisa. En algunos casos en los que el mantenimiento de datos resulta deficiente, Ecotrack puede incluso utilizarse para recrear recientes intervenciones de bateo partiendo de los datos de geometría de la vía