Spoornet moves to automate train health monitoring

INTRO: Condition monitoring of a range of equipment on moving trains can lead to fewer delays, lower maintenance costs and smaller component stocks. The key is to marry detection equipment with automatic vehicle identification so that management knows precisely where and when to target resources

BYLINE: J J Marais Pr Eng

Senior Manager, Materials Engineering, Spoornet

TRAIN PERFORMANCE can be seriously affected by failing or malfunctioning components, with the consequences ranging from a minor delay to a serious and expensive derailment. Failed components or subsystems may also lead to damage of other equipment on the train, or to the track or infrastructure. The need to reduce costs in today's competitive transport environment is increasingly important, and the opportunity to cut costs by monitoring the health of rolling stock is now opening up. It will also be possible to base maintenance requirements on the condition of critical components, so eliminating the need for costly time-based maintenance plans.

Spoornet has developed monitoring systems to detect malfunctioning or failing components on a train. Some of these operate automatically, alerting the section dispatcher in the control centre to an imminent failure. The CTC then instructs the train driver on procedures to follow, determined by the criticality of the failure.

Hot box detectors warning of overheated plain or roller bearings are currently used in conjunction with axle counters. The driver has to count the train axles to find the failing bearing, which often leads to errors and bearing failures in service.

Manual detection

A number of manual train condition monitoring systems have been developed, but these require the presence of trackside operators who feed data into a computer at specified measuring points. Interpretation of the results is slow and immediate reaction to problems very difficult. To identify the vehicle with the failing components, wagon and locomotive numbers are recorded on video. A time-consuming process then follows to match the recorded data with the vehicles.

For these reasons, manual measuring points are used on an irregular basis. The next step is to integrate and expand them to create an integrated train health monitoring system.

Measuring points

The positioning of a trackside measuring point should be carefully considered. In a typical situation wheel temperatures are measured at two locations, for example on the uphill approach to a summit and at a point on the following descent. At Measuring Point One, all wheel temperatures should be low (slightly above ambient temperature) as brakes are not normally applied at this location. A wheel significantly hotter than the others will therefore indicate sticking brakes. Should the temperature of all the wheels on the same wagon be higher than other wheels in the train, the hand brake might not have been released on that wagon or its brakes may be malfunctioning.

At Measuring Point Two, on the descent, all wheels are expected to be hot in relation to ambient temperature as the brakes have been applied. Should a wheel be significantly colder than the others, it would mean that the brakes on that wagon are faulty.

The positioning of measuring points for other components also requires careful consideration. Measurement of the position of wheels on axles, to detect broken or shifted wheels, should be undertaken well ahead of points and crossings. Should a shifted wheel be detected, the driver must be notified at a point where the train can be stopped before the crossing and a possible derailment avoided.

The parameter to be measured at each point should be clearly defined. To facilitate computerised processing of data, a minimum range of factors should be specified. These are:

Â?the type of rolling stock on which the measurement is to be made, including the bearing type fitted;

Â?the parameter to be measured, such as temperature or force;

Â?the position of the measurement on the component, such as taking wheel temperature at the field side of the rim at a certain height above the rail;

Â?the maximum speed at which measurement is to be made;

Â?the highest number of readings to be recorded per train;

Â?the best trackside position for the measuring point;

Â?the number of measuring points per corridor or section of track;

Â?the expected upper and lower limits of the readings;

Â?the alarm limit - the level above or below which an alarm signal should be sent to the CTC;

Â?the difference limit, which is the maximum difference expected between readings on the same train;

Â?the action to be taken if one or more of these limits is exceeded;

Â?how the data is to be processed.

Spoornet currently operates automatic hot box detectors and dragging equipment detectors on all its major routes. Manual systems are also used as required to measure other parameters. These include bogie alignment, obtained through measuring wheel contact forces to the rail, wheel temperatures to detect brake defects, and track impact forces to detect wheel flats. They are also in use to measure dynamic wagon mass, wheel flange height and hollow tread wear.

Future developments

In the near future, Spoornet hopes to establish measuring points for other parameters. The initial list covers wheel tracking width to improve the detection of shifted wheels; detection of wagons loaded outside the limits of the moving structure gauge; detection of leaks in brake systems; measurement of the thickness of brake blocks and pantograph strips; assessment of the condition of roller bearings; and measurement of wheel diameters and locomotive suspension bearing temperatures. This list is certain to grow as an integrated train health monitoring system emerges.

Automatic vehicle identification

Train condition monitoring systems used without automatic vehicle identification do not always give satisfactory results. Axle counting to identify the vehicle with a defective component or system is very cumbersome, especially on long trains, and trackside conditions do not always favour inspection on foot.

AVI facilitates the identification of the vehicle with the defective component and the retrieval of vehicle condition data. This can be related to a subsystem or component and its history updated on the maintenance database. Vehicle identification data can also be used to track vehicles and update train movement and positioning databases.

Fig 1 shows the flow of condition data and vehicle identification in Spoornet's planned train health monitoring system, which consists essentially of AVI and measuring points. These are integrated by data links to the CTC and the Maintenance Organisation Network for Availability, Reliability, Costs & Stock (Monarcs), based on SAP R/3 software. The make-up of the train is recorded by AVI as it leaves the departure yard, and entered on the Sprint database used for commercial purposes, rolling stock tracking and positioning. This information is also made available to the CTC.

As the train moves past a measuring point, the condition parameter is measured on each vehicle. Depending on the nature of the condition data, it may be interpreted at the measuring point, and if any limit is exceeded the information is passed to the CTC.

In the case of a safety critical fault, the CTC will inform the train driver about the nature of the problem and issue operating instructions. The condition data will be forwarded to Monarcs and to the wagon maintenance or breakdown crew. Should the condition be non-critical, but require attention to the wagon at the arrival yard, Monarcs will also be informed; it will then notify the wagon maintenance depot at the train's destination.

Identifying trends

Spoornet plans to store the condition data in various databases. A medium-term database will be used to identify trends in condition data over a particular journey and failing components and subsystems. If the temperature of one bearing on the train shows an abnormal rise in temperature during the trip, it could indicate a failing bearing.

A long-term database will evaluate trends such as wear rates. Wheel diameter or hollow wear data are typical of the information which can be used to distinguish long-term trends.

For research purposes, Spoornet proposes to capture raw condition data. This will be used to analyse signal waveforms to improve signal processing and quality control at the measuring point.

At the moment measuring points are mainly located on Spoornet's heavy haul lines, and these will be the first to move to computerised condition monitoring. The first step will be the introduction of AVI, now being evaluated on the Richards Bay line (p225).

Another project is in hand to examine the best way to computerise measuring points, which are to be used to analyse raw condition data. The data will also be analysed in terms of the requirements for that measuring point and the CTC will be notified if alarm limits are exceeded. Analysed data will be forwarded to Monarcs if maintenance is required on a particular vehicle. Data storage will also be considered.

Seamless interaction of train health monitoring with Sprint, Monarcs, the CTC and other systems will require careful planning.

Use of the monitoring system for commercial purposes is also being considered. Possible applications include determining the payload on each wagon and tracking consignments. Further development could lead to its use to monitor infrastructure and the overhead power supply, and work is currently in hand to investigate this.

Train health monitoring will bring several benefits. It holds the promise of significantly reducing rolling stock maintenance costs, cutting train delays and, most importantly, reducing the number of derailments. All these should help improve financial performance.

Another benefit is likely to be a reduction in stocks of components. At the moment stock levels are based on unscheduled maintenance, the need for which should diminish as condition monitoring is implemented. For the first time it will be possible to plan purchase of components on a long-term basis. o

CAPTION: Fig 1. Layout of Spoornet's planned train health monitoring system, indicating the flow of data from trackside to control centre and maintenance depots

CAPTION: A bolometer used to measure wheel and bearing temperatures; this unit is set up to measure the temperature at the wheel rim

CAPTION: Spoornet has installed a number of temporary measuring points. The heap of scrap metal shows why they are needed

Spoornet prêt à automatiser le monitorage de l'état des trains

Le monitorage de l'état de toute une gamme d'équipements sur les trains qui roulent, peut résulter en des retards et déraillements moins nombreux, permettant ainsi des réductions significatives en terme de coût. Ceci devrait aussi déboucher sur des frais d'entretien réduits et de plus petits stocks de composants. La clé en est l'informatisation des liaisons entre les équipements comme les détecteurs de boîtes chaudes et l'identification automatique des véhicules pour que les véhicules à problèmes puissent être identifiés rapidement et de manière précise. On peut alors prendre toutes les dispositions nécessaires pour arrêter un train si la sécurité est en danger ou effectuer des travaux de maintenance lorsque le train arrive à destinationSpoornet geht auf automatisierte Zugwartungskontrolle über

Durch Zustandsüberwachung an einer Reihe von Einrichtungen fahrender Züge kâ€?nnen Verzâ€?gerungen und Entgleisungen eher vermieden und erhebliche Kostensenkungen erzielt werden. Gleichzeitig müßten sich hieraus geringere Wartungskosten und kleinere Komponentenbestände ergeben. Der Kern der Lâ€?sung besteht darin, Verbindungen zwischen Einrichtungen wie Detektoren heißer Achslager und automatischer Fahrzeugidentifizierung auf Rechner umzustellen, um Problemfahrzeuge schnell und genau zu erfassen. Bei einem Sicherheitsrisiko kâ€?nnen dann Maßnahmen zum Anhalten eines Zuges getroffen oder Wartungsarbeiten am Zug durchgeführt werden, wenn dieser seinen Bestimmungsort erreichtSpoornet automatiza el seguimiento de la fiabilidad férrea

Realizar un seguimiento sobre una gama de equipo en líneas férreas puede resultar en una reducción en retrasos y descarrilamientos, permitiendo una reducción considerable en los costes. Al mismo tiempo, debería dar unos costes de mantenimiento m s bajos y unos menores niveles de recambios en stock. La clave reside en informatizar los enlaces entre el equipo tales como detectores de la caja de ejes e identificación autom tica de vehículos para que los trenes problem ticos puedan ser identificados r pidamente y con precisión. Pueden entonces tomarse las acciones necesarias para detener al tren si est en peligro la seguridad o realizar tareas de mantenimiento al alcanzar el tren su destino