PanCam inspects locomotive pantographs as the locomotives pass an inspection point. The inspection point has two digital cameras – one captures a side view of the pantograph while the other captures the top view. The side view camera uses a white backboard to obtain a silhouette of the pantograph. This image is analysed to detect wear and damage to the carbon contact blocks that the pantograph uses to contact the overhead wire. The top view image is analysed to monitor the pantograph’s horns for damage. At the same inspection point, the locomotive ID is read by an AVI tag reader and associated with the image data by the PanCam software.
The PanCam software analyses the images and stores them to a file server. The analysis results are stored to a database. A web interface, developed as part of the PanCam system, promptly notifies QR’s Rollingstock Defect Coordinators (RDCs) of any pantographs that are worn or damaged. The RDCs can then view the images online and forward maintenance requests as appropriate to maintenance staff.
Challenges
Train speed. Typical coal trains employ 5 locomotives in sets of 2 and 3. The in-service locomotives may be travelling at up to 80 km per hour at the inspection point. PanCam’s software must quickly capture the pantograph images as the locomotives pass the inspection point, then analyse them during the time available between the sets of locomotives. PanCam’s software uses a multi-threaded design and rapid image analysis techniques to achieve these goals.
Illumination variations. PanCam operates in a semi-controlled environment. Illumination of the pantographs and the backboard is affected by the time of day and the weather. At night, artificial illumination is required. PanCam’s software automatically controls the camera’s exposure parameters depending upon the illumination conditions, and also adapts the image analysis to those conditions.
Long-term stability. The PanCam system must operate unattended 24 hours a day, 7 days a week. To achieve this goal, PanCam includes self-monitoring software and hardware that autonomously recovers from network downtime, hardware errors and system crashes.
Pantograph types. PanCam recognises different types of pantographs and analyses each pantograph type appropriately.
Research and development
The development of PanCam was incremental, using an agile development approach over a period of two years. Initially, basic image capture and analysis capabilities were developed, and then the capabilities were improved and new features were added as the project progressed. In developing the PanCam software, Hamey Vision Systems used an agile programming methodology – as each new capability was added to the system, a new software version was delivered for on-site use and testing by QR Limited. This methodology provided useful system capabilities early in the development project – even when the analysis techniques were not yet well developed, the genuine cases of wear and damage detected by the system could be manually forwarded to maintenance staff for their attention, yielding immediate benefits. As the development proceeded, the quality of analysis was improved and the reporting mechanisms became more automated, culminating in the present Web-based notification mechanism.
Implementation
PanCam software was developed using HALCON, a commercial machine vision library distributed by MVTec Software GmbH, and programmed in Visual C++ under Microsoft Windows. Results are stored to a Microsoft SQL Server database, and accessed from Microsoft Internet Explorer via an Apache Web server with PHP. The imaging components include 2-mega-pixel CCD digital cameras from ImperX, metal halide discharge lamps, backboard, infrared sensors and AVI tag readers. Communication is provided by existing network infrastructure.
Collaboration
PanCam was developed by QR Limited, Hamey Vision Systems Pty Ltd and Access Macquarie Ltd. QR Limited supply rail services in Australia and rail systems locally and overseas. QR conceived the PanCam concept and contracted Access Macquarie Ltd to provide development services which were subcontracted to Hamey Vision Systems Pty Ltd. QR provided all on-site development including sensors, tag readers, installation of imaging equipment and lighting, and onsite testing. Access Macquarie Ltd provided project management and contract services. Hamey Vision Systems Pty Ltd provided consultancy, imaging system design, and software development. The parties used the Internet and telephone communications to facilitate remote development.
Outcomes
Direct benefits
Reduced dewirements. QR estimates that dewirements cost at least AU$1,000,000 each in lost service time of track and locomotives, lost rail system throughput and repair costs. In 20 months of operation, PanCam has prevented all pantograph-related dewirements on the rail network it monitors, yielding benefits in excess of $20M. This is a significant benefit to QR and the Australian coal export industry.
Increased rail network availability. Reduced dewirements translate directly into increased availability of the rail network, resulting in increased rail traffic throughput.
Reduced inspection costs. PanCam provides an automatic inspection process that replaces manual inspection. This reduces labour costs and increases the availability of the locomotive.
Improved maintenance process. PanCam notifies maintenance personnel when pantographs may require maintenance. This ensures timely maintenance of worn and damaged pantographs and improves the quality of pantographs in the operating locomotive fleet.
Historical data collection. PanCam’s photographic database provides a unique pictorial history of each locomotive pantograph, showing the development of wear and damage patterns over time, and the effects of maintenance. The potential benefits of this data are still being explored.
Early identification of overhead wiring damage. The historical database of pantograph damage is useful for identifying problems in the overhead wiring such as loose fittings and hanging or damaged equipment. These problems produce consistent patterns of minor damage to the pantographs of locomotives that pass through the section of track with the wiring problem. By cross referencing the historical pantograph data with train paths, QR can accurately estimate the location of the wiring problem, saving the workforce time in costly “walkouts” of the track and reducing the possibility of dewirements.
Sustainability benefits.
Reduced environmental impact of dewirements. Dewirement events impact the natural environment through the necessity to bring in heavy equipment to repair track infrastructure and remove the non-powered locomotive. Derailment may further impact the environment through spillage of materials being carried by the train. The interruption of train service also wastes electrical power required to stop and restart the trains which, in the case of coal trains, are up to 2km long and weigh up to 12,000 tonnes.
Improved rail network operation. Use of PanCam ensures the timely maintenance of pantographs, improving the overall quality of in-service pantographs. The improved quality benefits operation of the electric locomotives. The contact surface of the overhead wire is normally smooth and coated for good contact and minimal wear. Pantographs in peak operating condition maintain this coating. In contrast, worn and damaged pantographs remove the coating, reducing the lifetime of the contact wire and causing other pantographs to experience inferior contact (skipping and arcing) due to roughness of the contact wire. PanCam helps ensure that the pantographs and the overhead wire are maintained in peak operating condition.
Improved maintenance process. Improvement of the maintenance process also improves environmental sustainability through reduction in component wastage and transport fuel.
back to Zenith Awards 2008.