Automated Inspection of Rolling Stock Dramatically Improves Maintenance Workflows
The move to automated inspections is driving change in the longest, heaviest, and highest value rail operations in the world. Many of these operations use Trimble technology to improve how they operate, maintain, and utilise their assets.
Trimble condition monitoring solutions enable the detailed condition assessment of rail assets from wheel surface condition to full train inspection. Pinpointing issues minimizes incidents and interruptions to keep rolling stock on the move, reducing operational costs.
Machine Vision Technologies Support Automation
In the application of machine vision technology to automate inspections using machine vision wayside detectors, such as the Trimble® Beena Vision® systems, there are three key phases.
Image data acquisition
In rail applications, imaging has become an increasingly important source of data for the condition monitoring of components.
In general, image quality is a significant factor in the success of condition monitoring quality. For the successful deployment of vision-based systems, image quality and fidelity must be independent of environmental conditions such as lighting, temperature, precipitation, etc.
This means that these imaging systems must be designed and built for a railroad environment. Trimble systems are designed and built from the ground up for the harshest rail environments in the world.
To deliver the best possible imaging outcomes Trimble systems use infrared lasers and illumination in conjunction with infrared area and line scan cameras. This helps ensure that image quality is optimised and consistent.
It is important to understand that any vision system requires a line of sight to the component and the quality of this perspective is also important to the quality of the image or automated examination that can be provided.
Trimble Beena Vision assist with this by providing multiple (up to 13 individual cameras) in trackside imaging systems. This is partially for redundancy but also to allow images from different angles and at differing focal lengths to be recorded. Cameras can be positioned to record a degree of forward/backward and left/right angling as required.
Image and data processing
After properly capturing, labelling and storing images, Machine Vision Algorithms (MVAs) are deployed to process these images and calculate condition information for the identified components. The information extracted from each image can be used to measure objects such as a wheel profile, brake shoe thickness or the coupling angle of a brake hose.
Additionally, the information can be used to evaluate the condition of a component. The possible outputs of MVAs may include identifying components that are as missing, bent, broken, displaced, rotated, etc.
The final stage of the process is decision making for which the outputs of MVAs are used to create different levels of warnings and alarms. Trimble Beena Vision have developed hundreds of MVAs that process image and provide the data that is used to understand condition and predict the faults that impact train operations.
Trimble works with clients to understand specific requirements and build necessary MVAs. Often an existing algorithm can be adapted for most purposes however Trimble are also continually developing and adding to the extensive library of Trimble MVAs.
Complex detections such as oil leaks, axle scoring, and complex geometry assessments are in real world operation. Components as small as split pins, nuts and bolts are also being assessed by MVAs while the train passes the equipment in its normal operation.
Fault detection and alarm generation
MVA based alarms are either instantaneous, planning level, or trend based. Instantaneous alarms can be generated to indicate significant failure modes such as coupler securement failure, a broken centre sill or a condemnable wheel. In these cases, depending on the applicable rules, the train may have to be stopped immediately for urgent corrective action. In other less urgent cases such as a bearing cap bolt missing or a broken truck (bogie) spring, the train is usually moved to a more appropriate location for proper repairs.
Planning level information refers to the identification of less serious conditions that require maintenance to be planned and executed in certain timeframes. These types of events could refer to, for instance, minor structural damage, missing earth straps on wagons or other events which can be dealt with when “next in shop or yard”.
Trend-based alarms are usually created based on relatively slowly changing measurements of a component where the trend of the change can predict a fault and also predict the date at which the “fault” will occur.
An example of this is the back-to-back measurement of a wheelset where an increasing measurement requires alarm generation while the individual measurements may remain within the acceptable operating range of operators and regulators.
Trend-based alarms are usually generated from a database where all historical data is available and analysed periodically. Alarms and alerts from Trimble Beena Visions systems can be viewed and managed using the Trimble CMMS™ (Condition Monitoring Management System) and TrainWatch software solutions.
The CMMS rules engine is user driven and managed user customizable. No additional investment is required to build rules.
Importantly rules can be and are built across multiple data sources. These are called “Composite rules” and could for instance consider acoustic bearing, bearing temperature and missing end cap data in a single rule. Data added to CMMS from other systems such as for component and maintenance history can also be added and utilised in rules.
Rules can also consider how data is trending and a rate of change. This can be used to project when condemning limits will be reached for instance.
In some cases, for small or complex detections types, machine vision may not detect the condition on each pass however CMMS rules functionality includes the ability to still generate alarms and warnings the detection occurs three out of five passes, for instance.
The Trimble Beena Vision range of vision-based wayside measurement and inspection technologies enable the efficient monitoring of rolling stock delivering the data that is needed to understand rolling stock condition. The comprehensive solution suite includes:
- Imaging units to inspect almost all rolling stock components visible while a train is in motion. Specialised systems for wheel profile and tread, brake components, pantographs, trucks (bogies), and system groups that can provide 180०, 270० and 360० views of trains as they pass through the inspection site.
- Machine vision algorithms (MVAs) to process and provide information related to captured images. Trimble MVAs can return conditions as small as missing nuts, bolts, and pins, broken components, component structures and geometry, and accurate point to point measurements.
- Databases and user interfaces to access and view data with short and long-term information for trending and prediction.
- Advanced software tools to trend data such as wheel and brake measurements and project “change by dates” that maximize component life and improve safety.
- Data management, storage, and access solutions.
- Reviewing and reporting solutions.
- Ability to interact with operators’ internal maintenance management systems.
- Flexible solutions that expand to inspect required rolling stock components, and that cover a large variety of car types, both in the freight and passenger market.
These solutions can maximize the life of expensive components like wheelsets and brake components, enable condition based and predictive maintenance cost savings, and improve asset performance.
Operators using Trimble solutions are achieving real step change in how they inspect, operate, and maintain their rolling stock. Inspections can be almost completely, or in some cases fully, replaced by inspections carried out with Trimble Beena Vision machine vision technologies and image review. The frequency of shop or depot inspections and their scope can also be significantly reduced whilst at the same time improving operational reliability and improving safety outcomes.