Andy Kunzmann, Tracsis Chief of Technology, gave railroaders an overview of the current state of RCM technology and its future applications, including the shift to AI-driven predictive maintenance, during an event for the IRSE (Institution of Railway Signal Engineers) North American section.
Hosted by IRSE Member Representative Stuart Maddock (CEng MIRSE), the webinar tackled a broad range of connected topics, from ROI and building a business case to how advanced applications of AI and Machine Learning are enabling prescriptive data analytics. Here is a summary, based on Mr Kunzmann’s presentation and his answers to participants’ questions.
1. What is Remote Condition Monitoring?
RCM is a maintenance strategy that monitors the health and performance of assets from a remote location. It vastly improves the visibility of rail assets, allowing maintenance to be proactive rather than reactive and enabling resources to be used where they are needed most in the near-term.
Remote Condition Monitoring provides decision-makers with insight into the effectiveness of asset maintenance and overall asset lifecycle performance. This allows them to compare performance before and after maintenance events, and to get notified when an asset is operating outside of its expected parameters.
In the longer term, RCM allows for better planning of maintenance operations and for capital expenditures. Decision-makers can view real time and historical data to understand the individual performance of each asset, allowing them to see unique performance characteristics and how performance changes over time in response to weather and other events.
2. How does RCM work?
Data loggers and sensors are installed wayside to collect data from wayside signal assets. That data is sent to the cloud, where it is analyzed and compared to predefined thresholds. When an asset performs outside its expected range, an alert is sent to maintenance personnel to take action. Dashboards provide trending analysis and various metrics on asset performance.
Discover more about how Remote Condition Monitoring works.
3. What are the benefits of RCM for railroads?
Increased Safety: identifying and mitigating potential hazards, while minimizing time in the field
Reduced Costs: enabling railroads to focus on assets in decline and to avoid unnecessary maintenance
Reduced Delays: Improve the reliability of rail equipment and systems, leading to increased uptime and productivity
Better Decision-Making: a structured framework with objective data for making decisions about railroad maintenance
4. Who benefits from the use of RCM?
Railroad decision-makers, planners, maintenance managers, and maintenance engineers. Decision- makers can use RCM data to better plan for capital expenditures; planners can get insight into individual asset lifecycle performance; maintenance managers can understand the effectiveness of asset maintenance; and maintainers on the front line can get notifications when an asset is operating outside of its expected parameters before it fails, and can also evaluate the anomaly before going out to the field, which can reduce on-site time.
5. What does the future look like for Remote Condition Monitoring?
The use of large data sets will enable RCM to look at related activity and data to understand the assets’ environment. The use of open APIs will enable these large data sets to be used in RCM analytics. Some examples include using weather data to understand environmental effects on rail asset performance, or referencing maintenance activity to understand what has been done and how that might relate to asset performance, as well as maintenance activity in the surrounding area. An example could be track maintenance affecting a power switch machine’s performance.
Advanced applications of Artificial Intelligence and Machine Learning will enable railroads to implement prescriptive analytics, so we are able to understand what is going to happen before it happens, and what should be done to prepare.
With more data going into the RCM platform, it will be possible for more curated information to be out-put, such as user persona dashboards that provide decision-ready information to various roles within an organization. This will help with evaluating the effectiveness of maintenance programs and to understand the general landscape of assets, such as how many good versus bad assets we have.
State of Good Repair will identify which assets are performing well, which are not, which assets fail the most and the least, which assets show more frequent performance variability, what types of anomalies and failures occur the most, and what correlations can be made between failures and root causes.
Capital spend allocation will be better defined by understanding when a given rail asset will need to be replaced and which assets cost the most to maintain.
For remote condition-based maintenance, which tests and inspection activities can be done remotely? And what kind of trouble-shooting can be done remotely before going out to the asset location? These are just some of the possibilities for the future of RCM.
6. How far away are we from having enough data available for AI to determine appropriate intervention levels and maybe even advise on courses of action?
That is the ultimate goal. From a practical standpoint, the data probably already exists to do this – it’s a matter of putting the system together, training it, and verifying that it’s actually giving good information.
As we’ve seen within the last year or so, AI capabilities have really taken off. We’re not talking about Gemini and ChatGPT and the human-machine interface; we’re talking about machine-to-machine AI.
Today, we’re solving these problems manually by going out to the field, making visual observations or taking measurements. What we’re talking about is automating this and advancing the process through AI and machine learning.
7. What is your experience with railroads and transit agencies in the US utilizing RCM today?
It's been a moderate adoption so far but we’re seeing more interest. Remote Condition Monitoring is not a new concept – it’s used very heavily in the IT world with servers, data centers and power systems, but it’s somewhat of a new concept in the signaling and communications world. These are some potentially large data sets that can be mined; I don’t think we have even scratched the surface of what we can do.
8. How can railroaders get buy-in from decision-makers for implementing Remote Condition Monitoring?
It's always about Return on Investment – if you’re going to spend money, you need to make that money back and more. And it’s also about improving quality: quality of life, quality of the work environment, quality of the service you’re providing to your customers. You need to focus on pain points, which may be related to safety, resources, or reducing delays. RCM really touches on all of those topics for railroads.
9. What measures should be taken to ensure that the sensors themselves do not become a source of failure or erroneous data?
The first thing – and this goes for any type of RCM solution – you need to make sure it’s non-intrusive, so, if you do have some type of monitoring equipment failure, it has absolutely no effect on the mission-critical systems being monitored.
As far as erroneous data, you need to make sure that you’re paying attention to your data, looking at your thresholds, and adjusting thresholds for the operating conditions, such as temperature and humidity.
10. Are there signal assets that you would recommend be monitored over others (i.e. track circuits over switch machines or signal lighting circuits over track circuits?)
That really depends on the operator’s pain points. Start with assets that you struggle with, that you are constantly chasing a good state of repair on, and that’s where the true savings and performance improvement can be realized. What type of asset problem is having the biggest negative effect on your business? Start there.
Find out more about Remote Condition Monitoring. Download our brochure.
