Operational optimization
Harnessing tech to make day-to-day operations more efficient
Running a network of trains involves more than managing their routes from A to B. It needs a whole infrastructure to check and maintain trains and tracks, ensure trains can leave on time, and to respond to problems quickly.
Here, again, technologies can help ensure efficient incident management and improve traffic flow. Transforming rail operations is less about complete transformation, and more about hundreds of incremental gains that add up to big cost savings. In this article, we’ll look at three, before drawing some broader conclusions on setting up organizations to continually use digital approaches to optimize operations.
Predictive maintenance (PdM)
Probably the most talked about digital solution for operational efficiency is predictive maintenance.
Predictive maintenance uses data to proactively detect anomalies and predict equipment failures before they occur. That allows maintenance to be scheduled at convenient times to repair issues before they become a more complex or expensive problem.
An example could be a door sensor that shows incremental slowing of the closing mechanisms that would not be detected by the human eye until it became an issue, or a sensor that detects subtle vibrational patterns that indicate a flattening of the wheel, or defects on the track. By establishing normal work conditions for assets, PdM can compare the actual infrastructure state to baselines, thus detecting the start of any drift, and so plan targeted maintenance to address the issue.
Further algorithms can be used to predict when this will actually become a problem. A wheel flat may affect safety, by damaging tracks and switch needles, and by reducing comfort (through extra noise and vibrations) and thus need to be corrected as early as possible - whereas a slowing door may still have weeks of usable life before it actually affects the customer experience. Predictive algorithms can tell you with greater accuracy what is failing and when the required intervention is needed to avoid such failure.
This can be backed by whole systems that optimize that scheduling - to ensure appropriate maintenance, at the right place, at the right moment, by the right people.
Implementing predictive maintenance can be a long process, which requires an understanding of each maintenance type, and the possible solutions. In each case, the first step is to set up systems to collect historical data to set the baseline, and live data to analyze failure patterns. We must also digitalize the whole maintenance process - from fault detection to maintenance scheduling, repair tracking and updating the parts history book.
Making day-to-day tasks easier with AI assistants
Thousands of rail engineers work around the clock to maintain vast and complex railway networks, so that passengers are safe, and the railway operates smoothly. Behind every operational procedure they conduct, are a slew of minor processes that take time. Some are burdensome but safety-critical, some are there out of habit. New technologies could be applied to streamline many of these processes, and make engineers’ lives easier.
Having the right crew at the right time and with the right skills is critical - so predicting resources, crew availability and resilience is an important capability. AI can be used to help with this kind of activity by anticipating sickness rates based on seasonal data, and potential service disruption or the lack of particular competence for an anticipated repair work can help to optimize resource use and network efficiency.
According to a 2023 Capgemini Research Institute survey [4] across 800 organizations, 67% of executives see the most potential for generative AI in the IT function to drive innovation and create value.
Gen AI’s ability to parse human language also means it can be used to analyze day-to-day interactions and propose improvements. For example, it can be used to analyze emails or transcripts of phone calls between engineers – thousands of which happen every day – and to identify whether they met guidance for being clear and concise when providing safety critical information. Capgemini is currently working with major rail operators to develop a proof of concept for auditing maintenance safety- critical phone calls – with appropriate consent of course – to learn such lessons.
Tracking and inventory management
Running a railway involves a lot of physical ‘stuff' that moves around, from tools and technologies for maintenance, to freight that needs to be loaded onto and off of trains. A surprising amount of time can be lost just looking for things. Tracking technologies, backed by digital management technologies and digital ledgers can bring some much-needed order.
For freight trains, there is a need for yard optimization to facilitate efficient loading and unloading of freight. That starts with knowing where everything is, all the time, through tagging freight, and having connected trains. For example, every crate has an RFID tag and unique ID that can be scanned when it leaves a customer’s premises (like a factory
or grain facility). When it arrives in a yard or boards a train, it pairs with the train’s GPS. All of this is tracked in a digital ledger - so there is a live view and record of where everything is at all times.
This approach can be applied to the entire freight transportation value chain, where a unit of freight is tracked from its point of origination (eg. factory, mine, grain facility) through its various transport modalities (eg. truck, train, ship, air cargo) - all the way to its final destination.
Such tracking provides customers with an answer to the common question ‘where’s my stuff?’, which provides reassurance and supports efficient planning of loading/ unloading. Similar tagging and tracking systems are also hugely valuable right across networks for quickly locating the tools needed to perform upgrades or fixes. We could also use AI to optimize the order of the carriages and wagons that constitute the train - tailoring them exactly to the needs of the journey, the kind of cargo being transported, and so on.
Whilst the immediate benefit is organizational efficiency, over time such systems gather rich data on which to design better systems. For example, historical data about grain transports will show patterns of transport needs linked to harvest season, which can be mapped to market demand models, and other forecasts about seasonal and event-based train demand, which allow long term schedules to be designed to maximize train use.
In summary
Today's rail maintenance and operations centers often rely on aging applications and technology.
Once again, the theme here is joined-up systems, which consolidate various legacy systems into a single, standardized approach, with an effective set of tools to draw data and inform the relevant workers of where things are, and what needs doing in the most efficient way possible.
Digital solutions like Enterprise Asset Management (EAM), Computerized Maintenance Management Systems (CMMS) or Asset Operation Management (AOM) all play a vital role in planning and scheduling. But true optimization relies on connecting everything up – from sensors, to management software, to engineer’s mobile devices – all joined by a cloud-based system, applying the right analytics, and designing the right user interfaces.
All of this must balance efficiency with keeping people safe, and respecting people’s security and privacy. But with the right approach to people and technology, there is lots of room for small improvements across the board that will add up to big improvements and savings for rail companies.
