The future of the Railway Sector is Edge Computing

Obstacles to digitisation in the railway sector

Studies such as the Harvard Business Review "Which industries are the most digital and why" place the transport sector in general at the bottom of digital maturity. And the rail sector, in particular, contributes significantly to this low digitisation. This is not only because it is subject to strict regulation, but also because technically it presents a multitude of challenges that make it difficult to integrate existing systems in trains or railway infrastructure into modern digital architectures.

In the article written in 2020 by several specialist members of the IEEE (Institute of Electrical and Electronics Engineers) they identified the technical challenges of integrating big data in the railway sector. They did so by grouping these challenges into what they called "the 5 uves", which include:

• Variety: it is a highly fragmented supplier and technology environment. There are hundreds of different systems on a train, and in many cases they are obsolete due to their long useful life and the low technological turnover typical of traditional industries.

• Volume: along trains and infrastructure, huge amounts of data are generated, which in aggregate can reach Terabytes per day and cannot be processed by traditional hardware and software.

• Speed: not only is a large volume of data to be processed, but processing times must be very low to be considered in real-time or near-real-time processes.

• Accuracy: due to the critical nature of railway operations, it is essential to ensure the accuracy and reliability of any advanced data processing.

• Value: Most of the systems to be considered have a high value, so the infrastructure is widely dispersed geographically.

Edge Computing the answer to the digitisation of the rail sector

If we analyse the strengths of Industrial edge computing as a digitisation technology, we realise that it emerges precisely as an answer to most of these challenges faced by the railway sector. Edge computing enables the analysis of large volumes of distributed data, in real time, in a cyber-secure, scalable way, and allows the integration of field equipment from a multitude of suppliers, technologies and protocols.

There is no doubt, therefore, that the penetration of industrial edge computing platformsthat enable optimal processing of data from sensors, radar, cameras, on-board equipment and other sources to provide more granular and holistic visibility and operation of railway operations will increase.

David Purón CEO Barbara IoT

Here are five use cases that show the potential of Edge Computing in the railway sector:

1. DIGITAL TWIN AND PREDICTIVE MAINTENANCE

The digital twin, or more technically speaking, digital twin instances, understood as the individual, real-time digital representations of each of the working parts of a physical system, is one of the most pioneering innovations of Industry 4.0, but also one of the most profitable in business terms.

These digital instances are essential tools for optimising product value chains, from manufacturing to maintenance and after-sales service. They can be used for remote monitoring and diagnostics, which saves high travel costs for specialised personnel. Even more important is the automatic anticipation of possible incidents, combining historical data, human experience, machine learning and simulations, which makes for better forecasting results every time. The use of digital twins and simulations is already key to the prediction and identification of components that may present problems in the short term on tracks and trains.

McKinsey has calculated that 51% of companies using artificial intelligence in their operations have seen their operating costs reduced, in one in ten cases by more than 20%.

This cost reduction has always been the big workhorse of railway companies. Decisions on where to locate supply bases, how and when to plan overhauls, what materials to use, etc. make hundreds of millions of dollars in operating cost differences for operators and manufacturers. Therefore,the digital twins and their impact on maintenance are hitting directly at the waterline of these companies.

Digital twins in real-time operations need to process very large volumes of data with low latency. For example, a vibration sensor for motor error detection requires an algorithm that processes data at a minimum rate of 1kHz (1000 data per second). This, coupled with the inherent coverage issues of any transport, makes edge computing the most suitable technology for these use cases.

2. OBSTACLE DETECTION

Safety in rail transport is traditionally another major challenge facing operators. To cope with human error, poor visibility due to bad weather or obstacles, or the like, computer vision is increasingly valued as one of the lines of improvement towards smarter and more automated transport. A good automatic obstacle detection system helps drastically to improve the emergency handling capacity, and thus the safety of travellers or pedestrians.

So much so that the Shift2Rail organisation, whose mission is to define and deliver digital capabilities that make European rail transport a more customer-centric and sustainable mode of transport, is dedicating a working group and an entire project to the subject.

However, obstacle detection is too critical and computationally intensive an issue for which only edge computing architectures can cope, saving significant network costs, as well as increasing data privacy by not having to be processed and stored in centralised infrastructures.

3. SOFTWARE AND FIRMWARE UPDATES

On-board systems on trains are becoming smarter and more "software ". Anyone travelling on a modern train today is used to seeing touch screens, Wi-Fi networks or digital security cameras.

The requirements and use of these systems is changing over time and it is therefore common to require many iterations to be able to optimise the use of these systems. Likewise, these IT systems are often quickly rendered obsolete by the discovery of security vulnerabilities, which can be a gateway for cybercriminals. This is why the ability to update configurations, software and firmware on these on-board computers is becoming increasingly important.

A unified system and a seamless remote equipment update process can save hundreds of thousands of euros per vehicle in downtime and technician hours. As a result, manufacturers such as Alstom, are already deploying containerised applications on the EdgeThese architectures reduce manual processes by automating the lifecycle of devices to deploy new versions and patches in real time or on demand.

4. IMPROVING THE SAFETY OF TRAINS IN MOTION

To ensure the safety and stability of a moving train, severe and real-time monitoring of parameters such as speed and load are of great importance. The combination of IoT sensors with high computing power is an optimal solution for this.

With sensors that can be as small as 5cm2, high frequency rail vibration information caused by the wheels of passing trains can be collected, and using edge computing, both the speed and state parameters associated with the current running risk can be calculated.

Scientists at the University of Hong Kong have demonstrated that continuous 24-hour monitoring is feasible with this architecture, with impressive results such as speed errors of less than 0.2 km/h, and with the advantage of taking up very little space on tracks and trains, and at a much more controlled cost than traditional systems that have traditionally performed these functions.

5. IMPROVEMENTS IN HEALTH AND COMPLIANCE

There is no doubt that the COVID-19 pandemic has suddenly and abruptly forced many industries to change their priorities, and to deal more efficiently with related issues such as health and social distance.

Edge computing enables real-time monitoring of aspects such as air quality, compliance patterns for wearing face masks, or social distances and gauges in stations and trains. With more advanced algorithms, it could even identify areas that require cleaning or disinfection, and even self-guide automated cleaning systems to clean or disinfect areas.

Barbara's Industrial Edge Platform

Edge computing is being implemented as one of the most relevant enabling technologies in the digital transformation of the industrial fabric. According to Gartner, by 2025, 75% of data will be computed and processed on the edge, giving rise to new opportunities for services and applications. To be able to govern this new model of distributed intelligence it is necessary to have an Edge platform, such as Barbara's, which allows:

1. Connect, analyse and manage data from any industrial asset and integrate it with remote business logic.

2. Develop, deploy, debug, operate and maintain edge computing applications and algorithms.

3. Protect equipment and data with cyber security mechanisms designed in accordance with industry standards such as IEC-62443.

If you are interested in this article, we encourage you to contact us to explain how Barbara can help you in your edge computing project.

‍