THE CHALLENGES AND OPPORTUNITIES FACING THE RAIL INDUSTRY
Andrew Smith, Bentley Systems, talks exclusively to Civil Engineering Surveyor
The UK rail industry stands at a crossroads. Ageing infrastructure, rising demand and environmental pressures demand bold solutions. In this exclusive interview, Andrew Smith, industry solutions manager, Bentley Systems, shares insights on how artificial intelligence (AI), digital twins and cross-sector collaboration are driving a new era of rail innovation.
Looking at the medium to long term, almost taking a step back from rail, one of the most pressing challenges we’ve got is the environment. Rail is easily the most environmentally friendly form of mass transit that we’ve got, so by making sure we’ve got a rail service that is able to meet our needs first is essential. We need to make sure that the network we have is safe, reliable, cost-effective and sustainable for the future. Electrification is key, but so is modernising operations and embracing digital transformation.
As the demand is growing faster than engineers can design and build, what can innovation do to help combat those challenges?
Globally, there is a shortfall in the hundreds of thousands of qualified engineers to be able to work with civil infrastructure projects. It’s not a problem unique to rail. We’ve also got financial and time constraints. During the operations and maintenance phases, we need to run longer hours so there’s less availability to sustain the network and improve the quality of services we deliver.
We need to work differently and we have to work smarter. We have to take advantage of improved insights that we’ve got in order to be able to improve the efficiency of what we do. Innovation means changing processes, using better data and simplifying frameworks. Digital technology is central to this shift. Engineers still make decisions but now they’re informed by richer, more frequent and higher-quality data.
Rail is already strong environmentally, but we need to modernise and adopt emerging technologies. The big ticket item is AI. It’s a disruptive technology and by that I mean it’s an opportunity for the industry to fundamentally evolve. It’s an enabler for a revolution in the way that we currently work – a step change in problem solving and efficiency. Thirty-nine years ago I was studying AI at university. I’ve watched its evolution through to today – specifically in the last few years, where it has exponentially ramped up and become mainstream. The AI tools and techniques that have been perfected over those decades are now being applied very broadly across a wide range of problem spaces and coming up with great results. I would still argue, in many cases, that AI is very well placed to support and augment engineers – and not to replace them in the work that they do in turn.
It’s a great challenge and it is a genuine concern that some people may actually lose their engineering expertise and ability. Engineers must understand the principles behind their work. For example, when my daughters were studying mathematics they were allowed to use a calculator, but only when they could prove to me they didn’t need one. The same applies with AI. It can be used to simplify a process or to streamline a workflow but the user must be able to do the task longhand. AI can streamline processes, but engineers need to validate outputs and maintain expertise. AI is a tool, not a magic wand.
We’re in a world where there is a huge amount of data that can help make better decisions about how to sustain infrastructure.
It’s vital. Technology companies like Bentley Systems, policy makers, universities and research institutions all play a role. Academia pushes the research boundaries generating large language models, policy makers set the framework and the companies realise the advancements in technology to create deliverable digital platforms.
Industry then transforms these into something that can deliver practical value. Collaboration ensures innovation aligns with real-world needs.
We’re in a world where there is a huge amount of data that can help make better decisions about how to sustain infrastructure. However, there are challenges with the accessibility of that data. It sits in silos. It uses different languages. The systems don’t talk to each other. The departments that have those systems don’t necessarily talk to each other in the way that they should, which is reducing the ability for us to gain external insight and value out at the far end.
Many assets that we have in rail actually pre-date digital transformation; they pre-date the CAD world. We need to take these legacy assets and apply a baseline digital transformation to them. From that we can build the digital representation, the digital twin, alongside the physical asset that’s in place. Once you have that framework, it means you can simulate and test scenarios to build more sustainably and efficiently.
Globally, we’ve got a large amount of infrastructure that is approaching or beyond its design life and there’s been historic under-investment in the maintenance. People are using a range of emerging digital technologies to be able to assess infrastructure. We’re getting a much higher volume of data, at a much higher frequency, giving a more accurate representation of the current state of the asset. If we start linking this to other datasets, such as historic maintenance works, historic renewals, subsurface data, the design intent, the as-built record, we can start building a complete history of the asset. Once you have that, you can start predicting performance into the future, managing and understanding the scope of resources that you need to sustain the asset by being proactive and predictive, rather than reacting to failure.
The UK has been towards the forefront of transformation in digital technology. It was very much an early adopter of BIM and digital twins technology. Globally, there are also some local champions that are particularly advanced in terms of the work they’re doing. Some have the advantage that they’re building completely new networks. In Asia, for example, you have newly built networks that have beautiful, rich datasets from the outset that they can use to maintain against. Some major North American metros are pushing hard towards digital transformation by integrating new technology with the legacy assets that they own. That’s one of the key challenges.
Looking at it from two different angles, you’ve got the data security itself in terms of maliciousness and then you’ve got data security in terms of the intellectual property associated with that data.
When you look at the digital twin technology, there’s a distinction between a greenfield world, where you’re free to design an operational system that takes advantage of all this data, and a brownfield world, where you’ve got a legacy dataset that requires updating with new data from emerging technologies. You end up with a quality and completeness of data that varies over your network, depending on when you last touched it in terms of major infrastructure works.
By and large, the UK is in a quite advanced state in terms of the collaborative work that it does. There is always more that can be done in terms of improving integration further, better reuse of data, avoiding the concept of dark data and opening up data safely and securely. It’s not a free-for-all, however, but an effort to expose all the data out there so that people can take advantage of that data and use it to make more informed decisions going forward.
Data security is a major priority in a world where we’re seeing very high-profile cyber attacks taking place. Looking at it from two different angles, you’ve got the data security itself in terms of maliciousness and then you’ve got data security in terms of the intellectual property associated with that data. From the malicious side, we’re following all of the latest security protocols in terms of encryption, two-factor authentication and so on.
More importantly, in terms of the IP of the data that belongs to a user, we are very explicit that a user’s data is their data and their data alone. One of the key examples for this is the use of AI for training. What Bentley will not do is take a user’s data without permission, train an AI against it and then pass the resulting AI model out to other users.
In turn, we’ve got the ability for users to train our AIs on their data. They train it up and they own the model. If they want to share that data out so that others can take advantage of it, then there are mechanisms we have in place to allow that. Even within organisations, care has to be taken if you expose data to people who are inside the organisation – who are not malicious in any way, but are not trained.
There is a big risk of misinterpretation. When looking at a digital twin for all these datasets being brought together, care needs to be taken that people understand what it is they’re looking at. As well as externally, it’s the security internally that needs to apply across and within an organisation.
Yes, but we’re only at the start of that fundamental transformation that’s taking place. If we look at the kind of emerging technologies that are coming along, we’re moving towards interfaces where you’re either typing or simply talking to the system with an AI copilot sitting alongside it. In Bentley OpenSite, for instance, you’ve got the ability to speak to the platform and instruct to it to carry out tasks.
For example: ‘Tell me whether the number of car parking spaces I’ve got in the design are compliant with local building regulations?’ The system will then look at the CAD model, count up the number of spaces, look at where the where the building is situated, look at what the local standards are and then answer the question for you.
Now, a person could have done all of that, but there’s multiple steps that we need to go through in terms of that process. AI is taking a lot of that drudgery and searching work away from you. My expectation is that across the board, AI is going to be embedded more or less as standard, in terms of operating as a copilot concept. It’s not replacing the engineer that’s in place, it’s actually helping them understand the volumes of data that are getting to a point that they are too big for a human to process.
One of the key things is to make sure that people realise AI is a tool that can help them do their job more efficiently and not to fear it. Nvidia CEO Jensen Huang famously said: ‘You won’t lose your job to AI – you’ll lose your job to somebody who uses AI’. That message needs to be communicated from the top down.
Right at the start, we talked about the huge shortfall of engineers within industry. By using AI to increase efficiency, we’ve got more of a fighting chance with the engineers that we’ve got of being able to do all of the work that actually needs to be done – rather than being able to use AI as an excuse to actually reduce the size of the workforce significantly.
One of the key challenges with this – from the AI side in particular – is that the technology is advancing at such a rate, it would be dangerous to look into a crystal ball and predict the next five years, let alone 200 years in the future. If we look back 200 years in the past – even pre-dating Charles Babbage's invention of the analytical engine – the concept of automated computing of any kind didn’t exist. The transformation we’ve seen has been absolutely huge.
Going forward, there’s going to be more and more data. That data foundation is going to grow and grow and grow; in terms of the breadth of datasets that we have, in terms of the quality of the data that we’ve got and the frequency with which they all get updated. That very rapidly gets us to the point where humans simply cannot drill into that amount of data and try and interpret it.
One thing that AI is very good at is drilling into huge amounts of data and identifying patterns to help with decision making.
One thing that AI is very good at is drilling into huge amounts of data and identifying patterns to help with decision making. The data that sits underneath this becomes vital. We have to treat the data that we have as an asset in its own right and manage and maintain it to make sure it’s fit for purpose.
Data is going to be one of the highest value assets organisations have. It’s the core asset that is used to determine the work that we’re doing. Given that data has a longevity associated with it, we have to make sure that we’ve got open standards in place so that the data can be preserved across tools and over the supply chain.
We’ll certainly see a transformation in terms of how repetitive and dangerous tasks are carried out. The inspection process will change significantly. I expect to see fewer people physically on the track and more automation and digital inspection taking place. The industry needs to be more proactive, moving towards having more contextual tools that understand the design process, the requirements, the regulations and the environment that we’re building in. We’re already starting to see this and it will happen more and more with engineers utilising AI to handle vast amounts of data and turning that into information.
What we have to remember through all of this is that technology and digital transformation are tools. They’re immensely powerful, transformative tools, but they are still tools. The ultimate aim is to improve safety, reliability, cost-effectiveness, accessibility and perception of the rail network itself.
The engineers will then extract the insights from that, optimising the decisions that they’re making going forward into the future.
What we have to remember through all of this is that technology and digital transformation are tools. They’re immensely powerful, transformative tools, but they are still tools.
The ultimate aim is to improve safety, reliability, cost-effectiveness, accessibility and perception of the rail network itself. We need to make sure that we’re not using technology for technology’s sake.
We’re using technology as a tool to improve the quality of life of the people who are actually using the system that we’re applying that technology towards.
AI is not the future. AI is the present. It is already here. You are going to have to engage with it or be left behind, plain and simple. Make sure when you’re looking at going through the AI process, you have got a firm handle on data and security.
Manage your intellectual property and make sure that all colleagues within your organisation are on board with that digital transformation – people, processes and technology. Technology usually isn’t the sticking point, by and large – it does what it says on the tin. It’s making sure that people are engaged and motivated to see the value that the systems that are being put in place can actually make a difference to the service that they are offering.
AI is an immensely powerful tool that has the potential to transform your organisation, to make it more profitable. But, as with any other tool, if you don’t know how to use it properly, it’s not going to give you the value that you’re expecting it to.
AI and digital transformation are not distant concepts – they’re here today. For the rail industry, embracing these technologies is essential to meeting environmental goals, improving efficiency and delivering a resilient network for generations to come.