Is it possible to model a station in 3D without being familiar with the location, having limited information and not holding face-to-face meetings with the development team? Just a few years ago, it might have seemed like fiction, but nothing could be further from the truth. Bentley Systems is the protagonist of this feat, and COVID-19, the main trigger. Pandemic has driven digital transformation in many sectors, and construction is no exception. In this context, the software developer tells us how artificial intelligence and digital twins stand as allies for tasks such as optimizing asset maintenance.

On the occasion of the Year in Infrastructure conference in October, we interviewed Steve Cockrell, Industry Marketing Director of Rail and Transit at Bentley Systems, and Andrew Smith, Solutions Executive for Rail and Transit at Bentley Systems, to explore post-pandemic construction, how the data and behaviors in transit have changed and what we’re to expect from the times to come. Covering the particularities of rail and transit infrastructure, we also emphasized what the restrictions of movement have meant for construction and asset management. Especially, in Europe or the United States, where the number of new infrastructures will not be as important in the coming years. 

What has changed or is changing in rail and transit due to Covid? And how it has and will affect the construction of the new infrastructures?

As an example of how lockdown and movement restrictions have changed the way of working, Bentley’s Industry Marketing Director of Rail and Transit, Steve Cockerell uses the project of Exeter station where the team had to model the station in 3D for the purpose of signal sighting. In this case, Network Rail Wales and Western Region (NR WW) had to launch to the task without being familiar with the location, having limited information and holding a face-to-face meeting with the development team. 

So, over the course of three days, they used drones to collect different data types, point clouds, laser scans, photogrammetry, and bring that information together to build a 3-D model. In effect, that same 3D model will be the foundation of a digital twin and can be used for signal sighting from distributed locations.

Aerial view of Exeter station signal sighting digital twin built with MicroStation, Descartes, and OpenRail Designer in three days (Image courtesy of Network Rail)

Similarly, in the London Paddington station, the NR WW had to catalog over 100 elements of existing signage to develop and devise new ones for wayfinding renewal that included visibility and commuter navigation. For that, the team conducted point-cloud scans of the station layout and all assets. 

In these projects, digital workflows permitted the teams to complete the tasks faster and with a greater level of confidence in their accuracy. “It has enabled them to deliver improved service to their client and to different stakeholders. On the other hand, it has also enabled them to attract further investment into the modeling team,” Steve concludes. So without a doubt, Covid-19 has been a driver for digital transformation in many sectors, and construction is one of them.

Paddington station digital twin used for wayfinding upgrades and renewals. (Image courtesy of Network Rail)

Paddington station digital twin used for wayfinding upgrades and renewals. (Image courtesy of Network Rail)

What’s the role of AI in Asset Management? And what’s the potential of digital twins in this scenario? 

The restrictions on movement have altered the way we move around and carry out our lives, and all that affects maintenance and rail asset management. As Solutions Executive for Rail and Transit Andrew Smith explained, the linear analytics capabilities and coming up with a new framework that allows us to measure the condition of the asset will help us to optimize the maintenance and adapt it to the level of usage. 

When it comes to rail asset management, it’s absolutely vital that we can count on technology to sort through the data for us. That’s because an individual sample might not tell us what’s the problem, but the relationships and history will. “AI is very good at identifying those patterns, being taught those patterns and then being able to churn through vast amounts of data and flag up areas where it’s seeing similar characteristics and grow over time,” Andrew goes into detail.  

The outlook is that we will have more and more connected devices and more and more data to sort through. Andrew concludes: “This is where the digital twin and then the AI building as a layer on top of that can be used to place the data and see it in context. We can rely on it to build the information that allows us to make decisions to be able to optimize the way that our systems work, which ultimately adds value to the end-users.” 

How to adapt the new technologies to existing infrastructure? Or is it the other way around? 

Over the life cycle of the infrastructure assets, it’s not uncommon to hear that 80 percent of the total spend on the asset is in its maintenance phase and 20 percent in the design and construction phase. So, saving one or two percent in maintenance compared to design and construction actually gives a much bigger return over the life of the asset. 

However, building a data-driven maintenance model for an existing infrastrucutre is not an easy task: the datasets are often fragmented, incomplete, and not up-to-date. They sometimes contradict each other, so the operators end up making decisions based on incomplete and inaccurate data. So, if we were to apply digital twin here, we would have a skeleton on which we can actually build and relate the datasets in order to be able to support decision-making. 

As an example, Andrew brings one of the YII Award finalists, SMRT Trains. They operate and maintain over 140 kilometers of rail track in Singapore and count with a daily ridership of over 2 million people. With the help of AssetWise Digital Twin Services and AssetWise Linear Analytics, they went from manual maintenance planning across separate data silos to overlaying multiple data sources within seconds rather than hours. All that allows SMRT to optimize the efficiency of a crew’s maximum work capacity during one shift and ensurie the reliability of the rail network in general.

 

No more articles