SALFORD University’s industry link-up THINKlab has been working with Network Rail to develop a 4D simulation tool for track renewal. The system is built on Epic’s Unreal Engine 4 (UE4) platform and uses gaming visualisation to aid planners as they explore options when scheduling works. THINKlab director Professor Terrence Fernando spoke to Abigail Tomkins about the new program.
First things first, what’s THINKlab?
THINKlab was established about 14 years ago to support interdisciplinary research with industry partners. It is about exploring possibilities – people come to us to see what new technology and concepts are available and how they can use that knowledge to enhance their competitiveness through innovation.
In the past, we would publish our research findings in journals. However, we realised that in order to have lasting impact in society and industry, we have to go the final mile with the end users. You have to take that research knowledge, package it and apply it in industry contexts.
How does this work in rail track renewal?
The rail industry has experienced overruns in track renewal programmes, mainly because planning errors are not identified before starting work on site. This has cost the industry financially, as well as its reputation with customers and commuters through disruption to train services, and it has an impact on the UK economy. We’ve seen the aerospace and automotive industries use advanced modelling, simulation and visualisation technology to avoid such errors and improve efficiencies – we are taking this approach to track renewal scheduling in 4D.
Lots of complex engineering activities need to be considered both in time and space during scheduling. Planners need to explore various options and choose the most efficient and cost effective. Furthermore, track renewal scheduling needs to interface with upgrades to signalling and OLE (overhead line equipment). Teams involved in these disciplines need to work together, otherwise you could get situations where the PWay engineers take the track out and there are no tracks for the engineers putting in the OLE.
How did you come to work with Network Rail?
Network Rail approached us around five years ago to provide modelling support. They were looking at a nine-day blockade on a track renewal programme, and wanted to demonstrate why it had to happen and how it was going to be handled. Track renewal is a complex process. It involves heavy plant, and health and safety and space requirements need to be considered. You need to have a sequence of activities – first bring in the machines, take the tracks out, flatten it and go through layers to reconstruct it. There can be lots of clashes between various tasks involving machines. Many parallel and sequential tasks need to be managed in time and space to run the renewal process without any errors. The standard approach was to use 2D drawings, PowerPoint and animations to explain to the team what was going to happen. But often by the time the animation was delivered, the schedule had changed, and it was no longer valid. This kind of system did not allow you to explore any space-time clashes or different alternatives.
There are some commercial 4D simulation tools for building construction, however these require you to break down the building information model into various sub models and bring them into project planning software. We’ve found this approach is not intuitive to the planner. It takes a lot of cognitive load and time to deconstruct the model to prepare a visual schedule. For rail, it is overkill. We wanted to create an environment which is in line with planners’ thinking. We felt that it needed to be a 4D planning package that allows users to think in 3D and plan in 3D in an intuitive manner, linked to the planner’s thinking process.
How does the system work?
You begin by bringing in site data – digital elevation models, aerial photographs and track data – into a virtual workspace, which is built in the Unreal game engine. You could use Bentley software or any other system to draw your tracks, the only thing you need to import is the centre line. You can then bring in elements like temporary facilities, entrance points, bridges and OLEs to create a realistic view of your site. The data could come from many sources, including Network Rail’s Geo-RINM system which holds GIS, track and OLE data.
Once the site is prepared, the underlying logical network of the rails is automatically constructed to handle various movements of plant on this virtual site. There are plant and task libraries, which allow you to define the task schedule using a time-distance diagram, and as you plan you can really see what’s happening visually in front of you. The planner can use the task and plant libraries to define the schedule and attach appropriate resources against each task. Once the tasks are specified, the planner can play with various schedule and resource options and make sure the schedule is clash free and cost effective. As the planner makes changes to the schedule, the resulting animation is updated straight away.
What difference has gaming technology brought to the system?
It’s all about visual realism. We use Unreal Engine for nearly all of THINKlab’s visualisation work. We started migrating most of our projects to UE4 shortly after it became public, as it provides much easier and faster iteration of our scenes.
Where is the system being used now?
Currently, the system is being used at North Wembley. The Central Alliance has done a very comprehensive planning process using our platform. It’s been applied to look at which part of the track is going to be taken possession of, which plant is going to be used, for how long and what space it is going to operate in. It’s not just about plant manoeuvring. Certain sections need to be blocked to support welding and the safety aspect around that.
How hard is it bringing innovation to the rail industry?
Innovation within major industry sectors takes a long time. As academics, we go from project to project, because we cannot wait until the industry catches up with the innovations that we create through our research. However, we managed to keep the team on this one for about five years, with around 15 projects coming through from Network Rail, which gave us some funding to enhance the functionality. We are now at the stage where we have brought really pioneering work into Network Rail, and our challenge is about how we turn this into a product.
To bring this innovation throughout an organisation like Network Rail, you are going to have to train people in a new way of working. We’re doing lots of presentations and providing training to those alliance partners and contractors who work with Network Rail.
We are also in discussions with rail organisations in France and Denmark because, internationally, it’s a very similar problem that people are facing. Going forward, we’re looking at the huge possibilities of the system in terms of supporting design reviews and maintenance in the rail industry.
Professor Terrence Fernando, Director, THINKlab, University of Salford, was talking to Abigail Tomkins for Civil Engineering Surveyor, June 2020.
A simulation of the program at North Wembley can be viewed at: https://vimeo.com/420202951
Read more about the digital work with industry partners.