Engineers across the world suffer regular headaches over infrastructure buried underground. Whether it is the stress of making sure their team avoids striking hidden electricity cables, not knowing where ageing assets are during disaster responses, or costs soaring after stumbling upon unexpected pipes, having a clearer picture of what lies under the surface would ease a lot of minds.

For years a group of leading data scientists at the Open Geospatial Consortium, an international panel of experts, have been grappling with this problem and how to solve it.

They formed a working group to create a concept – the Model for Underground Data Definition and Integration – known as MUDDI for short. Its purpose was to create an international standard for mapping geospatial data underground. The model visualises subsurface infrastructure assets and characterises the underground environment that contains them, then defines a data store for it all. This information can then be shared easily among all parties working in the same underground space and turned into a map.

Ordnance Survey’s Chris Popplestone and Carsten Roensdorf have been key contributors to the project, lending their expertise and thinking to create the MUDDI model.

Chris explained the MUDDI model was developed from existing geographic data standards INSPIRE, CityGML and the International Standardization for Organization, and will eventually become an international standard in its own right.

He said the concept of a ‘harmonised data model’ can be applied to lots of diverse data challenges and MUDDI was an example of that.

“The harmonised data model is that central target into which all of the source infrastructure data is transformed,” Chris said.

“The source data comes in all different shapes and sizes, but it all gets transformed to this harmonised model.

“We start off with a conceptual model of how this transformed data will look and we use various tools to focus that down into a physical model that can then be implemented as a data store. It’s a process from a high-level conceptual design down to an actual physical schema which can then be deployed by a development team.”

Proof of concept

The first chance to prove the MUDDI model worked came with pilot testing for the National Underground Asset Register (NUAR), a programme led by the UK Government’s Geospatial Commission.

Two pilots were successfully user tested in both North East England and London, and work has started to build the system to roll out nationally under the leadership of engineering consultants Atkins. The national roll out of NUAR will involve developing a UK profile of the MUDDI model, focussing on the safe digging use case.

Data is collected from 650 or so data or asset owners by software company 1Spatial, which is delivering the data transformation and data ingestion part of the process.

OS’s development team is shaping the harmonised NUAR data model, based on MUDDI, as well as building the data store and creating the platform that will serve up the transformed infrastructure data through a map-based user interface.

If all goes to plan, it will give people in the field a tool where they can draw an area on the map for where they are planning to dig up a road, to reveal the pipes and cables that they need to know are already there in a clear, uniform way.

Chris said: “The NUAR pilot was based on an earlier version of MUDDI, and the feedback we got from the pilot phase has fed into a new iteration of the MUDDI model, and that is now forming the basis of the national roll out of NUAR.

“There has been a lot of international involvement and that is being looked at very carefully for other international projects that are also planning to follow a similar approach.”

Challenges

What makes the MUDDI concept exciting is it can be adapted to suit different use cases, and different national regulations and standards. This is an idea designed from the ground up to support replication around the world. However, to make it work, the model must be fed, and that is where the challenges lie.

Collecting data is a messy, complicated process. It gets even messier and more complicated when it has been collected by hundreds and hundreds of different organisations and companies of various shapes and sizes. Each one may have classified its data a different way, meaning there can be several ways of describing the same type of pipe. The same type of cable found in West Sussex can be called something completely different in the West Midlands.

Chris said: “Because we know that is a risk, we have tried to make everything about the data model and the implementation of the data model as responsive as possible. Working under the Geospatial Commission we helped to build in flexibility to the UK’s ‘excavation profile,’ alongside strong governance structures, to try and streamline any additions that need to be made, so we don’t have to completely tear down data stores and rebuild them because we have found a new type of pipe.

“We have tried to make it quite modular and quite flexible.”

The problem is prevalent across the UK, where almost all of the country’s infrastructure is privately owned and split between different companies' region by region. Each of these use different data systems and the data they record is different. In some cases, they have radically separate ways of describing and attributing things.

Chris said: “That is part of what the data model is designed to solve. There is potentially so much variation that we have to be able to make tweaks and add things that we haven’t come across in other data sets. So, things like terminology can be quite different between different companies – we must make sure when we come across a new term it is a genuinely a new thing rather than something else that we have already modelled called by a different name.”

Other domestic uses

The harmonised model concept is in use for other national infrastructure projects in the UK.

OS is just finishing its work supporting a national project looking at energy systems.

Chris said: “This is a completely different data model to MUDDI and NUAR, but it is the same idea. Again, you have several different data providers with completely different data sets and the data gets transformed to this target and then presented in a consistent way.”

The concepts behind MUDDI can be applied to the management of diverse types of data. Examples could be land use, land ownership or flood risk. Again, the model would be built along the same framework, with diverse data sets and different data providers, all coming through a transformation pipeline to a central harmonised target.

The MUDDI working group has also started work examining issues created by movement within sub-surfaces underground. Problems can develop rapidly, such as flooding in sub-surfaces, or gradually with the spread of contaminant. The group is testing two use cases using MUDDI - surface water flooding underground structures via vertical shafts and leakage from underground storage tanks.

International

Another benefit of the MUDDI model is that it can be adapted to build national profiles, taking into account the differing regulatory environments found underground around the world.

Voices at the Open Geospatial Consortium that helped shape and influence MUDDI come from as far and wide as New Zealand, the Netherlands, Belgium, the US, Canada, and Scotland.

Carsten, who co-chairs its working group and continues to advise on the development process, believed that meant there was a great chance for MUDDI to succeed globally.

He said: “Through MUDDI and other forums we have built-up a good picture of the worldwide activities in the underground data integration space.

“We are watching the development in other countries, across Europe, in North America and parts of Asia. Through these activities we are also getting as better understanding about a variety of use cases, including answering environmental questions and increasing a city’s resilience, for example in dealing with bigger storm and flooding events that affect underground infrastructure.

“It is interesting to see the differences in the implementation in different countries and cities and even better to spot commonalities between the different approaches. “

The hope is, with the progress the MUDDI model is making underground, engineers of the future will endure much milder headaches while pondering how to dig up street corners in years to come.

For more information about Ordnance Survey data insights visit the OS Control Today, Shape Tomorrow webpage.