With the rapid-fire implementation of BIM (Building Information Modelling) in the AEC industry over the past decade, computer models are increasingly replacing traditional methods of data provision – gone are the days of handwritten records and two-dimensional sketches. The use of models allows workflows to be streamlined, and the virtual construction site developing alongside the analogue increases the quality of the final product whilst greatly increasing transparency for actors as the project develops.
The software itself has developed exponentially in terms of capability, as the producers of such suites seek to take advantage of emerging technology and trends, and fold ever-more-numerous dimensions of project management, higher automation and greater processing muscle into their offerings.
The Internet of Things (IoT), Virtual Reality, Augmented Reality, and 3D printing are technologies already being integrated into BIM workflows. Time management tools are now packaged (Known as 4D BIM), costings (5D) and increasing aspects of project lifecycle management (6D). While we can’t cover the entirety of developments in this fast-moving space, join us as we lift the lid on some of the current trends, their drivers and projected trajectories.
4IR + machine learning
The Fourth Industrial Revolution (4IR) refers to high-level integration of advanced automation and digital technologies to create true cyber-physical systems. While present-day 3D and BIM processes aren’t this evolved, some of the core techniques and driving technologies of the 4th Industrial Revolution, like cloud computing and artificial intelligence are shaping the next generation of tools for the construction industry.
BIM and modelling giant Bentley Systems acquired machine learning and IOT-focused start-up AIWorx in November last year, and has been busy leveraging its IP portfolio for integration in Bentley’s vast suite of offerings. The start-up has introduced advances in data compilation and evaluation to leverage digital twins in infrastructure engineering, continually accessed with contextual information in real-time, to optimise productivity, activities and servicing.
Phil Christensen, senior vice president of reality modelling and cloud services at Bentley Systems Australia, said that the firm seeks to utilise machine learning in a number of ways.
“As we capture more and more data about projects then we can use artificial intelligence to find trends in that data. If we look at the stream of data that comes in during the engineering phase, alongside the rate that the BIM model is developing and all the different teams are working on it, how it’s evolving – it can look at that data stream and find patterns to identify projects at risk or identify where the weak points that are in the overall supply chain,” he said.
Zooming in somewhat, machine learning can be applied to the creation and analysis of 3D models in a range of different ways. Christensen says that using machine learning to identify objects within a reality mesh can translate directly into deliverables.
“If it was a civil project you could use machine learning to identify construction vehicles over then to get an accurate terrain of it would be one example,” he said.
Automating model creation from site photos with machine learning is a game that others have cottoned on to. Leica Geosystems’ acquisition of Melown Technologies has granted them access to their Vadsten 3D model, meaning users of Leica’s Aibot aerial capture platform can now fully automate the creation of digital landscape models from sensor data, according to Craig Robertson, construction segment manager for Australia and New Zealand.
“The potential of this type of technology is unlimited for applications such as corridor design and presentation of construction data from different sensors to present to all stakeholders,” he said.
“Some of the current advancements we are seeing in the 4th Industrial revolution are aiding in the speed, accuracy, quantity and quality of data available. Cloud computing is helping UAV reduce process times, cloud hosting is enabling collaboration on data sets and share of data between the field and the office.”
Greater data flows
Indeed, vastly increased data transmission between job sites, stakeholders and models is a core driver for many of the individual innovations and new feature sets of BIM and 3D tools currently being developed. More numerous data sources are one dimension, as the number and capability of sensors – UAV-borne and terrestrial laser scanners, for example – rapidly grows.
“UAV-sourced point clouds are changing the way projects are run from start to end,” said Robertson.
“The existing conditions of a site can be validated in an efficient timeframe far more accurately than in the past. This leads to better value engineering and design changes taking place before the commencement of construction. In turn, having more complete and accurate information earlier in the construction process leads to more competitive project bidding as some risk can be mitigated before a project starts,” he said.
There are many more dimensions for this increasing transmission and interoperability, however. The levels of the BIM maturity model essentially refer to the richness of data contained in the model itself, which by extension allows the potential for sharing those dimensions to all stakeholders with access to the model.
As hardware and software iterate on, the capabilities for sharing and processing data increase further. Total stations with 4G modems can transmit data straight to the office, and IFC files can now be opened and processed by field instruments, allowing manipulation of 3D models and complex data objects with handheld devices onsite.
“For both the building and heavy construction industry there is a greater need to be more digital and to have direct lines of communication for data flow between the office and field and back again,” Robertson said.
“Often on construction projects, there is a data block of getting the right design to field crews at the right time, either by slow communication or the silos of traditional workflows of having to download data in the office and transfer to the instruments by USB.”
The digitisation of field instruments now extends to heavy plant. Machine control and automation of excavation machinery isn’t new in the resources and heavy construction industries, but regular use of these systems now extends well beyond tier one contractors and early adopters. Becoming more commonly seen on major construction sites, these capabilities now integrate with off-the-shelf modelling suites to facilitate autonomous actions directly from the model.
“Products such as Leica ConX allow designs to be downloaded directly from the cloud to equipment in the field without the requirement of manual survey set out. Data can then be sent from the field back to the office, reporting production and as-built information back to project managers and the design team allow for close to real-time collaboration,” said Robertson.
Digital Twins: from physical to digital
At the nexus of these nodes sits the concept of the digital twin – that virtual doppelganger of the project and the end result that gets more detailed, closer to the real thing as the technology moves forward.
Phil Christensen of Bentley Systems, the foremost champion of the digital twin, holds that there is no completely agreed-upon definition of a digital twin, it’s a concept that is still evolving – but one that won’t be disappearing any time soon.
The term has certainly entered the lexicon. According to a new study from Grand View Research, Inc., the worldwide ‘digital twin’ industry size is anticipated to achieve $US26.07b by 2025, across economic sectors. Another report by Global Market Insights puts the figure at $USD 20b by 2025 – not chicken feed, in either case.
Bentley has invested heavily in the concept, looking to make seeking to facilitate the creation of digital replicas as close as possible to any fully functional asset. Last year, the company unveiled PlantSight digital twin cloud services, which was developed in collaboration with Siemens. PlantSight allows both physical reality and engineering information to be synchronised with as-operated and up-to-date digital twins, providing a holistic digital background for continuously understanding digital parts across different information sets for any working plant.
It then moved on to introducing mixed reality apps for infrastructure construction projects using Microsoft HoloLens 2. SYNCHRO XR is Bentley’s new app for immersive visualisation of digital 4D building twins, taking advantage of the considerably more capable new HoloLens headset.
A notable recent Bentley acquisition is Keynetix, a UK-headquartered provider of cloud-based tools for modelling and visualising geotechnical data, enhancing their subsurface modelling capabilities for their ever-expanding digital twin services for infrastructure projects and assets.
Illuminating the dark
Christensen posits that these added dimensions, the increased depth and richness of the model have completely altered its role in the project, representing the essence of the benefits of BIM – the medium has become the message.
“So the really the fundamental change with the digital twin is that the model and the data is the means of communication rather than drawing is the means of communication,” he said.
“Having data in a BIM model, but then creating a PDF to create drawings for construction can potentially throw away 90 percent of that data in the model. This data is valuable data, which isn’t finding its way downstream to the people who can actually consume it.”
This is essentially the definition of a current buzzword in the BIM and 3D modelling space – dark data. This refers to data created in one phase of a construction project, but is inaccessible to the other phases of the project that can benefit from it. It could be data created from one discipline which other disciplines can’t benefit, or other forms of siloing that can segregate project actors.
The Bentley vision of digital twins attacks this problem by combining many data layers and distinct models that to form a very powerful dataset. A BIM model is just the beginning – apply analytics to it to predict structural and thermal performance, for example, combined with subsurface data and surrounding reality modelling data – and you now have an object that can be queried for business intelligence.
“The endgame is a very high fidelity digital model of the physical asset – it has the data, the algorithms and the connections to the real world to predict the future performance of the asset, and the performance of the construction projects to create the asset,” said Christensen.
“The most important thing is that it’s not static. The digital twin continues to evolve through the engineering phase, through the construction phase and through the life of the asset as it’s modified,” said Christensen.