Trial PointFuse

Case Study

DPR Construction deliver faster project collaboration & coordination.

2021 | 5 min read

a 3d model of a room showing coloured elements such as pipework

Merging existing 3D models with reality capture tools can be an inefficient process. DPR Construction often faced having to recreate models to ensure they had accurate representation of the existing conditions of the site to use as backgrounds for the next phase of coordination. DPR deployed PointFuse’s meshing technology to accelerate production of 3D models during ongoing projects. Models were rendered in days not weeks, with estimated cost savings of over 80% compared to conventional scan-to-BIM techniques.

Background

DPR Construction is a self-performing general contractor and construction management firm based in Redwood City, California, with offices across the US, in Europe and in Asia. It particularly specializes in technically complex and sustainable projects for the life sciences, advanced technology, healthcare, higher education, and commercial markets.

The company has invested in technologies to help it deliver better buildings more quickly and cost-effectively, with increasingly routine use of Virtual Design & Construction (VDC), on-site data capture and as-built visualization tools.

a labelled 3d model of a ceiling showing where proposed pipework and ventilation could be

Labs + office coordination incorporating PointFuse mesh

The challenge

DPR Construction professionals frequently work on both new-build and improvement projects with high levels of complexity. Particularly in refurbishment works, it identified that laser-scanning of facilities could be used for reality capture, providing millimetre-accurate data about existing structures and equipment. The point cloud outputs from such laser scan could then be used to create and coordinate 3D design work, helping to identify and avoid ‘clashes’ between new additions and existing elements.

However, DPR’s Virtual Design and Construction (VDC) manager Shane Raber, based in San Francisco, California, found conventional industry workflows to convert point clouds into useable design information often proved expensive and inefficient.

  • The high accuracy of laser scans come with a digital burden: every image is formed from millions of data points, with individual files measured in gigabytes, requiring computers with high-powered image processing capabilities to make use of them.
  • Clash detection using point clouds in conventional BIM software applications also tended to identify 1,000s of issues – many of them false results arising from the scanning process. These needed to be resolved before any meaningful insights could be gleaned.
  • Converting even a ‘cleaned up’ point cloud to a design model was expensive, required a skilled professional, and delivered an output that was often over-detailed and which could not easily be classified and managed by different design disciplines (architectural, structural, MEP, for example).
  • Moreover, any ‘scan-to-BIM’ conversion often took days to finalize, limiting the value in supporting daily decision-making during construction delivery.
a 3d model of an empty building showing coloured pipes and support columns

Mesh used for clash on exiting C&S relocations

The solution

In early 2020, Varun Raja, VDC Engineer for DPR searched online for a software application which could be used to make as-built point cloud data more manageable. He identified PointFuse’s ‘meshing’ technology as a potential solution and tried out the software during a two-week free trial.

Individual laser-scan files can be anything from 2GB to tens of gigabytes, and multiple files are typically combined to produce accurate 3D imagery, multiplying the data challenge. As a consequence, a single surface in a point cloud might be represented by 100s of thousands of data points. Meshing enables much more efficient representation. By capturing surface edges and combining data points from across each surface, a point cloud can be converted into a mesh of geometric shapes that can be more quickly assembled, viewed and explored. PointFuse can reduce files to 1/100th of their original point-cloud size – measured in MB rather than GB – so that they can be viewed and shared via other applications.

At DPR, Raber piloted PointFuse during construction works relating to a $200m laboratory refurbishment project. Without training in the application’s use, he quickly learned how to categorize objects by project discipline, creating ‘layering’ conventions that separated structural elements from MEP elements, for example, so that clashes between different aspects of the design could be quickly identified.

Floors, walls and other structural and architectural elements could also be easily tagged as single continuous elements, helping reduce the number of coordination issues from tens of 1,000s to 100s. DPR staff experimented with classifying elements to detailed levels – breaking down structures into components, for example – but with hindsight they found it easier to keep classification at more general levels.

PointFuse also helped in quickly clashing proposed designs. For instance, an existing pipe was captured as a single discrete element and moved to a new position; clash detection could then be re-run to test the design with the pipe in its new location.

The potential time and cost savings from using PointFuse were substantial, Raber said.

“We estimated that, in this case, modelling the original point cloud data to produce a detailed building information model would have taken two professionals two weeks at a cost of $54,000. Deploying PointFuse to produce a mesh-based model took one professional two days, at a cost of $9,400.”

Additionally, the rapid turnaround of outputs meant DPR could take advantage of regular scanning and always have up-to-date meshes of as-built work, helping feed into ongoing design work. And as many architects and GCs share their BIM work in cloud-hosted platforms, the manageable file sizes were also a benefit, said Raber:

“We share information on projects where everything is live and collaborative, but gigantic point cloud files prevented us from doing that. PointFuse Mesh offers a way to keep us up and running in BIM360 cloud so we can include it in our BIM workflows.”

Following the pilot project, PointFuse has become a part of DPR’s software portfolio and workflows, available to its project teams wherever they work. From initial use by a handful of self-trained users on a single project, DPR has expanded routine use of PointFuse to a pool of 17 users, ten of whom are now regarded as very competent, with use extending to other US DPR offices, to work in Europe, and to DPR’s partner firm in India. By late 2021 it was being used to generate time and cost savings on three projects at any one time. Clients and subcontractors have welcomed DPR’s efficient management of design, construction and installation processes, which have also made DPR more competitive.

Benefits

  • Free use during two-week trial
  • Fast adoption – PointFuse is simple-to-use, requiring little or no training
  • Faster turnaround of point-clouds into 3D models – outputs in days not weeks
  • Time and manpower savings – 3D model outputs achieved in 10% of time
  • Cost savings of over 80% vs. baseline in pilot
  • Faster production and sharing of up-to-date as-built information resulting in less rework
  • Improved information sharing via cloud-based collaboration platforms
  • Competitive advantage in managing project design and construction workflows.