ClearEdge3D

ClearEdge3DClearEdge3D's Verity software compares as-built structures to CAD models and BIM to identify discrepancies and improve project quality.

By Tim O'Connor, Senior Editor at Knighthouse Media

Every builder knows that what's shown on a blueprint or in a 3-D rendering never exactly matches up with the end result. Something as small as a misaligned bolt could alter the placement of a support beam in a way that impacts how the structure fits together and its overall strength. Understanding those imperfections and having an accurate picture of the real-world building requires a way to map an as-built model of the structure.

Laser scanners are the first step in building a precise 3-D model of an environment. A high-end terrestrial laser scanner digitally captures the shape of physical objects in a 270-degree view, capturing up to one millions points of data each second. Those points are then merged into a point cloud made of billions of dots to create a precise 3-D model of the existing conditions, in this case a building. Multiple scans can be conducted to create a complete point cloud map of the structure.Clearedge3D info box

However, the resulting point cloud doesn't contain what each grouping of points represents. It doesn't know the differences among a structural beam, water pipe or window. Contractors must manually identify those structures within the model, a laborious and time-consuming process that reduces the usability and value of reality capture for construction or operation.

The founders of ClearEdge3D realized that if they could train a computer to recognize building components in a point cloud the way a human does then they could displace a lot of the manual labor and time needed to produce an as-built model. The company began developing algorithms to solve that problem in 2006 and it launched its first software product, EdgeWise, in 2011.

EdgeWise uses object and pattern recognition technology to automatically search for shapes in the point cloud that match standard objects; such as piping, structure, windows, walls and ducts. Items with non-standard shapes, such as chairs and toilets, are not currently identified by the system.

With a comprehensive scan, the software can usually identify 90 percent of pipes in the point cloud. That means that on a project with 1,000 pipes, the software will automatically extract 900 of them into an intelligent model that can be opened using CAD software such as AutoCAD or Revit. A human then checks the software's work, removes false positives and can add any pipes that were missed.

Even though EdgeWise doesn't capture every component within the building, clients have seen a 50 to 90 percent reduction in the hours spent creating the as-built model. "It's kind of a no-brainer if they have existing facilities they need to document," Vice President of Product Management Kelly Cone says. "It's the only tool right now that does that kind of bulk automated extraction."

EdgeWise's efficiency in identifying piping and structural elements made it particularly suitable for the plant and process space, where it has been used in building petroleum and life sciences facilities.

Building upon the success of EdgeWise, the ClearEdge3D team then turned its attention to another major problem in the construction industry, the quality and consistency of the installation of work. The new targets were contractors and trades who needed better ways to validate the installed work against their design, coordination and fabrication models.

Solving Verification

Before joining the company in late 2016, Cone realized the value of EdgeWise as a ClearEdge3D customer. He had used the software while working at a large integrated design and construction firm and became impressed with its accuracy and time-saving capabilities. He began to think of other ways to leverage its capabilities, and during a dinner with ClearEdge3D representatives pitched them on developing a system that would compare the point cloud to the design or coordination model to identify quality control issues.

Rework is a major concern on most construction jobs, eating into profits and causing setbacks that can affect the overall construction timeline. Studies estimate that 5 to 12 percent of a construction project's cost is spent on rework; including material costs, schedule delays and downstream clashes, totaling as much as $450 billion in waste annually. "Even if we just nibble off the top of that, it's a significant amount of money," Cone says.

Ultimately, the cost of rework stems from poor tools for quality control in the field. Cone understood the potential to use ClearEdge's automated modeling algorithms to improve that quality control piece by ensuring the as-built structures lined up with design plans. Other customers soon began suggesting the same thing and ClearEdge went to work on developing its next product, Verity. Cone joined the company about a year ago to help develop and test the software.

Verity takes some of the concepts behind EdgeWise and repurposes them to focus on the general contractor market. Where EdgeWise functions as a computer aided as-built modeling tool, Verity is a a comparison tool focused on quality assurance. It is a deep dive to solve the core problems of quality in the construction industry. "The biggest change was moving from an analyze-everything[process with EdgeWise] to letting the user choose what they want to analyze," Cone says. This workflow change aligns Verity with other construction industry processes like clash-detection.

Verity works with any scope of work on a project, from mechanical, electrical and plumbing (MEP) systems to the concrete, walls and finishes. The software can detect installation status and inform project stakeholders whether a component is in place, missing or obstructed from the scan. Knowing the status of that piece helps contractors know whether that part of the project is on schedule. Clients can also set the tolerance level for the work they are analyzing, and the software will compare the as-built point cloud against the CAD model, then automatically flag components that exceed those tolerance limits.

The software is a companion application to Navisworks, a 3-D design review package. This integration allows users to update the Navisworks elements to their as-built positions, clash-detect the position of the as-built model against the design or fabrication model and visualize the out-of-tolerance work using a color-coded classification scheme.

Unlike other comparison tools that offer heat maps and measurements between the model and the point cloud, Verity fits solid model elements to the as-built point cloud, allowing it to make solid-to-solid model comparisons and provide more accurate and detailed variance data. The software then generates reports to stakeholders to communicate problems with the building. The system is not designed to replace quality assurance professionals and inspectors, but rather to dramatically improve their efficiency. Says Cone: "Our goal was to give quality professionals a tool to let them check 100 percent of the installed work in the same time it now takes them to spot check 5 to 10 percent."

Utilizing scanning as the basis for the comparison also provides tens of thousands of data points on any given element being tested. Compare this to measuring only a few spots on the bottom of a beam using traditional methods. The completeness and accuracy of the data empowers clients to make better decisions on how to solve quality issues.

Quality Control

Because it can examine the entire structure, Verity solves a fundamental problem with how the construction industry approaches quality inspections: the spot check methodology. Spot checking randomly tests small sets of work to gauge whether there are more widespread quality issues. It's an effective methodology in manufacturing facilities and on assembly lines where the consistency of production means that a small sample of a batch is likely to reveal any present faults.

Employees take care to manufacture every part correctly because they can never know when the component they work on will be the one that is tested. "Spot checking works when it's random because you can't find ways to beat the system," Cone says.

But on a construction job, Cone argues that spot check sample is rarely ever random. Supervisors know what problems to look for based on past job experience. If someone had a blow-up on a curtain wall during their last job, they are more likely to check the curtain wall embeds on the next one. Workers learn quickly what the superintendent in charge of their area is worried about and what parts are most likely to receive inspection attention. Those items are likely to meet standards, but other areas can be easily overlooked and quality suffers as a result.

With more than 100 clients using Verity to aid inspections, ClearEdge3D has seen a clear pattern in those quality issues. "What we find are the biggest mistakes are where no one is looking," Cone says.

He recalls one customer that was beta-testing Verity to check for installation issues on the perimeter of a project. The software found some issues with the perimeter that Cone believes the contractor would have likely identified through the regular inspection process, but it also discovered a major issue on the interior of the building. The project's design called for a slope in the middle of the building's footprint to facilitate drainage, but the steel and metal deck was fabricated and installed flat.

Verity picked up on the deviation from the expected design and the quality team alerted the project stakeholders. Fortunately, the concrete had not yet been poured so the contractor was able to fix the problem within a day and stay on schedule. "If the slab had already been poured, that's a $50,000 mistake, and there's no way they were going to catch it early because they weren't checking there," Cone notes.

"We short-circuit that whole loop because we check everything and do it very quickly. We expose errors in those areas where no one is looking," he continues. Verity identifies those mistakes before they become costly problems; most customers have gotten a 100 percent return on investment from the software after only one job. "It's really neat to see these kinds of problems that I lived and pulled my hair out over now get caught early enough to do something about," Cone adds.

Speed is the other reason contractors are jumping onto Verity. Like EdgeWise, Verity's automation capabilities save substantial time in the construction verification process. One customer used to dedicate one person and eight man-hours to validate penetrations were installed correctly for a single 40,000-square-foot concrete slab. "Running our software, they got that down to less than an hour for the same task," Cone says. "Our software can check all of that work in a fraction of the time it takes to check it all manually."

"Laser Larry" Kleinkemper, CTO of Lanmar Services, which models historic buildings for architects, engineers and contractors that are managing renovation, historic preservation or addition projects, plans to utilize Verity in upcoming jobs to improve quality control checks. "Verity will make sure my staff are accurately modeling and matching all the building elements that are there and finding the items they haven't modeled yet," he says.

In previewing the software, Laser Larry was impressed with its ability to verify modeling location. "We haven't really had something like this before," he says. "It's about giving our clients the most accurate models that are the most complete and making sure our work is the highest quality possible."

Anton Dy Buncio, COO at VIATechnik, a company that creates as-built models and virtual reality simulations for construction projects, says Verity has been helpful in retrofits and renovations. "When you're trying to build something around existing conditions, that's when the as-built matters," he explains. If a hospital is planning an addition or equipment installation, for example, having an accurate as-built model can help it design the project around real conditions to avoid disturbing critical life support systems.

Using Verity enables VIATechnik's team to better use their time while the software validates the model against the point cloud. "Instead of having multiple people wok on it, we can have Verity check while working on other things, then we'll look at the report," Dy Buncio says.

Increasing Accessibility

The launch of Verity signals the next leap for ClearEdge3D, but the company is also expanding by acquiring complementary software systems. Earlier this year, ClearEdge3D purchased Pericept, the Texas construction verification software firm. Pericept is the developer behind BIMtrace, a product that captures recently installed scopes-of-work on a construction site and compares the as-built work to the design model to check for accuracy and proper installation. BIMtrace shares similarities with Verity; however, instead of point clouds derived from laser scans, it uses iPhone or iPad images to compare the real-world work with the CAD model.

Technologies such as BIMtrace are lowering the barrier to entry to ClearEdge3D's software solutions. In the past, the high cost of laser scanning meant that general contractors typically had to outsource scans to survey companies that specialize in providing measurements to the architecture, engineering and construction (AEC) industry.

That has begun to change as costs have come down. Even as recently as a year ago, a high quality terrestrial laser scanner would sell for $100,000. The same-quality equipment now goes for $16,000, opening the door for general contractors to more easily invest in their own laser scanners at the jobsite level. "That certainly makes scanning much more accessible, which is good news for us," Cone says.

The types of scanners that can be used to populate the point cloud are also expanding. ClearEdge3D has partnered with GeoSLAM, a leader in 3-D mobile mapping technology, to enable Verity to work with as-built construction data measured using GeoSLAM's hand-held laser scanners. "What might take you eight to 10 hours to capture with a terrestrial laser scanner might take you an hour with a mobile scanner," Cone explains.

Aside from more affordable laser scanners, Cone believes the biggest impact on automated verification software will come from machine learning. When the original computer vision algorithms used to create EdgeWise were developed in the late 2000s, machine learning was not yet relevant to desktop computation.

However, advancements in hardware and machine learning mean there is potential to use it to train the software to identify building components and improve automated extraction, or to more accurately identify whether the installed work is the correct type. "We've put all the feedback infrastructure in so we can really leverage machine learning down the road," Cone says.

Machine learning could be the key to increasing Verity's accuracy from 90 percent to 95 percent, or 99 pecent. "It turns out that it is actually pretty hard to teach a computer to see things in point cloud data the way a human can, so it's really fun to see some of the early applications of machine learning in the AEC space and it's something we want to deploy in the near future," Cone adds.

Current Issue

Check out our latest Edition!

 

alan blog ct

Contact Us

Construction Today Magazine
150 N. Michigan Ave., Suite 900
Chicago, IL 60601

  312.676.1100
  312.676.1101

Click here for a full list of contacts.

Latest Edition

Spread The Love

Back To Top