Description
Principaux enseignements
- Gain modeling tips for quality assurance
- Learn about Revit and ReCap visualization tools and methods
- Learn about QC Revit templates
- Learn best practices and common mistakes in Scan to BIM
Intervenants
- Maria Victoria OrtegaI´m an architect since 2011 and I have been working in the BIM world for the last 10 years. I started as a BIM modeler and from then, I have developed in different positions inside the company. Currently, I'm the Director of the Architecture & Structure Business Unit, leading several technical and also adaptable teams. It´s great to feel part of this industry, in constant growing and moving
- AMAlfred MortTed is the Vice President of Eco3d (eco3dusa.com) where he provides direct oversight for the full scope of laser scanning and modeling on all projects. He also sits on the Board of Directors for the U.S. Institute for Building Documentation and on the Board of Directors of the United Surveyors of Arizona.In addition to an expertise in laser scanning, Ted leans on a deep professional history of practicing applied measurement on a variety of construction projects while operating in the fields of Land Survey, Mechanical and Electrical Engineering. The wide spectrum of projects he’s worked on range from Bridges, Dams and Massive Civil Infrastructure to High-Rises, Hospitals and Industrial Complexes.His commitment to the healthy growth of our industry is highlighted in his volunteer work through industry nonprofits and as a lecturer at Arizona State University and Grand Canyon University.At ASU and GCU he exposes students in engineering, architecture and construction management to new technologies while detailing the impact they have in the current and future world of A/E/C.
VICTORIA ORTEGA: OK. Good afternoon everybody. And thank you all for joining us today. The funny accent is because I'm from Argentina.
My name is Victoria Ortega. I am an architect. And also I've been project manager for AEC Resource, which is a company that develops quality beam projects. And we have clients all around the world. Today, I have a huge co-speaker. Ted, can you answer this?
TED MORT: Huge? My goodness.
VICTORIA ORTEGA: Yeah.
TED MORT: Well, I'll tell you what, that's quite the setup. My funny accent is from Arizona. I'm the Vice President of Eco 3D. We are a big service provider, one of the biggest in the country here. We do a lot of laser scanning, scan-to-BIM conversion, and then some BIM services lightly. So our focus is on getting a lot of quantity scanned to BIM completed with a high quality.
VICTORIA ORTEGA: So, what are we going to see today here? Hopefully some tips and tricks for us to be able to take those concepts, such as QA and QC. And those concepts are going to help us out to get through the entire workflow on a scan-to-BIM project.
First, some tips for the quality assurance. To talk about this, I would like to start by showing you the typical phases that a scan-to-BIM project has. We have the capture phase, which is the reality capture through some kind of device of our election. Then this will give us some raw information that we have to be processing for us to be able to have at the end of the process a point cloud file.
That point cloud file is the one that we should be inserting in some kind of BIM platform. We're going to talk about Revit today. So the idea here is for you to replicate those information that we have from the point cloud files and convert it into intelligent elements that Revit can read. Finally, we have the quality control phase, which is the correction and the validation of this model for us to be able to guarantee its accuracy.
Let's take a look now how the QA affects each one of these phases. And to talk about QA over the capture, I would like to show you a project that we have done this year at the office. This is a low-income settlement over Buenos Aires. It's called [SPEAKING SPANISH], like a big favela, equivalent to a Brazilian favelas over Buenos Aires.
And here, the government had a renovation plan for it. So they committed to us to do the survey for the facade of all the area. This was 50 miles linear, equivalent for 280 football fields for you to have an idea. So at the time, the operator developed a global planning for the processing and the capture of this entire area. And when he finished the work and came back to the office, he was a reference for the [INAUDIBLE] team when we had issues between point clouds that were noisy, or when they got overlapped and we had missing some elements.
Some time after that, the client ask us for us to do the same with the roof. So at that point, we use a drone that flew over the area and took that information as photogrammetry, which also can be converted after into point cloud files. So this kind of global plannings at the beginning of the project is what you have to have in mind for decide the scope of work, the device election, and to have really qualified operators.
Following with this, the QA of the process phase, we should talk here about the hardware and software, and for you to know that this operation cannot be performed in any kind of computer, or device, or software. You can delegate this task to a specialist. But any of the ways that you prefer to do that, take into consideration the times for the processing because it can take a while. So as soon as the person in charge gets into contact with the wrong information, he can run some quick analysis and confirm with you or not the times that you have previously estimated for this task.
Last but not least, the information access. What does this mean? If you have a scanner, your scanning a site in another country, in another city, how long it's going to be for that information to get into your hands?
Take into consideration the size of the files for download and upload from the platforms. Is it going to be one big file? Is it going to be several smaller files?
Or maybe they are sending the files to you through a hard drive. So in that case, you should take into consideration the post office times. It can sound silly, but it has happened before that we had three, four delays because of the post office. And that is going to impact in all the planning for your project.
Regarding the modeling, we should talk about here-- well, we recommend here for you to advise your client for you to model a representative area of your project and send it over for some validation. He can come back with comments, with some markups, or just a confirmation that you are on the right track and continue to modeling.
So at this point, the communication with your client is key. And try to figure it out, besides the technical systems, and the LOD, and the scope of work, what is the client expecting by the end of the process? So at that point, we need expertise people to be able to communicate with the client and also expertise on the technical area for the modelers. Scan-to-BIM projects can be a little bit tricky if you're a newbie. But in the best-case scenario, your modeler should have a large background of software or tools and maybe some kind of training on interpretation of these kind of files.
Finally, the QA over the quality control phase. We recommend you here for you to have at your office some kind of checklist and templates, like standards. And according to your projects, and to your needs, and a lot of the parameters that we have seen before, you can customize that so that way you don't forget anything.
And we're going to talk about the software a little bit later. But today, we have in the market several options to do the quality control phase a little bit lighter for us. And this is going to help us to win some time and resource. So the key here is for you to try to figure it out what's going to fit best for your own project. Ted.
TED MORT: Thank you.
VICTORIA ORTEGA: You're welcome.
TED MORT: So we'll get started talking about the capture process and some steps that we can take during that to help ensure that the quality's going to be maintained through the rest of the duration of scan-to-BIM. So the first question I have for really the group is, what makes us different? And it's a holistic us. It's everyone in here that's going to be taking part in some sort of a capture-to-BIM conversion.
I think that we're different because we're going to be accountable for the quality. A lot of folks say we preach high quality. We live by high quality. But we are really the only true representation of a group that is going to be consumed and requires that quality to be prevalent throughout the process because if it's not, we're going to get burned by it, guaranteed. So I think that's a little bit unique. And it means that we really have to be thoughtful about what we're going to do when we're heading out to these sites.
So the first part of the evaluation is before we even get mobilized, we start looking at equipment selection. Now, when we talk about equipment selection internally, we're not a one-brand shop. So we don't just get behind one particular manufacturer and try to make that tool fit everywhere.
We consider laser scanners like hammers. So there's a lot of different types of hammers out there. Each one of them has a very unique purpose that it's specifically designed for. It doesn't mean that it's a bad hammer if you can't take the sledge up to the roof and start putting nails in. It just means that it's a wrong tool for the job. So we spent a lot of time understanding the different components of the hardware, what it really excels at, where there might be limitations. And then we go through and select a specific unit depending on that project itself.
So it's difficult to do if you have capital restraints, and you're looking to make a purchase. We actually rely on vendors. So we do own a lot of scanners internally. But if we're looking at a site that we may not have that opportune piece of equipment, we'll go through a rental company and actually acquire that piece of hardware so that we can execute it properly.
Another thing that we look at before we even get moving is we like to take that hardware out and perform a scan with it before we pack it up and send it to the site. It doesn't happen very often, but on occasion, we get that call. Somebody didn't follow this step, went out to the site, and found out that maybe it got bumped on the way back in or while it was in storage. And now it's out of calibration, or the hemispheres aren't looking right. And worst-case scenario, the entire project gets performed before you find this out. And you have to go do it again.
So we always pre-scan. We check that to make sure that it's still within calibration. And then we actually store that scan again.
Because there are ways that you can go back and forensically correct scans that have gone wrong based on that original one. So maybe in transport, it falls out of calibration. If we have that original scan, and then we see what's happening, in some cases we can work together with the manufacturer to provide some sort of a calibration transformation.
The kit, right here in the middle is a little shot of something that we put together. Too many times we checked our baggage, and we looked out the window to see what's happening over here to the side. They're kicking it off the side of the plane or knocking it all over. And this is really high-calibrated equipment. And it's very sensitive.
So we said, OK, we're never checking a bag-- or we're never checking a bag again. We're going to carry it on every time. So we went out and bought bags that are specifically designed for carry-on, right to the max. And then we've manufactured foam inserts that go in.
And it holds all the equipment that we need. So not only the scanner, but this also has-- maybe they need a DISTO, a digital camera. We have all the different pieces that they need here to be successful, including even an iPod for digital notes. And we'll get into that a little bit later.
So carry it on every time I got to implore you to do that. Don't let it out of your possession.
When we start talking about a strategy to attack the site, so some components that go into the strategizing are going to be methodology primarily. So this is acquisition methodology. You've probably heard a lot about the-- the BLK360's out there. And it's running off of a cloud-to-cloud type registration process. You've got different types of targets. You've got all different methods of going through and acquiring information.
You can typically tell before you even get to a project what type of methodology or what combination of methods you're going to need out there to be successful. So we'll sit down with the crew before they leave and talk through what we're going to do when we get out there to the site. So they go out with a game plan in mind. And already we're expecting something from them that is going to be reflective of what we've determined to be the best method to deliver high quality.
And the last thing, anyone not familiar with the USIBD? Great, everybody's familiar-- just kidding. This is a fantastic organization, all right? The USIBD has an LOA guideline out there. It's a great certification.
You want to talk about predictive quality, one of the biggest things that remains amiss is an easy way to articulate what kind of accuracy we're going to expect, not only from the scans, but also in that scan-to-model conversion. The LOA way is the only thing out there that I know of that's designed to do that. And it's actually designed also to fit directly with the AGC BIM form LOD specification. So this is a great cert to get.
Now, when we get on site-- so we've got our equipment packed up. We know what we're going to use. We know the method that we're going to approach the project. We first have our technician go through a site walk every single time.
So, he's going to call back to the office. Hey, I got here safe. All the gear is here.
And then he's going to go on that walk. And what he's doing on that walk is he's watching for things that might prohibit him from following through with the methodology that we decided we were going to deploy. So I'm not sure if you all have been on any sites that were different than what it was described to you as?
AUDIENCE: No, never
TED MORT: Not really, right? Pretty close. So, what this does is it gives us a great opportunity to put a flag in the ground and say, one, are the site conditions is going to compromise our ability to meet the requirements of the contract? That's the first and foremost. Because if that's the case, we need to get in touch with the client and start working through it.
Now assuming that they go through the site plan-- or the site walk, everything's great, or maybe there's a few minor things that we need to walk through, then they get started. So we've been communicating. They say, OK, this looks good. We're going to get going. They are going to then keep in mind where they're going to move for this methodology.
So when we're on a site, there's a lot of phasing going on. There's a lot of areas that might be busy in the morning or in the afternoon. We try to keep that in mind.
We try to work with the people that are there. Because the best thing you can do when you get on that site is explain to folks that this is an activity that is in support of everybody that's doing work there to install. So what we don't want to do is we don't want to get in the way. We don't want to say, hey, you've got to get the framers out of here so we can come through and scan. We want to understand what that schedule is so that we can adapt our plan to it.
The next thing, take notes. If you send five people out, and you leave it up to them, two of them are going to come back without notes, and three of them will come back with something that you couldn't put together to equal the notes. So what we try to do is we try to give them some sort of a template or format so that when the data comes back, the people that are processing it understand what they're looking at. And it's more predictive.
So this is a little tip. Evacuation plans are amazing when it comes to identifying a floor plan to start your notes with. We rarely get floor plans or some sort of existing documentation. A lot of that's the reason we're out there on-site because we're going out there to create it. But almost every one of them has something that represents at least that floor plan so your guys aren't out taking the time to sketch out a footprint and then try to fill it in.
So the iPads that you saw earlier, we try to digitize all the notes. So we have apps on those iPads that we've imported our templates too. So they'll go through and leave their notes digitally. So you don't not to worry about handwriting.
They can take photos like this and impose them right into the notes. So they can start notating on top. In fact, if you take a look at this plan right here, there's some writing up here in green and red, and then some notes on top. This is actually field notes from that site when they went out there. So they brought the image in and started taking notes on top of it.
The last thing that we want to look at here on-site is we want to go through the process of a couple of good best practices. One of those is going to be closing loops. So what I mean by closing loops is starting your plan out with the idea of making small circles that overlap each other, almost like the links of a chain.
Because when you're working with laser scanning, and you're moving in more of a traverse methodology, one in front of the next, what you're doing is you're giving yourself an opportunity to start building up error. So this error starts to accumulate through these. It may only be a couple millimeters. But by the time you move 100 scan placements, that couple millimeters turns into hundreds. So now we have a huge issue.
If we're closing the loops, the way that this technology in the software is built is that it's helping to balance that when it sees two commonalities in the scan. So if you go through and you create that individual loop, maybe it's a 10-scan loop, it's balancing that error and lessening the impact of all that accumulation. And then you're getting overlapping with these chains with the chain links that then, again, also work to help balance that error out.
Blind checks is a great one that we use. We will go out and leave a target up and not include that target in the registration. So we perform our entire registration, and then we go back through and we extract that target coordinate.
We do that because in a lot of cases, we have a surveyor go out with us and actually collect that data. So that becomes something that they keep, we keep it. And at the end of the entire project, we try to put them together, and we say, OK, did we meet that accuracy standard based on this blind coordinate that wasn't used? That's also a component of the USIBD LOA document.
Though maintaining direction, another little tip here, when you are scanning, that scan itself is assuming 00. So it's got a bearing that it just thinks is zero every time. It's based on the physical component there. So if you're not paying attention, you're moving that around, what you're doing is you're creating deviation in every one of those scan placements. So now you're not only having to solve for where that scan goes, but you're also solving for the rotation.
You can take that one element almost completely out of it and at least get it very close if you instruct the people that are going out in the field to face that scanner in roughly the same orientation every time. So we'll send him out with a compass. And they sit there with that compass affixed, and they tried to just point it north.
At the very least, what that does is it brings it back in the office, and it satisfies another one of those degrees of freedom for us. So now we're looking at elevation. We're looking at horizontal in the x and y. But it really secures the other components of the six degrees.
So losing the looks, another one. Everybody loves RGB. I mean, it looks pretty. It's just an amazing, intuitive environment when you get it back. But my gosh, one of those cameras can take twice as long on-site.
So when I told you earlier we do a lot of scanning, we do thousands of scans every week. So if I cut my scan time in half, now I'm only getting half that amount of work done. That's a huge component when you start to scale.
If you're just going out to do three scans, and it's a supportive function of your business, that's great. But when it is your core business, we have to maximize efficiencies on site. So by cutting out that camera element, we are removing the photo, but we still have a great intensity value. And it allows us to get maybe three extra setups that help really secure visibility on a target area.
We also like to encourage targeting past the setup. So right here, you have an example of the closing loop procedure. This is actually a developed website that we have. It's like a Wiki for us internally.
This is available with all of our training material for our people. So they go out in the field, and let's say they haven't done a target pass or setup. They can look that up in this Wiki on their phone and get some of those best practices and methodologies down, so a really cool tool.
It's fairly easy to set up. If anybody's interested in it, let me know. I can let you know what we're using. It's off-the-shelf stuff-- great.
But getting back to the targeting past your setup, so what they do here is when you're setting up that scanner, you're looking ahead and saying, OK, where am I going next? And I'm going to put a couple of spheres because I need at least three targets if I'm using spheres. So I've got those out there.
I move to the next setup. And I scan those three spheres, and I look for it. I put three more down. But as soon as I move again, I pick the three that were behind me.
The problem here is that I'm not doing myself any favor with that accumulated error because all I can see is that scan that's right behind me and the one in front of me. If I leave some scans, or I target one or two places ahead, I might not use them necessarily for all the registration, but I'm going to get enough data on some of those targets that are really from three scam placements that are going to be occurring in the future. And it really helps secure those chains down the links. So by using some of these best practices, you can flush out almost all of this naturally-occurring systemic error in your registration process.
The last thing here for the guys on-site, we can't walk away from the scanner. Unfortunately, we've lost a lot of scanners. We do a lot of scanning, we lose a lot.
This one was a tough one, though, because if you guys can see it, that's a C10. And this was when C10s were brand new. So we're buying them for $150,000.
Had it out here. This is Camelback Mountain in Arizona. We were scanning the trail head. And we had the thing set up. The guy went back to get a picture of it. Goes bouldering back over to it. And when he got there, it wasn't there anymore.
And unfortunately, what had happened was he set the legs up-- and you see we're using metal legs with some brackets. My best guess, because there is nobody else out there, is that it started expanding a little bit. We had a little thermal expansion that didn't suit right. The leg gave way. And that scanner took a tumble.
It doesn't look too far there. It went down about 35 feet. You know what? Kudos to Leica, though because the thing they keep scanning. You couldn't use it after that, but, you know.
So we get the information back into the office. And now we're looking at the pre-registration. So like we talked about, getting that information back and take some time. But it's imperative that we don't jump right into registration because if we do, we're missing out on some critical steps.
So, how many surveyors? All right. I have some survey background. I've gone through some of the licensing process. So I'm not beating up on surveyors here.
But we're all human. And there's going to be some error in surveying. You're going to have that one point that maybe didn't make it in with the rest as far as accuracy.
Here's the problem with the software that's out there. If you're using survey control as your primary mode of registration, that software thinks that every surveyor's perfect. So it says, I am holding that survey control no matter what. And what it does is it balances the error across the rest of your scan registration.
So the problem happens when that one scan-- or that one survey target is maybe three inches off, something happened. Unfortunately, it got moved over. You're never going to see that because if there's enough scans in between that survey target and the next one, it balances all that. And it tells you that the registration's good.
So what we like to do is register everything together, take the survey control then, extract those same coordinates out of the scan, and then we do an analysis of that. So we take both coordinate sets, and we look through them. We look for varying oddities.
So if something is really sticking out, we're going to call the surveyor back and say, hey, here's the thing. We've looked through it on our end. We think there might be deviation in this point. They'll go back either confirm it or say, no, you need to look at your information again.
In a lot of cases, it got moved after the fact. We're in a lot of industrial spaces where we're affixing targets, temporary targets to the ground. It's not unheard of and somebody comes back through after we're done scanning, but before the surveyor gets there, or vice versa, and they peel one up because they're looking for some writing on the ground, and they go and put it back pretty close. So we always like to do an outside inspection before just trusting that survey information blindly because the software doesn't do you any favors there.
A scan inspection, when we get the scans back in, just like we did a scan before we sent it out, we like to do a scan again. We check it to make sure that it's still in calibration. And then every one of the laser scans that was performed, we open that up and we look at it.
When you open these scans up, you have sort of that PanoView. So you're in the bubble, and you're looking around. I like to tell the guys to find some sort of a line and to watch that line because if you see that line, there's a rhythm in it, like a heartbeat at some point, in some cases it's because maybe there was a water hammer next to where the scanner was located, and it gave it a little shake.
Now, the way that the hardware works, it's not going to catch that shake if it landed back to about the place it was before. Because what the scanner's doing is it thinks and it says, OK, this is where I'm at right now when I start. It does the scan, it comes back, and then does that little spin again, it's saying, am I in the same place level-wise as when I started? And if I am or I'm close enough, then we're good to go. But when you're in an industrial environment especially, you're going to see impacts. So we typically throw out about 1% of our scans that happen in industrial sites, even though all the lights were green and everything looked good.
We're going to go through clustering. Methodology used in the field is going to be similar to closing loops. Each one of those loops will be defined as maybe a cluster. Or maybe there's two components of a building, or maybe a room with done with one methodology and another was done with a separate. So all that needs to be conveyed in the notes.
And the people that are going to be registering this information should take that into consideration before they even get started. I know it seems like a no-brainer, but I can't tell you how many times people are excited just to start registering, and they don't take into consideration what was done out in the field. So if we can take some time, you split it up, you organize it into the right files, it'll save you a world of pain later.
We also maintain all of our raw scan data. So when it does come back into the field on that hard drive, the last thing you want to do is to start playing with it in the registration and have something corrupt. So we always keep the raw data, and then we go through, and we start processing it. If anything happens, we can go back to that raw information and start over.
Now, once that registration's complete, we move into what we're going to call the post-registration period. So a lot of the software that's out there for the registration is going to give you sort of an automated indicator as to whether or not it's accurate. The challenge being that, like we said before, it doesn't have the intelligence to see an outlier in a lot of cases and identify it properly. So we like to push a second set of eyes.
Now, we call it second set of eyes. It's really getting your buddy next to you to come over when you think you're done with the registration and take a look at it. When he does that, we typically turn each one of the scans a different color, cut a cross section, and then we start looking for deviation.
We don't look scan to scan because you're probably not going to see you there. But we look out to the fringes. So if I'm doing a rooftop, I might look at the building across the street. It's not part of my scope. It's not going to make it through the scan-to-BIM process. But if I see an angular error, that's going to be really dominating when we get across that street. So it becomes very apparent and allows us to make an adjustment or at least an assessment back here where you might not see it within the confines of 30 feet of that scan.
Setting alignment is an important thing to do before you turn it over. I know when you get started in the modeling process, the last thing you want to do is you find out that maybe it wasn't exactly ortho, or it wasn't lined up in Revit to the site coordinates. It's not easy to fix, Is it?
VICTORIA ORTEGA: No.
TED MORT: No, all right. Still not easy to fix. Anyway, it's a good thing to be cognizant of because you're really helping the people out downstream. Ready to take over?
VICTORIA ORTEGA: Yeah.
TED MORT: All right.
VICTORIA ORTEGA: Thank you. Before we jump into the modeling phase, let's take a look at some of the most common mistakes that we have detected so far. I believe that the most common mistake is to model from only one view. We have projects that we have checked, and we find out that the modelers went directly to the plan view.
So in this image, we have a point cloud. So they model the wall and assume it's done. But if we take a look at this wall in a 3D view or an elevation view, we will be able to see that it's not regular, and they did some adjustments at the top. So these kind of partial and regular checkings is the one that you should be doing during all the modeling phase.
Another thing that we have detected is the lack of knowledge regarding the constructed system. This happens often on the joints between elements like we can see here between beams and columns, or even though when two different disciplines get together, for example structure and architecture.
Finally, we have interpretation issues. Point cloud files can be tricky to read and to interpretate. So at this point, this kind of error, like missing elements or mismatches on the perimeter of the elements, are very common to find out.
Let's move now into the modeling process. And we will give you some tips to spend the processing-- the modeling process the best way easy-- the best possible way, sorry, and hopefully not get into the previous mistake that we have seen. So, we should be knowing by now the LOD and the final use of our project because it's not the same to be modeling for 2D implementation than for MEP coordination. So try to figure it out here, which are going to be the hardest elements to be modeled, and maybe if you should be regarding-- not modeling some of these categories or elements.
We can talk here also about the conversion tools. Anyone familiar with conversion tools for Revit? Only one? OK. Two. Great.
Conversion tools are add-ins that, through Revit, automatically detect elements in a point cloud and convert it into Revit elements that you can after import and adjust into your Revit model. So today, the market offers several brands and uses. So it's going to add to basically the LOD and the scope of work that you are using.
Here, we can talk with the client about the deviation tolerance, as mentioned before. And this image, we had this project for us to check that was already modeled. So when we found out this almost three inches deviation, we talked to the client about it.
But particularly in this case, he did not care because his scope of work for this project was the confirmation, the validation for some ADA sidewalks. But the building, he just needed some representative walls. So once again, the communication at the beginning of each of the phases with your clients, it's a real big key.
Let's move on to the insertion of the point cloud file into the Revit. Revit cannot accept them natively, so the native file. We have to index it first. We can through these with ReCap software. Which one of you use ReCap? Oh great, awesome.
So once we have our index file, we can insert it. If you want to do it through Revit directly, a pop-up sign will come out and ask for the confirmation that is going to be indexed. But anyway, once that, we do have our file inserted, we should check if we have to follow some coordinate system. This is regarding-- and most importantly, if we have several hundred point cloud files or maybe coordination models to link, because this is going to give us the accuracy, all the files are going to be aligned between each other. So manually movements, you should not going down there that way.
Once we have our point cloud file inserted, you must lock down with the pin tool. This kind of sound silly, but we have seen it out there. One of the users accidentally moves the point cloud and continue to modeling, losing all kind of accuracy. So this, besides to be a headache and a lot of time of rework, can be avoided with a single click from your mouse. So please, pin your files.
We should be moving on to some worksets. We recommend you here for you to have one workset for each point cloud file in order to be able to turn them on and off regarding the area that you are working with, or the user that is modeling that. You can have another workset for all the linked elements such as coordination models or maybe some CAD files for the background.
If you have used conversion tools to bring some elements into your model, those has to be in different workset also because you're going to find out very quickly and very fast and be able to adjust those elements to the point cloud by having those in a separate workset. And the rest of them, it's up to the project. You can have worksets by discipline, by users, maybe by floors if you are working on a tall building, or maybe by building if you're having a project for a large complex, for example.
Moving on with the modeling process, we should be setting up some levels. A cool way to set up several levels at a time is for you to pass through a section through a circle so that way you're going to be able to see all the floors at once. Pin the floors too, please, please please, please.
We should be setting some plan views according to the templates that you might have been using. A client can give you a template. Or maybe you can have one of your own office. Grids can be optional. You might be having this information from a CAD file or a PDF file.
And the following is to work according to the best modeling practices. The best recommendation here is for you to use multiple views in order to avoid the kind of mistakes that we have seen before. Avoid the generic models and the modeling place. So when you have found with that issued area or that issued elements that you cannot be able to recognize, you should be working with some visualization tools for you to have a background and a backup for all the things that you are watching there.
So at this point, we as ReCap. At the office, we work with two monitors, one with Revit and one with ReCap simultaneously. So at this point, ReCap allows us to create sections, makes scope boxes to define little areas, take measurements to back it up what we are trying to make into Revit.
So we have our model done. And now we have to deal with the quality control phase. Usually what we do is to throw some sections over both directions for the building and pass it over so that way we can find some errors that we are not able to see on RCPs or in floor plans, such as hides displacement of elements and that kind of error. So it's like a cool and a quick, fast way to do that.
Other thing simultaneously with these sections, we threw some checklists. Also we have standard checklists. And we adjust those according to the project needs for the checking.
For global, maybe view for some rough elements, you can use some templates and make your point cloud to look with a single contrasting color. And orbit around the 3D, you can find funny elements mishosted flying around like crazy things. We have found toilets flying around. So at that point, it's a real quick way to find out this role mistakes.
Also use of filters, we can see here how we've being filter beams by type with different colors. So at that point, you can have that information and check out the information such as the type with the point cloud localization of the elements.
For purge and clean, we should be opening our detach file so that way we can conserve our working file. And this one's going to act like the delivery file. So once we have our file detached, we can purge some working views, sections, all kinds of auxiliary elements that you have already done, like reference planes, detail lines. You name it, you have to erase all the working stuff.
And then we can go through a Navis walkthrough. Navis is more friendly. It's going to be a lighter file to work on. And here's where we can detect clashes between elements, for example, or duplication on some kind of elements.
TED MORT: I'll take that. So I think a lot of us use that very similar QA process. We're cutting through planes. We're no different. That's a primary function of our scan-to-BIM quality assurance is going through and cutting through planes.
Now, we consider that a challenge because the other thing that happens when we're doing this is I feel like we're never getting it all. So we're going through, we're catching the errors, and we're cutting these planes, and we're removing the model and the scan through. And we fix all the errors, and it goes through QC again. And oh my gosh, it comes back with more errors. What the heck?
It's not that they weren't fixed the first time. It's that there's so much information that starts occurring is that you end up missing some stuff. You become a little bit snow blind to it.
So we would find ourselves going through this process over, and over, and over again until finally one of my guys says, this is about it, man. I can't take it. We've got to do something differently.
And what we did to change this for us is we leaned more towards a tool that was being used in the medical field. So there's an open source platform out there-- everybody familiar with Cloud Compare? No? Yes? Well, if you're not you, should be. Because Cloud Compare is something that we used as a foundation to start building our own algorithms for quality assessment.
Now, this is just the interface. The way that it ends up looking now, it sits on top. Cloud Compare acts as our engine, and then we have different programming that happens in there to help make things a little bit smoother, run some custom algorithms to start checking things.
So what this does for us is it takes us out of this realm of looking at something like this and saying, what's wrong with it? Or where are we deviating? Because if I look at this right now with you all, and I say, how many things can we find that weren't modeled? We've got a model and a scan overlaid with each other here.
First thing, let's look for omissions. I can see a couple of things. But it's difficult to do.
After I've run it through the first process of what we're calling EcoVerify, EcoVerify I tell it, I want you to identify omissions and turn them all red. So what we've done here, everything in red, that's the point cloud without an associated model to it.
So how much easier is this to look at and say, OK, I've just identified all the omissions in the model. Some of them are OK. There's a forklift sitting here, some spools over there. That's all right. That stuff's going to move.
But then I look up, and I see, oh man, there's a bunch of piping. Maybe that's conduit running through here. Is that something that we omitted on purpose? Or do we need to go back and fix that? So it makes for a very quick, a very efficient assessment of omissions.
But that's only part of the battle because we've got omissions, but we also have variance. So one of the big things that we talk about at the beginning when we're scoping a project is, what's the accuracy of the laser scan? And what's the accuracy of that model that we're going to get after we go through the scan-to-BIM process?
So here's the answer that we came up with for accuracy validation. Now we apply a color code. So we type in what our accuracy tolerance is. And it applies a color code to it.
So here we can see, now all of a sudden, what the heck? We didn't do the flanges. But that's OK for this because within this scope, we were just looking for spatial coordination. So the flanges just weren't that imperative. They were more concerned with the pipe locations.
But then we can start to see here, maybe we missed a couple of the supports. We need to get those in. But it becomes something that-- it's much more interactive for the QA/QC person to go into and start identifying things that are outside of tolerance.
The other thing that this does for us is it allows us that confidence that when we turn something over, it's right. And it's all right. It's not just right from what we were able to find during our cutting, and pasting, and going through all these cross sections. It's right because we run this again at the end to verify that everything's done.
Now, here's a little example. This is a local one, a casino down here. It was partially built, sat for a while, and then it went under the construction process again. Now, when it did it, started up with a new set a team.
They wanted to use a BIM VDC sort of methodology for construction. But they didn't have anything that accurately depicted the exact conditions of this. So we went through, modeled it up.
And you can see this is a big building. So we had almost 1,000 scans that were conducted. We didn't do the entire tower. But we did a couple of template floors. We did almost 1,000 scans for this.
The area was approximately a million square feet. And we QC'd this entire thing in less than a day. That would have been impossible for us before going through this process. We'd probably still be QCing it right now as I'm sitting here talking to you.
But it was just a really compelling way that now we talked about in the scan process maybe turning off the color to increase efficiency. Well this is increasing efficiency again on the back end. So when you apply this to scale, as a service provider, it means that we can start making healthier margins. As a consumer of the data, it means that there can be a higher level of confidence in what's being delivered.
A couple of things I'll just hit real quick. We're running a little bit over. But we started tweaking it. So after we did that, and we were so inspired, we said, all right, what else can this thing do? And we started building in different little tools.
So we're doing things like ADA compliance. Where's the slope exceeding what's tolerable? So this is an overlay of a mesh and then with some point cloud over top of-- it's hard to see there-- a city block. Actually, there's a theme park block. So you can see there there's a couple of areas where it's red. That's where slope exceeds what's tolerable.
The other thing that we start looking at is new build. So if you've got new built construction, you got the civil engineer or the surveyor out there doing asbuilt on the hard escape, but they're asbuilting the individual spots that were on the grading and drainage plans. What's a huge concern for the owner? It's ponding. In this instance right here, the owner was very concerned because this is the entrance to their building.
And this was in an area that'll freeze. So it wasn't Arizona. We traveled for this one. But it would freeze. And there's just a sheet of ice right at the entrance.
They had a big problem because that's where they were dropping off people from the senior home. So the bus pull up, and they'd get out. You break a hip, that's worth a lot more than getting this kind of work done. So this would identify ponding areas that they can go look at. It also helps in being more expeditious about correcting things within the asphalt before it starts to fail. Over here, this is just an asbuilt verification for the site itself. So it shows deviation, let's them go through all the surface area.
We did some called an Agile Cloud up here. So if I have a design, and I have a laser scan, and I want to identify some routing through that scan, it's very difficult. So if I'm in there, they're just scanned out all over the place, I have a hard time telling the depth, so what this does, it goes through and it cuts out just scanned data around your proposed routing. Now it's a little bit more intuitive for you to go through and start seeing detail. It's also lighter so that you can start integrating this into more of a workflow with the design team.
A couple other things we're doing, Smart Filter. We're really psyched about this one because it's allowing us to bring in just information we want now into the modeling process. So it gets rid of all the scanned-out on flat surfaces to lighten that file size up and allow us to use a much denser cloud for the things that are really important to the modeling team. So I'm encouraging everyone here, use the tools that are available to you. But don't be afraid to think outside the box and start applying different techniques, different procedures that can speed you up.
All right, I think we're ready to go into questions.
VICTORIA ORTEGA: Yeah.
TED MORT: All right. Yes sir.
AUDIENCE: This EcoVerify software [INAUDIBLE] is that just [INAUDIBLE] used to turn on, or [INAUDIBLE] shared that software [INAUDIBLE]?
TED MORT: No, good question. So the EcoVerify software-- let me make sure I'm on, yep-- is that something internal, or is it something we share with others? We share it with others. I did a presentation about four years ago where I just talked about-- this is right when we found it. And we went through and showed everybody how to do it.
So, if anyone's interested in replicating it or a function of it, look me up. Let me know. I'll have my team talk you through it and just show you some of the high level details of creating that yourself. Yes sir.
AUDIENCE: Have you tried the VR for QC for bigger scanned versus models?
TED MORT: Oh, this is probably a good one for you, VO for QC.
VICTORIA ORTEGA: Not yet. We were really interested in to get into the office, but not yet.
AUDIENCE: I can tell you that the NavisWorks check and walkthrough takes it to a whole nother level when you put it into scale.
TED MORT: Very cool.
VICTORIA ORTEGA: Yeah.
TED MORT: I saw another one. Yes sir.
AUDIENCE: You had mentioned you used a bunch of different hardware for different hammers.
TED MORT: Yes.
AUDIENCE: Do you really do your registration in ReCap then?
TED MORT: OK. Question is, we use a lot of different hardware for acquisition of data. Do we mainly do our registration in ReCap?
AUDIENCE: Because each one has their own software registration.
TED MORT: Correct, because each one has their own. We use a combination of different things, not just ReCap. And a lot of it is stipulated to what is being developed by the manufacturer of the hardware is typically it's a little bit of a friendlier relationship through that registration. But ReCap's an incredible tool to implement as well holistically.
AUDIENCE: Thank you.
TED MORT: Yes. Yes ma'am.
AUDIENCE: So you mentioned the different conversion tools going from scan to BIM. So I was curious about how much you guys leverage those plug-ins to automate some of that BIM data, and if you find it useful.
TED MORT: OK. Question is, do we use third-party plug-ins to help automate and make the process more efficient in the scan-to-BIM?
VICTORIA ORTEGA: It's really up to the project. These days, there are clients that ask for it, and clients that they don't want to hear about it. And they trust the old-fashioned, human way. So between one and another, again, it's up to the project, to the LOD, and the kind of elements that you are trying to convert. Yes.
AUDIENCE: About halfway through, you showed a slide of some modeling that you had done in Revit using Point Cloud. And in the slide, you built the exterior wall of the bu8ilding, and the red wall, and--
VICTORIA ORTEGA: The deviation?
AUDIENCE: The deviation.
VICTORIA ORTEGA: Yeah
AUDIENCE: And then you made some adjustments to the wall. How did you make those adjustments? Are they just movable, or did you use some other--
VICTORIA ORTEGA: Yeah, moving the wall on Revit.
AUDIENCE: The whole inspection. [INAUDIBLE].
VICTORIA ORTEGA: Yeah.
AUDIENCE: Just curious.
VICTORIA ORTEGA: On that project we have done it mildly, because it was a small project.
TED MORT: Yes sir.
AUDIENCE: Modeling in regular objects, do you guys like to work with different-- of course, it all depends on the project. But say if you're approaching historic preservation, and there's a lot of case work that you have to go and model, do you guys tend to stay in Revit, or do you go into a more focused 3D model where [INAUDIBLE] can be the [INAUDIBLE] and bring them in as families?
TED MORT: Question-- I'll repeat the questions because they are recording them. So question being, when there are very custom items specifically like in a historic preservation, and you go through the scan-to-BIM, are we doing that in Revit or using an alternative platform?
VICTORIA ORTEGA: If the LOD requires, we do that on Revit also because the elements, it's not going to be intelligent in another software like 3D or something like that. It's going to be like a dummy element. So if the LOD requires, certainly it's going to be intelligent. So in that case, we do it by Revit, yeah.
TED MORT: Are you creating mass elements within it?
VICTORIA ORTEGA: Yeah.
TED MORT: Revit? OK
VICTORIA ORTEGA: Yeah.
TED MORT: Yes sir.
AUDIENCE: What's your [INAUDIBLE] like? Do you have a log for a pipe or something? That type obviously has the [INAUDIBLE] turns, ups and downs and that sort of thing. What kind of workload do you work with for your designers and how often they bury [INAUDIBLE]? So when they're gone, if they're trying to [INAUDIBLE], you don't have all that data?
TED MORT: No, great question. So with something like a long line of pipe where it deviates, there's sag, there's minor movement, how do you determine whether to just model it straight, or to break it into segments? And is there a standard operating procedure for that? Do you mind if I take this one?
VICTORIA ORTEGA: No.
TED MORT: All right. We do a lot of pipe modeling. And an example that I'll give you, we went through and broke it into five-foot segments through this massive ammonia plant, only to find out that when we came out of it, we delivered it, we were proud of ourselves. And they said, oh, we really can't use this. We just need to be straight.
So the way that I would answer that is that you want to have that conversation with the client and talk about what's the use case, what's tolerable. In a lot of cases similar to the wall where there's-- walls aren't going to be plumb and ortho. You need to have that conversation. What kind of deviation is going to constitute either us raising our hand and saying, how do you want us to approach this, or making a geometric correction through the model to better match that.
So with piping, the standard that we use is we'll say that in most cases, we're going to model the pipe to within one inch of the scanned data. And that allows us to play. And it also allows us enough of a variance to be safe.
AUDIENCE: Do you put that data in that element so if [INAUDIBLE] and modeled and you put the tolerance to model that to. Can that help?
TED MORT: No, but that's a great idea.
AUDIENCE: [INAUDIBLE].
VICTORIA ORTEGA: OK, thank you all so much for joining us tonight-- afternoon. I'm already at night.
TED MORT: Thank you.
Downloads
Étiquettes
Produit | |
Secteurs d'activité | |
Thèmes |