Navisworks Has a Point to Prove, But What Is the Point?

IAN PHILPOTT: My name's Ian Philpott. I'm a civil BM lead for J Murphy and Sons, which is the UK heavy civil engineering contractor. This topic is something I've been working on from my previous employment. It really came about from a lot of highway engineering-type projects. I'm going to skip through this stuff quite quickly because I've got a lot to cover. Please don't expect the PowerPoint to be all singing or dancing, it's just really a placeholder to keep me in check.

The notes are up on the AU site. There's basically everything I'm going to do today and some in here, and all the steps you would ever need to repeat what I'm going to show you. And at the end I'll put up my email address and Twitter handle, etc, so if anybody wants to get in touch and talk about anything, please do. So, class summary. We won't dwell on that too long. You've seen that all before.

Again, with the key learning objectives, we're just going to look at the various bits of software out there. So to recap, Navisworks, Civil 3D, bit more Navisworks, and a bit of subassembly composer crept in there for one of the more advanced sections of the workflow. So this is me. 20 years in civil engineering, 10 years of 3D modeling, consultant contractor, and wide-ranging projects, anything from 20 houses up to about four billion pounds of urban regeneration type work.

I also sit on the Civil 3D UK user group. So there's four of us that manage that across the UK and we do a number of presentations across the country, about three or four times a year. So just quickly flick these down. So why? This all came about from a project I was working on. We had a problem and needed to fix it. Very complicated, I'll show you some pictures in a minute.

We were already using Navisworks to measure things, but people kept changing the design. So I wondered whether I could automate that process so that we didn't have to keep going back and repeating the same tasks again, and manually checking all these head rooms and clearances. Most of all what we wanted was a fast, repeatable, and auditable solution.

So this was the project I was going to use. But I went and changed jobs and didn't think it was really right for me to go and use a data set from a project from my previous employer. But this is the kind of thing we're talking about, quite complicated structures, a number of overhead structures, and some railway arches. So fortunately, my friends at Autodesk and Topcon gave me some more data to play with, so we do have something to show you today.

What I want to touch on first is a bit about data management. When you're dealing with this kind of data, if you don't manage it properly, you're going to get in a mess very, very quickly. The job we were looking at was four billion quid worth of infrastructure works. And the first phase of that was something about 400 million pounds. So you can imagine the amount of data I had-- about 180 gigs worth of point cloud, and somewhere about 2 and 1/2 billion points in that dataset.

So there was some initial processing to do. A fair bit of central London work, so you can imagine it's any big city, lots of noise, vehicles going past, no matter what day or time-- day or night, so you've got some noise to clean up. Then you've got to make the data manageable. You can display it very well in Navisworks. It will cope with those 2 and 1/2 billion points very easily. But you try and crack the surface out of a road with anything more than about two million points in it, you're going to come unstuck very quickly.

And even-- don't know what that was. Even at the kind of levels we're looking at in this model, which is about 1.1 million points, I won't do the surface creation today because it would take 15 to 20 minutes to create a surface from it. So what we're going to look at first is the basic work flow, which what I call quick and dirty clearances. So we'll look at some Civil 3D exports, a bit of Navisworks import, and I'm going to just measure some clearances, some very, very quickly. So I'll just-- and recording our results.

So I'll quickly just run through what we were doing with that, previously. So let's go to-- find the right window. Got lots open. So this is the data set we're going to be looking at. This is around Shrewsbury in the UK taken from a mobile mapping vehicle. And you can see it's quite a-- oh, sorry. That's better. All right, so I'll zoom back out again, back to where we were.

So this is Ring Road around Shrewsbury, a small town in the UK. And taken from a mobile mapping vehicle, so [INAUDIBLE]. You can see that there's a fair amount of data lurking in there. And you can see some-- we go here, you can see some noise, as well, that's some vehicle noise in there. And again, we see that moves around fairly smoothly in recap if we just flip across to Navisworks it's the same thing inside Navisworks.

So you get the same sort of behavior in Navisworks. There's quite a lot of points there. That's probably about five or six million points in this cluster. So let's get across to-- so what I've decided to do is we'll pick a piece of this data set where we had a bit of-- there's a couple of bridges that fly over in this area here. And we were going to do just a typical kind of thing. You do some road rehabilitation.

You're going to put some new surfacing on. You're going to change the levels on the road. So we created a bit of a surface. I'll just switch that surface on so you can see what-- I won't try and create this today, because it is quite-- it does take quite a long time to-- and even now, as you can see, there's 1.1, 1.2 million points in this-- Lost a bit of [INAUDIBLE].

Let me just unload that point cloud data for a minute. So you can see it's quite an intense triangulation in there. So there's lots of points for you to do. So we created a surface from it. This technology only came about, so if you're running anything before Civil 3D 2015 productivity pack 1, then you'll need to export it from recap and bring in as a PTS file and do it that way. From productivity pack 1 in 2015, you can actually take it directly from recap files. That's all explained in the handouts in there.

So we took that into Navisworks. Let me switch to the other file now. So this is our file. So we've actually brought across a very basic corridor that we created for our road rehabilitation. And we've got our point cloud data. So if I spin around you can see we've actually got some-- you can actually see these bridges in here. And the very basic workflow is just to measure it.

So you've got your measurement tools. And then remember to lock the zed value. And you can actually pick out the point cloud. And you can see the change in the cursor, so you know you've got the point cloud. And it will come straight down. And when you get the square, you get height clearance. It's not the total answer, but I found it very good when you're doing the design work just to be able to flick back and forth, make sure you've got your head rooms right.

But it doesn't just stop there. If you convert that to a red line, you can see over here in the save view points, it saved a view point for me. So I generally like to just rename that. Call it basic. And you can come back to it at a later date. So it's always there when you want to come back to it. That's where we started. Everybody can do that. It's a very, very simple workflow.

Let me switch back to Powerpoint. So that's where we were with the very basics of it. Then I thought, when I was actually doing this, this intermediate workflow came to mind. And this is something that really has, again, this wasn't available to me when I actually was working on the project because I was working on the 2014 platform. This is, again, only came about in 2015 productivity pack 1, when we were able to create solids from Civil 3D surfaces.

So you have to create a little bit of a subassembly. And we'll go and look at that in a minute. I've got a little trick that I've come [? across ?] with that that actually helps you create a subassembly that's dynamic to your row center line corridor. So if your road moves, your little envelope is going to move with it. Then we'll extract some surface solids, a little bit of export again, and then we'll stylize it in Navisworks and then we'll do a bit with Clash Detective.

So again, we'll switch back to Civil 3D. Civil 3D intermediate one. So I'll go in and turn that surface off and get it out of the way. So this is our little bit we're talking about. But the first thing we've got to look at is this little subassembly I've created up here. And this is what allows us to create our corridor envelope.

So what we've got is a little vertical link. And this is the really most important piece of this, because this link here is what keeps it tied to your center line. So it's always going to work from your vertical profile you've created. And this is one of the things that a lot of people don't use a lot of inside Civil 3D. So if we just look at-- let me just expand my tool palette out a moment.

OK, down to-- so in here, this is the generic link in here, link vertical. But if you put that in like it's shown in there at the moment, when you try to create a surface from it, you're going to get some quite wacky results, because it tries to triangulate down to this bit and up to these bits. But there's a little bit of an unknown command hiding in here. So this little command here called omit link. So if you omit that link when you're creating the-- it will basically ignore that vertical up stand when it tries to create the surface.

So again, what we do, just like any other corridor, I won't repeat it. I'm sure you are all well-versed with Civil 3D. You create a Civil 3D corridor of this item. And this is my 3D solids. So it's basically created this Civil 3D corridor in here. And you created a surface from it very much like you do any other. But now, with this extract surface solids, you can create a surface solid from it.

So you we pick on your surface. Go to extract from surface tab. And extract solids from a surface. And you'll get this window come up. What you have to remember here is, obviously, your plane is now up in the air. So when you set your depth here, you're going to need to set it to whatever you want your clearance downwards to be. So I think in my case here, I think I had it-- I think I was at 5.7, which is standard for the UK.

Lots of other parameters you can set. And then you create these solids. So if I pick on this solid now, you've actually got a box. But that will move with your center line. If you try and export this now to take into Navisworks, a lot of information will come with it. So the center line and all these markings. You don't necessarily want to do that.

One of the things when I was working on this project with my team, previously, was we-- doesn't seem to like having two windows open today for some reason. It loses focus. I can do it from-- I'll show you from here. But this Navis export layerstate I find very, very useful. So go in through your layer manager and turn off all of the stuff you don't want, which predominately is every layer bar the surface layers that Civil 3D creates. Then you'll be left with a nice, clean export to work with.

So again, if we go into Navis and we come away from our-- so if I just turn off everything for the moment. And I'll turn on my intermediate. So this is the same object brought into Navisworks as a polygon object. Again, once you're in here, you can-- if I reset these items back to when it comes in, it's obviously a plane object. I found it quite useful to change the color on it a bit, and change its transparency as well, so that when you're looking at it your point cloud data, it actually shows up better where the clashes are.

Because although this is the workflow that we came up with, it's far from ideal and the way the point clouds come in at the moment is not the best. But there are technologies that are on route to help with that. So I switch back on my point cloud data. So very simply at the moment, you can actually see, because this bridge is-- I've only scanned the underside of it. So you can actually see this area in here is actually clashing with the bridge deck.

So what we'll do now is we'll go into Clash Detective. This is something we've sort of been experimenting with for some time as to how to make this work a bit better. And it relies on how you break up your point cloud data. And again, it very much depends on how many you've got to do. If you've got a lot to do, I would suggest you want to break up your bridges from your surfaces.

And if you're really worried about the clearances, separate your bridge piers and columns from the decks, so that when you class check you can actually class check against individual objects, rather than just class checking against an entire model of the point cloud. But really, for speed of this today, we'll just do the two. So we've picked up our intermediate. We've gone and found our surface solid. And we've gone and found our point cloud data.

And we'll run a test on it. This is where live demos come into their own and you run our of screen real estate. But you can see, because of the way the point cloud in this model has been broken up, it doesn't give you exactly the right answer. But it does point out that there is a clash. And for me, when you're working with bridges and clearances, there is either a clash or there's not a clash. If there's not a clash, you're all right. If there is, you've got to do something about it.

With this example, it's never really easy to do what you want, which is where the advanced workflow comes in. There's a bit more flexibility to the clash check different objects. And with a bit more work on the point cloud, you can do that. So you can see, there's a couple of locations and a couple of bits of data. So there's clearly some kind of issue over here that we need to go and investigate a bit more.

And then we've got this issue up here, which in this case, is fairly blatant to see. It's in this area here. So we'll just switch that one off again. And we'll switch on the advanced one. So I go back to PowerPoint again. So this is where we start to look at a bit more, some complex clearances. What have I done? Advanced workflow. That's better.

So we're going to look at a little bit about subassembly composer. How many people have used subassembly composer? A few, but not very many. This is one of these bits of software within Civil 3D that it's not the easiest thing to use, but it does give you some power of creating things. What we're going to show you here is not a hugely complex subassembly. And I'll talk you through the process of what's going on.

But it does give me some benefits over the intermediate work flow. So then we'll import our subassembly. We'll do the same kind of thing. We'll do a corridor. But this time we will extract the corridor's corridor solids, rather than surface solids. Again, that gives us some benefits that we'll talk about. And we'll do a similar class detection.

But I'll show you why this one gives you more flexibility. So first, we'll take a look at subassembly composer. So for those of you haven't used this before, this is a way of creating those subassemblies that you get from your-- so this is the inside of subassembly composer. It looks a bit daunting to start with. And it certainly daunted me the first time I tried to use it. But like I say, it gives me some benefits in this regime.

So you've got-- on the left hand side, you've got a number of tool boxes where you've got various geometry items you can use to create your thing. You got your flow chart in the middle that shows you how everything's built. This is your preview of what your subassembly is going to look like. And then we've got our parameters over on this side and some more parameters on various tabs. This bit in the bottom right hand corner is where you've got to start where you need to have a think about what you're going to do first.

And if you look at the handout notes that I provided, there's actually a little sketch in here. And doing the sketch is the absolute only way you can do this, because you have to sit and work out what all of your parameters you want to put in are going to be. So for example, if you look at this, the first one that comes in is side. So for those of you familiar with Civil 3D, you've got your left and your right side when you put things in. That's a default. We don't have to do anything.

So then I've got clearance right, clearance left. And I've got width left, width right, and this parameter of what I've called sag correction. So what we've got is, obviously, the clearance on the right hand side. So this is this link L4. Just about see it on the right hand side. Clearance left is this [? link ?] L6. So that controls the height of your clearance envelope.

The width left and right, so we start off with our base width. So I've gone with what I would use in the UK, a standard carriageway width. So you can make that whatever you like. So I've gone with 3.65 meters, so [INAUDIBLE] metric. And this sag correction, I don't know what your standards are like over here, but in the UK, if we've got a curve, if we've got a bridge and we're doing the clearance, if that is on a sag curve, then we need to put a correction factor on the clearance to allow for any-- as a long vehicle goes through it, we make sure we miss the headroom.

So that allows me to put that sag correction. And that comes from a set of tables we have in the UK, hence why I've got that in as a parameter. We've also got targets. And again, if you're used to using Civil 3D, you'll be familiar with working with targets. But what I've got here is-- so obviously I've got a target for the existing road. I'll explain that a bit more in a minute. And then I've got some off-set targets [INAUDIBLE] left and right. So again, you're not limited by having a fixed width for these two lengths L1 and L2.

So you can use your offset channel lines to target and force this subassembly to go wider to suit whatever your road is doing. So this will work with any situation, any width of road. So where you start is with these lengths L1 and L2. So obviously, the Red Cross is where your alignment is. And we've used this over here. Where is it? Surface link. So these links, L1 and L2, are surface links.

So if I click into here, then you'll see where these parameters will start to fit in. So there is your link. So you've got this target surface. So it's going to target this parameter called existing road. So that allows us to make this bottom link follow your existing road profile. And the same with L2. We then move on to these links, vertical links in the side. But if you look at this bit here, you can see that we don't just have to have a plain output parameter in there. So we've actually got formula in there.

So it looks at our clearance left and our sag correction and sums them up, and makes that the length of those two, the leg there and the leg there. And then we've just got a little leg to close this off at the top. So that's a very, very basic subassembly. You can make these incredibly complicated. I've got another one back in the office that's got a similar thing that's tagged on the side for what we call a structure-free zone. So we're not allowed any structures that are outside within, sort of, two meters of that thing.

But we have a reduced headroom requirement on that. So we can have that programmed in, but I didn't think that would really work for this class. So that saves that file. And it saves that as a PKT file. So if we go back into Civil 3D-- let's just close recap, because we don't need that anymore. Close subassembly composer because we don't need that. Get back into our advance one in Civil 3D. And to get that in, you can just come to this import subassemblies, and find your file, and create yourself a little tool palette where it lives.

So I put mine on a tool palette somewhere, buried in here somewhere. It's probably on my old computer before it expired, because I did that before that. So let's just do that. So import, subassemblies, let's pick our file. Clearance, there we go. Create a new tool palette. Call it AU 2015. So we've now got our subassembly item in here. So if I just quickly create a subassembly, we can then drag that onto our-- so there's our subassembly.

And then you'll start to see all of these-- so all of those parameters we set. So the headroom left, headroom right, width left, width right, sag correction factors are all start to appear in here so you can customize them. Let's get rid of that so I don't get confused. So again, let me just remove that. So that's our main corridor we had on. So let's just switch that off for a moment.

So again, we did a very similar process. So we created ourself a corridor. We've used the same-- so if I come in to our dialogue box here, we've used the same alignment, which is this road centerline left. We've used the same profile. But we've used our new subassembly and just created this short section of corridor. Let me just get rid of these-- so we've got our short section of corridor.

So at the moment it just looks like a flat object because it's just set up at the top. Still got some legacy stuff. So if I select that corridor now, we can now go to this, correct surface solids. This you could do back in 2014. So this is where I started off. And the intermediate workflow only came along after the event. So you can create your surface solids. So we create all of those. So you add all your baselines to it.

And then you can either insert it into this drawing or insert it into a new drawing, depending on what you want to do. For this one, I will insert in the drawing. But if you were doing a big project, again, you'd want to do it into separate drawings. As the software develops, I would expect that in the future to become a dynamic object. At the moment, once you've created those surface solids, they are not dynamic anymore. So if the road moves, you need to remove them and go and recreate them.

So if I just quickly window over this area here, you can actually see we've now got-- What happened there? [INAUDIBLE] Add all my baselines. OK, what's going on there? I knew there was something was going to go wrong today. Extract surface solids. [INAUDIBLE]. OK, maybe I've done something wrong somewhere. That's disappointing.

What's going on? Don't understand that one. So we'll just skip over that momentarily. But it does work, because I did it about an hour and a half ago. So let's just hide all of these again. So we're getting some more data coming through that we don't want. I'll just hide that. I'll just hide that. There's something on subtly wrong here. But the bottom is largely irrelevant.

This is where this method is an improvement over that intermediate method, because what we've got here is we've actually got, in that previous example-- so if I go back to-- let me just freeze that off a moment and go back to our intermediate one-- Hit some wrong buttons in my haste to do things.

All right, OK. We'll just get back to open the file separately. I knew it was too good to be true today. Unfortunately, this is what happens when you decide to do things live. And I always jump with jeopardy and go live. So if I actually look at this, it's actually much harder to-- so that's the corridor surface that's come through. So I need to hide that. But that 3D solid is one complete object.

So you're very limited. You can only class check against the entire object. Whereas, if we go back to-- my point cloud, in this example it came straight in from recap. So it's reading a recap file directly, yes, directly from the RCP. That came in in 2015 productivity pack 1 in 2016. If you're in a version before that, you'll need to go into recap and export it from recap as a PTS. And then bring it in through the point cloud menu in Civil 3D.

AUDIENCE: [INAUDIBLE].

IAN PHILPOTT: Yep. Again, all that stuff is covered in the notes, both the old workflow for 2014, and the new workflow for 2015. So you can pick it up from there. Have I lost my-- there we go. So let's get back to where we were. Let's turn this back on. So as you can see now, this is my advanced corridor into Navisworks. And the advantage here, if I click on that, you can see we've actually got the entities separately.

So if we go and do our clash detective again, in this example, so I'll get rid of-- delete that test. We'll go and we'll create a new test from scratch. So we can go to the here down into our-- and we knew that. So that bottom one there is the-- so we've actually selected the same one that we've got selected on the screen now. So that's the [INAUDIBLE] of the thing against the point cloud.

And you have to remember to set this clearance. And you also have to remember, in here, to set this item as point. If you don't do that, then it won't work. So again, we can now say that it has produced a clash. We can also run another test. So if we go and run another test now, we can go back to our advanced one and we'll use these other three objects, which are the bottom and the side. Let me just minimize-- so that's the side. That's the side. So it's those two.

So we'll get rid of that one. Again, against point cloud, make sure that's checked, clearance check. So we've actually got a clash somewhere over on this side. So this is where this advanced workflow with that subassembly in it, allows you to check the [? source of it ?] and the side separately from the [? thing. ?] Particularly at the moment, that was important for me because the way the workflow, and you can't break up this point cloud very easily. It can't determine where it is.

So at least with this advanced workflow, you can determine whether it's the side or the [INAUDIBLE] that's actually doing the clashing. So that's really a very, very quick run through of what we've got going on. What I had coming up with my workflows. But what I thought I'd do as well is look at what else we can do with this technology. So we looked at our--

I went back one too many. Visualization. So what other value can we get from these models? So visualizations, I've worked with a number of clients, particularly private developers, that are very, very-- when you're doing this kind of work for them, industrial clients, doing bridge clearances, because they want to take a big piece of equipment from their factory down to the dock somewhere. They now have to build equipment, but they're not very good at reading drawings. This is a very visual wire showing them where the issues are and why they've got to do something about it if they want to get their piece of equipment through.

I found it invaluable with Navisworks to get that point across to them. Again, stakeholder engagement, a lot of the work I do has been done with private developers. So you're dealing with what we have in the UK, which is highway authorities. You guys called it DOTs over here. They're very well educated, but you have to explain-- you still need to clearly demonstrate what you need to do and why you need to do it, because it's obviously their asset. And it's very valuable to them.

And what we've come on to more now, is I since I moved from working for a consultant to working for a contractor, is health and safety briefings. When you're working on live carriageways, and you've got traffic management going on to do some work, or you're trying to get a crying or a large excavator under a bridge, you can actually show this and show the working areas and stuff using that visualization method.

And the same with construction planning. In the UK, our road networks are very, very tight. There's not a lot of room. So we have to do an awful lot of forward planning, trying to make sure that we've got room for both the right number of vehicle lanes to be left over, all but narrowed down barriers, then working safety zones before our construction areas. Again, we can show that. We can use the measurement tools in Navisworks against those objects to try and demonstrate and show people that we thought about that.

You're seeing now-- so what I've been showing you is what you can do now. But what next? Recognition of point cloud rather than cloud in its entirety. Referred to a scant bin, commonly. And if you've been down to the exhibit hall-- anybody been down to see Edgewise, Clear Edge? They've got a piece of software that recognizes beams and pipes. I don't think it's, personally, I don't think it's going to be long before that kind of technology is available to us in the infrastructure world.

Imagine taking that scan of the bridge and [INAUDIBLE] being able to create a bridge object. That suddenly becomes even better for us. So this is-- clearance envelopes. I use Autodesk vehicle tracking for a lot of my work, because that came from a company called Savoy in the UK. And was our baseline vehicle tracking software for many years before Autodesk bought it. I can see that becoming more advanced with clearance envelopes attached to your vehicles.

So imagine being able to do that, putting that on your alignment. And make it drive down there. And then instead of all the work I've just had to do, the vehicle envelope is left behind from your vehicle tracking. It's just a layer you can switch on and off as an object. That really comes to the end of what I had to say on this. And like I said, this was a very, very quick run through. There is an awful lot more detail in here, and very, very detailed procedures as to how all of this was done.

But if anybody's got any questions, I'm happy to-- yeah?

AUDIENCE: You were talking about changing [INAUDIBLE]. You use the subassembly and you have all those different diagrams, can you change the [INAUDIBLE]. The subassembly that you created.

IAN PHILPOTT: Yes.

AUDIENCE: So those numbers and things that you get [INAUDIBLE] numbers you change.

IAN PHILPOTT: Yes, absolutely. So I'll just repeat the question for the recording. The session's being recorded. So the question is, the subassembly I've created, can we change the numbers that we were using in there? Absolutely. So if we go into our subassembly over here-- oh, sorry. Every time I switch it seems to revert back to extend rather than duplicate. [INAUDIBLE] So all of these, those are the parameters there.

So if we suddenly say, it's not 5.7 meters. It's 7.5 meters for both of those. Instantly see that jump up. But what you can also do, as well, so the width parameter. So at the moment, that width is set across here at the bottom by those two values. So if I quickly draw-- say our road is actually wider out here. And we've got a couple of polylines that represent the edge of our carriageway, for example. So if I go into there and-- where's my regions? I'll do it from the other.

So I'll go into my corridor properties. So I'll set my targets on-- so here, so here's my target left. I can actually-- it does show in here because I haven't got the co-set set inside, but that will actually force that sub-- just the same way you do inside Civil 3D if you want to force a road width, those same parameters. So you can be forcing that subassembly out to another alignment, a couple of polylines that you put inside. So that's the flexibility you get.

If you create those subassemblies, you can put those target parameters in. So if I go back to-- minimize all this. Get back to subassumbly composer in here. So that's those two targets. So those are your horizontal targets. So you can use that to force the width on the left or right hand side of your centerline. Any more questions?

AUDIENCE: Can you run the analysis in [INAUDIBLE] where you show ranges of clearance complex [INAUDIBLE].

IAN PHILPOTT: You can't in-- it doesn't seem to work very well inside the point cloud. It actually reports-- so I go back to clashes. I go back to my clash report. Results, if I expand this window across. When you're working with normal civil and revit objects, it will actually populate that distance box. But for some reason, when you're working with point cloud data, it doesn't do that. It's not relying on an entity.

So there's lots of points. And that entity is quite big. So it can't work out what that clearance distance will be. This is where I'm hoping the point cloud recognition-type software will come forward for us in the next sort of 12 months or so. I don't think it's very far away, because they're obviously doing it with building objects. I say Clear Edge are doing it with pipes, beams, walls, windows. It's not going to be long before somebody can do the same thing with engineering structures.

And once you do that, you've then got a native Revit or Civil object that you can do a proper class check with. So we'll be much better off. We'll be able to class check and know whether it's hitting the pier or the [? soffit ?] of the deck, or even as much as a light fitting that's mounted on the underside of the deck. I've had occasions in that previous project where we had-- it was an arch, brick arch bridge. And it had drainage scuppers coming down from the rail track drainage above that rain around the shape of the bridge.

And where they came right round and met the dis of my clearance envelope, it was those that were fouling. But this method, although it didn't tell me what it was, it told me there was a clash. And we're looking at an object that's probably only four inches wide in reality, and about three inches off stand from the soffit. So it took some finding, but at least it told me there was a problem.

Whereas if you'd done that and not noticed it in the field, the first big optic that goes steaming through there and rips all that drainage off, you're going to have a problem. And you're going to end up with a bridge strike. So at the moment, the workflow is not ideal. But it gets me-- it gets me here a true or a false result. And that's, I think, is the best we can do at the moment.

AUDIENCE: Can you rename-- like you have four bodies [INAUDIBLE] two sides [INAUDIBLE]. Can you rename those just so--

IAN PHILPOTT: Yeah. You can go through the objects and change the parameters. That's done in the bit where you are creating the surface solid.

AUDIENCE: [INAUDIBLE].

IAN PHILPOTT: Yeah, it's one of those things. I've tried to-- I didn't do it here on purpose, because it's one of those things I've worked on a lot with my team, is creating BIM standards for object naming. And you've got to get the object naming right. And the whole idea of that is to work. If you get it right, you can then work with your Navis search sets. But you have to have a clear idea. So if you're going to do this a lot, like you say, you name it left, right, [? soffit, ?] for example.

You can then save that search set in Navis, so that the next time you come to do it, you can go import search, and it will find all of those left things. And you run a clash check on it. But it's quite a complex process. You do have to sit and think about your naming strategy. But if you do, you can then save all these searches and tests. Even these tests can be exported from here, and imported again. So if you're naming strategy is consistent across your entire portfolio of work, so the next bridge job is the same, you'll never have to do this again once you've done it right the first time.

AUDIENCE: Oh, good.

AUDIENCE: Following up to my previous question, is there a way to [? find out the ?] maximum clearance?

IAN PHILPOTT: Other than going back to the manual method at the beginning and then measuring it, not really at the moment. Again, that's something that's on the way, I'm pretty confident of. Again, we're seeing people creating objects from point clouds. Once we can do that, then it becomes much more intelligent to do that. I would keep your eye out. That's all I can say. I've seen some preview type stuff. But obviously, it's not there just yet.

But it will come. And like I say, if people are doing it with buildings, it's not going to be long before somebody can translate that into civil engineering structures. Yeah?

AUDIENCE: Can you do the same thing with wires, like [INAUDIBLE].

IAN PHILPOTT: Exactly the same. The clearance envelope would be the same. All I would do with that was I would isolate in recap the wire, and save it as a separate object so that you're purely just checking against that wire, because you could get quite a lot of noise. When it's scanning wires, you can get a little bit of a reflection sometimes. And if it doesn't get it quite right, you'll not necessarily know where it's hitting.

Whereas, if you just isolate the sagged wire as a point cloud object and do it that way, that would be the way I would do it.

AUDIENCE: We're the only ones that have on-site company right now. And they-- it's a rail yard.

IAN PHILPOTT: Yep.

AUDIENCE: And they have a lot of wires going through. And like you said [INAUDIBLE]. So [INAUDIBLE].

IAN PHILPOTT: Yeah.

AUDIENCE: [INAUDIBLE].

IAN PHILPOTT: I very rarely do manual surveys anymore. I get it laser scanned because the technology is there with people like Topcon, [? Lyca. ?] They've got software that will turn it into a conventional survey for you.

AUDIENCE: We have the technology we just--

IAN PHILPOTT: Yeah. It's certainly things like that. If it comes to anything like-- because those are going to be high voltage wires. I wouldn't put anybody near those anymore. Why do you put anybody near that and put them in any kind of risk when you can laser scan it and get the same information out of it. It's the same reason why that project I showed was laser scanned. That's got a main UK motorway. That was done from the back of a truck at 50 miles an hour.

And then they terrestrial scanned the roundabout around it. But that meant we could do that at any time during the day when we could find a clear spot in traffic. And not one single operative had to stand anywhere near a live carriageway, which is, to me, fundamentally the right way to do it. We go exactly the same information, but there's also the fact that if you do that with a conventional survey, you get your conventional survey, but you can't guarantee--

If you get to when you're trying to do a [INAUDIBLE] pavement [? tie-in. ?] And you think, is that quite right? Is it a little bubble in the data? With the point cloud you've got masses and masses of data. You can go back and just look at a bit, a couple of feet this way, a couple of feet that way, and just make sure that it makes sense. But all that data is still there. And it's still available to you, all the way through the life cycle of the project.

And the scheme I was showing there is hugely complex. It's a three layer grade separated junction. And you've got-- we had a brick arch railway going that way, a fly over going that way, another one going that way, and a whole roundabout underneath. One issue with your headrooms there could make for massive knock on effect on all of those junctions as you go up. And you could end up with a really big issue at the top, because you've had to lift all of this up.

You can't get your ramp to fit in before you get to a building that someone down there. This is in central London. And it's the only way, that's the only way to do it. So we had the entire thing just laser scanned. And I'm working with companies now that the first they do, contracting companies, the first thing they do when they get on site is just laser scan everything. As much as anything, for conditions survey, because with a modern laser scanners you can capture photographs as well, high resolution photographs, as well as the point cloud data.

So you can go back and look for-- you get accused of making a crack in a pavement. Well, you go back to your scan of photographs. No, this crack was there when we started. It's just assurance. I think it's assurances for everybody. It protects everybody. It doesn't just protect you as the designer or you as the contractor. It just makes it much easier and removes the confrontational issues we have in the construction industry.

AUDIENCE: [INAUDIBLE].

IAN PHILPOTT: Yeah. The biggest cause of RFIs, ambiguities. Hence why I was sort of working with this. Again, I go back to the repeatable, traceable, auditable. So again, with this you can go back to it any point. If that road moves, with the second two examples, those envelopes will move. Those clash checks are saved, you just hit rerun. And if you had no clashes to start with and you suddenly got two clashes, you know you've created a problem. You can go back and fix it.

But once you save them, you don't have to do any more work. You just hit that rerun test over there, and it will do it.

AUDIENCE: So what is it exported out to?

IAN PHILPOTT: It exports-- those go out as-- they're XMLs. So it saves the test as an XML. You can also save the reports. So if I go to the reports here, current test, XML-- I'll just open it as a-- documents. So that's the-- so this is the report you can export. So this is saved as an-- you can save that as a CSV-type file, so you can suck it into Excel. This is saved as an HTML file.