Infrastructure Condition Assessment and Monitoring with Point Cloud, Images, and 3D Model

ALI AFRASIABI: OK, it's official. They closed the door. Good morning. How are we doing?

AUDIENCE: Good.

ALI AFRASIABI: Well, thanks for attending our presentation. This is my second year. I'm flattered to be in AU. And every year is getting better. So it's 25th year. My colleague and I-- Vincenzo-- we are presenting the condition assessment for specifically for a structure type structures. And then the goal is to utilize the technology-- high definition imaging, point clouds, and some 3D modeling. We have a lot to cover, so if you don't mind, please keep the question-- [INAUDIBLE]-- keep the question for the end of presentation. We will leave 5 to 10 minutes to answer your question.

With that in mind, let me start with the introduction. I'm the engineering manager at AECOM in California. We do a lot of multidisciplinary projects. Most of them is done in-house because we have all the expertise. We have variety of the clients with different preferences. So hopefully, this presentation cover-- good morning. There are some chair here. There's one there. Sorry, guys, for interruptions. OK.

Let's start again. So I'm engineering manager with AECOM in California. We do a lot of multidisciplinary project. Most of them are done in-house. And then that would give us the edge to collaborate more as is the team of AU in 2017. Also to bring the new technology in the process that we have been doing.

The other item is that I just want to put it out there for grass root evolution. AECOM is a big infrastructure company. But having said that, my observation is usually engaging technology, it happens from bottom up. It doesn't come up up down. So we use the projects to work with other disciplines, other companies, or other clients to bringing the new technology and try to advance them. And one of this presentation is showcase that what we have been doing.

And also, the company is an advocate of IPD-- Integrated Project Delivery-- so we are lucky on that. And we have worked extensively with Autodesk in our several projects to bringing their new products. Having said that, I'll pass it to Vincenzo, who will take you through the first section.

VINCENZO MELCHIORRE: Thank you. Good morning everybody. Thank you for being here. My first time presenting here at Autodesk. Let's see if I can get this figured out. Bear with me. Experience with drones, and structure assessments, and infrastructure designs-- those are the main focus.

When we were selected for presenting here at Autodesk, we were thrilled. Here, there are technologies and tools such as reality capture, centralized model for collaboration, virtual reality, and augmented reality. Those are the tools that we want to leverage to innovate the space that we're going to look at together. So water infrastructure, we'll talk about water infrastructure as a main focus. DSOD is the department, the division that oversees the dance.

There are 87,000 dams in the United States, which 1,200 approximately are in California. One third of this pose a high significant hazard to people and property. So more than 180 failures since 1880s, with fatalities from 0 to large numbers. Generally if it happens, the damage is disproportionate for how could be avoided. So in this case, in South Fork, in Johnstown, Pennsylvania 2,200 people, a 37 feet wall hit the town. This is what it looked like after.

It looks like there is something wrong with that photo. So what we're looking at here is a concrete dam. We're standing on the downstream portion. There is a steel part of the dam, there is already a gate, that opens upward and the water comes from underneath. In 1995, one of these gates failed and released the water downstream without any control.

We have these large infrastructures that we need to monitor because things can happen. This is a more recent event, Oroville dam, the spillway, apparently had some crack on the top portion of the spillway. The spillway is a concrete shoot. The water passes on top of the concrete at a very high speed, called velocity for dam, and this creates damage to the concrete if it is already damaged.

And then the erosion did the rest. It's a very fast event. And it can be extremely catastrophic. Aging infrastructure and older control and monitoring, this is what we want to focus on. And we think we can innovate the workflow that has not been innovated for many years.

I'm going to pass it on to Oli.

ALI AFRASIABI: Thank you. So let's get a sense of audience what we have. Because I don't want to get too technical but if you don't have a technical background, then maybe I just touch on what was-- so how many of you are in a design or construction-related [INAUDIBLE]? Awesome.

So how many of you, specifically, work with dams? Great, so less number. So for those of you who know, be patient. [INAUDIBLE] some technology, so [INAUDIBLE].

So if you look at the dam, this is one of the projects that our office is working on for [INAUDIBLE]. This is an earth damn. And then what we're focusing on is not the dam, but what is here. This is called a spillway, and the reason they put this because if anything goes wrong with the damn, we have a backup system to let the water out in a controlled manner. Usually the dam has starting from 30 to 40 feet of water UP to, I know 600 or 500 [INAUDIBLE].

So you're talking about lots of energy behind this. So you need to control it, otherwise it will go wild. This is an earth dam. Whatever is behind it we call upstream, what is after that is called downstream. So this structure here, it's called a spillway. For earth dams, they have a cross at the top, which this cross is lower than the dam. So when water goes up, the water comes here, and it goes through the spillway. The picture that Vincenzo just showed for [INAUDIBLE] dam, it was some failure actually down here where the spillway couldn't tolerate the load of the water.

The next one is, this is a concrete dam. This is Folsom Dam next to-- it's in Sacramento, north California. This is the original dam they built a new one on the side, but as you can see, this we are looking at the downstream. The water is behind this. For a concrete dam, the system is a bit different. They usually put radial gates at the top. You see there's the radial gates. And then comes in different [INAUDIBLE]. What they use will control the water, very off and on. They let the gate open to have the water [INAUDIBLE].

And there is a spillway after that, and then come down. This one doesn't have a spillway, this has a chute, but there are ones which has a spillway. Again, it depends on the [INAUDIBLE]. And it comes over here, this structure called stilling basin. You can guess by the name that it stills the water, let it go, and then they will take it downstream from there. Let's move on.

With that in mind, let's go to what we do in condition assessment. And then first section we will talk about what we have been doing traditionally, and then in the next section says what we have been doing [INAUDIBLE], and the last part we'll get into what else we can do and the opening to the questions. So the traditional way of the condition assessment or the [INAUDIBLE] come to contact us and said we need to do an assessment or a routine [INAUDIBLE].

The selecting for [INAUDIBLE] example in this case, it's a dam with the radial gates. It will go through some workflows and [INAUDIBLE] will go and look at them, and usually it involves monthly [INAUDIBLE]. It's usually involved through [INAUDIBLE] water, [INAUDIBLE], structure, sometimes energy. So they get together, they go through the checklist to make sure everything gets done right.

To do that, they have to go to the field. As you can guess from the photo, it's not as easy as building assessments. They're a tall structure, usually they do not have access to them, and then even though it's hard to get into them. One of the ways to do it, is on the ground [INAUDIBLE], which you see the long drop, they use the rope access.

You need to be certified to do that. Develop quality control groups to make sure nothing falls, and then you go down there with your camera, with your notes, 200 feet drop, and go ahead do your assessment. Take a photo and then your field notes, and then bring it back to the office. This is a view of the

VINCENZO MELCHIORRE: Kerckhoff.

ALI AFRASIABI: Kerckhoff Dam. This is again in Northern California. As you can see, this is a concrete dam. And then there is some structure here with no road access. And then what I will go through is that to show you how many [INAUDIBLE] downstream. This is a downstream site, as I stated again, these are the steel radial gates doors flowing down, and then what we do is that [INAUDIBLE]. I'm hoping [INAUDIBLE], people go through there safe with their hardhats, camera, [INAUDIBLE], notes, [INAUDIBLE]. I'm trying to get into the detailed to picture how difficult this is. And then, you want to go through all of these and you want to have a good little [INAUDIBLE]. If you're talking about the crack, you're talking about a quarter of an inch. And then we want to measure the land. Consider yourself hanging there, you want to capture all of this.

Then this is at the off-stream site. It's water, so what you can do, you can get a boat, go out there and usually they try to do it at different times, because water levels go up and down, so make sure they capture all factors. Then, [INAUDIBLE], just walk around, [INAUDIBLE]. You can see the boat here is going through taking notes, and then bringing back [INAUDIBLE]. Then what we do with that, there is a engineering process.

AUDIENCE: Can you turn on another microphone, please?

[INAUDIBLE]

ALI AFRASIABI: Sorry about that. You should have told me earlier. [LAUGHS] So you get all of the information, go to the office, and try to bring in the existing documents you have. Usually what it is, is as for a dam, there are some single drawings which shows the different elevation, it shows the stationing of the dam, and different radio gate information. You go through the blueprints to see what things are on the radial gates you look at. There's [INAUDIBLE] to look at. And try to make a note of them. At the end of the day, you give them a very thick report, with a bunch of photos and appendices, and then it's a static report. That's what you give to the clients. And that's what has been done historically.

Let's look at what are the challenges and problems here. The first one is site accessibility. I think that the photo that I showed you, kind of showed the facts here. The other one is the data collection. So you have access, you have limited access to the data which is there. There's a lot more to capture. Then you bring those limited data to the office, and you've got to go through the process. So again, not having enough data, now you go through the process, and you want to link in that to the previous information that you have, and then all of these things will be done manually.

Consider for example, you capture a crack in one of the radial gates. You take a picture of it, you go to the office, you want to locate where it is, you look at the blueprints, you take your photos, and you just define an issue. And you've raised it to the client's attention. And then the accuracy, like what was the crack size? Where was it? Where exactly it is? Hopefully this showcased the challenges which currently we have and hopefully it can be improved. By then I'll pass it to Vincenzo.

VINCENZO MELCHIORRE: So how do we leverage all this with, how do we leverage new technology? We talked about reality capture and centralized model. So, let's look at the field operation. Field operation, We can introduce drone HD images and LiDAR. In different ways. Either everything on the same place, different data sets, overlain data sets. But where it matters is the stuff from the site access and safety, OK? Now if we look at the downstream portion of the same dam as we were looking at before. We think that we don't need to access it physically. We either downstream or upstream.

So there is not going to be either a person or a crew of people rappelling from the top. Plane view of the same thing, will identify like five points that have an advantage view with a straight line of sight. LiDAR can detect most of the geometry, then the detail the images they will cover 100% of the area of interest that can be retrieved with a drone.

So the data collection at this point can pass through, can be completely specified ahead. What is the size of the images? What is the size of the pixels on the structure? What is the overlap of the images? All this gives us a standard data set that we can rely on. And since we know that that is going to be the minimum quality that they're scanning into the process, then the post-processing can leverage that quality with softwares out there that we all know most likely, if you are in the drone environment and the 3D modeling. It's possible to create 3D models fairly easily now just from the photos. We can then have a consistent product and just out of the field.

What is the advantage? Now that this documentation goes through to the office, first of all, there's not going to be just one engineer that repels from the rope. There can be a team, can be a team of engineers, team of geologists, and who else? Material specialties, steel specialty, concrete, and the inspection is performed by a specialized team at this point. Not just one person that needs to know five or six different fields very well, or the history or the dam, because now we have 100% of the information in front of us.

The people there can bring some relevant conclusion to the project are looking at it. So tools has been 360s, other data sets can be overlaid, and-- oops, I'm sorry. --overlaid. And that's when the engineering process starts. Access the information, overlay to new information, and here we go.

So here we have the engineering findings, cracks, displacements, on a [INAUDIBLE] condition. And engineers-- I keep pushing the wrong button. --engineers, geologists can access this type of information. Now the conclusion are using a much larger data set.

ALI AFRASIABI: Vincenzo, before you move forward, let me add some notes in these slides. So basically, I think it has lots of information here, which I would like to point out. This was a portion of the report which we give to the client. Specifically, it's Anderson Dam we go through and we identified the deficiencies. So what we did here was that it was a hybrid approach. It was not fully automated. You get information, you get all the point clouds data, and photos, and then what we did was-- we because we were working on the dams-- part of our projects for seismic retrofit. We had the Civil 3D model of this. So Civil 3D model was our base. Then we got the historic data from the clients to make sure that, at this point of time, we capture what's out there. And the focus is that to build that 3D model and put as much information you have. It might be costly, but it is pay off down the road, because now you have a model. Not all the time, you can go back and retrieve those as built. Just put as much as you can and share it with the clients.

Again, I'm going back to the keynote yesterday for-- I don't know how many of you guys at Forge-- but the emphasis of AU 2017 is collaboration. We've got to start changing the boundaries and share the information together. Again, legality come to picture, we've got to address those, but the more we have in one model, it will help us all better. So going back to here, we saw the data. We got it from previous inspections, so we make sure is those ones getting worse or they're still there? And also, we laid it out with the Civil 3D model we had also the GIS information. They did some aerial photographing and then they got the photo from there. So we rely on all of those, make sure the lay out is there, make sure that the [INAUDIBLE] we are putting with the new design it matches. You don't want a hand in something down the road to the contractor and they come up, It's no right. The time to do it now.

And then just get into the detail of it, some of the items identified by our engineers we saw some cracks. And then when the crack was more than certain sizes that triggers some more investigation, then we will be getting sat down with a geotechnical and geologists to see why are these things happening. One of the reasons. Let me go a segue here is that for the failure Oroville Dam, were off specifically working with the owner agency to figure out what went wrong, and team of experts worked on it. And what happened was, the soil below the spillway started washing out. You don't see that, but this spillway start going down, and you see the crack in the surface. There were only one layer of rebars in the spillway. Usually the rebar-- [INAUDIBLE] I use number four or number five, and this says no, no, no. Water structure is generous. Put rebar put thick concrete. So putting one layer of rebar in spillway, not a very good idea. But it was there and the cracks start happening, correct? You put one layer, you don't have enough covers, and then all the time these cracks start opening up. But nobody is there. Nobody goes there to look at these things. So, that's why we are really vigilant here, to make sure if crack is there, and that crack being monitored. It's like a symptom. Sorry to get--

VINCENZO MELCHIORRE: Thank you. Thank you for--

ALI AFRASIABI: Sure, no problem.

VINCENZO MELCHIORRE: No, I mean, the cracks are seen. They've been detected. But it's important that the right person is involved at that point. If somebody says that there is a critical crack, there is a type of person who needs to be involved. In a spillway on a concrete structure where there is a crack, you want to involve a structural engineer, who is going to say this is going to affect the stability. OK? In this system, this system involves a team of different backgrounds. OK.

So naturally where all this those data sets can live, can live in a canvas [INAUDIBLE] there is a GIS, probably a GIS environment.

Well actually, Ali, do you want to come back on this one?

ALI AFRASIABI: Sure.

VINCENZO MELCHIORRE: Go back on the Anderson dam.

ALI AFRASIABI: We did a 3D model of the LiDAR and the drone photography.

VINCENZO MELCHIORRE: Great. So if you don't want, if you don't mind, I'll take over.

ALI AFRASIABI: Yes.

VINCENZO MELCHIORRE: Yeah, sure.

ALI AFRASIABI: So, Oops, I went forward. Yeah, let's go back here. This shows the geological data and some of the fault. There some of them is not very critical, some of them are. So what we have done was that we went through identified those, and go very deep boreholes to make sure they're captured right when we come up with the retrofit approach. And then, as Vincenzo mentioned, we usually want to have them in a GIS platform. The reason is that our clients like to see things not in a paper anymore. They want to have access to the data real time. To do that, we have a strong GIS team in our office, which we collaborate with them, we use all of those products like BIM 360. Which the engineering output can resides there, as Vincenzo mentioned. And then real time communicate it with that GIS platform, which they can have it without having the softwares.

One of the complaints we had from clients was before they said, "Oh, why don't you send me your Civil 3D model?" I'm like, I'll send it, but a lot of things, there's some legality involved. But really what they want to see is that what the final product looks like. What are the areas they want to look at? And then I think GIS gives us this edge. You can communicate the information, they don't have to have all the information, and it's real time. So it's a happy medium. I'm going back here. This is Anderson Dam. That's what you see here is the photographing of the information and then the GIS teams put this area on top of it. They did a really good job here. But you can see some of this information has been overlayed to basically show them where they are and then what information needs to be added on. Anything else you want to add, or?

VINCENZO MELCHIORRE: Let's move one slide down from here. Yes. So we really covered accessibility, we don't need to go back on that one. We'll definitely see the advantages on that. Standardized data acquisition, data collection, leads towards data processing that is also studied ahead and somehow standardized. So it starts to becomes automatic. The options that we have out there are identifiable ahead. And we can tailor individual processes to cover concrete cracking, to cover steel deformation, to cover erosion, [INAUDIBLE], and soil displacement. So that an automatic process study ahead is going to accelerate the assessment result. And now there is another factor that comes in, which is time. This assessment can be done a lot faster than before.

So, we're starting to enter a kind of 4D dimension where time is not a variable that is as important as before, where to do an assessment-- to complete an assessment takes six months or three months. Now the time can be accelerated and the assessment can almost feel like they are on the map. Almost like you are going towards pushing the bottom the process starts. You know that within three, seven, ten days-- whatever days it will be, depending on the structure-- but you know that you will have a result. You are anchoring a workflow they study ahead.

ALI AFRASIABI: Yeah, thinking in terms of, if it's a rosy picture, but again as I mentioned, we are doing hybrid at this point. It's not all automated. Eventually things will go there, but again, we are scratching the surface. And then Vincenzo walked you through all the advantages we have. So if I'm going back to this workflow, basically if you compare the two, you can see that there are advantages to engaging in new technology. And then cut back on the traditional field work. But still, when we do the LiDAR for some of these projects, we identify some deficiencies, and we send the [INAUDIBLE] inspection guy to go look at it. But this time, he doesn't have to go look at everything. We give him a map that says these are the three locations we want you to go take a look, but get a closer look because, [INAUDIBLE]. So he'd bring with more accurate information. Same as [INAUDIBLE] and other things.

And then, as far as the engineering data, as Vincenzo mentioned, we have more data sets to work with and also the presentation to the clients. But having said that, what does it do for us? This will get us to the last segment of the presentation, which these are good for assessment. Our clients usually will say, "Oh, it's cool, but we are not ready for it." But how does this enhanced workflow will help us to not have another Oroville Dam, to not have another Folsom Dam. What can we do to get that alert? I think that's the main goal here. And then the answer is monitoring. So what we' are proposing here is that if we go back to the enhance workflow, which we come up with. What we can do is that we-- for example, I'll go back to the radial gates example which failed the Folsom Dam in 1995, what we could have done to avoid it.

So to do that, all of us know these are old infrastructures. We can go through the previous failures, also previous inspection, and see what was the deficiencies. And with that, you can specify some benchmark. You can say, these bolts are vulnerable. Every 20 years, they fail, let's add them to the list. And then this plate connection, however, it's strong, but it needs to be visited more often. So that will go to the three months check. The bolts go to one year check. So you start identifying all of those. So you have a very advanced checklist, which that checklist specifically is altered for each component to make sure they are being taken care of.

The next one is to establish sets of rules and regulations. So when you go there, now you have more data set. You're not hanging up there looking at the bolts. You have as much of the information you want. You can capture, if for example, for Folsom Dam, if that [INAUDIBLE] has been tilting a bit. You have the 3D view of that, you can overlay it and see if it's tilted a quarter inch or not. So there's no excuse accepted, correct? So we got to see, but we have too much data. So what we're going to do with it? We've got to go back, set the tone, and then come up with new rules and regulations for this assessment. This needs to be enhanced. We have more data, it's not acceptable to use the same criteria. This criteria needs to be looked at.

But as all of you know, usually these dams are not owned by the real estate owner. They are owned by government agencies, which they have their own rules. It's not an easy move, but the movement is happening and we are the one who can work with them. And probably some of you guys are with some of those agencies to work it out and see what we can improve with these sets of information.

The next one is make it automated. When you make it automated, what happens is that you can do it more often, you can do it more accurately, and you can do it as needed basis. So you-- originally, you have to go for inspection, they will give you a laundry list, and you do it once a year. But this time, it's not like that. It can be weekly check, it can be monthly check, it can be yearly check. For yearly check there are certain things you look at. For a monthly check there are certain things. So that the data collected is not necessarily needs to be processed for the whole thing. It can be processed on as needed basis.

And last but not least, what they can give you is that when you set the tone, you come up with the criteria which can raise the red flag. It will give you early alert system. Specifically for the radial gates in Folsom Dam. If they would have seen that this seal is failing, I don't know how many engineers do I have here, but still as you know, the failure is ductile. It doesn't fail overnight. So, that poor radial gate has been under pressure for years, but nobody listened to the radial gates until it's failed, correct? So if we can start capturing data, say oh it's moving. It's tilting. The affliction is coming, it gives us the sign. It's the symptom. That's our early assessment. And then when it's happened, we go for a visit. It's like a doctor appointment. You see something is wrong here, it's bending too much, but why is bending too much? What's the reason? Maybe the reason is that the bolt. So that's a symptom, now you can go and come up with what else needs to be here? You're deep in.

Now you are beyond assessment. Now you are coming up with what will cause that early-- [INAUDIBLE].

VINCENZO MELCHIORRE: But I want to advertise that the tools are here, now.

ALI AFRASIABI: Yes.

VINCENZO MELCHIORRE: That's why this is possible.

ALI AFRASIABI: Yes. Exactly. This is a Kerckhoff Dam again. So I cherry picked this one for monitoring. So as you can see, it has I think 18 radial--

VINCENZO MELCHIORRE: 14.

ALI AFRASIABI: If I'm no mistaken. Oh 14. 14 radial gates, and then what can be done is that we can select certain location. And this location, as I mentioned, is based on the past data and it will be worked out. So you go through and you look at this specified location with specific criteria. For example, are we missing nuts here? Are we cracked developing here? What is this dimension here? And we deal some information for privacy reasons, but you can see that all of these things can be measured. You can measure the thickness of the plate. You can see that if the plated start bending up, or bottle necking by measuring the accuracy of eighth of an inch, I think. If you do a local. Then some of the welding and other.

So having said that, you will go ahead and do that for specific location for certain time. Like you're doing it today. And then you will do that on a yearly basis, for example, and you will do a year from now. And you can compare the data sets. And figure out what's wrong. Having said that, let me go back to the white board which, the same flow that we have. Now see how we can implement the monitoring here. So as Vincenzo mentioned, this is our fieldwork. Basically we go collect the data on as needed basis. And then we do the engineering in-house and we put the data on BIM 360 using cell 3D and pull these databases.

So what can happen is that we can have this regular monitoring along the way. And again, this is pre-defined. You go through for specific cases. And then any red alarm, you will just communicate to the clients. There is something beeping on your phone saying that this bolt is moving or this steel is thinning out. And that's a red flag. It doesn't mean you know what it is, but I will tell you or there's something wrong, let's give it some love. Let's see what was going on wrong there.

Having said that, there are challenges. Again, I don't want to paint a rosy picture here. There are challenges we need to deal with. One of them is policies and regulations. Give you some example, you want to fly a drone. For some of the agencies in California, I don't call the name, it's six months internal permitting to fly the drone. It's not easy. But again, things are going that direction, but different clients have different expectations of different regulations. And for some of these dams, the dam is owned, for example, with bore reclamation. PG&E is taking over on the downstream sites. DSOD has some jurisdiction, FERC has some jurisdiction, and if you add up all of these peak regulation, there is no wiggle room for you. But the good news is that there are people like us over there who thinks that it makes sense to use this thing. So things are changing. And I think we are on the verge of these changes.

The other one is the initial cost. So as I mentioned, we collect all this information and build a model. Its initial cost. It was same things for the BIM. It's a first high cost, which usually clients back off saying, so we don't have funding How should I pay for this? But if they look at it in the long run, it will pay back. But it needs to go to the policy maker, it needs to go to the higher management to see these things that really putting the initial cost, making those models, it will help you down the road. It has a lot of added value.

VINCENZO MELCHIORRE: In this initial cost, if I can add, there is also a portion that is related to the deliverable. The deliverable is not a report that they can report anymore. The deliverable is a complex entity. It's live. It can keep being updated. So it's much stronger deliverable. So that's why the cost is also larger than would be in the traditional approach.

ALI AFRASIABI: Mm-hmm. Yeah, thank you.

VINCENZO MELCHIORRE: There is more work. Yeah.

ALI AFRASIABI: And then it's a learning curve. Learning curve on all sides. All the agencies, they have a reviewer, correct? You hand them the report, they will go through it, and they have been doing it for whatever they have been working there. Now, you give them something different, correct? So it's a learning curve there. The process will be different because if you go to any of these agencies, they have them been there. The process is predefined. If you want to change that, it takes a lot. It can happen, but it takes energy to go through that. They have to go through workshops. They need to start learning how to use these data sets. This goes to the path forward that it takes a lot of client collaboration. And again as I said early on, it is all about collaborating to kind of gain something better. If we do the same things we used to do, we will get the same results. So that needs to change for better results.

VINCENZO MELCHIORRE: Open conversations. Open conversations to look at the road ahead clearly. We are going towards a route that is fairly new. It's going to take effort. Are you in it? And it needs to be clearly open. It's not going to be easy road. It takes energy from both sides.

ALI AFRASIABI: And then the last, but not least is deliverable. So deliverable, the good news is that deliverable [INAUDIBLE] can be customized for the clients. This is another example related to the infrastructure, but we been work with other clients, which they totally, for example, [INAUDIBLE]. They like to see the assessment of their buildings and we use that as proof of concept to show our infrastructure clients that now they have all the data set on their buildings in a GIS platform. So that can be done for infrastructure in the same manner, but tailored the way they want it. Everything can be tailored for each client's because they have different priorities. They have different type of structures. They have different type of regulation.

With that, I'll open it to questions. Hopefully at the end of presentation, you gain something with you and then as I mentioned this is a two way communication. It needs to be. We are all just working a long way to gain to something better, and then and then hopefully it will give you some insight that you can take it, and run with it, and make it to the better way. And the end, I just want to also thank Gary who coached us from Autodesk along the way to get to the presentation. So I want to acknowledge your input. I appreciate that. I would be happy to answer any questions or comments you might have.

AUDIENCE: I have a question. All the expense part, is [INAUDIBLE] all that initial expense has to come [INAUDIBLE].

ALI AFRASIABI: It goes both ways. Sometimes [INAUDIBLE] does it, but sometimes it goes to the [INAUDIBLE] to use this technology. It depends on the size. But usually it's according to the clients. So see what they want. But it has been done both ways. Yes?

AUDIENCE: With the advancements of the technologies [INAUDIBLE] you guys [INAUDIBLE]?

ALI AFRASIABI: The question is, if with use of those technologies possible to see the structure below the water level?

AUDIENCE: Yeah.

ALI AFRASIABI: It is possible. There is a LiDAR that goes under water, as well, and that accounts for that defraction, I think it's called.

AUDIENCE: Yeah.

ALI AFRASIABI: The change of direction on the [INAUDIBLE] site. So there are methods. [INAUDIBLE] It's been done in some of the projects, not this one specifically, but it is one of the techniques that is used. So, yes. Yes?

AUDIENCE: Can you elaborate on [INAUDIBLE] that you used to [INAUDIBLE] the information.

VINCENZO MELCHIORRE: Yes. So we used BIM 360 to try to keep things together, but the outward canvas where everything lives is generally GIS and GIS environment. The photos are taken by-- let's say, we have a specialized [INAUDIBLE] for specific work, OK? So they do the data acquisition and the initial data works processing. And we generally ask for a 3D model made out of, depending on the [INAUDIBLE] they're using, some of them-- this company in particular is using a LiDAR. And then it puts the imaging straight on the LiDAR and is able to cover the gaps that the LiDAR wouldn't cover depending on the quality of their data set with images and the pixel used in the images. So once that model is created, it goes into either see Civil 3D or, I don't know, [INAUDIBLE]. We haven't tried to bring it to Revit yet.

ALI AFRASIABI: It's no, not Revit. The infrastructure is not very Revit friendly. We use Civil 3D. Also our geotechs, they use some other software, which I'm not familiar with. And then I'm trying to mention all the Autodesk software since we are at AU. BIM 360 is one, and I don't know if we start using Forge or not, but BIM 360 is the main platform for data collaboration and Civil 3D is our go-to. Both we are exploring other softwares, also.

AUDIENCE: Thank you.

VINCENZO MELCHIORRE: No problem. Yes?

AUDIENCE: So it seems like within [INAUDIBLE] you guys are doing [INAUDIBLE] it's also obviously people that do the inspection [INAUDIBLE] Have you seen the cost difference versus just having the one guy [INAUDIBLE] down. And if it is more, is it possible to still have that done?

ALI AFRASIABI: So the question is how is the cost related between the traditional and then enhanced, right? How does it vary?

VINCENZO MELCHIORRE: Do you mind if I answer that?

ALI AFRASIABI: Oh yeah, go ahead. Sure.

VINCENZO MELCHIORRE: We actually did a cost comparison. And if you notice, one of the items I had was the initial cost, correct? Because if you know the initial cost, we can just give them the report. But that's not the real intent they want to have. For Anderson Dam, we could afford to do it because we were doing seismic structure field of the dam. But in general, there is a high spike of initial cost. But if you show the long term profit to the client, it usually tapers off. And then, in the long run, enhanced is more economical. But if you want to look at your nose? No, it's not.

AUDIENCE: How long does it usually take for you to start seeing it pay off?

VINCENZO MELCHIORRE: I would say after second or third inspection.

ALI AFRASIABI: Yeah.

VINCENZO MELCHIORRE: Yeah. It depends how detailed you want to go, correct? Because with enhanced, the deliverable it has a lot of flavor into it, which for the previous one, you don't. But again, you are changing the mindset, correct? They say, I've been doing it for 30 years. I'm happy with what I'm doing. Why should I change?

ALI AFRASIABI: Already the second round of inspections for the same item, the process completely figured it out. And even the flight path is figured out. So they just need to follow the same procedure. So all the time spent to create the procedure itself and the specification and discuss if it's possible or not, there have already been discussions that have happened, so.

VINCENZO MELCHIORRE: But good question. Client asked the same question. [LAUGHS]

AUDIENCE: So then after the second or third time you do the inspection, it's basically you guys have been monitoring them?

VINCENZO MELCHIORRE: It's just a breeze after. Because you have everything, correct? You know what to do. And by then, the client's crew they know what to look at.

AUDIENCE: Sure.

VINCENZO MELCHIORRE: Correct? So it's just, you already paved the road.

AUDIENCE: [INAUDIBLE] from that point on.

ALI AFRASIABI: The intent is enable them to do it even themselves. So there is an evolution of what maintenance approach is.

VINCENZO MELCHIORRE: That's a very good point. So the other selling items, if you want to do this with your client, is that again collaboration. You can tell them that this is a tool, you can use it, too. So if you want to do things yourself, do it. For example, we only do the review for you. You can do it on a monthly basis. Go check it out, then if it's a problem, come talk to us. Or what you can do is that you can work together. For example, maybe they don't have a strong GIS. What we'll say is that, you do the inspection work, you fly the drone, or contract it to anybody you want to do it. Get a very good deal. And then get this data set based on this criteria, and then you can pass it to us, or you can have your internal folks to look at these things. Because maybe they have been doing it themselves. And then put it into system. And then we're just only doing the back end. The rest can be done by themselves.

ALI AFRASIABI: And when we see 87,000 things, there is a lot of work ahead of us. We're just trying to come up with a system that is repeatable. It's fast. Because sooner or later, we're going to have to have multiple inspections. We're seeing them for different reasons. And that 87,000 number multiplies by two, by three, I mean who knows?

VINCENZO MELCHIORRE: Yeah.

ALI AFRASIABI: It can easily go out of control.

VINCENZO MELCHIORRE: Yeah there is a mandate after Oroville Dam for-- just to sec --for Oroville Dam that it's sent out from FERC and the DSOD. That's all the jurisdiction they give them a list. You've got to go through the inspection. But they didn't say what to do. Yeah sorry, go ahead.

AUDIENCE: I was wondering, for the user would it be cost effective to put up stationary scanning stations at dams, so they've could actually scan them periodically instead of you sending a guy in there or getting a drone to come in every year or whatever? I mean, [INAUDIBLE]

ALI AFRASIABI: The question is that it does it make sense to have it stationary, LiDAR couldn't do anything down.

VINCENZO MELCHIORRE: It's just there to system. To capture the [INAUDIBLE]. I love it. It's a great idea.

AUDIENCE: Yeah, well LiDAR was actually the planes.

VINCENZO MELCHIORRE: Here we go.

AUDIENCE: But I'm talking about scanning on site like you're--

ALI AFRASIABI: Absolutely.

VINCENZO MELCHIORRE: Yeah, it's going to be a system. Because this is multiple data sets, multiple data sets overlayed. Therefore, there is going to be LiDAR, there is going to be photos, something stationary, something can be-- I mean from above-- dams are great for taking photos from above. The structure is exposed, you can see it without going to weird spaces. And so is going to be a mixed system. Yes, please?

AUDIENCE: What software do you use for point cloud obstruction?

ALI AFRASIABI: I will pass it to him.

VINCENZO MELCHIORRE: I didn't hear the question.

AUDIENCE: What software do you use for point cloud obstruction?

VINCENZO MELCHIORRE: So to elaborate the point clouds, if it's taken with a photogramatic system, Pix4D is one of the software that can be used, but we generally rely on the subcontractor that we involve. And sometimes they have their own post-processing because they use their devices. They're not just market devices syntax. OK.

ALI AFRASIABI: We do have some in-house. You can send me an email I don't know. It's not my expertise. And I don't want to try to say that I know, point cloud.

AUDIENCE: It's not mine, either. I was just--

ALI AFRASIABI: But it you are interested, I think my email is there. Just shoot me an e mail, I'll hook you up with the folks. If they can share the information with you, we'll be happy to share with you.

VINCENZO MELCHIORRE: Yes? [INAUDIBLE]

AUDIENCE: Yeah, we used [INAUDIBLE], but have some difficulties.

ALI AFRASIABI: I see. Yeah, no absolutely. In back?

AUDIENCE: [INAUDIBLE] Can you elaborate a little bit on how you guarantee confidence of the data capture? [INAUDIBLE]

VINCENZO MELCHIORRE: Yes. So the question is, how do we guarantee that we have a complete scanning of their portion. So, one method that we used is a scanning in two sequence. The first one was the LiDAR scan that creates a 3D point cloud, point cloud. This point cloud is then uploaded on the drone, the drone has the capabilities to see the point cloud as it flies. So it's not relying on the GPS anymore, but it sees the 3D structure as it flies. One of the upsides is the accuracy of where the photos can be taken. Since it's not relying on GPS, there is an accuracy of three feet, one foot, depending on the type of GPS. But now we're going to take the photo exactly there. And we're going to take it now, we're going to take it in a month, in two months, in three months.

So, as far as completeness, once you have a scan of the area that you were looking at, then the process is just going to keep going until that area is fully covered. And then there is some quality control.

ALI AFRASIABI: Just to add one thing to that is, that our strategy currently is that you-- because you get more data, is more expensive, correct? So what we do is, we go with a less accurate way to capture the whole thing, and as needed we zoom in and pay for it. Which, it makes sense.