Models are Great, But Industry 4.0 Needs a Map to Achieve Operational Excellence

ERIC ALLEN: Thanks for making it today. My name is Eric Allen. I'm the CEO of Environmental Intellect. I'm here to present on "Models are Great, But Industry 4.0 Needs a Better Map to Achieve Operational Excellence." I found this comic last week, and it really resonated with me, not just from this is all too often the plight of us bringing technology to big, heavy industries like oil refineries, but I think this kind of resonates with a lot of what I hear with a lot of cool innovations that are happening in industry to bring this industry 4.0.

And, just like here where this company, this hypothetical company, has a desire to bring digitization to their company, yet they clearly kind of lack the architecture, the understanding, the technical backbone to do that, I think that's very similar to what a lot of heavy industry faces today in that they may want to introduce these new technologies and innovations-- machine learning, online process analytics, all that stuff-- but the fundamental thing they're missing is a lot of them still don't understand what they have.

And, today, through the presentation, hopefully, you'll get an understanding of maybe more of that challenge-- I think how models are somewhat broken for a lot of the use cases. And, at the end, I'll go through some demonstrations of how we're adapting some mapping technologies to augment models to really solve that challenge and to bring a lot of the potential of this industry 4.0 to light.

A quick background on myself, I'm a chemical engineer by education from Purdue University. I started my career in environmental consulting. So I had the opportunity of traveling around to upwards of 100 oil refineries across the globe to help them with various work practice compliance requirements around environmental, safety, health where I got a lot of exposure to the challenges that plague these very large companies.

I'm also a self-taught software developer. So I head up our development team, developing some of these solutions. Fingers crossed, I might do the dangerous thing and do a live demo later today. So, if it breaks, a lot of that's on me.

So a quick-- so a quick background on our company, Environmental Intellect. We were founded in 2008 to bring technology to industry to solve compliance challenges. One of the things that myself and the founder of the company had a lot of experience with were work practices related to environmental inspections that occur at these refineries. And one of the industry standard ways to achieve compliance is to highlight your P&IDs.

So does anybody not know what a P&ID is? OK, piping and instrumentation diagram-- it's one of the most foundational documents in an oil refinery chemical industry. It really spells out all of the instrumentation, all of the pipes, everything related to an oil refinery. And the work practice at these plants is to highlight with Sharpie markers. Still to this day, that is a work practice.

And, years ago, with the advent of plant or Autodesk P&ID, we discovered that, wow, there's maybe a better way to do this. What if we integrate process data into these P&IDs and develop a simple tool to generate highlight?

So this was one of our first products. It still solves a lot of challenges for some customers, but it basically gives folks the ability to integrate process data into their P&IDs and generate applicability terminations automatically.

So, in the eight years that happened after 2008, we furthered that technology, and we started expanding really our capabilities at digitizing 2D assets. So we've kind of become one of the best companies at taking a plant's flat, vanilla AutoCAD or any other format drawings and converting them into these digital plant 3D drawings to digitize those 2D assets.

Now, in 2017, we recently started-- we got some exposure to laser scanning, point cloud modeling, photogrammetry, and 3D. We acquired KTMN enterprises, which had a lot of expertise in this area. And we really think that was the last puzzle piece that was missing from our mission to help digitize plant and industry. So our customers right now include Shell, Chevron, Exxon, Dow, BASF, Marathon, most of the very large oil refineries in the country and globally. Oops, wrong way.

So what are we going to talk about today? High level industry overview and challenges. I'm going to-- who in this room has been to an oil refinery before? Anybody? OK, got a couple of folks. But, for everybody else, we'll give a quick rundown of what the industry is, how they operate, and how we think oil refineries and petrochemical and oil and gas are one of the most ideal environments to bring forth these innovations.

We'll talk about data that exists in this industry. And I think it's abstracted to apply to just about any industry folks are in. We'll talk about the models that are used in this industry and the challenges that face them. We'll introduce our kind of solution that we're beginning to deploy with maps. And then we're going to go through our four-step process for digitizing and mapping a facility, go through some case studies, and then, finally, leave with some key takeaways and lessons learned.

So what is industry 4.0? From Wikipedia, you get this nice graphic. And really it means interconnection, information transparency, technical assistance, and decentralized decisions. That's a lot of really cool techno jargon, but, to me, I think this is bringing forth the concepts of machine learning, online instrumentation, process analytics, machines and manufacturing that sometimes makes a decision for itself, internet of things, a lot of really great things that there are folks presenting here at this conference on that are just it's absolutely fantastic technology.

But how do you achieve this? And I kind of alluded to this at the beginning. I think step one is for these facilities-- they have to first understand the equipment they have. And I'll speak to some examples where this management of change process and the as-builting process, the handoff from the construction phase of a project to the operation phase is very much broken. And, until we solve that problem, I think there's going to be a lot of challenges to bringing these technologies to industry.

So step two is do all the easy stuff. Machine learning, that's easy, right? So our industry-- petroleum, petrochemical, oil and gas industries-- I just kind of eyeballed an average refinery that I've been to. And this is why we think they're just the perfect testbed for bringing forth these technologies.

Assets, they're incredibly dense in assets. Average one has a quarter of a million valves. I've been to some that have a million valves, thousands of pumps, compressors, vessels, process equipment, just online instrumentation, just a ton of assets and live data feeds. Workers, in a very small footprint, 5,000 to 10,000 to 20,000 workers-- sometimes that balloons up with turnaround to even higher than that-- incredibly dense work area, 24 hours, around the clock, three shifts-- these facilities are just rotating workers in and out.

Activity, early in my career, one of the original-- or one of my first projects was at a large British Petroleum refinery implementing some work practices related to the management of change process. So I was the lowly engineer on the totem pole who had to go through 10,000 management of change documents, which I'll talk about what that is, and identify which ones actually involved piping changes because that was relevant to the environmental compliance area that they were working on.

So 10,000 plus changes in a year-- I think 10 million is probably a low number, more than that in the inspections every year. I'd say 100,000-plus maintenance activities, just a really incredibly dense work environment.

And then, on top of that, these refineries and chemical industries are incredibly hazardous. You're heating up gasoline to 2,000 degrees Fahrenheit in a reactor, treating it with hydrofluoric acid. If you're in a chemical industry, you're using dangerous chemicals like phosgene and hydrogen sulfide and all kinds of crazy things that you've got to keep in the pipe.

So, with all of that combined, I think oil and gas is a really great industry to bring forth these technologies. I don't really touch on safety and compliance, but they're very heavily regulated. And I'll talk about my models everywhere. There are just tons and tons of models at that plants like these.

So this facility you see here is actually one of our customers. Just last week, we had a company retreat in New Orleans. And anybody who has flown into New Orleans, this is one of those refineries that you get on the landing, beautiful, right there on the Mississippi River.

And I've got a captive audience. I'd be lying if I didn't take the opportunity to brag a little bit. We had our company outing and took about eight fishing boats out with the team. And this was the haul last weekend, which was pretty incredible. I think, however, I probably need to keep to software and engineering because, as much as I've been traveling to New Orleans for about six years on and off, apparently, I still don't know how to fish on the bayou, according to my team.

But, all joking aside, industry challenges and data challenges-- so this is kind of my bullet point list of the challenges that exist in this industry. It amazes me how these are such heavily designed and regulated and controlled environments, but the amount of time that is spent simply identifying and finding a piece of equipment related to a job-- how much time that takes.

I know from some of the work practices we implement where you're having to find specific inspections or go replace and do repairs on one piece of equipment, upwards of 40% of your day can be spent searching with plot plans and P&IDs and trying to identify what you're working on.

Furthermore, the time required to enter the field and look at equipment, I wouldn't say it's as difficult as going to the moon, but, sometimes, it feels like that. You've got to put PPE on. You've got to travel out to the unit. You've got to talk to operations, get a permit.

If you're doing any sort of hot work, generally, they have to go have an operator walk around the unit and sniff it for hydrocarbon leaks. By the time you finally get approval to walk into the unit, sometimes, you have to identify what kind of staging you need. You need to build a scaffold. You need to get a scissor lift out there.

Just to go look at a piece of equipment-- and I'll talk about this later in terms of turnaround planning-- sometimes can take half the day. And, when you're working with turnaround planners whose job it is to identify these work packages and plan the turnaround, we see that, typically, a good turnaround planner can maybe accomplish two, get two work packages done in a day. And that's because they've got to go out in the field, grab a picture, take some measurements, do work like that. So the time required to enter the field to look at equipment is incredible.

Worker safety and minimizing exposure, like I said, these are incredibly dangerous industries. One thing that's present in this industry and a lot of other industries is the OSHA TRIR, Total Recordable Injury Rate. And, for those who don't know, basically, if you're a company working in this industry, you report how many hours you've worked and how many reportable injuries you've had.

We know this from a lot of our customers. You're in the green if your TRIR is two or more, which, if you do the math on that, basically means, for every 200,000 hours worked, you're allowed to have one recordable injury.

So, when we talk to plant managers, and we talk to people in the safety department with that number in mind, if you can eliminate hazardous work-- and that includes, as they say, walking is work. Walking out into the unit is work. You can start reducing the number of injuries that occur in this industry. And I think that's really a big thing that resonates with me a lot.

Managing equipment up-times, short amount of time in a turnaround, planning that work, and making sure that what needs to happen within the five to seven year turnaround, which is when they shut down these facilities, is always a challenge. We see this all the time where somebody in some department knows that they need to replace a piece of equipment or improve it. But, in the midst of a four-week turnaround, somebody drops a ball on planning a work package. And they start the unit up, and they didn't get that replaced. And they've got to wait another five to seven years to get that done.

Now, on the right side, I have more kind of data and information challenges. Maintaining as-built schematics, I talked about the number of drawings in these plants. I have yet to go to a plant where they have good as-built P&IDs, as-built schematics and isometrics.

A lot of paper-based workflows, there is a handoff I would say breakdown that occurs in a lot of facilities. And, generally, there's 5, maybe 6 or 10 folks in the drafting department that always have a stack of papers on their desk. And that's a big challenge I think with this industry.

And a couple others-- communication, collaboration across roles, departments, and inside/outside contractors-- I'm sure this is something that resonates with everybody in the room. If you're in any sort of industry, I think this is always a challenge that folks have.

And then the last part is managing dozens of different databases. I would say an average refinery has between 50 and 100 different databases, whether it's inspection databases, operations databases, SAP, Primavera, process data, history and laboratory information databases. Hundreds of these databases exist in industry, but they all represent the same equipment.

There may be a hundred different references at one plant in different databases to pump 101 in the crude unit. And I think that breakdown in data siloing that happens between these roles and departments is a big challenge and a big opportunity as well.

So let's talk about that a little bit. This is kind of a diagram we've developed that spells out kind of the hierarchy of data and information at plants. And I think this is very applicable to whatever industry you're in.

The very topmost of the pyramid, the most valuable, is the asset information. And this is what I said at the beginning. I think folks don't really have-- still don't have a good understanding of what assets they have at the plant, maybe the major pieces of equipment, but, when you start getting down to control valves and hand-valve level, I think that's where there's a lot of opportunity to improve.

And, in fact, sometimes, it's not even the minor piece equipment like a hand valve. I was at a facility about a couple months ago where we uncovered an entire benzene metering skid about the size of this entire room that it was between two units. One unit thought the other operations room was responsible for that. The other one thought the other one was.

And, obviously, there were workers that were going between and taking care of it. It wasn't like an explosion hazard. But the drafting department didn't have a single drawing for this.

So it was constructed. It was built. Somebody designed it. And then they installed it. But, when they went to go incorporate that into the CAD in the facility's drafting system, there was a drop-off.

The next level down is entities. So, to us, those are maybe representations of that pump in an SAP database, maybe an inspection program record of that pump. Further down the list are tasks. So, if you look at databases like Primavera or SAP, a lot of times, they'll include tasks. Do an inspection. Perform a repair. Do an installation.

And at the very bottom is where you finally get to data. And I think there's a lot of just incredible amounts of data at these plants, but, until they're kind of put into a hierarchy that makes sense, I think there's always going to be challenges with really truly optimizing and bringing forth all these great innovations to industry.

So further explaining that, what are some databases that we look at when we go to the plant? This is sort of our kind of risk ranking matrix or a prioritization matrix. Whenever we go to a customer, they want to complete part of that pyramid and understand how to integrate data.

Obviously, I think the best representation of assets at the plant are the models. So, for us, it's the P&IDs. The P&IDs are supposed to represent all the equipment, the as-builts that are installed in the field. Sometimes, it's isometrics. I've been to plants to do that very rarely. It's facilities that will maintain and manage a fully integrated 3D model.

We go down to the task level. I mentioned Primavera. That's a big one right now with planning work packages and doing turnarounds, IMS, some other databases.

Moving down, SAP, SAP to us has kind of a combination of assets, as well as different tasks and inspections that are run through that. And then, further down, we have some different inspection databases that process data historian. That's the online analyzers that are pulling data every sometimes second, thousands and thousands of these, laboratory management system, which basically pulls in sample data and inspections.

But you start mapping out these dozens and sometimes hundreds of databases at a plant, you can really generally assign them into a matrix like this. So, when we go to a new facility, we'll generally make up a matrix like this and start talking to folks at the plant to understand and map out where is their data. Where is their information?

So, going back to maps, how do you fix this problem? How do you integrate these data systems? How do you fix the as-builting problem? And this has been something that we've been focused on for years and years is how do you do that.

And one of the things early in our history that we uncovered with AutoCAD P&ID and Plant 3D was, wow, everything that goes in a P&ID is not just a dumb block. It's actually an object.

And, for folks that know, there's a PNP ID. There's an engineering tag associated with those. The vision behind that in our early days was, wow, we're going to use that as a globally unique identifier.

Now the challenges that we faced in our early times implementing that was that, sometimes, that PNP ID changes, actually pretty frequently. And all it takes is a wrong keystroke in the CAD department or an outside contractor. And, sometimes, that engineering tab will change.

So, if you're really looking to integrate these data sources at the plant, you've got to establish a globally unique identifier. And, in the last few years since we started implementing facility-wide laser scanning and modeling at the facilities to update the as-builts, we have learned that there's really only one true address. And that's the GPS coordinates.

And with laser scanning these days, you can get down to millimeter accuracy. And, unless you're in a place like San Francisco where I live where maybe occasionally there's a little bit of drift that happens in the plates, generally, that address is never going to change. So we'll map that to the plant coordinate system. And, basically, for every piece of equipment at the plant, we can get an address.

And I'll talk about how we're integrating that later, but, basically, when I look in the consumer space, this is what we all probably did yesterday to get here-- use Google Maps. And what's the power of that is they've established a global address book for every address, my home address.

And it's quite incredible that somehow they've integrated the data about this conference in Google Maps. And it's all because they're able to map that back to an address. I don't know the wizardry that happens to make that happen. I was pretty impressed yesterday when I pulled up the address of this Casino, and that's what it pulled up.

OK, yeah, so one last thing, this animation you see here, this is sort of our model for how you begin integrating these data systems. The center part of that is to map that assets. So I'll talk through kind of how that works a little bit later.

So what are the models that exist in this industry? I'm going to kind of go in reverse. 3D models, we see those are heavily leveraged for the design and construction phase. There's no better source of information to support the building and design of a unit than a model. We get that. However, what we often see-- and I'll show a video later of how, sometimes, like I said, that as-builting is whenever the handoff to operations and facility owner operator. It's kind of broken.

Isometrics, generally, these are kind of derivation of the model. We see these heavily used in the engineering and in some of the post-construction maintenance. They show the spatial representation of actually what's in the field.

But I generally see that this is somewhat limited, especially when you're dealing with-- if you're an operator, and you're trying to make sense of your unit and understand your process flow and understand how to lock out, tag out a piece of equipment. The most important thing for you to understand is the overall flow.

Where is this pipe? Where does this relate to the process? Sometimes, that's hard to identify out of an isometric, which is where P&IDs come into play.

So P&IDs basically show the flow here. So you'll see here here's a control valve and transmitter loop. One of the challenges with P&IDs is, unlike 3D models and isometrics, they don't actually show the spatial representation. So, although P&IDs are standard in our industry, and they're actually regulatorily required, like I said, locating equipment is often a challenge because it may look like this flow transmitter is right next to the control valve, but, in actuality, that could be 100, 200, 300 feet away in a maze of pipes.

So kind of a quick overview of the models that exist in this industry-- so what are the challenges we face with models? Like I said, they're the most detailed representation of what exists at the facility and hands down make perfect sense through the design and construction phase.

But, when it comes to the handoff that I talk about to giving it over to operations and the end users, outside of maybe the design group and the CAD department and a couple of engineers, I don't see many folks using models. You get into the pyramid of workers at the plant. I would say 98% of workers at the plant aren't generally using models for their work. And I'm going to talk about some of the reasons why, but these are some of the challenges that our customers faced.

How do you maintain them cost effectively? Years ago, early in kind of our ideation of some of these solutions, I was convinced that models were what we had to drive forward with. I was convinced that we could provide the value add for a CAD department itself to maintain there as-built models. That's changed a little bit recently with some of the innovations we've come up with.

Another challenge that we see is how do you give access to folks to these models. There may be 5 or 10 seats of AutoCAD Plant, 3D at a facility. But, generally, the solutions for viewing these are some legacy tools like Autodesk Design Review.

BIM 360 is fantastic. I've yet to see too many folks at the plant using this. But, generally, if it's harder to use than a Xerox copier, you're going to find a lot of folks in operations at the plant aren't going to use these complicated models.

And then the last part is how easy are they to-- pardon me. I don't know why this has gone away. It's hiding. How easy are they to navigate and use? Again, if you're familiar with the CAD drafting platforms, I think it's very easy to zoom in and out of models and navigate around.

But, if you're in operations, a lot of times, all you know is what you see. You're most familiar with how do I operate this control valve loop. You know how it looks. You know how the procedure maybe references a picture of that. But, when you look at a model, sometimes, that's just-- it doesn't make perfect sense and especially so if you don't believe that that model is maintained and updated. So we just don't see folks using these.

So I talked about earlier model management of change is broken. I'm going to play a quick video here. The project was at a facility. They had constructed a new-- an expansion of the hydrocracking unit. So basically a new side of the process unit had gone through the design phase, maintained a 3D model, built it, constructed it.

After startup, we went in and did a very quick laser scan to identify were there things in this as-built model that the engineers handed over to the CAD department that were wrong. And we actually did discover quite a few of those. So what you'll see here-- ah, it works-- is basically we just took this scan.

And, at the top of this vessel-- and this was just one of the many instances we identified-- you'll see the laser scan we identified a couple of valves that were installed where there actually in the design model was nothing more than blind flanges. This was one of dozens of cases we identified where what the owner/operator got, if you will, was not what was fully designed. And somewhere in that process was a handoff drop where, basically, somebody didn't communicate a change.

So what is MOC? I'll say MOC a lot. So, to our industry, MOC is Management of Change. All facilities have this. It's one of the key tenets of process safety management under OSHA.

And it basically requires that, for any process change, any operational change, any procedure change at the facility, they're supposed to adhere to this management of change process, which generally is there's a requester. A department or an engineer says I want to make a change. I want to install a valve here or change a set point.

And what they'll do is they'll generally include maybe some red line P&IDs and isometrics, a description of what's going on. They'll use databases like KMS is one that we see often to basically pass this work package or this MOC around the plant to just about every department. And everybody has to look at it and say does this affect my compliance program, my operational procedure. Does this affect anything related that I'm responsible to? And, if so, then there's a process to incorporate that, say if you're in the CAD department, incorporate that in the IFC drawings or the as-builts.

One of the last parts post-installation of equipment is the PSSR, Pre-Startup Safety Review, which is you've installed it. But, before you turn that valve to turn a piece of equipment on, you get operations out there and engineering to basically walk through and check off everything that was done. A lot of times, we'll see that's where a lot of changes to models are incorporated where a ball valve or something new was installed, but, through this management of change process, facilities are supposed to not have an issue with their as-builting at all.

However, like I said, I've yet to go to a facility that has accurate as-builts. One facility I was at just recently, their solution for this is they have a team of I want to say like seven engineers and basically field drafters that go into the field. And, through this whole facility, they're rotating every five years, full-time job, walking out with paper P&IDs, basically identifying where things have changed. And that's their solution, knowing that, hey, somewhere upstream there's a breakdown in that process.

And funny story there, the human eye sometimes is deceiving. I was in fact in a meeting with a plant manager who was furious because he's paying good money for these folks. And, for anybody who's been in a facility, the top of a column, a big distillation column, a lot times will have a swing arm, which they can put rope on and get equipment up and down. And it looks like a big elbow at the top of a column. It's sometimes this big around.

These drafters basically drew that in as an open-ended line saying basically this whole column was open to the atmosphere. So the plant manager looked at that and just about stormed out of the room. This is what we see a lot of folks are doing to try to fix this as-builting problem.

So, like I've kind of mentioned before, model, why is this broken? The fact that there's only a few folks in the facility that have access to electronically request changes and mark up and draw outside of paper processes I think is a big bottleneck. Like I said, most drafters that I see at these plants have a stack of paper on their desks half my height.

Another big challenge that we see is just making sure the workers have access to the valve as-builts is surprisingly not always the case. Just about every compliance program I go to is a couple of revisions back because they've got to do their work on top of the drawings. And, by the time they do their work, they're two revisions ahead.

So, finally getting to maps, what's the solution? I think one of the solutions is to really democratize access to these models. Give folks access to models when and where they need it. And really just drive electronic processes.

I think BIM 360 Docs is an incredible solution. It's a mobile platform for folks that haven't been using it. I think it's fantastic, but, again, BIM 360 is not much more different than just an electronic version of paper. If you can't really provide value and reason to get that model to the worker, to the job site, people are just not going to do it. So that's what we're going to talk about next. Oops, wrong way again.

So maps are the solution. Like I said, they're easier to understand than models. If you're in construction, if you're in operations, the most powerful thing about a map is that little blue dot that appears in your Google Maps right now that says I am here. And, generally, I see that they're just easier to understand.

The standard map that we see used in industry is a plot plan. Most control rooms you go to get a permit will have a whiteboard that just shows the overall plot plan, very simple. And they'll put magnetic push-pins on a map, which was actually one of the original things we saw and had the idea for one of our mapping solutions.

But the fact that that is still used, a whiteboard, I think shows that, for operations and a lot of people at the plant actually doing the work, the map is what they care about the most. It's easy to mobilize, a very low data footprint, just very easy to make touch friendly-- I think that's a big barrier to overcome for some of these big, complicated models is navigating around with touch screen-- and, like I kind of talked about, more useful and relevant to workers.

So this kind of image that I'm showing here basically I think is kind of a big representation of the challenge, which is there's only a couple folks in the plant that are using I would say and updating the models. But I think most of the people at the plant are more interested in the map.

So what do I mean by-- I keep going the wrong way. What do I mean by a map? So I've been talking about Google Maps and the plot plan before. Our idea and solution for the map is kind of a combination of that augmented with some other mapping technologies.

So first and foremost is the 2D maps-- the plot plan, the high-level overview. Our current platform uses a Google Maps API just to get spatially referenced. The pros of this, like I said, are easy to understand, but one of the more important ones is that it's geospatially referenced, which, talking about surveying equipment and getting GPS coordinates, that's the easiest way to display where those locations are. And not data rich is one of the cons-- obviously, a map, very limited information.

I've talked briefly about point clouds-- incredibly spatially accurate, but very high data footprint. Most clouds that we deal with for a unit are upwards of 100 gigabytes. So there's obviously challenges in getting that out to where the worker is.

Now photospheres, visually rich, and they're very easy to understand, but there's a lot of limitations. And, when I say photospheres, that's the Google Street View for folks that are familiar. We see that eyes light up when we show folks in operations and outside of the CAD department the photospheres because they look at that and say, hey, that's the control valve. I know that.

They may look at a point cloud, or they may look at a model, and it doesn't quite look what they're looking at. But, when they actually see the picture, their eyes just light up. So that's one of our kind of tenets of our mapping solution.

And then, finally, models, models are incredibly important, and they still are to this process. And that's the fact that they're data rich. So, when we look at developing kind of a solution for a plant, we think to kind of cherry pick these solutions, combine these together.

And really, from the map, we get the ease of understanding and the geospatial reference. From the point cloud, we get just the spatial accuracy and detail. From the photospheres, we get the visually rich. And then, finally, by combining that with model information, we're getting data rich.

So how do you implement this at a facility? I've already talked about how management of change is broken, and we want to fix that. And I've kind of postulated that mapping is a solution to fix that. But how do you approach this within current budgets?

And what I've seen from some very, very large tech companies and players in this industry is they'll pilot a full facility-wide digitization initiative and mapping solution at a plant, but it doesn't fit within their op budgets. It's funded by a big tech company. They're going to go in and digitize the plant and do these case studies and pilots, but it's not really a scalable approach.

So how do you make that scalable? How do you actually fit that within the budget cycle? You've got to leverage existing plant needs to meet short-term value adds and budget cycle needs. And it helps, obviously, to have a long-term focus on the bigger goals at the plant digitization.

And this is one of the key things with I think getting this solution into more departments outside the drafting department is you've got to really focus on the end user. So you've got to really focus on what are the use cases that they need to interact with this solution. And that's how we eat the elephant.

So I'm going to walk through our four-step process for digitization really quickly. We start with digitizing the 2D assets. We then move on to digital backbone scans. So that's laser scanning and photogrammetry.

Step three is the magic where we combine those two and basically identify GPS coordinates for the equipment and revalidate, which is one of the value drivers that we have. And then finally is the fun stuff and implementing the use cases in an evergreen maintenance process.

So 2D asset digitization, generally, what we'll do is we'll take a facility's drawings and convert those to AutoCAD Plant 3D. That allows us to standardize the blocks and symbology and the format of the drawings. A lot of times when we go to these facilities, we're talking to folks in drafting or to engineering that know that they have a challenge and want to fix this through kind of a holistic approach starting from scratch, which is kind of where we start with digitization. One of the key parts of this and why we use AutoCAD Plant 3D is that it generates that master asset list very easily and seamlessly.

Step two is a digital backbone scan. So we use a combination of mobile and static scanning. A lot of times, we'll go through and push-- we have a NavVis M6, which was recently lost by American Airlines. It just about gave me a heart attack. It was recently found. So we have it again.

We'll push that through the plant very quickly to collect a digital kind of backbone and then augment with static scanning very detailed areas to kind of zoom in on what we actually need to get more detail out, especially for elevated points. It's kind of hard to push a mobile scanner that's wheeled around. For folks that don't know, one of the nice things about the NavVis that we found is how high quality the photospheres are that it collects, which supports one of those tenets of mapping.

And this is something that we see at a lot of facilities. We talk about laser scanning, and they, oh yeah, we know that for capital projects. Our focus is to do this for operations, so downstream of capital projects.

And I mean, it just happened a couple days ago. They were blown away by how quick and cost effective it was because we're focused on doing the Google Street View and not just focused on one little area, a pipe reroute or so forth.

This is the magic part, one of the magical parts. This is the revalidation process. So we've got the digital P&IDs that are all object driven. Every object is an actual object in there. We have the point cloud.

And what we'll do is we'll basically kind of do what somebody would do in the field, which is find a pump and walk out that line and make sure it's in a red line. We'll do that digitally. So we'll link each and every one of those objects to the position it is in the point cloud.

Through that process, we're able to identify, hey, there's a valve shown on the P&ID here that's not in the field. We probably should remove that in the model or in the P&ID. Or, vise versa, there's something here that's not in the P&ID. We probably need to update that. Like I said before, the value, when we're done with this process, not only does the customer get updated as-builts, but we're getting the GPS coordinates for all that, which serves kind of the backbone of this digitization solution and mapping solution.

So kind of a summary of these three steps-- the fourth step omitted so far-- we've standardized their schematics to AutoCAD Plant 3D. We've revalidated those. And we've implemented a Google Street View-like viewing application and use case for folks.

Now starts the fun part. So this was a kind of a data dump from one of our databases. I kind of prettied it up a little bit, but, basically, the concept of what we're doing, just to kind of show it visually, is, for every single piece of equipment that we have in the P&ID, we assign an address. And then, basically, in that table, we're able to link to each one of those assets what is the represented asset ID or database ID in each and every other database.

So I show a couple here-- SAP and MMS. By building this out, basically, this central table kind of for our solution becomes if you want to call it the API gateway to enter and transact data between these databases. I'll talk about that a little bit later.

I'm going to walk through three case studies. I've got some videos that kind of break up the monotony of my voice. We're going to walk through three-- maintenance planning and turnarounds, environmental compliance inspections, and global asset inventory and compliance.

So I talked about earlier what a turnaround is. One of our recent successes has been supporting a couple of facilities' turnarounds where the turnaround planners are interested in I think right now two main things. One is how do you augment the planning process to make that more efficient and make that more I guess of a standardized approach.

And, second, when you're planning a turnaround, one process unit, not much bigger than half this hotel, we're going to bring some thousands of workers in a four-week period and shut down that process unit. If you can proactively identify kind of troubled areas and hot spots and realizing that, hey, we can't do this job at the same time as there's a crane over top of us, there's a lot of opportunity there to improve that and increase efficiency and reduce schedule creep in these costly turnarounds.

So I'm going to start this video. This was the results of just one of our quick mobile scans. And what we were able to do is, after we had linked the assets to where they were in GPS coordinates, we were able to integrate to their Primavera database. So their Primavera database has their scheduling points. It references each piece. Each Primavera work order will reference a piece of equipment. We're able to just map to where that is.

And what we've done for operators here was we've basically, for some critical jobs, we've uploaded-- and I'll show the next one. We've basically been able upload nothing more than their simple PDF work package that they have right now to where it is in the map that they can search for. They can find. And one of the key benefits out of this, like I said, was a heat map.

So one of the successes we had out of this was they were interested in mapping out the jobs that they had in Primavera throughout the turnaround and not just where the jobs were, but how many workers were assigned to that job. And one of the roles of a turnaround planner is to plan a logistical area. So where are the welding shacks? Where are the trailers? Where are even the porta potties and lunch rooms?

And, by mapping out where the work was at different times, we were able to identify quicker paths and more efficient paths. If you can eliminate two or three minutes of each worker's walk time throughout a turnaround, and you've got thousands of workers, you can get a little more efficiency out of your turnaround.

AUDIENCE: Was that using ReCap software, the [INAUDIBLE] feature, to identify [INAUDIBLE]?

ERIC ALLEN: So this video was actually-- this was the NavVis IndoorViewer was the implementation that we had.

AUDIENCE: It's called NavVis?

ERIC ALLEN: Their IndoorViewer. It was pretty seamless. We didn't have to do any kind of post-processing. This was just navigating within photospheres, yeah.

One of the other benefits that we got out of this project was, for some critical path jobs-- this is ReCap-- we were able to basically sit down with a turnaround planner very quickly and find some critical path jobs. This was a demo and removal of this exchanger. And, to make sure there was positive communication, instead of depending on just P&IDs and written notes and a work package, we were able to identify all the steps that were related to that work package in Primavera.

We didn't do this for all the jobs because, sometimes, they're just replace a control valve. But, for a big, critical job like this, if something goes wrong in this process-- a contractor doesn't know exactly what they're supposed to be doing-- you're looking at schedule creep that could trickle down throughout the entire turnaround. So the key deliverable that we brought out of this was really we mapped out each and every single step of their Primavera work item for positive communication.