Last year, we reconnected with Zacc Dukowitz from Flyability, John Delp from AECOM, and Larry Barnard from Chevron to understand what it means to define, build, and then scale a drone program in energy and utilities. During the interview, they discuss how to present the benefits of using drones from an end-user and manufacturer's perspective to key stakeholders to initiate a drone program, how to measure and communicate the value of a drone program, and the use cases that are most valuable in the Energy and Utilities sector.

Working with AECOM for more than 27 years, John Delp is a DCS NA Chief Unmanned Aircraft System (UAS) Pilot and a level one UAS thermographer. He has multiple years of experience in infrastructure inspections and extensive experience in both government and commercial UAS flight operations. He has led UAS projects all over the country for critical infrastructures such as wastewater facilities, earth and concrete dams, and many bridges, leading to extensive experience in FAA waiver filing and airspace deconfliction for clients all over the country.

Larry Barnard has been with Chevron for more than 35 years now. Currently, he is the Product Owner for the company's manufacturing segment, which is within the downstream and chemicals organization responsible for drones and robotics in inspections and various other use cases.

As Content Marketing Manager at Flyability, Zacc Dukowitz talks to people like Barnard and Delp every day to understand how Flyability's Elios 2 drone is used in the field. He writes case studies, press releases, and articles for Flyability, but he also works as a contributing writer at UAV Coach. Since 2016, he has written almost 500 articles on the drone industry for UAV Coach alone.

While this interview is available for free as one of our Podcast Episodes, below is a quick run-down of the major takeaways.

What were the original needs that led a business like Chevron or even a smaller business to use drones? How are they looking into drones as a tool, and how were they thinking about this when they first started that program?

Larry Barnard: It didn't start out that way. Like many industries who have implemented or developed an in-house program versus wholly contracting out for services, Chevron started with somebody who was an aficionado - somebody who was interested in the tech or who otherwise utilized the technology outside of work and tried to bring it inside. But moving forward to now and how it's being used, it's very much about reducing risk hours and increasing that safety margin by doing historical tasks slightly differently. It's about gaining a level of efficiency over how we've done things in the past. There are a lot of challenges after you've identified how these robots might help us. It's marketing, funding, training, permissions, and more. There are a lot of swim lanes associated with getting it from, "I got an idea" to "We have a mature program that we're getting regular value from."

John Delp: Back when I first got into it, I was asked to go out and photograph a particular sign structure in the middle of a river. It was about 200 feet from the shoreline, so I thought to myself, "If I could just fly my camera out there and get the perfect shot..." That's what started the venture. AECOM has been involved in drones for a long time, mainly from a military perspective. However, it wasn't until around 2015/16, when there were rumors of the 107 coming in, that we started going into the commercial side of things. It took off quickly from the construction documentation sector and then infrastructure inspection, which is a primary focus of our group.

Zacc Dukowitz: Drone analysts' market sector report from last year looked at how a lot of commercial drone programs start with an excited person, who does it on their own, and then they're like, "There's a lot of potential here." That's a cool story, but it also highlights a problem with adoption. You know the technology can be powerful and what it can do, but you face cultural issues. It's not that you can't prove it, because it's been demonstrated several times over, but sometimes you have hesitation with cultures, where people say, "We don't need a new a new solution. We've always done it a certain way". But having that one person who says, "No, I do this on my own, I know what it could do," that's often the starting point.

And what does that mean for adoption?

Larry Barnard: The first time that I brought this to folks, at my business unit in the manufacturing sector, the refining sector, I got a hard no. They had no bandwidth to look at or otherwise test this technology to help them. If it's not on the business plan for that year, you're probably not going to get a lot of permissions to move forward, but that's also the challenge. You need to brush up, you need to get it in front of people, you need to do a better job of potentially selling what it can do, and you need to be honest about what it cannot do.

John Delp: Once you can get past the "It's cool" and "How much does it cost?" phases, your project managers see what is possible and what it can produce; it opens up their eyes. Once you've proven that, it can just go off on its own from there, and once you start brainstorming all the different ideas, your possibilities are endless.

For those who don't have that kind of background but are very interested in what drones can do, they might not know how to bring them to their business. What are some of the key things that you found were very successful, beyond putting it in their hands, that helped make that return on investment case to define the value these drones can bring?

John Delp: Start small. The Elios is an awesome tool, but it does come with a nice little price tag. Don't run out and buy an Elios, don't run out and buy an M300. Don't run out and buy these larger airframes. Go out and buy a $1,000 or a $2,000 airframe, and start doing some pretty pictures, some marketing, some video. Show what drones are capable of, and once you have proven that, whether it's to clients or your internal project managers, then you can start expanding.

Larry Barnard: There's a lot to unpack there. There's no one solution; it depends on the environment, what type of vehicle you're using, what airspace you're in, and more. It helps to have a technology-forward kind of sponsor to help promote it at the higher levels of the organization or to get it into that business plan, so it gets funding and appropriate support. Apart from John's recommendation, you can also try contracted services. Somebody who does this for a living may have a waiver in your location, and they may be able to write one or otherwise get access to the airspace quickly. They have the equipment, you don't have to guess what to buy or bring, and maybe you experience some benefit or have that "aha" moment. Once you see the bill starting to add up, or when people start talking about the cost of services, you start to realize you might be able to do this yourself.

Talking about the ROI of using drones—if you don't have to build a new road or stop traffic, or shut down your asset to inspect it, that's a huge return on investment. But what are some of the other things you discovered along the way?

Zacc Dukowitz: I never knew about any of this, but the alternative to a drone for sewer inspections is sending someone out in scuba gear. That's not ideal, and sometimes it's impossible since you might not be able to send a person in because of noxious gases, or the water flow is too powerful and dangerous. Otherwise, there's also the possibility of using a raft with a CCTV camera and hoping it's pointing in the right direction. If it gets jostled around, that's just the footage you happen to get, so you don't have any control. Or you can put a huge robot in there, and then you face access issues.

John Delp: Getting that robot in and out of the sewer requires a crane to be set up on-site and multiple people. With a drone, you're typically talking to a third person from the client to open the access. It cuts down on the crew, setup, and downtime.

Larry Barnard: There's a lot of ROI in many aspects of the business. One of them is the reduction of costs regarding extraneous equipment that we may need, like scaffolding. If you think about this, a typical flare stack at 170 feet to enable inspectors or repair crews to get up to the top is going to run you $3,000 to $500,000. You're going to run it for every day that it's installed, and you're going to pay for the folks to erect it and then remove it. There are significant costs associated simply with scaffolding. When you look at these process units, the expenses they bring to a company every working day adds up significantly. So, every hour you can shave off of a maintenance turnaround scheduled for a particular unit, or an individual piece of equipment within that unit is worth real hard money. Additionally, regarding the environment and our permits to operate, utilizing these robots to give us a better look at the assets and understand what's going on, what might be a problem that is not a problem yet, provides a significant amount of ROI.

Do you believe drones' capabilities have been oversold? Are there any opportunities for improvement?

John Delp: A drone is just a tool for visual inspection, and it's never going to replace hands-on inspections. There are various levels when we go out and do bridge inspections, and I'm not an engineer, I'm not a certified bridge inspector, I'm a pilot. For instance, if we're doing a cable-stayed bridge, the rappelling apparatus that the team would have to use to check the cables can damage the housing on the cable. Using a drone to see the condition of those cables and pinpoint where the problem is, makes it easier to address that particular area versus repelling every cable. However, you're never going to be able to replace that hands-on inspection.

Zacc Dukowitz: We see a lot of claims ahead of the actual capability. Delivery is like an obvious place where we'd see a press release about a single delivery made. But the questions here are "Do you have a drum program up? Do you have a delivery program? Are you doing it every day? Is it providing value?" - and I think we're still seeing that. For me, the place where we're overselling is autonomy, but it's also a place where there's just huge potential. The idea of having that tool that goes and flies and gets you your visual data, maybe other kinds of data down the road, and then comes back, is very promising, but we're not there as much as we'd like to be.

Talking about AI and Machine Learning (ML), how are these technologies helping right now? And what are your thoughts about them in the future?

Larry Barnard: ML could be very beneficial. However, the challenge that we've seen so far is that you need an excessive amount of imagery showing something bad. And we do too good of a job of maintaining our assets that we end up not having a lot of pictures of stuff going wrong. That's a real challenge with ML, and it's almost like we need an algorithm to train our algorithm. It's a lot of man-hours to sit there and say it's a banana, it's not a banana, it's a banana, it's not a banana. Once that improves, I think you'll see it getting adopted more quickly. One of the things we're seeing that is very interesting is the combination of ML with GIS tools, so not only can you scour or automate the inspection of the images or detection of anomalies within the image, but you're also able to show where that image was taken. For an oil and gas asset owner or a refining site, those two things together could be very beneficial. If you think about power poles, identifying and setting up triggers and mechanisms to alert you to a potential problem based on an ML finding is much more valuable if you can show that anomaly's location. But I think there's a long way to go with that.

Looking at the drones themselves and what companies have been asking for, especially in the Energy and Utility and Infrastructure space, how has the drone industry responded to that?

Zacc Dukowitz: At Flyability, we have been on a closely supported journey by our customers, who always give us data and feedback. The first iteration of the Elios presented a lot of limitations. Flyability didn't try to imagine what people needed for the second iteration. We went and talked in-depth to people like Larry and John and went out in the field and said, "Okay, we have this drone that can fly indoors. What do you need as an inspector? How would this become a tool that would be useful for your needs?" From those conversations and a lot of research, the Elios 2 rolled out with many new features. We created a drone that was very use case-specific. I would never tell you to buy our drone for a cell tower inspection. If you're flying outside, and need to zoom in from a distance, maybe don't pick our drone - it can get you really close to something, you can even roll along the surface, but other tools do other jobs much better than our drone would. And we're starting to see a lot more drones made for specific purposes like the Wingtra for surveying and mapping. We're starting to see these flying robots solving problems instead of just as a cool flying toy.

Speaking of solving problems, we don't usually just talk about the drone anymore. We're also talking about how drone data's collected, transferred, and processed through software to make a data product for o a client. How has the industry changed how it deals with data?

John Delp: The drones and sensors are getting better, which means more data. For example, I collected 26 gigabytes of data in six hours, and our client, in most cases, doesn't want that. We provide the client with whatever deliverable they are looking for. That may be a digital twin, elevation model, CAD files, etc. The raw data will remain on our network unless the client specifically asks for it, which doesn't usually happen since it is too much information to handle.

Larry Barnard: When you either initiate an in-house program for your company or utilize contracted services to provide drone operations, you immediately realize the volume of data generated by this hardware. Then you start to think, "How are we going to make sense of all this data? How long do we store this data?" Suppose you're talking about storage and how to classify or tag that data. In that case, you're talking about partnering with somebody like Amazon Web Services or somebody else to get some help with that storage. All of those are real challenges and problems that need to be solved regarding these drones and robots' data. I haven't seen a one-size-fits-all/one-stop-shop solution for that, and I think it'll be a while too.

Zacc Dukowitz: I noticed that every person has a different solution for the data problem, which implies that there aren't any good solutions, right? Fliability did a webinar on Data Management, and we had one guy who'd made a proprietary system - he made the whole thing to do exactly what he wanted, but it took a huge investment. And we also had another one who just kind of shoehorns three or four different pieces of software. I almost feel like we will have autonomous drones before we have the data solution solved.

What is the best piece of advice you can give to your audience, whether they're starting from the beginning or they're working through building out their program in industrial facility-type space?

John Delp: Start small, build your program up, get smaller airframes, get the experience that you need, get the buy-in from senior management that is showing the return on investments. Make sure you're building a solid program, have a UAS ops manual written up that you're following, have all your protocols and checklist guidance in that manual. Have some fleet management software where you can track your flights, pilots, airframes, and payloads so that you have a well-established program, and you can document it and show the return on investment.

Larry Barnard: If somebody were to ask me for advice, I would say to get the right tool for the job, have a hybrid approach where you're not fearful or concerned about, or there's no pride associated with having your in-house programs with your coworkers operating the technology versus an outsourced trusted contract service provider. Be prepared to do some marketing and promotion. If you're the person who has brought this technology in and you're confident in the value it will provide your company, you need to be willing to do the long hours. Have discussions, inject yourself into team meetings, including teams that you don't belong to, and do a presentation to help them understand the capabilities you're trying to bring in. Aside from having a sponsor on a leadership team, or other decision-making body, that's always a good idea to have somebody who has your back when you're not around by the water cooler. Somebody that could push you into a direction and share with you what your leadership team might be looking for or what might be the concerns moving forward so you can either address them or otherwise mitigate them. When you look at commercially available UAS versus things you can buy at Best Buy, the technology works very well; it's reliable but not wholly reliable that you don't have to worry about it. It's also essential to have somebody who manages that program and is ultimately responsible for the airspace, the use of that technology, and the other employees at the facility who might be driving, biking, or walking around the area. Nothing kills an internal drone program more than incidents.