How ExxonMobil inspects oil & gas infrastructure using drones
In the dynamic and challenging landscape of the oil and gas industry, maintaining the integrity and functionality of expansive asset infrastructure is paramount to support strong reliability and performance throughput.
This talk, led by Shankar Nadarajah, Global Drone Inspection Lead at ExxonMobil, will delve into specific challenges faced in overseeing remote inspection of vast infrastructures and how ExxonMobil is addressing these challenges using aerial robotics.
Learn about the evolution and benefits of oil and gas workflows and critical response through the use of aerial robots, including:
- Standing up a drone inspection program and process for getting executive buy-in to scale inspection operations.
- Illustrating the value of aerial robots for inspection of oil and gas infrastructure vs. traditional methods
- Enhancing inspection workflows with higher quality insights powered by AI-enabled aerial robots.
- Expanded uses for innovative technology solutions outside of inspection, such as critical response and security
- The future vision for persistent aerial inspections with dock technology and remote operations.
Learn more about Skydio Solutions for Oil & Gas Operations.
- Appreciate the invitation. Um, I'm the Global Drone Inspection
- Lead for ExxonMobil. Uh, been with the company 22 years.
- Uh, so been in this particular role for, uh, plus or minus
- about two years or so. Um, and my previous
- background, I have, um, you know, started off in IT,
- technology, networking, uh, servers and data centers, but then
- quickly moved into a wide variety of assignments outside of my comfort
- zone, uh, everything from HR to chemicals, sales,
- um, and into manufacturing and refining, supporting our, um,
- our large maintenance turnaround and execution.
- So, uh, these are maintenance turnarounds where we
- have anywhere from a few hundred
- to a few thousand people on site, um, for a period of time, about 30
- to 60 days, taking down a unit and putting it together
- and getting it pre-
- prepared for, um, you know, um, operations again.
- Uh, so yeah, no, appreciate, uh, the opportunity.
- Thank you for the invitation.
- Yeah. Thank you for, for joining us today.
- Uh, so, uh, being part of the solutions engineering
- team, uh,
- allows me, uh, to go into the field, um,
- uh, such as with Sha- Shankar and, and perform some really amazing
- autonomous, uh, aerial robotic tasks in, in
- complex environments. Uh, so we'll be covering some of those
- successes, uh, that we've had in, um, the local
- Exxon refineries. Uh, so Shankar, he will, will first,
- uh, cover some of the, the high level Exxon overviews,
- and then we'll
- get into some specific, uh, Skydio captures.
- Yeah. So I think, uh, just to kinda kick things off and, uh,
- kinda set the scene
- for the meeting, right? Just a little bit of background around,
- um,
- ExxonMobil and our entry into the, into the drone operation space,
- and really why did that happen. Um, we're gonna talk a little bit about a-
- an acronym called GAVIS in the next slide, but it stands for Global Aerial
- Visual Inspection Service. And, and really it was, um,
- it, it was really this whole, uh, strategy that helped
- us launch where we are today in the areas of,
- um, scaled out drone operations for the
- corporation.
- Um, uh, essentially the, the, the business problem
- that led to this is, is really the, the fact that we had a
- very interesting technology back in early as even 2015,
- 2016, but it was mostly used for hobby purposes.
- A lot, not a lot of commercial use. Um, and there was just a lot of,
- um,
- uncertainties around what you can and can't do with a drone on a, on a
- commercial property, especially in oil and gas facility
- that has a lot of different
- various risks and hazards, um,
- to manage. And, and so that's, so that's where we, you know, we really
- needed to have a, a program in place to help us work through
- that.
- And so we'll talk about that as we go from the left to the right
- of the screen.
- Um, first and foremost, we wanted to understand where can we apply
- this
- technology and where do we see, um, the need?
- And a lot of this has been, um, for areas where we have
- assets at heights, and you can see some of those examples on
- your, um,
- on the top right-hand corner as w- which looks like pipe racks.
- We've got reactors, furnaces, flare stacks, and then you've got
- confined space entry. So if you look at those spheres, uh, over
- there, and we have to, if we have to inspect those
- and get in the insides of these,
- um, these assets to look at the integrity on
- the in- on, on
- the interior, uh, that is where we're getting into some of these
- confined space
- scenarios, where again, there are a l- number of, you know, safety and, and
- personnel risk mitigation there at, at play that we wanna manage
- through the use of this technology, [smacks lips] and, uh, not to mention, you
- know, better quality data. So this, this, um,
- GAVIS program, um, really helped us set the stage to
- help standardize, uh, how we wanted to, uh, expand this
- program globally. We, we are now having this, um, I'm
- proud to say, in over 100 plus ExxonMobil
- affiliate locations and sites around the world where we
- operate.
- Um, and so we've got some deep penetration of
- the technology, and
- we're, we're, we're... It's paying dividends.
- Um, I, I do also wanna acknowledge, though, that a
- lot of this program was only successful, uh, due to, due to our
- partners. And so when I look at our partners and, and team
- members there, there's a lot of, uh, a lot of people here.
- Uh, first and foremost, we took a strategy
- and approach to scale quickly.
- And to do that,
- we established some really good strategic third party drone operators
- that were vetted. You know, it really ensured they had the right,
- you know, safety
- metrics and performance in that area, but also had the right level of
- equipment, the latest equipment, and knew actually how to operate it and
- produce good quality results. 'Cause ultimately, data's the
- prize at the end of the day for ExxonMobil.
- We wanna be able to extract those insights to help, um,
- quickly put together remediation repair plans or, or, or planning
- purposes for projects and, and future, um, up- updates to our
- equipment. Uh, the second piece of that also is the,
- the data management side.
- We wanted to make sure that our folks in, uh, Baytown, Texas, at our
- largest integrated petrochemical plant were looking at the same kind of
- format and results as our partners in our
- upstream con- conventional space in Australia.
- And so that way we, we worked clo- closely to screen and selected, um,
- a company called Hover Data based out of Austin, Texas.
- We just had an excellent quarterly visit with them yesterday, and
- they provide us the backbone to help us manage, organize,
- store, analyze, and really produce the, the, the,
- the high-quality reports that our inspection teams around the world
- can
- collaborate and solve the business problems with.
- And, um, last but not least, the two things I wanna cover are there...
- is our aviation department that's been instrumental in partnering with this GAVIS
- program to ensure that we've got the right operators, um,
- who are meeting all the aviation operation guidelines that
- ExxonMobil Aviation has in play. Um, and, uh,
- also our community of practice, our drone COP-
- Who are es-essentially business champions at the sites around the
- world that are all, you know, are, are, are, are
- spectators and also are main users that are taking this technology
- deep into the sites, promoting awareness, and coming back and
- sharing with that, through that feedback loop, where
- are the o-opportunities for
- improvement? What is the new technology we need to
- start looking at to solve
- other business problems that a current drone technology can't
- solve today?
- So they've been
- awesome to work with, and we meet on a quarterly basis as well.
- So we have a, a huge sounding board, uh, to help us feed to
- the COE, the center of expertise group here at
- headquarters in
- Houston, Texas, how we want to position the strategy
- and the roadmap
- for our drone platform going forward as we go into the future.
- So again, just kinda going back to...
- Like I told you I was gonna tell you a little bit more GAVIS
- and, uh, it stands for
- Global Aerial Visual Inspection Service.
- I don't wanna say that 10 times fast, so that's why I came up with an acronym.
- But we have a million acronyms at ExxonMobil for [laughs] everything from
- our
- projects to our services. So it's definitely one we have to, um,
- yeah, sometimes use an online dictionary to k-
- [laughs] keep up with.
- But, um, the, the, the things that drove, again, GAVIS, and I think it's
- important for anybody to take a step back and understand as they
- think about starting their own drone program if they have not got one set up at
- their site, is just first understanding where the,
- uh, opportunities, uh, to move the needle in terms of taking, uh,
- hold of some of these challenges, and understand how you can break down
- those b-
- silos and barriers to, to push towards value.
- And for us at ExxonMobil, value captures everything.
- Um, GAVIS really paved the way for us to
- organize
- how we would structure and scale a, a large-scale drone program by
- first identifying where the opportunities are.
- So we found out that
- from, um, an, an inconsistencies perspective that we
- had different things happening at different sites before.
- Um, we had sites that could operate a drone
- and had people with Part 107,
- you know, uh, certificates to, you know, to effectively
- manage and use those drones. We had other sites that
- were learning how to spell
- drone. So we really wanted to close that gap by putting a
- standardized program in place and having the tools, education,
- skill set to
- do that, and really leaned on third party, uh, you know, in the short
- term especially. We do have in, in-house site-based
- drone programs, especially for our emergency response
- where you can't call a
- third-party drone operator that quick to go address a major incident.
- Um, so we have hybrid teams set up, but ultimately, you know,
- going third party for us was, was ideal to help us move the needle
- quickly and, and capturing that value and, and, uh, taking advantage of those
- opportunities on the table. And also it helped us in the regulatory
- front,
- 'cause the civil aviation agencies around the world have different, um,
- rules and regulations where w-what you can and can't do with a drone
- for commercial purposes, and how high you can fly it and, you know,
- um, you know, what kind of sensors perhaps could be on
- there.
- So we really leaned, again, and working with our third parties who had
- very
- close relationships with the civil, respective civil aviation agencies,
- to help us understand and quickly know how to stay in compliance
- around the world
- where we operate. Um, and that helped a lot.
- One thing you don't see on the slide that I wanna also add is the fact that
- we also identified where the opportunities to save money, right?
- So we... You know, you can't put a price on a life, so,
- you know, safety was
- paramount for us in terms of moving forwards and having a drone program to,
- you know, really de-risk, uh, uh, some of the work we're doing in some of
- the most
- dangerous, hazardous areas where we're operating.
- But also, we looked at cost of scaffold reduction,
- motorized
- equipment, rope access, um, you know,
- and, and, and it goes on and on, right?
- So, so there's a lot of, uh, you know, the areas that we can reduce our costs as
- well and be more efficient in how we do our inspection work around the world.
- And then last but not least, I can't even stress the, the quality of
- data and the order of magnitude
- of change we saw in that, right? We went from binoculars and
- digital cameras to like, you know, to high quality
- 1080p4K, you know, uh, high resolution
- video. Uh, so that has also helped pave the way in
- producing
- very precise, um, very accurate, um,
- uh, data that we can now
- produce some better insights from.
- So you may be asking, "Well, what's next for ExxonMobil now
- that we've got a
- program, it's scaled out to 100 locations around
- the world?
- What is going on with GAVIS in 2024 and
- beyond?" And we're gonna talk a little bit about that, especially with some of the
- next slides that Sterling's gonna be covering.
- But we ultimately now are, our, our key word and messaging,
- uh, for our drone program at ExxonMobil is automation.
- We are looking to, you know, continue to, um,
- really get polished in terms of our ability to digitize that
- en-end-to-end inspection workflow so somebody can hit play
- from the time they want that drone to go out and inspect a,
- an asset, to
- the point the data gets transferred, um,
- um, quickly, timely, and processed so we can
- produce what we see a, a more visual way of working. Okay?
- So today you're looking at images right now in these slides that are, you know,
- you
- know, 2D, kind of flat, and you can see clearly maybe some of
- these
- anomalies and defects we're looking at.
- But we really wanna start getting towards a 3D view of how
- we're, um, looking at this as if somebody was walking
- inside the plant
- doing that inspection, but now having the ability to see this
- asset like you see on the screen from either the very top, the middle, or
- the bottom. And to do that, the drone technology, especially
- from Skydio, has really helped us to produce
- that level of
- view. And more importantly, ExxonMobil is now
- very much into looking for ways on the basis of
- theme of automation to bring in some of that data, ingest
- it, and have it processed by AI and machine learning algorithms.
- So we don't have to manually do that or put that o- onus on our
- third-party drone operators to do that as well.
- So we're taking, you know, that into our hands and working
- again with, um, select and screen third
- parties. Um, we're continuously looking at this space.
- AI is definitely new. There's a lot in-- going on in
- this area, and so we're taking a very pragmatic
- approach to making sure
- that we find the right parties to work with.
- Not only them, but also our internal parties to support
- our AI, AI
- machine learning efforts to automate some of this part of the analysis.
- And looking for those damages and defects, and quickly being able to tell the
- inspector where exactly on that pipe rack you see is that
- corrosion or that, or that paint loss or, or,
- um,
- or missing components like bolts.
- So these are some of the things that, um, we're very keen on in our
- GAVIS two point O journey.
- Think with that being said, I'm gonna pass now this, uh,
- baton back to, to Sterling, who's gonna take the time to kind of
- talk about some real specific examples of where some of the Skydio
- technologies have played a very, uh, crucial part in
- helping us on that journey that I just mentioned to you.
- And, uh, Sterling, I'll let you take it here--
- take it on from here.
- Yeah. Thank you so much, and, um, definitely appreciate again the,
- the opportunity going back, it feels like several years now,
- when we first
- started a lot of these projects and initiatives.
- So,
- um, going out into the field, I remember the first time, uh, you
- basically gave us five to six assets to, to cha-
- uh, as a challenge to tackle, um, for a proof of concept within
- a day. And that really, you know, set the tone, uh,
- not only for the initiative, but also the success criteria
- that we followed
- and, and really hit in a successful, uh, way.
- So,
- um, as I approach any situation drone-wise,
- um, I,
- I do, you know, site assessment, standard job safety
- analysis, risk mitigation, hazard assessment.
- Um, and with Skydio's autonomy,
- that really resonates within obstacle avoidance and also our three-D
- scan skill sets. So previous to Skydio AI,
- individuals and pilots such as myself would have to be, you know, intimately
- familiar with, with the aircraft, um, how it operates, how to safely
- operate in different environments, and then pair that with the
- knowledge to capture inspection or photogrammetry-grade
- datasets.
- I can proudly say that, you know, I wouldn't discredit all of that ten years
- of experience I had, but anyone on this call or in a
- crowd with basic understanding of our UI,
- uh, can go into complex environments and using
- Skydio AI, collect really, uh,
- complex, detailed datasets that can be used for a
- variety of, of, of use cases. So with
- that being said, I was confident and successful in, in
- going into a completely new, challenging, complex environment
- with Skydio technology and was able to really showcase to
- Shankar and his team on-site, um, how we're able
- to collect data in, in a new form within their
- facility.
- Uh, so today we'll be seeing a couple, um, slides here of,
- of some models, uh, raw photos, and,
- um, images as well paired with some GIFs.
- Uh, so we are able to set up an area of
- interest, um, within,
- uh, augmented reality in three-D scan.
- And from there we have, uh, an exploration phase where the, the
- aircraft is actually going to explore the assets within that
- defined space, and it will intelligently store and create
- a model, an environmental model of that asset,
- uh, that can b- then be used for, uh, ground sampling
- distance,
- uh, overlap, uh, different scan pass settings, and
- all of that will be maintained in flight while the aircraft is
- intelligently interacting with those assets.
- And so I again, can, can confidently say that, um,
- when it comes to collecting autonomous data, there's no other tool out
- there, um, like this that provides the level of safety,
- the level of
- detail, and the level of coverage and quality, uh,
- across your, your drone organization with very little
- to almost
- none, uh, in terms of skill required entry for, for
- scale of these operations. And in terms of the data collection,
- I'm, I'm at the point now where, you know, it's just-- it's, it's almost impossible
- to beat the AI. Uh, I would have to spend so much time
- on-site manually capturing the impact that I would have,
- and then consistency of, of, of the data is I'm, I'm
- trying to make the drone interact with, with the environment
- manually where in Skydio's AI, it's intelligently
- aware and interacting in real time with that asset.
- Um, so that's how I was able to just, uh, streamline through
- these types of, uh, complex structures and within one
- day, uh, capture five different, uh, pretty
- large, uh, assets to high inspection
- quality and even crossing into that three-D, uh,
- reality, uh, uh, model, uh, for, for
- post, uh, analysis, design work, et cetera.
- Um-
- So just looking at the angles that we're able to, to achieve here.
- And just to take a quick pause and, and throw a question to you, Shankar,
- uh, in terms of the data collection that we're able to do with, with Skydio,
- how, how is that fitting within, um, your normal
- day-to-day inspection workflows, whether that's terrestrial, aerial,
- internal on some of these assets, or LIDAR-based?
- Um, have you seen an impact with Skydio data
- in helping not
- only blend, but also bring some of that data together?
- Yeah, no, ab- absolutely. I think, you know, one of the greatest things,
- again, and
- this is ag- again, going back to what I said around having a
- foundation to
- start with at your, uh, respective corporation on how do you
- ingest
- data from different sources and, and being able to have
- kind of a unified platform, again, regardless of where you are in, uh,
- the world we're operating in, to have a, a standardized platform to bring in that
- data, whether that's coming in from a Skydio,
- uh, feed or another,
- another source to help w- like what Scott, uh, what Sterling was saying, is
- blending that through. And so we've been very fortunate to have
- a, a great team
- here within our, um, uh, ExxonMobil, um,
- uh, digital organization called, uh, the Digital Reality Ecosystem, and the
- dr- we call it DRE, D-R-E. And the DRE team has been instrumental
- in, in helping us on that journey and making it easier for folks like
- Sterling at Skydio and then the rest of the team to bring in
- this data, um, and
- being able to blend it in with, uh, you know, CAD and o-
- and other models so we can
- start to superimpose and start to see as fit.
- And if you look at the accuracy of, you know, some of the, um, the models
- that Skydio's able to scan, I think we're, we're very, very happy with the, with
- the, uh, accuracy of those results.
- And, um,
- uh, for me, it's just the fact that I can now turn this model
- any which way and have a very good purview into some of
- the details of, uh, some of the anomalies we're looking for.
- Um, that level of detail to me, you know, we, we historically was never able to
- get without having to put in some, you know, additional infrastructure and
- support. And so this is a, a nice, neat way of really getting a
- high-quality product without having to put in like a, a, a very
- tremendous and costly infrastructure to support it.
- Thank you.
- So going on to some, some other assets that were captured.
- Uh, this is a, a fairly large one, um, in terms of,
- you know, some of the instances that we're able to discover, um,
- post-flight, post-process here.
- You can imagine in, in the area, we've actually had some pretty large
- storms roll
- through South Texas in, in the last weeks, and even damage, uh,
- significant damage to downtown Houston, right?
- So we'll just use that as a, as a mock here.
- So, uh, being able to step out into the field, uh,
- with the Skydio
- aircraft and quickly set up an autonomous mission, uh, and
- capture the entire asset, full coverage.
- And again, because the aircraft is intelligent of its environment and is
- actively interacting with those assets to maintain those settings
- that we as a user can define, uh, for quality or coverage.
- Um, and then being able to then quickly assess
- those through,
- uh, Gavis and those structures that Shankar has highlighted today in
- the post-process environment. We can look at in finding, you know,
- damage to insulation, potential, um, insulation that
- was, you know,
- removed from, from post-storm.
- Uh, but again, as, as, as Shankar said, we have now the ability to
- look at this asset in 360-degree
- unrestricted space, and this is just the, the reality model.
- There's, you know, tons of, of raw data that can be
- used in various capacities. So all of our metadata,
- EXIF data, telemetry, it's all open, uh, for us to stay
- agnostic. Uh, so this one dataset can be processed in
- many different ways. Um, we do have streamlined Skydio
- Cloud integrations and also direct partnership integrations,
- and have
- recently launched Skydio Extend for professional
- services on that back-end integration side of things.
- So
- again, being able to step out into the field and,
- and easily capture,
- uh, a, a complex structure like this is possible with Skydio.
- Yeah, I think as, you know, Sterling, it might be interesting to add, like,
- just to
- give, like, just an order of magnitude o- on a model like this, like how many
- actual images this is when you stitch it all together.
- I think on this particular furnace, I remember Skydio,
- um, you know, X2, I think shot around, like, 2,000 to
- 3,000 images.
- Mm-hmm.
- And then it was those images that were then carefully, you know,
- processed properly and stitched together, um, to build this model.
- So, you know, for any given area we wanna hone in
- and look at the
- actual photo, the high-res photo, we can do that
- now with, um, you know, our, our, the platforms that we have, uh,
- working with Skydio.
- Yeah, I mean, that's, that's a great call-out 'cause, you know,
- some may think,
- wow, that's a lot of images. Well, [laughs] imagine trying to do
- that manually.
- Uh, you could capture 2,000 images on just the
- stack or the
- front side of, of the facade of, of this structure, right?
- So that's 2,000 images roughly of, of full
- coverage end to end as, as Shankar mentioned, as any surface
- that is recognized by the aircraft during exploration and in
- flight will be captured. And we actually have tons of augmented
- reality for you in AR observers and overlays to
- actually see and view insight into the
- autonomy, uh, for overlap coverage, areas that, you know,
- may or may not be, uh, fully overlapped, and
- also manual capture workflows within that as well, if you see any significant
- damage.
- Uh, I will say, uh, you know, there is a high level of autonomy here,
- but there is
- also a tremendous amount of flexibility for the user, uh, during
- setup, overlap, distance to surface, uh, the ability to
- pause the mission at any time. Uh, you can move to maintain
- visual line of sight. Uh, you can pause to, uh,
- let's say, conduct a safe operation to allow a
- non-participating
- personnel to move through. You know, this is an active refinery, right?
- So full control of being able to pause an autonomous mission, take
- action, and resume, uh, back in is completely possible.
- Um, and, and that- those number of photos
- are then the highest coverage while,
- while being, uh, fully efficient on time-wise to
- provide, uh, data sets like this that we're seeing on the screen.
- This is another angle here.
- Excellent.
- So then even moving into smaller, but yet still complex
- en- you know, environments, uh, we're able to, to look at, you
- know, our, our smaller aircraft even still today is,
- is quite relevant in,
- in the pipe rack inspection or more confined space.
- Our obstacle avoidance is 360 degrees, roughly one meter from
- propeller tips. And so the S2+ is a little bit smaller, so that
- footprint, um, for, for standard obstacle avoidance will allow
- us to, to
- operate in a, a pipe rack like this a little bit more
- freely than an
- X2 or an X10. Uh, so during th- this day we really
- put the S2+ to, to the challenge and, and found some really interesting data
- on, on live and decommissioned lines here,
- uh, captured. Uh, it might be a little challenging to see,
- but, uh, in terms of
- all the, uh, photo positions and refreshment points here, green
- highlighted and raw imagery to even a shot from,
- from one of the models here on one of the decommissioned lines, I believe so.
- Jumping into,
- um, what a 3D model, uh, can look like
- from a pipe rack inspection. I think this was actually, uh,
- one of my mo- more
- favorite inspections on site, uh, 'cause there was an active,
- uh, inspection happening not, not too far off from, from where we were
- flying, uh, of, of an individual actively going
- through this complex and challenging area to work in
- to, to inspect it. And I'm sitting there thinking, "Wow,
- if only he had, you
- know, the 1,000 images, 2,000 images, or even 100,"
- depending on what we're scanning
- in Shankar's end-to-end workflow to where he could have just
- reviewed,
- uh, maybe the, you know, top 10 key highlighted points for
- focus and targeted those versus actually having to go out
- and find what's wrong first. And so that, that was, uh,
- r- really resonating with me in the field as
- the S2+,
- uh, was flying through this entire complex
- pipe rack capturing,
- uh, a couple thousand images within a short
- amount of time.
- So in terms-
- Yeah
- ...
- of, uh, you know,
- you know, not necessarily quicker, but better quality, better
- coverage, uh, the autonomy and scale across your
- pilots, um, that you're able to, uh, grow your network
- within data collection.
- I would like to just ask, you know, like, what type of impact does this data
- here have on decision-making within your
- different business units within Exxon at a, a refinery like this,
- Shankar?
- Yeah, no, I think, um, you know, Sterling, good question, right?
- I mean, if you look at, first of all, um, just take a step back and just take a
- look at that, uh, that asset right there.
- That is a very detailed pipe rack. Limited, um,
- you know, limited distance between each
- and every other one of those pipe racks, so
- getting in and out of that area is already a risk in itself.
- A lot of it's done, you know, typically by rope access teams.
- Um, and, uh, and, and a lot of them would have to use, like,
- a pole on the
- stick kind of setup and have to manually go through there,
- like Sterling was
- mentioning. So, you know, speed to market, right, for us is being able to get
- a, a very good perspective of that asset using, you know,
- the Skydio S2 drone technology that has a collision detect and
- avoid like no other, uh, to be able to properly and safely
- scan that, uh, entire pipe rack for us, um, to a point where
- we can perform the basics of an API 571
- inspection and, and kinda go through this
- and look at some of the key
- anomalies and pain points that we're, we're focused on, you know,
- especially corrosion under piping, uh, we call it CUP
- inspections, as well as other, other, other parts of the, um,
- checklist that we have for pipe, uh, pipe rack inspections.
- And so,
- uh, you know, for, for us, this is, this is a game changer, um,
- essentially to be able to have a, a robust technology like this to quickly
- scan the asset, get the, the level of quality we need to
- do that analysis, and then porting this, you know,
- high-quality data through, um, AI and machine learning algorithms-
- Mm-hmm
- ... to speed up the identification of those
- anomalies and defects, and
- then localize them so I know where on this pipe rack these defects
- even are. Um, that is, you know, as if, you know,
- we're in there in the field doing this ourself, and that's, that's to me the, the
- big game changer right now. Um, we're remotely being able to
- inspect a lot of this stuff.
- Excellent. Yeah, so we, you know, highlighted there from capture into the
- field
- through our integrations into third-party ecosystems,
- visualizations, and post-process down to individual
- reporting
- for, you know, in, in-field inspectors to, to focus on those
- areas. I did wanna highlight, you know, I don't have the, the camera positions,
- um, on, on the screen here, but, you know, the aircraft was actively flying
- through- That support structure in the entire
- entirety of that pipe rack. Uh, and we do have an unrestricted
- gimbal across all of our aircraft at Skydio, so that does allow for, you know,
- straight down and straight up, uh, unrestricted limitations
- during capture. So in, in terms of overhangs, under,
- under portions of the piper ra- pipe racks to straight overhead nadir
- shots, we're able to, to capture and coverage, uh,
- cover those areas.
- So in terms of, you know,
- proactive
- of, of, of data capture, being proactive in data capture,
- where do we stand, you know, moving forward if, let's say, someone
- was to go
- out and inspect this pipe rack tomorrow?
- Are we looking at capturing drone data prior to looking at
- historical data that was captured, if it exists, um,
- or moving forward with that day-to-day inspection?
- How, how do we view that happening today and,
- and in the future, Shankar?
- Yeah, I mean, I think this is the beauty of, uh,
- what we're doing right now in
- terms of digitizing, you know, our, our inspection
- workflows. So the historical stuff that now that we're capturing, we can
- use it, uh, to determine, um, for example, change
- detection, right? If we identify, um, you know, an
- anomaly or, you know, from their, from the data that was captured and
- analyzed, especially from our AI algorithms,
- we'd wanna be able to measure whether
- that be, you know, the severity of a crack or severity
- of, um, s- specific corrosion that's taking place,
- um,
- understanding the rate of that change.
- And, and so now we're just being, you know, more laser,
- you know, precise on exactly some of the key areas
- we're looking at, and so we
- don't forget and have to, like, lose sight of what we have checked,
- you know, a
- prior inspection because our inspection intervals are all different
- depending on equipment type and-
- Mm-hmm
- ... and age and whatnot. So, um, this is just giving us a,
- just a level of transparency now across, uh, a
- wider, uh, duration of time so we can now piece this
- story
- together and, and then really start to build a
- predictive maintenance strategy.
- And I think that's, um, you know, y- you think about the, the, the
- size of these plants, the amount of equipment we're h- we're ha-
- we're having to
- maintain, having a way to drive a predictive
- maintenance
- strategy and, um, repair strategy using these kinds
- of technologies is gonna be, um, you know, a, uh,
- really a game changer to help us pave the way on how we're gonna, uh,
- run our operations and our sites on, on, on a future basis.
- So it's-
- Yeah
- ... um, definitely a lot more intelligence going
- into this now than ever before.
- Thank you.
- All right, moving on. So in terms of forward thinking, as, as
- everyone knows, we have, uh, recently, uh, released our,
- our next generation aircraft, our Skydio X10, which, which really
- resonates, you know, more powerful autonomy paired with higher
- grade, uh, sensors and modularity.
- Um, so we have a, a tremendous amount of runway for
- use with an X10, uh, from RTK.
- Uh, we have Night Sense. We have, uh, parachutes coming
- out, right, for operations over participating or
- non-participating personnel,
- uh, and also the ability to upgrade into
- different sensor packages
- as they're released. So in terms of forward thinking and
- scale and adoption, uh, autonomy is definitely going to be
- key in those conversations as we continue to look for
- more flights, more data, uh, and more, um,
- accountability within inspections, uh, and progress
- monitoring, et cetera, based on that data.
- Uh, so with that being said, definitely, uh,
- uh, proud and excited to, um, talk about some of the remote
- operations that we are starting with Exxon,
- uh, in terms of Skydio Dock operations within
- their facility,
- uh, which allows for, um, some pretty groundbreaking
- regulatory work, uh,
- within Skydio's professional services offering in partnership
- with Exxon in order to achieve the proper
- waivers, um,
- and meet the right regulation in order to a- achieve the
- end result and desired data.
- And that, uh, is possible through Skydio hardware, our
- autonomy, our obstacle avoidance, our safety,
- and our deep and, and intimate understanding of,
- of how the FAA
- operates. Uh, we're able to take Shankar's, uh, full,
- you know, program documentation, standard operating procedure, and
- historical drone operations and really resonate that into brand-new
- technology and, and start doing this, uh, next
- step into the desired fully autonomous workflow
- here.
- Uh, so just take a moment to, to ask you, Shankar, of, of really how you
- see, you know, some of those inspections
- that we did in the field across
- five assets, you know, in one day to now in a remote
- operations environment,
- um, being able to have scheduled missions, predetermined
- missions, and then also the ability to have a
- multi-use asset, right,
- where you're inspecting assets during the day.
- You have incident response when needed and also security as well
- across many different use cases as these are online
- connected, um, you know, IoT devices, flying
- sensors. So there's a lot-
- Yeah
- ... a lot of different,
- uh, values that we can add when we,
- uh, go and scale up into full autonomous
- remote operations.
- Yeah. No, you said it well, Sterling. Absolutely.
- I mean, if you see what we've gone through the
- slides and we went from a specific
- asset or assets to now we're looking at a site-wide approach in how we do
- inspection. I think the fact of having this
- Skydio Dock technology or just drone-in-the-box technologies in
- general
- just give us an ability to do a lot more in terms of, uh, periodic-
- Um, inspection, and it really touches on this entire
- predictive maintenance
- strategy, right? Of being able to be more in a proactive
- mode than
- a reactive mode. Um, you know, maybe we don't have an
- interval per se that we have to inspect certain assets, but maybe there's a, you
- know, a heavy rainstorm hits and we wanna go and send our, um,
- you know, um, drone-in-a-box solution here, uh,
- into our tank farms and be able to look at the tank roofs
- and just check and
- see that everything's looking safe and sound from an integrity perspective
- after
- that rainstorm. That's just one of many examples where
- we're looking at this
- from a, you know, um, kinda site-wide perspective
- in terms of identifying use cases that will return value
- for us and maintaining, you know, strong performance and
- reliability of our entire operations.
- And I think this example here is kinda showing you a footprint of a,
- a typical
- refine- integrated refinery chemical plant set up.
- But we also have operations, as you guys know, in the Permian Basin,
- which is our
- unconventional business, that is...
- Got a lot of, um,
- different, um,
- uh, equipment classes and types, but also it's spread apart.
- It's very much spread apart for those who've been to the Permian Basin
- and then, and, and witnessed, uh, all our well pads and, uh, other, other
- companies' well pads spread apart like that.
- You've just got a lot of coverage to, to, to maintain and to inspect.
- And having a technology like this to help us automate
- some of that inspections, um, is gonna, it's gonna help us greatly.
- Uh, today we're, you know, we, we have to have people driving in pickup
- trucks
- and high, you know, heavy duty vehicles inside, um, these
- locations, and have to travel a far distance to get
- to the next
- asset to do that inspection. And so we're trying to, you know, really
- mitigate the, the need for some of that extended travel and
- the risks associated with that with using technology like this.
- And so we're real excited about just where the future is about
- automation and being able to use kind of Dock technology, or
- drone-in-a-box technology, to support that, um, across our,
- um, asset fleet and sites, um, uh, as we, as we move
- across the world.
- Excellent. Really appreciate that.
- And yeah, I've, uh, had several [laughs] trips out to the, the
- Midland-Odessa area and, you know, it is few and far between, you know,
- driving out to each of those sites.
- And you can imagine, you know, with our, our professional services offering, both
- in the regulatory space and also in API and
- integrations, being able to have a, uh, a
- Skydio aircraft and a Dock environment that is fully integrated
- to where you actually have a triggered, uh, alert through your SCADA system.
- Instead of triggering a, a truck to roll out, you've just,
- uh, had an automatic API call that
- initiated a scheduled mission or a predetermined mission that
- flies to view that gauge or that sensor that could be in question.
- So I've, I've had several very fun, successful POCs in that capacity as
- well.
- Um, so when it comes to, you know, remote operations here, again, I
- truly do believe in, in the c- in the, uh, in the safety
- and scale that Skydio can offer in
- beyond visual line of sight, VO, non-VO, skills-based or
- performance-based, uh, beyond visual line of sight remote
- operations. So,
- um, it's, it's been really fantastic to see the
- growth.
- Very excited, uh, for expansion, um, within
- different use cases and the improvements in these environments
- and
- data collection with X10 as we've offered now,
- uh, much, uh, more robust sensor packages,
- 640 radiometric thermals, telephoto, um, lenses,
- narrow and wide, and also the modularity of it.
- So, so quite ex- excited to explore, uh, new
- opportunities and data collection types, uh, within your,
- uh, facilities, Shankar. So really it's been a fantastic journey
- that we've been on, again, starting in, in the drone space in
- 2015 and, and teaching myself every day how to be
- the
- best drone pilot, to now teaching others about autonomous
- drones. It's, it's been really a f- fantastic journey, and I would not be able
- to go into any environment, any complex environment,
- and feel so confident if it wasn't for Skydio.
- So
- let's, uh, hit the next slide here.
- And I think we're pretty good on questions
- time.
- Just wanna make sure-
- Yes. We've saved-
- ... we're not a little early or late.
- No, I think we've saved plenty of time for
- questions.
- Um, just as a reminder to everyone, if you have not already done so, to
- ask your questions, please use the Q&A box and hit
- send to submit them. And let's see.
- Our first question here, "Is Skydio's AI
- ready to fly in dark indoor spaces?
- For example, inspections of furnaces from the inside?"
- It's a great question. So first, you know, before a Skydio
- X10 was released, I'd be thinking about, how much space do we
- have? How much lighting do we have, right?
- Well, with Skydio X10 and, and our, uh, groundbreaking
- product that goes with it, uh, a payload accessory, NightSense,
- we actually can physically illuminate our f- our flight area,
- both in the infrared spectrum or the visible light spectrum,
- with a module on top and bottom of the aircraft.
- So now that question really is, how much space do we have to operate
- within
- those environments? And Skydio's standard obstacle avoidance,
- again, is one meter, 360 degrees from
- propeller tips, and that's driven by those six onboard
- navigation
- cameras that feed our neural network and process those world
- iterations and allow us to be intelligently aware of our
- environment. So to, to directly answer your question again,
- through Skydio
- NightSense- Both in the infrared or the visible light spectrum,
- we are able to actually have obstacle avoidance at night.
- It is, uh, very groundbreaking,
- um, and one of the most advanced features, in my opinion.
- Awesome. Thank you. Um, and are the deliverables
- for reporting from a Skydio software or
- from a third party?
- Yeah, great question. So that is going to be in that
- partner ecosystem, right? So Skydio is, is an amazing tool
- that provides the solutions and the software and the hardware to
- capture
- the data. Um, and then there we have, again, more
- software and integrations in Skydio Media Sync, Skydio
- Cloud, Skydio Extend to port into
- existing or new workflows, uh, for individuals such as
- Shankar or businesses like Exxon to seamlessly
- integrate. So, uh, if you wanna quickly, Shankar, hit on, again, just
- hover or some of the, uh, softwares that you use to
- integrate and, and how you do that with Skydio data, I think
- that would be great.
- Yeah, no, I mean, I think in general, you know, we've,
- we've got a wide variety of
- different platforms, uh, you know, to help support different business
- problems we're trying to solve, right?
- So not every inspection ne- necessarily requires per se,
- um, certain, certain features and capabilities.
- So we, we are very, uh, you know, careful in how we wanna manage that,
- but that's
- the, the beauty of having that level of flexibility
- with a foundational platform that gives us that restaurant menu to
- choose when
- we wanna use something that requires a 3D model versus what
- needs, uh,
- some particular AI machine learning algorithm to scan that first, right?
- So I think that the, the unique thing is here is having the ability to have the
- data, um, you know, easily to be, uh, shifted and
- managed without having, um, too many, uh, you know, chefs in
- the kitchen, as I call it, and having that data smoothly come in
- and, and then be able to kinda go in- into the different platforms we choose
- depending on the business problem we're trying to solve. So yeah.
- That's one of my favorite things really about, uh, the backend of, of
- the API and, and Skydio Cloud ecosystem.
- And, and I hit it earlier on, on the, uh, call is
- one dataset can, can be used for an inspection team
- through an AI machine learning and processing workflow,
- uh, for reporting. And that same dataset could also be used for
- modeling and integrations with terrestrial or aerial-grade
- LiDAR, uh, for design teams, uh, you know, registering
- and, and recognizing as-builts to design.
- Um, so we have that ability to seamlessly port
- one
- capture through many different processing engines so that your, uh,
- end-to-end clients or your users are working in that
- data ingest area and not the drone capture
- data workflow, processing workflow.
- Uh, so we have, uh, again, some pretty streamlined workflows to where
- you're living
- in that end-to-end goal versus having to worry about the
- drone and the capture and the processing.
- Great. And someone else is wondering, how long did it take
- to capture the 2,000 images on the furnace, and
- how many battery swaps?
- Yeah, great question. So,
- uh, we do have the ability to break up
- missions, really.
- You just set them up and then also think about maintaining
- overlaps, same settings, et cetera.
- So I broke up the mission for a couple of reasons.
- One, maintaining visual line of sight, radio connectivity,
- right? And then also coverage and efficiency, and also being able to
- resume the mission back into those complex environments.
- So I did, I believe it was a total of, um, about three,
- three and a half batteries, so about 30 minutes of flight.
- Um, so we were able to break that up pretty quickly and, and capture a
- large structure like that, uh, quickly.
- Keep in mind, I was out there, um, doing a
- POC for both inspection, uh, showing the
- technology, but also being able to walk away with a
- high-quality model,
- uh, and also be able to capture five assets
- within that
- day. So, uh, with that being said, there's a t- a
- tremendous amount of flexibility outside of default Skydio,
- uh, 3D scan, 2D scan settings that a user can enjoy.
- Uh, if you are kind of heavy in photogrammetry
- and you understand a
- lot of those settings, then you'll, you'll f- be quite excited to see that
- flexibility in setting up only a Z-axis only or X and Y,
- adjusting all of your scan passes, stop for
- photo. Uh, there's a tremendous amount, uh, that a, a
- higher,
- um,
- uh, end user, I guess, uh, more historical-based in
- photogrammetry would find quite useful to
- change in terms of inspection versus a pure
- end-to-end photogrammetry dataset used for reality
- models or design work, right? That, that different initiative would
- change a little bit if Shankar just said, "I needed, you know, the chimney stack,"
- or, "I need, you know,
- uh, the latest photo collection," versus a
- full end-to-end inspection
- workflow or a reality capture. That flexibility is totally
- designed within the system for the user to define number of
- images, number of footprint of overlap, et cetera, with
- insight prior to starting the mission.
- Great. Thank you. Um, and do you operate the drones in the
- rain? And what gust or wind speed can you expect at
- the oil platforms, and how do the drones cope with that?
- Yeah, great question. Um, you know, all, all of this information is, is always
- paired within our Skydio support articles and,
- uh, overview and manuals just for y'all to
- have as reference.
- But we'll find, um, about 25 to 28 miles an
- hour with the X2.
- Uh, S2+ in terms of wind tolerance that we suggest, and, and really
- that's in a comfortable range. We have to think about flying in new
- environments with closer proximities and autonomous action, so we have to be
- cautious of how the aircraft is behaving in
- winds that are higher than that, right? Much higher than that.
- I actually did spend, uh, several days, uh, offshore,
- um,
- and we, we had a, a safe metric of about
- 26 knots, uh, to maintain, and we experienced a little bit more,
- but we were able to,
- uh, mitigate and perform safely within those
- environments.
- So both onshore and offshore, uh, everyone knows the Midland,
- Odessa area can get quite windy, et cetera.
- Um, so for current dock landings, right, we're about that 30, 30 mile an
- hour threshold for current dock.
- Uh, the X10 is a little bit heavier.
- The aircraft itself can fly 45 miles an hour with a 40-minute
- flight time. Uh, so it can handle, um, more wind and
- is more robust with an IP55 rating,
- um, which does allow us to fly in that inclement weather, dirt,
- dust, debris, rain. We do have to keep in mind
- that, you know, we do have computer vision, so if we do,
- uh, impede our computer vision, we do have those workflows defined
- in terms of disabling obstacle avoidance and listening to GPS or vice
- versa, uh, for tho- those environments where, um,
- again, it's inclement weather and, and our vision navigation sensor
- may not have a reference, right?
- So hopefully that answers your question.
- I did do a, a nice, uh, orthomosaic the other day, um,
- in, in some, in some light to moderate moisture,
- and it,
- it was, uh, quite an enjoyable flight as some other
- aircraft were
- left in their case.
- Thank you. Um, and someone else is asking, um,
- "Shankar, referencing your 100 locations that are
- using drones, how does that translate into number of
- drones and pilots?"
- Yeah. So, uh, we actually don't even track that, um,
- because again, like I said, our model from the start is, you know, to
- scale quickly. So we partnered with, you know, third-party strategic
- drone operators that we had vetted through our ExxonMobil aviation
- departments, um, globally. And so for that matter, you know,
- we, we don't, you know, we, we actually are, uh, quite lucky in a
- sense that we have a, a interesting model that lay, allows us to
- scale up or scale down. Um, and so these, these
- strategic drone operators are essentially the ones that'll help us,
- uh, with the latest equipment and even more pilots as needed,
- depending on the demand. So yeah, we don't track that, that
- information to tell you the size. Now, our internal teams are different,
- but, like, those, those are far, far than few, right?
- Because they're more focused around certain, like, areas of the world
- that maybe
- are very, um, uh, remote and, uh, depending, and having a lot of
- dependence on a third party would be challenging.
- Uh, but for the most part, we, we are heavily scaled out with a third party
- s- uh,
- support model for these 100-plus locations around the world, and we
- rely on, on those, those, those great partnerships to, to help us
- manage our program.
- Excellent. Thank you. Um, and how has drone
- technology improved safety standards during
- inspections at ExxonMobil? Can you provide specific
- examples where drones have mitigated risks?
- Yeah. I mean, I think, you know,
- like I said before, uh, some of the biggest areas where we're
- deploying
- drone technology is into confined space locations where you typically have,
- you know, um, whole, whole watch monitors and you,
- you're gonna have all
- kinds of, um, you know, you know, PPE requirements to be able to
- enter some of these, these, these, um, these assets internally to do those
- confined space inspections. And so a lot of that, you know,
- carries a lot of safety risks, right?
- And so when you have personnel inside the vessels, and so now having
- robots be... uh, in this case, drones being able to do
- that really kind of,
- um, you know, de-risks the situation in many ways
- and also reduces the
- nes- all the associated costs of having all
- those additional
- services like, you know, the PPE req- requirements and then
- also the, the whole watch m- uh, uh, requirements, motorized
- equipment, um, and scaffolding for,
- especially when you start looking at, um, assets at heights.
- Your scaffolding costs will quickly, uh, amount to quite a bit
- depending on how high we have to get up there and inspect some of these assets.
- So a lot of that is, uh, mitigated, and on top of that, some of these
- inspections are done online. You know, in the past when we have to
- send personnel, we'd have to take the unit down,
- especially on some of these flares
- and stuff, and now a lot of our flare stack
- inspections are done li- uh,
- online while the unit is up and running, and we can safely, um,
- be able, be able to capture all the, the high-quality data we need to build a,
- a proper, uh, model, uh, and look for those anomalies while
- the, you know, the unit's, uh, the flare's, uh, on.
- So, um, you know, these are some of the big things that support,
- um, overall value capture because, you know, every, you know,
- every minute our
- equipment is down, you know, we're, we're not, we're not producing product,
- and we're not then therefore producing any earnings.
- So this is a way for us to continue our operations safely,
- uh, as well as be able to understand the
- mechanical integrity of, uh,
- of our asset structure.
- And has... Let's see here.
- Sorry, my screen [laughs] just adjusted.
- Um,
- and is Skydio inspecting offshore installations as well?
- Yeah, I'll take that one. So specifically with Exxon, uh, we,
- we haven't done
- anything offshore that I'm aware of.
- Correct me if I'm wrong, Shankar, in terms of-
- Sure
- ... of something I may not be aware of.
- But, uh, we have had other instances,
- uh, again, I had mentioned that I ha- I had spent several days offshore,
- um, in other clients that are utilizing the technology offshore.
- Yeah.
- You'll, you'll find that there are a lot of similar assets offshore,
- just more in
- a condensed space.
- Yep.
- Excellent. Thank you. And someone else is wondering, are FAA
- permissions or waivers ongoing and permanent, or
- do you need to apply regularly?
- That's a great question. So definitely, you know, a conversation for the Sky-
- Skydio regulatory team, but we definitely look, um, both from a,
- uh, operations perspective, a historical record of your
- organization, um, and what you plan to do, what your
- intent is, and de- determine in which pathway is
- best for you to achieve that goal, whether it's skills-based,
- performance-based, um, BVL- BVLOS
- operations, VO, C mode, et cetera.
- There's a lot of different pathways, again, depending on the environment that
- you're in, the operations that you're trying to conduct, and
- if you have any
- historical records of operations that are deemed, you know,
- safe and valid within the eyes of the FAA.
- Uh, Skydio, again, professional services offering, uh,
- architects
- that entire workflow, um, and defines those waivers, which
- usually are good for a, a much longer set of time.
- You know, it could be two to four years where you'd
- have to then do a
- resubmit, uh, for that specific waiver or any
- type of exemption. Again, would, would love to, uh, you know,
- get your contact information and, and discuss regulatory services at
- Skydio further.
Watch Time
56 Mins 55 Secs
Related Webinars
Demo: 1 Remote Operator. 2 Inspections. 1 Response. No Field Dispatch
Watch webinar
Pre-Incident Planning for FIFA: Turning Drone Data into 3D Operational Readiness
Watch webinar
Citywide drone response in action - Lessons from Wichita
Watch webinar