"Big Data" transcript
00:20 Michael Gaines: Hello and welcome to NOV Today, a podcast designed to give you a deeper look at the people and technology shaping our world. I'm your host, Michael Gaines. "Big data" is a phrase that's truly revolutionizing the world around us at an astounding pace. From wristwatches that can measure our heart rate and sleep patterns to the ever growing reality of self-driving cars, leveraging the data revolution is quickly becoming the wave of the future and pulling what was once science fiction into present-day realities. But what does big data look like in oil and gas? How is this industry using data to make smarter, more informed decisions? That's what we'll be discovering through an ongoing series of podcast episodes that will look to shed light on the people and technologies making an impact in this space. On this first episode, our big data journey starts in Navasota, Texas at the Grant Prideco Navasota Plant.
01:33 Michael Gaines: The sound you are hearing is that of drill pipe being upset. This is a process by which the end of the drill pipe is being modified to increase its resistance to fatigue and increase its usable service life. This is one of over a dozen major manufacturing steps that helps Grant Prideco, the world's largest premium drill pipe and drill stem accessories manufacturer, produce high quality products for customers around the world. Taking a tour of the facility, I was shown the full range of manufacturing development, from the receipt of raw material to the office personnel who produce the final documentation for customer acceptance.
02:10 Michael Gaines: I personally spoke to over two dozen of the several hundred employees based out of this specific location and was impressed that a facility with as many machines, people and processes could be running so smoothly. With their overall equipment effectiveness or the summation of throughput, equipment reliability and quality reaching 85%, it was evident to me that there was more going on behind the scenes than met the eye. I wanted to understand what was under the hood of Grant Prideco that made this whole operation run so smoothly. I had heard that they were using big data and not really having had a lot of exposure in this area, decided to visit Carl Fehres, NOV's Vice President for Engineering Technology, to understand what big data is and how it impacts NOV.
02:58 Michael Gaines: So I guess, I'll start out with a very simple question that may or may not have a simple answer. Could you define big data? What is big data and what does it mean to NOV?
03:09 Carl Fehres: Yeah, that's a tough question to answer, actually. It sounds simple enough but...
03:14 Michael Gaines: Yeah.
03:14 Carl Fehres: I've heard it described many ways and for me, it's this big umbrella term that covers a lot of different technologies. And so I've heard it described as the 3 Vs, or velocity, volume and variety. So if you think about data, you're talking about massive quantities of data, so volume. Velocity, which is getting tens or millions or trillions of data points every second from all kinds of places. And then variety, which basically means data from any type of data source, whether it's a website or a PDF or a report or a drawing file or an ERP system. I think when people think of data systems historically, they revert back to SQL-type technologies or Oracle ERP-type databases, where you're using some of the traditional database technologies. Big data is fundamentally different in that it can handle any variety of data type, which is pretty unique.
04:22 Michael Gaines: That's really helpful. I have never heard it mentioned that way. So volume, velocity and variety.
04:28 Carl Fehres: Yeah, I didn't come up with that. [chuckle]
04:30 Michael Gaines: Okay. Well, I wasn't gonna give you credit anyway, but okay.
04:33 Carl Fehres: If you think about where we are today, everything in the world is pretty much connected. There's sensors on every device that are out there. How many devices at home do you have connected to your WiFi? And all the things that each of those devices is generating is data and so you've got trillions of devices out there and just massive amounts of data being generated. How do you manage all that? And so, these big data technologies have been around for, I would argue, decades but have been only available to most companies for the last 5, 6, 8, 10 years roughly.
05:18 Michael Gaines: So when we're looking at big data, and keeping in mind the three Vs, I keep thinking V for vendetta, but that's a side bar. [chuckle] But when we're thinking of those three areas, how does that then apply to NOV? What are we doing in that space?
05:37 Carl Fehres: Yeah. So, well, at NOV we've got this great history of creating data. We were the first ones in the industry to create a control system for our rig. And so with the control system, you've got sensors on the rig that are basically used to determine where you're at with where the equipment is at, where the drilling process is at, and to basically control the equipment. So we've used sensors for decades for the purpose of control. So if you look at that, we've got this treasure trove of assets in the field that are generating massive amounts of data on a daily basis. And what we're looking to do now is to harness that information coming in and to use that data to derive other benefits: Productivity and performance gains in operations, improving maintenance, understanding equipment health in real time, so assessing the health of a piece of equipment as it's operating; stretching out your maintenance, so that maintenance is much more economical, but still, the equipment is as reliable or even more reliable than it is today. So really using that data for outcomes like that.
06:50 Michael Gaines: So when we think about big data and NOV's role in this space, the question immediately comes to my mind, well, how did we start? What were the beginnings and what precipitated our start in this space?
07:06 Carl Fehres: First, I'd say that there are many groups in the company that have invested and built products on big data-type technologies for the last decade. I think of things like RigSense, for example, that have been around for a long while. But two years ago, Hege and I started the group that we built out to focus on big data. And the idea was that we would foster an environment and a technology that would enable engineers to bring new products to market much quicker than if they had to do it on their own, so build a center of excellence around big data and build technical expertise around the new technologies that have come out in the last several years. So two years ago, we started building two core functions or platforms. One was Max, which is our big data infrastructure platform. We're focused mostly on the Industrial Internet of Things type of data, which is again, sensor data. So we've got this massive infrastructure and platform, which we can use to aggregate data and do lots of things with data like analyze and calculate and present and visualize. And then we use that technology to basically build products on.
08:23 Carl Fehres: The second core component that we've built out is called the Access Portal. It's a new NOV web application that's customer-facing, that allows customers to log in and interact with applications that interface with the big data platform. For example, you've got Rigsentry BOP, it's a real time monitoring and analytic system for BOPs. And so the customer, our users, need to interface with that data, with all that data, to look at trend information, what's coming up maintenance-wise, what's happened in the control system that may have caused some kind of an issue, to do diagnostics, that type of thing. And so that user interface is the Access web portal. The Access web portal is, it's built on modern technologies. Again, it's mobile-friendly. And it interfaces with the engine, which manages all the data and the analytics themselves. And so Max is the back-end engine and Access is the front-end user interface for the customer, essentially.
09:27 Michael Gaines: Okay. When we're talking about what we should do, and we're looking at NOV and looking to the future, maybe not so much what should we do but what can we do, what does the future look like for NOV when we're looking at big data?
09:44 Carl Fehres: Yeah. That's a great question. I think we're still searching it out, but there's no doubt that artificial intelligence comes into this. It's the latest buzzword that sits on top of big data. Artificial intelligence is really powered by big data, if you will. It's in our roadmap, and you read a lot of industry buzz and news about this. A lot of it's science fiction still, even though it's marketed as here today. A lot of it is actually here. So if you look at the prediction models we built with the Rigsentry BOP product, we use artificial intelligence inside of that. So there's some of it that's real in industry and there's some of it that's really future-thinking.
10:28 Michael Gaines: Rigsentry, NOV's answer for predictive maintenance and condition monitoring of blowout preventers, is a technology that is using big data to help customers improve their bottom lines and operational efficiencies.
10:41 Nick Morriss: So on a land rig, BOPs sit underneath the drill and rig. And in a lot of ways, they're out of sight, out of mind.
10:49 Michael Gaines: This is Nick Morriss, New Product Commercialization Director for rig solutions.
10:55 Nick Morriss: Where that gets amplified is offshore, particularly in deep water, where they are underneath the rig floor and they're also potentially 12,000 ft of water, beneath 12,000 ft of water away from the rig floor. So what that means from a maintenance and a criticality standpoint is that, in order to do any maintenance on that BOP, you've gotta bring it to surface. That's a job and, depending on water depth, can take a number of days to get it to the surface. And so, when an operator is paying a day-rate for an asset and it takes several days just to get to it to maintain it, then you've got the maintenance time and then you've got the time to get it back to the sea bed, it puts a criticality on maintenance that's much greater than most other assets on a drilling rig because of the time associated with maintenance offshore.
11:51 Michael Gaines: So, given that level of cost impact to operations downtime, how does Rigsentry fit into the picture in answering or addressing the pain point of customers, especially offshore with regards to their need for increased uptime, and a level of understanding of when they really need to bring the BOP to surface 'cause you don't wanna do it just because the calendar says you have to do it, I imagine.
12:26 Nick Morriss: That's right. And that's where a lot of our customers in the industry is today. We do calendar-based maintenance on most of our assets. And that's where Rigsentry comes in, is to try to address a transformation from a calendar-based maintenance program to something that's smarter. And there's a number of ways that can be done. It can be done by performing maintenance on a usage-based level instead of a calendar. So an equivalent to, let's say, the automotive industry there would be, you don't change your oil 12 times a year, you change your oil based on the number of miles that you drive your car. So that's a usage-based example of maintenance. Likewise, Rigsentry has ambitions to offer, and in some cases, offering the ability to do predictive type maintenance. If we go back to the car analogy, an example would be that your car tells you when it needs an oil change, you don't go by a calendar or by your odometer, you go when a light tells you to get your oil changed. So, Rigsentry is a platform based on big data, where we are trying to accommodate or, I should say, we are accommodating those different types of maintenance philosophies.
13:42 Carl Fehres: So now having a better understanding of how Rigsentry works and what its value proposition is, is this something that we have already deployed in the field already? Where is it in terms of its offering position?
14:03 Nick Morriss: Yep. We have a lot of experience in the condition monitoring space with customers. We have, I'd say, around a dozen legacy condition monitoring contracts active today and we're in the process of converting those over to various levels of Rigsentry, what we call a soft rollout. It's very, I would say, customer-specific the way that this platform is rolled out. And so we've started that process on a customer by customer and also on a tool by tool basis. So, we've focused today here on the BOP, there's also Rigsentry suites of products that are available for top sides equipment, so top drives, mud pumps, draw works, critical path equipment on the top side of a rig. We're primarily offshore at the moment, but we do have ambitions to expand that in terms of equipment and also categories of rig, so into land rigs, into fracking equipment, and so forth.
15:07 Carl Fehres: Back in Navasota, Heat Treat Manager Michael Nelson described the temper furnace, a process by which the drill pipe material is initially heated and then rapidly cooled as a part of the overall heat-treat process.
15:19 Michael Nelson: This bar is 1600 degrees when it hits that water. Do you see how fast? It's not moving very fast, but actually this is moving pretty fast. This quench right here puts out about 2400 gallons per minute, and most of that 2400 gallons per minute is right there within 2 feet. So that puts a lot of water really fast, and that's how you get the quenching. Just like... I'm sure you've seen people heat stuff up, just stick it in a bucket of water, they don't even realize what they're doing, they're quenching it, making it harder. [chuckle]
15:51 Michael Gaines: Right. Right.
15:54 Michael Gaines: This massive machine is one of the critical components that must be expertly maintained. And using sensor information and big data produced by this machine can truly help keep this vital process running. But, as we all know, a system may sound great in theory, but sometimes may prove to be different in reality. That's why I wanted to talk to Larry Ritchie, Grant Prideco Navasota's Plant Maintenance Manager, to understand from a real-world perspective how using NOV's approach to big data helps the plant in achieving its goals. Larry shared that there was a benefit to the impact that the changing market conditions had on how he and his team viewed equipment maintenance.
16:36 Larry Ritchie: I think at the time there was so much profit and work being done that the concern about the equipment and the cost of the repairs, whether it'd be catastrophic or not, were not a concern. Just get it and so we can keep on producing the pipe, we got the money, let's go. So, once that slowed down and the profits got smaller then we gotta find a way to do this better. And the way to cut cost in my element is to look at the equipment and treat it as something you take care of continuously, to monitor it, understand what's going on and why we need to do what we're gonna do. Unneeded repairs, unneeded preventative maintenance measures like changing oil annually, stuff like that, those all incur a bunch of cost in the end and if you don't pay attention to the cost of doing business, then the cost of doing business will eat you up when it comes down to, when the profit margin's not so large.
17:32 Larry Ritchie: So for me, we started out first by putting in place a preventative maintenance system, whereby we do checks on equipment every 90 days. And when we started there, what we started to realize that there were certain conditions that we would find that we could address smaller problems before they became much bigger, and cheaper. At first, it seemed intrusive to some of the operations. I'm not gonna stop [18:06] ____, my job is to get the numbers out. And I said, "Give me a second and you can produce a lot longer between those times." So I used the AIMS system, it's Asset Information Management System, and what that does, it does report out. So any time a piece of equipment goes down, there's a work order open, my mechanics go fix it and then it also would do the preventative maintenance side of it, the reliability side of it, where I can do timed or planned events, and they automatically tell me when those times come up, 'cause we're looking at 200 acres across the street from each other, 44 buildings, roughly 1,800 pieces of equipment. So, one individual trying to look at that on a paper trail is very difficult.
18:55 Larry Ritchie: So, what the AIMS system does, it gives us that data and we can reformat it, report it, and look at it, and then we can also do the analytics as performance measures. How many times has this equipment failed over a period of time? And how many times does it fail? At the end of the month, what I do is I lay out all the equipment downtime and I narrow it down to... I take the average of a score of how many times it's been failed, how much money we spend on it, how many times it's been addressed, and I give that score and I average that out. And anything above that average, I address. So, we go down, we pay closer attention to it. And what I'm finding is, as I do that, the equipment performance indicators, that average, that number grows. So I'm getting better at doing it and I'm letting the data take me to where I need to fix and concentrate on. I'm not making that decision myself. I don't need 30 years of experience on a welder or whatever, to know what I need to fix, the data tells me.
19:53 Michael Gaines: If you're still wanting to know more about big data and how our company is innovating and growing in this space elsewhere, we'll be coming back to this topic again in future episodes.
20:02 Michael Gaines: Thanks for listening to this episode of NOV Today. We'd like to hear your feedback. Share your thoughts by tweeting us @NOVGlobal and using the hashtag, #NOVToday, or you can contact us by sending an email to [email protected] For NOV Today, I'm Michael Gaines. Thanks for listening, and we'll talk to you later.