Oral-History:Roland N. Horne

About Interviewee

Dr. Roland N. Horne is the Thomas Davies Barrow Professor of Earth Sciences in the Department of Energy Resources Engineering at Stanford University, and Director of the Stanford Geothermal Program. He was formerly the Chairman of the Department of Petroleum Engineering at Stanford from 1995 to 2006. He is best known for his work in well test interpretation, production optimization, and tracer analysis of fractured reservoirs. To date, he has supervised the graduate research of 37 PhDs, and 120 research MS students. He is the recipient of the Lester C. Uren Award from Society of Petroleum Engineers (SPE), as well as the John Franklin Carll Award. Horne was elected a member of the U.S. National Academy of Engineering in 2002, and in 2007 he was designated an Honorary Member of SPE. Horne is the 2010-2013 President of the International Geothermal Association, Technical Program Cochair of the World Geothermal Congresses 2005, 2010 and 2015, a Guest Professor of the China University of Petroleum, and has served as adviser to Texas A&M University, University of Auckland, and Universiti Teknologi Malaysia.

About the Interview

Roland N. Horne: An interview conducted by Amy Esdorn for the Society of Petroleum Engineers, October 28, 2014.

Interview SPEOH000117 at the Society of Petroleum Engineers History Archive.

Copyright Statement

The content of this oral history transcript, including photographs, audio and audiovisual clips, and biographical information, and are intended for noncommercial educational and personal use only. Copyright restrictions apply. Commercial publication, redistribution, or use of the content is not permitted.

Anyone wishing to use the text, image, or audio or audiovisual files associated with this oral history transcript for publication, commercial use, or any other use not expressly permitted, must contact the SPE Historian.

Society of Petroleum Engineers Historian
10777 Westheimer, #1075
Houston, TX 77042

It is recommended that this oral history be cited as follows:

[Interviewee Name], interviewed by [Interviewer Name], SPE Oral History Project, Society of Petroleum Engineers History Archive, [Interview Date].

Interview Video

Interview

INTERVIEWEE: Roland N. Horne
INTERVIEWER: Amy Esdorn
OTHERS PRESENT: Marco Blomsma
DATE: October 28, 2014
PLACE: Amsterdam, The Netherlands


Background, Education, and Entry into the Petroleum Engineering Profession

ESDORN:

All right. Well then, we’ll start. Today is Tuesday, October 28, 2014. My name is Amy Esdorn. I am conducting an oral history interview with Roland Horne. Dr. Horne, thank you for being with us today. My first question for you is just starting at the beginning, where did you grow up?

HORNE:

Where did I grow up? That’s actually not such a simple question because I was born in London, England and I lived in England till I was eight or nine years old and I moved to New Zealand, where I really grew up. Then in the 1970s, I moved to California, where I now live.

ESDORN:

What do you think were some of the factors that influenced your interest in the industry?

HORNE:

My interest in the oil industry actually was somewhat serendipitous. I’m sure that’s true for many people. I always, however, since I was a child, I always wanted to be an engineer because my father was, and that was something that motivated me a lot. I was always interested in what he did.

In New Zealand, at that time, when I was a student, large civil projects, energy projects, were at the kind of the pinnacle of what engineers did. Geothermal energy was our space program, if you like, in New Zealand in the 1960s and 1970s. I actually did the reverse of what many in the geothermal industry did, who they sort of retooled petroleum engineers. I actually came into the petroleum industry as a geothermal engineer, and actually I moved to California in the 1970s in the first and second oil shocks because that was the first burgeoning of interest in renewable energy. At that time, the oil industry were the people who were driving geothermal development because the technology’s sometimes the same, sometimes different.

But at that time, the big projects in the US, at least, were being operated by or were developed by oil companies. I went to Stanford University to join the geothermal program that happened to be housed in the petroleum engineering department, and then over the following 30 years, I had become a petroleum engineer actually kind of by osmosis. But I was never formally educated as a petroleum engineer, except on the job. That’s how I -- you asked me how I became interested in it, but I became interested in it just by being part of a program that was centered around.

ESDORN:

That’s great. That’s really interesting. You said your father was an engineer. What kind of an engineer was he?

HORNE:

My father is a civil engineer. He built sewers and bridges and large civil construction. Somewhat interestingly, perhaps by circumstance, the reason that he moved to New Zealand with my family when I was small was when the Wairakei Geothermal Power Station was being constructed. He actually worked on the civil works of the Wairakei Geothermal Power Station. And although I was nine years old at the time, ultimately, I followed down that same path into geothermal energy but in subsurface side.

ESDORN:

When you were little, did your father encourage you in your interest in engineering? How did he encourage you, if he did?

HORNE:

That’s a hard question. I won’t to say that my father encouraged me to be an engineer, although it makes it sound as if he discouraged me, which he absolutely did not. He encouraged me by example. He always took a lot of interest in the things that I was interested in. I liked to build model airplanes when I was a kid, and he had done that, too. He, I think, to some extent, relived some of his youth as an engineering student while I was an engineering student, and he said he took a lot of interest in that. If we talk about parenting rather than petroleum engineering, I think my parents were very skillful, hands-off parents who managed to get all of their kids in the right place without pushing them too hard in any direction.

ESDORN:

You said that you originally got into petroleum engineering through geothermal. But how did you choose geothermal to begin with? Then once you did -- well, let’s answer that first. How did you choose geothermal?

HORNE:

How did I choose geothermal as a career path? I mentioned before that my path into the oil industry and the energy industry was serendipitous, and I am, as you know, a professor. This is an interesting part of the story, too. I chose geothermal energy as an interest, principally because I chose the professor that I wanted to work with, who is Professor Mike O’Sullivan at the University of Auckland. I decided that he was the one who is doing the most kind of interesting research and he was the person I wanted to work with, so I went to him and said, “Well, what have you got?” He said, “Well, this is geothermal energy.” I said, “Yeah, okay.” That’s how I began in geothermal energy, and ultimately, in oil and gas.

ESDORN:

Great. We’re going to move on now to some of your contributions. Is there anything else you want to say before we move on, any thoughts, how you got involved in the industry or your background or early education?

HORNE:

I don’t think so.

ESDORN:

Okay. What have been some of the biggest challenges you faced in your work in well test interpretation? Which of these challenges has had the biggest impact in the industry?

HORNE:

Okay. So, challenges in well test interpretation, which is a field that I followed for 30 years or so… well testing is an interesting subject because it’s been followed by a great many people over a long time. It goes through sort of branching steps from time to time. I’m sure that’s true of all fields. You follow, the industry follows, everybody follows a path for a period of time, longer probably than they should, until somebody breaks out into a new direction. Not all of them. Some people keep doing the same old thing. But in the 1970s to 1980s, a lot of the very difficult mathematical problems were addressed and solved by some of the great names—Ramey and Gringarten, and others—and it seemed by 1990 that all of the hard and interesting problems were solved, and people predicted the end of pressure transient analysis research several times.

Finding something new, not to look for something new, but to break out of that path and go in a different direction, was a challenge, as I’m sure it is, in all research. One of the directions that just happened along in the 1990s was the advent of the permanent downhole gauge. That’s an example of a technology or tool which came into an existence before an interpretation and analysis method really was available to make use of it. That was of course a wonderful opportunity for research as a tool which looks like it would be very useful. There’s no standard methodology. It’s completely different from the kind of tool that we had before. Developing interpretation methodology for permanent downhole gauge data was a big new challenge that we began in the mid-1990s. Interestingly enough, it was one of the most satisfying parts of my career is that I and a couple of students -- I had a very prominent student by name of Suwat Athichanagorn. He is now a professor at Chulalongkorn University in Bangkok. He was the initiator of most of the ideas that finally became useful in the area. But he worked almost uninterrupted by anybody else, so we worked sort of independently in isolation. There were few other people working around the field, but it was an unusual circumstance in that we were able to sort of move the field forward without the sort of competition that we’d always faced with the mathematical modeling, that who did it first or what got added and what was useful and what wasn’t useful.

But that was a difficult research challenge because we actually had no idea how to solve the problem. The problem as it turned out was very different than we thought it was. We thought the problem would be about the mathematics of the solution, which turned out to be rather easy, but the problem was in fact the data themselves. Once we got to look at data from permanent downhole gauges, we recognized that it wasn’t the interpretation of the data that was the problem. It was the manipulation, handling of the data, to put them in a form where they could be interpreted. So that was a challenging but very satisfying problem. I think that’s probably one of the more important contributions that we were able to make out of the research group that I had.

ESDORN:

How did you solve that challenge of interpreting the data? What was that process like?

HORNE:

How we came up with a solution to that problem is like many research problems, the answer came completely out of left field. I always tell my students and people who can’t escape listening that actually the best idea that you will ever have is where you’re listening to a talk or a conversation about something totally different. Because we all think deeply about subjects that we’re working on all of the time. We’ve had most of the sort of good ideas, but then you listen to a talk about medicine or electrical engineering or something like that, and then the idea comes to you, okay, we could take that idea from this field and use it in our field, and that solves our problem, too. In the case of permanent downhole gauges, the application of the wavelet approach, which comes actually out of electrical engineering and signal processing, that was the key that Suwat came up with, adopting a technology from a different field and using it in ours. He in fact tried many things. He was frustrated after a couple of years of dead ends, as often researches are at the PhD level, before actually hitting that golden idea that made it work.

ESDORN:

Thank you. Can you give a little bit of background on the permanent downhole gauge as well as you can, sort of how it’s developed?

HORNE:

Yes. A permanent downhole gauge is an interesting industry phenomenon because they were originally useful not for reservoir engineers–I’m a reservoir engineer—but for the production engineers, the production engineers principally in the North Sea. They first became popular in Norway in the early 1990s. Norway has a regulation or a principle that they don’t wish their reservoirs to go below the bubble point pressure. Therefore, the wells at that time had to be produced relatively conservatively to ensure that the bottom hole pressure didn’t go below the bubble point. Therefore, the advent of the permanent downhole gauge was a powerful tool for them from a production standpoint because they could place a gauge close to the reservoir and be sure at all times that they were not going past that minimum pressure level. Once they had the gauge in place, they could sneak up on that minimum pressure and thereby increase their production, sometimes by a lot, to the point where they could pay for permanent downhole gauge installation—they’re not cheap—but they could pay themselves back in a couple of months of increased production.

That being said, most permanent downhole gauges were the property and instigation of the production engineering department. Reservoir engineers thought we could certainly use that data, too, but actually going from taking data that was actually developed for the purpose of production monitoring and using it instead for reservoir monitoring and reservoir engineering, that was sort of a migration of interest or a migration of procedure and technology that took place industry-wide over quite a long period of time. Permanent downhole gauge started being sold in 1992. We didn’t see very much use of them. Certainly, people did use them, but the use of them for reservoir analysis didn’t really take hold in earnest until about 10 years after that. Nowadays, growing out of its early stages in the North Sea, we see the use of permanent gauges almost universally in certain environments—Saudi Arabia, deep water, Brazil, and the Gulf Coast. Many, many installations of permanent downhole gauges, and that has enabled much better control and better management and monitoring of the reservoirs than was possible 20 years ago.

ESDORN:

That’s great. Thank you. Would you like to take a sip of your beverage?

HORNE:

I would, actually.

ESDORN:

Well, anytime you need a break, you just let me know. Please discuss your work in the area of production optimization. And how has your research in this area advanced innovation?

HORNE:

Okay. Production optimization is another interesting subject from the point of view of bifurcation of direction of where the industry has gone. Research has done that, too. Twenty years ago, in the 1980s, when -- “production optimization” is a term which has been used in the industry for decades, but in the 1980s, at least, when people talked about production optimization, what they really were talking about was looking at two choices and picking the best one. It wasn’t optimization, really, at all, which means looking at all choices and choosing the best one.

The beginnings of production optimization research at Stanford began also with a student, of course, and I had this very good student called Jim Carroll, who came actually from Texas A&M to Stanford, and he was interested in nodal analysis, the analysis of well performance. And it happened to be at a time when artificial intelligence was kind of all the rage. People were talking about that as Silicon Valley was getting started over in California. When he came to Stanford, he said to me that he wanted to look at artificial intelligence applied to nodal analysis and well calculations and things like that.

I at the time didn’t know what nodal analysis was. In looking at it, I actually couldn’t see a way to use artificial intelligence for nodal analysis. That was when, in the process of discovering what nodal analysis was, I also discovered what I just said, that people were doing production optimization, choosing between two things, and realized that actually, production optimization wasn’t being prominently used in the oil industry in the true context of the term. Jim Carroll subsequently did something quite different than artificial intelligence in nodal analysis, and instead looked at optimization of well control, choosing wellhead pressure, choosing choke size, separated pressure and things like that in order to truly optimize the output of the well.

Nowadays, the kind of technology that we’re using and researching in the industry -- now it’s done in Smart Field, I-Field, E-Field, whatever. That’s the context that is referred to nowadays, but it really comes out of trying to develop methods and measurements and techniques, automated or assisted techniques to really get the best out of every well or the best out of every reservoir. But that actually began from actually taking, once again, ideas from another field, operations research and profit optimization, the kind of things that businesses do. Oil refineries do optimization, have done it for decades, too. They’re designing what size pipe has to go from vessel A to vessel B, design it in such a way that overall output is maximized. That technology, the kind of technologies that oil refineries use didn’t really start to be applied in depth to subsurface optimization until about that time.

ESDORN:

How did that advance the industry, in what ways specifically?

HORNE:

How optimization changed the industry or enhanced the industry was by, number one -- well, we have optimization provided for maximization of output, although one of the things I always found intriguing about the research that we did in production optimization was that early on, it wasn’t clear necessarily in advance what you optimized. You can talk about maximizing oil production, but in fact, the way to maximize oil production is quite easy to think of. You simply inject infinite water, and that will maximize the oil. That of course is not a practical solution. Therefore, you have a multivariate optimization. You have to maximize your oil production and minimize your water handling, et cetera. That makes it a much more difficult technical problem, and the solution of that problem actually turned out to be quite simple, which is instead of maximizing oil production, you maximize profit, and that’s something that every company understands because you can convert everything effectively into a cost or an income stream.

So, oil production becomes an income strain, or the handling cost becomes a cost. And if you maximize your net present value, putting the two together—and of course a lot of other things as well—that’s a very simple, convenient, and practical way of conducting optimization studies without having to consider all of the variables independently, which is very difficult. How that changed the industry was, first of all, maximizing profit, that’s a good thing, but also took advantage of tools and technologies that have been continuously developed over a long period of time. But we talked earlier about permanent downhole gauges. Permanent downhole gauges have become part of production optimization and reservoir monitoring that has gone hand in hand with optimization of output and profit, taking advantage of measurements that we didn’t have before for financial advantage of the company, or in the case of a country if you’re talking about a national resource, maximizing the benefit of the national resource for the citizens of the nation is what production optimization is to a national oil company.

The advantages that have been gained from new tools and new measurements has helped propel interests and development of those tools of measurements. Service companies find ready customers when the advantages, the financial and sort of societal benefit of those measurements become evident for companies, corporations, and governments, then it helps to advance the overall technology everywhere in the measurements as well as in production. That was a long answer.

ESDORN:

Perfect. We talked about permanent downhole gauges, but what were some other tools and measurements that kind of came on the scene that were utilized in this production optimization to sort of optimize production?

HORNE:

Other kinds of tools that have come into play along the path of research, not only in Stanford but in many other places, too. One of the ones that I found most intriguing is temperature, which doesn’t seem exotic at all. But as it turns out, every pressure gauge that’s run downhole for the last 50 years has been accompanied by a temperature gauge, and the principal reason for that in most cases is that the pressure gauge has to be compensated for that variation of temperature in the hole. We’ve been measuring temperature together with pressure for decades, and for most of those decades, we’ve largely ignored or discarded that temperature data and focused our attention instead on the pressure data.

This, as it turns out, was not such a good idea, because the temperature data contains a rich vision of the reservoir which we’ve only in the last 10 years started to pay attention to. There are reservoir properties and reservoir behaviors that are evident in the temperature data that are not evident at all or even visible in the pressure data, and so over the last 10 years or so, the industry has been finding ways and seeking ways to actually take advantage of temperature data that in fact had been sitting in the drawer and on the floppy disks for a very long period of time. As a consequence of that, now people have gotten that idea, began to see new kinds of tools like distributed temperature surveys, where people are measuring temperature on purpose with the specific intention of using those data for reservoir analysis. Temperature transient analysis system, temperature versus depth and time, are new measurements that we were getting a lot of advantage from in the industry today.

ESDORN:

Are there any other tools or measurements that you’ve worked on, perhaps, or that you’ve contributed to that have changed the way that the industry has…?

HORNE:

Well, not so much a tool. We’re talking about new tools that have changed the way we do things. It’s not so much a tool as a technique, but a technique which has changed our vision, at least of geothermal reservoirs, which still remains an interest of mine, is tracer testing. Because 30 years ago, everybody understood that geothermal reservoirs occur in volcanic environments, and the flow-through of such rocks is governed by the fractures, the fractures that principal conduits and permeable possible flow.

The world imagined in the 1970s that the interaction of all those networks of fractures in combination acted more or less like a porous medium, and we learned to our amazement that actually that was not true. And once there were some difficulties with some of the early developments in geothermal reservoirs in the 1980s with permanent fracture flows over distances of a kilometer or more with fluids that were being ejected in one place, they traveled very rapidly and came up in other places much, much earlier than everybody thought was possible. We recognized that that was in fact a problem that had not been anticipated. Everybody thought that the one kilometer between those two wells would be described by an interacting network of fractures. In fact, that wasn’t the case at all. It’s just one fracture that went from one well to the other. Fluids passed from the injection well to the producer well in hours sometimes. That was a huge detriment to the production. How was that discovered? That was discovered by use of traces, and the application of traces through geothermal development is pervasive through until today. And more than that, the understanding of the physics of the flow in fractured reservoirs is now understood to be the principal part of the problem.

Actually, if you’re going to do reservoir engineering of a fractured reservoir, you have to know where those fractures are, what direction they’re going to carry the fluids in, and what the interaction between the wells will be. That said, our interest at Stanford in fractured reservoirs suddenly kind of took a new direction because most of the development of oil and gas and the shale, unconventionally, that we’re in now is about fractures. And so, what we understand about flow-through fractures has now grown massively in importance in a whole different area. Let me take a sip.

ESDORN:

Yes, please do. My next question actually was all about tracer analysis, and I’m going to ask it because it’s just a different way of asking the same question and it might trigger something else. What aspects of tracer analysis have you been involved in, and how has your work impacted oil recovery?

HORNE:

Okay. We’re talking about tracers some more. Tracers are used across the board in geothermal developments. They are used also to some extent in oil and gas, too, although not as much, but in different ways I’ll mention in a moment. In geothermal, one of the intriguing ways in how we discover or realize things in engineering and technology -- I had an experience in the early 1980s where I was basically sightseeing in the geothermal reservoirs in Japan, of which at that time, there were three or four. And in discussions with -- most of them were a few years old by that time. The oldest one was perhaps 10 to 15 years old, and some of the newer ones had only been producing for a couple of years.

In discussions with the engineers that I visited with, each of them was describing difficulties they were having with the reinjection wells. And the problem they were having is that the reinjection water was coming back much sooner than they had expected, and that was a surprise at that time in the industry. People hadn’t imagined that that was possible. But what made it intriguing was that after going from one field to another, I was on a trip that last three or four days, each of the developers described more or less the same problem but gave the impression that they thought that they were the only ones who had that problem. Because the problem was a surprise, they thought, well, the injector shouldn’t be behaving that way, and therefore it was something about their field or something about their development that caused them to be that way, without realizing that the other four fields were all having a very similar problem.

That was I feel like a research result that came by accident without being researched at all. To actually hear the same story four times from four different developers, each of whom didn’t know that the other three developers had the same problem, that shone a light on the problem and the people. I didn’t do it, but the developer who solved the problem for themselves, and hence ultimately for all the others, too, was the one who made use of tracers. They identified what the problem was by applying tracer technology to identify where the fractures were, which combinations of producers and injectors were unfortunate once, and therefore they cut off the ones that had the fast breakthroughs and ended up with a project that until that time was really about to die completely. They lost 35 percent of their production over a year in a project that was supposed to last for 30 years, so they were in serious concern to identify the problem with tracers. They shut in the injectors that were strongly connected. They drilled a few outfield injectors in order to get rid of the water they had and basically solved that problem. They regained their production probably a year later.

So, that was an entire project which was basically rescued. Two things. One, they identified the problem with tracers. They figured the solution using tracers and recovered the project and basically provided a precedent for technology that was applied to a lot of other fields, too, many of which either had the problem and solved it a similar way or identified that they could have a problem and watered it off by doing something differently that the tracers showed them that they should do. In the oil industry, tracers are used, too, but not as prominently, mainly because of the different nature of the reservoirs. Fractures are important in our industry, too, but well-to-well flows through fractures have less of an impact, if you like. We do that kind of analysis, but not so much.

ESDORN:

Please discuss how you got involved in the SUPRI-B and the SUPRI-D research consortia, and what kinds of research do you conduct?

HORNE:

The mechanism or one of the mechanisms by which we conduct research at Stanford University and many other universities too is the formation of research consortia. And a research consortium is a group that is organized by professors at a university to gather companies that have interest in that particular area of research because of the kinds of fields they have and the kind of problems they are facing. The consortium members contribute an annual financial donation to the group, and they basically share in the results that come out. The consortia arrange -- the companies join for different reasons, I’m sure, but the research results are what many are interested in. Some of them do it simply for philanthropic reasons. They want to support the universities. I think a big advantage for many companies is the opportunity to actually track very good students, of which we have many, as they pass through their research career. So they can see what the student is doing. They may be interested in the result of the research itself, or they may simply be interested in the students. When they graduate, they see that’s a good person and we want to give them a job.

In the case of SUPRI-D, which is the research consortium that I still work with at Stanford, which was originally called Innovation in Well Testing, that’s more or less what it is today. At the time that we formed it, well testing research at Stanford was famous for well testing research. At that time, principally by the work of Hank Ramey, who is a famous name in well testing -- he was the chairman of the petroleum engineering department at that time and a good friend and mentor of mine. He had done work on well testing for a long period of time, but we wanted to kind of expand the research in well testing and go in some different directions. Hank Ramey was a very prominent and well-known engineer in the industry for a very good reason. He had a lot more draw than I did because I was simply a young assistant professor at that time. He could call up a couple of friends and get support for student more or less anytime he needed to, but I could not.

I actually instigated the formation of SUPRI-D consortium as a way to gather funds for my students and to follow the ideas I had. It was a consortium that formed not necessarily to address the specific problem that was happening at that time but to allow for an expansion of research activity in that particular area. It was a time when well testing research was very popular—not that it’s unpopular today. But in the 1980s, many universities and many students liked well test research. It was very mathematical. It was very technical. It was the opportunity to sort of do cerebral problems and come up with solutions that didn’t require building lab equipment, which is always problematic. It didn’t require going out in the field and sort of spending hot summers making measurements or things like that. There were many students who wanted to research in that area, and coming up with the funding for them was the principal reason to create that consortium. The consortium still exists today. We still have member companies. The directions of the research have changed over time, but we still have companies interested in that. We still have plenty of students interested in that, too.

ESDORN:

What were some of the areas of research that you’ve overseen over the years in SUPRI-D consortium?

HORNE:

SUPRI-D has had influence in quite many different directions associated with well test interpretation. When it began, principal interests at that time in the late 1980s was nonlinear regression, which is now pervasive in the industry. There are many software companies who have created wonderful tools for well test interpretation, most of them relying a lot on nonlinear regression. But in the 1980s, although the idea had been actually proposed much earlier in the 1970s and kind of played around with, the concept of nonlinear regression were not widely implemented or, for that matter, widely accepted. It required two things to happen. One of them was computer technology, which if you think back to the history of computers, the 1980s was the PC boom. IBM PC was created, Apple Macintosh, et cetera. So those tools provided the actual methods and devices that could be used for nonlinear regression and computer-related interpretation. That was our focus in the 1980s in SUPRI-D and in many other places, too.

The second thing that was necessary, interestingly enough, was acceptance of the idea. Our industry can be quite conservative sometimes. People keep on doing the same thing in the same way for a long period of time, and they like to keep on doing it that way. Nonlinear regression was a bifurcation of that path, and people feared that once you had automated analysis, nonlinear regression, then you could have secretaries doing well test interpretation. Reservoir engineers will be out of a job, which of course is actually nonsense. But those are the kinds of defects that people imagine. Automated or computerized interpretation was a strong focus of SUPRI-D in those days, and to some extent even today.

Second area that we sort of -- second trail that we blazed was permanent downhole gauges. I talked about that earlier. There was a lot of new measurements become available, and figuring out how to make use of them was the whole direction of that. In the 2000s, we moved in somewhat different directions again and started focusing on temperature transient analysis and distributed temperature surveys. We’ve also actually grown out of our interpretation of complex problems in more recent times in looking at adoption of the kind of techniques that come from computer science in data mining and machine learning instead of actually doing elegant mathematical solutions to problems, which is what we always liked to do. Take a look at the data themselves. We have tools that are generating gigabytes of data, sometimes per hour, at least per year in many cases instead of trying to take those gigabytes of data and filter them down to a few hundred or few thousand to do the kind of analysis that we did throughout history to actually try to capitalize on those millions and billions of data and actually try and figure out what they’re trying to tell us. So that’s a migration away from parametric nonlinear regression that we did in the 1990s to nonparametric regression, where we don’t seek a history matched to a model but we tried to have the data suggest what the reservoir model should be.

And the way to think of that problem is the way they use data mining and machine learning in industry today. Something that everybody knows is character recognition or voice recognition when you call your bank on the phone and you are asked whether you want to speak to department A or department B, you don’t talk to a person. You talk to a computer, and that computer is recognizing your response, even though people talk with different accents, there’s noise in the background, they run their words together. Computers are now able to recognize signals from complex, noisy, difficult signals that they receive. There’s no reason to expect that we can’t do the same thing with the reservoir data, too. We have these massive amounts of data. They’re coming in from permanent downhole gauges. They’re coming in from surface gauges. They’re coming from seismic surveys. The problem that we have in hand today is to try to divert or adopt some of those techniques from computer science and use them in reservoir analysis.

ESDORN:

If I understand properly—please correct me if I’m wrong—we are picking up a lot of information and data that we don’t know how to interpret yet, or in the past that has been the case, and then research comes in and learns how to then interpret it. Is that correct-ish?

HORNE:

Yes, it’s correct. Quite correct.

ESDORN:

That’s very interesting to me. My question then is how much currently or in the past has that data -- I guess it’s just been growing, as you said. So, how do you focus your research to decide what to interpret? How do you, I guess, give it priority?

HORNE:

Well, the question is about data and where we get it and how we decide how we’re going to use it, and there are many answers to the paths that we’ve followed over the years in the industry. There’s a lot of the new technology that’s developed all the time every day in our industry. Sometimes it’s a result of a company that is in the business of acquiring data will have a new idea. They’ll think about the new kind of data that they could measure and provide us the services, the business they’re in.

Then the second question is how would it be of value to the producers to actually make use of that data, because otherwise, of course, they don't want to provide the service. So that’s one path. Another path is the other way around. The companies want to know something about the reservoirs, and there isn’t a tool available to actually make a measurement that allows them to know those things, and therefore, the service companies will seek a method of acquiring that kind of data.

Separating yourself for a moment from the energy producers and the service companies to a university professor, our part of the problem is rather different. We see data which is being collected. We see the methods by which those data are analyzed and interpreted. From our point of view, we can look at how to do those interpretations better. That’s a common kind of research problem. But a more exciting research problem is how to make use of data that nobody ever used before. We talked about temperature earlier on. People didn’t use it very much. Or, to propose a new management. We’re unconstrained by reality in a university, so we can imagine any kind of measurement that we like in order to create something useful for the point of view of increasing energy recovery efficiency.

We don’t do that in vain. If there’s something that’s a clear advantage to measure, then it can become apparent to these companies or others that this would be a good tool to develop, because it would provide this kind of value to the oil producers and gas producers. Our task in the university is education, actually. But in the process of that education process and doing research, trying to come up with new ideas is one path we follow. Trying to improve on old ideas is another path that we follow. That replicates the kind of things that people are doing in the industry itself directly, trying to find new measurements and trying to find new uses of old measurements. Can I have a drink?

ESDORN:

Yeah, go ahead.

HORNE:

Okay.

ESDORN:

We have discussed your research in the SUPRI-D consortium. What about your research in SUPRI-B consortium? And what were some of the outcomes there?

HORNE:

SUPRI-B -- actually, I don’t do so much work in SUPRI-B. I’m a part of it, but…

ESDORN:

Okay.

HORNE:

I don’t have much to say about that.

ESDORN:

Okay. Then we’ll move on. Please discuss how your work in geothermal impacts your work in petroleum engineering and vice versa. You’ve kind of discussed that a little bit, but if you could go into more detail…

HORNE:

Yes. The relationship between geothermal reservoir engineering and petroleum reservoir engineering is kind of interesting because there are prominent similarities and there are prominent differences, too. One of the important differences that I might mention is the fact that in geothermal reservoirs, we have non-isothermal flow, and that’s no big surprise. We’re talking about thermal reservoirs. We’re talking about temperature changes. When we’re talking about the flow of steam and water, the difficult physical problem is the interchange between steam and water during the flow.

Unlike oil and water, which are different materials, they basically remain separate at all times. The interesting and difficult thing about steam and water is that that actually interchanges from one phase to the other without still being the same stuff. That’s the difficult research problem, one that we actually focused on for quite many years. And measuring steam, water, relative permeabilities are very difficult experimental process that we faced in our lab. However, oil and gas, as it happens, have this problem, too, without having been too aware of it for quite a long period of time because hydrocarbons, specifically oil and gas, also interchange between the two phases. Actually, in an even more complex way, but the fact that you can have the same material in either one of the gas or liquid phase with an interchange between the two sometimes has impact for the oil industry, too.

The principal similarity or overlap between steam/water flow in geothermal and oil/gas flow in the oil industry comes in most prominence to gas condensate reservoirs, because gas condensates are very light hydrocarbons which interchange very easily between the two phases, actually in a reverse fashion if they condense upon a lot of pressure reduction. But the manner by which gas condensates form liquids and gas phases, which affects the flow in an important way, is completely analogous or at least very similar to that in geothermal. The work that we did on phase changing multi-phase flow in steam and water in geothermal reservoirs turned out to be very useful for gas condensates, too.

Some of the research that we did in the geothermal group migrated into SUPRI-D when we’re looking at the flow of gas condensates. That was an interesting and important and relevant problem to those companies producing gas condensates, but more importantly still is that in the last five years, the oil production of the State of Texas has doubled, and most of that oil production comes from so-called tight oils. Some people call it shale oil, and a very significant fraction of that is gas condensate. So the manner by which we’re producing large amounts of liquid from formations like the Marcellus and the Eagle Ford really are affected strongly by those phase changing multi-phase flow problems, which the oil industry didn’t pay too much attention to 20 years ago, but the geothermal industry did. We focused on this problem for 30 years and turned out to be in a very good place when tight oil and shale oil became prominent in the oil industry.

ESDORN:

You actually took the words right out of my mouth. I was going to ask you what were some of the analogs between the two, and also how has the petroleum industry benefited from research in geothermal. Can you think of any other than what’s been going on right now, or of how the petroleum industry has benefited from research in geothermal?

HORNE:

The interaction between ideas in the geothermal industry and the oil industry have been in both directions. There’s been technologies that have been developed in one that had been used in the other, and there are several good examples of that. While logging for fractures, image logging was actually something that’s used thousands of times a day in the oil industry today but was probably first tried in the geothermal industry—again, because of the importance and impact of fractures in volcanic rocks. There were drilling technologies actually developed in geothermal, too. Because the rocks are very hard and very abrasive, polycrystalline diamond cutters have been developed quite early in geothermal drilling. So those are good examples how -- what I was talking about before, you take an idea from one place and use it in a different place. That’s what we do so often in engineering.

ESDORN:

Let’s see. You discussed a little bit nonlinear progression, permanent downhole gauges, temperature transient analysis, and that sort of thing in your research and the things that you’ve been a part of developing. Are there any other contributions that you would like to discuss at this time?

HORNE:

Yes, there is. We talked a lot about technology and ideas that changed the industry or changed the way that we do petroleum engineering or geothermal engineering in a better way that’s advantageous for everybody. My life doesn’t always hinge only on that. One of the things that is an important part of being a professor in the university like Stanford is not only the research but the people who are doing it. One of the ways that research in a university like Stanford—and other universities, of course, too—provides advantage to the industry is that the people who do the research just don’t deliver a report and move on. They actually join the industry and take the ideas that they’ve developed, they implement them in practice. And in some ways, it’s the ideas that change the industry. But in more ways, it’s the people who provide the ideas and implement the ideas that change the industry. Research is absolutely important to me, but I could be a researcher in an oil company and probably do much the same sorts of things. What’s different about being a professor in the university is that the process of the research itself becomes important in that it engages students. They are developing the ideas, but sometimes there is some participation required in developing their process or their approach or the way of getting to these ideas or taking advantage of those ideas. That’s it. That’s a different part of the research altogether and one that’s important to me as a professor.

ESDORN:

Sounds like what you’re talking about is almost… not quite mentoring but kind of guiding and showing, helping them to kind of find the way to research that is the most efficacious for them and to develop the whole student.

HORNE:

You’re right.

ESDORN:

Okay. So my next question is which innovations or milestones in your discipline do you consider to have had the biggest impact on the industry, and why?

HORNE:

In my area or in general? In my area?

ESDORN:

Yes. You can also, by all means, discuss in general if you would like as well, but specifically your discipline.

HORNE:

Okay. So impacts in the industry of the kind of research that I’ve been involved in -- I talked a lot about fractures. Fractures has been a focal point of a lot of the research we’ve done. In time, we have come to, in the case of geothermal reservoirs, we’ve come to understand, we think, how they work, where we previously didn’t. That realization that reservoirs aren’t the way that we thought they were was an important impact of the research we did on fractures and the tracer analysis of those fractures. In the oil industry, the things that we have learned about reservoirs that we didn’t previously know that’s associated with research that we’ve done at Stanford, permanent downhole gauges are very important part of that.

Permanent downhole gauges are a vision of the reservoir that we never had before. The problem with them was initially—and to some extent even today—we don’t fully know how to capitalize on the data that they provided. Developing methodologies to make use of the kind of massive data sets that we have available to us for now is part of what the industry needs in order to improve recovery to lower costs and to gain increased fracture recovery. That’s what everybody wants. Instead of leaving 60 percent of the oil on the ground, we’d like to leave 40 percent of the oil on the ground, or hopefully zero percent of the oil on the ground. The methodologies that we need as an industry is to understand how the subsurface works, first of all. That’s not in itself a goal. The goal is to actually take advantage of that understanding of how the reservoir works in order to increase the efficiency of the recovery that we have.

Coming back to research and how we get to that point, there are many other ways than ours, but the use of real-time data such as from permanent downhole gauges and for that matter, from permanent installations of DTS fibers and things like that, that’s what we have to have if we are ever going to have Smart Field operations—some people call it I-Field, E-Field. If we want the kinds of real-time control and understanding of our reservoirs such as people have in mechanical installations like factories or an aircraft which is flying—they know exactly what each engine is doing and what the air pressure is, everything—that’s what we want to have about an oil reservoir. We want to know exactly what’s going on everywhere in the subsurface. We’re nowhere close to that.

In a perfect world, we like to know the pressure at every location, the fluid saturation in every location so that we could take advantage of that in order to maximize recovery. A transparent reservoir or a transparent earth above the reservoir is what we’d like to achieve. We’re never going to have that, so we want to get as close to that as we can. We want to take measurements of whatever kind. Pressure is what we started with, but we can go to other kinds of measurements, temperature. I don’t know what else. Resistivity. Pull them all together and try and make a picture of the reservoir in order to capitalize on that picture and improve our recovery. That’s sort of the Holy Grail of reservoir engineering is to know everything about the reservoir, soon enough to take advantage of it.

ESDORN:

Okay. What do you consider to be some of the biggest challenges facing the industry going forward?

HORNE:

Yeah, so if we talk about the challenges of the industry today, there are many, as there always had been. Our industry sometimes changes slowly, and it sometimes changes rapidly, and basically it’s been changing rapidly, especially in North America with the transition, or at least the refocusing of attention on unconventionals. So, the reservoir physics involved in producing hydrocarbons from nanopores requires an understanding different than we’ve been accustomed to seek.

We have a pore which is so small that it contains a countable number of molecules than we’re obliged to consider molecular dynamics, whereas previously, we were considering bulk properties of fluids. I don’t think we currently fully understand what those reservoir physics are, and that probably contributes to the very modest recovery factors that we’re seeing in unconventionals right now, and we clearly want to improve on that. Understanding the reservoir physics in this very small space is something that we need to do in order to get to that point.

At the same time, we also have an incomplete understanding of the way that shales actually behave geomechanically, how exactly they shear or fracture, what the kind of properties of those fracture networks are like, how is it even possible to produce commercial quantities of oil and gas from material which has a permeability in the nano Darcy range. Those are problems that the industry is addressing widely. Lots of people focusing a lot on that, but the problems are by no means solved. This is something that we continue to address, and I’m sure we have a lot of room for advancing the understanding before we get to 35 percent recovery in unconventionals, which we’re far short of at the moment.

ESDORN:

What has made working in the petroleum engineering industry meaningful to you?

HORNE:

Well, I love working in the oil industry. I don't always have the opportunity to say that too loudly in California. But the great thing about the oil industry is that it’s big but not so big, so actually it’s quite possible to get to know a lot of people in the oil industry. And of course … it’s completely international. You can know people all around the world who are working on similar problems, and sooner or later, you meet many of them. Our industry has a lot of different facets and corners and areas of technology that people work in. I don’t claim to know all of the well testers in the world, but I know a lot of them, and one of the great things about our industry is the opportunity to actually do that, to actually know the people who work in your field in China, in Russia, in Japan, in Mexico, and everywhere, basically. It’s kind of the human side of the industry that I think people outside often don’t fully understand.

ESDORN:

What are some of your favorite memories about working in the industry?

HORNE:

That’s a good one. When I think about the people that I know and have known in the oil industry and the places that they are today, one of the things that I have… the stories I like to tell people are the peculiar places that I have met people that I know unexpectedly in the world, to be in a third country and bump into somebody I know from yet a different country on the street or in an airport or somewhere like that is unexpected and always memorable. And I can think of several occasions. I once visiting Neuquen in Argentina on a distinguished lecture tour for SPE, and in a very small airport which was probably 20 meters square, I bumped into a friend and former student from Brazil who happened to be visiting that same place at the same time. I once sat behind a friend of mine on a plane between Singapore and Jakarta that I knew previously. Meeting people in unexpected places and sometimes very unusual ones is always memorable.

One of the things that I do remember from Hank Ramey, he told me the first time I ever went to Indonesia, he said, “Well, if you’re working in the oil industry and you’re in Jakarta by yourself, if you ever get bored, just go down and sit beside the pool in the Jakarta Hilton, and sooner or later somebody you know will show up.” I happened to be staying in the Jakarta Hilton and I did exactly that. Sure enough, that’s exactly what happened. I sat down there and ordered a beer, and somebody called my name from the other side of the pool. It was somebody I knew sitting drinking coffee in a deck chair.

ESDORN:

Are there any other things that you can think of that have made working in the industry meaningful to you?

HORNE:

I like to travel, and one of the interesting things about oil is the great diversity of places where it’s actually found. So, working in the oil industry provides the opportunity to actually go to a great many very different kinds of places. So, places that you would never go to as a tourist, sometimes for good reason, adds a lot of texture and color to your life.

ESDORN:

Definitely. This is our last question. It has to do with SPE. How has being a member of SPE affected your work and your career?

HORNE:

I joined SPE in 1977, so I’ve been a member for a long time. SPE has had a lot of impact on my life and my career, but the thing that I might mention most prominently is that coming to the fall meeting and technical meetings, which I go to quite many, I have good friends whom I have met at SPE meetings, people who work in the same technical area as me, or sometimes not. It’s kind of my stamping ground, if you like. Being a member of SPE provides the opportunity for, first of all, technical interaction between people, but also people interaction between people. That’s an important part of anybody’s life. People help others technically with ideas and also on the human side too, introductions. If you want to visit somewhere or you need to know who is working in a given area, when we’re looking for new faculty, we know people to call up and ask, “Do you have somebody to recommend?” SPE is most important to me in bringing people together, and in many circumstances, when they’re bringing people together for a technical meeting of some kind and we’re talking about technology and engineering in a focused and specific way, and so a lot of ideas come from meetings like this. That’s the reason why I go to so many.

ESDORN:

Well, thank you so much.

HORNE:

Is that it?

ESDORN:

Yes. That’s it.

HORNE:

Good. All right.

ESDORN:

It was a pleasure…

HORNE:

Thank you.