There is no time for us to sit and wait for someone to bring us a ready-made solution
Recently, the Gazprom Neft Science and Technology Center (STC) has established an in-house research engineering unit with the main objective to shorten the path for the implementation of R&D results in practice. The head of the new unit, STC Deputy General Director Alexander Sitnikov told the Siberian Oil magazine about the role of R&D in the oil industry today and why it is more prominent than ever before, what oil producers can learn from thermophysicists and aircraft engineers, and what way of thinking a petroleum engineer of the future should have.
Siberian Oil Magazine
– Alexander, could you tell us what the research engineering is about, and why today this discipline has gained interest from the oil industry?
– Research engineering is the application of knowledge accumulated by the fundamental sciences in the production process; it brings together scientists and engineers to solve business problems. In the 1950-70s it was not unusual that some company could develop a know-how, have it patented and capitalize it for the next 20-30 years. Development of technologies used to take decades. Today, as the quality of reserves deteriorates and oil companies start getting access to unconventional, low-permeability reservoirs, such as the Achimov Formation, new challenges are emerging so fast, that we have to radically cut this time. We can only be competitive and take leading positions in the oil industry if we cut the whole cycle as much as possible — all the way from the emergence of an applied scientific idea to its implementation in technology and application in practice. It is no longer the time when the academic and business communities are separate, when businessmen absorb new ideas only during rare meetings with scientists. Today, science, technology and business should always be in close contact. The world is changing so fast that science, if it does not promptly respond to business needs, will lose its appeal. Five-six years ago, we noticed these trends and began building partnerships with various industrial R&D institutes, traditional universities, research centers and service companies in order to minimize the gap between science and business. This is how the research engineering discipline was taking shape at Gazprom Neft.
– Is the current situation unique, or were there any periods earlier when the role of the fundamental sciences in the petroleum industry would see a dramatic growth?
– Petroleum engineering was formed in the 1930-50s, and the fundamental science played an important role in this process. Such scientists as Morris Muskat, Leonid Leibenzon, Vladimir Shchelkachev came to this area of knowledge from the classical sciences. For example, Muskat studied thermophysics. He noted that the equations that describe the propagation of the thermal field can be adapted to describe the propagation of the pressure field between wells when petroleum is being produced from them. So he adapted the thermophysics equation to the underground hydrodynamics equation, thus having laid the foundations for the petroleum engineering. Wells were being drilled long before that, but as long as the quality of reserves was such that oil just flowed out on the surface no one thought about how to describe it, what physical laws the oil production is subject to.
This revolutionary breakthrough was followed by several decades of evolutionary development.
Then, in the 1980s, information technologies came in. It was the time when large oilfield service companies started adapting them to petroleum engineering and, based on the physics described in the 1930-50s, created tools that gave a new impetus to the development of petroleum engineering in the 1985-2000s when the geological and hydrodynamic modeling tools were being actively introduced.
But the classical petroleum engineering does not have answers to the challenges facing the industry today, so to optimize development systems with horizontal wells and multi-stage hydraulic fracturing we need again to turn to the fundamental sciences, to create new theories. Meanwhile, since the middle of the XX century the fundamental science has accumulated a tremendous potential that has not been fully realized in petroleum engineering yet.
– Where has Gazprom Neft already applied the new science to successfully meet business challenges?
– One of the examples is the development of oil rims with horizontal wells. The reservoir physics there is totally different from the traditional one. The classical petroleum engineering did not have a theory to describe it, so we needed to create this theory: using the knowledge accumulated by the fundamental physics we had to understand how the fluid behavior in oil rims differs from conventional oil field. This helped us come up with a successful concept for the development of the Novoportovskoe field with its under-gas-cap oil reservoirs. For instance, the relation between vertical and horizontal inhomogeneities has a significant effect on how the gas cone will be formed and, therefore, determines the main reservoir development parameters. In conventional fields, this relation does not have such an impact on the development system efficiency while in the Novoportovskoe field, where oil reservoirs are overlaid by gas caps, it was critically important.
Messoyakha is another example, but the problem there was not the gas cap, as in the Novoportovskoe, but the underlying water, so the laws and mechanisms there are a little different. However, in the same way as in Novoportovskoe, understanding physics, accurately identifying all factors that affect the efficiency allowed us to make the right decisions — for instance, expand the fluid treatment facilities at the Messoyakhinskoe field. Now we understand that it was a critical and timely decision that saved us billions of rubles.
– Do you need any special competences to implement scientific engineering in the petroleum industry?
– Unfortunately, petroleum universities today cannot give fundamental knowledge of this kind to their students, so our mission is to bring it there. This is about developing human resources. We want our petroleum engineers to think from the reservoir physics standpoint, not from empirical rules, with deep understanding of what they are doing and why.
The petroleum engineer of the future is a person with a deep understanding of physics and mathematics. Without this, we cannot achieve the leadership. Realizing that there are very few people in the labor market who would meet our requirements, in 2014 we started establishing fundamental science departments at such classic educational institutions as the St. Petersburg Polytechnic University, Moscow Institute of Physics and Technology, Moscow Physical Technical Institute, St. Petersburg State University, etc. This allowed us to set up training of full-fledged petroleum engineers of a new type. Many of those whom we recruit today are graduates from our fundamental sciences departments. And when they solve problems of petroleum engineering, they already think in a profound and accurate manner. We just will not be able to find the keys to the reservoirs and the fluids that we are facing today if we take a different approach.
– In the 1980s, the industry was changed by oilfield service companies. What about today? Has the development initiative passed to vertically integrated oil companies?
– Back then, the technology development rate increased dramatically, but still there was enough time to develop the technology and bring it to the market. Today, we can no longer sit and wait for someone to bring us a ready-made solution. The solution is needed right now, so we begin to look for it ourselves, create prototypes, sometimes in cooperation with service companies.
As an example, we can take algorithms for selecting the best parameters: well length, number and location of hydraulic fracturing ports. To optimize our development systems, we have created a hierarchy of models that allow us to engineer the optimal field infrastructure and field development system in an integrated manner. We can do it quickly, in three to four months, because we start with the reservoir physics and outline the key factors only which allows us to work with simpler models. We adhere to the principle that we need the procedure to be as simple as possible so that we could make the decision on time but not too simple so that we would not lose the key drivers of efficiency.
– Who are your key partners from academia and how does your cooperation evolve?
– In recent years, we have established many partnerships. We have more than a hundred technology partners. There are more than twenty universities which we cooperate with. Keeping in mind our business objectives reflected in our technology strategy, we have formulated nine major areas of interest, where the biggest prize is hidden. Meeting these objectives will bring us the keys to unlock the reserves that we are going to access in the next 5-10 years. Within each of these areas, respective challenges have been identified. As we meet these challenges, we set tasks that need to be addressed. Having formulated the task, we clearly understand what competencies are needed to solve it. Of course, our internal resources are not enough to meet so many tasks. That’s why we pass them on to our partners who possess best expertise in respective areas.
– Can you give us some examples of ongoing projects as part of the research engineering discipline?
– One of the brightest examples is new materials. For instance, drillers need superhard bits, so we model composite materials that are not inferior in strength to diamond but much less expensive.
Another example of research engineering is the spectral geological modeling. Spectral analysis as a field of science involves a fundamental mathematical tool of breaking down a signal into spectrum. Such methods are widely used in chemistry, radiophysics, nuclear physics, acoustics and other fields of science, but until recently they have not been used at all in geology. Under the classical approach, geologists took two intervals (let us call them interval 1 and interval 2) and drew a correlation between them on the basis of various physical parameters defined as a result of interpretation of recorded signals (logging curves): thickness, net-to-gross ratio, resistivity, etc. On the basis of this correlation (comparison) they made conclusions about properties of these intervals, which depended very much on the experience of the interpreter. Spectral analysis of the same signals and comparison of the main components of the spectrum, rather than the signal as a whole, allow us to make more accurate predictions.
Another example is geomechanics. Until recently, this field of science was not used in the oil industry although it had existed for decades. Today we are using it to solve our problems. On the basis of the solid state physics, understanding fragility of the rock and its strength limits we can build geomechanical model of the reservoir taking into account natural fractures created in the course of formation and maturation of the reservoir rock. Then, for example, in the process of drilling we can induce new fractures in the reservoir layer. As a result, we are facing a complex task: create such conditions that, on the one hand, the network of primary and secondary fractures would not cause excessive loss of drilling fluid and, on the other hand, in the course of further hydraulic fracturing we could get the widest sweep area possible. Geomechanical model allows us to predict what changes will occur in the reservoir if we change the bearing, rate of penetration, or direct the well anywhere else. We are now able to determine the optimum, minimize drilling fluid loss, find the right points to initiate hydraulic fracture, ensure the most efficient geometry of these fractures. All this can be achieved by combining the laws of mechanics, hydrodynamics and geology.
Or let us take the enhanced oil recovery methods. Very few people today can describe with equations and formulas how displacement agents, polymers, surfactants interact with oil and surface of the rock, or how efficiently the displacement occurs, although it is vitally important. After all, creating a model and comparing it with the actual results we, firstly, check its correctness and, secondly, can determine what properties the fluid should have to displace oil better. This is the chemical engineering approach that allows us, using chemical and mathematical modeling, to synthesize the substance with the most suitable properties by quickly analyzing thousands of formulations. And then we go to the laboratory and on the basis of several resolving experiments determine what needs to be changed to get the desired properties.
– These are examples from physics and chemistry. But how can mathematics help?
– One example is simulation of fluid inflow to wells based on PEBI grids. In modern hydrodynamic simulators, orthogonal grids are normally used: the reservoir, without account for its structural features, is cut into identical “cubes” in the simulation program. This eventually results in a significant reduction in the simulation speed and accuracy. Unlike the above, PEBI grids are irregular grids that provide a more accurate simulation of the reservoir hydrodynamics. The simulator automatically adjusts the grid to any elements in the development system, be it a horizontal well, hydraulic fracturing, fishbone, in order to describe them as adequately as possible. The scientific component of this project is a mathematical tool that allows us to determine how cells in such a grid should be defined.
Another project, which we are running today jointly with Skoltech, is aimed at solving the dimensionality reduction problem. What is it? Here is an example from the aviation industry where such solutions are already in place. The aircraft wing profile is described by several dozens of parameters. Previously, solving the problem of wing optimization we had to take into account all these parameters which required a huge number of calculations. But then it was noticed that under certain conditions the wing profile can be described, with a good degree of accuracy, only by a few parameters thereby reducing dimensionality of the problem by ten times. In turn, it speeds up the calculations hundreds of times and, accordingly, reduces the optimization cycle from a year to several months. We want to apply the same approach to meet our development system optimization tasks. This means that after receiving new data it will take us a few months, rather than a year or two, to update the model.
There are a lot of such projects already, and they are making an increasingly significant contribution to improving our performance.