Archive

No. 3, 2012


ADVANCED MATHS FOR OIL INDUSTRY


Oil of Russia talks to Andrew Conn, Research Staff Member and Research Relationship Manager for Chemicals and Petroleum, IBM Thomas J. Watson Research Center

IBM IS THE LARGEST IT CORPORATION in the world with more than a century-long history. The company has been working in Russia for almost forty years now and its technical solutions in our country enjoy immense popularity. One of the priority areas in IBM's activity in Russia is cooperation with enterprises of the oil and gas industry, the largest consumers of IT services. The company's customers include almost all the key oil and gas companies - LUKOIL, Rosneft, Gazprom, Surgutneftegaz, TNK-BP, and Tatneft, as well as many small independent oil and gas enterprises. IBM's highly powerful computers and competent mathematical analysis help oil companies to optimize oil field operation.

Q: How did IBM begin operating in the oil and gas industry? And how did work in this sphere influence IBM development?

A: The oil industry was the first branch where IBM used optimization processes. At the end of the 1950s-beginning of the 1960s, we first used the linear programming method in the interests of the refining industry for optimizing technological processes and delivering raw material. At that time, we were able to resolve tasks with one hundred variables in approximately ten seconds. Today we solve tasks with millions of variables in fractions of a second.

In fact linear programming tasks were solved back during the Second World War. While in peace times, it was in the oil industry that this method was first used the most. George Danzig and Soviet scientist Leonid Kantorovich are rightly considered the pioneers of the linear programming method. As computer technology developed, their ideas became increasingly sought after in the industry.

It was the activity of oil companies, particularly in the area of seismic research, that gave one of the most powerful boosts to the use of highly productive, extremely fast computer systems. So IBM's interest in the oil and gas industry is entirely natural.

Q: Which of IBM's solutions are in particular demand today among oil companies in the world and in Russia?

A: Specialists of the oil industry have to deal with solving extremely complicated tasks relating to linear programming, and IBM has immense experience in this area. IBM mainly offers optimization and regularization according to the highest standards and data processing, including seismic and geological, for making the best choice of models for developing and exploiting reservoir basins. Where to place wells, how to reduce the amount of water pumped into the basin, how to enhance production - our optimization estimates help to answer these and many other questions. Optimization is extremely important, since it helps to save resources.

For example, our services are widely used in servicing offshore oil platforms. Imagine a typical situation for an oil company. A group of experts gets together and discusses how to organize service work for the next two years. The oil workers usually hire contractors and use outsourcing for carrying out service work. Service companies receive a schedule from the customer based on the solutions of its experts. And no matter how intelligent the experts are, they cannot objectively take hundreds and sometimes thousands of variables into account when carrying out optimization. If he is lucky an expert may be able to draw up an optimal solution if the number of variables is four, eight at most. Working by themselves, experts cannot take into account all the parameters and variables, which is where IBM's mathematical methods come in. Problems are often so complicated that it is impossible to unequivocally obtain good mathematical models of their solution. Let's take electric circuits, for example, their physics are very simple and it is easy to choose very good mathematical models for them. But if we are talking about an oil field with complicated reservoirs situated many miles under ground or in the subsea, with heterogeneous layers, and the information about the processes going on deep within a reservoir basin can only be obtained from wells, it is much more difficult to create a mathematical model. It may be impossible to create an ideal model, but even if operation is improved by at least 1%, this is going to save millions and millions of dollars.

There may be many solutions to a problem, be it where best to drill a well or maximize the production volume over the entire exploitation time of a reservoir basin, but the most correct one must be chosen. Geologists may have numerous solutions based on the available data. But it is much too expensive to study and appraise all of them. This is where mathematics comes to the rescue, simplifying the choice and reducing the number of alternatives.

We are helping oil companies to optimize in many vectors. For example, they need a competent mathematical analysis to understand all the subtleties of the warning signals that come from sensors in the wells. This is particularly important for offshore platforms with their higher security requirements.

Many oil companies have extensive databases, but these data are not always unified even throughout the same company. It is asking too much to create a new computer program for each separate oil field. We also help to unify these data.

IBM's efficient algorithms of optimization and complex mathematical instruments are used throughout the world to search for optimal solutions. Oil fields in the Middle East, the North Sea, and Siberia, for example, require different approaches. We will use the same mathematical methods, but the solutions will significantly differ.

Many real problems involve continuous variables as well as discrete ones which are much more complicated. The best mathematical techniques often have to be used for bad models in order to obtain small improvements. But these improvements are capable of bringing immense profit.

High-tech companies are trying to automate oil field management as much as possible. And IBM offers them this service - the intelligent oil field.

Today IBM has five powerful research centers in which intelligent systems are being developed for oil fields: in Canada (Calgary), Norway (Stavanger), the United Arab Emirates (Abu Dhabi), China (Beijing), and Russia (Moscow).

Q: Please tell us more about the opportunities that intelligent oil field management systems provide.

A: The intelligent oil field presumes using sensors situated in pipelines, pumps, and throughout the entire field. The data obtained from these sensors makes it possible to optimize the technological parameters of the wells. Information is fed into the company's analytical systems in real time and makes it possible to improve the oil and gas withdrawal rate from the fields. The intelligence of this approach is manifested not only in work organization, but also in the possibility of predicting difficulties and even preventing them by automatically making corrections and lowering the risks for people and the environment.

Throughout this entire production pro-cess, companies have to gather and process large data sets in real time. Masses of data comparable to the amount that can be stored on 200 DVD discs can come from just one reservoir basin a day. Thorough analysis of this information is important for making optimal decisions related to field survey, production, and organizational management.

Intelligent field development envisages integration and processing of geophysical and other data that are then used in three-dimensional modeling of reservoir basins. It makes it possible to find previously unattainable oil and gas reserves lying in hard-to-access places and in the depths of the ocean.

Drilling wells is preceded by gathering information about the volumes and quality of the oil and gas reservoirs. This minimizes the drilling area and the risk that surveying will not yield results, while raising operation security and reliability.

Q: What other essentially new areas in IBM service would you note for the oil industry?

A: A new interesting vector in which IBM is becoming increasingly involved is working with unstructured data. You probably remember that a few years ago an IBM computer played chess with Russian world champion Garry Kasparov and the computer won. This was a great challenge and achievement for IBM. And just recently we built an even more powerful computer, Watson, capable of working even more efficiently with unstructured data. Oil companies may be interested in processing unstructured data. Let us suppose that a company has an oil field with specific characteristics. The oil workers would like to know how many other fields there are in the world with similar properties. These and many other data relating to the oil industry may be available globally, but they are not structured. Correct analysis of these data can help a company's management to make correct decisions on field development. And such data can easily be analyzed with the help of supercomputers.

We are actively working on introducing the principles of unstructured data analysis in medicine. Watson can help to diagnose illnesses. Based on the symptoms and a medical analysis, the computer compiles the list of possible illnesses, then methodically asks clarifying questions to generate a more precise diagnosis. We are not saying that the computer is now ready to replace physicians, but the availability of this kind of computer program could be of immense help to medical institutions. A computer is capable of analyzing all the available medical literature in the world and providing physicians with real assistance. Such data-processing principles can also be applied to the needs of the oil industry, and we at IBM are intensively working in this direction.

Q: In that case, can it be said in business today that "technology solves all problems?"

A: Not entirely. The global trend today is such that information technology is largely receding into the background, while analytics aimed at optimizing business management is coming to the foreground. Of course, seismology is an unshakeable exception in the oil and gas industry - for objective reasons it is strongly oriented toward IT.

Of course, ideally it would be good to resolve increasingly difficult problems and do this all the faster, using the best machines. But whereas 15 years ago the main emphasis was indeed on computer science, "hard" and "soft," today mathematics and analytics are increasingly moving into the center of attention.

How do we use particular algorithms, how do we explore large data sets, how do we deal with unstructured data, how to mix discrete and continues variables, how to make fast decisions?

Of course, to answer these questions we use computers and data, but the most important element is algorithms, analytics.

Q: What do you think about the potential of information technology in the sphere of alternative energy? Is IBM participating in projects involving green energy?

A: IBM is showing immense interest in green energy. For example, in 2010, one of IBM's research laboratories constructed an element for an efficient solar battery. The main layer in the photo element, which absorbs a large amount of light for transformation into electricity, was made entirely out of relatively available elements and a special approach was used to create it using nano particles, in contrast to the widespread but more expensive vacuum deposition process. This element of a solar battery established a new world record in efficiency, and its appearance demonstrated potential for more mass and commercially profitable production of solar energy.

On the whole, green energy is influencing the situation in the world. Nevertheless, many alternative sources of energy are unstable. We still do not have truly efficient ways of storing the energy generated by instable solar panels or wind turbines. We have to admit that we are still children in this sphere.

One of the bottlenecks in alternative power generation today are batteries. IBM has made enormous investments in creating accumulator batteries that operate on entirely new principles, particularly lithium-air batteries. Such batteries make it possible for an electric drive car to travel at least 500 miles on one charge, and these batteries are much lighter than the traditional batteries used in electric cars today.

Our company is also doing a great deal to improve power supply systems as a whole. In 2007, IBM created the Global Intelligent Utility Network Coalition with the goal of accelerating the introduction of intelligent utility network technology. This structure includes eleven innovative energy companies servicing around 100 million electricity consumers all over the world. The development of intelligent utility networks will make it possible to reduce power outs and failures, efficiently manage consumption, and integrate renewable energy sources into the electric power system.      - Vladimir Akramovsky



© 1998-2020, "OIL OF RUSSIA".
"OIL OF RUSSIA" magazine welcomes comments and ideas from its readers.
Letters should be sent by regular mail, fax or e-mail.
All right reserved, including right of reproduction in whole or in parts in any form.
Design: exdesign.su
@Mail.ru