Rapid technological developments and emerging digital opportunities promise to transform the oil and gas industry for good. The Industrial Internet of Things (IIoT), big data, cloud technology, and advanced analytics will come together to provide companies with renewed insight into operations, increased productivity, and improved efficiency.
Many of these technologies will also fuel the development of the next generation of reservoir characterization.
Reservoir characterization methods have improved drastically over the last years, equipping geoscientists with better computing and software tools that produce models to understand both the potential and the risks of prospective drilling sites.
During the next five to ten years, however, reservoir characterization methods will continue to undergo further improvements. A happy marriage between cloud computing and advanced analytics methods such as machine learning will make up the building blocks of the next generation of reservoir characterization – promising a radically faster and effective process.
The next generation of reservoir characterization will rely heavily on the high computing power of the cloud. For the uninitiated, cloud computing is a catch-all phrase for everything from data processing and data storage to software on servers made available via the Internet. Microsoft describes it as the delivery of computing services, whether that is servers, storage, databases, networking, software, analytics, or intelligence, over the Internet.
For instance, the cloud offers significant efficiency opportunities for the oil and gas industry, as it promises to change and improve how companies access their applications, process their data, and manage their infrastructure. European natural gas operator GRTgaz, for instance, has used the cloud to raise its environment availability from 54 percent to 90 percent and reduce provision testing environments from up to 12 weeks down to roughly ten days, as highlighted by consultancy Accenture.
As one example, E&P companies, specifically, can combine the cloud’s high computing power with recently developed sophisticated seismic inversion technologies for improved performance and reduced costs. Instead of performing a seismic inversion on local workstations, algorithms such as ODiSI and other heavy number-crunching methods can take place in the cloud. This provides the interpreter with nearly unlimited resources and the ability to execute those algorithms within seconds.
But cloud computing also opens up a new avenue of possibilities, as the cloud is capable of storing massive information which can be analyzed for renewed insight into the reservoir characterization process.
Much in the same way as cloud computing is a term that encompasses several functions, advanced analytics is an umbrella term for techniques that models internal and external data to yield valuable insights and uses advanced computing techniques – such as machine learning. Whatever the method, the goal of advanced analytics is to extract useful information and insights from various data types to enhance decision making.
Simply explained, AI and machine learning are techniques that help us to manage large data sets intelligently. Often considered a subset of AI, machine learning enables computers to learn without being explicitly programmed, finding hidden patterns, and generating new insights.
Considering the enormous amounts of data that the oil and gas industry has acquired and collected over the last decades, AI and machine learning offer companies a tremendous opportunity to gain new insights, improve decision-making capabilities, increase productivity, and reduce costs. According to Accenture, oil and gas companies can see a return on investment that is almost four times the baseline from advanced analytics.
Consequently, AI and machine learning are steadily gaining a foothold within the industry. Consultancy McKinsey highlights a North Sea operator that applied advanced analytics to optimize production settings and raise production output. Shell is also heavily investing in AI. Among its many AI initiatives, Shell is in the process of deploying reinforcement learning for its exploration and drilling program, to reduce gas extraction costs. BP is a third example. They are using machine learning to reduce the time needed to pinpoint prediction models and boost the productivity of their data scientists when predicting the percentage of retrievable hydrocarbons in reservoirs.
In the near future, AI and machine learning will be fundamental building blocks of reservoir characterization. It will help geologists and geophysicists speed up the characterization process and draw on an extensive search net by searching through massive amounts of data on similar events that match the problems at hand.
The next generation of reservoir characterization will be somewhat similar to how intelligent search engines functions today. Take Google as a comparative example. When you type something in Google, they combine the typed-in words and use intelligent machine learning patterns to collect relevant information from everywhere matching your search queries.
We want to apply the same idea to geology. By combining the high computing power of the cloud and recent developments in advanced analytics, we aim to develop a meta-knowledge database for interpreters – a Google search-style database for reservoir characterization, if you will.
Imagine a geologist starting to work on a reservoir characterization project somewhere in the Barents Sea. Today, he or she would open Petrel* and upload his or her seismic data. Then, they would begin manually interpreting, drawing on a range of knowledge sources both internally and externally. In the next generation of reservoir characterization, however, the geologist will instead send the project to an intelligent system. The system will recognize that the reservoir is in the Barents Sea, and it will understand the geology based on previous experience. The geologist can then enter a search query – for example, “show me the potential sand channels that contain oil” – and the system will highlight appropriate locations for drilling within minutes.
The application of cloud computing and advanced analytics for reservoir characterization has yet to mature, and we still have some ground to cover before it becomes a viable and reliable alternative for interpreters. But we are getting close, and when it happens, we in Cegal are determined to be in the driving seat of the development of next-gen reservoir characterization.
Drawing on substantial experience with cloud computing, Cegal is currently in the process of developing a new software environment that allows you to perform reservoir characterization tasks at an unprecedented speed. A working prototype is already in place for any forward-thinking E&P company who wish to investigate its opportunities.
*Petrel is a mark of SLB