Digital Transformation in Oil and Gas Industry and Application of AI
The use of AI and Data Science in Oil and Gas Industry is not covered and popular in most of the data science community site. For this reason, I would like to give you an overview of the types of data used in Oil and Gas Industry and how Oil and Gas Industry is preparing for Digital Transformation especially in times of uncertainty by undertaking cost optimization using various automation processes and AI implementation.
The digitization in the oil and gas industry can be traced back at least a decade when most of the major oil companies like BP, Shell and Chevron started their smart field transformation journey. Though they named it differently; however, the concept was the same. Chevron named it as i-Oil field (Intelligent Field), Shell called it Smart Field and BP approved it as Field of the Future.
Since the beginning of the smart sector and the revolution of digital transformation, we have come a long way and COVID-19 has further accelerated the adoption of digital transformation. However, with data being the cornerstone of such a transition, the lack of structured, high-quality, easily accessible data in a timely manner has been one of the biggest challenges to this digital transformation process, making it difficult to realize the full potential of digital transformation. High quality and timely data are crucial, particularly with regard to closed-loop optimization, to ensure that the models are still ‘green’ and accurate, and therefore the optimizations based on them are reliable; otherwise, it would be catastrophic to enforce these ‘optimal’ decisions.
Application of AI and Data Science in Oil and Gas Industry
The decisions relate to Oil and Gas exploration, production and development are based on a huge amount of data acquisition, processing to interpretation and finally, integration of all interpretation to build a predictive reservoir model. The data volume increases daily with new data acquisition, processing, and interpretation. Many types of data captured in different formats and have used them to create a subsurface model (different layers ranging from 5,000–35,000 feet below the surface). The data ranges from — 2D-3D seismic data (.segy format), wireline log data from drilled wells (.las files), borehole image logs (XRMI or FMI format), core data from wells, fluid sample data from well, mud log data, etc. However, it is a challenging task to keep all this data on one platform so that everyone working on a project from processing geoscientist to interpretation geoscientists, petroleum engineers and reservoir engineers can access it from the same place. The software companies have utilized some of the technological developments from 1990 by enhancing the modeling software by integrating the data into one platform. The concept of using Data Science and Machine Learning techniques is a very new concept in Oil and Gas Industry now compared to other industries. During the next decade, it is necessary to focus on the ways to use all the data that the industry generates to get automated simple decisions and guide harder ones, which will ultimately reduce the risk in finding Oil and Gas. This technique will also be beneficial in producing more oil and gas with less environmental impact. The completion time of a project will be reduced at the same time increasing the accuracy in project deliverables.
Data mining, Data processing, Data management and Predictive analysis
The huge number of datasets is rich in the three Vs of big data (volume, variety and velocity). One raw seismic dataset is usually hundreds of gigabytes, resulting in terabytes once the processing and interpretation and after building the geological models. The increase of dataset is not limited here, once the drilling starts every minute keep on adding mud log data, wireline log data, borehole image data, production data, core data, etc.
Data mining and data processing are the 1st step to start with to build any predictive model which can help to make any decision. Recently, various machine learning, and deep learning techniques were used for the Reservoir Simulation model preparation and predictive analysis. Some of the widely used methods are Natural Language Processing techniques, Computer vision, image classification methodologies, Time series analysis, Bayesian Statistics, etc.
Some useful links to follow about AI applications in Oil and Gas Industry
Some of the YouTube videos will give you an overview of how Oil and Gas Industry is using AI. An interview with Dan Jeavons, general manager for data science at Shell can be found here. An overview of BP’s new AI-based platform known as ‘Sandy’ can be found here.
To get updates about AI techniques, Data Management and Data Science topics in Oil and Gas Industry in the UK you can follow the following links: OGTC website, Sword Venture news and blogs page.
To get an idea about the types of data I am also including one open-source oil and gas data source here. If anyone is interested to join any data science competitions in Oil and Gas industry topics by registering at xeek.ai.
I hope you enjoyed reading and learning about data science and AI application in the Oil and Gas industry.
References:
https://www.intellias.com/digital-transformation-in-oil-and-gas-a-remedy-for-market-volatility/
https://www.api.org/news-policy-and-issues/blog/2012/06/13/innovation-chevrons-i-field-links-perfor
https://www.geoexpro.com/articles/2015/08/smart-fields-people-processes-and-technology
https://www.youtube.com/watch?v=VxYpyPZXyFk
https://www.maritime-executive.com/article/sandy-joins-the-dots-for-bp
https://www.youtube.com/watch?v=9gSw-gw-zb8
https://www.venture.co.uk/sword-venture-blog/
