The Human Side of Artificial Intelligence
November 05, 2018
Author: Andrey Bogdan, Advisor, Advanced Analytics and Technology
We live in very interesting times as we witness technological revolution. Advances in such areas as deepwater exploration, 4D seismic, horizontal drilling and multistage fracturing along with the tools, resources and technology innovations are bringing incredible results to our industry.
Expansion of new practices in drilling and completion have not only made once uneconomically-viable wells, viable, but is driving the need for data analytics and machine learning.
Recently, I presented during the ShaleTech™ Hydraulic Fracturing Forum webcast
- Over the last decade, daily production from US shale has grown by a factor of 20, creating a new horizon for North American land-based operators
- The growing drilling and completion activity brings more dense and diverse data and an increased host of public data sources
- Planning and optimizing today’s complex multiwell design pads requires new approaches.
These combined factors are creating an opportunity for pure artificial intelligence (AI) methods to revolutionize pad planning and completion design. AI has the capability to speed up the entire process. It can analyze data that would normally take 6 to 8 weeks of manual research in minutes or even seconds, ultimately leading to reduction of nonproductive time.
In addition, AI is powerful in predicting well performance scenarios. As these technologies progress and become even more sophisticated, the data insights will become even more powerful. AI has the possibility to boost our analytical capabilities, inform the decision-making process, and elevate human creativity. However, a live person is still the one who ultimately makes the decisions.
Man’s Collaboration with AI
This is where a partnership and collaboration between man and machine can become invaluable. This collaborative intelligence allows human experts to achieve a better outcome than either one could achieve alone.
When looking at optimizing completion designs, there are several areas when collaborative intelligence can lead to better results:
- Curating the data: Each data type is complex and requires an experienced subject matter expert to deal with acquiring, formatting and managing the data sets.
- Data interpretation: Equally important, IT and domain experts must collaborate to ensure that the data are always correct, integrated and accessible. To be effective, a data manager must have a strong blend of domain and technical expertise that requires a degree in a relevant discipline such as geoscience, engineering, or information technology plus knowledge how to script in AWK, sed, Perl, R or Python.
- Informed Real-time Decision Making: In order to fulfill the outcomes of a proposed model or well design, people will have to determine if it is indeed the best course and take action.
For AI to gain full utilization and make the biggest transformative impact, artificial neural networks will need to be blended and stimulated by human creativity. Our industry will need to embrace this collaborative intelligence as it can transform decisions, improve operations and truly change the way we work. To achieve the next breakthrough for our market and industry, this combined predictive capability has the potential to spark revolutionary creative solutions and further unlock shale reservoirs.
To learn more about how data analytics and machine learning are impacting shale production, download our white paper, “Will Machine Learning Drive the Next Shale Revolution?"
About Andrey Bogdan
Andrey Bogdan is Advisor of Advanced Analytics and Technology for BJ Services. He brings over 11 years of leadership in technology development. Combined with expertise in data analytics and strong upstream domain knowledge, Dr. Bogdan’s specialties include inventing and leading new technologies startup, machine learning, completion optimization, hydraulic fracturing and wellbore isolation.
Dr. Bogdan holds a Master of Science degree and Ph.D. in High Energy Physics from Novosibirsk State University and Budker Institute of Nuclear Physics. He is an inventor of more than 15 patents and patent applications, two trade secrets and has co-authored 27 publications.