Oil the new Black Gold and the Gas Industry can be said Black Gold Mine. We all know about yellow gold and how the fight for it has created and destroyed many dynasties. Therefore, is very vital to properly manage gold be it Black Gold by visualizing the bigger picture around it.
The oil & gas industry can be transformed through automated drilling and real-time data. Yes, this is the current need.
Unprecedented situations like the COVID-19 crisis have made us realize that the industries need to prepare not only for Supply shock but also for Demand shock.
COVID-19 has accelerated the trend of decline in oil prices which started a few years back, due to the global efforts for the transition towards renewable and sustainable sources of energy. But due to modernization and globalization, few estimates suggest that oil demand could peak as early as 2025, rather than 2040 as projected by the reports earlier.
European oil giants like BP, Shell, Total, Equinor, and Eni have been ahead of the transformation curve when it comes to making the transition to alternative and renewable sources of energy.
But this transition is critically challenging given the complexities in the processes and the cash flows involved at a time when the need for research and development is more.
So, is there a way out? Well no. But Innovation & digital transformation will have a big role to play here. AI will play a significant role in order to drive the transformation in the Oil and Gas industry. AI will lead to the reduction in the cost of processing, production, and transportation, by transforming their value chain and operations.
Narrowing the view of data
In order to achieve efficiency and cost optimization, the oil and gas companies have structured themselves. Due to this, they do not have a single view of data and applications because they usually operate in siloes. This type of traditional structure makes it practically impossible to take on the larger cross-functional ML (Machine Learning) problems.
As discussed above AI is the way out but, in such industries, AI adoption has been restricted mostly to localized point solutions so far. Thus, it could not lead to creating a broader impact. As a result of which only 29% of the organizations in the energy sector have concluded that their AI deployments have been working satisfactorily.
For example, A common use case for Machine Learning (ML) is bringing down non-productive time (NPT) and Improving drilling efficacy.
However, rather than taking a comprehensive approach that uses data from across different rigs, most algorithms address very specific aspects such as torque on mast motor for stuck pipes.
Data quality & Cross-functional approach
Cross-referencing, transformations, duplication, and merging of cross-domain data sets remains a huge challenge for oil and gas organizations because the majority of their data resides in the form of documents, scanned assets, and reports. They have abundant data, but little is in the usable form.
Oil & gas giants are transitioning towards integrated data lakes managed by a cross-domain governance team.
With the help of new standards such as Open Subsurface Data Universe (OSDU), companies are consolidating data operations. Companies are taking to integrated data lakes that are equipped to make cross-domain data accessible. Above all this data could be transactional, unstructured, real-time, IoT, or even Petro-technical data.
There are instances where with the help of integration of operational data for ML-Driven analytics companies have made considerable progress on predicting the operating parameters for enhanced oil recovery.
Integrating Machine Learning Models with Physics and Engineering
We all know that operations related to Oil & Gas sector are highly critical with respect to safety concerns. Therefore, decisions arising from ML Models should be aligned & tested consistent with the principles of Engineering and Physics along with reasonable explanations.
Therefore, in order to enable AI-powered operations in decision-making, the alignment of machine learning (ML) models with physics-based simulation models are key to the solution.
We’re seeing the rise of a different programming paradigm that allows for the augmentation of data-driven AI with physics including,
1. Physics informed neural networks – A deep learning framework that enables the synergistic combination of mathematical models and data specifically applicable for problems involving nonlinear partial differential equations. (Raissi, Perdikaris, & Karniadakis, 2019)
2. AI Feynman – A recursive multidimensional symbolic regression algorithm that combines neural network fitting with a suite of physics-inspired techniques.
While the technologies are still at the nascent stage, it is evolving in this direction which can help the explainability of the machine learning (ML) models used in decision making.
Unplanned downtime can be mitigated, bottlenecks in the process, the workflow can be limited and safety incidents and be reduced with the help of deep insights derived with the help of (machine learning) ML models.
See into the Future
Now as the world is moving towards a new normal of lower oil demand, the emergence of unlikely alliances and a greater push towards driving process efficiencies are expected.
As cross-functional silos melt away and the broader picture of consolidated data becomes clear there will be greater emphasis on integrated data and governance.
These models will be governed by real-time data from wells on drilling & production, as well as past drilling reports. We can also expect the emergence of new engineering workflows that are governed by reservoir and geotechnical interpretation.
The emergence of goal-driven automation in drilling, where the system dynamically calibrates and plans logistics including personnel, equipment, and consumables is likely to be seen in the near future. The role of AI in the oil & gas industry will certainly be front and center in the future.
However, the oil and gas industry will continue to demand more scalability and functionality from AI adoption. The ability to integrate engineering technologies & physics with data-driven analytics, machine learning (ML), and automation will prove to be a huge success factor for the oil and gas sector.