https://www.consultancy-m...drocarbon-reservoirs

Date - 08.23.2022

"Using technology that has its origins in the world of flowers (where AI-engines identify flowers among thousands of different species), ADNOC deployed a solution that identifies classes of carbonate rock.

The solution works as follows. First, geologists feed high-resolution rock images into a database. The solution then matches the images with a database to classify the rock. Using IBM Watson – IBM’s artificial intelligence engine – ADNOC’s geologists can then more effectively digitally construct reservoirs – and simulate all kinds of decision-making factors.

Engineers construct reservoir simulation models to test reservoir behaviour, including storage space (porosity), the ability to flow (permeability) and the amount of oil (potential recovery). The models allow engineers to consider different development characteristics, including well spacing, the type of well, the number of wells and pressure maintenance schemes.''

Observations -

ADNOC use case is a neat example of transfer learning.- branch of ML that focuses on storing knowledge gained while solving one problem and applying it to a different but related problem.

Likewise, we can take ML model trained for speech recognition (some ordered series) and adapt to do decline curve (time series) analysis. Example - Long Short-Term Memory (LSTM), a kind of Recurrent Neural Network, fits that bill.

If you have come across other examples, feel free to share with the community.