Estimation of the reservoir quality of Triassic sandstones offshore U.K. from cuttings analysis and machine learning: A CASE studentship with PETRIVA Ltd.

Triassic sandstones in the southern North Sea, U.K. contain an extensive saline aquifer that has a huge potential to store CO2 and hence contribute to the UKs NetZero targets. There have been very few petroleum discoveries in this sandstone so very little core has been taken and so little is known about its reservoir quality. Numerous wells have penetrated the Triassic on their path or Rotliegendes and Carboniferous gas reservoirs. These wells have brought to the surface small pieces of the Triassic sandstone known as cuttings. This project aims to study the microstructure of the cuttings using scanning electron microscopy and then use machine learning techniques such as convolutional neural networks (CNNs) to estimate porosity-permeability relationships. This would allow reservoir characterisation to be conducted rapidly and cheaply so could potentially have a significant impact on the expansion of CCS in the U.K. 


The project would provide the student with a thorough training in microstructural and petrophysical property analysis as well as machine learning. This would be a CASE award in which the student would spend 3 months working with the University spinout company, PETRIVA Ltd, in a commercial environment. So as well as potentially making a significant contribution to reaching Net Zero the student would get the opportunity to be involved with industrial applications of petrophysics, data visualization and data-mining.


Ideally, the successful research student will have a strong background in computer science. Knowledge of petrophysics or geoscience is not a pre-requisite as this will be taught during the research project.

Figure 1. General work flow being developed to classify the microstructure of images using convolutional neural networks (CNNs).