Machine Learning Assisted Real-time Fluid Identification

Nominated for The Prize

This project has been nominated for the Exploration Innovation Prize 2023.

Read more about the prize and be sure to vote for your favourite nominee by May 10^th.

Identifying the reservoir fluid content while drilling has long been the goal for geoscientists for an integrated fluid-rock reservoir characterization. This is by some claimed to be the last puzzle in the reservoir characterization which recently has been accomplished by Equinor.

Real-time fluid identification is direct proof of how a combination of geoscience discipline expertise (PVT and geochemistry, along with rock lithology and properties characterization), and machine learning can generate a technology that benefits the industry. To fully characterize a reservoir from its lithology and its fluid content while drilling, unlock great and new industry opportunities.

While the rock properties can be characterized by standard logging tools (e.g., gamma ray, density-neutron, and resistivity log), information on the fluid content traditionally require geoscientists to take samples from a few points in the reservoir and perform PVT post-well fluid analysis in a laboratory  This is both time-consuming, expensive and this practice creates a gap for a fully integrated digital reservoir characterization due to the delay and the scarcity of fluid information (only analyzing a few samples).

Utilizing mud gas data in a new way

Equinor has demonstrated that they can identify reservoir fluid content in real-time by analyzing mud gases released during the drilling process. Mud gas refers to the gas that is released during the drilling process. It is a mixture of gases, such as methane, ethane, propane, carbon dioxide, and hydrogen sulfide that are trapped in the formations being drilled. Mud gas measured continuously and generates a data set through almost the entire well.

Using an internally developed machine learning model, Equinor’s reservoir-fluid-identification system compares a database of more than 4 000 reservoir samples against samples of mud gas collected in real time. The output not only discriminates between water and hydrocarbon, but also hydrocarbon properties.

Equinor’s new digital innovation changes the way the industry utilizes mud gas data from traditional post-well analysis to real-time interpretation, and from primary exploration focused to broad implementation for production wells. The value creations are well recognized.

In summary, some of the main benefits are:

1. Improved drilling efficiency: Traditional methods of fluid identification can be time-consuming and expensive, requiring physical sampling and laboratory analysis. Equinor’s real-time identification system provides rapid and accurate identification of reservoir fluids, enabling drilling operations to proceed more efficiently.
2. Reduced costs: The ability to identify reservoir fluids in real time can significantly reduce costs associated with traditional fluid identification methods, such as laboratory analysis and sample transportation. Simultaneously, it provides large reductions in CO2 emissions.
3. Enhanced safety: Identifying the presence of contaminants such as hydrogen sulfide in real-time can help ensure the safety of workers and reduce the risk of environmental damage.
4. Valuable reservoir insights: Real-time identification of reservoir fluids can provide valuable insights into the composition of the reservoir, including the presence of hydrocarbons, other fluids, and contaminants.
5. Versatility: The real-time identification system can be applied beyond drilling operations to production logging, reservoir monitoring, and stimulation operations, providing a range of potential applications.

Their new technology was featured in the Journal of Petroleum Technology in 2021, and in 2022, Equinor was given the Best Data Management and Application Solution Award during the World Oil Awards in Houston, Texas. Equinor’s reservoir technology specialist Tao Yang was also awarded the OG21 Technology Champion 2021 for this work. In April 2023, Yang was also given an honorary award by the Norwegian Academy of Technological Sciences.