Subject ▸ reservoir simulation

Reservoir Simulation Datasets

README Purpose of these datasets is using them for testing using MRST and its derivative Proof of Concept (POC) in Pyhon using the machine learning library PyTorch. Since PyTorch has built-in functionality to work with Graphics Processing Units or GPUs, we expect demonstrating -before embarking in the full porting of MRST-, that PyTorch GPU-based tensors could significantly reduce compute time during reservoir simulation. Evaluation for Proof of Concept The steps are the following:

Read More…

Integrating Python and R for data science. Converting Eclipse binary files to dataframes in the Volve dataset

Introduction Python and R offer a good combination of powers: dozens of proven engineering, data science, and machine learning libraries, also a science oriented approach towards full reproducibility. As I have told you before, I started my coding journey with Python many years ago. I even wrote a large application for production optimization using OpenServer, Prosper, GAP and MBAL by Petroleum Experts while I was on my 3-year tour with Petronas in Kuala Lumpur in Malaysia.

Read More…

A reproducible comparison of the Volve reservoir model

Introduction Continuing with the previous article The fabrication of an artificial intelligence agent for reservoir history matching from the Volve dataset, and the generation of a master dataset for an AI agent to perform history matching of reservoir models, we will extract additional data from the output of the Volve reservoir model, the PRT text file. This is the output “as-is”, as we found it. No additional simulation runs have been performed over this model.

Read More…

The fabrication of an artificial intelligence agent for reservoir history matching from the Volve dataset

Introduction History matching is one of the core activities performed by petroleum engineers to decrease the uncertainty of reservoir models. By comparing real data -production data gathered at the surface-, with the output from a reservoir simulator, the engineer starts filling in the gaps in reservoir properties of those block cells in the model. And this what makes it so interesting in data science, and ultimately, in the fabrication or construction of an artificial intelligence agent.

Read More…

Volve dataset. First look of the Reservoir simulation models.

The Eclipse reservoir models from the Volve dataset working like a charm. The compressed file is 399 MB in size.

I was able to open the models with ResInsight (thank you Matthew Kirkman). The software is open source and relatively easy to use.

Here is the Eclipse case opened.