The Blueback Toolbox suite contains close to 100 tools complementing standard Petrel. It covers the geophysics, geology, reservoir engineering, and the data management domains, in addition to assisting the Petrel user with a series of tools for better managing the Petrel projects. In 2016 we introduced the Python tool, a built-in integrated development environment for Python that let you interact with domain objects and create bespoke work steps which manipulate data.
Previously, the Python distribution (IronPython) was bundled in with the tool where it was self-contained and worked out of the box. However, it could not take advantage of all the work done in projects such as Pandas and scikit-learn.
To cater for the increased interest to apply data science and machine learning on geoscience data, we are introducing support for the reference implementation of Python CPython via the Anaconda data science platform. This allows for the Python tool to use external libraries such as NumPy, Pandas, Keras and Scikit-Learn. Example workflows:
- Pandas - manipulate tabular data, a simple way of exporting data from Petrel to a Jupyter notebook
- Sci-kit learn - Machine learning using Petrel data (clustering, regression etc.)
- NumPy - Numerical python, basis for scientific computing in python and the basis for e.g. pandas and scipy
- SciPy - Implement fourier transforms and perform signal processing
- Keras/Tensor Flow - Solve geoscience challenges with neural network functionality
For more in-depth details on the functionality and examples please read this LinkedIn article by Espen Knudsen, Digitalization Portfolio Manager in Cegal.