Streamlining subsurface workflows with Cegal Prizm

Modern subsurface exploration demands the efficient integration of diverse datasets, research breakthroughs, and machine learning (ML) models into established workflows. In environments where platforms like Petrel* are central, the challenge lies in incorporating advanced computational techniques without requiring every geoscientist to become an expert Python programmer. In response, Python APIs such as Cegal Prizm have emerged as a practical solution to streamline data workflows, automate processes, and bridge the technical gap between research and daily operations. 

Subsurface projects are inherently complex. Geoscientists routinely work with data from seismic volumes, well logs, horizons, and fault interpretations, all of which can be highly heterogeneous in format and scale. Key challenges include: 

• Technical complexity: Integrating new ML models, recent research findings or even proprietary algorithms into existing commercial platforms often demands substantial programming expertise. 

• Data interoperability: Moving data between Petrel and Python-based environments involves multiple conversion steps, from exporting data to converting it into formats like NumPy arrays or Pandas DataFrames and eventually reimporting the processed results.
  
  
• Collaboration hurdles: Managing multiple, distinct Python environments can hinder effective workflow sharing. Additionally, the successful use of these workflows generally depends on the end user's familiarity with Python.

Harnessing Python APIs for seamless integration 

Python’s status as a leading language for scientific computing and ML provides a natural foundation for addressing these challenges. By leveraging Cegal Prizm, subsurface workflows can be significantly simplified: 

• Direct data exchange: Cegal Prizm enables the retrieval of various data types from Petrel, converts it into a Python-friendly format, and support applying any modification directly to the data. For example, a pretrained fault detection model developed using synthetic 3D seismic images, can now be deployed directly within the Petrel environment. This approach drastically reduces the time required for fault segmentation compared to conventional methods.  
  
  
• Minimized coding requirements: The Prizm Workflow Runner tool provides an intuitive interface within Petrel, enabling non-programmers to execute Python-based workflows effortlessly. Developers can encapsulate complex routines behind a user-friendly UI, ensuring that advanced processes remain accessible to all team members.
  
  
• Streamlined workflows: The back-and-forth data handling (export, process, and reimport) is consolidated into a few intuitive steps, enabling a focus on interpretation rather than technical logistics. 
  
  
• Centralized environment: By hosting Python environments and scripts on cloud or on-premises platforms, teams can access consistent configurations without each member needing to set up and maintain their local systems. 

Illustrating using the Cegal Prizm Python API to connect a Jupyter Notebook to a Petrel project, retrieving a seismic volume, applying a pre-trained ML model, and writing the results back to Petrel. 

Bridging application silos 

Subsurface workflows frequently require the integration of third-party applications. For instance, connecting Petrel with mapping tools like ArcGIS typically involves laborious data export and format conversion steps. Python scripts can automate these tasks by: 

• Extracting well and seismic data from Petrel. 

• Converting the extracted data into formats such as GeoDataFrames, shapefiles, or geospatial rasters. 

• Enabling direct import into ArcGIS for further analysis and visualization. 

This automated process reduces manual intervention, minimizes errors, and accelerates the overall workflow, allowing geoscientists to concentrate on analysis and interpretation. 

Additionally, this integration method is versatile—other applications equipped with a Python/open API can be incorporated into the workflow, enhancing overall automation. This not only reduces manual steps and minimizes errors but also accelerates processes, enabling geoscientists to focus on interpretation and decision-making. 

This image illustrates exporting a surface from Petrel as a raster file and subsequently using ArcPy, ArcGIS's API, to import the raster into an ArcGIS project.

Conclusion 

By embedding advanced Python tools into Petrel workflows, Cegal Prizm exemplifies how modern ML techniques and streamlined data processing can overcome long-standing challenges in subsurface interpretation. This approach not only simplifies data integration and workflow automation but also enhances collaboration across teams. Ultimately, the focus shifts back to the data itself, allowing geoscientists to derive meaningful insights and make informed decisions with greater efficiency. 

Cegal Prizm’s technical innovations pave the way for a more agile and productive subsurface exploration process, where the complexities of data management and coding no longer stand in the way of subsurface interpretation and modeling advancement. 

Read more about Cegal Prizm >
Release notes Cegal Prizm - Python Tool Pro 2.6 >

*Petrel is a mark of SLB