Explor, Halliburton and AWS Collaborate on Cloud-Based Seismic Data Processing

By Tim Shea

Category:
Acquisition or Partnership

In collaboration with Halliburton and AWS, Explor has achieved a milestone with seismic data processing in the cloud.  Explor successfully ran Seismic Processing, a DecisionSpace 365 cloud application powered by iEnergy on AWS, leveraging a range of different elastic compute instances to optimize key seismic processing workflows.  In the first phase of a proof of concept, multiple benchmarking tests were run which demonstrated:

  • An 85% decrease in CDP sort order times: Tested by sorting 308 million traces comprising of 1.72 TB from shot domain to CDP domain, completing the flow in an hour.
  • An 88% decrease in CDP FK Filtering times: Tested with a 57 million-trace subset of the data comprising 318 GB, completing the flow in less than 6 minutes.
  • An 82% decrease in pre-stack time migration times: Tested on the full 165 million-trace dataset comprising of 922GB, completing the flow in 54 minutes.

 

For this project, Explor provided the 3D seismic dataset, data science and geophysical expertise.  Landmark, a Halliburton business line, provided the Seismic Processing, a DecisionSpace 365 cloud application powered by iEnergy and technical expertise, and Amazon Web Services (AWS) provided the cloud computing resources and a team of technical experts and Solutions Architects.

Working together on the project, Explor, Halliburton, and AWS were able to optimize the cloud solution to reduce total processing timelines by 90 percent.

The challenges driving this project were the need to process ever-increasing sizes of seismic data while delivering higher quality results at lower cost.  In recent years, the seismic industry has dramatically increased 3D seismic trace densities, with several companies (including Explor) breaking over the 100 million trace/km2 threshold.  Surveys exceeding a billion traces/km2 are now being planned.  This exponential growth in acquired seismic trace densities poses new challenges for seismic data processing as processors must deal with data volumes of several hundred terabytes for small or medium sized surveys, with large high-density surveys producing petabytes of data each month.  Difficult and volatile market conditions make capital investments in high performance computing infrastructure challenging and risky for processing companies.

Halliburton's Seismic Processing, a DecisionSpace 365 cloud application powered by iEnergy, built on SeisSpace software technology delivers a flexible, extensible, user-friendly seismic processing application/foundation.  This project delivered the first scalable cloud-deployed seismic processing engine as part of Halliburton's strategy to deliver lowest TCO in the age of digital transformation.

To achieve the desired scale and reduce the seismic processing timelines by 90%, different steps of the project required access to a variety of compute instances (total of 38K cores) and high throughput storage, which the teams were able to provision on-demand as needed.  AWS provides the most elastic and scalable cloud infrastructure to run these types of HPC applications.  With virtually unlimited capacity, engineers, researchers, and HPC system owners can innovate beyond the limitations of on-premises HPC infrastructure.

A diverse, 20-person, multi-disciplinary team completed the project during the COVID-19 pandemic, with team members working mostly from their homes in three different cities in two different countries, further proving the value of cloud computing to minimize risk, drive innovation, support collaboration and reduce costs.

Engage with ARC Advisory Group