Cloud-enabled Subsurface Modeling

Project Details

Project Lead
Ravi Vadapalli 
Project Manager
Ravi Vadapalli 
Project Members
Akash Pargat, Alan Sill  
Institution
Texas Tech University, High Performance Computing Center  
Discipline
Petroleum Engineering (113) 
Subdiscipline
11.07 Computer Science 

Abstract

Cloud computing has a significant promise in hydrocarbon reservoir modeling and exploration studies. Reservoir modeling involves sparse data and therefore, multiple realizations, stochastic models are necessary to reduce uncertainty in reservoir characterization and management. Localized cluster computing environments can harness the scalability of reservoir simulations and reduce the simulation runtime. However, they are not growing as fast as the amount of work and data demands in this application area. As a result, cloud computing is becoming an emerging computing standard and the main purpose this project is to deploy a cloud based solutions for reservoir modeling applications and evaluate various standards and APIs for their interoperability, and areas for future development. 

Intellectual Merit

Harnessing the scalability reservoir modeling and simulations, the intellectual merit of this project is to (1) evaluate the interoperability of cloud standards applicable to this application area, and (2) facilitates the comparison of performance characteristics between local cluster, grid, and cloud computing environments.

Broader Impacts

The effort could be used to demonstrate the scope of advanced distributed computing in petroleum engineering. A portion of this work will be used to develop/demonstrate educational content for the future engineering workforce in petroleum engineering, computer science and high performance computing.

Scale of Use

Initially the resources will be used to test existing capabilities with FutureGrid and their interoperability with reservoir modeling applications. During the course, simulation models will be developed and at that time the usage may increase. Until then, the scale of usage is minimal.