Oil & Gas

ENERXICO aims to have an impact on the exploration, production, transport and refining areas of the oil and gas industry.

ENERXICO enables the development and application of state-of-the-art computational and experimental tools for efficient industrial-scale hydrocarbon exploitation. HPC seismic modeling and inversion codes, highly consistent with real wave physics, are enhanced and evaluated within the project. Their main application is to assist on field assessment and imaging of complex Earth models, as demanded by modern oil and gas (O&G) exploration. Multiphase flows are precisely modeled during hydrocarbon production in naturally fractured reservoirs and during pipeline transportation, by means of novel and massively parallel fluid simulators with high physical fidelity. This project also aims at suitably exploiting in-silico models of new catalysts for crude oil refinement in order to reduce the dependence on traditional time-consuming laboratory experiments. Four industrial and academic partners in Europe (REPSOL, BSC, TUM, UGA) and seven similar institutions in Mexico (PEMEX, ININ, UAM-A, UNAM-FC , UNAM-IINGEN , CINVESTAV, IMP, IPN-ESIA) are jointly working in ENERXICO to improve the O&G efficiency cycle by providing and tuning these sophisticated exploitation tools.

Mitigation of uncertainties in reservoir simulation by improving the multi-physics modules of large-scale massively parallel multiphase modeling codes, to automatically combine the effects of interrelated physical and chemical phenomena. These improvements would enable accurate simulations of the in-situ recovery processes, accounting for data provided by the industrial partners.

ENERXICO´s objectives in O&G energies:

Explore data processing of fine meshes to capture geological features.
Forecast hydrocarbon production at reservoir scale.
Simulate experiments that can substitute expensive real laboratory exploration for new catalysts.

SEISMIC MODELING AND INVERSION

Specification

Energy type	Oil & gas
Collaborating institutions	REPSOL, BSC, TUM, UGA
Software involved	BSIT (BSC), ExaHyPE and SeisSol (TUM), SEM46 (UGA)
Main objective / mission	Highly accurate seismic modeling and inversion for geophysical subsurface imaging in complex and realistic geological scenarios
Target TRL	TRL 10
Relevant stakeholders	REPSOL (BSIT), UGA (SEM46) and researchers communities of seismic imaging (ExaHyPE and SeisSol)
Achievements up to M24	Surface wave modeling under different topography complexities and result comparisons under suitable accuracy measures SEM46 extension to fluid-solid coupled media highly amenable for seismic modeling/inversion of multi-component data in marine scenarios SeisSol extension to anisotropic and poro-elastic materials with intense validation ExaHyPE uncertainty quantification studies for subsurface parameters (Bayesian inversion)
Related work and further information	Two journal papers (in preparation, submitted) and three conference talks (SIAM 2021: GS21)

Exascale outlook

Exascale target (capability/capacity)	Exploiting exascale computing for seismic wave modeling in 3D real-scale complex geologies in the presence of hydrocarbon bodies, as those found in Mexican oil and gas prospects
Main benefits expected from large increase in compute resources	Computationally achievable numerical modeling, being more consistent to wave physics in coupled fluid-solid, anisotropic and poro elastic materials Processing and inversion of large seismic datasets as those acquired by industrial-scale marine surveys
Main challenges in harnessing large compute resources	Portability of our physically general and CPU-based codes may be challenging to small- or moderate-sized HPC systems
Specific actions towards exascale at present time	Optimization of the core propagator for accelerated architectures such as GPUs and FPGAs is an ongoing work

Computational details BSIT

	Number of Cores / GPUs	Memory (GB)	Storage (GB) both temporal and permanent	#Files written both temporal and permanent
Minimum	1	1GB	0	0
Average	N/A	N/A	N/A	N/A
Maximum	48000	18TB	200TB	~100,000

Computational details SEM46 (Restriction for distribution)

	Number of Cores / GPUs	Memory (GB)	Storage (GB) both temporal and permanent	#Files written both temporal and permanent
Minimum	32	1.2GB	2GB	6
Average	N/A	N/A	N/A	N/A
Maximum	19200	18TB	1TB	~100

Computational details SeisSol

	Number of Cores / GPUs	Memory (GB)	Storage (GB) both temporal and permanent	#Files written both temporal and permanent
Minimum	48	96GB	N/A	N/A
Average	4800	9.6TB	500GB	~100
Maximum	19200	38TB	500GB	~100

Images & videos