DI G ITAL TR AN S FOR MATION
Supercomputers and CCS key
to Petrobras’ efforts to build
a low-carbon energy ecosystem
NOC relying on powerful data processing
to reduce number of wells drilled while seeking
regulatory support to develop industrial CCS
BY STEPHEN WHITFIELD, ASSOCIATE EDITOR
Meeting the demands of the energy transi-
tion will require operators both to adopt
new technologies that can improve effi-
ciency in oil and gas E&P and to deploy
low-carbon solutions like carbon capture
and storage (CCS).
Petrobras is aiming to reduce its envi-
ronmental impact by combining low-car-
bon energy development with comput-
ing systems designed to reduce geologi-
cal uncertainties in exploration and, thus,
reduce the number of wells needed to
develop new fields. This will lead to sig-
nificant cuts in emissions while enabling
increased efficiencies, which will allow
the company to continue exploring new
offshore frontiers, said Joelson Mendes,
E&P Executive Director at Petrobras.
“In our vision, oil and gas is an enabler
of the low-carbon economy,” Mr Mendes
said. “We’re revisiting our strategy and
bringing light to our priorities regarding
the energy transition, but at the same time
we’re reinforcing that oil and gas produc-
tion is compatible with the energy transi-
tion. A digital and data-driven exploration
and production strategy will enable us to
create positive results here.”
During the 2023 Offshore Technology
Conference in Houston on 2 May, Mr
Mendes described Petrobras’ vision for a
“low-carbon energy ecosystem,” or a busi-
ness that integrates E&P in frontier areas,
such as the Brazilian Equatorial Margin,
with low-carbon energy sources such as
offshore wind energy and hydrogen produc-
tion, to reduce its environmental footprint.
Data analytics and computing power
are lynchpins to this ecosystem. Over
28 the past couple of years, Petrobras has
launched a pair of new high-performance
computers (HPCs), Pegasus and Dragon,
that are running machine learning algo-
rithms to process geological and geo-
physical data. Pegasus has a processing
capacity of 21 Petraflops, roughly equiva-
lent to the processing power of 150,000
laptop computers, while Dragon provides
14 Petraflops.
Pegasus and Dragon currently rank 35th
and 73rd, respectively, on the Top500 proj-
ect’s list of the most powerful non-dis-
tributed computer systems in the world.
Petrobras also has two older HPCs, Fenix
and Atlas, that came out in 2019 and 2020,
with less processing power.
Because processing geophysical data to
create seismic images involves complex
mathematical equations, having these
HPCs has led to a significant reduction in
data processing times, Mr Mendes said.
He also noted that the four HPCs provide
Petrobras with enough processing power
to deploy high-usage third-party software
programs like EXP100, which uses AI and
machine learning to predict the likeli-
hood of hydrocarbons being present in a
given reservoir. That reduces the number
of exploration wells needed and, in turn,
reduces the company’s emissions.
“We have a database with 70 years of
operations, a collection of a vast amount
of data from different fields and develop-
ments. Machine learning is necessary to
help us capture and process all of this data
that’s available to us, and it’s also impor-
tant to have sufficient processing capacity.
High-performance computing is a step-
As Petrobras expands its CCS efforts,
the NOC is seeking to establish a regu-
latory framework in Brazil around eco-
nomic activities related to CCS, said
Joelson Mendes, E&P Executive Director.
Mr Mendes spoke at the 2023 OTC on 2
May in Houston.
ping stone to achieving greater efficien-
cies in our operations. We’re foreseeing an
increase in the use of predictive models
that can help us reduce the number of
wells we have to drill,” he said.
Low-carbon energy efforts like CCS
make up the other side of the business
ecosystem that Petrobras is building for
the energy transition. In its 2023-2027
Strategic Plan, the company outlined its
ambition to store 80 million tons of CO 2 by
2025. In 2022, it stored 10.6 million tons of
CO 2 , which accounted for approximately
25% of the CO 2 stored globally that year,
according to the Global CCS Institute.
Mr Mendes noted that Petrobras is
studying the implementation of a CO 2 cap-
ture and geological storage hub. The proj-
ect would involve the construction of pipe-
lines from industrial facilities and E&P
projects into a saline aquifer reservoir with
a potential storage capacity of 25 million
tonnes/yr. A pilot project is under development
near a Petrobras-owned natural gas pro-
cessing facility in Cabiunas, Brazil, to
sequester up to 100,000 tonnes/yr of CO 2
J U LY/AU G U ST 2023 • D R I L L I N G C O N T R AC T O R