27
TECHNOLOGY Bloomberg Businessweek October 30, 2017
THE BOTTOM LINE Barrick says its Cortez mine in Nevada has
already cut per-ton digging costs by more than 25 percent by
analyzing data from thousands of sensors throughout the mine.sensors—hundreds of thousands rather than the
tens of thousands likely to wind up in Cortez. But
Williams says the advantages over traditional walkie-
talkies are obvious. Sensors track how many bolts a
given worker fires each hour and how many buckets
of waste rock his teammate moves. At a micro level,
the mine’s analysis software tells managers which
workers are underperforming and whether that’s
because of employee error, fatigue (if, say, perfor-
mance flags over the course of a shift), or a type of
rock that’s especially tough to handle. At a macro
level, all those data feed into executive- level reports
on overall production, but they also arm workers
with the information to suggest changes that can
be made midshift in a matter of hours rather than
in days or weeks.
For example, miners might use Wi-Fi-connected
mobile devices to flag a certain chemical in cases
where it’s being depleted faster than expected.
While the system would try to figure out why, it
would also generate a list of potential consequences
and alert suppliers and others to head off those
risks. At a mine in Argentina, where Barrick has
faced pressure and criticism from local officials
and the public for repeatedly spilling cyanide solu-
tion, the company is testing real-time monitoring
of water quality.
At Cortez, logistics changes such as rerout-
ing trucks and reassigning miners are relayed to
work crews midshift through the C0dem1ne team’s
Short Interval Control app. A small fleet of under-
ground loaders essentially runs on autopilot, con-
tinuing to work when humans can’t. Another app
specifically monitors the maintenance of Cortez’s
350-ton haul trucks, each of which requires enough
diesel to light a small city block. The trucks moved
125 million tons of rock at Cortez last year, and
their upkeep totaled about $60 million, so avoid-
ing the domino effects of a breakdown can save a
lot of money.
Beyond traditional predictive maintenance,
such as engine-temperature and fluid-level indi-
cators, Barrick’s systems route data from each
truck to the central computers to assess overall
fleet issues well beyond what other software can
do, according to Jamie Dwyer, who led develop-
ment of the predictive-maintenance app. “Because
we’re building it ourselves, we can merge it all with
our other data, right down to who was driving the
truck when it broke down,” Dwyer says.
Even so, skeptics question whether Barrick’s
C0dem1ne team can really outperform the pro-
fessionals. Other mining companies, including
Barrick’s chief rival, Newmont Mining Corp., have
implemented a range of high-tech upgrades but
stopped short of in-house coding. “Are you better
off spending $100 million for a proven tool that
somebody else de-risked or gambling that you’ll
get it right the first time?” says Mark Mills, who spe-
cializes in energy technology analysis as a senior
fellow at the Manhattan Institute, a think tank.
Barrick employees say they opted to turn off
Caterpillar Inc.’s predictive-maintenance software
on their trucks once it was clear their own app
performed better. “Some of the vendors we were
dealing with told us we were crazy to try this,”
says David Yazzie, a C0dem1ne project manager.
“We’re absolutely going to prove them wrong.”
—Danielle Bochove From left ○ Barrick’s
Cortez gold mine
in Nevada is being
extensively wired for
Wi-Fi and data analysis
○ Workers are using
more semiautonomous
and remotely operated
vehicles ○ A custom
app helps predict
maintenance needs
for the mine’s fleet of
350-ton haul trucks
○ Some 1,300 workers
operate beneath the
desert sands ○ A
central operations
center can analyze data
from the mine’s sensors
to suggest efficiency
improvements midshift