American Oil and Gas Reporter - November 2022 - 39

SpecialReport: Smart IT And ESG
Laredo is in the midst of a robust digital transformation
initiative with Amazon Web Services that is contributing to our
focus on ESG as well as revenue generation. In 2020, Laredo
migrated its legacy systems to the cloud, beginning with
serverless, centralized data in a " data lake. " The data lake stores
both structured and unstructured data, providing a vast repository
of information that allows for scalable data processing, realtime
analytics, machine learning, and modeling through dashboards,
visualization, and mobile applications.
Employing Internet of Things (IoT) and smart devices helps
us collect and fully understand our data. These devices include
smart cameras, sensors and other digital devices that transmit
data in real time for storage in the data lake. The solution uses
computer vision-a field of artificial intelligence that enables
computers to interpret images-to help accurately identify and
classify camera images. Data feeds deep learning models that
detect adverse events. We are able to match these reactions with
operational data to develop a probability metric that then can be
monitored to predict when an adverse event may occur.
We believe predictive analytics may be a major revenue
driver in the near future, while also contributing to ESG
measures. Currently, Laredo deploys more than 70 smart cameras
and other IoT devices that can detect events related to pressure
relief valve emissions. This information is paired with real-time
equipment surveillance to detect an event up to 20 minutes
before it actually happens. We already have been able to prevent
some adverse events and are working toward implementing
even more robust processes. The plan is to use automation to
proactively prevent even more adverse events, which means
more product will be kept in the sales pipeline.
Another benefit of IoT cameras has been improved physical
security. We have been able to recover stolen or missing items,
such as $20,000 worth of wire that never showed up on location.
In the first four months of 2022, we saved $250,000 on legacy
camera solutions, and we continue to save about $1,200 a day.
The budding partnership between the IT department and
field engineers enables rapid iteration of digital solutions.
Engineers realize that the use of automated solutions frees them
to focus on more substantial tasks rather than spending valuable
time monitoring equipment, and now reluctant vendors are
letting us tap into their systems.
For example, one vendor who supplies compression services
initially did not allow Laredo to access digital readouts from
their units. Instead, we pointed cameras at the gauges to capture
the analog readings, then used computer vision to convert
analog gauge readouts into digital information. Our system
ingested the real-time digital data and sent it to the cloud for
advanced analytics.
As a result, we improved compression efficiency by 50%,
directly correlating to increased revenue. Our successful implementation
convinced the vendor to allow us to tap into its
system, and internally our solution was a great proof of concept
to show how we could use data to increase revenue.
ESP Optimization
In 2022, Laredo has two primary areas where digital
technology is driving revenue and simultaneously supporting
ESG initiatives: electrical submersible pump optimization and
dynamic routing.
ESPs can achieve high daily production rates, but they are
susceptible to high gas-to-liquids ratios, solids in the production
Smart cameras, IoT sensors and other digital devices are critical
in collecting and transmitting data in real time for storage in
Laredo's data lake. The company is using artificial intelligencebased
computer vision to accurately interpret and classify
camera images to feed into deep learning models to detect and
even predict adverse events.
NOVEMBER 2022 39
stream, etc. By leveraging data and machine learning, we
worked to tune the pumps to operate more effectively, starting
with a 20-well pilot program. In the first 30 days, production
increased so significantly that the uplift translated into $2.5
million in additional revenue.
ESP optimization relies on algorithms built around data
from multiple different sources housed in our data lake. Laredo
data scientists can tune algorithms to specific business strategies,
whether that is pump longevity, increased production or another
desired outcome. The algorithm goes a step further to show the
expected result from a suggested change and data scientists use
the results to continually tune the algorithms. We built the algorithms
with what we call a " human in the middle. " For now, engineers
can accept or reject the algorithm's recommendation
and gradually build confidence as we move toward fully autonomous
solutions.
We have two ways of preventing an algorithm from making
drastic recommendations. First, we constrain the inputs to human-specified
operating parameters, such as limiting the temperature
to below 160 degrees Fahrenheit or restricting the
speed of the pump to between 45 and 65 hertz. Second, the recommendations
cannot result in changes that exceed a specified
number of units at a time. For instance, the pump cannot
increase or decrease speed by more than three hertz or 40 psi.
Because our pumps are more efficient, they have longer run
American Oil and Gas Reporter - November 2022

Table of Contents for the Digital Edition of American Oil and Gas Reporter - November 2022
