Model Behavior: Powering a Brighter Future with Industrial Data Science

Welcome to the fifth installment of Model Behavior. This ongoing series provides an inside look at Uptake’s industrial data scientists and the challenging work they do on behalf of our customers. Why is this topic so important? Because tough problems don’t solve themselves. See how model Uptakers build and deploy Uptake’s models — and how both of those elements must work together in harmony to create a world that always works.

Heavy industries are experiencing a fundamental shift. That shift is particularly evident in the energy sector.

Not only are energy businesses facing tough economic challenges around the world, the industry’s traditional model is being upended because:

  • The grid is becoming increasingly decentralized.
  • Renewables and retail power are on the rise.
  • Consumers are disrupting the status quo by gaining more choice and control over their energy usage.
  • Today’s industrial workers and energy subject matter experts — including the institutional knowledge they hold — are exiting the workforce at a rapid rate.

Energy companies are also having to do more with less, given the rising costs of operations and maintenance. The good news is that data is here to save the day.

Now Here’s the Catch

Uptake industrial data scientist Matt Dzugan is an expert at making catches. As a member of our Energy Team, Matt builds models that enable wind turbines to produce as much renewable power as possible.

Recently, we made a big catch on one of our customers’ wind turbines within the first few months of deployment. Our model correctly identified a calibration issue with the turbine — a problem that’s quite difficult to detect, even to the trained eye. Here’s why:

Despite the fact that turbines might still be online, rotating and generating energy, even the slightest miscalibration of its components can result in subtle inefficiencies. Those inefficiencies add up to a lot of lost energy over time. Not only does that mean lost revenue for the business, it means less renewable energy for society and a potential greater reliance on non-sustainable power sources.

In terms of the physics, each of the turbine’s three 100+ foot long blades are actively controlled by the turbine’s computer. While the blades are rotating, they’re constantly pitching ever so slightly, so that they’re capturing as much wind power as possible. In this catch, the turbine’s controller mechanism — which sends signals out to the blades to inform their correct positioning — was miscalibrated, causing a blade to consistently point out of the wind. In some sense, this is similar to an airplane attempting to reach cruising speed with its flaps deployed. The additional increase in drag prevents the plane — or in our case the wind turbine — from performing as efficiently as possible.

Uptake was able to quickly identify this issue because of the deep library of models we’ve built for wind turbines. Our models analyze data from hundreds of different sensors — including advanced weather sensors that aren’t located on the turbine — in order to understand the amount of power that a particular turbine is physically able to generate at any point in time.

Because our models factor in wind conditions, nearby turbine performance and other key variables, they can answer a fundamentally tough question: How much power should a turbine be producing right now? Most importantly, if our models discover that a turbine is actually underperforming — i.e., generating less power than is expected given the conditions — the issue gets flagged to the customer right away so that appropriate action can be taken.

Because Uptake is agnostic to the original equipment manufacturer (OEM), our technology can catch productivity issues that are universal to any turbine maker. We go to great lengths to pre-train our models to master the key variables that affect and explain power generation, such as wind speed, wind direction and other weather conditions, as well as historic turbine performance. With this depth of understanding of each individual asset, we deploy our proven models to detect when turbines are producing less power than they should be.

Armed with this actionable insight, our customer was able to dispatch a maintenance crew to the turbine and send a technician up the tower to repair the issue. Ultimately, it was an easy fix once our customer was alerted to the problem.

In terms of outcomes, here’s what Uptake enabled:

  • Our customer’s turbine was able to generate a lot more energy than it was previously producing. How much more? Specifically, the additional amount of energy that this one turbine can now produce in a year is enough to provide one hour’s worth of energy to 1 million homes.
  • Here’s another way to measure the impact of our catch: Generating that much more renewable energy means 700 metric tons of carbon dioxide emissions will be prevented from entering the atmosphere this year. That’s something everyone can feel good about.

At Uptake, we frequently hear from our customers that they count on us to be the honest broker. As a company, we don’t have any financial incentive to sell parts or labor. We share one common goal with our customers: To help their assets work as efficiently as possible.

Uptake’s unbiased approach to industrial data science helps us extract every useful insight out of the data, whether it’s something that’s known to the OEM or not. When data is used to its full potential, it empowers businesses to know the truth about their critical assets and to pinpoint issues that need to be addressed before they turn into bigger problems down the road.

Ask Me About My Failures

Watch this video to hear Matt explain the catch in his own words:

Want to meet more of Uptake’s industrial data scientists? Check out our past episodes to see how they’re using their skills to:

Keep manufacturing lines up and running.

Make sure trucking fleets stay on the road and deliver on time.

Help locomotives run better for more reliable rail transport.

Keep wind turbines turning to produce more renewable energy.

New to Our Model Behavior Series?

Here are some helpful definitions to know:

A machine learning model is a codified structure and digital representation of an industrial asset that physically exists in the real world. It organizes the data generated by the asset and standardizes relationships within that data — and, importantly, how all of that data relates to the physical properties of the actual asset. A model targets a specific component of a specific asset, growing increasingly accurate over time with each new data point.

At Uptake, our machine learning models are the output of our AI engines, which are tools that automate the development and deployment of our models. Our engines use valuable insights gained from different industrial use cases to come up with approaches that solve common sets of problems.

A failure happens when an asset or one of its components shuts down unexpectedly in its operating environment, causing problems like unplanned downtime, lost production, service disruptions, unsafe working conditions and more. Failures are what models try to predict and prevent from happening.

A catch is when one of our models — running in a live production environment for our customers — generates an alert that correctly warns of a failure or performance issue that’s bound to happen. Or, in other words: A catch means our model did what it set out to do.

Think you have what it takes to become an industrial data scientist at Uptake? Visit our Careers Page to view our open positions.

Check out the sixth installment of Model Behavior here.