Model Behavior: Predicting and Preventing Failure on the Front Lines of Heavy Industry

Welcome to the first 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.

There’s a common misconception about new technologies: That humans somehow aren’t part of the equation. Not only is this fundamentally false, but more often than not humans play a monumental role in building and deploying new technologies.

At Uptake, we strongly believe that technology should empower people to be even better at their jobs. But how do those innovations even come to be in the first place?

People — wildly talented, industrious people. And we’re here to set the record straight with Model Behavior.

Solving Industrial-Scale Problems Takes Industrial-Grade Expertise

First, let’s cover some baseline definitions you need to know:

What’s a model? 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 how that information relates to each other — and, importantly, how all of it 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 a tool that automates the development and deployment of our models. Our engines use valuable insights gained from different industrial use cases to come up with approaches to solve common sets of problems.

What exactly is a catch? 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.

What constitutes a failure? 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.

Now Here’s the Catch

Uptake industrial data scientist Kevin Zen knows a thing or two about catches. As a member of our Energy Team, Kevin makes catches for our customers’ wind turbines that, if left untreated, can result in lost revenue.

In this example, Kevin’s model identified a failure in a wind vane on a customer’s turbine. The wind vane was exhibiting anomalous behavior that deviated from normal patterns, causing the turbine as a whole to behave irregularly and inefficiently.

Solving this problem was particularly challenging for a few reasons:

  • On top of it being a rare occurrence with few historical examples to study and model, the issue can’t easily be seen with the naked eye.
  • Data that comes off of a wind turbine’s sensors tends to be inaccurate in a variety of ways. One of which is that its internal compass — which sets its true north and guides its direction — is frequently miscalibrated, changing erroneously over time.

Uptake’s software alerted the customer to the impending failure, who then inspected and confirmed the faulty wind vane. Not only that, the customer caught it operating 45 degrees out of the wind, which had resulted in power loss. For utilities and independent power producers, that unused capacity equals a great deal of money being left on the table.

By making a smaller repair to the wind vane in the near term, the customer was able to avoid any further production loss. The repair also extended the life of the gears on the turbine — no small feat, as the gears are what enable the turbine to rotate, and any irregular movements reduce the turbine’s lifespan.

Ask Me About My Failures

Check out this video to hear Kevin explain the catch in his own words:

Want a more in-depth exploration of this catch? Visit the Uptake Tech Blog to read a post from Kevin himself.

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 second installment of Model Behavior here.