GE Global Research Uses Model-Based Design to Quickly Move Ideas from the Laboratory to the Field
Dr. Rajendra Naik, GE Global Research
Over the last decade, the speed to move an idea from the laboratory to the field has been reduced drastically, with no compromise in accuracy, using Model-Based Design.
Published: 22 Dec 2020
[MUSIC PLAYING]
You know, in today's world, you essentially need speed. You need to design, test, and put things in the market as quickly as possible. And tools that companies like MathWorks make help us immensely in this journey.
What's changed over the last 5 to 10 years is the amount of data that we are getting from these assets. And these, then put through AI learning models, so physics and then data, combining together to create hybrid models. So we are actually progressing from a fleet level modeling to individualized, customized asset-level modeling or digital twin.
So one of the examples that we have here is, in a wind farm, you want to know where you exactly locate your wind turbines. I want to make sure that I design my wind farm layout so that I can maximize the annual energy production. If you see the picture here, the black dots actually represent the wind turbines, and what you see, the little blue streaks, are what we call as wake. You want to minimize, on an entire farm level, the effect of all these wakes.
What you essentially have to do is, in real time, sacrifice some of the energy that the front turbine captures, which might seem counter-intuitive. But, on a overall farm level, you actually increase the annual energy production. You need to do a coordinated turbine control across a wind farm, where we use a lot of the tools from companies like MathWorks. A lot of the development has happened in MATLAB. And that, again, is a framework that we want to adopt in our industry.
[MUSIC PLAYING]