This is the third in a series of blogs on the “AI is here.” podcasts. Each blog in the series will highlight the insights from a specific industry leader as they describe how their organization is deriving significant value from AI today.
In this edition of the “AI is here.” podcast , Dan Faggella, Founder and CEO, of market research and publishing company Emerj, speaks with Peter Tu, Scientist at GE Global Research. Peter and Dan talk about how GE is using AI in manufacturing for aerospace and power systems.
In the podcast Tu shares how AI is providing a significant benefit to the manufacturing of components, such as the blades in a jet engine. According to Tu one of the biggest advantages that AI has provided has been the ability to use statistical inference.
Tu states that there are many use cases for AI in industry. He outlines areas where AI can help control processes, predict what went wrong when there is a failure, and prepare for future outcomes, such as computer vision, sensor data, and more.
One area he speaks about is inspecting components. He identifies different factors that can be analyzed over the life of a component, including:
- The shape of the object
- The size of the object
- What the tolerances are for defects
- Meeting specification requirements
All of this involves incredible complexity. For example, when he looks at a factory there is a large number of complex processes going on, each reported on by an even larger number of sensors. Managing how that volume of structured and unstructured data is received, processed, and acted upon is an extremely challenging process.
Beyond just performing inspections during the manufacturing process, once the component has been used, it may need to be continually inspected to see if it is still within specifications. Damage may have occurred, but does it need to be repaired/replaced? What is the probability it will fail in a given time frame? Proactively answering these questions is critical to maintaining operation of a given component.
The components also need to be inspected as part of a larger system, for example a jet turbine blade. This component needs to be inspected at every step, and then within the context of how it has been used. An engine that has been heavily used in a sandy, desert environment will not have the same wear characteristics as one used in a cold, wet environment. Ultimately, this enables organizations to make better predictions about predictive maintenance.
According to Tu, where these types of inspections once required human involvement, advances due to algorithmic improvements have made these types of analysis using AI far more effective. This is particularly true for deep learning, which is programmed with data and does a better job taking us from data to outcomes than human programmed models.
AI is here. Discover how you can fundamentally transform what is possible for your organization with the power of AI in weeks, not years, with the industry’s most powerful, full stack deep learning platform while reducing the need for significant investment in infrastructure, personnel, and other resources.