How AI Enables 20X Faster CT Scanning
Can CT scanning fill a role in production validation similar to the CMM?
Even though people in manufacturing are generally aware it is an industrial measurement tool, most associate computed tomography (CT) with medical applications rather than manufacturing.
But imagine CT scanning as a daily industrial metrology tool validating ongoing production. Such a tool would do more for measurement than a coordinate measuring machine (CMM) can do. CT scanning can deliver fine-detail measurement of the exterior geometry of the part, like a CMM, but then also adds fine measurement of any detail hidden inside.
Speed has been the limitation on this scenario. CT scanning moves too slowly to keep up with production. Except: Researchers at Oak Ridge National Laboratory’s Manufacturing Demonstration Facility have succeeded at making CT scanning faster.
As you might have expected, AI is the key. Part of why CT scanning is slow is the amount of sampling needed. Measuring many different angular sections provides enough data to overcome inherent noise. But in place of all this redundant measurement, deep learning via the Oak Ridge technology, called Simurgh, can now deduce accurate measurements past the noise with far less data.
Parameter development in additive manufacturing was the first beneficiary. Part of what makes AM parameter sets effective is their ability to print fully dense metal without internal pores. Using Simurgh, an array of laser powder bed fusion test specimens in a new metal, all made with different parameters, can be CT scanned rapidly to now determine in a relatively short lead time which parameter set is best.
But scale production, whether additive parts, machined parts, or castings, is the application where fast CT might deliver even greater value.
In the latest episode of Manufacturing Unlimited, I talk with ORNL MDF researcher Obaid Rahman about the promise of fast CT and whether we might see CT scanning begin to fill in for CMMs.
Watch the episode:
Manufacturing Unlimited is produced by ASTM International and sponsored by the International Conference on Advanced Manufacturing (ICAM), where Rahman will present on this technology. Learn more at the button below.