NIX Solutions: AI in defective hard drives detection

NVIDIA, Hewlett Packard Enterprise, and Seagate talked about the Edge RX platform, which quickly detects malfunctions in off-line hard drives with the help of artificial intelligence, says 3dnews. Thanks to new technology, the manufacturer wants to accelerate the release of products, improve their quality and increase revenue.

Seagate Edge RX technology was developed in conjunction with NVIDIA and Hewlett Packard Enterprise. Most often, in the manufacture of hard drives, arise problems with magnetic heads, the production of each requires 1400 steps. The manufacturer needed a technology that analyzed 17 million images of magnetic heads daily to identify defects quickly. There was no such technology on the market, so it had to be created jointly.

Since there are a lot of malfunctions of the heads of hard drives, ordinary artificial intelligence could not be used simply, working according to strictly defined rules. The manufacturer needed a deep machine learning technology that could quickly adapt to completely new defects in order to report on their occurrence in time.

Hewlett Packard Enterprise’s Edgeline platform, based on the NVIDIA T4 GPUs, was used to collect data, says NIX Solutions. The Apollo platform with NVIDIA V100 Tensor Core GPUs was chosen for deep AI training. Ultimately, it took a year to create the technology for capturing and analyzing images of disk heads.

Seagate representatives believe that thanks to the Edge RX platform they will be able to increase the speed of production of hard drives by as much as 10%. And by improving product quality, profitability of production can increase up to three times. According to Raghavan Srinivasan, Seagate’s senior global marketing director, mistakes cost them too much. They could find out about the presence of defects only after the completion of production processes.

At the moment, Seagate wants to deploy technology at all of its production sites. It also wants to find out how the Seagate Edge RX is reflected in other manufacturing processes. Along the way, the company shared its experience with other manufacturers so that they also had the opportunity to develop such a technology.