FINDLAY — Robots are helping Whirlpool identify defective products before they leave the assembly line, the latest example of how companies are adopting artificial intelligence.
The Findlay production plant, which today employs about 2,200 people, uses the AI image recognition tool to identify faulty parts by analyzing the shadows of an image, all within a quarter of a second. The information is then reported back to an operator, while broken or improperly seeded units are sent back to a repair bay.
“It allows (the operator) to within a few seconds quickly identify something they wouldn’t normally see, and then they’re able to go correct the issue,” said Mike Surratt, who oversees Whirlpool Findlay’s manufacturing execution systems.
The AI is self-learning, similar to facial recognition used in security, breaking images down into tiny pixels for what Surratt says is a more efficient, consistent and accurate analysis than tradition vision systems.
“We created this from scratch … We said, ‘We know that there’s a technology that can recognize images. Let’s see how we apply that to our manufacturing process,’” Surratt said. “This is something that normally somebody would do a visual inspection of. A human would do it. We were able to replicate that and even improve.”
The tool was developed by employees who had been working on the assembly line, according to Findlay plant leader Ramsey Aljahmi.
Those opportunities are available for other shop floor employees too, Aljahmi said, whether it’s helping develop new AI tools or programming routes for auto-guided vehicles.
While the plant has adopted many new technologies — such as the auto-guided vehicles that can transport parts across the facility, eliminating congestion – Aljahmi has found that sometimes manual intervention makes more sense. That’s why the plant still uses man-powered vehicles for some tasks and auto-guided vehicles for others.
“It’s value-based,” he said. “We look at benefit and cost and what makes the most sense.”
Reach Mackenzi Klemann at 567-242-0456.