Companies Tap IoT to Drive Quality and Save Money

Manufacturers such as Standard Textile leverage IoT tools to pinpoint the location of production slowdowns and process problems — and better understand why they happen in the first place — ultimately saving time while boosting the overall quality of their products.

“The Internet of Things allows us to know what we need to do faster,” Heiman says. “It gives us insight into things that are happening throughout the process that we never had before.”

Heiman and his manufacturing team leaders can view the current state of machines at all times via a dashboard that uses applications written using Microsoft programming language. 

For example, sensors now measure key elements such as the heat, temperature and composition of the chemical bath in which Standard Textile pretreats yarn, ensuring it will withstand the weaving process. Sensors also collect data on power consumption that is used to estimate current and future productivity of machines. 

“You can see what’s happening in each machine in the plant, from weaving into finishing and final fabrication, cutting and packaging,” Heiman says. “If there’s a problem, it will not only show you that the machine is down, it will also show supervisors and technicians exactly what the problem is.”

The company also sees gains in its efforts to improve sustainability. For instance, engineers now have a better idea of how much heated water is left over at the end of a production cycle, which can be used at the beginning of the next cycle to reduce water and energy use throughout the entire production process.

Connected sensors also help companies maintain and improve quality by ­identifying elements of the production process that may be problematic. Flex, which designs and builds custom ­products for other companies, deploys connected ­sensors and artificial intelligence for automated inspection of its production processes, such as looking for scratches in the casing of a product.

Robotic devices the company created to perform manufacturing tasks also receive information via sensors to adjust production elements as needed.

“We have a standard assembly of DIMMs,” says Mike Doiron, Flex’s CTO for global operations, referring to dual inline memory modules. “At times, both the DIMMs and sockets they’re inserted into can come from different suppliers, which can mean the force required to insert a DIMM into one socket can deviate from another. By modeling and ­profiling these characteristics, we have an intelligent, reactive process, rather than just using a fixed amount of force.”

Production equipment delivers sensor data to PCs within Flex’s ­facilities that act as edge devices, transmitting the information to the cloud, where it’s processed using algorithms.

“That’s a really big benefit of IoT and connectivity to the cloud,” Doiron says. “You’re starting to collect data across ­production, so you’re not requiring an engineer to go analyze each ­individual machine.”

IoT Has a Favorable Future in Manufacturing

Deploying an onsite IoT solution doesn’t require an inordinate number of components — typically, just a gateway device, sensors and wired or wireless networking capabilities. That’s why businesses will continue to integrate IoT creatively into manufacturing, maintenance and other processes, O’Donnell says.

Investment in IoT solutions will reach an estimated $1.2 trillion by 2022, according to IDC.

Still, companies should decide early in their IoT journey what data they want connected sensors to gather, and how it will be used, says Nick Barendt, executive director of the Institute for Smart, Secure and Connected Systems at Case Western Reserve University, and co-director of the IoT Collaborative, a joint initiative of Case Western Reserve and Cleveland State University.

“We’ve seen some companies put sensors on everything, collect all of this data and plan to figure out what it’s used for later,” Barendt says. Instead, he suggests, “really try to be thoughtful about the business problems you want to solve and what data will help solve them.”

“In the last year, we’ve seen companies move away from the generalization of IoT to more clear and definitive use cases,” says Ravin Sanjith, program director of intelligent authentication at Opus Research. “As long as it’s a problem you can state, technologists and engineers can solve it.”

One problem frustrating Standard Textile was machine maintenance, but that’s no longer an issue.

“The predictive side of the system will say, ‘We detect a vibration in a certain part of the machine a human being can’t hear or see; from experience, we know the last time there was a vibration like this, a major motor went out,’” Heiman says. “Our technicians can then fix it, which will take 15 minutes, rather than having a major breakdown three days in the future.”