Page 19 - Logistics News - March_April 2022

AR TIFI CI AL INTELLI GEN CE

AI brings precision and

speed to the production line

With the latest developments in artificial intelligence (AI) technology integrated
with modern cloud platforms, embedding AI into business processes is no longer an

expensive or lengthy process.

utomotive industry leaders are displaying a Creating agility in the sector, the AI
rapid approach to innovation by deploying applications can scale quickly and easily across
Aan AI application in the manufacturing multiple production lines and factories. In
production line. “As an example of the efficiency that addition, OEE is the gold standard for measuring Paul Bouchier, Sales

AI can introduce, let’s consider production line workers manufacturing productivity. For example, an Director at iOCO.
and engineers who manage reject rates in the pulley OEE score of 100 percent means only good parts

assembly process. This used to be a multi-step, manual are being manufactured as fast as possible with no stop
process,” says Paul Bouchier, Sales Director at iOCO, time. In the language of OEE, that means 100 percent
within iOCO Software Distribution, an Infor Gold quality (only good parts), 100 percent performance (as
Partner. “Now, through AI-driven anomaly detection, fast as possible) and 100 percent availability (no stop
the process is radically changed because machine time).
learning actively checks every 10 minutes by processing
millions of records of Internet of Things (IoT) sensor “Measuring OEE and asset utilisation is a

data on the pulley assembly production line for a manufacturing best practice to gain important insights
potential increase in rejection rate. By suggesting into systematically improving the manufacturing process.
the root cause for failure, workers can quickly resolve Now AI delivers integrated business intelligence and
the issue in the production line. In practice, with AI, reporting dashboards to track rejection rates, OEE and
automotive leaders are recording the lowest levels of additional key performance indicators (KPIs) in real time.

rejection rates than ever before.” This reduces manual load, simplifying and automating the
reporting and self-service process,” concludes Bouchier. •
The precision and speed of the AI models are based

on two years of production line and machine sensor
data brought into the Infor Data Lake and used to train
the machine learning (ML) model to observe when the
pulley tightening process falls outside normal behaviour,
increasing the rejection rate.

“In practice, we’ve seen 99 percent faster detection
and diagnosis of failure (from one day to 10 minutes),
lower rejection rates and improved overall equipment
effectiveness (OEE) and asset utilisation. As a result,

better products are delivered, and these are passing

quality checks on their first pass. This then leads to a

reduction in scrap and parts rework, leading to even more L O GI S T I CS NEWS L O GI S T I CS NEWS
consistent on-time order fulfilment to customers,” adds

Bouchier.

17
P
R
IL 2022
H/A
M
www .l o g ist i csn e w s .c o .z a M A RC H/A P R IL 2022 17
RC
A