Case Study
UST Vision AI solution helped global manufacturer optimize processes and decrease loss by 26%
OUR CLIENT
This Middle-Eastern company is one of the largest manufacturers of rubber-based products, such as tires, conveyor belts, and sports equipment. With nearly 20,000 employees, the company generates almost $3 billion in annual revenue.
THE CHALLENGE
Manufacturer needed to improve operations, boost production, reduce waste, and business disruption
As this manufacturing company dealt with market volatility and a rapidly changing business landscape, it wanted to streamline complex processes to improve key performance metrics. With only 70-80% capacity utilization and 86% first-pass yield, the company needed to boost productivity and quality control while reducing waste and manufacturing costs. The other challenge was to avoid business disruptions as people retired. While the organization’s aging workforce had deep institutional knowledge, it wanted to decrease its dependence on these employees.
THE TRANSFORMATION
Deployed a real-time, non-intrusive, AI-powered vision intelligence solution
To help the company optimize manufacturing processes, UST experts implemented a three-dimensional artificial intelligence (AI) vision intelligence solution combining:
- Advanced visual patterns-from the smart cameras that provide visual context
- Real-time machine learning (ML) data-from a supervisory control and data acquisition (SCADA) application
- Reference data-from the company’s existing SAP systems
The comprehensive solution encompassed an area spanning 80 hectares with an accuracy rating of 99% for visual detection models. It covered over 200 pieces of heavy machinery and approximately 2,000 mobile machines outfitted with at least 2,500 sensors as well as an inventory footprint of over 10,000 units. To keep costs in check, the project team integrated the vision intelligence solution with the company’s existing network of more than 1,000 surveillance cameras. The AI engine used object and advanced texture detection, object and people tracking, and trajectory mapping to predict shop floor and process anomalies. Custom data from the vision intelligence AI engine reduced the need for ML training data in the SCADA application.
THE IMPACT
Improved shop floor safety and optimized processes, boosting productivity 16%
The company achieved:
- 80% increase in material movement and handling compliance
- 35% reduction in unscheduled equipment downtime
- 26% decrease in loss
- 16% productivity improvement