Emerson Revolutionizing Performance in Manufacturing with Self-Healing and Scrap-Minimizing Edge AI

[Semih Asil]

### Increasing Challenges in Manufacturing and the Search for Solutions
Globalization, labor shortages, supply chain issues, and sustainability pressures create a complex picture for manufacturers. While improving quality, reducing scrap and emissions, and maintaining productivity have become imperative, the lack of skilled personnel makes achieving these goals even more difficult.

### Digitalization in Industry with Edge AI
Artificial intelligence (AI) and machine learning (ML) technologies offer significant solutions for achieving operational excellence in manufacturing. However, cloud-based AI systems are often unsuitable for industrial environments due to latency, connectivity issues, security constraints, and cost. Real-time production line feedback cannot be provided through cloud connectivity.

### On-Premise AI is No Longer an Option
Real-time industrial AI analyzes data instantly in factories and makes immediate decisions, enabling closed-loop controls. This allows errors to be detected immediately and automatically corrected, almost completely preventing scrap. Especially in image processing applications, where dozens of images need to be processed per second, milliseconds are vital.

### Connectivity and Security Issues
Some manufacturing facilities, located in remote areas, experience problems with cloud connectivity and do not want to send data to the cloud due to data security and intellectual property concerns. The cost of cloud services can also create a significant burden for image and video transfer, which requires high bandwidth.

Since all processing is done on-site, data does not leave the facility, latency is reduced, cost decreases, and data security is ensured.

### Specialized Edge AI Processors and Industrial PCs
Instead of the cloud, industrial computers (IPCs) equipped with specialized AI accelerators perform AI operations in the factory with real-time and low latency. These processors are specifically optimized for high-speed image processing and large language model inference. They consume less energy than traditional GPU solutions and can operate 24/7 without requiring fans.

### Emerson's PACSystems IPC Platform
Emerson's robust, high-performance IPCs are designed for harsh factory conditions. Resistant to challenges such as dust, heat, vibration, and humidity, these systems offer visual analysis, control, and data processing tasks on a single platform, enabling on-premise AI applications.

### Technical Specifications

• On-premise, real-time AI processing with millisecond-level latency
• High energy efficiency with a dedicated AI accelerator
• Fanless cooling system
• Durable and certified IPC chassis
• Multi-core x86 processor support
• Optimized for high-speed image processing in industry

### Example of AI Application in Industry
A thermoplastic pipe manufacturer monitors the pipe wrapping process with an AI-supported visual inspection system. This system, which prevents human errors and labor shortages, instantly detects errors, provides automatic intervention, and minimizes raw material waste. Additionally, the system generates video recordings for quality certification.

### The Future of AI in Industry
These technologies can become widespread in many areas such as laser welding alignment, automotive adhesive applications, medical packaging inspection, oil and gas leak detection, and mining analysis. Edge AI is opening a new era in automatic control and predictive maintenance in production lines.