Utilizing Industrial Edge Computing for Smart Manufacturing

Inexpensive sensors with high-speed communication capabilities—described as Internet of Things (IoT) devices—have led to an explosion in data acquisition. However, data is of little value unless analyzed to yield useful actions. Many IoT sensors and devices send their data to the cloud for processing. Edge computing is an alternative to cloud-based processing that can benefit manufacturing in several ways.

Industrial edge computing is especially valuable in situations where manufacturers need fast and reliable insights from their equipment data. Rather than waiting for every sensor reading to travel to a centralized cloud platform, edge-enabled systems analyze critical machine data closer to the asset. This helps teams detect abnormal conditions, receive alerts and respond before small equipment issues become disruptions.  

What is edge computing?

Picture a network of sensors and computers, all sending data to and receiving instructions from a central server. In simple terms, this is how the Internet is organized. At the “edge” of this network are the IoT devices that interact with physical hardware. Pumps, motors, rolling mills, robots and packaging machines are just some examples of the machines that might be connected in an Industrial Internet of Things (IIoT). Every one of these machines could stream data back to a central server. This might include production statistics such as cycle time or operating conditions such as temperature and vibration. Inspection systems might be transmitting images for analysis and storage.

Such data can help operators monitor performance and identify improvement opportunities, but there are two limitations: latency and bandwidth. Latency is the delay that occurs when sending and receiving data. It can vary a lot, leaving cloud computing systems unable to perform real-time control. Bandwidth is the capacity of the “pipe” from device to server. As more industrial IoT devices are added to a system and sampling rates go up, the volume of data being transmitted grows. At some point the “pipe” becomes full and any additional data is delayed.

Industrial edge computing means manufacturing data from these devices are processed and analyzed closer to the machine, device or production line where it is generated. This avoids the latency and bandwidth concerns that come with cloud-only processing, resulting in faster analysis and responses. Here’s how edge computing compares to the alternatives: 

How edge computing benefits manufacturing

Use cases for edge computing in manufacturing include:

• Gauging and inspection

• Production planning and control

• Process control and optimization

• Predictive maintenance

In gauging and inspection applications, industrial edge computing allows for analysis of data and immediate response to any issues observed. For example, a thickness measurement system might detect a drift and implement corrective action in a rolling mill or coating process. Image-based defect detection systems increasingly use artificial intelligence (AI) to find mistakes or excessive variability. This previously required the computing resources of a server but a growing number of edge computing IIoT devices can perform this at the machine.

Edge computing saves machines from reporting every aspect of production data to the server. Instead, the IIoT device might only report trends when limits are exceeded. Some manufacturing processes have large numbers of variables that make it a challenge to maintain product consistency. IIoT edge devices can monitor parameters such as viscosity, flow rates and temperature and use AI to optimize inputs in pursuit of an output goal. Industrial edge computing is especially useful for bringing analysis and response times as close to real time as they have ever been, thanks to less communication with a centralized server and in turn, a reduction in latency.

Smart factory edge computing addresses one of the most pressing needs of the connected facility, cybersecurity. With more localized processing, the facility can maintain more control and oversight over how data is handled. With local processing and onboard computers in newer equipment also comes a lower cost for centralized servers and related infrastructure, which can not only yield direct cost benefits but can also reduce energy and maintenance costs. The largest and most mature use of edge computing in manufacturing may be in predictive maintenance. This addresses one of the biggest challenges in manufacturing: how to eliminate unplanned downtime.

Industrial edge computing use cases

Here are some hypothetical edge applications that can deliver the most value for manufacturers:

Quality inspection and defect detection 

• Edge AI can analyze images or sensor data near the inspection point. 

• Useful for identifying defects faster without waiting for cloud processing. 

• Can reduce scrap, rework and delayed detection. 

Process control and optimization 

• Edge systems can monitor variables such as temperature, pressure, flow, viscosity, and speed. 

• Useful when immediate process adjustments are required. 

Predictive maintenance 

• Edge-enabled sensors can analyze vibration, temperature, current draw, and operating conditions. 

• Supports faster anomaly detection and maintenance alerts. 

Robotics and industrial automation 

• Edge computing can support local decision-making for robots, automated cells and vision-guided systems. 

• Useful where response time matters. 

Production monitoring 

• Edge systems can track cycle time, throughput, downtime, micro stops, and machine utilization. 

• Helps operators and supervisors identify performance changes quickly. 

Safety and environmental monitoring 

• Edge systems can support local alerts for unsafe conditions, overheating, pressure changes, leaks, or environmental threshold events. 

Edge computing and predictive maintenance

Many manufacturers combine reactive and preventive maintenance to maximize equipment availability while simultaneously attempting to minimize costs.

Reactive maintenance—running a machine until it fails and then making repairs—is a valid strategy in situations where downtime costs (lost output, quality problems, delayed deliveries, overtime working) are not excessive. However, when downtime drastically affects business performance, preventive maintenance is the norm.

Preventive maintenance involves taking a machine out of service on a scheduled basis and replacing components and adjusting in order to keep it running within specifications. The risk in this strategy is doing more maintenance than is needed to prevent breakdowns, and yet not doing the right maintenance.

Predictive maintenance offers an alternative. By using IIoT sensors with edge computing capabilities to monitor vital signs on manufacturing equipment—the mechanical equivalent of checking heart rate and blood pressure—it’s possible to detect and react to wear or faults before they affect production. Examples of the “vital signs” that can be monitored include:

• Temperature

• Vibration

• Current draw

• Fluid levels

• Flow rates

• Noise

• Speeds and cycle times

This type of monitoring can be applied in both discrete part and process manufacturing. Turning and milling centers, robots, and integrated assembly and packaging lines are all good candidates for predictive maintenance, as are mixing, reacting, processing, curing and coating processes. Wherever predictive maintenance is applied, the benefits can include:

• Improved equipment availability

• Protection against unplanned breakdowns and micro stoppages

• Reduced process and product variability

• Better safety

Industrial edge computing is a prime example of the potential of Industry 4.0 and the smart factory. With edge computing, Industry 4.0 foundations such as connectivity and responsiveness become all the more effective, with more bandwidth available for the most critical, data-intensive processes and more agility at the point of data collection to make independent, autonomous decisions. Edge computing analytics are one of the leading ways that the smart factory is becoming even smarter, approaching true real-time analysis and response.

Need faster insight into equipment health? Talk to ATS about predictive maintenance and machine health monitoring. 

Edge computing architecture in manufacturing

What follows is a typical architecture for edge computing in an industrial environment:  

1. Sensors collect data from machines and processes. 

2. Edge computing devices process data near the machine or production line. 

3. Local systems respond to urgent conditions or send alerts. 

4. Filtered data moves to cloud or enterprise systems for dashboards, reporting and analysis. 

5. Maintenance systems receive insights through alerts, dashboards, CMMS/EAM workflows or technician review. 

6. Teams act on the information through inspections, corrective work, PM updates, or reliability improvements. 

Ask ATS about comprehensive maintenance solutions

As a leading provider of technology-based industrial maintenance, ATS understands the role that IIoT plays in effective predictive maintenance. The results are clearly measurable and include reduced unplanned downtime and higher productivity. Contact us today to find out how we could help improve the effectiveness of your maintenance operations.

Next Post

The Digital Factory: What it is and Why it Matters