3 key uses cases driving edge computing for manufacturing
Innovative uses of AI, IoT, and edge computing are transforming the manufacturing industry. This article explores some uses cases that are driving demand for smart manufacturing, and the challenges that are preventing widespread adoption.
What is smart manufacturing?
Smart manufacturing refers to the integration of advanced technologies and data-driven processes to optimize production and improve efficiency in traditional manufacturing. It leverages solutions like AI, the Internet of Things (IoT), robotics, and automation to create more agile processes. By using real-time data, manufacturers can make informed decisions and automate processes to improve operational efficiency.
Given the number of IoT devices present on a manufacturing site, combined with the unique processes for each organisation, the demand for automation and intelligence is generally higher than other verticals. Although examples of completely automated, smart factories are few and far between, manufacturing is widely considered to be fertile ground when it comes to edge computing adoption (though this remains predominantly on-prem today).
Use cases that are driving demand for smart manufacturing
Amongst the many use cases that leverage edge, there are some which stick out as significant opportunities for efficiency gain.
1. Worker safety: Video ingest and analytics
2. Workflow and asset tracking
3. Advanced predictive maintenance
Each of these tackles specific challenges faced by manufacturers, offering practical solutions that leverage high-definition cameras, real-time data processing, 5G connectivity, AI, and edge computing.
Worker safety: Video ingest and analytics
Worker safety is a top priority in manufacturing. To help mitigate the risks created by hazardous environments and fast-moving machinery, video ingest and analytics has emerged as a critical solution which can save costs for the enterprise. By using high-definition cameras and real-time image processing, businesses can identify dangerous situations and non-compliance with safety protocols.
Video ingest and analytics allowing manufacturing to detect potential hazards quickly without having to integrate expensive equipment across the site. Off-the-shelf cameras can be installed quickly and easily, with video footage then captured and analyzed in real time, allowing for immediate alerts or even the halting of machinery for the prevention of injuries and accidents. Certain systems can also identify workers not wearing the correct safety equipment, triggering alerts to ensure adherence to safety protocols.
Foghorn used this use case to minimize virus transmission during the COVID-19 pandemic. They utilized video ingest and analytics to detect:
- Monitor mask compliance
- Body temperature
- PPE usage
- Track handwashing and social distancing.
There are, however, barriers to widespread adoption of this solution. Installation of cameras must remain as light-touch as possible to not disturb the processes of the factory, and overly sensitive alerts can lead to unnecessary downtime, disrupting production processes. Given the horizontal applications of this use case however, solutions are already relatively mature.
Workflow & Asset Tracking
Workflow and asset tracking can leverage 5G and edge computing, navigating the challenges posed when devices are moving in a complex physical environment. Traditional processes often suffer from human error, leading to lost parts and delays, whilst inefficient inventory management without asset visibility hinders overall production efficiency.
By leveraging 5G connectivity, low-power tags on equipment and materials can transmit real-time location data to the cloud, offering manufacturers greater visibility of their assets. Edge computing enhances this by enabling real-time processing and localized decision-making, optimizing network bandwidth and improving security.
While the benefits of workflow and asset tracking are significant, use cases are limited to outdoor scenarios due to the cost of 5G end points. When operating in an indoor environment, the number of 5G end points required to cover an area increases as a result of the high frequency of 5G signals. The resulting costs are high for most scenarios, meaning that many will be wary to invest. As the ecosystem matures we expect these prices to drop. As it does so, the adoption of this solution will likely accelerate.
Advanced predictive maintenance
Advanced predictive maintenance combines sensor technology, AI and edge computing to proactively monitor machinery, detect faults, and prevent breakdowns. Prevention of errors is preferable in a manufacturing environment, where factory downtime of even a few hours can incur costs in the millions of dollars to an organisation. Traditional maintenance practices are costly and labour-intensive, often resulting in unplanned downtime and compromised worker safety. By integrating sensors and AI predictive maintenance platforms, manufacturers can limit unplanned downtime significantly.
Sensors attached to critical machinery, either via retrofitting or with specialised machines, collect real-time data on the condition of equipment. This data can be auditory, visual, temperature, etc. The data is transmitted to an AI predictive maintenance platform, which analyses patterns using machine learning models which are often trained on proprietary data relevant to the specific machine or process. Timely updates and alerts are issued when maintenance is needed.
The primary concern with predictive maintenance is the training of models. Given the variance in machinery used in manufacturing sites, sourcing the data to train models can be time-consuming and costly. This is especially true if a process is highly specialised, requiring the model to be trained especially for each specific deployment.
Manufacturing remains an important vertical for edge computing. The key use cases we have outlined (worker safety measures through video ingest and analytics, efficient workflow and asset tracking, and the progressive use of advanced predictive maintenance) stand out as major opportunities for enhancing efficiency. Nevertheless, the journey towards broad adoption isn’t free of hurdles. Challenges ranging from potential disruptions during technology installations to the substantial costs of adoption, as well as the intricacies of training predictive maintenance models, pose significant barriers. Yet, with ongoing advancements and foreseeable cost reductions, the outlook for smart manufacturing remains optimistic.
Edge computing market overview
Edge computing market overview
This 33-page document will provide you with a summary of our insights from our edge computing research and consulting work:
Is AI accelerating the edge opportunity?
Part one of a four-part series where we examine exactly how the advent of a new wave of Large Language Models (LLMs) will change the edge opportunity. We explore the AI use cases that edge computing could help to facillitate, the use of AI in edge orchestration, and whether there will be new entrants to the edge market.
What is SCADA? Supervisory Control and Data Acquisition
This article defines SCADA (Supervisory control and data acquisition) and the demand for edge computing to support SCADA systems. Across heavy industries, automation and analytics are transforming process management and OT, integrating intelligence into the operations of enterprise. Due to its specific compute demands, SCADA benefits hugely from the computational and security benefits of edge computing.
7 Open source projects shaping edge computing
Edge computing open source projects are an important part of accelerating edge adoption because of their role in standardising edge computing.