https://www.delltechnologies.com/en-us/blog/ai-set-to-turbocharge...

https://www.delltechnologies.com/en-us/blog/ai-set-to-turbocharge-the-factory-floor/?dgc=SM&cid=10094&lid=spr4733353115&RefID=dtw21_spr4733353115&linkId=116294725

The edge and the cloud

We’re all aware that it has become increasingly difficult to store, manage, process, and secure the ever-increasing tsunami of data. Savvy manufacturers are responding by performing analysis at the edge and only moving the most relevant data to the cloud for further analysis. Apart from benefiting from real time decision-making and insights at the edge, this reduces the cost of transport and storage.

Making faster decisions with AI at the edge

As a result, we’re now entering an era where compute is following the data and where AI has become the main technology for data-driven innovation. Certainly, AI has the potential to deliver tangible benefits, significantly improving all aspects of manufacturing, including quality control by identifying potential issues earlier in the process. For example, applying AI to images and measurements taken during the various stages of a manufacturing process can help identify problems with a product as it is being assembled and before it’s finished. If AI can identify early on that the product is no longer viable, it follows that you get to save many further steps that involve time, material and resources. According to Forbes, AI has alreadyincreased defect detectionin quality control by up to 90%.AI enabling predictive maintenance

AI enabling predictive maintenance

AI at the edge is also transforming predictive maintenance. Picture a motor in a factory line that continuously produces telemetry data as it runs and rotates. This motor creates vibration patterns in three axes and can be measured by computer. In the past, a summary of the data, including averages or outliers, would have been sent to the data center or Cloud for analysis. Today, with edge technologies, the compute resides right at the motor, where an AI inference algorithm can monitor and instantly send alerts, when something out of the ordinary occurs. Having this immediate access to unfiltered raw data is the key to identifying subtle trends that can indicate or predict failure. This would simply not be possible by performing analytics with consolidated or aggregated data.