Another way to phrase this question is: does your business need real-time data – and therefore, edge computing – to improve key processes or overall operations? The answer is not necessarily. Data that is not real time – but that is processed with cognitive automation (CA) and analytics software – can still provide many actionable insights that can improve processes and throughput by identifying bottlenecks impacting yield, increasing overall equipment effectiveness (OEE), and determining the root cause of complex plant problems. Once best practices are modeled from a set of data, you may also be able to apply these to other machines, shifts, or plants.
That said, some parts of a manufacturing process can only be improved through access to real-time data and analysis. To make a change in a process as it is happening, you need to process the data as it is streamed – enabling real-time, automated feedback loops in the process – and this requires edge computing.
The first step a business must take is to strategically determine what processes, if made more efficient, would contribute the greatest ROI to the bottom line. Those goals vary by company, but they include items such as increased product quality or consistency, reduced energy consumption, greater OEE, lower transportation costs, and others.
Best practices in the implementation of CA and analytics software is to start small and focus on specific operational data sets and results. For example, look at a specific challenge such as a well-known quality, throughput, or process issue such as machine downtime in operations. Solving a specific problem offers multiple advantages: clearly defined data sources, easily measured success, and easily calculable ROI.
Use cases for real-time data and edge computing
Some manufacturers need real-time monitoring and diagnostics to optimize a machine’s performance or specific processes and therefore need machine learning (ML) and advanced analytics software on-premises to enable data processing in real time. As mentioned earlier, edge computing enables data to be processed as it is streamed, unlike sending data to the cloud, and it is less expensive for processing very large amounts of data than sending it to the cloud.
For example, edge computing is necessary in order to analyze sensor data in real time to identify values that fall outside of a previously defined threshold and automatically stop the production of defective parts. It is also necessary if you want to use data from sensors to monitor the condition of machinery and then speed up or slow down operations to optimize usage; or for motion, temperature, and climate sensors to adjust lighting, cooling, and other environmental controls to make the most efficient use of power in a plant.
To learn more about cognitive automation, analytics, machine learning, and advanced analytics applications and the value they bring to operations, read AMT’s white paper here.
