Artificial intelligence (AI) is all around us. Soon, AI will make decisions and take actions for you. The availability of production data gives the intelligence necessary to make it happen, but you often need to "work that data" to understand it.
Fortunately, it’s becoming easier to make it useful and actionable. Instead of manually mapping the data throughout the plant, software—such as Rockwell Automation's FactoryTalk Analytics for Applications—is available to automatically discover devices and locations, such as lines and even entire manufacturing plants.
The raw data found is then turned into storyboards that display usable information for either a person or a controller. Some of the first steps on the path to AI can be found in The Connected Enterprise, a Rockwell Automation IIoT vision in which smart assets and technology enable making decisions closer to the point of information. These subjects will be discussed in the many sessions at the sold-out Rockwell Automation TechED in San Diego in early June.
System- and human-input components of AI analytics can help to determine how much the system does automatically vs. how much user input is needed to make decisions, and then the software can take actions to impact the process. As is often the case, data is reviewed to determine what happened and why it happened. This is called descriptive and diagnostic analysis, but the user must do much of the analysis.
As more AI analytics are added, the system becomes more predictive and prescriptive, and it does more of the work needed to get to a decision. The AI starts predicting what will happen and prescribes what should be done with much less input from the user. At this point the analytics are approaching closed-loop control.
Now is the time to look closely at AI from a device level to an enterprise level. In the past, with a little work, the data showed what machine, system or plant was running the best. It also provided answers on why faults happened and where poor throughput was occurring. By adding AI, such as FactoryTalk Analytics, it is possible to predict that a motor will fail, quality will be out of tolerance or a plant will be behind schedule, soon. As the scale of AI analytics increases, these predictions become actions needed to maintain a motor, improve or correct quality or stay on plan.
Check out FactoryTalk Analytics for Applications and discover how data can provide more actionable information than ever with little work needed to understand device health and machine performance. Presenting the data and understanding it for optimization can be just a few clicks away.
AI is now moving into controllers. Rockwell Automation has a data-driven analytics algorithm in a module that plugs into the controller chassis. The AI in the module can learn the application automatically, or users can choose what they want to model. The automation itself, with the help of add-on instructions, will be learning the system, and, when anomalies occur, alarms can be triggered. This AI engine is just the start of an ever-expanding offering from Rockwell Automation, and Microsoft's vision of AI is helping to lead the way. The connections between FactoryTalk Analytics and AI in devices, machines, systems and beyond will make all parts of The Connected Enterprise smarter.
About the author: Dave Perkon
