The digital age has changed how companies interact with customers, create product and execute new ideas. What’s powering all these changes is data, and many industrial producers are investing in software to access and understand the increasing quantity and availability of data.
With more data, there is now untapped potential to inform engineers and plant managers about what applications should be deployed, and where, in their facilities. Their big challenge, however, is that with so much data, organizing and making sense of it can take more effort and time.
Efficiently tapping into these benefits first requires users to eliminate irrelevant data and focus on what matters most to their operations and current critical business objectives. There are a few capabilities industrial producers should look for to get more value from data.
Seeing the difference in operations
People often find it easier to work with information when they can graphically interact with and explore data from connected operations. Creating histograms, or xy plots that visualize constraint behaviors can provide even greater context when pursuing answers to questions like: How can I optimize one section of the plant? Or, which section should I optimize?
One opportunity for manufacturers is determining where MPC (Model Predictive Control) could best be applied and how much improvement is achievable. Once deployed, the capability to graphically interact with data can support MPC as it monitors the relationships between quality, quantity and performance. Where operations consistently drift from the ideal conditions previously identified, MPC is used to continuously drive back to that ideal.
For example, the ability to visualize data can help identify where a bottleneck is slowing capacity and yield. At what section of the plant is money being left on the table, because production is lower than it could be? With this knowledge, decision makers at the plant know where to apply MPC or other machine learning technologies and improve performance.
When looking for a tool to help graphically interact with data, it also helps when the technology can merge manual lab data, or other manufacturing databases like a manufacturing execution system, with real-time process data. This allows users to compare and integrate data from different systems.
So, what does that mean in the actual facility?
Making the difference in food and beverage
The ability to graphically interact with and explore data is essential to food and beverage producers, especially because they suffer from the same data quality and relevance issues prevalent across the manufacturing industry.
Whatever problem you’re trying to solve, you will need to explore and filter data to focus on what’s relevant while removing bad data from your sample. If you are studying time needed to clean equipment, for instance, clean-in-place (CIP) data may be relevant, but if you are studying what processing changes extend necessary CIP cleaning time, processing data is relevant.
With a better idea of what challenge to solve, you can then better use technology like MPC. With a focus on throughput and ingredient yields, a significant change on one portion of the line often requires multiple adjustments over time downstream. In these situations, MPC helps coordinate and facilitate adjustments by anticipating when changes will appropriately maintain the final product quality.
For example, interacting with data can help someone in the facility identify the blanchers, driers, fryers and other units on a french fry line as a problem area that would benefit from MPC. By enforcing the line’s constraints, employees can ensure a plant is producing the most fries possible without sacrificing quality or tripping sections of the line.
MPC coordinates sections of the plant, such as unit temperatures, to eliminate disturbance impacts, even as throughput or raw potato quality changes. Other disturbances could include steam pressure or refrigerant temperature, which MPC can sense, and then leverage its model on fry quality to make appropriate corrections.
Where to start with exploring your data
Before exploring and learning from data, users need to filter and clean information. By removing irrelevant outliers and unusual conditions through filters, users can focus on what’s most important to each line and challenge.
Bundling all these capabilities together streamlines data exploration and the process of seeking solutions. And now, Rockwell Automation is releasing the FactoryTalk® Analytics™ Data Explorer solution to do just that.
Data Explorer is targeted to release near the end of March 2019. At that time, we’ll offer a free 60-day trial, so you can confirm the technology provides the expected value for your operations.
If you’d like to learn more or get started with your trial, please learn more here.