Working with Multiple Frames

One of the more interesting parts of Layered Frame Analytics is that it can break up your frames and loop over them.

For example, you can start with a frame that represents your entire 24 hour day.

Then you can split that day up into the times your machinery was running.

You can then go even further, splitting things into the different products made during that time.

Any layer after you've split your data up into extra frames is run on each individual frame.

This means you can stack your analytics in one way to get KPIs for your individual production runs. But stack them differently, and you can get KPIs per-batches, per-crews, per-day or per-product.

You can even manually enter and upload data to create the information needed to build virtual sensors.

In this case, we're going to start with a frame that represents our day.

Then we use a query layer to fill that frame with a table of data that includes what state the machine was in across the whole day.

Next, we use a timeframes layer to split that frame into running and stopped times.

We can then filter those results to remove all of the times when the machine was stopped.

Now, instead of a single time-frame for the day, we've got many small time-frames for each distinct time that the machine ran during the day. We can use this to perform a number of analytics, such as calculating…

• How many times the machine started,
• What the average duty cycle is,
• The total running time throughout the day etc.

You could then dive deeper - if we added more query layers, they would be run individually on each duty cycle rather than on the entire day, so you can extract figures like the min, max and avg amount of vibration on the machine for each cycle without contaminating your data with the times your machine was stopped.

Find out more in Why Capture?.