Golden batch workbench allowing to swop analyses of the different variables within the asset tree

[MArgarida]

The idea is we define a golden batch profile for one or more assets, defined by asset trees.

Golden batch would be a reference profile+/-SD, for the batches performing as expected

Through a button on top we would be capable of selecting other variables from that asset, running in the range defined, with the same calculations. (see example attached)

At the moment, this has to be set up manually and perhaps in different workbenches to run properly. it would be a great troubleshooting possibility

[Synjen Marrocco]

Hi MArgarida,

Thank you for the post! There is currently a way to do this leveraging Seeq's Asset Groups. We can treat the variables that you would like to apply your analysis to as the assets themselves. 

The steps are as follows: 

Begin by adding all the variables that you would like to compare/monitor to a new asset group. You can name the signal column with an alias 'Signal' and then name each asset(row) by the variable itself. Link each new assets to the corresponding signal.

• Create the Asset Group:

 

• Add an asset for each variable and then a single column that will be called the 'Signal' column. Click the '+' sign to link each asset to the corresponding raw signal (in this case it would be the level, pressure, temp, and vacuum signals from a historian).

 

Next, add columns for the reference profile upper limit, lower limit, and average if desired. 

• Add each of the items by clicking Add Column > Add Calculated Item > Existing Seeq Item. Locate the reference profiles (the ones you previously created for your view linked to your post) to add to the asset group. Here you will swap the referenced variable in the formula to the signal variable in the asset group. 
  • Note: If creating a new reference profile, I recommend first creating it in the workbench with the reference profile tool, then coming back to the asset group to add the item as done above. Again, swapping in the signal from the asset group for the original signal. 

 

• The final tree should look something like the following. Of course you may have more or less assets/variables.

A couple notes:

1. You can also add column to the asset group for conditions and link the reference profiles to those if you happen to have different ones for each signal. 
2. This workflow can be leveraged in creating and scaling deviation monitoring by creating a condition that monitors the signal against the upper and lower limits. You can use Treemaps to scale and monitor this deviation condition across all your variables. The same is true for tabulating the deviations in a table.

Now, go and create a display that overlays your batches/capsules with the signal, upper, lower, and average profiles from within the asset group you created. Leverage the asset swapping capabilities to swap between your variables with one click. 

 

Edited November 4, 2022 by Synjen Marrocco

[MArgarida]

Thank you, i'm glad there is a way. i was, however, expecting it could be easier. in this case we talk only about 4 assets and maybe 10 variables (to get started), but we have sites with more than 10 parallel assets and many more variables. i will anyway try to set this up as you propose, see what it gives.
 

Just to be clear, this "golden profiles" asset, what exactly means all variables?

- does this mean i no longer use the asset tree structure? 

-or do i create one golden profile per asset?

thank you

[Synjen Marrocco]

The method I mentioned is creating a new asset tree using the asset groups function. In this case, the variables themselves (temp, pressure, etc.) are the assets in the asset group tool. Asset groups only allow one layer, so if you wanted to apply this to multiple reactors for example, you would need to create an asset group for each reactor. The major benefit of the asset group tool is it will scale the calculation across all the assets (your variables in this case).

If you are looking to apply this more holistically across many pieces of equipment, each with many signals you would like to monitor, I would recommend using the Spy Assets functionality. Spy Assets allows you to create complex asset structures with built in Seeq calculations. 

https://support.seeq.com/space/KB/1590165555/Asset+Groups

https://python-docs.seeq.com/user_guide/spy.assets/index.html

[MArgarida]

Thanks, it works somehow. it's a matter of defining asset groups in a different way, and creating different asset groups and different workbenches. It's not a solution that would be easily implemented and used in 50 sites 😞 while it has a massive potential, the heat maps are nice, but not so insightful

It would still be neater if it could be consolidated in one workbench, based on one PI tag file, much better scalable and used at local level, it seems so close to get it and i would still like to suggest to consider developing it

[MArgarida]

actually, i found that if we just define the variables as assets in a "group asset", and the different assets as signals it is possible to swap variables, so case solved 🙂