Best practices for OPC UA logging, Dataset naming, and asset model for one OPC server that includes multiple machines
I have a plant with 3 OPC UA servers. One of the servers has 7 "blendlines" (with 9 tags each) and 5 tags in a "utilities" group. I'm looking to understand what "best practices" that I use here (e.g. recommendations on setup). In the past, I have typically used a separate dataset per machine but that was more when I was using OPC DA and had a logging session/file per machine.
Right now, I have a single OPC UA session for this OPC server with 2 groups where 1 group is called Blendline and has all 63 blendline tags (e.g. BL1.State, BL2.State, etc... so I have separation by blend line in my tag picker under the local historian name) and the other group is called Utilities. With that setup, all 68 tags are in the same DataSet since the dataset is at session settings level.
I then have a virtual view/asset model setup for that dataset for the 7 blendlines.
My question is.... Is this the best setup or should I have 8 different OPC UA sessions where all point to the same OPC UA sever but each covers one machine and points the 7 blendlines and 1 "utilities machine" to its own dataset?
Seems like I'm probably doing it the right way already (1 OPC UA session and dataset for all 8 machines) but I just wanted to try to confirm that.
1 reply
-
Hi James Wise !
What you are doing is good. 68 tags is very small. Splitting those up into their own dataset would be overkill, and since it's all coming from the same server, we recommend writing it to one dataset. Typically, it is best practice to have a 1-to-1 ratio of data source to dataset provided the tag count doesn't explode (like 50,000+). And since you have the structure built into the tag name, building the model will be a breeze.
Print this page