Labels
Labels answer the question: where does this data come from?
A factory quickly collects hundreds of tags. Without structure, finding the right data becomes impossible. Labels give every datapoint its place in the factory: which site, which hall, which line, which machine. In dashboards and alerts you then filter on labels: “show all temperatures in Hall 1” or “alert me on everything of Line 3”.
Each datapoint inserted into UpBlue contains:
- A tag name
- One or more labels
- A value
- A timestamp
Each label consists of a key and a value. The key can for example be machine with a value of printing1.
One or more datapoints with the same labels and tag is called a time-serie.
For example, a time-serie contains the following labels:
| Key | Value |
|---|---|
| Enterprise | UpBlue |
| Site | Factory1 |
| Area | Hall1 |
| Cell | Line1 |
| Unit | Step1 |
Tag name is temperature.
Let’s say there are multiple units and each unit has one temperature sensor. In a dashboard you can list all the temperature sensors in Hall 1 by filtering on Hall1 and tag equals temperature.
Working with labels has an advantage over a traditional tag tree: time-series are not restricted by the one dimension of the tree. The example above is hard to do with a tree. Grouping and filtering is much easier with labels.
Which labels should I use?
Labels are free to choose. You can use any key and value that fits your factory.
Our advice: use the ISA-88 structure. When creating a connection, UpBlue suggests the labels Enterprise, Site, Area, Cell and Unit. These are based on the ISA-88 standard and fit almost every factory. The Enterprise and Site labels are prefilled with your company and organisation name. All labels are optional — use the levels that make sense for you and skip the rest.
The ISA-88 physical model
ISA-88 is a standard developed by the International Society of Automation (ISA). Its physical model describes a factory in levels:
| # | Level | Description |
|---|---|---|
| 1 | Enterprise | Your company, the business entity |
| 2 | Site | The factory (site), the production location |
| 3 | Area | A place inside the factory, a hall or a specific area. |
| 4 | (Process) Cell | The production space. A production line. Multipurpose. |
| 5 | Unit | A part of the production line. Singlepurpose. |
| 6 | Equipment Module | Set of components inside the machine. |
| 7 | Control Module | A valve, actuator |
Each level has a one to many relation. One Enterprise can contain one or more Sites, one Site can contain one or more Areas and so on.
Levels 6 and 7 are rarely needed; they can be added as extra labels on a per rule basis.
You are also free to add extra labels when needed, for example machine or sensor_position.
Where labels are set
- On the connection — every datapoint of the connection gets these labels. Set here the location of the machine or data source: Enterprise, Site, Area.
- On the rule — a rule can override or add labels. Use this when one connection delivers data for multiple machines, for example an MQTT broker that carries data of a whole hall.
Labels are not meant for dynamic data, like a batch number or a production order number. Dynamic data belongs in Operations, as metadata on a production run. There you can store the order number, customer and article per run.
Think carefully about the labels you choose! Labels cannot be changed in the database after they are stored. When changing labels in the connection, new data gets the new label from that moment on. Old data keeps the old labels. In effect a new time-serie is created. When changing labels, check if your dashboards keep functioning.