One machine, five systems: joining up the data you already have
Your asset register, warehouse system, maintenance system, historian and sensors each know something about the same machine. Here is how to join them up without a data warehouse project, and what to check before you trust the result.
In short
- Agree what each kind of thing is known by, and map every system’s fields onto your own words.
- Use whatever each system offers: an API, a read-only database query, an export, or readings posted as they happen.
- Test each connection before it saves anything, and keep the keys out of sight.
- Joined-up records are what make monitoring, numbers and AI worth having.
The same machine, five times over
Take one drive motor on an outbound conveyor. The asset register, a spreadsheet, knows it is a 4 kW geared motor. The warehouse system knows which line it drives. The maintenance system knows it was lubricated last month. The historian knows how many hours it has run since. A vibration sensor on its bearing reports every few minutes.
Nobody has the whole picture, so the question that matters, whether this motor needs attention, takes five screens and someone who knows where to look.
| Equipment | Geared motor, 4 kW | Asset register · file |
|---|---|---|
| Line | Outbound conveyor 3 | Warehouse · API |
| Vibration | 3.4 mm/s · 2 min ago | Sensors · readings |
| Last service | Lubrication · 12 Sep | Maintenance · API |
| Since service | 2,140 running hours | PLC historian · database |
| Next service | Overdue by 140 hours | Calculated |
Step 1: agree what everything is known by
Before connecting anything, decide what each kind of thing is known by: the asset tag for machines, the work order number for jobs, the device ID for sensors. Where two systems use different identifiers, choose the one printed on the equipment itself and map the others to it.
Evalu8 Pro merges records on that key. Syncing again updates a record rather than duplicating it, and a blank value from one system never erases what another one filled in.
Step 2: map their words onto yours
Every system names things its own way: wo_number, equipment.tag, raised, state. A connection rule says which of its fields means what in your business, and how often to check.
APImaintenanceevery 60 minutes
wo_numberwork_order.numberequipment.tagwork_order.assetraisedwork_order.opened_atstatework_order.status
4 pages386 work orders12 placeholdersSynced 4 minutes ago
When a record mentions something that isn’t on file yet, such as a work order for an asset the register hasn’t sent, Evalu8 Pro keeps a placeholder and fills it in when that asset’s own record arrives. Nothing is dropped for arriving in the wrong order.
Step 3: use whatever each system offers
- An API
- Evalu8 Pro fetches from JSON APIs page by page: on a schedule you set (as often as every five minutes), at the press of a button, or when another system asks it to.
- A database
- For a historian or production database on PostgreSQL, a query runs in a read-only session, whatever the query says.
- Details to look up
- For records you already have, Evalu8 Pro can call a system once per record to fill in more fields, such as a device’s model and location.
- Only an export
- Upload the spreadsheet. AI maps its columns onto your model, you check the preview, and next month’s file with the same columns goes straight through. PDFs and photos are read straight into records.
- Devices and gateways
- Sensors, drives and gateways post readings to one documented endpoint, up to 5,000 at a time. Sending a reading again overwrites it, so retries are safe, and readings kept through an outage slot into place. A new sensor shows up by itself. Equipment that speaks an industrial protocol comes in through a gateway.
Step 4: test before you trust it
Run a connection once and look at exactly what it would bring in, before anything is saved. Then check:
- Do the keys line up? The same motor from two systems should be one record, not two.
- Are the units and dates right? Times in milliseconds, or dates without a time zone, are the commonest surprise.
- What happens to blanks? They should leave what is already stored alone.
- Who can see the keys? In Evalu8 Pro they are set by the account owner, encrypted, kept out of the model, and never shown again or included in an error message.
- Can it change anything? Connections only bring data in, and database connections can only read. The one way Evalu8 Pro sends anything back is a control a person presses.
What you get for it
Once the systems agree, the record assembles itself: the motor’s make from the register, its line from the warehouse system, its last service from maintenance, its running hours from the historian and its vibration from the sensor, with the hours to its next service worked out from them.
Then everything else has something to work on. Monitoring watches the motor, the dashboard counts its line’s stoppages, and when someone asks Ask which drives need attention, it answers from the joined-up record rather than five partial ones, and names the records it used.
If you make the equipment
OEMs and system integrators can write the connection rules for their product’s API or device cloud once. Each customer adds them to their own model and connects with their own keys, and the equipment’s data arrives joined up with everything else they run. See how connections work.