An AIoT architecture must preserve the origin and timing of plant data. PLC status, online readings, laboratory results and shift records answer different questions and need a shared structure that retains their differences.
Inventory each source before choosing hardware
Record the device or software, owner, protocol, access permissions and available history. Confirm compatibility on the actual source. Include asset identifiers so records from different stations cannot be confused.
Separate collection, history and display
The collection layer maps signals and records connection quality. Historical storage preserves source records. Dashboards and analytics then use checked data. Agree network boundaries and role access with the responsible plant team.
Retain source and receipt times
For laboratory data, keep sampling and result-availability times. Use sampling time to understand the process and availability time to evaluate information accessible to a prediction. Preserve quality states alongside values.
Test a small approved scope
A read-only pilot can check source mapping, units, freshness and access. Test connection loss and any buffering or replay capability on the selected device. Remote control requires a separate approved operating and acceptance scope.
Records to check
| Record group | Evidence to review |
|---|---|
| PLC and sensors | Address, asset, unit and quality |
| Laboratory | Sample identity, time and method |
| Work records | Shift, equipment and confirmed action |
Prepare an assessment with X-IRIS
Bring process and network diagrams, equipment details, sample data and the owners of each source. Review the implementation roadmap and request an assessment to agree a scope that can be tested.
Technical reference: OPC Foundation's DataValue specification separates the value, status code, source time and server time.
