A wastewater AI pilot needs more than a collection of sensor values. Operators must be able to identify the source, time, quality and operating context of each record. Web-SCADA provides a practical starting point for organizing that evidence.
Map the sources already available
List PLCs, sensors, existing SCADA, laboratory results and shift records. For each source, identify the owner, access method, units and available history. Confirm compatibility with the actual equipment before selecting an integration approach.
Show data quality alongside the process
A dashboard should distinguish current readings, the last received value and missing or invalid data. A displayed number does not prove the sensor is still communicating. Include update times and a clear route from each signal to its equipment.
Separate monitoring from control
Start with an agreed monitoring scope when appropriate. Remote actions require their own permissions, plant operating conditions, command feedback and action logs. The local operating procedure remains part of the acceptance criteria.
Choose one AI question that can be checked
Define a specific event to forecast and how the plant will verify it. Review operating modes and confirmed incident records. Run an observation pilot and measure false alerts, missed events and useful warning time before expanding.
Records to check
| Record group | Evidence to review |
|---|---|
| Signal list | Source, equipment and units |
| History | Timestamps and known gaps |
| Events | Confirmed incidents and operating changes |
Prepare an assessment with X-IRIS
Bring a process diagram, source inventory, representative data and one operational problem to investigate. Review the implementation roadmap and request an assessment to agree a scope that can be tested.
