Compare sources without confusing measurement location, aggregation, and processing differences.
Manufacturer, controller, and grid-operator totals can differ even when every system is functioning correctly. First determine whether the sources measure the same physical point and period.
| Cause | Example |
|---|---|
| Measurement location | Inverter production is measured before cable and transformer losses; the grid meter is measured at the connection point. |
| Different boundaries | One source uses local midnight while another aggregates in UTC. |
| Sampling and interpolation | A portal estimates short gaps while the raw source leaves them empty. |
| Counter resets or rollover | Energy difference calculations cross a device reset. |
| Late or missing files | A daily upload arrives after a report was generated. |
| Unit or scaling error | Watts interpreted as kilowatts or a meter multiplier omitted. |
| Asset mapping | A file is attached to the wrong inverter or installation. |
Do not assume a fixed acceptable percentage
The expected difference depends on measurement class, topology, cable losses, transformers, aggregation, and source processing. Investigate the physical and data boundaries before applying a generic tolerance.