The Energy Conversion Model (ECM) is the description of a semi-scheduled solar or wind farm that AEMO uses to forecast its output. For a solar farm it tells AEMO's forecasting system how irradiance, temperature and other weather inputs become active power at the connection point, so that the plant's unconstrained intermittent generation can be forecast for dispatch. It is a registration requirement, prepared on AEMO's template, and supported by plant data and evidence.
The ECM rarely gets the attention the GPS does, but an incomplete or inconsistent ECM can hold up registration, and a poor one degrades forecasts and therefore the plant's dispatch for as long as it operates.
What the ECM Describes
The ECM breaks the plant into clusters: groups of equipment that behave identically enough to be modelled as one. For each cluster it records the technology, the conversion chain from module to inverter output, and the losses. AEMO also needs the meteorological measurements the plant will supply in real time, and how they relate to the clusters.
Defining Clusters
A cluster is a set of inverters and the arrays behind them that share the same characteristics. Typical criteria for grouping equipment into one cluster are:
| Criterion | Why it matters |
|---|---|
| Same inverter model and rating | Inverter efficiency and clipping behaviour must be uniform within the cluster |
| Same module technology and model | Temperature coefficients and spectral response differ between technologies |
| Similar DC capacity per inverter, within a tight tolerance of the cluster average | DC to AC ratio drives clipping, which is a major forecasting effect |
| Same mounting: fixed tilt and azimuth, or the same tracking type and algorithm | Orientation and tracking change the irradiance each module receives |
| Limited geographic extent | Cloud effects vary across large sites |
Check the tolerance and extent limits in AEMO's current ECM guideline. Plants built in stages, with a different module or inverter in a later stage, or with mixed DC ratios to fit an irregular site, will need several clusters. Grouping dissimilar equipment to simplify the spreadsheet is a common reason for a submission to be returned.
Data Required Per Cluster
ECM Cluster Data
- Module technology, model, count and rated power
- Tracking type, axis orientation, rotation limits and backtracking behaviour, or fixed tilt and azimuth
- Inverter model, count, AC rating and DC capacity per inverter
- DC to AC ratio (overprovisioning) totals
- Inverter efficiency curve as tabulated values
- Operating temperature range and any temperature derating of the inverter
- Module temperature coefficients and spectral response where requested
- Degradation assumption over time
- Loss factors between module and connection point as the template requires
The inverter efficiency curve is a frequent stumbling block. AEMO wants it as a table of values, not a picture copied from a datasheet. Use the certified efficiency data for the inverter model and the DC voltage range the plant actually operates at.
Meteorological Data
The forecasting system uses real-time measurements from the plant's meteorological stations alongside numerical weather predictions. The ECM records which stations exist, what they measure, where they are, and how they map to clusters. Typical measurements include global horizontal irradiance, plane-of-array irradiance where relevant, ambient and module temperature, and wind speed.
Data must reach AEMO with a short update interval, as near-instantaneous values or short averages. Check AEMO's current requirements for the update rate and averaging, and confirm early that the met station, data logger and SCADA path can deliver it. A met station that logs ten-minute averages for an energy yield study is not the same specification as one that feeds real-time forecasting.
Supporting Evidence
The spreadsheet is supported by documents that let AEMO verify it. A typical evidence pack includes:
- inverter efficiency certification
- module and tracker datasheets
- meteorological station specification and layout
- DC sub-array layouts showing how arrays connect to inverters
- the AEMO SCADA points list for the met data
- an energy yield assessment, as a cross-check on the expected output
- the construction and energisation program, where clusters come online in stages
Common Mistakes
| Mistake | Consequence | Fix |
|---|---|---|
| Dissimilar equipment grouped into one cluster | Submission returned, or poor forecasts | Split clusters on inverter, module, DC ratio, mounting and location |
| Efficiency curve supplied as an image | Data cannot be loaded | Provide tabulated values |
| DC capacity per inverter not consistent with the layout drawings | Questions on clipping and capacity | Reconcile against the DC sub-array layouts |
| Tracker backtracking not described | Wrong irradiance model for early and late hours | Describe the algorithm and limits |
| Met station update rate too slow for real-time use | Data rejected or forecasts degraded | Specify the station and logger for forecasting, not just yield measurement |
| Staged commissioning not reflected | Forecast assumes capacity that is not online | Define clusters to match the energisation stages and update as stages complete |
| ECM inconsistent with the GPS and registration data | Registration queries | Check capacity, inverter counts and ratings against the other documents |
That last point links the ECM to everything else. Capacity, inverter numbers and ratings in the ECM should match the GPS, the registration data and the design drawings. A late design change, such as a different DC ratio or an extra inverter, must flow into the ECM too.
Why ECM Quality Affects Revenue
A semi-scheduled plant is dispatched against AEMO's forecast of its unconstrained output. If the forecast is consistently high, the plant receives dispatch targets it cannot meet; if it is consistently low, available energy is left unused. Both show up as deviations from target that affect how the plant performs in the market and how it is assessed under the frequency performance payments arrangements that replaced causer pays in 2025.
The ECM is the starting point for that forecast. Forecast providers and AEMO's own systems refine their models with operating data over time, but a structurally wrong ECM, with the wrong DC ratio, the wrong tracking behaviour or clusters that mix different equipment, builds a bias in from the first day. Errors in clipping behaviour are particularly costly, because they matter most on clear days at high output, exactly when the plant's energy is worth forecasting accurately.
It is also worth keeping the ECM current. Inverter replacements, a change of tracker control algorithm, module replacements after damage, or a later stage of construction all change the conversion chain. Update the ECM when the plant changes, not only when AEMO asks.
When the ECM Is Prepared
The ECM is usually finalised during the registration phase, once equipment is procured and the design is fixed, and updated if the as-built plant differs. Starting it once procurement decisions are made, rather than at the last moment before registration, gives time to resolve data gaps with OEMs. Our article on commissioning a 5 MW solar farm covers the wider commissioning sequence, and AEMO Generator Performance Standards covers the GPS side of registration.
Get the ECM Done Right
ElectraGlobe prepares Energy Conversion Models and the supporting evidence packs for semi-scheduled solar farms, reconciled against the GPS and registration data. See our AEMO GPS applications service, or contact us.
Frequently Asked Questions
What is AEMO's Energy Conversion Model?
It is a description of a semi-scheduled solar or wind farm, prepared on AEMO's template, that tells AEMO's forecasting system how weather inputs convert into active power at the connection point. It is required for registration.
What is a cluster in the ECM?
A group of inverters and arrays with the same inverter model, module technology, similar DC capacity per inverter, the same mounting or tracking arrangement and a limited geographic extent, so that they can be modelled as one unit.
Why does AEMO want the inverter efficiency curve as a table?
Because the data is loaded into the forecasting system. A picture copied from a datasheet cannot be used, so the efficiency values must be provided in tabulated form.
What met station data does AEMO need?
Real-time measurements such as irradiance, temperature and wind speed, delivered at a short update interval, with the station locations and their mapping to clusters. Check AEMO's current guideline for the required update rate.