Australian Solar Energy Forecasting System

ASEFS phase 1 (ASEFS1) produces solar generation forecasts using a combination of statistical methods and Numerical Weather Prediction (NWP)-based models. It uses the following inputs to produce solar generation forecasts for large solar power stations:

• Real time Supervisory Control and Data Acquisition (SCADA) measurements from the solar power station.
• NWP data from multiple weather data providers.
• Standing data from the solar power station as defined in the ASEFS Energy Conversion Model.
• Additional information submitted by the solar power station, including inverter availability and an upper MW limit on the solar farm.
• Imagery from Himawari satellites.

AESFS1 produces solar generation forecasts for all NEM forecasting timeframes as follows:

• Dispatch (five minutes ahead).
• 5 Minute Pre-dispatch (five minute resolution, one hour ahead).
• Pre-dispatch (30 minute resolution, up to 40 hours ahead).
• ST PASA (30 minute resolution, seven days ahead).

The ASEFS Energy Conversion Model and relevant guides can be found here.

ASEFS phase 2 uses a combination of statistical and physical methods and NWP-based models. It uses the following inputs to produce aggregated regional solar generation forecasts for small-scale PV systems:

• NWP data from multiple weather data providers.
• Output measurements from selected household rooftop PV systems from PvOutput.org and Solar Analytics.
• Static data from selected systems from PvOutput.org and Solar Analytics, such as inverter size and model.
• Aggregate kilowatt capacity by installed postcode for small-scale solar systems as recorded by the Clean Energy Regulator.
• Imagery from Himawari satellites.

ASEFS2 produces solar generation forecasts for the following NEM forecasting timeframes:

• Pre-dispatch (30 minute resolution, up to 40 hours ahead).
• ST PASA (30 minute resolution, seven days ahead).

Forecasts and actuals for aggregated small-scale (rooftop) PV systems as produced by ASEFS2 can be found here. There are several types of estimated actuals produced by ASEFS2, described in the table below.

The estimated actuals and forecasts for rooftop PV are published via the Data Interchange through the following fields described in the EMMS Data Model:

Rooftop PV data files

Half-hourly Actual Rooftop PV
filenames: public_rooftop_pv_actual_measurement*.csv, public_rooftop_pv_actual_satellite*.csv
These files contain the historical rooftop PV fields from the ROOFTOP_PV_ACTUAL table by region for the previous half-hourly interval.

Half-hourly Forecast Rooftop PV
filename: public_rooftop_pv_forecast*.csv
This file contains the half-hourly forecast rooftop PV fields from the ROOFTOP_PV_FORECAST table, by region covering the pre-dispatch and ST PASA timeframe.

Rooftop PV by area data files 

Rooftop PV generation data is also published for the individual areas that are used in preparing the load forecasts for pre-dispatch and STPASA. These load forecasting areas are defined in the SO_OP_3710 – Load Forecasting procedure.

Half-hourly Actual Rooftop PV by Area
filenames: public_rooftop_pv_actual_measurement_area*.csv, public_rooftop_pv_actual_satellite_area*.csv
These files contain historical rooftop PV fields similar to the ROOFTOP_PV_ACTUAL table for the previous half-hourly interval, however they are split by load forecasting area instead of by region.

Half-hourly Forecast Rooftop PV by Area
filename: public_rooftop_pv_forecast_area*.csv
This file contains half-hourly forecast rooftop PV fields similar to the ROOFTOP_PV_FORECAST table covering the pre-dispatch and ST PASA timeframe, however it is split by load forecasting area instead of by region.