The dataset presents estimates of thermostatically controlled load (TCL) over multiple years for zone substations in the AusNet supply area. Estimates reflect, for each substation, the half-hourly aggregate TCL demand, drawn from a mathematical model developed by the NEAR Program team. The approach uses cold, mild and hot days to identify typical heating and cooling trends for each zone substation, and then estimates the distribution of TCL model parameters such as air-conditioner set-points and power. The resultant TCL model is then used to provide half-hourly estimates of TCL demand based on local weather data. See Related Resources for further methodological detail and for the total load for substations in the AusNet supply region.

Note that thermostatically controlled load is dominated by air-conditioning and the dataset therefore rests as a suitable proxy for air-conditioning. The estimates provided will include refrigeration and other thermostatically governed loads.

Note that results include approximate quality estimates to guide use of the dataset. The mean absolute percentage error (MAPE) values provided indicate the relative accuracy of the AC estimates. As ground-truth aggregate air-conditioner load data is unavailable for substations, accuracy reflects how close, for each substation, the sum of the estimated air-conditioner load and an estimate of baseline load is to the measured total substation load. High MAPE values indicate that the AC estimates (when summed with the baseline load) are meaningfully different from the measured load and should be treated with caution.

Note that zone substations that had more than 40% of load data missing for each year have been excluded in this set.