Python for Network: Tariff to Load Profile (PyNTLP)
Three hours of free electricity a day is something no load forecast has seen before.
Since July 2026, households in parts of Australia can choose a plan that makes electricity free from 11:00 to 14:00 every day[1]. Network forecasts learn from years of history, and none of that history contains a standing free window. PyNTLP is open-source software that turns a tariff like this into a change to the load profile a forecast can use.
The problem
A forecast learns from history, and the history has no free window
Electricity networks forecast net load, the power their customers draw from the grid less what rooftop solar sends back, to plan the network and decide where to invest. These forecasts learn from years of past readings, weather and the calendar, and assume the future will look like the past.
A new tariff breaks that assumption. When a whole region can use electricity for free in the same three hours every day, load in those hours can change in ways the history cannot show. It is a structural break, and a forecaster needs to know what happens inside the window before adjusting for it.
- What does load do in these three hours?
- Which households respond?
- How should a forecast allocate it?
Key facts
Two policies, one year apart
Retailers were selling free midday plans before the Solar Sharer Offer began, such as OVO Energy's Free 3 and GloBird Energy's ZeroHero, both free from 11:00 to 14:00[6, 7].
Why it matters
Today's estimates are assumptions, not measurements
Forecasts learn from history
A tariff that was never in force leaves no trace in past readings, so a forecast trained on them cannot learn its effect.
Estimates assume a share
The government's impact analysis assumed households would move a share of their electricity use, 10 to 30%, into the window[8].
Elasticity breaks at zero
Economic models move load in proportion to the percentage change in price[9]. When the price falls to zero, that percentage is undefined.
Nobody has measured it
What load actually does in a standing daily free window has not been measured; the first study of the scheme lists the actual load-shift curves as unknown[10].
Networks plan for peak and minimum demand years ahead, so the question cannot wait for the history to catch up.
The research question
Three questions about one free window
What happens to load inside the window, which households respond, and how should a forecast allocate the change across the window?
A journal article in preparation answers them with smart-meter data from Ausgrid's network, in three steps.
Measure
Measure the load response to a standing daily free window from half-hourly smart-meter readings, across Ausgrid's network.
Attribute
Compare groups of households defined from Ausgrid's register of rooftop solar and other distributed energy resources, to find where the response comes from.
Model
Build a model that allocates the response inside the window, and test it against the allocations forecasters use today.
The findings will appear with the article.
A separate study
Adaptive Quantile Flexibility (AQF)
AQF is a separate piece of research, not part of the PyNTLP package. It asks how much of an aggregate load profile is flexible without hand-picking a quantile to split it.
At each time of day, AQF fits two Gaussian curves to the daily readings and computes from that fit the quantile that lands on the underlying load, with a test that flags the times when the two cannot be told apart.
On synthetic data, AQF cut the error in the estimated underlying load to about a third of the best fixed quantile's, about halved the median error in recovered flexible energy, and came close to an oracle that knows the true quantile wherever the separation test was met[11].
Research
Built on a PhD study with Ausgrid
PyNTLP grows out of a PhD study at UNSW Sydney in partnership with Ausgrid's Demand Forecasting team. A journal article on the Solar Sharer Offer is in preparation. The separate AQF study is accepted for the 2026 IEEE PES Innovative Smart Grid Technologies Asia conference[11].
The team
Who made it
UNSW Sydney
M. Syahman Samhan, Anna Bruce and Baran Yildiz, School of Photovoltaic and Renewable Energy Engineering
Ausgrid
Navid Haghdadi and Daniel Sim, Demand Forecasting
PyNTLP is part of M. Syahman Samhan's PhD study, supported by the University International Postgraduate Award (UIPA) from UNSW and top-up scholarships from Ausgrid, RACE for 2030 and the NSW Decarbonisation Innovation Hub. It is open source under the MIT licence.
References
- Australian Energy Regulator, "The Solar Sharer Offer (SSO): What you need to know," Energy Made Easy, last updated 30 June 2026. energymadeeasy.gov.au
- Department of Climate Change, Energy, the Environment and Water, "Energy plan now provides daily free-power window," 1 July 2026. dcceew.gov.au
- Clean Energy Regulator, "Small-scale Renewable Energy Scheme," Quarterly Carbon Market Report, December quarter 2025, last updated 5 March 2026. cer.gov.au
- Clean Energy Regulator, "Solar batteries," last updated 31 July 2026. cer.gov.au
- Ausgrid, "Distribution and Transmission Annual Planning Report," December 2025, pp. 6 and 14. ausgrid.com.au
- OVO Energy, "The Free 3 plan." ovoenergy.com.au
- GloBird Energy, "ZeroHero." globirdenergy.com.au
- Department of Climate Change, Energy, the Environment and Water, "Options to Provide Solar Sharer Offer: Impact Analysis (Second Pass)," Office of Impact Analysis, 2026. oia.pmc.gov.au
- D. S. Kirschen, G. Strbac, P. Cumperayot and D. de Paiva Mendes, "Factoring the elasticity of demand in electricity prices," IEEE Transactions on Power Systems, vol. 15, pp. 612–617, 2000. doi.org/10.1109/59.867149
- R. Galvin, "Free midday electricity in a Solar Sharer scheme: Should Germany follow Australia's lead?," Renewable Energy, vol. 267, 125735, 2026. doi.org/10.1016/j.renene.2026.125735
- M. S. Samhan, N. Haghdadi, D. Sim, A. Bruce and B. Yildiz, "Adaptive Quantile Flexibility: Recovering Candidate Flexible Load from Aggregate Profiles for Network Tariff Modelling," accepted for the 2026 IEEE PES Innovative Smart Grid Technologies Asia (ISGT Asia), Wuhan, China, 30 October to 1 November 2026.