In the December 2024 AEMO intermittent Generator Forum, the Australian Energy Regulator (AER) accepted a welcome opportunity to present its perspectives on self-forecast biasing.
Self-forecasting is the provision of an unconstrained intermittent generation forecast that represents a unit’s power generation availability for the end of the upcoming dispatch interval. Sometimes it is also referred to as the dispatch self-forecast as it is used in dispatch to determine the unit’s generation availability.
Significant concerns
In short, AER communicated that it holds significant concerns with the biasing of self-forecasts for commercial benefit. It expects participants ensure that self-forecasts contain an accurate reflection of expected generation capability.
Ongoing work on compliance and enforcement
The presentation leveraged the work of the AER’s Wholesale Compliance Team. It is motivated by Priority 3 of AER’s Compliance and Enforcement (C&E) Priorities for 2024/25 which is to:
Support power system security and an efficient wholesale electricity market by focusing on generators’ compliance with offers, dispatch instructions, bidding behaviour obligations and providing accurate and timely information to Australian Energy Market Operator (AEMO).
Readers can find the C&E Priorities for 2024/25 through the link below:
Self-forecasting, a valuable addition to the market
The concerns around biasing raised by AER don’t represent call to shut down the self-forecasting capability for participants.
AER said that self-forecasting helps AEMO access more accurate wind and solar availability forecasts and this improves power system security.
It’s clear AER sees self-forecasting as a valuable addition to the market. This is a sentiment shared widely across industry.
The problem with biasing
The AER concerns over biasing – where the unconstrained dispatch availability forecast is deliberately under or over stated – were highlighted in relation to:
- Adverse impacts on system security and dispatch outcomes,
- Reductions in the accuracy of AEMO’s forecasting,
- Competitive advantages due to targeting lower ‘causer pays’ costs (and therefore those costs shared among fewer others), and
- Potential higher costs being passed on to consumers.
In targeting lower causer pays costs by employing self-forecast biasing participants are treating self-forecasts are commercial parameters. AER clearly mentioned that semi-scheduled generators should seek not to treat self-forecasts as commercial parameters.
AEMO’s Manager of Operational Forecasting also noted, in ensuing discussion, that biases do adversely impact power system operation, and make AEMO forecasts less dependable impacting control room operations which must trust self-forecasts.
A further expectation of AER’s, related to having better forecasts, was that
- semi-scheduled generators should regularly assess the performance of their self-forecasts and apply learnings to continually improve the accuracy of the forecasts.
WattClarity interest in self-forecasting continues
Here at WattClarity we’ve long been interested in the performance of intermittent generator self-forecasting and have collated a series of articles on the topic over the years at this link. Often we’ve reflected on the ”the purpose of self-forecasting” when observing unusual patterns.
In 2024, Global-Roam partnered with a new startup to support wider access to high quality intermittent generator forecasts, and detailed understanding of performance. This new startup draws upon the experience of a highly capable and knowledgeable team who have extensive experience in operational forecasting and deep market understanding owing to prior roles in AEMO and working closely with with industry. The service provides participants with an opportunity to source accurate forecasts and avoid exposure to the risk of the biased forecasting that AER is monitoring.
Going forward, FPP to the rescue?
AER anticipates the upcoming Frequency Performance Payments (FPP) mechanism will deter self-forecasters from employing biases.
FPP is the replacement to the causer pays mechanism for allocating costs of regulation FCAS. FPP also aims to compensate units for the provision of primary frequency response.
Conservatively, AER remains vigilant for potential gaps in outcomes of the new mechanism, relative to expectations.
Some behavioural change is anticipated (following today and with introduction of FPP). And we are still yet to even test the water with the FPP mechanism – the FPP non-financial operation period commences 9 December 2024. So we’ll probably have to wait and see whether the practice of biasing continues in any form.
The AER plans to continue monitoring for biasing, carefully. The door to considering rule changes, if issues persist, remains open.
Leave a comment