Marginal Loss Factors
How we model MLFs, and how they're integrated into the dispatch model
Marginal Loss Factors (MLFs) quantify the impact of electrical losses on energy transfer between a generator or load and the Regional Reference Node (RRN) within the NEM. They are crucial for accurately modelling revenue and dispatch outcomes, especially for battery energy storage systems (BESS).
1. Estimating loss proportions
We trained a regression model to estimate loss proportions for each settlement period. The model incorporates predictors such as:
- Proportion of renewable generation
- Transmission flow proportions
- Temporal factors (e.g., hour of day, day of week, season)
These estimates provide a dynamic view of losses across the network.
2. Calculating volume-weighted losses
By integrating the inferred loss proportions with BESS import and export data, we compute volume-weighted losses. This involves:
- Adjusting import/export volumes by the estimated loss factors
- Aggregating adjusted volumes over the financial year
- Deriving Average Loss Factors (ALFs) from the adjusted volumes
- Calculating MLFs by squaring the ALFs, based on the assumption that ALF is the square root of MLF
This approach aligns with the AEMC’s guidelines on loss factor calculations.
3. Location-based adjustments
Recognising the spatial variability of MLFs, we refine our estimates using location-based adjustments:
- Distance to RRN: We apply a linear regression model that correlates MLFs with the distance from the asset to the RRN.
- Historical MLF data: By identifying the nearest historical BESS site to the proposed battery site, we adjust the MLFs to reflect local network conditions.
4. Integration with dispatch modelling
Incorporating MLFs into dispatch modelling makes revenue forecasts and operational strategies in the model more accurate. Our approach ensures that:
- Location-specific factors are accounted for, providing tailored MLF estimates per asset
- Historical trends inform future projections, aiding in risk assessment and planning
Updated 6 days ago