6th International Conference Energy & Meteorology: Abstract Submission

Quantifying the value of improved probabilistic power forecasts for power system applications using stochastic dispatch models (725)

Bruno Schyska 1 , Lueder von Bremen 1 2 , Wided Medjroubi 1 , Detlev Heinemann 1 2
  1. DLR-Institute of Networked Energy Systems, Oldenburg, LOWER SAXONY, Germany
  2. University of Oldenburg, Oldenburg, Germany

Objective and Background

Probabilistic forecasts have been promoted by meteorologists for years. But although forecast skill has steadily increased, the use of probabilistic forecasts in the economy and industry is still limited. One reason for that is that improvements in forecast skill – as measured by the continuous ranked probability score (CRPS) for instance – are hard to grasp. Benefits arising from these improvements can hardly be assessed, in particular as real-world power system applications, which use probabilistic forecasts, rarely exist. The objective of this study is to contribute to a better understanding of the value of probabilistic power forecasts. Therefore, we introduce a novel approach to translate probabilistic forecast skill into costs based on network constrained stochastic dispatch models.


Method

Usual approaches to derive the cost-optimal power dispatch within a market region do not account for the potential balancing costs arising from the dispatch decision. It has been shown by Morales et al. [2014] that this leads to sub-optimal market clearing. To overcome this issue they proposed a stochastic market clearing model. In this model, average balancing costs are estimated from a set of scenarios. We use this approach to translate improvements in forecast skill into costs. Scenarios are derived from the 50 member ECMWF ensemble forecasts [Leutbecher and Palmer, 2007] and different ensemble calibration techniques for a simplified European power system. Based on the resulting costs, the scenarios can directly be compared. Furthermore, the understanding of the underlying improvements in skill can be increased by deriving a relation between forecast skill and costs.

Principal Findings

We found that increases in forecast skill can clearly be translated into reduced costs. For the illustrative example of a highly renewable European power system, for instance, reductions in CRPS of 175% translate to a reduction of balancing costs of 20%.

Conclusion

Based on calculations for a simplified European power system, we showed that stochastic market clearing models are a valuable tool to link the skill of probabilistic power forecasts with costs. We hope that describing this relationship contributes to increase the understanding, value and use of probabilistic power forecasts.

  1. Morales, J.M., Zugno, M., Pineda, S., and Pinson, P. (2014): Electricity Market Clearing with Improved Scheduling of Stochastic Production, European Journal of Operational Research, 253(3)
  2. Leutbecher, M., and Palmer, T.N. (2007): Ensemble forecasting, Journal of Computational Physics, 227