An Adaptive Robust Optimization Model for Power Systems Planning with Operational Uncertainty

Felipe Verástegui, Álvaro Lorca, Daniel E. Olivares, Matías Negrete-Pincetic, Pedro Gazmuri

Research output: Contribution to journalArticlepeer-review

63 Scopus citations


There is an increasing necessity for new long-term planning models to adequately assess the flexibility requirements of significant levels of short-term operational uncertainty in power systems with large shares of variable renewable energy. In this context, this paper proposes an adaptive robust optimization model for the generation and transmission expansion planning problem. The proposed model has a two-stage structure that separates investment and operational decisions, over a given planning horizon. The key attribute of this model is the representation of daily operational uncertainty through the concept of representative days and the design of uncertainty sets that determine load and renewable power over such days. This setup allows an effective representation of the flexibility requirements of a system with large shares of variable renewable energy, and the consideration of a broad range of operational conditions. To efficiently solve the problem, the column and constraint generation method is employed. Extensive computational experiments on a 20-bus and a 149-bus representation of the Chilean power system over a 20-year horizon show the computational efficiency of the proposed approach, and the advantages as compared to a deterministic model with representative days, due to an effective spatial placement of both variable resources and flexible resources.

Original languageEnglish
Article number8718350
Pages (from-to)4606-4616
Number of pages11
JournalIEEE Transactions on Power Systems
Issue number6
StatePublished - Nov 2019
Externally publishedYes


  • Generation expansion planning
  • renewable energy
  • robust optimization
  • transmission expansion planning


Dive into the research topics of 'An Adaptive Robust Optimization Model for Power Systems Planning with Operational Uncertainty'. Together they form a unique fingerprint.

Cite this