Abstract

In this paper, we propose a three-stage robust generation and transmission expansion planning model considering generation profiles of renewable energy sources (RES) affected by different long-term climate states. Essentially, we extend the broadly utilized two-stage modeling approach to properly consider partial information of climate states with conditional short-term scenarios of RES output and outages. The proposed model is formulated as a five-level optimization problem. The first level determines the optimal generation and transmission expansion plan under uncertainty in climate conditions, RES generation, and contingencies. Given the selected expansion plan, the second level identifies the most severe climate state. Following the decision-information hierarchy, in the third level, the system operator optimizes the generation schedule of energy and reserves under perfect information of the climate state, but yet under uncertainty in the RES generation and contingencies. Then, the fourth level identifies the worst-case combination of contingency and conditional short-term RES generation adjusted to the current climate condition. Finally, the fifth level determines the optimal redispatch of reserves to react against the worst case RES generation and contingency scenario considering the uppermost decisions. Within this multilevel structure, the optimal investment plan considers a more realistic decision setting, where system operators adapt RES forecasts based on the observed climate conditions before planning the operational schedule. To solve the problem, a variant of the nested column-and-constraint-generation algorithm is proposed with global-optimality guarantee in a finite number of steps. A case study based on the Chilean system illustrates the applicability of the model in a realistic network. Published by Elsevier B.V.