Stochastic optimization applied to Brazilian hydroelectric seasonalization of assured energy

The Brazilian power system is mainly composed of hydroelectric generation. Due to the stochasticity of inflows each hydroelectric plant receives an assured energy certificate, also known as physical guarantees. In general, the assured energy certificates correspond to the energy that each plant can produce at a 95% confidence level., They also represent the maximum energy each hydro unit can contract. So, the hydro generators are constrained by the assured energy and they determine their portfolio of energy contracts with free consumers, other generators, distribution companies and trade companies accordingly. The portfolio is composed of bilateral contracts that are, mainly composed of long term contracts negotiated significantly before the power delivery. Even considering that the assured energy is constant, each generator can seasonalize it on a monthly base, but keeping to the total annual assured energy. This paper aims to use stochastic dual dynamic programming to optimize the seasonalization of this assured energy considering the bilateral contracts previously assigned and the projection of the spot price. The results obtained show a considerable increase in the revenues of a generator utility from the Brazilian system.