ALBERT

Description: The highly temporal variability of hydrological response in Mediterranean areas affects the operation of hydropower systems, especially in Run-of-River plants located in mountainous areas, where the water flow regime strongly determines the failure, defined as no operating days due to either, flows below the turbine minimum discharge and environmental flow requirements, or flows above the turbine maximum discharge. A Bayesian dynamics forecast model is developed from statistical modelling of both, forcing agents of runoff generation, and water inputs to the plants, as dependent variable. Failure frequency analysis and its related operationality, along with their uncertainty associated at different time scales is carried out through the Monte Carlo 250 stochastic replications of the 20-year period of forcing agents. Finally, a tool for forecasting the occurrence of failure from 1 to 7 months ahead is developed that allows to analyse different scenarios of storage in the plant loading chamber. The approach is applied to mini-hydropower plants in Poqueira (Southern Spain), where the snow and rainfall regimes determine the hydrological regime.Results revealed the snow influence as the lowest probability of failure was found in April and May when snowmelt is outstanding. Regarding the forecast tool, the analysis of different scenarios successfully reflected the decrease of the probability of failure on a monthly scale with greater stored water volumes, while on a weekly scale the improvement is only appreciable in the dry season (June-October) when the snow influence is negligible.