Abstract
Reservoir operation is studied for the Daule Peripa and Baba system in Ecuador, where El Niño events cause anomalously heavy precipitation. Reservoir inflow is modelled by a Markov-switching model using El Niño-Southern Oscillation (ENSO) indices as input. Inflow is forecast using 9-month lead time ENSO forecasts. Monthly reservoir releases are optimized with a genetic algorithm, maximizing hydropower production during the forecast period and minimizing deviations from storage targets. The method is applied to the existing Daule Peripa Reservoir and to a planned system including the Baba Reservoir. Optimized operation is compared to historical management of Daule Peripa. Hypothetical management scenarios are used as the benchmark for the planned system, for which no operation policy is known. Upper bounds for operational performance are found via dynamic programming by assuming perfect knowledge of future inflow. The results highlight the advantages of combining inflow forecasts and storage targets in reservoir operation. © 2014 © 2014 IAHS Press.
| Original language | English |
|---|---|
| Journal | Hydrological Sciences Journal |
| Volume | 59 |
| Issue number | 8 |
| Pages (from-to) | 1559-1581 |
| Number of pages | 23 |
| ISSN | 0262-6667 |
| DOIs | |
| Publication status | Published - 2014 |
Keywords
- Water Science and Technology
- Ecuador
- El Niño
- genetic algorithm
- inflow forecast
- optimization
- reservoir operation
- Atmospheric pressure
- Dynamic programming
- Genetic algorithms
- Optimization
- Heavy precipitation
- Hydro power production
- Inflow forecast
- Management scenarios
- Operational performance
- Optimized operations
- Reservoir operation
- Forecasting
- benchmarking
- El Nino-Southern Oscillation
- flow modeling
- inflow
- Markov chain
- numerical model
- precipitation assessment
- reservoir