Multi-objective Optimal Operation For Reservoirs

Reservoir operation mode will determine the economic performance of generation and other performance of power stations, such as navigation, irrigation and so on. In this paper, focused on the status of small and medium sized reservoirs, the characteristics of the current operation mode was analyzed, long-term multi-objective stochastic optimization of reservoir operation was deeply studied. The solution to the problem of reservoir operations was proposed, in order to get the maximum Comprehensive Benefit of the reservoir.In this paper, small and medium sized multi-objective stochastic optimization of reservoir operation was studied, the key works was mainly divided into two parts, one is Multi-objective decision-making method, and the other is random description of runoff series.Based on analysis of today's various Multi-objective decision-making methods, it can be seen that classical multi-objective decision-making method is difficult to be used in Description of the multi-objective problem. The modern multi-objective decision-making method can make up for that deficiencies, but not with the optical result obtained the corresponding changes in scheduling program. In this paper,a comprehensive linear multi-objective algorithm for multi-objective decision-making was presented. The optimal scheduling program can be found by searching in the target space. Customers can make their decision based on their needs. In this paper, the Jinjiang Reservoir, for example, the method of Range Variable Algorithm (RVA) was improved to get the ecological goal. These multi-objective problem include two objectives, such as ecological and generation goal. Results showed that the multi-objective decision-making method can be very good Coordination between the two goals and guide clients in making decisions.The random description of Runoff will determine the results of stochastic optimal operation of reservoir. Due to short of data of medium and small reservoirs'common characteristics, it is difficult to statistical analyze the data when the runoff series was randomly described .In this paper, AR model was used to extend the data, combined with transition probability formula, a PⅢ-AR model was presented. It can be clearly seen from the two transition probability matrix that from the original AR model data and the PⅢ-AR model data that the PⅢ-AR model was better in extending the hydrological data. Those two transition probability matrix was applied to stochastic optimization scheduling, results showed that the PⅢ-AR model can improve scheduling performance.