Research On Optimal Operation Methods For Xinfengjiang Reservoir With Utilization Of Energy Generation And Water Supply

Under the premise of satisfying the requirements of flood control, reservoirs in China are facing dual pressure from energy generation and water supply because of the increasing proportion of hydroelectric in power system and soar demands of water in industry, agriculture and people’s daily life. Moreover, the frequent occurrence of extreme weather brings huge challenge to reservoirs’operation and dispatching.Reservoir optimal operation refers to seeking the optimal operation scheme and control strategy for reservoir under different demands through applying specific optimization method and technique. Current reservoir operation schemes mainly consider the reliability and economy of energy generation, in order to maximize the using of hydropower resources controlled by reservoir. However, for those reservoirs under multiple utilization demands, an overall consideration of water supply reliability and energy generation reliability has important realistic meaning. And making reasonable operation rules is one of key problems to realize complex objective requirements during operation period of reservoirs and hydropower stations.According to runoff randomness and optimal operation problems caused by reservoir’s changing of function, this paper summarizes relative research on reservoir optimal operation and takes Xinfengjiang reservoir as an engineering background. Firstly, this paper studies reservoir level control problem while considering surplus discharge risk. Then, this paper studies optimal operating model for reservoir with utilization of energy generation and water supply. Finally, by introducing genetic algorithm, this paper further optimizes the operation rule generated by the previous model. The main contents of this paper are as follows:(1) Aiming at the problem of reservoir optimal level control in actual operation, this paper proposes an optimal trajectory optimization model, which is aimed at maximizing the annual average energy output and constrained by the probability of surplus discharge. Then the model is solved by simulation-optimization method. Practical application in Xinfengjiang reservoir shows that the proposed method has comprehensively considered the coordination problem between controlling surplus discharge risk and improving energy output while searching reservoir optimal level control line, which is beneficial to practical application. (2) Aiming at the optimal operation problem of reservoirs which undertake both energy generation and water supply task, this paper proposes a stochastic dynamic programming model with triple constraints of energy generation reliability, water supply reliability and minimum water supply flow. The two reliability constraints are addressed using the Lagrangian relaxation, and the minimum water supply flow is addressed with penalty function. The updating process of The Lagrangian multiplier and penalty coefficient is guided by simulated operation results. Case study for Xinfengjiang reservoir shows that the proposed method can effectively avoid extreme failure on a basis of satisfying both energy generation and water supply task and can be applied to reservoirs with urban water supply task.(3) Based on an improved genetic algorithm, this section implements further optimization of the previous operation rule while choosing annual average energy maximization as optimization target. Decimal float is chosen as genetic coding and a specific strategy is adopted for population initialization. Genetic operation include championship selection, two-point crossover and non-uniform mutation. Calculation results of Xinfengjiang reservoir show that the previous operation rule is well improved and can obtain more energy.Finally, this paper is summarized, and problem is prospected to be studied in the future.