Genetic Algorithm In The Reservoir Long-term Applied Research, Optimization Scheduling

The optimization dispatch of reservoir is the central content of economic operation for hydroelectric station, and is the major measure for managing and controlling the safety and reliability of reservoir, rational using the water resources and displaying comprehensive efficiency of reservoir, which is the key problem for formulating and implementing the long-term optimization plan of hydroelectric station and reservoir. After consulting large domestic and overseas documents of reservoir dispatch, this thesis researched the situation of long-term optimization dispatch of Xiaowan hydroelectric station.This thesis researched the mathematical model of long-term optimization dispatch of the single reservoir firstly, based on analyzing the definite model and random model specially, the optimal model of Xiaowan hydroelectric station was established which is objecting to the maximal electricity production. According to the runoff data in possession, this thesis did the work of traditional calculation and optimal calculation. Traditional calculation is based on the equal flow while the time-interval is month. Optimal calculation is uses modified genetic algorithm to calculate the maximal electricity production model of Xiaowan hydroelectric station. Based on analyzing and researching the principle of standard genetic algorithm, this thesis put forward an improved genetic algorithm through modifying the select algorithm and mutation algorithm. Especially in modifying the mutation algorithm, this thesis put forward the method of two-direction selected mutation, which improved the ability of local search of mutation algorithm.In order to simplifying the calculation, this thesis used the combination of pluriennial regulation and annual regulation while using genetic algorithm to calculate the optimal model. This method calculated the model based on pluriennial regulation firstly, then calculated the model based on annual regulation for some representative years, which is to amend the result of annual regulation. At last, this thesis analyzed and compared the results of traditional calculation and genetic algorithm, and draw the coherent charts of reservoir dispatch. Based on analyzing and comparing the results, it arrived at the conclusion that improved genetic algorithm can improve electric production 3.03% compared with the traditional calculation, which supplied a practical method for the work of long-term optimization dispatch of the reservoir.All calculation programs in this thesis are developed with visual basic 6.0, which is oriented to objects. Visual basic 6.0 is a real object-oriented and visual method. It can be correlated to data base easily, which can read data from the table of excel and write data to the table by some easy codes. In some implementation procedure of complex calculation like reservoir dispatch, it's a simple and effective tool.