Improved Shuffled Frog Leaping Algorithm And The Reservoirs Optimal Operation Model Research

Energy and resource is the indispensable production factors and material basis for the development of national economy and society, which is related to production development and people’s life. The hydropower, as a kind of renewable clean energy, occupies an extremely important position on Chinese energy history. The positive, reasonable and effective development of hydropower resource which is abundant in our country will play an important role in alleviating the country’s energy crisis, reducing pollution, achieving energy conservation and emissions reduction and the economic and social sustainable development. According to the situation of Chinese hydropower development and energy situation, the five-year power plan puts forward the principle of priority to the development of hydropower. During this period, hydropower enterprise has an unprecedented development and the river basin water resources is under further large-scale development and utilization, forming cascade hydropower group systems in major rivers. Therefore, at the background of energy shortage and advocating energy conservation and emissions reduction, research on optimal operation of cascade hydropower station is an important way to achieve comprehensive tilization of the river basin hydropower.Reasonable, optimized reservoirs operation can coordinate flood control and impounding, and improve the utilization efficiency of water resources, which is an essential way to alleviate supply and demand of water resource and hydropower. To study the optimal scheduling of cascade hydropower station and flood resourcification, combining the theory of system engineering, swarm intelligence optimization theory, credibility theory, his paper did in-depth research in improvement of intelligent optimization algorithm, the single goal optimal operation, optimization of cascade total output dispatching chart, parallel algorithm, dynamic control of the water level in flood period and obtained some research achievements of theoretical significance and engineering application value. The main research results are as follows:(1) Improved shuffled frog-leaping algorithm. The traditional shuffled frog-leaping algorithm may fall into a local optimum. To solve this problem and enhance the convergence speed, in this paper, it combined the shuffled frog leaping algorithm with the cloud model, then the Normal Cloud Mutation Shuffled Frog Leaping Algorithm (NCM-SFLA) is presented for the cloud model being stable but variable and having both randomness and stable orientation. The paper tested the algorithm with some complex testing function, proving that NCM-SFLA is effective, precise and stable in both precision and speed.(2) The cascade reservoirs optimal operation based on NCM-SFLA. For the complex relationship between hydraulic and electric power, in order to improve the optimization efficiency of intelligent optimization algorithm in solving cascade reservoirs optimal operation, this paper gives the model structure and calculation steps to solve optimal operation at the condition of analyzing the period water level’s feasible region’s building method and the initial solution’s generating method. Taking the long-term cascade reservoirs optimal operation and short-term cascade reservoirs optimal operation for example, compared with the conventional optimization method, it shows the result of solving the reservoir operation with NCM-SFLA.(3) Optimization of cascade total output dispatching chart with NCM-SFLA. Because of the compensation function between different reservoirs, the power generation efficiency and power reliability of cascade reservoirs will have a certain degree of ascension. To maximize generated energy at the condition of meeting the minimum guarantee output and reliability, the initial dispatching chart is figured out. But while calculating the dispatching chart, the systemic optimization is ignored. So NCM-SFLA can optimize the cascade total output dispatching chart. In the paper, it analyzed the way of encoding, management of constraints, the period water level’s feasible region’s building method and the initial solution’s generating method, and proved the capability of the algorithm using Li Xianjiang basin as an example.(4) The parallel algorithm of NCM-SFLA and its application. The capability of intelligent optimization algorithm will be weaken as the dimensionality coming up, when solving the cascade reservoirs optimal operation. In this paper, it presented PNCM-SFLA based on Fork/Join parallel execution model and the process of solving the cascade reservoirs optimal operation using PNCM-SFLA. In this paper, it analyzed the rationality and computational efficiency and performance of PNCM-SFLA in Li Xianjiang basin(5) The research on dynamic control of the water level in flood period considering the flood forecast error. Considering that the simple handle of the runoff forecast using the probability theory cannot reflect the uncertainty of error, affecting the precision of the forecast and scheduling decision. In this paper, it explored the method of quantizing the randomness and fuzziness of runoff forecast, using probability theory to describe flood volume which is regarded as a random variable, using credibility theory to describe forecast error which is regarded as fuzzy random variable, and establish forecast error quantization model based on credibility theory. According to fuzzy chance program theory, using flood operation theory, in this paper, it built fuzzy chance constrained programming optimization model of water level in flood period. The model calculated the dynamic control field of water level in flood season which is not lower than flood control standard, operating design flood under variant flood period imited water level and forecast error. In the paper, it is verified that the rationality and feasibility of the model based on credibility theory by the example of Fengtan Reservior.