The Study For Optimal Operation Rules Of Lancang River Cascade Reservoirs

As the rapid increase of Centralized production of giant cascade reservoirs, the focus of water conservancy development has been experiencing the transition from construction to management. These cascade reservoirs have characteristics of great numbers, large installed capacity and wide range of electrical transportation. Reasonable operation and management can fully play the role in bringing the benefit and abolishing the harm, thus to maximize economic and social benefits. Operation rules of cascade reservoirs is an effective tool to guide the daily operation and management of the reservoir. Under the background of the national energy generation scheduling, how to make operation rules of cascade reservoirs has become imminent water development problem to be solved.The operation rules of hydropower reservoirs are an important tool to guide reservoir operation. In this paper, the cascade hydroelectric plants on Lancang River are chosen as the study subject. The implicit stochastic optimization principles are utilized to screen the state variables and the decision variables, and use multiple linear regression analysis to establish two kinds of operation rules, one with and the other without runoff forecasting. Initial solutions are used for operation simulation. Based on simulation results, coordinate rotated method is adopted to improve operation rules. The main contents are as follows:1. Combined with the58years (1953-2010) runoff data of five cascade reservoirs on Lancang River, which are Xiaowan, Manwan, Dachaoshan, Nuozhadu, Jinghong, the deterministic optimal operation model of cascade hydropower stations are established. Then use successive optimization algorithm (POA), combined with state-by successive approximation dense discrete differential dynamic programming algorithm to solve the model, and obtain the required data samples, which needed by regression analysis.2. According to the obtained samples of the optimal scheduling, the beginning regulating storage capacity of each reservoir at face period are selected as state variables, and the average output are selected as decision variables. Then use multiple linear regression analysis to obtain the initial scheduling function, and use the initial scheduling function to simulate operation. The simulation results are used to optimize the initial scheduling function, and the optimal scheduling of cascade hydropower stations function without considering the forecasts are obtained. 3. In current research of operation rule, historical run-off data is used and that has resulted in no consideration of the influence of prediction error to simulating operation. To solve this problem, an operation function and optimization model with predicted value of available water is proposed. In this model, the predicted value is effective storage in the face stage plus inflow forecasted by seasonal ARIMA model, which considers the uncertainty of forecasted run-off.At last, the conclusion summarizes the results of research and gives some suggestions of further study.