Short-Term Optimal Operation For Large-Scale Hydropower Plants

During the past more than 60 years, especially in the past decade, hydropower system has been undergoing the most concentrated and intense development boom in China. Hence, some large-scale hydropower systems for provincial even regional power grids have emerged, which are composed of many huge cascaded hydropower plants. In general, these hydropower systems are characterized with many plants, large installed capacity, discrepant running demands, wide transmission, and complex hydraulic and electric link. Consequently, a lot of general difficulties or problems in operation and management have been brought about. First, the huge hydropower plant and power grid are facing great challenges due to multi-vibration zones with high head. Second, regulating peak load for provincial power grids becomes more difficult while the operation limitation in transmission lines is considered. Third, the great difference in the inflows, total capacity of plants located on different streams, and operation requirements among hydropower plants, single-objective optimization is hard to express the discrepant demands. Furthermore, there is a pressing need for large-scale inter-provincial electric power transmission because of uneven distribution of resources and power consumption in the regional network. Therefore, a number of engineering in one regional power grid, five provincial power grids, and several cascaded hydro operation centers are chosen in the present paper. A series of effective solution methods and strategies are proposed for addressing some key problems faced currently, showing valuable results. The content of present paper is summarized as follows:(1) A short-term operational model based on unit commitment method is presented for huge cascaded hydropower plants. In this model, the objective is to maximize effective generation production of cascaded hydropower plants. According to load demands, generation weighting coefficients in different periods are provided so that all plants can generate in the order of system loads. A solution method based on the unit commitment is provided to optimize the model. During optimization, the assembled mathematical principle and dynamic programming are taken advantaged of to generate combined units and their vibration zones and optimal power curves respectively. Furthermore, the heuristic search method which has incorporated a rapid and efficient strategy for jumping out the vibration zones designed according to the relationship between vibration zone and water head is used to generate the operational schelduling of hydropower plants. The above solution model and technique are implemented on the Hongshuihe River hydropower system with 13 hydropower plants, generating effective and satisfactory results for real projects. (2) An objective of maximizing peak power generation is used to cut down peak load and secure stable operation in power grid, a solution method called successive approximation of dynamic programming with associated searching strategy is developed for solving operation problem of large-scale hydropower plants. The proposed model considers the novel limitation of multi-layer generation limitation and complex running constraints. The main objective of this model is to cut down peak loads in order to smooth the remaining loads for thermal systems. Thus, the times of operation and shut-down of thermal units should be reduced. The proposed solution method effectively combines the advantage of reducing dimensions and the strong searching capability to deal with the curse of dimensionality and complex operations and security requirements. Also importantly, the generation limitation are loosed during optimization and considered after optimization in order to simplify the original problem and keep the optimized results basically. The proposed model and method are tested by the hydro system in Yunnan Power Grid. The results indicate that the substitute of original objective function is reasonable, and that solution technique developed is efficient than successive approximation of dynamic programming in solving the present problem.(3) A novel scheduling method with discrepant objectives is proposed to consider the complex characteristics and demands among large-scale hydropower plants. Coorespondingly, all hydropower plants are classed into several types, and every plant group is responsible for a specific task which decides the operation objective function. To enhance the solution efficiency and improve the optimal solutions while considering complex generation fluctuation constaints, a multi-step of progressive optimality algorithm is proposed to solve short-term hydro scheduling problem. Within the framework, the original problem is first changed into a sequence of problems with different time intervals bigger than fifteen minutes. Next, the optimal solution is gotten by iteratively optimizing SHSP with an initial solution from last bigger time interval using POA. The efficiency of this algorithm is enhanced because the feasible range of state or decision variables will be greatly increased by weakening even eliminating some constraints mentioned earlier and synchronously the number of computational stages also reduced. Three different case studies are utilized to test the discrepant-objective method and MSPOA. The results show that MSPOA can get better results than POA within shorter time, and discrepant objectives among the hydropower plants can express the operation characteristics of large-scale hydroelectric systems more accurately than unique objective or multiple objectives.(4) Because of uneven distribution of resources and power consumption in the regional network, power shortage and water spill are caused in different provincial power grids. A short-term optimal operation method for several power sources if proposed for the regional power grid. Based on the decomposition and coordination of large-scale system, this problem is decomposed into two subproblems, load distributition problem for each provincial power grid and coordination problem among several multiple grids. On the one hand, the provincial scheduling problem is addressed by resolving several operation sub-problems for each type of plants. It should be noted that each sub-problem is sovled independently. On the other hand, reasonable acceptant and delivered generation schedules are generated based on the results obtained from each provincial operation problem. The proposed method is implemented on the case study called China Southern Power Grid (including five provincial power grids with 188 plants). The sactisfactory optimal solutions are obtained from this method, avoiding the power shortage and water spill of provincial power grids.Finally, a summary is made and a few issues which need to be further studied are listed.