Improved Binary Particle Swarm Optimization And Its Application In The AGC Of Cascade Hydropower Stations

With the continuous development of hydropower energy construction, the main cascade hydropower stations have been formed. In order to play the important role in the flood control, navigation, power generation, irrigation and other huge social and economic benefits, joint operation of cascade hydropower stations and the optimal operation has become a hot academic and engineering research. Automatic Generation Control (AGC), as a joint operation of cascade hydropower stations, and an important part for optimal scheduling, has been widespread attention by researchers.In recent years, with the increasing installed capacity of hydropower and the increasing number of units, the former optimization methods of AGC for small and medium hydropower stations have been unable to meet the actual requirements of large-scale hydropower projects. In solving high-dimensional, multi-objective, nonlinear, strong coupling and discrete optimal operation of cascade hydropower stations combined problems, the traditional algorithms such as the equal incremental algorithm, dynamic programming algorithm, genetic algorithms are limited, such as: long calculating time, large amount of the required memory capacity, algorithms curse of dimensionality and local convergence of the algorithm. Binary particle swarm algorithm with wide application in solving discrete combinatorial optimization problems has increasingly evident advantages. In order to solve the unit commitment, load distribution optimization problems of large-scale cascade hydropower stations AGC, this paper, based on binary particle swarm optimization algorithm and the introduction of dynamic inertia weight factor, puts forward an improved binary particle swarm optimization, which has made some theoretical and practical value of the results. Some research results have successfully applied to the decision support system for the joint optimization of the Three Gorges cascade hydropower station. The main research work and innovations are as follows:(1) As to the unit commitment and load distribution optimization problems in high dimensional, complex nonlinear characteristics and the security, economic, practical requirements of cascade hydropower stations, an improved binary particle swarm algorithm is proposed. This algorithm can not only avoid the curse of dimensionality, but also overcome the initial iteration of local search and global search ability of the late iterative weak shortcomings, and the introduction of a new species update mechanism helps maintain the population diversity, while improving the algorithm global optimization capabilities. This algorithm is applied to the Three Gorges cascade hydropower stations AGC instance, results show that the algorithm reflects a good performance: short computational time and high accuracy. It provides an effective means for hydropower optimization problem especially for the large-scale hydropower cascade.(2)In the solution process of the mathematical model of AGC, besides considering the power balance constraint, reserve capacity constraint, the unit output constraint and the unit opened down time constraints, but also give full consideration to the unit maintenance scheduling and unit cavitation vibration area. The optimal allocation to meet the load and to avoid frequent opening stop unit can not only ensure that all stations within the unit head in the design of stable operation and significantly reduce wear on the unit and other damage factors, by avoiding the scheduling of the power units are often separated from the low efficiency in the operation area, but also can avoid the vibration of the joint area between stations, which significantly increase the output of the plant.(3)In the past, to the adjustable hydropower station, in the optimal load distribution calculation, the selected power head is generally the initial period of the power head, and heads of the follow-up period remain the same generation, it is called the fixed optimal load distribution head. For run-of-station, the incoming water running in the station is changing all the time, so the traditional fixed-head calculation of optimal load distribution is no longer applicable, so this paper puts forward variable head innovation of load allocation method, which is not only for Run-of-station, but also for the adjustable hydropower stations. This method provides an effective solution for establishing hydropower generating plan.