Multi-core Parallel Algorithm For Long-term Optimization Of Cascade Hydropower System

These years, with the rapid development of hydropower construction in China, the number of giant hydropower station is on the rise. Simultaneously, the calculation amount for optimal operation of hydropower system is increasing. As a result, the tradition algorithms are unable to solve the complex optimal operation of hydropower system. So it's important to explore new methods and ways to improve the efficiency and quality of system. In recent years, multi-core processors are widely used in projects and constructions because of its excellent properties in computing. However, the solving approaches of serial optimal operation of hydropower system actually based on single thread mode so that computational resources have not been fully used. It's leading to one cause of the complex to solve engineering problems, on the other hand the existing parallel resources is wasted. So how to bring out the parallel algorithm of optimal operation of hydropower system combine with the technology of multi-threaded and parallel algorithm has become the problem that needs to be solved urgently. This paper studies the issues as follows:(1) According to the curse of dimensionality of tradition dynamic programming in optimal operation of reservoir, the author proposes the parallel dynamic programming using in multi-core to improve its compute performance. The results of implementation of parallel dynamic programming show that it can give reasonable and efficient solution and reduce greatly calculating time to the original algorithm, and it provide a basis for the next study in paralleling algorithm of optimal operation of cascade reservoir.(2) A fine-grained parallel discrete differentiation and dynamic programming algorithm (PDDDP) for long-term optimization of cascade hydropower station is proposed which based on the analysis of normal discrete differentiation and dynamic programming algorithm (DDDP) and implemented in multi-core calculation environments. The proposed algorithm has been practically tested for the hydropower system of Lancangjiang Basin. Results demonstrate the PDDDP algorithm is easily implemented and greatly improve solving efficiency due to making full use of parallel resources, showing its efficiency and practicability for solving optimal operation of large-scale cascade hydropower station.(3) A new algorithm named parallel progressive optimality algorithm (PPOA) is proposed which combines with the problem in optimal operation of cascade reservoir in Lancangjiang Basin. It's an improvement to tradition progressive optimality algorithm. The result shows that improving algorithm can not only give reasonable and efficient solutions to increase the power generation benefit of power grid, but also is a feasible pathway for full utilization of the hydraulic power.Finally, a summary is given and some problems to be further studied are discussed.