Two Stage Particle Swarm Optimisation for Long-term Operation of a Hydroelectric Power System
Cheng Chuntian,Liao Shengli,Wu Xinyu,K W Chau
DOI: https://doi.org/10.1080/1023697x.2009.10668161
2009-01-01
Transactions
Abstract:The problem of determining the optimal long-term operation of a hydroelectric power system has been the subject of numerous publications over the past sixty years. A major problem encountered in operating long-term hydroelectric power system is their dimensionality. A great effort to decrease or eliminate possibility of dimensionality problem is addressed through developing innovative optimization techniques, such as genetic algorithms, artificial neural networks, and so on. Particle swarm optimisation (PSO), a newly developed evolutionary technique, is a population based stochastic search technique with reduced memory requirement, computationally effective and easily implemented compared to other evolutionary algorithm. However, there exist some difficulties in applying PSO to hydropower system. Constrained by complex constraints and hydraulic relationships between upper and lower reservoirs, it is unfeasible to use stochastic search algorithms of PSO directly for most initial populations. In this paper, a two stage PSO algorithm is presented to solve the optimal long-term operation of a hydroelectric power system. The maximisation of electricity generation and maximisation of minimal mean power of the hydropower system are alternatively used as the objective of longterm planning of hydroelectric power for the two stage problem. The maximisation of minimal mean power of the hydropower system is chosen as the objective at the first stage and an initial feasible solution will be generated using PSO. The system objective, ie the maximisation of electricity generation is selected as the objective at the second stage and the optimal result of the first stage will be used as the initial feasible solution. The proposed method is implemented to the optimal long-term operation of a hydroelectric power system in the Yunnan Power Grid which is located in the Yunnan Province of China and consists of 17 dominated hydropower plants with an installed capacity of 3,942.5 MW. The results show that the two stage PSO can give reasonable and efficient solution and that applying PSO to the long-term operation of a hydroelectric power system is feasible.