Font Size: a A A

Particle Swarm Optimization Method Based On Complex Network Problems In Unit Commitment Optimization Research

Posted on:2013-01-29Degree:MasterType:Thesis
Country:ChinaCandidate:J R TangFull Text:PDF
GTID:2218330374965509Subject:Power system and its automation
Abstract/Summary:PDF Full Text Request
Unit commitment is an important part of optimal operation of power system. Since a reasonable unit combination scheme can reduce the coal loss, extending the use time of the unit and bring in significant economic benefits, so it continue to be the main optimization task in plan of short-term operation. In the view of mathematic, unit commitment is a large-scale, discrete, multi-constrained and non-linear complex engineering optimization problems, it is difficult to get the optimal solution in theory, so far, there is still not exist a practical algorithm which can not only close to the realistic system model completely but also achieve an ideal computational accuracy and speed. Therefor, how to build a crew mathematical model which can be close to the actual system and how to improve the accuracy and speed of solving the unit commitment problem has an important significance to the optimal operation of power system.Particle swarm optimization is an intelligence optimization algorithm which is studied by the predatory behavior of bird populations besides the ant colony algorithm, the shuffled frog leaping algorithm and the fish algorithm. It is improved to solve the problem which is easy to appear the "premature" phenomenon so that the algorithm into a local optimal value of standard particle swarm optimization in this paper. First, the dynamic of acceleration factor and inertia weight is used as a new parameter adaptive strategy to improve the accuracy of convergence of the particle swarm optimization; Second, the two models of the complex network of small-world networks and scale-free network are introduced into the neighborhood structure of the particle swarm optimization, then formed the new algorithm of particle swarm optimization based on the small-world network model and scale-free network respectively, they speed up the rate of algorithm convergence.In this paper, we mainly focused on the thermal power plant unit commitment optimization. A mathematical model is presented and consuming minimization of total coal is chosen for its objective function with considering the load balancing constraint and so on. Then designed the detailed framework for the two improved algorithm of particle swarm optimization, and implement it by the MATLAB programming, used in ten-machine system simulation respectively, through the analysis and comparison of numerical results show that the improved particle swarm optimization which based on the complex network model can achieve a better optimal solution, both of which do not exist in the case of the curse of dimensionality. In comparison, the PSO algorithm with small-world network has a better optimization ability, the PSO algorithm with scale-free network has a higher convergence rate, they has the feasibility to solve the unit commitment optimization effectiveness.
Keywords/Search Tags:Unit Commitment, Complex network, Particle Swarm Optimization, Small world network model, Scale-Free network model, Neighborhood structure
PDF Full Text Request
Related items