Optimization For Flash Vaporization Process In Chaotic Particle Swarm Algorithm

As a swarm intelligence algorithm, particle swarm optimization (PSO) algorithm has provided a new way to find the solution of complex problems. PSO has been applied efficiently, but as a new and developing intelligent algorithm, PSO is still far from mature on systematization and standardization theory and application extending, also has some disadvantages, such as easily trapped into local optima and bad local search ability. To overcome these disadvantages, the following works have been done:The thesis summarizes chemical process optimization's necessity and the problems faced. The theory and application of intelligence algorithm as genetic algorithms, particle swarm optimization, ant colony optimization and simulation annealing are described. It puts emphasis on the research actuality and analyzes the problems faced of PSO. The principle and algorithm flow of PSO is introduced.In order to overcome the disadvantage of easily trapped into local optima, a modified method of adaptive hybrid chaos particle swarm optimization (HCPSO) algorithm based on chaotic search and N-M Simplex is introduced. During the iterative process, according to the variance of the population's fitness, the chaotic update of the particle is performed adaptively. The modified algorithm makes use of the stochastic property and ergodicity of chaotic search. When the fitness value is trapped into local optima, chaotic search will help the particle get out of local optima and N-M Simplex will enhance the ability of local search.HCPSO is applied to the benzene-toluene flash vaporization process optimization. On the based of building flash vaporization, adiabatic flash vaporization and the benzene-toluene flash vaporization process mass transfer mechanism model, HCPSO and CPSO algorithm is used to solve the benzene-toluene flash vaporization process optimization. By comparison with HCPSO and CPSO algorithm, HCPSO algorithm has higher optimization efficiency, better global performance, and more stable optimization outcomes, accorded with mass transfer mechanism model.