Research On Particle Swarm Optimization Algorithm And Some Issues About Power Stations

High energy consumption and serious pollution are important issues in the development of the power plant. Reducing coal consumption and pollutants while meeting the security standards and the demand load is attracting more and more attention. The optimization method based on modem intelligent algorithms is considered to be an effective method to solve these issues. Therefore more and more electrically technical workers focus their research on the optimization technology of power plants.On the basis of the development course and the present state of particle swarm optimization(PSO), economic load dispatch and NOx reduction optimization, particle swarm optimization and some issues of power plants were emphatically studied. And some useful conclusions to provide reference for the operation of power plants have been achieved. The main contents can be described as follow:(1) To overcome the premature problem and convergence problem, orthogonal design study strategy was introduced to design a new particle by extracting valuable information from the optimum and suboptimum particles. Then simulated annealing method was introduced to strengthen the ability of overcoming premature by searching around the best particle stochastically. At last, the improved particle swarm optimization algorithm was tested on several benchmark functions. Results showed that the improved particle swarm optimization algorithm outperformed other algorithms in the aspects of mean values, standard deviation, function evaluations, success rate and success performance.(2) A new function was proposed to map diversity to inertia weight for the first time to overcome the premature problem. Then the improvements of particle swarm optimization provided in this study were integrated and the convergence ability of the new algorithm was proved theoretically based on stability criterion for linear discrete systems and probability theory. T test and Wilcoxon test showed that the improved particle swarm optimization performed better after the improvements were done. The improved PSO outperformed some state-of-the-art PSOs and other intelligent algorithms.(3) The parameters of support vector machine were optimized by the improved particle swarm optimization algorithm. To deal with large samples, an ensemble support vector machine was introduced and improved by the parameter selection method proposed in this paper. Some operation data related to NOx emission and boiler efficiency were got from the thermal test of some 600MW unit. At last, NOx emission predicting model based on improved support vector machine was constructed. Results showed that the improved support vector machine has great generalization ability and predicting ability.(4) To handle large systems with many units and prohibited zones problem better, the improved particle swarm optimization algorithm was introduced to economic load dispatch. Studied cases showed that the improved algorithm can give better dispatch scheme. Cycle method was proposed to calculate carbon content of fly ash and the coal components calculation method based on real data was improved. Then the coal components were calculated based on the operation data. A new prediction model of coal consumption was constructed based on the improved ensembled support vector machine considering coal components, environment temperature, the aging of equipment and operation level. The performance was verified by corresponding cases. At last, the demand load was dispatched based on the new prediction model for some plant. Results showed that the new prediction model was feasible and effective.(5) The NOx optimization model was studied and improved to cut down NOx emission, which can handle the coal component changes and meet the demand load at the same time. Studied cases showed that the new optimization method was feasible and reliable. The improved particle swarm optimization was employed to cut down NOx emission based on the new optimization model. Compared with other state-of-the-art algorithms, the improved optimization algorithm was more effective and robust. The optimized operational parameters were consistent with the physical nature of achieving low NOx emission.