Research On Modeling And Optimization For Combustion System Of Large Capacity Utility Boilers Based On Intelligence Computation

In a long period of time, the thermal power will be the main force of power industry in our country. And the thermal power units will always be the key field of energy saving and emission reduction. Compared with traditional optimization methods, the optimization method for combustion system based on intelligence computation has unique advantages. Especially in the case of increasing unit capacity and parameters and thermal process becaming more and more complicated, doing the research on modeling and optimization for combustion system of large capacity utility boilers based on intelligence computation will offer an effective solution for the operation of generating unit with high efficiency and low emission.This dissertation focuses on the intelligent methods of modeling and optimization for optimal operation of large capacity utility boilers, the main content includes:(1) Research on the modeling and optimization of coal blending system. The significance of coal ash fusion temperature for the coal blending system is described. The modeling frame and performance indexes of coal ash fusion temperature prediction model is given. The performance of different models are compared and the optimization is launched on the better model.(2) Research on the optimization of combustion system for a330MW tangentially fired boiler. The influencing factors of thermal efficiency and NOx concentration are discussed. A model for the combustion system is given based on the thermal samples with the One-Factor-at-a-Time plan under full load. The model structure is described and the effectiveness is verified. The optimal solution is obtained by the optimization searching for adjustable operation parameters.(3) Research on the optimization of combustion system for a1000MW ultra supercritical boiler. The field combustion adjustment experiment is launched with the adjustment of load, fuel, burners angle, over flow air, secondary air, burners compound mode and etc. Based on the above thermal data a sharing model is set up. And the satisfactory prediction results for thermal effciency and NOx concentration are obtained. The optimal scheme is obtained by the optimization searching and the rationality of the results is analysed.(4) Research on the intelligent modeling and intelligent optimization methods. A regression method suitable for the combustion system with multiple outputs is proposed based on the study on the support vector regression. The model parameters are optimized so that the prediction performance is improved. Some research is launched on two swarm intelligence methods. An ant colony method for continuous functions optimization is proposed based on the research about basic ant colony algorithm and the algorithm frame is given. Some improvements are made on particle swarm optimization method to accelerate the convergence, enhance the diversification of the population and strengthen the local search ability.Some innovative achievements are obtained and they are:(1) An optimization scheme for the combustion system of a330MW tangentially fired boiler is proposed.(2) An optimization scheme for the combustion system of a1000MW ultra supercritical boiler is proposed.(3) A least squares support vector regression method is proposed that is suitable for the combustion system modeling. An ant colony optimization method suitable for continuous function and an improved particle swarm optimization method are proposed.