Study Of Modeling And Optimization For Combustion Process Of The Boiler In Power Plants Based On Neural Networks

With increasingly severe situation of energy and environment, coal-fired power plant is energy-hungry and faces growing pressure on energy-saving and emission reduction. Boiler, the main energy-consuming equipment of power plant, is attracting more and more attention. The existing boiler combustion control technology of power plant has increasingly apparent shortcomings, and has become more and more difficult to meet the actual needs. So high-efficiency and low-emission combustion optimization technology is increasingly becoming the focus of the study.At present, there are two methods to improve combustion, which are traditional method and intelligent control method. Boiler combustion system is a complex large system with nonlinearity and strong coupling. The traditional method only takes single factors into account, not fully reflecting the true state of the system. However, the intelligent control is very suitable for modeling and optimization of nonlinear system with strong coupling, so it has more and more attention.In order to obtain high-efficiency and low-pollution combustion method, a model of the boiler combustion system is established with output variables of boiler efficiency and NO_x emission, and genetic algorithm is used to optimize the system.Modeling of combustion system is achieved by two kinds of neural networks. They are BP network and the dynamic fuzzy neural network (one kind of RBF network). The two kinds of modeling methods and their respective applications are analyzed, namely that online model in industrial field can be established by RBF network which is faster and better-performing in non-linear function approximation, while the off-line model with multi-dimensional input data can be achieved by BP network with stronger flourishing ability. Optimization system calculates optimal values of two important parameters, which are fuel volume and air supply. The comparison between value of optimization system and outputs of original system can be seen apparently in the graphs.In order to put the modeling and optimization system of combustion process into practice to guide the operator, C++ langue is used to program, and OPC interface is set to connect the program and DCS. Then the research can be applied in the practice.