Industrial 35t / H Circulating Fluidized Bed Boiler Combustion System Modeling And Simulation Research

At present in order to solve the burning low-grade coal and low-cost areas over the issue of pollution and meet the demand for central heating, a large number of small and medium-sized thermal power plants which have circulating fluidized bed boilers have been completed and will be the construction. As the circulating fluidized bed boiler combustion efficiency is generally higher than ordinary chain furnace, wide adaptability of coal, good environmental benefits, and load regulation than large and difficult to fire for the steam, so much so that by the users. In China, control of circulating fluidized bed boiler while the introduction of DCS control system, but the combustion process because of their complex, multi-factors, making access to justice on the combustion system of "self-control." In such cases, this article focuses on the combustion system identification model and simulation, advanced control strategies for the study provided a reliable basis.In this paper, the characteristics of circulating fluidized bed boiler, the combustion process and control objectives of the mechanism analysis of the boiler combustion system is a multi-variable, large time delay, strong coupling system, the input for a given volume of coal, primary air volume, and secondary Feng-volume output of the main steam pressure, bed temperature.Circulating fluidized bed boiler as a result of the combustion process is more complex, ever-changing characteristics of the object at the same time, the mechanism has more difficulties in modeling, but also difficult to obtain a good control effect; through the system so we can identify ways to create a cycle of fluid bed boiler dynamic model, the establishment of an amendment to expand the "black box" model. And artificial neural network approach any nonlinear function with the ability to multi-input multi-output nonlinear system provides a common modeling method. We are concerned about the model as long as the input variables and output variables relationship. Neural network is based on input data and output data, the right to keep the value of the network be modified to achieve from the input parameters to the output parameters of the relationship between non-linear mapping.Therefore, Kunming Jinggu paper mills in accordance with the actual production of the accounting data to Toolbox for MATLAB neural network tool, focusing on the establishment of the boiler combustion system of the dynamic BP network model and simulate the actual production data. If using a static model of BP neural network, the greater the change in condition occurs, it can not accurately reflect the relationship between input and output. In that case, the output in order to be able to make better predictions, it is necessary for the introduction of dynamic BP network characteristics. Usual practice is the introduction of data for historical information, and the combined effect of the current input to the network input, the actual system in order to make accurate forecasts. The simulation results show that the model has higher accuracy and expansion.